HARVARD UNIVERSITY

Library of the

Museum of

Comparative Zoology

B R E V I O R^K'''

UNlVFRc;|yv

MmseiLmi of Comparative Zoology

Cambridge, Mass. l 5 September, 1969 Number 326

ANOLIS INSOLITUS, A NEW DWARF ANGLE

OF ZOOGEOGRAPHIC IMPORTANCE FROM THE

MOUNTAINS OF THE DOMINICAN REPUBLIC

Ernest E. Williams

and
A. Stanley Rand '

Abstract: Anolis insolitus. a new dwarf species occurring in the Cordil-
lera Central of the Dominican Republic, is closely related to both A. oc-
cidtiis known from several montane localities in Puerto Rico and A.
darlingtoni of the Massif de la Hotte in Haiti, but is closer to A. darlingtoni,
a much larger species. The three species are basal members of the carolin-
ensis group (sensu Into) in the West Indies, of which A. occultiis is the
most primitive known member.

INTRODUCTION

In the spring of 1963, the Museum of Comparative Zoology re-
ceived from the Cordillera Central of the Dominican Republic a
single small anole that was immediately recognized by E. E. Wil-
liams and by James D. Lazell, Jr., as a very distinctive and im-
portant new species. Even at that time the name insolitus —
"strange or unusual" - — was chosen for it.

Because, however, the new species was represented by a single
specimen, its description was delayed, to wait upon more material.
An attempt by Lazell, during the last week of December of 1963
and the first week of January 1964, to collect at the exact locality
where the first specimen was taken failed because of bad weather.

Not until late summer of 1968 did E. E. Williams and A. S.
Rand succeed in visiting the pertinent locality and in collecting a
small series that fully confirms the "unusual" nature of the animal.

^Smithsonian Tropical Research Institute, Balboa, Canal Zone.

BREVIORA

No. 326

which also turns out to be of considerable phyletic and zoogeo-
graphic significance, its external indications of relationship veri-
fied by an osteological investigation made by Richard Etheridge
(pers. comm.) :

Anolis iNSOLiTUS n. sp.

Type. MCZ 60144, Paraje La Palma, Seccion La Palma, Municipio Con-
stanza, Provincia La Vega, Repiiblica Dominicana, C. E. Ray and R. R.
AUen coll. 19 March 1963,

Paratypes. (Same locality as type) MCZ 107014-18, A. S. Rand and
E. E. Williams coll. 30-31 July 1968. (MCZ 107015 skeletonized, MCZ
107017-18 used unsuccessfully for chromosome study.)

Diagnosis. A dwarf anole related on the one hand to A . darling-
toni Cochran of southwest Haiti, from which it differs in size and
in several features related to size, e.g. lamellae under fourth toe),
and on the other to A. occultus of the mountains of Puerto Rico
(which it resembles in size but from which it differs especially in

Fig. 1. Anolis insolitus, Type, MCZ 60144: dorsal view of head.

1969 ANOLIS INSOLITUS 3

the much larger head scales). The new species differs from all
known species of Anolis in the presence of small but distinct
postorbital, supratemporal, and occipital spines, bony in nature
and especially prominent in males (Fig. 1).

Description. Head: Narrow, elongate. Head scales large,
smooth, smallest at tip of snout, three scales across snout between
large second canthals. Nostril oval, nasal scale separated from
rostral by a single oval scale. Rostral scale wide, low, in contact
with five scales posteriorly.

Supraorbital semicircles large, weakly convex, rugose laterally,
separated by a single row of scales as large or larger. A much
less distinct row of seven large oval granules or scales on each
side at the supraciliary margin, no elongate supraciliary. Posterior
and internal to the supraciliary row, some smaller granules or
scales. A single scale raised into a spine just beyond these smaller
granules at the posterior end of each supraorbital semicircle. An
elongate supraocular disk of ca. six to nine enlarged scales, two of
them about one-third to one-half the size of the scales of the semi-
circles. Canthal ridge of six scales well defined, second canthal
scale largest, diminishing in size anteriorly, anteriormost below
nostril. Loreal rows three with some irregularity in size. A dis-
tinct supratemporal line of four to five enlarged scales, the fourth
replaced by a spine. Temporal scales small, smallest at center,
flat. Supratemporal scales above supratemporal line becoming
larger toward a ridge of protuberant — almost spiny in <J —
scales forming a U-shaped crest behind the interparietal region,
in S with a larger spine at base of U. Interparietal ovoid, much
larger than the ear opening, separated by one flat scale on each
side from the supraorbital semicircles. Scales surrounding inter-
parietal large, flat, with some tubercles in <^ (tubercles also on
surrounding scales in 5 ). Ear small, subround, placed far ven-
trally, directly behind the comissure of the mouth.

Suboculars in contact with supralabials, anteriorly grading into
loreals, posteriorly grading into temporals. Seven supralabials to
center of eye.

Mental large, semidivided, wider than deep, in contact with
four granules between the infralabials. Two large infralabials on
each side in contact with sublabials. Throat scales granular,
smooth.

Trunk. Dorsal scales granular, smooth, subequal on flanks and
middorsum except for a crest of small, triangular, swollen scales
continuing middorsally from the U-shaped crest behind the inter-
parietal to a point a little behind the insertion of the forelimbs.

4 BREVIORA No. 326

Ventrals larger than dorsals, smooth, round, in transverse rows.

Gular Jan. Moderately large; present in both sexes and well
developed even in juveniles, lateral margins slightly inset, scales
granular, smaller than throat scales, much smaller than ventrals;
lateral scales about as large as edge scales in well-separated rows
( $ ) or less well-defined rows { $).

Limbs and digits. Limbs short, tibial length ca = distance tip of
snout to middle of eye. Fifteen to sixteen lamellae under phalanges
ii and iii of fourth toe. Scales of limbs smooth, those of anterior
thigh larger than those of ventrals. Supradigital scales smooth.

Tail. Round, with a distinct dorsal crest of a median row of
enlarged, keeled scales, interrupted at intervals of two to four scales
by paired paramedial scales, usually the most distal scale in any
small series largest but with some irregularities. No enlarged
postanals, but scales nearest vent larger in i than 9 . Scales be-
hind vent and round base of tail smooth, grading into keeled scales
distally. Four ventral rows distinctly enlarged.

Color in life. The general body color is greenish or grey-brown,
mottled, lichenate, with the dewlap, present in both sexes, blue-
grey in front, orange behind. Detailed notes on two specimens
follow: (1) 9 . Dull green with a dark grey middorsal zone en-
closing a series of dark grey spots. Traces of a sacral butterfly pat-
tern overlying a dull orange sacral spot. Blurred barring on tail,
barring hardly visible on limbs. A light yellow streak under eye;
reddish color on upper eyelid. Faint indication of a light streak
from ear to arm. Flanks mottled. Dewlap blue-grey anteriorly,
dull orange posteriorly, crossed by rows of white scales. (2) i .
Dark butterfly-shaped blotches dorsally, less distinct on sides. One
such blotch above shoulders, two on back, and one on sacrum.
The sacral blotch crossed by a light orange spot. Tail and limbs
crossbarred. A light yellow streak under eye; skin around eye
reddish. A curved yellow streak from above ear to above base
of arm. Sides and belly lightish cream, lightly speckled laterally
with brown. Dewlap as above.

Field observations. (Compare with observations on A. occultus
by Webster, 1969). Four of five specimens were caught asleep.
One adult was taken about six feet from the ground on a broad,
nearly horizontal leaf of a bush, its head facing toward the stem.
Two other adults were found eight to ten feet apart, sleeping along
slender, nearly horizontal twigs of bushes. One was about four
and the other about five feet above the ground. Two juveniles
were found at the edge of the forest about 15 feet apart, sleeping

1969

ANOLIS INSOLITUS

o

N
U

to"

3

6 BREVIORA No. 326

along the very slender terminal twigs of bushes; one was about
six and the other eight feet up.

The single adult seen during the day was eight feet up on a
horizontal, quarter-inch branch of a small tree. It was in a rather
open part of the forest on a sunny morning. It moved in and out
of sun flecks, apparently paying little attention to them in the few
minutes that it was watched. At first it pressed itself against the
branch and squirrelled to the other side. Its mottled pattern
matched the lichen-covered branch, and it was hard to see. Soon
it resumed a more erect posture and slowly climbed along one
branch and out on one even more slender. When an attempt was
made to catch it, it ran along this branch and jumped to another
a few inches away. When captured, it threatened with open mouth
and erected the blue-grey and orange dewlap.

Behavior (A. S. Rand). On the morning after capture, the fol-
lowing observations on the locomotion and defensive behavior of
two specimens were made : on the ground or along a perch, move-
ment was either a slow, very deliberate walk or a series of rapid
hops. Jumps from perch to perch were made willingly and accu-
rately, and were of considerable distances for a two-inch lizard.
One jump measured about 12 inches with little loss in height; an-
other jump was six inches, and the lizard landed on a perch higher
than that from which it started.

A lizard climbs willingly and without difficulty on small twigs
but seems more at home on twigs of about body diameter than
on those much smaller. When it is approached by a hand, a lizard
will squirrel slowly around its perch, even a slender one, pressing
its whole length against the substrate. On a horizontal branch the
lizard may go completely underneath. Very closely approached, it
may jump away or to the ground. When seized, the mouth is
opened threateningly, and there is an attempt to bite.

Except when hiding, the head and shoulders are raised away
from the substrate, and the neck is straight, whereas the vent is
almost, but not quite, in contact with the substrate.

In walking, the toes of the fore feet are spread widely so that
toes 1 and 5 point in approximately opposite directions. On very
slender perches the toes oppose one another: 1, 2 versus 3, 4, 5 or
1. 2, 3 versus 4, 5.

On flat surfaces, the toes of the hind feet spread so that toe 4
extends the axis of the foot, toe 5 is at ri^^ht angles to this behind,
toe 1 at right angles in front, and 2 and 3 between. On slender
perches, sometimes toe 5 opposes the others, sometimes 1 and 5

1969 ANOLIS INSOLITUS 7

together oppose 2, 3, 4. The tail is frequently carried in an upright
curve, but is more often straight and rests against the substrate.
While climbing, the tail is frequently used as a sliding hook. Usu-
ally at about half way along its length, the tail hooks in a semi-
circle over and behind some projection. As the lizard moves for-
ward, the hook slides backwards along the tail until the tip of the
tail reaches the projection, crosses it, and drops off. The tail may
be used on a straight branch without projections by being bent
to one side and around the branch. If the branch is shaken, the
tail may strengthen its hold by forming a complete loop around
the branch. Apparently only the very tip is flexible enough to grip
a small branch tightly.

Relationships. Anolis insolitus is almost as distinct in Hispan-
iola as A. occultus is in Puerto Rico (Williams, Rivero, and
Thomas, 1965) : it is quite impossible to confuse it with any other
Hispaniolan species. However, it does have resemblances in two
directions, with A. darlingtoni of western Hispaniola and with
A. occultus of the mountains of Puerto Rico. These resemblances
indicate that it is an annectant rather than an isolated form.

It is best placed in its proper group on osteological characters.
Table 1 records the pertinent comparisons (information provided
by Richard Etheridge — pers. comm. — from X-rays of occultus,
insolitus, and darlingtoni and from dry skeletons of occultus and
insolitus). According to the informal groupings suggested by
Etheridge (1960), A. occultus em.erges as a very primitive (and
somewhat aberrant) member of the carolinensis group (those
with "T-shaped" interclavicles; compare Etheridge's 1960, fig. 4)
of alpha anoles (those without caudal transverse processes) . The
primitiveness of occultus is manifested by a high number of at-
tached inscriptional ribs ("parasternals" of Etheridge 1960, but
see Etheridge 1965). A. darlingtoni is a somewhat less primitive
member of the same group (fewer attached inscriptional ribs),
but is peculiar in the specialized character of non-autotomic caudal
vertebrae. A insolitus is again a member of the same group, but,
like A. darlingtoni, is specialized in having non-autotomic caudals;
it is, however, more advanced than A. darlingtoni in having only
three attached inscriptional ribs and one free one. (Most of
the carolinensis group show the latter condition; a few are still more
advanced and have only two attached inscriptional ribs and two
free.)

A. insolitus has pterygoid teeth. These are absent in A. occultus,
and we lack information about A . darlingtoni.

8

BREVIORA

No. 326

The occurrence of pterygoid teeth (primitive for lizards in gen-
eral) in Anolis is somewhat erratic. They are usually absent in
dwarf species. At least in the West Indies, however, they are
frequently present in the more primitive members of any group.
In the carolinensis group, their presence can be verified in A.
chlorocyanus, A. coelestinus, A. aliniger, A. equestris, A. allisoni,
and A. carolinensis among the more primitive species, and in A.
lucius among more specialized forms. The presence of pterygoid
teeth in A. insolitus is presumably to be regarded as primitive and
is so recorded in Table 1 .

A. darlingtoni and A. insolitus are unique among Hispaniolan
anoles in having non-autotomic caudals, and, indeed, are the only
West Indian members of the genus Anolis that lack tail autotomy.
(Chamaeleolis and Chamaelinorops, the only other West Indian
anoles to lack tail autotomy, are very distinct genera. ) Though loss
of autotomy has occurred several times in anoles, its occurrence in
the West Indies only in two species on one island suggests affinity;
it does not, of course, demonstrate it.

The enlarged plate-like head scales of A. insolitus and A. darl-
ingtoni (compare Fig. 3 and Fig 2) provide the most obvious
external resemblance between the two species. This character,
however, is not unique to these species even within the West Indies.
In fact, A . darlingtoni and A . valencienni were formerly united in
the genus Xiphocercus solely on the basis of similar large plate-like
head scales. A. darlingtoni and valencienni, however, belong to

Fig. 3. Anolis darlingtoni, MCZ 38251 : dorsal view of head.

1969 ANOLIS INSOLITUS 9

two different sections of the genus Anolis as analyzed by Etheridge
(1960), alpha and beta respectively, and the character of plate-
like head scales is now recognized as having arisen many times
within the anoles. Again, therefore, this resemblance is not proof
of affinity between darlingtoni and insoUtus, but the presence of so
special a character in two species of one island is suggestive.

It would be a possible argument against the close affinity of the
two species that darlingtoni (72 mm) is approximately twice the
snout-vent length of insolitus (33 mm). Differences in size be-
tween closely related species, particularly if they are sympatric,
are not unusual, but as far as is known, these two species are
widely allopatric, and the size difference is extreme.

However, a number of external characters suggest not only an
affinity between darlingtoni and insolitus but also suggest relation-
ships with occultus: ( 1 ) the simple annular nasal scale separated
by one round scale from the rostral (=prenasal not fused with
nasal); (2) the small, round, ventrally-placed ear; (3) the long
head and short limbs; (4) the low number of loreal rows; (5) the
smooth ventrals.

Table 2 lists not only those characters in which all three species
are similar but also all other pertinent external characters. The
significant resemblances between any two or among all three
species are italicized. It is easily seen that insolitus occupies a key
position. In many critical and sometimes quite special characters
(e.g. absence of an elongate supraciliary scale; presence of a slotted
gular fan in both sexes), insolitus resembles sometimes darlingtoni,
sometimes occultus.

It is clear, of course, that insolitus is not just an intermediate
between the two species. It has very striking specializations of its
own. The small spines and rugosites on the head are the most
remarkable feature; these are as visible on the skull as they are
externally. The crest of enlarged scales on the nape and the
peculiar tail crest are almost as singular. In another regard,
insolitus is not intermediate; the inscriptional ribs show a condi-
tion more advanced than that shown by either occultus or darling-
toni. However, such a complex of primitive and advanced adapta-
tions is just what we should expect of the surviving representative
of the stock that was at one time intermediate between the ancestral
grade now represented by occultus and the more advanced grades
represented by darlingtoni and by other still more advanced mem-
bers of the carolinensis group. The importance of insolitus as an
annectant form phyletically and zoogeographically is not diminished
by admission of its specializations.

10 BREVIORA No. 326

Figure 6 shows the known locaHties for A. darlingtoni (still
known only from the unique type from Roche Croix, Massif de
la Selle, Haiti) and A. insolitiis (known now from six specimens
from La Palma in the Cordillera Central of the Dominican Re-
public). Such a map reflects more ignorance than knowledge.
Though Hispaniola has recently been assiduously collected, it is
obvious that the fund of new information and of new taxa is not
nearly exhausted, and the need for further collection and study is
abundantly clear. The genus Anolis is only one fraction — even
though an important one — of the herpetofauna of Hispaniola.
The current count of species (including insolitus) is 21. Of these,
no less than seven have been described in the last ten years {chris-
tophei Williams; koopmani Rand; cochranae Williams and Rand;
whitemani Williams; singularis Williams; rimarum Thomas and
Schwartz; insolitus Williams and Rand). In Table 3 we list the
known species with comment on degrees of distinctness and on
geographic variation (the latter may in some cases conceal valid
species). We confess to a lack of belief that the list is complete.
An asterisk marks those species that are especially inadequately
known.

Certainly the most plausible assumption based on current evi-
dence is that darlingtoni and insolitus are geographic representa-
tives — south island and north island respectively — of one stock.
This assumption, however, leaves the extreme size disparity of these
allopatric species without easy explanation. A discussion of this
point and of the possible history of Hispaniolan anoles is deferred
to a future paper.

A. darlingtoni has not previously been adequately figured.
Cochran ( 1941, pi. 11) provided only a photograph, which showed
little more than general shape. Figures 3 and 4 permit comparison
with the similar figures of A . insolitus.

Figure 5 diagrams the probable relationship of A . insolitus within
the carolinensis subsection of alpha Anolis. It and A. darlingtoni
appear to be the earliest radiation of this stock within Hispaniola.
Three further radiations have occurred within Hispaniola, one of
these, that of the Hispaniolan grass anoles, being the result of a
back invasion from the complex radiation of the carolinensis group
in Cuba (Williams, 1961).

ACKNOWLEDGMENTS

This study was pardy supported by National Science Foundation
grant GB 6944 to E. E. Williams. Figures were prepared by Joshua
Clark and by Laszlo Meszoely.

1969

ANOLIS INSOLITUS

11

'>

•4— »

oo
m

N
U

Co

12

BREVIORA

No. 326

I,

U5

s

w

<

5

OS

C3 "4-

t o

Q

o

BJ

s

1— 1

o

hJ

•4-t

o

<

^

1-^

3

o

rt

u-l

c

cu

o

CA)

c

CO

<

BO

s

c

1)

C3

i

SI

<

•a Oh

o
« <u

Q

u

PJ

s

o

hJ

■«-t

o

<

^

1-^

3

o

1

m

c

0.

o

C/5

c

3

Q

Q

Q

P-l

UJ

UJ

N

N

N

l-H

h^

HH

a

hJ

^

<

<

<

HH

r^

i-H

o

O

q

u

PJ

pj

Oh

o^

CU

00

c/o

(/5

^ — ^

■ — '

.^

^^

c

d

C

0)

tu

<u

W5

(/3

in

X)

JO

XI

C3

rt

oj

Q
PQ
N

<

O

PJ
a.

t/5

<u

a

XI

(/3

.. S

T3

x:
o

VO

(«

>

I/)

c

3
O

(«

o

(J

o

o

S

o

o

3

Oh

(U

*ca

f-H

_o

c

-o

O cfl

^

is

■g.

ipti
ron

3

c
ca

(U

tn

a>

inscr
chev

3
u

X)
(U

1969

ANOLIS INSOLITUS

13

TABLE 2
occultus insolitiis darlingtoni

34 mm snout-vent length 33 mm snout-vent length 72 mm snout-vent length

9-13 scales across snout 3 scales across snout

nasal scale separated
from rostral by a
single scale

supraorbital semicircles
weak, separated by
2-4 scales

No differentiated
supraciliaries

no postorbital, supra-
temporal, or occipital
spines

no distinct supra-
temporal line of
enlarged scales

nasal scale separated
from rostral by a
single scale

supraorbital semicircles
strong, separated by one
row of wide scales

A distinct supraciliary
row, but no scale
elongate

postorbital, supra-
temporal, and occipital
spines, especially
prominent in males

a distinct supratemporal
line of large scales
ending in a spine

5 scales across snout

nasal scale separated
from rostral by a
single scale

supraorbital semicircles
strong, separated by one
row of narrow scales

A distinct supraciliary
row, but no scale
elongate

no postorbital, supra-
temporal, or occipital
spines

no distinct supra-
temporal line of
enlarged scaled

Ear small, round, low Ear small, round, low Ear small, round, low

interparietal small,
round, ca = ear, sepa-
rated from semicircles
by 2-6 scales

canthal ridge weak,
barely differentiated

loreal rows 2-6

suboculars in contact
with supralabials

10-1 1 supralabials to
center of eye

mental in contact with
4 scales between sub-
labials, no differentiated
infralabials

interparietal ovoid, much
larger than ear, sepa-
rated from semicircles
by one large scale

interparietal ovoid, much
larger than ear, sepa-
rated from semicircles
by one large scale

canthal ridge strong canthal ridge strong

loreal rows 3

suboculars in contact
with supralabials

7 supralabials to
center of eye

mental in contact with
4 scales between well-
differentiated infra-
labials

loreal rows 3

suboculars in contact
with supralabials

7 supralabials to
center of eye

mental in contact with
2 scales between well-
differentiated infra-
labials

14

BREVIORA

No. 326

occultus

insolitiis

darlingtoni

middorsal scales smooth, a low crest of triangular scales on nape somewhat
flat, subequal enlarged scales on the smaller than middorsally

nape to a little past

insertion of arms

ventrals > dorsals,
smooth, juxtaposed
in transverse rows

giilar fan large, present
in both sexes: inset,
scales in rows in
females, not in row
in males

ventrals > dorsals,
smooth, juxtaposed
in transverse rows

gnlar fan moderately
large in both sexes,
inset, scales in distinct
rows in females, rows
less distinct in males

ventrals ca = dorsals,
smooth, subimbricate
in transverse rows

gular fan large, not inset,
scales evenly distributed
in males

limbs short, tibial length limbs short, tibial length limbs short, tibial length
ca = distance snout tip ca = distance snout tip less than distance snout
to center of eye to center of eye tip to center of eye

14-20 lamellae under
phalanges ii and iii
of 4th toe

scales of limbs smooth,
always smaller than
ventrals

supradigital scales
smooth

tail round without
dorsal crest

no enlarged postanal
scales in male

15-16 lamellae under
phalanges ii and iii
of 4th toe

scales of limbs smooth,
of anterior thigh
larger than ventrals

supradigital scales
smooth

tail round with
dorsal crest

no enlarged postanal
scales in male

24 lamellae under
phalanges ii and iii
of 4th toe

scales of limbs weakly
carinate, of anterior thigh
larger than ventrals

supradigital scales
multicarinate

tail round without
dorsal crest

enlarged postanal scales
in male

1969 ANOLIS INSOLITUS 15

Q

u

•2 u ^ — o

■2 ^ ^ .^ g- -S) M

S '" „> — p o

^ b -y -H 5- § §

t/) -C

1/3

WC/3(«(ua'L«3 3 CXC O-C

D-

3

O

C3

u

OX)

"o

c/o

d

•H

d

cu

D-

Ml

c«

• »-<

CO

X

B

UJ

— 1

o

c

<

1/2

(J

Oh

u

(/5

1—

"o

■**-

tn

CO

<

Q

_C

(U

43

H

«

(U O 43 -S

C I -2 ^ ^ S C g ^

tf £ ^ s I ^ g fe f I ': p. .

<u

sis- IS- :go5|^o III

X

x:

o

c

*-<

3
O

lU

U

o

c

o

>;^

o

>^

6

;s

a

iZ2

73

o

«

o

03

C

u.

(J

Lh

O

vw

_g

<4-l

_o

—

vi

o

c

c
o

•T3
C

J3

3

y

s

."'

g

^

;>

s

a

•S

^*»«

to

1^

"S

a

O

•~

■Q

i>

c 6 ;= g ;i: 73 -a

o

c

J5 := 2 oil "C u"

16

BREVIORA

No. 326

*- «

s

.0

s

o

c

o

c

c

3

c .a

O

B

C

p

u

X)

1

C

•a

u

c«

(U

^^

C

^

03

3

P3 ca (u — C

c c j: • - o

"O 9- =" —

>^ 3 u. (U iS

C - CI. c

O rt -O c^

3

D-

O

Cfl

o

•^J

c

o

■J

c

."'

>%

"5

T*^

u

o

e

E

o

C ~: o
=2 5 —
o .2 -a

N

c -i: .^

o

to

1^

S

E
o

T3

C
03

o

ca

X

c

00

Ui

to

£

3
O

\-

ca

■.-»

E
o

0

•S

x:

-*— »
o

X)

c

u
.I"

o

>

ca

O

•3

a

13

•4-t

1-

c
o

la

c

u

u

•S

•3

1)

03

E

E

o

w

E

E

.—

E

Q

ca

oo

.■^

03

."^

•5

73

o

<u

o-

_C

(L>

>^

wT

,"^

<u

C«

o

l«

ca

QJ

v^

>

\-

"3

u

O

"o

CJ

u

o

c

C

ca
44

o

^

^

ec

a.

o

o

_c

k.

ca

c

c

1^

ca
E

(U

"1

c

c

OJ

3

3

c«

u

•u

o

GO

_>,

O

1)
>

>.

o

3

"ca

•4-1

i5
13

c

3

"ca

"?

"Sd
o

4>

O
u

ca

O

u

D.

U

E

o

u

o

■♦-•

13

c«
O

13

c

3

(L>

u

u

^

U5

3
X>

.&3

x;

c

■£

ca

•3

O

CJ

C/3

4-»

CJ

3

CJ

S

.§

_c

^

_C

>

_C

03

to

>

-«-)

.^

g-

.^

e

3

_«

^

-5

^

•3

o

"Sc

X

"3

•ti

-Q

•2

a

B

2 c

u •-

•5 °

? O

B ■-

en O

"2, '="

x: I-

>
<ca

B

o
O
-o

E

ca

•a

c
ja
en

3

o

<4H

1) o

•S «

<*-( E

o o

(/, («

?? C

o c c«

X u c

'i = 3

■o -j^ '£

i2 o «

.i£ U •£

o

•o

(«

B

ca

s.

(U

u>

v>

CO

ca

*^"

b.

X

JU

12

ca

■•-»
3
O
or)

"3
b.

E

C

6

OJ

ca

u

.c

•4-^

J3

(U

U

(4-1

_c

g

O

V)

2

X

M

1.4

ca

1)

o

1)

CJ

X

a

ed

3

CO

3

"O

CO

B

B

_o

ca

■*-•

"^

ca

'•*

3

<u

*-»

x;

C/3

-4-»

<«

_o

o

■^

u

o

E

E

_B

e^

>-.

•i

■D

O

.£0

1^

5

,Cl

'^

^

I

5

3t

a

:6

1969 ANOLIS INSOLITUS 17

c

(U

en

M

u

(U

u

o

en

(U

a.

C/3

2

u

kH

f^

u

o

Ui

o

C

B

O

C

,2

u

&

^

&

C

o

c

§

O

c

&

O

^

^

O

M

u

Lu

c

c

^

-«-<

c

"d

cn

p

p

;3

2

o

T3

T3

o

ally separ
ative disti

13
c

'5b

u
03

O

2
.5

0

1

2

o

OJ

5

3

5

C
OS

3

o

s

5

c

en

en

3

.9

limatic
lose rel

E
E

-4^

en

4— •

en

2
c

^

1!

en

en
OJ
X

-4—*

o

dj

_>

'>

X

o

-2

o

.'"

S

_o

.22

1

_o

-5

CJ

3

03

-4— »

3

o

CJ CJ

c«

>>

■3

CJ

•5

a

"o

"o

X

X3

.g

t5.^

^

O

_>^

1

O

O

c

■*-i

c

en
<U

CJ

§ B

'?

O-

E

c/3

c

c

■;-;

_c

.C

tn O

3
X)

en

3
Xi

C3

x;

en

3
X

X
en

s

3
SD

en

a,

_'en

■3

o

en en

3

a
o

o

—

-*— »

13

i2

en

13

O

c

■4— >

CJ

_n

en

C3

■^^

^

o

_>,

.o

■♦^

"a.

^

E

0

5

■a

u. X « CO «iJc3S

c -a -a T3 ^ -a o 'aOC;:::Ogo

oc c c*^ c"" c'"o«'*"oE

'«« .22 nj .22 iJ „-=■« =>^_

x:^ x'o xi'2 &'2'2 ^ := ' '

*-« troS —03 >Co3 > "^

ex ex 5x^.52x^5<S

i i r\ y so susracJSCJ

o o Si- ou otSi^fcE

C c

^ =: V. Si ?N r o

18 BREVIORA No. 326

.0

!c

x:

Oh

Q-

i»«

rt

03

■2

u

u«

«.«

ex)

00

c

O

O

<4j

u

4)

Is.

00

00

■5

o

c

c

o

c

c

.o

_g

.2

_J-

^—t

C3

c3

■*..

^

*-<

C

C

i-

"3

(U

■3

<u

o

o

o

c

c

■3

c

•3

3

1
03

D-

CS

>

3

o

^

(U

'C

o

kH

-4~*

■«— »

ca

u

00

03

o

T3

tn

3

O-

■*-»

c

"55

"«

3

-4^

3

o

-3

•5

VM

(^H

3

C/3

s

o

XI

>>

o

^

£?

^

■♦-»

13

5

u

'*-*

^

<U

-*-»

c

03

>i

^

J2

a

X

03

i^

lU

c

-4— •

O

o

-*— t

■«-»

a

-4— »

■*...

^

"H

*u

r

C

Si

. r

■♦-»

c/)

*^

'c3

tj

CJ

(M

C/5

a.

1

5

_u2

O

Q

-3

t:

0

>>

'•B

Si

1j

•3

to

•Si

en

.2 u

c

oj

o

c

O

(4-1

^-t

rt

O

,—1

o

c

O
o

•o

U5
C

03

•^^

JH

c

_o

C

3

x>

(U
T3

-4—1

3
O

s

i2

03

>

(U

U

03

T3
C

3

•3

a

o

1)

op

o

v:

(4-1

(—

IS

I

U

s

2

o

Jo

tj

^

0

1969 ANOLIS INSOLITUS 19

REFERENCES CITED

Cochran. D.

1941. The herpetology of Hispaniola. U. S. Nat. Mus. Bull. 177: 1-
398.

Etheridge, R.

1960. The relationships of the anoles (Reptilia:Sauria:Iguanidae) : an
interpretation based on skeletal morphology. University Micro-
films, Inc., .Ann Arbor, Michigan, x + 236 pp

1965. The abdominal skeleton of lizards in the family Iguanidae.
Herpetologica21: 161-168.

Webster, T. P.

1969. Ecological observations on Anolis occult us Williams and Rivero
(Sauria, Iguanidae). Breviora No. 312: 1-5.

Williams. E. E.

1961. The evolution and relationships of the Anolis semilineatus group.
Breviora No. 136: 1-8.

Williams. E. E., J. A. Rivero, and R. Thomas

1965. A new anole (Sauria, Iguanidae) from Puerto Rico. Breviora
No. 231: 1-18.

(Received 29 April 1969.)

20

BREVIORA

No. 326

protocarolinensis

Fig. 5. Diagram of relationships within the carolinensis subsection of alpha
Anolis. A. darlingtoni, A. insolitus, and A. occultiis are primitive
relicts within this subsection.

1969

ANOLIS INSOLITUS

21

to
:5

to

• ♦

s

•T3

•^ ^

O

a
o

(U

g^

so

(Vff -CLa\-nbr.c(^€ J ,.-,US. COMP. ZOOL

LIBRARY

B R E V I O RepAi969

MmseiiiiiM of Contiparsitive ZooiogysiTY.

Cambridge, Mass. 15 September, 1969 Number 327

THE ANGLES OF LA PALMA:
ASPECTS OF THEIR ECOLOGICAL RELATIONSHIPS

A. Stanley Rand' and Ernest E. Williams

Abstract. The ecological relationships of the anoles known from a lo-
cality in the Cordillera Central of the Dominican Republic are described
in terms of the climatic and structural habitats found useful in describing
these relationships for the anoles of the Hispaniolan lowlands and of other
West Indian islands. The montane fauna of the central Dominican Republic
is closest both ecologically and phyletically to the montane fauna of the
southwestern portion of Hispaniola. Discrepancies may be due to incom-
plete knowledge of these montane faunas, which have only recently been
carefully collected. Both the Hispaniolan montane faunas are now relict,
although locally occurring in dense populations.

INTRODUCTION

Reports of the ecological relationships of closely related sympat-
ric species provide one of the basic lodes of raw material for
studies of competition, adaptive radiation, and evolution. This
paper, concentrating on a relict population of montane anoles in
Hispaniola, is one of several describing the ecological relations
among sympatric anoles at various places in the West Indies
(e.g. Ruibal, 1961; Collette, 1961; Rand, 1962, 1964, 1967b;
Rand and Rand, 1967; Schoener, 1967, 1968; Schoener and Gor-
man, 1968). It fills an important gap in the series and is pre-
liminary to a study comparing the patterns of ecological adapta-
tion shown by anoles in different areas and discussing their evolu-
tionary significance.

The area we chose for study — La Palma in the Cordillera
Central of the Dominican Republic — is one in which there are
small relict patches of broad-leaf montane forest surrounded by

^Smithsonian Tropical Research Institute, Balboa, Canal Zone.

2 BREVIORA No. 327

cultural steppe. Though most of this type of forest has been cut
all over Hispaniola, relicts like those at La Palma occur at a num-
ber of places. La Palma provided us with the opportunity to study
the interrelationships of seven species of anoles, four of them
montane and occurring in the relict forest, and three lowland, oc-
curring primarily outside the forest. In this paper, we have con-
centrated on one locality instead of discussing relationships be-
tween montane and lowland anole faunas in general terms, be-
cause an approach in terms of a local fauna will allow the reader
to distinguish more clearly our factual information (i.e., what we
saw) from our ideas about what probably happens.

Our observations were made at La Palma during eight visits,
two of them at night, on five days between 23 and 31 July, 1968.
The brevity of our observations and their concentration at one
time of year limits their generality, but they are much more exten-
sive than any yet available for this specific montane situation, and,
since there is litde prospect of extending them in the near future,
they are offered here.

Our La Palma observations have been supplemented by our
observations at nearby areas and elsewhere in Hispaniola and by
the unpublished observations of Drs. James Lazell, Clayton Ray,
and Albert Schwartz, to whom we are grateful.

We are deeply indebted to Dr. E. de J. Marcano and his asso-
ciates and family for their help and kindness in the Dominican Re-
public. Dr. Marcano's enthusiasm and generosity in helping visit-
ing scientists are extraordinary, and we have benefited largely.

This study was partly supported by National Science Founda-
tion grant GB-6944 to Ernest E. Williams.

GENERAL REMARKS AND DEFINITIONS

Our study at La Palma was undertaken in order to determine
the ecological relationships of the anoles sympatric there. Since
the data are intended for comparison with earlier studies at other
localities, we have tried no new methods. We have assumed that
careful censusing of microhabitats (structural and climatic niches,
see below) and a record of adult size will give a sufficient approxi-
mation of the most important ecological relationships between the
species. As an aid to future workers, and because of their rele-
vance to interspecific social behavior, we have also noted general
appearance and dewlaps. (A comment on the signal significance
of dewlaps is being published elsewhere.)

1969 THE ANGLES OF LA PALM A 3

Structural and climatic habitats. These terms are used in the
senses previously defined (Rand, 1964, 1967b; Schoener, 1967,
1968). Structural habitat refers to that aspect of the microdistribu-
tion of an anole that can be described in terms of the physical
structure of the environment. It is recorded as the height, diameter,
and other characteristics of the perch used during the day, with
mention of the normal foraging areas.

Climatic habitat describes an anole's relation to temperature and
moisture in the environment. It is usually described in terms of
li^ht and shade distribution without reference to moisture. Obser-
vations on basking behavior are noted here. Earlier papers on
anoles (Ruibal, 1961, and Rand, 1964) suggested that the most
important factor in determining the climatic distribution of anoles
was temperature, and certainly lizard body temperatures taken
with a Schulteis thermometer correlate well with the light and
shade aspects of the climatic habitat occupied. Even in the early
papers, however, it was realized that humidity might be very
important, and our observations and some experimental work of
Sexton and Heatwole favor this. Sexton and Heatwole (1968)
have shown that forest anoles in Panama lose water much more
quickly under conditions of desiccation than do open-country
anoles.

For both structural and climatic habitat some standardization of
terms has been found useful:

Within the aspect of anole adaptation to structural habitat there
are several types of anoles recognizable in terms of the perch area
or areas on which at least adult males spend most of their active
day. Such types are:

Crown anoles: anoles, typically of very large size, that are most
frequently seen very high in trees, i.e., within the crown.

Twig anoles: dwarf anoles characteristically lying along thin
branches and twi^s, often near or in the crown.

Trunk-crown anoles: anoles usually found on the upper trunks
of trees or in the crown.

Trunk anoles: anoles mainly confined to tree trunks and rarely
climbing to the crown or descending to the ground.

Trunk-ground anoles: anoles found on the lower trunks of trees,
foraging from there onto the ground.

Grass-bush anoles: anoles whose characteristic perch is on
grass stems and bushes, away from trees. (This type of habitat is
sometimes subdivided, and there may thus be grass anoles and
bush anoles.)

4 BREVIORA No. 327

Within the concept of cHmatic habitat, the convenient terms for
the anoles adapted to the several habitats are sun anoles, half-
shade anoles, and shade anoles. These must be used, however,
with the express caveat that the situation in terms of insolation is
not nearly as simple as these terms imply and that moisture, as sug-
gested above, is of very real, perhaps primary, importance. Sun
anole, half-shade anole, or shade anole imply only that the species
so designated are most frequently (but not always) found in sunny
situations, or in half or full shade.

A particular point must be made of the fact that observation of
a single individual in sun or shade at a particular moment is not
primary evidence of climatic habitat nor of basking behavior. Bask-
ing behavior must be defined as deliberate movement into a spot
of sunlight, movement with the sunlight, and special postures as-
sociated with being in the sun. Non-basking animals may avoid
sunny areas or move into and out of them almost randomly.

LA PALMA DESCRIBED

The specific area studied is adjacent to the tiny village of La
Palma (ca. 800 meters) about 12 km east of El Rio on the road to
Bonao, La Vega Province, Dominican Republic. The small stream
crossing the road here is clear, fast moving, and rocky, with a
succession of pools and riffles about six to ten feet wide and shallow
enough to be waded.

We worked upstream in the small patches of forest that bordered
the stream. The area most thoroughly worked is up to 50 yards
wide between the stream and a trail roughly paralleling it. This
area extends about 200 yards from the road to the first crossing
of trail and stream. Though the forest here had been hacked at,
it had probably never been completely cleared at any time. Small
trees form a nearly complete canopy 15 to 30 feet tall with a few
open, swampy spots with only three- to five-foot high dense vege-
tation. There are scattered emergent trees 30-50 feet tall. These
larger trees are most common and closest together along the banks
of the stream.

Undergrowth is moderately dense in most places and, in addi-
tion to saplings and bushes, includes tree ferns, smaller ferns, and
nettles. Vines are common but do not form vine curtains in most
places. Movement is possible in any direction along the small
paths or, except for a few dense fern patches, by following a wind-
ing course. Some moss and many lichens occur on tree trunks.

1969 THE ANGLES OF LA PALMA 5

Bromeliads are only moderately common. The ground is nearly
completely covered by a shallow leaf litter, much scratched, par-
ticularly near the main road, by chickens.

Above the first ford, in the area between the stream and the
trail, the undergrowth had been cleared and coffee planted be-
tween large trees retained as shade.

Along much of the trail beyond the first ford there are fences
and frequent dense fern thickets. In one place an outcrop of rock
briefly replaces the fence on one side. Away from the stream and
beyond the fences, the slopes are covered with pasture, or are
under cultivation for beans in places. High on the ridges are bits
of forest and scattered pine.

During one day we followed the trail for several kilometers as
it crossed and recrossed the stream. The conditions of the first
area repeated themselves until, beyond the small savanna settle-
ment, even this gallery forest disappeared.

THE ANGLES
Table 1

Seven species of Anolis are known from the area of La Palma
in the Cordillera Central of the Dominican Republic, making this
among the richest in anole species of any area of comparable size
in the West Indies. All species are restricted to Hispaniola except
one (A. distichus), which also occurs in the Bahamas. Three
species are widespread and quite well known in the lowlands {A.
distichus, A. cybotes, A. ricordii) (Mertens, 1939; Rand, 1962).
The remaining four are montane species, much more restricted in
distribution and much less well known {A. aliniger) (Williams,
1965a); A. christophei, A. etJieridgei (Thomas and Schwartz,
1967); A. insolitus (Williams and Rand, 1969). Two species {A.
chlorocyamis, primarily lowland, and A. cochranae, strictly mon-
tane) that have not been found at La Palma are known from the
general area and might occur at La Palma.

As mentioned above, the La Palma anoles fall into two natural
groups, those known only from montane areas and those that also
have a very wide lowland distribution. Three of the four montane
species are primarily forest forms, the fourth may also occur in the
forest but is certainly tolerant of open conditions, since it was
under these conditions that it was taken at La Palma. Two of
the three lowland species occur principally in the open, the third
certainly occurs in the forest but is probably tolerant of relatively
open situations.

BREVIORA No. 327

m o

•I 1

I ?

r \ «

UJ

_j

CO

<

"t3

u

euo

T3

U

B

O

•o

c

CS

.4_,

,^_,

t/5

Ui

(U

(U

;-<

k^

,o

o

c

c

c

c

lU

OJ

<u

<u

D-

a-

c-

o-

o

o

o

o

C

C

c

3

(O

^

&

O

O

o

in

x:

o

s

in

c

C3

13
C
3
O

o

•a

13

UJ

c

c

C3

C

a.

en
■*— '

C

a

u.

u^

ex)

^'

.V

<u

o

en

tt;

_c

£>

2

•a

c

3

C
3

OJ

O

!c

•*—

^

tj

c

^

u.

o

2

go

s

;-
o

L^
CJ

4>

2

M

«

(U

u

a-

D.

c/5

1)

E

c

Lh

M

13

c^-

n.

O.

lU

r.

<U

1)

C

o

U5

c/5

3

3

U

T3

"O

•n

■*»*

cn

o

c

C3
en

c

3

(U

u
c

<u

<4-l

o

en

C

C
3
O

c

en

-4— t

4J

c

c

c

3

c

en

en

d

C

e

1)

<u

u

u.

JJ

c

Uh

o

O

3

Iri

Xi

-o

"in

«

60

EE EEEEEE E

EE EEEEEE E

oo o>ooooc o

^ OX)

E

E

E

E

3

?

3

3

*U

4>

•5

T3

■o

01)

•a

&£)

0)

<u

<u

u,

(U

u

E

E

E

C3

E

JS

^i EEE«E^ g

a

1969 THE ANGLES OF LA PALMA

MONTANE SPECIES

Anolis insolitus. A small species (ca 40 mm snout-vent length)
gray or green, more or less distinctly mottled. Dewlap large in
both sexes, skin blue-gray anteriorly, dull orange posteriorly,
crossed by well-separated rows of white scales.

This recently described species (Williams and Rand, 1969) is
known at present only from La Palma, where we found five ani-
mals in the forest.

The one seen during the day was about eight feet up on a slender
branch of a small tree. It climbed along slender branches and
hopped from one to another. It made no obvious efforts to seek
or avoid the sun while we watched it.

In structural habitat, A. insolitus appears to be clearly a twig
anole. In cUmatic habitat, it appears to prefer half-shade. In both
respects, it strongly resembles A. occultus of Puerto Rico and
probably resembles A. isolepis of Cuba (R. Ruibal, pers. comm.).
It is not unlike young A. valencienni in Jamaica. The extent of
ecological similarity to its much larger close relative, A . darlingtoni
of southwestern Haiti (Williams and Rand, 1969), is unknown.

Anolis christophei. A medium-sized species (a little under 50
mm snout-vent length), brownish or greenish, conspicuously
marked with dark dorsal butterflies, a pair of conspicuous light
spots on the flanks and a light yellow or green mark under the eye.
Dewlap very large in male, the skin light purplish gray, crossed
by distinct rows of light scales. Dewlap very reduced in female.

This species, widely distributed at moderate to high elevations in
the Cordillera Central and occurring also in the Cordillera Septen-
trional (Thomas and Schwartz, 1967), occurs at La Palma
throughout the forest and occasionally on the edge, and also on
the shade trees in coffee plantings. Usually it was in shade but
occasionally in a sun fleck. No basking behavior was seen.

It perched on the trunks of trees, from saplings to very large
trees. When smaller perches were used, these were typically next
to a big tree or steep bank. Some individuals were seen on vertical
earth banks along a stream and others on damp, vertical walls of
a small outcrop. No close association with complex fissured tree
trunks (such as that characteristic of A. lucius in Cuba, Ruibal,
1964) was evident. Most individuals were between two and twenty
feet up; none were seen to go to the ground.

In structural habitat, A. christophei appears to be a trunk anole,
as is /4. distichus, but the latter occurs at La Palma in the open
instead of in forest shade.

8 BREVIORA No. 327

Anolis etheridgei. A medium-sized species {a bit less than 45
mm snout-vent length). Males plain greenish brown or with
darker cross banding {we do not know if one male can show both).
Females polymorphic with a highly variable dorsal pattern of
stripes and spots. Dewlap in male of moderate size, the skin me-
dium gray with evenly spaced whitish scales. Dewlap in female
much reduced.

This species, restricted so far as known to higher elevations in
the Cordillera Central, occurs at La Palma both inside the forest
and at the edge. Almost all those seen were in shade; there was
no suggestion of basking.

Almost all individuals were on slender perches (i.e., branches,
twigs, fern stems, etc.) close to the ground; they perhaps come
regularly to the ground. Many seem to live in dense bushes and
fern thickets, since they were occasionally seen at the edges of
these during warm, sunny periods and were commonly taken
sleeping on them at night.

In structural habitat A. etheridgei is a bush anole, resembling
A. krugi of Puerto Rico and, like that species, it prefers at least
moderate shade. It also resembles, and is more closely related to,
A. hendersoni of southern Hispaniola (Williams, 1963a).

Anolis aliniger. A medium-sized species (ca 50 mm snout-vent
length). Green with the capacity to turn brown. Dewlap extremely
reduced in both sexes. Perhaps replacing it in display is an orange
area in the axilla followed by a black spot. Both markings are
larger and brighter in the male.

This species is poorly known, although records indicate a wide
distribution at moderate to high elevations in the Cordillera Cen-
tral and across the Cul de Sac plain in the Massif de La Selle in
Haiti.

The two specimens from La Palma were taken along the edge
of the forest, one on a fence post in the sun, the other on a bush
at night.

Very litde is known about the ecology of this species, but from
the very limited evidence, it may be a trunk-crown species of the
montane forest, approaching the ground at forest edges.

A. aliniger is very closely related to A. singularis of southern
Hispaniola, the behavior of which is even less understood than is
that of A. aliniger. Both species are almost as closely related to
the larger species A . chlorocyanus, which is a trunk-crown species
of the lowlands known from near La Palma (Puenta Jimenoa, El
Rio) but not from La Palma itself. A fourth species, A. coelestinus,

1969 THE ANGLES OF LA PALMA 9

is the geographic representative of A. clilorocyanus in southern
Hispaniola (Williams, 1965a).

PRIMARILY LOWLAND SPECIES

Anolis ricordii. A very large species {over 100 mm snout-vent
length), mottled gray and green. Dewlap large in both sexes, skin
bluish gray in center, paling through light violet to pink at lower
margin, edge with dense yellow-white scales.

This species occurs throughout Hispaniola, though with con-
siderable geographic variation (Williams, 1962a, 1965b). It is
primarily an inhabitant of the lowlands and is not known from
high elevations.

At La Palma, one juvenile was taken about 10 feet up on a
small branch of a small forest tree. Two adults were brought us
from a large tree in a nearby agricultural area. It thus appears
to occur here both in forest and in more open areas.

In other places in Hispaniola this is a crown species, as it prob-
ably is here. Whether or not this species basks is unclear, but it
may well do so.

In its ecology, A . ricordii strongly resembles the other giant an-
oles, A. cuvieri of Puerto Rico and A. equestris of Cuba. It is a
little less similar to the somewhat smaller A. garmani of Jamaica.

Anolis distichus. A species of moderate size (ca 50 mm snout-
vent length) with considerable power of color change: green, gray,
or brown; all light phases are finely mottled with darker. Dewlap
in male of moderate size, with a rather wide, bright yellow border
and bright red center. Dewlap reduced in female.

This species, which has many striking geographic color variants
(Schwartz, 1968), occurs everywhere in Hispaniola except in situ-
ations of extreme aridity; it does occur at quite high elevations.

At La Palma, A. distichus is present in the open agricultural
steppe and at the forest edge. We saw it also, but very rarely, in
the forest. It is commonly seen sitting in sun flecks and seems to
bask.

It perches on trees, frequently very large trees, and on fence
posts, occurs at heights of one to twenty feet, and apparently does
not regularly come to the ground.

A trunk anole like A. christophei in structural habitat, it differs
in climatic habitat. The two overlap in microdistribution, but A.
christophei is much more common in the forest and A. distichus in
the open, each almost to the exclusion of the other.

10 BREVIORA No. 327

Anolis cybotes. A large species (ca 60 mm snout-vent length).
Brown with strong, dark, transverse dorsal markings and more
or less distinct, light, longitudinal stripes on the flanks. Dewlap in
male large, skin very pale yellow crossed by rows of similarly col-
ored scales. Dewlap in females much reduced.

A. cybotes is a very common lowland species replaced in cer-
tain arid localities by its close relative A. whitemani, and in the
highest areas of the Cordillera Central, by the related A. shrevei.

At La Palma, this species was seen in the open, at the forest
edge and very rarely inside it. It is not as common here as A.
distichus and seems associated chiefly with the edges of dense
thickets in the open. One juvenile was seen in the forest on the
ground.

In most areas, this is a trunk-ground species (Rand, 1961),
common on trees and fence posts. Obvious comparisons then are
with other trunk-ground species, such as A. sagrei on Cuba, A.
lineatopus on Jamaica, and A. cristatellus on Puerto Rico. These
comparisons are not so obvious at La Palma. The association with
thickets and with the ground invites consideration of A . etheridgei
as its forest counterpart. A. cybotes at La Palma is a bit anomalous
and will be discussed further below.

DISCUSSION

The ecological picture seen in the anoles of La Palma requires
discussion of three quite discrete topics, which we treat serially
below:

1. The ecological significance of anole microdistributions. As
has proved true of Anolis elsewhere, the anoles at La Palma differ
from one another ecologically in obvious and striking ways. Here,
as on other islands, the most conspicuous differences involve ( 1 )
size, which implies differences in size of food taken (Schocner,
1967); (2) structural habitat (i.e. size and height of typical perch
and associated foraging area); and (3) climatic habitat (at La
Palma, the differences between inside and outside the forest).
These last two imply differences in foraging area.

Taken together, these three parameters specify a different but
not exclusive feeding habitat for each species — not exclusive,
because significant overlaps exist among species: a large species
is small when young; both forest and open-country species occur
together at the forest edge, and each invades, though only rarely,
the climatic habitat of the others; finally, individuals of each species
sometimes use quite atypical perches. Despite these overlaps, the

1969 THE ANGLES OF LA PALMA 11

central portion of the food resources of each species seems unique
to it.

The patterns of differences among sympatric anoles seen at La
Palma are those usual in all well-analyzed situations in Anolis:
structural habitat differences among anoles with the same climatic
habitat; parallel structural habitats among animals with different
climatic habitats; size as a reinforcing factor for either of these
ecological patterns.

It is more usual than not for anole species to be constant, or
nearly so, in their ecology throughout their ranges. The example
of A. cy botes at La Palma reminds us, however, never to assume
this out of hand. A. cy botes at La Palma is more nearly a bush
anole than the trunk-ground anole it is over most of Hispaniola.
Why it should have diverged in this fashion at La Palma is not
clear, but it is clear that the only stringent ecological constraint
on any species is that its ecological specializations be relevant and
adequate in the specific biotic context in which it finds itself.
Beyond this, restraints upon divergence are genetic: the possibili-
ties for genetic change present in the genome of the animal, this
genome being the product of prior history, i.e., the selective stresses
of other times or places.

2. The relationships of the montane faunas. The most distinc-
tive montane anole fauna in the West Indies is that of Hispaniola,
and this is at present best known from the northern section of His-
paniola (the "north island" in the sense of Williams, 1961).

Four montane species are found at La Palma; three additional
montane species, known from the Cordillera Central, appear not
to occur there. One, A. rimarum (Thomas and Schwartz, 1967),
is known only from one forest locality in Haiti, where it coexists
with A. christophei. It was described from a very specialized
habitat (boulders of talus slope) that does not occur at La Palma.
The two remaining species, A. shrevei and A. cochranae are in-
habitants of more open country: A. shrevei is a cy botes relative
living in pine forests at very high altitudes; apparently it nowhere
occurs in broad-leaf vegetation, and it does not follow pine to
lower elevations. A. cochranae is a grass anole of high elevations.
It might occur in the La Palma area away from the forest, but it
has not been taken despite considerable night collecting.

Three of the anoles of the montane broad-leaf forest of the
"north island" (A. christophei, A. etheridgei, and A. rimarum)
are all members of one phyletic group (Williams, 1962b; Thomas

12 BREVIORA No. 327

and Schwartz, 1967). The fourth anole of the north island mon-
tane forest, A. imolitus, is not as close but is probably a deriva-
tive within Hispaniola of the stock ancestral to this group and to
the coelestimis-chlorocyanus group (green anoles, Williams,
1965a). The other two montane anoles, A. cochranae (grass)
(Williams and Rand, 1961) and A. shrevei (pine) (Williams,
1963b), are, in contrast, derivatives of two diverse stocks and are
each more closely related to a lowland stock than to the animals
of the montane broad-leaf forest.

The "south island" of Hispaniola also has a montane anole
fauna, but there are almost no species in common between "north"
and "south" islands. The sole known exception, A. aliniger, has
obviously established a rather recent beachhead across the Cul
de Sac plain in the Massif de La Selle in Haiti (Williams, 1965a).

The "south island" montane species are, with the possible ex-
ceptions of A. aliniger and A. singularis, all more or less shade,
i.e., forest, animals. One, A. monticola, is a very close relative of
A. rimarum and occupies a similar rock-ground habitat. Southern
A. darlingtoni is close to A. insolitus but is much larger, and its
ecology is unknown. (It is still known from a unique type.) A.
singularis is the southern representative of A. aliniger but lacks
the latter's bold axillary pattern. A. hendersoni is the southern
ecological equivalent of A. etheridgei but is not phyletically close.
A. koopmani (Rand, 1964) is a dwarf ground anole believed to
be close to the A. monticola, A. rimarum, A. etheridgei, A. chris-
tophei group, but it has no known northern representative. On
the other hand, there are no known "south island" representatives
of A. christophei or A. shrevei or A. cochranae.

Table 2 summarizes our present knowledge of the montane
anole faunas of Hispaniola and of their north-south relationships.
A question necessarily arises as to the completeness of this knowl-
edge. Is there really no southern representative of A. christophei?
No northern representative of A. koopmani? The faunas have
been separated long enough to have achieved species difference in
the two areas; they might also be old enough to differ genuinely
as faunas also. However, the total known montane anole fauna
of Hispaniola is twelve species; of these, six have been described
in the last ten years; A. christophei, Williams, (1960); A. koop-
mani, Rand (1961); A. cochranae, Williams and Rand (1961);
A. singularis, Williams (1965a); A. rimarum, Thomas and
Schwartz (1967); A. insolitus, Williams and Rand (1969). Very

1969 THE ANGLES OF LA PALMA 13

little work has yet been done in the Massif de la Hotte at the ex-
treme end of the southwest peninsula of Haiti. It is consequently
very possible that species are yet to be discovered.

TABLE

2

Montane

anole faunas in

HiSPANIOLA

NORTH ISLAND

SOUTH ISLAN

A. christophei
A . etheridgei
A . rimarum

A. (hendersoni)
A. monticola

A . aliniger

A . insolitus
A . cochranae

A . singular is
A. koopmani
A . darlingtoni

A . shrevei

We thus cannot at the moment determine whether the differ-
ences we now appear to see between the montane faunas of the
north and south islands are (1) historically real — i.e., the result
of long separation between the two areas, or (2) ascribable to the
accident of extinction following a relatively recent separation, or
( 3 ) only an artifact of our still imperfect knowledge of the perti-
nent (particularly south island) areas. Of these, the first seems
the least probable.

However, even on our present knowledge the point is clear that
the montane forest faunas of Hispaniola, excluding insolitus and
darlingtoni, make a close phylogenetic unit that seems to have only
distant relationships with other anoles both within and without
Hispaniola. A larger but still genuine phylogenetic unit includes
these forest forms with insolitus and darlingtoni and the green
anoles. All of these, then, contrast with the other Hispaniolan
anoles, which seem to represent several different phylogenetic
groups with more or less close relationships outside Hispaniola.

Such a set of relationships implies differences in times of origin
or invasion. The montane forest species are the end points of an
old radiation within Hispaniola. The widespread lowland species,
except for the green anoles (chlorocyanus group, Williams 1965a,
related on the one hand to insolitus and darlingtoni and on the
other to the montane forest species), are the results of more re-
cent invasions of Hispaniola and have provided more recent and
less extensive radiations within the island.

14 BREVIORA No. 327

This postulated history of the fauna suggests that anoles have
invaded Hispaniola and estabhshed themselves throughout both
the open and disturbed habitats, for these are the current habitats
of the widespread lowlands species (least true of ricordii). The
older species tend to be established in the montane forests and to
resist invaders there. This pattern is similar to that suggested
by Wilson, 1959, 1961, for the Melanesian ants.

3. Relict habitats and relict populations. On a very recent time
scale, the forest at La Palma is a relict habitat and contains a
relict of a once much more widely distributed and, as just sug-
gested, probably old anole fauna. The ridges and hillsides above
La Palma that now support scrub and pine were not long ago
largely covered with broad-leaf forest. Below the frost line, pine
appears to be maintained in most of the Cordillera Central by
fire (of which there are many evidences, such as charred trunks)
and cutting. In one place near Jarabacoa at moderate elevations,
we found a hillside with many large pines, 40-50 feet tall, among
which grew a dense broad-leaf undergrowth 15-20 feet tall. We
could find no pine seedlings or young pine trees. It seems clear
that if not disturbed, much that is now pine will be replaced by
broad-leaf forest.

The present disjunct distributions of A. christophei and A.
etheridgei (Thomas and Schwartz, 1967) are the results of forest
destruction. Almost all of the original broad-leaf forest in which
they live has now disappeared, leaving scant isolated remnants.
The original distribution of this forest is mapped in the "Mapa
Ecologico de la Republica Dominicana" in "Reconocimiento y
Evaluacion de los Recursos naturales de la Republica Domini-
cana" published by the Pan American Union in 1967.

However, though today occurring only in scattered colonies, in
pre-Columbian times anoles of this forest type must have been
one of the dominant anole faunas of the islands, and the lowland
and edge species that are now so widespread and abundant must
have been restricted to very much smaller areas.

We do not know how many of the anole species that once in-
habited the montane forests have become extinct during the period
of dissection of these extensive forests into isolated relicts, and
even among those still extant, it is certainly possible that there are
still undiscovered species. (As we have mentioned above, six
montane species have been described from Hispaniola in the last
ten years.) Probably, however, we have a reasonably good rep-
resentation of the original fauna in such small relict areas as La

1969 THE ANGLES OF LA PALMA 15

Palma — at least of the "north island" portion of that montane
fauna.

We do know that in these relict areas the montane species may
be quite successful, and quite high population densities are some-
times reached. At La Palma we saw 35 A. etheridgei asleep dur-
ing two hours of night collecting. (Whether this is quite as high
as the densities that are reached by lowland species in some
perianthropic areas we cannot be sure.)

In fact, anoles — at least some anoles — seem adapted to liv-
ing in very high densities in small areas, and this is probably an
important factor in their evolutionary success in the West Indies,
both in persisting in relict areas and in becoming established in
new ones. Anoles can live and thrive on small islands and in small
patches of suitable habitat. In fact, population densities may be
higher in such small areas because bird species specializing in sim-
ilar habitats are absent: (such areas are too small to support
them), and birds are probably the most important predators on
and competitors with West Indian anoles. In the West Indies
today, with its many small patches of suitable habitat (gardens,
fence rows, groves of trees, little areas of forest) anoles may be
generally more common than they were in earlier periods when
habitats were more uniformly and continuously distributed.

Though anoles can and do survive in some very small patches of
suitable habitat, they do not do so in every such spot. The ad-
vantages gained from the reduction in numbers of bird predators
and competitors may sometimes be outweighed by the numbers
of adjacent, potentially competing species of anoles. As at La
Palma, the area around a relict forest is usually inhabited by open-
country anoles with structural habitats similar to those of the
forest species. As we have seen, the forest and open species gen-
erally avoid competition by occupying different climatic habitats.
But in those species of anoles that have been studied in detail,
some individuals do disperse into relatively unfavorable habitats
and live there at least briefly. Thus at La Palma we saw occasional
open-country animals within the forest. Where both open and
forested areas are not too disparate in size and the population den-
sity in the open areas not too great, this invasion pressure is prob-
ably unimportant, but where one habitat is relatively much smaller
than the other and the larger one is densely populated, vagrant
animals from the larger population may well be an important
source of competition and, in marginal cases at least, tip the bal-
ance toward local extinction of the relict population. Of course.

16 BREVIORA No. 327

the smaller and more scattered the relict habitats, the more pre-
carious this balance must be for the relict populations.

We have stressed above the abundance of anoles in certain
situations, but it is important to remember that not all anoles, nor
any anole species at all places and times, are abundant within their
natural range. It is clearly as characteristic of West Indian anoles
as of other animals that their distributions are clumped or patchy.
Many of the lowland species of Anolis in Hispaniola or the other
islands give the casual visitor the impression of overwhelming
abundance. This is frequently the case in perianthropic situa-
tions, but even the commonest species are not everywhere present
even in habitats that to the human eye are ideal and are continu-
ous with areas of high population density. Rarer species may
require specialized habitats that are widely disjunct, such as the
rocky areas favored by A. monticola or A. rimarwn. For still
other species their habitat, perhaps high in the trees, e.g., A. ric-
ordii, or their cryptic coloration and behavior, e.g., A. insolitus,
reduce their visibility to the observer; for these we have no basis
for an estimate of population density. We have, however, no
reason not to believe that their distribution is as patchy, even
within favorable habitats, as is the distribution of better known
anoles. Variations in species abundance and patchiness in dis-
tribution are empirically obvious, but the factors producing them
are a major unsolved ecological problem.

PAPERS CITED

COLLETTE, B. B.

1961. Correlations between ecology and morphology in anoline liz-
ards from Havana, Cuba and southern Florida. Bull. Mus.
Comp. Zool. 125:137-162.
Mertens, R.

1939. Herpetologische Ergebnisse einer Reise nach der Insel Hispani-
ola, Westindien. Abhandl. Senckenberg. Naturf. Ges. 449:1-84.
Rand, A. S.

1961. Notes on Hispaniolan herpetology 4. Anolis koopmani, new
species from the southwestern peninsula of Haiti. Breviora
No. 137:1-4.

1962. Notes on Hispaniolan herpetology 5. The natural history of
three sympatric species of Anolis. Breviora No. 154: 1-15.

1964. Ecological distribution in anoline lizards of Puerto Rico.

Ecology 45:745-752.
1967a. Ecology and social organization in the iguanid lizard, Anolis

lineatopus. Proc. U. S. Nat. Mus. 122, No. 3595:1-79.

1969 THE ANGLES OF LA P ALMA 17

1967b. The ecological distribution of anoline lizards around Kingston,
Jamaica. Breviora No. 272: 1-18.
Rand, A. S., and P. J. Rand

1967. Field notes on Anolis lineatiis in Curagao, Stud. Fauna Cura-
sao 24:112-117.

RUIBAL, R.

1961. Thermal relations of five species of tropical lizards. Evolu-
tion 15:98-111.

1964. An annotated checklist and key to the anoline lizards of Cuba.
Bull. Mus. Comp. Zool. 130:475-520.

SCHOENER, T.

1967. The ecological significance of sexual dimorphism in size in
the lizard Anolis conspersiis. Science 155:474-477.

1968. The Anolis lizards of Bimini: resource partioning in a com-
plex fauna. Ecology 49:704-726.

ScHOENER, T., and G. Gorman

1969. Some niche differences in three Lesser Antillean lizards of the
[1968] genus Anolis. Ecology 49:819-830.

Schwartz, A.

1968. Geographic variation in Anolis distichus Cope (Lacertilia,
Iguanidae) in the Bahama Islands and Hispaniola. Bull. Mus.
Comp. Zool. 137:255-310.

Sexton, O. J., and H. Heatwole

1969. An experimental investigation of habitat selection and water
loss in some anoline lizards. Ecology 49:762-767.

Thomas, R., and A. Schwartz

1967. The monticola group of the lizard genus Anolis in Hispaniola.
Breviora No. 261:1-27.
Williams, E. E.

1960. Notes on Hispaniolan herpetology 1. Anolis christophei, new
species from the Citadel of King Christophe, Haiti. Breviora
No. 117:1-7.

1961. Notes on Hispaniolan herpetology 3. The evolution and re-
lationships of the Anolis semilineatiis group. Breviora No.
136:1-8.

1962a. Notes on the herpetology of Hispaniola 6. The giant anoles.
Breviora No. 155:1-15.

1962b. Notes on the herpetology of Hispaniola 7. New material of
two poorly known anoles: Anolis monticola Shreve and Anolis
christophei Williams. Breviora No. 164:1-11.

1963a. Notes on Hispaniolan herpetology 8. The forms related to
Anolis hendersoni Cochran. Breviora No. 186:1-13.

1963b. Anolis whitemani, new species from Hispaniola (Sauria, Iguan-
idae). Breviora No. 197:1-8.

1965a. The species of Hispaniolan green anoles (Sauria, Iguanidae).
Breviora No. 227:1-16.

18 BREVIORA No. 327

1965b. Hispaniolan giant anoles (Sauria, Iguanidae): new data and
a new subspecies. Breviora No. 232:1-7.
Williams, E. E., and A. S. Rand

1961. Notes on Hispaniolan herpetology 2. A review of the Anolis
semilineatus group with the description of Anolis cochranae,
new species. Breviora No. 135: 1-1 1.

1969. Anolis insoUtus, a new dwarf anole of zoogeographic impor-
tance from the mountains of the Dominican Republic. Breviora
(in press).
Wilson, E. O.

1959. Adaptive shift and dispersal in a tropical ant fauna. Evolu-
tion 13:122-144.

1961. The nature of the taxon cycle in the Melanesian ant fauna.
Amer. Nat. 59:169-193.

(Received 29 April 1969.)

1969

THE ANGLES OF LA PALMA

19

c3

(/I

o

c
a

a

J3

O

t/i
■*— »
•3

IS
a

'C

«

x:

tu

J3

E

OJO

- NA-C La-W;^e^ DO NOT ^'-^—^^-^ ^oou.

LIBRARY

B R E V I O R A569

Musemim of Contiparaitive Zoolog^uY.

Cambridge, Mass. 15 September, 1969 Number 328

A NEW FOSSIL DISCOGLOSSID FROG FROM MONTANA >

AND WYOMING'

Richard Esfes

Abstract. Scotiophryne piistulosa, n. gen., n. sp., is a small discoglossid
frog from the late Cretaceous Hell Creek Formation of Montana and the
late Cretaceous Lance Formation of Wyoming. It is probably also repre-
sented in the middle Paleocene Tongue River Formation of Montana. Scoti-
ophryne has a distinctive pustular dermal skull, but its postcranial remains
show resemblances to the Recent Eurasian discoglossid Bomhina.

INTRODUCTION

Fossil frogs of Mesozoic age are rare; Hecht (1963) has sum-
marized most of the occurrences. North American Mesozoic rec-
ords, so far, are based on disarticulated remains, which are difficult
to interpret.

Recent study of late Cretaceous and Paleocene samples of verte-
brate fossils has revealed the presence of several different kinds
of frogs. Lance Creek (Wyoming) and Bug Creek local faunas
(Montana) are rich samples of a once widespread late Mesozoic
vertebrate fauna that lived on floodplains of North American Cre-
taceous epicontinental seas (Estes, 1964). The Bug Creek mate-
rial is from the Hell Creek Formation of Montana (Sloan and Van
Valen, 1965) and is part of a Bug Creek collection in the Museum
of Comparative Zoology (MCZ), Harvard University. An Amer-
ican Museum of Natural History (AMNH) collection from the
Lance Formation of Wyoming is also utilized here. The Paleocene
specimens are from Princeton University (PU) collections from
the Tongue River Formation of Montana, and are part of a fauna
presently being studied in collaboration with Glenn Jepsen and
Marshall Lambert.

^Fossil Vertebrates from the Hell Creek Formation, Montana: Contribu-
tion No. 4.

2 BREVIORA No. 328

ORDER SALIENTIA

Family Discoglossidae

Scofiophryne pustulosa, n. gen., n. sp.

Holotype. MCZ 3623, left ilium.

Etymology. Greek, skotios, dark (referring to the darkness of
the fossil bones); phryne, toad; pustulosa, referring to the dis-
tinctive pustulose sculpture of the referred skull elements.

Para types. MCZ 3624, four left and two right ilia; MCZ 3625,
11 distal ends of humeri; MCZ 3626, 14 anterior and posterior
fragments of maxillae. All specimens collected by A. D. Lewis and
party.

Locality. Bug Creek Anthills, SW Va Section 9, T 22 N, R 43 E,
McCone County, Montana.

Horizon. Hell Creek Formation, Upper Cretaceous.

Other referred specimens. AMNH 8102, right squamosal;
AMNH 8132, left maxilla; AMNH 8137, right ilium; Lance For-
mation, Wyoming. University of California, Museum of Paleontol-
ogy (UCMP) 55703, left ilium. Lance Formation, Wyoming. PU
17037, left ilium; 16784, 16827-28, humeri; Tongue River For-
mation, Montana.

Known distribution. Known from the Hell Creek Formation of
Montana and the Lance Formation of Wyoming. A probable
Paleocene record occurs in the Tongue River Formation of
Montana.

Diagnosis. A discoglossid frog with ilia most similar, among
modern discoglossids, to those of Bombina, differing from the
latter in having a relatively thicker ilial shaft, slightly more ex-
panded subacetabular expansion, and in lacking a dorsal pro-
tuberance. Differs from all recent discoglossids and most fossil
forms in having a sculptured dermal skull casque. Humeri like
those of Bombina but relatively more robust.

Description. Ilium (Fig. 1) with robust shaft having a deeply-
marked groove dorsally that extends onto the dorsomedial side
of the shaft, this groove well defined on the type but less strongly
marked on other specimens; acetabular fossa (terminology fol-
lows Estes and Tihen, 1964) relatively large, its anteroventral
border strongly produced; no dorsal protuberance as such, but
dorsal prominence showing irregularities of muscle attachment;
subacetabular expansion large, markedly set off from acetabular
fossa and directed somewhat mediad; medially a tiny raised area
on midpoint of the suture of ilium with other pelvic bones.

1969

FOSSIL DISCOGLOSSID FROG

Fig. 1. Scotiophryne piistiilosa, n. g., n. sp.: «, lateral, and b, medial
views of questionably referred left ilium, AMNH 8137, loc. V5620, Lance
Formation, Wyoming; c, lateral, and d, medial views of holotype left ilium,
MCZ 3623, Bug Creek Anthills, Hell Creek Formation, Montana; all X 6.

Referred humeri (Fig. 3) with obHque olecranon scar (termi-
nology as in Hecht and Estes, 1960), small but deep fossa cubitus
ventralis; well-developed medial epicondyle and small, bi-tubercu-
lar lateral epicondyle; well-developed humeral ball, flanked by
prominent lateral crest leading from shaft to proximal tubercle
on lateral epicondyle, and by stronger crista medialis leading to
medial epicondyle. Variable development of flattened area for
muscular attachment on crista medialis probably reflects a sexually
dimorphic feature not uncommon in frogs.

Maxilla (Fig. 2c-f) with broadly-expanded anterior end and
prominent nasal process; posterior end expanded, pointed at its
ventral tip and with notch medially for quadratojugal; dorsally an

BREVIORA

No. 328

expansion and notch for squamosal; strong pterygoid process medi-
ally; teeth numerous, small, probably pedicellate; tooth row ex-
tending posterior to pterygoid process, external surface covered
with relatively fine pustular sculpture.

Fig. 2. Scotiophryne piistulosa, n. g., n. sp.: a, lateral, and b, medial
views of right squamosal, AMNH 8102; c, medial, and d, lateral views of
posterior part of left maxilla; e, lateral, and /, medial views of anterior part
of left maxilla, MCZ 3626; a - d from loc. V5620, Lance Formation,
Wyoming; e - f from Bug Creek Anthills, Hell Creek Formation, Montana;
all X 6.

Squamosal (Fig. 2a-b) compact; tympanic process expanded,
rounded, with angle at ventroposterior corner; pustular sculpture
as on maxillae; prominent pterygoid-paroccipital crest medially;
maxillary process with medial flange.

Discussion. Ilia of discoglossids are quite distinctive, and the
family reference of Scotiophryne is based on the similarity of the

1969

FOSSIL DISCOGLOSSID FROG

ilium to that of Recent Eurasian Bombina. The relatively large,
protuberant acetabulum and weak enlargement of the iliac sym-
physis region also resemble the relatively better developed, similar
features of the Recent Philippine genus Barbour ula (Estes, 1964;
Hecht and Hoffstetter, 1962). The humeri also show general
similarity to Bombina in shape of ball, epicondyles, oblique ole-
cranon scar, and dimorphism of crest development. The texture
of the sculptured skull elements is distinctive, but dermal sculpture
is also known in fossil discoglossids from the mid-Cenozoic of
Europe (Latonia, Zaphrissa ; Friant, 1960).

The ilium was chosen as the type specimen because it is more
often recovered than the relatively more delicate skull elements.

b.

f.

Fig. 3. Scotiophryne pustitlosa, n. g., n. sp.: above, dorsal, and below,
ventral views of distal end of three humeri, MCZ 3625; a - b, right; c - / left;
a - d male?, e - f female?; all X 6; Bug Creek Anthills, Hell Creek Forma-
tion, Montana. Medial condyle of e - f broken, cf. Fig. 4 b - c.

BREVIORA

No. 328

The cranial, girdle, and limb parts referred to Scotiophryne are the
most frequently-occurring frog elements in the Bug Creek sample.
Three other types of frogs are also present, but are relatively rare
in comparison; the Scotiophryne assemblage is probably a nat-
ural one based both on numerical and morphological factors.

The Princeton University specimens from the Middle Paleocene
Tongue River Formation are very similar to those of Scotiophryne,
and although worn and broken, are probably referable to this
genus (Fig. 4). The other lower vertebrates from this locality
closely resemble those of Bug Creek and Lance local faunas, and
represent a similar flood-plain ecological association. A dimor-
phism (probably sexual) similar to that in the Bug Creek speci-
mens is also shown by the Tongue River humeri.

Fig. 4. cf. Scotiophryne pustulosa: a, lateral view of left ilium, PU 17037;
b, ventral, and c, dorsal views of left humerus (female?). PU 16827; Tongue
River Formation, Montana; all X 6.

Lance Formation specimens of this small, distinctive frog appear
in material collected by the American Museum of Natural History
and Museum of Paleontology, University of California, Berkeley.
Frog remains from the Lance are more rare than from Bug Creek,
but elements referred to Scotiophryne are also the most frequent
in the Lance sample; this provides additional evidence that the
association made here is the correct one. The only known squamo-
sal referable to this species is AMNH 8102, which I figured and
described as "near Leptodactylidae?" in 1964 (p. 61, fig. 32).
The additional specimens described here indicate that such an
identification is no longer possible. The ilium questionably re-
ferred to Ascaphidae by me (1964, p. 55, fig. 32) is probably
from a small individual of Scotiophryne.

Scotiophryne is the second North American fossil discoglossid
to be reported (the first was cj. Barbourula, Estes, 1964). It

1969 FOSSIL DISCOGLOSSID FROG 7

resembles the Recent Eurasian discoglossid Bombina in some
girdle and limb features, but has a distinctive sculpture of the
dermal head casque. In having a broad, expanded squamosal with
a tympanic process that has a ventroposterior angle, Scotiophryne
resembles Zaphrissa (Friant, 1960) from the Oligocene of Ger-
many, but the sculpture type of the latter is not pustular. The evo-
lutionary history of Scotiophryne is unknown, and further comment
is postponed pending more detailed studies of other fossil dis-
coglossids.

ACKNOWLEDGMENTS

I thank Drs. Zdenek Spinar and Max Hecht for comments. The
drawings are by Mr. Laszlo Meszoly, except for Figure 4, which
is by Mr. Howard Hamman. This research was supported in part
by National Science Foundation Grant GB-7176.

LITERATURE CITED

ESTES, R.

1964. Fossil vertebrates from the late Cretaceous Lance Formation,
Eastern Wyoming. Univ. Calif. Publ. Geol. Sci., 49: 1-180,
73 figs., 6 tables, 5 pis.

EsTES, R., AND J. TlHEN

1964. Fossil vertebrates from the Valentine Formation of Nebraska.
Amer. Midi. Nat. 72: 453-472, 5 figs,

Friant, M.

1960. Les Batraciens anoures. Caracteres osteologiques des Disco-
glossidae d'Europe Acta Zool. 41: 113-149, 12 figs.
Hecht, M.

1963. A reevaluation of the early history of the frogs. Part II. Syst.
Zool. 12: 20-35, 7 figs.
Hecht, M., and R. Estes

1960. Fossil amphibians from Quarry Nine. Postilla, Yale Peabody
Museum, No. 46: 1-19, 3 pis.
Hecht, M., and R. Hoffstetter

1962. Note preliminaire sur les amphibiens et les squamates du
Landenien superieur et du Tongrien de Belgique. Bull. Inst.
Roy. Sci. Nat. Belgique 38: 1-30.
Sloan, R., and L. Van Valen

1965. Cretaceous mammals from Montana. Science, 148: 220-227, 6
figs., 1 table.

(Received 8 May 1969.)

■sn v%r^

DO NOT ^ir.O^^^^^^^^Ki

B R E V I O R A^i^

Mesemim of Coimparative Zoology

Cambridge. Mass. 15 September, 1969 Number 329

AMIA ( = KINDLEIA) FRAGOSA

(JORDAN), A CRETACEOUS AMIID FISH, WITH NOTES

ON RELATED EUROPEAN FORMS'

Richard Estes and Paul Berberian

Abstract. Numerous well-preserved disarticulated specimens from the
late Cretaceous Hell Creek Formation of Montana show that Kindleia
fragosa Jordan, a late Cretaceous and Paleocene amiid from North
America, is referable to the Recent genus Amia. A. fragosa resembles the
middle Eocene A. giirlexi from North America, the late Paleocene A. rus-
selli from France, the middle Eocene A. kehrcri from Germany, and the
Oligocene A. nuiiiicri from France in a number of minor but consistent
skull features. A. kclireri and A. giirleyi are relatively shorter-bodied forms
than A. calva; the two other species were probably also short-bodied, but
are known only from disarticulated or incomplete material. Most differ-
ences from A. calva shown by the fossil forms are minor and probably
primitive. These morphological differences, and close relationship of
Cretaceous seas of the western interior to the Mississippi River drainage
of today, indicate that A. fragosa is not far from the ancestry of A. calva.

INTRODUCTION

Jordan ( 1927, p. 145) described a fossil fish from the Edmon-
ton Formation (late Cretaceous, Alberta) as Kindleia fragosa, re-
ferring it to the Cichlidae. Estes (1964) demonstrated that it was
an amiid, utilizing extensive material from the Lance Formation
of Wyoming, and synonymized Stylomyleodon lacus Russell
( 1928) on the basis of specimens including the type bones of both
described species. Russell (1968) has apparently not accepted
this conclusion. Recently, Janot (1966, 1967) has described new
European amiid material and has suggested that features shown

1 Fossil vertebrates from the late Cretaceous Hell Creek Formation,
Montana: Contribution No. 5.

2 BREVioRA No. 329

by Lance Formation amiids indicate no more than specific dis-
tinction from Amia. We agree with this conclusion, which is doc-
umented below with some necessary qualifications, and Kindleia
is here included in the synonymy of Amia.

Recent collections from the Hell Creek Formation of Montana
(Sloan and Van Valen. 1965) have produced more material of
Amia jragosa. This study is based on a Museum of Comparative
Zoology collection made in 1964 by A.D. Lewis and party, from
Bug Creek Anthills, west half of section 9, T 22 N, R 43 E, Mc-
Cone County, Montana. Supplementary material from the same
locality was provided by Dr. Robert Sloan (University of Minne-
sota). An American Museum of Natural History collection from
the Lance Formation (localities V5711 and V5620: Estes, 1964;
Clemens, 1963) was also utilized in this study. Almost all skeletal
elements have been recovered in large numbers (except scales,
in contrast with the Lance Formation collection). At least 94
individuals are represented in the MCZ Bug Creek Anthills sample.

Estes (1964) studied this species in detail, and we therefore
discuss Bug Creek specimens only as they modify conclusions
reached by that study. Order of discussion of bones follows that
of Estes.

ORDER AMIIFORMES

Family Amiidae

Amia fragosa (Jordan, 1927)

Kiiulleia fra^osa Jordan. 1927. p. 125
Styloinyleoihm huiis Russell. 1928, p. 105

Basioccipital. Estes ( 1964) noted that Lance Formation basioc-
cipitals had only one fused vertebra rather than two as in Amia
calva. Twenty basioccipitals occur in the Bug Creek sample; nine
with one fused vertebra as in the six Lance Formation specimens,
and eleven with two fused vertebrae as in our six specimens of
Recent A. calva (Fig. 2). There is a weak tendency for fusion
of vertebrae to be correlated with increasing size in the Bug Creek
sample; since all of our A. calva are approximately the same size,
it is possible that such a variation exists in the Recent species as
well. However, Janot (1967) has shown that variation in this
feature occurs in her fossi! material and suggests that it is inde-
pendent of size-age variation, since even large bones may lack
the additional vertebra (ihid., pi. 12, fig. 2). Whatever the case,
we believe that the lack of a second fused vertebra in the six Lance

1969 AMIA FRAGOSA CRETACEOUS AMIID FISH 3

Formation specimens was a chance aggregation without taxonomic
significance.

The relatively short basioccipital and limited extent of the poste-
rior brain chamber impression on it were noted for Aniia fragosa
by Estes (1964, p. 29). These features are also visible on the
basioccipital figured by Janot (1967, pi. 12, fig. 6a), and differ
from the widely-open brain chamber impression and somewhat
longer basioccipital of A. calvci.

Pterotic. Estes (1964) stated that a parietal lappet on the
pterotic distinguished this species from A. calva. However, it is
present in some of our A. calva, and the condition is also variable
in the Cretaceous species. Pterotics of A. fragosa are consistently
shorter than those of A. calva, and have a relatively greater antero-
medial excavation for the frontal, reflecting the greater posterior
extent of the latter noted below.

Parietal. The parietals lack an opening for the sensory canal.
As Janot (1967) notes, this is different from the situation in A.
calva and is a specialization of A . fragosa. The parietal is approx-
imately square, rather than elongated anteriorly sls in A. calva.

Frontal. Estes (1964) noted that orbital excavation in Lance
Formation frontals was greater than in A. calva, and by comparison
with Sinamia suggested the presence of supraorbitals for A. fra-
gosa. Articulated specimens of A. kehreri (to be discussed below)
from the Eocene of Europe indicate that supraorbitals are lacking
in that related species, and they were probably also absent in
A . fragosa. However, the late Cretaceous amiid Enneles does have
supraorbitals (Silva Santos, 1960). The frontals appear to be
relatively longer in A. fragosa than in A. calva, an estimated 2.8
times the length of the parietals, as opposed to 2.4 for the Recent
species. This was determined in our disarticulated material by
matching parts of bones of similar widths and general proportions,
and by comparison with the related A. kehreri. The relatively
long frontal and short parietal proportion is a primitive character,
as judged by its presence in some other Mesozoic amioids (e.g.
Enneles, Megalurus).

Derniosphenotic. This bone was not identified in the Lance
Formation sample. Three specimens in the Bug Creek material
resemble those of A. calva but are less elongated anteriorly, re-
flecting the relatively larger orbit of A. fragosa.

Nasal. Estes (1964) noted no difference of nasals from A.
calva. Well-preserved Lance Formation specimens and the Bug
Creek specimens all indicate a slight bifurcation of the nasal around
the anterior nostril absent in our specimens of A. calva (Fig. 1).

BREVIORA

No. 329

Prenia.xilla. In 13 complete premaxillae, tooth count ranges
from 6-9 teeth [frequency 6( 1 ), 7(2), 8(9), 9( 1 )], as in ^. calva.

Vomer. Number of vomerine teeth was cited by Estes ( 1 964,
p. 32) as greater than in A. caiva. Counts based on Lance For-
mation and Hell Creek Formation specimens indicate a range of
18 to 24 teeth per vomer; range for the Recent species is from at

e.

Fig. 1. A mill ficij^osa: a. medial, and /', lateral views of MCZ 9291. left
operculum, unworn, unbroken; c, medial, and cl. lateral views of AMNH 9315.
left operculum, broken ventrally: e. dorsal view of right nasal, MCZ 9288:
/, lateral view of MCZ 9293, left interoperculum: g. ventral view of MCZ
9286. gular; all X 1. AMNH 9313 from Lance Formation. Wyoming: all
others from Buc Creek Anthills, Hell Creek Formation. Montana.

1969 AMIA FRAGOSA CRETACEOUS AMIID FISH 5

least 15 to 27, bracketing the fossil count. Increased vomerine
tooth count must thus be removed from the diagnosis of A. jragosa,
but the vomerine tooth patch in A. jragosa always extends farther
posteriorly than in the Recent species, as already indicated by
Estes ( 1964, fig. 17a, and Fig. 3b, this paper).

Dentary. As in A. calva, the alveolar border makes a right
angle with the external face of the bone (Janot, 1967, p. 146).
Few dentaries are complete enough to allow tooth count, but
two have 14 and one 16 alveoli, about as in our A. calva sample.
As m A. russelli Janot, coronoid attachment area is deep ante-
riorly in relation to depth of the jaw, and depth of jaw is greater
proportionally than in A. calva.

Giilar. This bone was not recovered in the Lance Formation
material, but two gulars have been identified in the Bug Creek
collection. This is a variable bone in A. calva, but that of A.
jragosa is consistently shorter and less concave when compared
with bones of the same width in the Recent species (Fig. 1).

Operculum. The few fragmentary opercula from the Lance
sample were broken and abraded posteriorly, resulting in an in-
accurate restoration by Estes (1964). The Princeton specimen
thought by him to confirm the rectangular restoration given {ibid.,
fig. 16e) is Eocene rather than Paleocene as noted (/7^/V/., p. 33),
and is also broken, as more recent preparation has shown. Well-
preserved Bug Creek specimens show symmetrically-rounded
posterior borders with an obtuse point, similar to that of the
Eocene Amia ("Paramiatus") gurleyi (Romer and Fryxell, 1928),
A. kehreri, and A. russelli. Well-preserved Hell Creek Formation
and Lance Formation specimens also confirm this shape (Fig. 1).
A. calva usually has the blunt point in a more ventral position;
the ventroposterior border of the bone in the Recent form is thus
relatively more elongated. Compared with height, length of
operculum is less than in A. calva (Fig. 3).

Subopcrcuhun. For bones of the same anterior height, length
is somewhat less than in A. calva (Fig. 2).

Interoperculum. The same as for suboperculum (Fig. 1).

Supracleithrum. This bone was not identified in the Lance
Formation sample. A single specimen from the Bug Creek col-
lection is more robust than that of A. calva and has an external
surface sculptured like that of the cleithrum (as described by
Janot, 1967, pi. V, fig. 4a for A. russelli).

Comments. The large Bug Creek sample and the American
Museum of Natural History Lance Formation sample thus con-
firm the suggestion of Janot (1967) that Kindleia is a junior

BREVIORA

No. 329

Fig. 2. A/nia ira^usii, late Cretaceous, Bug Creek Anthills, Hell Creek
Formation. Montana: a. ventral view of MCZ 9290, basioccipital with one
fused vertebra: b. ventral view of MCZ 9289, basioccipital with two fused
vertebrae; c, lateral view of MCZ 9292, right suboperculum. Abbreviations:
as = aortal supports, pas = parasphenoid articulation surfaces; sf = fora-
mina for spinal arteries: all x 3.

synonym of Amia. Operculum shape, vomerine tooth count, and
fusion of vertebrae to the basioccipital resemble conditions in
Recent Amia calva more than was indicated by Lance Formation
specimens studied by Estes (1964). However, these and other
features indicate specific difference from A. calva. A revised
species diagnosis is as follows:

1 . Amiids with an estimated range of body length about the
same as in Recent A. calva, as indicated by comparison of
disarticulated elements with those of the Recent species.

2. Dermal bones about 1.5 times as thick as in A. calva.

3. Frontals about 2.8 times length of parietals.

4. Relatively large postorbitals, probably filling cheek region;
lower postorbital much larger than upper postorbital; ver-
tical pit line present on lower postorbital.

5. Marginal teeth simple, pointed cones, palatal teeth usually
stout, styliform crushers.

6. Operculum with bluntly-pointed posterior border; opercular
series relatively short anteroposteriorly.

7. Supraorbital sensory canal not entering parietal.
Related forms. Janot (1967) has agreed with Estes (1964, p.

41 ) that A. numieri of the Olicoccnc of France is closely related

1969 AMIA FRAGOSA CRETACEOUS AMIID FISH 7

to A. frui^o.so. Another related form is A. kehreri from the middle
Eocene of Germany. Complete articulated specimens of this
species are in the British Museum (Natural History), collected
by Walter Kuhne (BMNH P33480, P33488, Messel bei Darm-
stadt), and in the Museum fUr Mitteldeutsche Erdgeschichte, Halle
(Saale), collected by Dr. Horst Matthes in the Geiseltal deposits
(fig. 4; pi. 1). These specimens conform to the species diagnosis
given above for A. jragosa, although the apparent absence of the
supraorbital canal in the parietal cannot be confirmed without
disarticulated material. In addition, they have only about 50-55
vertebral segments (counting diplospondyl centra as one pair per
segment). This low number of vertebrae is also seen in Amia
(" Paramiatus" ) gurleyi Romer and Fryxell (1928). A. gurleyi is
less distinct than its describers believed, and while relatively shorter
than A. calva, appears "deep-bodied" primarily as a result of the
shorter body and of crushing. Its opercular series is very similar
to that of A. jragosa and it has similar frontal-parietal proportions.

As noted above, Amia russelU Janot ( 1966) from the late Paleo-
cene of France is also close to this group in several characters.
The parietal Janot figures {ibid., pi. IX, fig. 5) is about as wide
as long, contrasting with that of A. calva (cf. e.g. Janot, 1967,
pi. IV, fig. 6) and resembling that of A. kehreri, A. mimieri, A.
gurleyi, and A. jragosa. The frontal figured for A. russelU (Janot,
pi. IX, fig. 3) is about 2.8 times as long as the figured parietal,
and the latter is of about the proper size to fit the frontal. Orbital
excavation in the frontal is also similar to that of A. jragosa.
Frontal-parietal proportions of A. russelU thus seem to have been
similar to the three species noted above. Operculum shape of
A. russelU is generally similar to that of Cretaceous and Eocene
species, being taller than wide, although the posterior angle is
slightly below the middle of the bone, as in ^4. calva.

The relatively narrow proportions of the opercular series of the
fossil species is a primitive character for amiids. In A. calva the
series has widened as a result of general body elongation, although
the operculum itself is always wider than or as wide as high regard-
less of size of the animal (Fig. 3, c-e).

It is thus clear that the late Cretaceous and Paleocene species
discussed above are closely related, and the similarities indicate
that litde evidence exists for maintaining separate species A. jra-
gosa, A. russelU, A. gurleyi, and A. kehreri. The oldest available
name for the species discussed here is A. kehreri Andreae ( 1892).

8

BREVIORA

No. 329

Fig. 3. a. Amid calva. ventral view of left vomer. /', A. fra^osa, the same.
Hell Creek Formation, Montana, MCZ 9287. c-e, A. calva. medial views
of left opercular series; sizes: c. MCZ 8970, Standard I.ent;th 45 mm. d,
MCZ 35780, SL 155 mm, c. unnumbered MCZ specimen. SI. 425 mm. /, A.
fragosa, restoration of opercular series. Note that in c-c operculum is
wider than tall, whereas in /, Figure 4. and Plate I, it is taller than wide.
a-h. X 3; e-f, X 1 ; (-</ not to scale.

1969

AMIA FRAGOSA CREl ACUOUS AMIID FISH

o

00

a.

X
Z

C/5

•a

Q

5!

3

o
X

XJ

O

c

O

W

•a

t3

■it:

C
^

c

t/5

Xi
C

HJ

03

D-

«

O
O

x:

Cl.

a

10 BREVIORA No. 329

We believe that A. kehreri and A. jragosa may be synonymous;
temporal and geographic separation are the only known differen-
tiating characteristics. The type of A. kehreri is from Messel, as
are the British Museum (Natural History) specimens of this
species noted above. The type is a partial skeleton, but shows the
expanded lower postorbital (Andreae, 1895, pi. 1, fig. 15) more
clearly seen in BMNH P33480 (Fig. 4) and the Geiseltal speci-
mens (PI. 1). A. gurleyi (Romer and Fryxell, 1928, fig. 1) also
resembles A. kehreri in close approximation of dorsal and caudal
fins, and this may be an indication that synonymy of A. gurleyi
with A. kehreri is in order. Since both are middle Eocene, only
geographic separation and a few minor details suggest that two
species be maintained. Presence of the supraorbital canal in the
parietal, somewhat larger size, and a few other superficial features
seem to distinguish A. russelli. We do not formally synonymize
any species here, however, until other early and middle Cenozoic
European specimens can be studied; our purpose is merely to
point out the close similarity of these Cretaceous and early Ceno-
zoic forms (see also Estes, Hecht, and Hofstetter, 1967).

There is thus a closely related group of species of Amia (some
probably synonymous) that is known from Cretaceous through
middle Eocene of North America, and late Paleocene to at least
early Oligocene of Europe. This group of species is distinct from
A. calva only in superficial and essentially primitive ways, includ-
ing relatively shorter body and skull and minor proportional dif-
ferences of skull roof and mandibular bones.

The morphology of A. jragosa and the close relationship of
North American Cretaceous seas to the Mississippi River drainage
(Estes, 1964) indicate that A. jragosa itself cannot be far from
the ancestry of the Recent bowfin. Further studies of European
amiid remains, of articulated specimens of "Protamia" at Prince-
ton University (Estes, 1964, p. 42), and of growth series of A.
calva will be of considerable interest in tracing the ancestry of
the modern species (see also Simpson, 1937, p. 59; specimens
lost).

Stratigraphic range oj Amia jragosa. Regardless of possible
synonymy with European species, as noted above, the stratigraphic
range of A. jragosa is remarkably long, extending from late Creta-
ceous through at least middle Eocene time in North America. No
criteria exist at present for naming more than the one species.
The major deposits in which remains of A. jragosa have been
found are summarized in Table I. Unpublished records are taken
from collections at the American Museum of Natural History,
Princeton University, and the United States National Museum.

1969 AMIA FRAGOSA CRETACEOUS AMIID FISH 11

ACKNOWLEDGMENTS

We thank Drs. C. Janot, Charles Meszoely, and Donald Baird
for data and helpful discussion. This research was supported in
part by National Science Foundation Grant GB-7176. to Estes.

REFERENCES

Andreae, a.

1892. Vorlaufige Mitteilung Liber die Ganoiden {Lcpidosteus und Amiu)

des Mainzer Beckens. Verb. Nat. med. Ver. Heidelberg (N.F.)

5: 7-15.
1895. Beitrage zur Kenntniss der fossilen Fische des Mainzer Beckens.

Abb. Senck. naturf. Ges. 18: 351-365.

Clemens, W.

1963. Fossil mammals of the type Lance Formation, Wyoming.
Part I. Introduction and Multituberculata. Univ. Calif. Piibl.
Geol. Sci. 48: 1-105.

Estes, R.

1964. Fossil vertebrates from the late Cretaceous Lance Formation,
eastern Wyoming. Univ. Calif. Publ. Geol. Sci. 49: 1-180.

Estes, R., M. Hecht, and R. Hoffstetter

1967. Paleocene amphibians from Cernay, France. Amer. IMus.
Novitates No. 2295: 1-25.

Janot, C.

1966. Ainia nisselli nov. sp., nouvel Amiide (Poisson holosteen) du
Thanetian du Berru, pres de Reims. C. R. Soc. Geol. France
1966(3): 142.

1967. A propos des amiides actuels et fossiles. Colloq. Intern.
C.N.R.S. 163: 139-153.

Jordan, D.

1927. Kiiulleia. a new genus of cichlid fishes from the Upper Creta-
ceous of Alberta. Canad. Field Nat. 41: 145-147.

ROMER, R., AND F. FRYXELL

1928. Paramiatus giirlexi, a new deep-bodied amiid fish from the
Eocene of Wyoming. Amer. J. Sci (5) 16: 519-527.

Russell, L.

1928. A new fossil fish from the Paskapoo beds of Alberta. Amer.
J. Sci. (5) 15: 103-107.

1968. Palaeontology of the Swan Hills area, North-Centra! Alberta.
Roy. Ontario Museum Contrib. No. 71: 1-31.

SiLVA Santos, R. da

1960. A posigao sistematica de Enneles aiida.x Jordan e Branner da
Chapada do Araripe, Brasil. Minist. Minas Energ.. Div. Geol.
Min. Brasil, Mon. 17: 1-25.

12 BREviORA No. 329

Simpson. G.

1937. The Fort Union of the Crazy Mountain Field, Montana, and
its mammalian faunas. Bull. U. S. Nat. Mus. 169: X -^ 287 pp.
Sloan, R., and L. Van Valen

1965. Cretaceous mammals from Montana. Science 148: 220-227.

(Received 8 Mav 1969.)

969

AMIA FRAGOSA CRETACEOUS AMIID FISH

13

■^'

-jsi.^:.

-c■^.^if^*S

-i*

J

m'.

■ i. ■ ' ■

.-r

'-«l

Plate I. Amia kehreii. middle Eocene, Geiseltal deposits. Above, skull
of one individual in lateral view, showing expanded lower postorbital,
opercular series, and frontal-parietal proportions; below, skull of another
individual in dorsal view showing frontal-parietal proportions; X about 1 .

14

BREVIORA

No. 329

WYOMING

MONTANA

S. DAKOTA

CANADA

o
o

L
EOCENE M
E

Bridger Fm.,
Green River Fm.,
Wasatch Fm.

Golden
Valley Fm.

o

LU

L

PALEOCENE M
E

Rock Bench
beds

Mantua
lentil

E

Ll_

.0

'c

0

Ll_

Melville beds

Tongue River
member

Tullock
member

Paskapoo

Fm.

o

UJ

MAESTRICHTIAN

Lance Fm.

Hell Creek Fm.

Edmonton
Fm.

o
<:

LU

o

CAMPANIAN

"Mesaverde" Fm.

Judith River
Fm.

Oldman Fm

Table I. Major deposits carrying remains of Amia fragosa in the West-
ern Interior of the United States and Canada. Nomenclature of units varies
with author; we have therefore deliberately chosen a conservative termi-
nology.

]-yA- C [arol.:^^-": DO NOT cmcuLA-TC

BREVIORA

Miaseem of Compsirative Zoology

Cambridge, Mass. 15 September, 1969 Number 330

THE CRANIAL EVIDENCE FOR HYBRIDIZAifofil^^
NEW ENGLAND CANIS -'^^^^

US^.CO.MP. ZOOL
RY

Barbara Lawrence SEP 1 1 303

and
William H. Bosserfi HARVARD

UNIVERSITY

Abstract. Using the technique of linear discrimination to compare known
dog-coyote hybrids, it is shown that skulls of these animals have a mean
discriminant function value almost exactly between those of the two parent
stocks.

Applying this same technique to the canids which are presently invading
the empty predator niche in New England, it is shown that this population
differs from the known hybrids. They are predominantly coyote and evi-
dence is given showing that they probably have some dog and wolf genes as
well. The New England animals are an extreme expression of a trend
already apparent in Canis latrans thamnos from Minnesota. The high degree
of variability demonstrated is evidence that the shift away from coyoteness is
the result of hybridization rather than of a rapid evolution to fit a new niche.

INTRODUCTION

Having shown (Lawrence and Bossert, 1967) that the three
species of the genus Canis: lupus, latrans, and jamiliaris, can be
clearly and significantly distinguished by the technique of linear
discrimination, the question arises as to whether or not this same
technique can be used to identify hybrids of these species. Part One
of the present paper discusses a linear discrimination study of known
latrans X jamiliaris hybrids; Part Two discusses the application of
both this study and our earlier work to the unraveHng of the ancestry
of the canids that have recently been moving into the empty predator
niche in New England. As in our earlier paper (1967), the meas-
urements used for this analysis were the fifteen found to be most

1 Division of Engineering and Applied Physics and Department of Biology,
Harvard University.

2 BREVIORA No. 330

diagnostic and the technique of using the linear discriminant func-
tion to reduce the multiple measurements to a single value for each
specimen was applied.

PART I
Known Hybrids

Specimens examined. Ten Fi hybrids were measured. Five of
these were from the collections of H. F. Gier at Kansas State Uni-
versity: two of a Labrador retriever and one of a collie crossed
with two different female coyotes, two of a coyote crossed with a
basenji bitch. Five more were from the collections of the University
of Kansas and were offspring of a small, mongrel terrier bitch
crossed with a coyote. A few additional specimens, unusuable be-
cause they were either broken or not adult, seemed by eye to fall
within the range of the above.

Twelve other specimens from the Gier collection, the results of
variously breeding the original hybrids amongst themselves or back
to a springer spaniel, are treated separately.

Comparisons. The discriminant function values of known Fi
latrans X jamiliahs hybrids were calculated, on the basis of the
pairwise discrimination of the two species described by Lawrence
and Bossert (1967). The resulting values are intermediate between
those for each species. Their range of variation is rather wider
than it is for each of the parent species, but there is no overlap with
either. One specimen at each end of the range is within three
standard deviations of each parent form; otherwise, the Fi hybrids
cluster around a point midway between the two species. Appar-
ently then, if the question is one of hybridization between two
known species, this technique, in the majority of cases, will prop-
erly show the intermediate position of individual specimens. To
what extent it would also suggest a relationship to other species
was next considered.

The difference between dogs and wolves, as was shown earlier
(Lawrence and Bossert, 1967), is considerably less than between
any other pairs of species. For this reason, to determine how wolf-
like these dog-coyote hybrids are, discriminations were also tried
using first the latrans-lupus discriminant functions and then the
familiaris-liipus. In both instances, where only one of the paired
species was actually involved in the ancestry of the hybrids, the
distribution of values overlapped the values for both species used
in the discrimination. While the latrans X jamiliaris tended to be
coyotelike rather than wolflike in the first discrimination, in the

1969

NEW ENGLAND CANIS

second they were more evenly distributed between dog and wolf,
with the majority of the specimens actually intermediate. These
relationships are shown in Figure 1 .

6.0r

Fig. 1. Linear discriminant values of known dog-coyote hybrids (H). The
latrans-familiaris discriminant function is used as the abscissa and the latrans-
lupus discriminant function is used as ordinate (from Lawrence and Bossert,
1967). The contours are extreme ranges of individuals of C. latrans (C),
C. lupus (W), and C. familiaris (D) used in computing the discriminant func-
tions.

Fo skulls were also studied. These were the result of subsequent
crossing in various combinations, using the collie and labrador
hybrids as well as a male springer. Discriminant functions were
evaluated for these as for the Fi hybrids. All fourteen proved to
be intermediate between latrans and familiaris but showed a larger
proportion falling within three standard deviations of one or the
other. Interestingly, in the latrans-lupus discrimination they differ
from the Fi series in being uniformly latrans- like, though in the
jamiliaris-lupus discrimination they coincide exactly with the Fi
series. The Fo relationships are shown in Figure 2.

The evidence then is that, while the discriminant functions of
both groups of hybrids are intermediate between those of the two
parent stocks, further discrimination to see whether the trend
towards dogness in particular individuals is expressed as wolfness

BREVIORA

No. 330

in a latrans-lupus discrimination is negative. The most doglike
specimens fall within the range for latrans in this latter discrimina-
tion and the few specimens which are wolflike are exactly inter-
mediate in the latrans-jamiliaris discrimination.

6.0r

Fig. 2. Linear discriminant values of variously bred F2 dog-coyote hybrids
(G). The coordinate axes and contours are identical to those of Figure 1.

In summary, it can be said that the technique of linear discrim-
ination can be useful for identifying hybrids between two known
ancestors on the basis of multiple characters. Specimens in this
category may be expected to fall between the two parent stocks.
Referring specifically to Canis, the population of known Fi hybrids
studied is characterized by having a mean latrans-jamiliaris dis-
criminant function value ( — 16.3) almost exactly between that for
latrans ( — 14.6) and that for faniiliaris ( — 17.8). The Fo generation
is also intermediate, but the discriminant value ( — 16.6) tends to-
ward that of jamiliaris.

PART II
New England Canis

Specimens examined. Twenty-two animals, offspring of siblings,
dug from a den near Croydon, New Hampshire, were studied; of
these, sixteen were included in the multiple character analysis. A
sibUng of the parents was also included, although the parents

1969 NEW ENGLAND CANIS 5

themselves were not, as they have been kept alive for breeding.
This entire series is referred to as the Boscawen series.

Of the animals collected in the wild, fifteen were suitable for
inclusion in the multiple character analysis and came from the fol-
lowing localities: New Hampshire, Croydon, 1 male; Temple, 2
females; Lancaster, 1 male; Haverhill, 1 male. Vermont, Wards-
boro, 2 females, 2 males; Brookline, 2 males. Massachusetts, Otis,

1 female; Colrain, 1 male; Leyden, 1 male, 1 female. Of these, all
but the animals from Lancaster and Haverhill were typically wild
in external characters. An additional fifteen specimens were either
subadult or too broken for inclusion but were studied and com-
pared with the first series. These include from New Hampshire:

2 from Croydon, 3 from Wiiton, 1 from Whitefield and 3 from
Colebrook, as well as six from Vermont: 1 each from Newfane,
West Dummerston, Townsend, Jamaica, North Bridgewater, and
Hereford.

Comparisons. Discriminant functions of skulls of the animals
raised in captivity and the wild shot individuals were similarly
evaluated in order to determine whether or not this population had
the characteristics of the known hybrids. While the population as
a whole was found to be somewhat intermediate between latrans
and jamiliaris on the one hand, there was considerable overlap with
latrans on the other. Even the most doglike is widely separated
from jamiliaris. The mean discriminant function ( — 15.2) falls rather
close to the range for latrans, and the Mahalonobis D- distance
statistic between latrans and the New England animals is less than
two-thirds of that between the latter and jamiliaris. On the latrans-
lupus discrimination, the population is more completely intermedi-
ate. While the overlap with latrans is less, a number of specimens
approach lupus rather closely (see Fig. 3). The average specimen
is close to halfway between the two and the D- distances are about
the same. On the lupus-jamiliaris discrimination the unknowns are
again more lupus-Viko. than the known hybrids. All, except one
intermediate specimen, have values which fall within the range for
lupus, vv'hereas the known hybrids are predominandy intermediate
and overlap about equally with both jamiliaris and lupus. These
pairwise comparisons are shown in Figure 5.

The conclusions that can be drawn from these comparisons are
that the unknowns differ from all three species and that they re-
semble coyotes more closely than the known hybrids do. They
also are more wolflike. Since dogness in known coyote-dog hybrids
se'dom shows up as wolfness in a latrans-lupus discrimination, it

BREVIORA

No. 330

6.0

5.2

4.4-

3.6

2.8-

2.0

N N ^% %-%
N N N "^ ^
N ^ N

■14

Fig. 3. Linear discriminant values of New England Conis (N). The co-
ordinate axes and contours are identical to those of Figure 1.

6,0 r

52

44

3.6

2.8

2.0

.._C ,

-18

-17

-15

-14

Fig. 4. Range of the linear discriminant values for latrans (dotted), latrans
thumnos (dashed), and New England Ccinis (line). The coordinate axes are
identical to those of Figure 1. Individual values for latrans thamnos specimens
are shown (M).

1969

NEW ENGLAND CANIS

is unlikely that the trend of the unknowns towards lupus can be
attributed entirely to an increase of dog genes in this series.

In an effort to determine what might be responsible for the dif-
ferences between the known hybrids and the unknown animals, a
population of coyotes from the eastern fringe of the range of latrans
was analyzed. Using the three pairwise discriminations described
above, the discriminant functions of this series of thirty-two
C. latrans thamnos from Minnesota were calculated and compared,
not only with the original three series but also with the known
hybrids and the New England population. While the Minnesota
population overlaps strongly with latrans and has a D- distance

a)

-18

-17

I

•16

■15

I

■14

I

familiaris

latrans

b)

2
1

3

I

4

I

i I iiiiiiiiHiMm

lupus

latrans

c)
-6

-5.5

-4.5

mil 111 Himmiiiiii

familiaris

lupus

Fig. 5. One dimensional linear discriminant values of New England Cams:
(a) on latrans-faiuiliaris axis, (b) on latrans-hipus axis, (c) on lupiis-famili-
aris axis. The range of the populations used to compute the discriminant func-
tions are indicated by brackets.

8 BREVIORA No. 330

from it that is not significant, six individuals fall well outside the
range of variation as determined (Lawrence and Bossert, 1967)
for the latter. Five of these are intermediate towards lupus in the
lupus-latrans discrimination and only one towards jamiliaris in the
latrans-jamiliaris discrimination.

Further, tJmmnos is intermediate between latrans and the New
England population and, while it overlaps strongly with the former,
it overlaps almost equally strongly with the latter, the D- differ-
ences between thamnos and each of the others being about the
same. The trend away from typical coyote and towards both lupus
and jamiliaris, already apparent in the Minnesota thamnos, has
progressed considerably farther in the New England population,
with the D- distance between the latter and both lupus and jamili-
aris much reduced, and a suggestion that the trend is slightly more
towards jamiliaris. The trend in the discriminant values is shown
in Fig. 4.

These comparisons suggest the possibility that the divergence
from the typical coyote pattern, culminating in the New England
population, is, in part at least, caused by some mixing with wolf
as well as with dog stocks. Further evidence is provided by the
rather high degree of variability found in the two not especially
isolated or reduced populations. There are several meaningful,
quantitative measures of total variability in a multivariate popula-
tion. We have used two that depend on the volume of the ellipsoid
of variation; they are the sum of the principal axes of variation
and the product of the ten largest principal axes of variation (see
Cramer, 1946, p. 406). The two measures are consistent over
the canid populations considered here for ranking the populations
as to variability and for demonstrating relative differences in vari-
ability. These variabilities are given in Table 1. The D- distances
between populations are given in Table 2. They show that when
coyotes are compared with wolves and dogs, the within-group
variation is relatively small and the between-group distances are
relatively large, whereas the reverse is true when thamnos or the
New England Canis are compared with wolf and dog.

In summary: although the multivariate analysis does not provide
definite proof of the genetic composition of the New England
population, a number of points may be deduced from it. The dif-
ferences between the New England population and the known
dog-coyote hybrids are sufficient to show that the former are not
"coydogs." It establishes that they are, in fact, predominantly
coyote, and that they are not a purely local phenomenon but are

1969

NEW ENGLAND CANIS

TABLE 1

Multiple character variability within populations: Two measures of
the scatter of the cranial proportions in multivariate space.

Population

Sum of principal

axes of variation

(X 10'')

Product of ten

largest axes of

variation

(X 10'^)

C. familiar is

3.40

13.5

C. lupus

1.00

0.14

C. latrans

0.66

0.02

C. latrans thamnos

1.28

1.42

New England Canis

1.69

8.22

TABLE 2
Mahalonobis D- distance statistics between populations.
C. familiaris

C. lupus

27.2

C. lupus

C. latrans

119.9

64.1

C latrans

C. latrans thamnos

—

—

6.7 1 C. latrans thamnos

New England Canis

44.55

29.84

26.83 9.12

extreme examples of a progressive change that had already begun
on the eastern periphery of the coyote's distribution.

Discussion. In external appearance, the specimens under con-
sideration are not unlike large coyotes. They have agouti hair,
and the rather common, wild-canid color pattern found in varying
degrees of intensity in all coyotes, in some wolves, and approxi-
mated in some dogs. The tail is carried straight, not curled up at
the tip, and has the rather bottle-brush appearance characteristic
of both coyotes and wolves and quite distinct from that of dogs.
The ears are always erect, and vocalization and smell are undoglike
(Silver and Silver, in press). On the grounds of external appearance
alone, it might seem a justifiable assumption that the New England
animals are examples of a rapid evolution of a race of coyotes
characterized by large size and more powerful teeth suited to preying
on large mammals.

10 BREVIORA No. 330

The multiple character analysis, which shows a definite trend
away from coyote towards both dog and wolf, slight in the Minne-
sota animals and more extreme in the New England population,
could also be interpreted this way. Such a composite picture of the
characters, however, masks the extent to which usually diagnostic
features of different species may be developed in the same indi-
vidual. These combinations of non-homogeneous characters
strongly suggest multiple ancestry.

In considering cranial variations, it is important to keep in mind
that the animals raised in captivity in Boscawen, offspring of a
single pair of adults, have a common genetic background while
the wild shot individuals come from a scattered, spottily distributed
population which may be anything but homogeneous.

Individual skulls of the Boscawen series (Fi litters) vary from
coyotelike (cf. MCZ 51726, 27) to ones which are coyotelike in
narrowness of rostrum, shape of brain case and interorbital region,
but are uncoyotelike (cf. MCZ 51865, 66) in their widely spread-
ing zygomatic arches, reduced bullae, shortened rostrum, and some-
what elevated forehead with a well-developed frontal concavity
between the broadly spreading frontal processes. Most agree in
having length of jaw incompatible with size of teeth and, in many,
the premolars actually overlap. All have M2/well developed as in
coyotes and, in general, the cusps of the molars are rather blunt
and rounded as in dog/wolf. Otherwise, variation in P/4 and M/1
and in P4/ and C/ spans the range between coyote and dog/wolf.

The wild shot specimens examined, setting aside those that on
the basis of external characters were at least part dog, are similarly
heterogeneous. Some have a broad brain case and zygomatic
arches, combined with a narrow rostrum. A number have the pre-
molars crowded and overlapping but none have the rostrum as
shortened, relative to the size of the teeth, as do some of the Bos-
cawen series. In addition, a given tooth row often combines,
interestingly, coyote with wolf/dog characters. The most coyote-
like of the characters are the rather uniformly large inner portion
of Ml/ and the large size of M2/. These are combined with short,
broad canines, a reduction of the metaconid of M/1, upper car-
nassials that tend towards wolf/dog not only in plumpness but
also in the slight development of the deuterocone, and an overall
massiveness of the teeth in relation to size of skull.

The most conspicuous and possibly significant variation is in the
size of these wild shot animals. The most wolflike one and one of
the most coyotelike, as shown by the multiple character analysis,
are also at opposite extremes in total size and, more particularly,

1969 NEW ENGLAND CANIS 11

in size of teeth. Both specimens were shot in Leyden, Massachu-
setts, and apparently belonged to the same small pack.

In the series as a whole, the largest skulls are at the upper ex-
treme of reported size for coyotes and in certain dimensions,
notably zygomatic width and width across condyles, are actually
larger than reported coyotes, (Young & Jackson, 1951). These
specimens often resemble closely skulls of some of the southern,
probably hybrid, animals sometimes called Canis niger gregoryi.
They are almost equally close in size and general appearance to
small specimens of the northern timber wolf, Canis lupus lycaon,
on whose range they have begun to impinge. The same is true of
certain tooth dimensions; extreme individuals have a massive upper
carnassial with a breadth to length ratio that falls outside of the
range for coyotes; even more conspicuous is the greater width of
the often shortened canine. These largest teeth again approach
certain of the specimens referred to above of so-called niger gre-
goryi and are close to those of some of the smallest wolves. Such big
teeth are not, however, necessarily associated with the biggest
skulls.

The similarities between both the wild shot and the Boscawen
animals support the theory that the two are closely related. The
multiple character analysis shows this total population to have
certain characteristics in common and, further, to differ more
from the known hybrids than it does from the Minnesota popula-
tion. That this is not evidence of rapid evolution of pure coyote
stock to suit the prey and habitat requirements of the Northeast
is suggested by the extreme and uncoyotelike combinations of cer-
tain characters described above as well as by the high degree of
variability discussed earlier. Further, it seems that some of the
traits found, such as the slender rostrum and crowded teeth, would
have no selective advantage for an animal preying on large game.
Finally, some of the unmeasurable characters usually diagnostic
for dog/wolf, such as reduction of the metaconid of M/1 and
shape of the postorbital region, or for dog such as flattened bullae,
suggest some heterogeneity. Probably what has occurred is that
animals of mixed but predominantly coyote ancestry have sur-
vived and bred amongst themselves, adapting rather easily, as Canis
does, to shifting environmental conditions as they have moved
east. The difi'erences between these animals and the known hybrids,
their intermediate position when compared with lupus, and the
rather large size, especially of teeth, all suggest further that wolf
as well as domestic dog is involved in their ancestry.

12 BREVIORA No. 330

Although studies have not been made in detail of animals from
areas between New England and the erstwhile extreme eastern edge
of the range of latrans thamnos, individual specimens from the
Adirondacks and the St. Lawrence Valley, as well as reports of
difficult-to-identify Canis from southern Ontario, all suggest that
we are dealing with a rather widespread phenomenon, which very
likely parallels that found in the southern states where, along the
eastern edge of the coyotes' extending range, "red wolves" are
reported. That these animals are not a distinct species and the pos-
sibility of hybridization were discussed in our earlier paper (1967:
230). While there is considerable resemblance between individual
specimens from the different areas, the New England population
on the whole seems less wolflike, though both populations agree in
being highly variable.

SUMMARY

Cranial studies of the population of Canis, which is presently
expanding into the empty predator niche in New England show
that these animals are predominantly coyote and probably have
some dog/wolf ancestry. A multiple character analysis shows that
they differ from known dog-coyote hybrids, which are intermediate
between the two parent stocks. The New England animals are
closely related to C. latrans thamnos, a Minnesota population that
has already begun to move away from typical latrans towards both
familiaris and lupus. That these changes cannot be entirely ac-
counted for as evidences of a rapid evolution of coyote stock is
shown by the high degree of variability of the population, the non-
homogeneous combinations of certain features, and the possession
of some particular characters usually considered to be diagnostic
for dog/wolf. The differences, as shown by the multivariate analy-
sis, between this population and that of known dog-coyote hybrids
further suggest that wolf as well as dog genes have been introduced.
The conclusions arrived at in these cranial studies are in agreement
with the behavioral trends noted by the Silvers (in press).

Because of our present imperfect knowledge of these animals,
their probable hybrid ancestry, and undiagnostically wide varia-
tion of cranial characters, no trinomial is proposed for them; rather
they should be called Canis latrans var. and may be referred to
as the eastern coyote.

1969 NEW ENGLAND CANIS 13

ACKNOWLEDGMENTS

This work is part of a joint study of New England Canis under-
taken in cooperation with Helenette and Walter T. Silver of the
New Hampshire State Fish and Game Department. Their work on
comparative behavior of these animals is in press (Wildlife Mono-
graphs). The Boscawen animals used were raised by them, and
the authors are indebted to them for also supplying most of the wild
shot specimens. We are grateful as well to the many individuals
who collected these.

The authors are much obliged to Dr. H. T. Gier of Kansas State
University for his kindness in allowing us to study his excellent
series of dog-coyote hybrid skulls, and to Drs. J. Knox Jones, Jr.
and Robert M. Mengel of the Museum of Natural History, Uni-
versity of Kansas, for the loan of additional hybrid material. Our
thanks also go to Dr. W. M. Breckenridge of the Minnesota Mu-
seum of Natural History for the series of C. I. thamnos, to the
Illinois State Natural History Survey, the United States National
Museum, and the Ministere de la Chasse, Faune Terrestre of the
Province of Quebec for supplementary specimens.

This work has been supported primarily by National Science
Foundation grant GB-1265 and also by the Department of Biology
and the Milton Fund of Harvard University.

LITERATURE CITED

Cramer, H.

1946. Mathematical Methods of Statistics. Princeton, Princeton Uni-
versity Press. 575 pp.
Lawrence, B., and W. H. Bossert

1967. Multiple character analysis of Canis lupus, latrans, and famili-
aris with a discussion of the relationships of Canis niger. Ameri-
can Zoologist 7: 223-232.

Young, S., and H. H. T. Jackson

1951. The Clever Coyote. Harrisburg, The Stackpole Co., and Wash-
ington, D.C., The Wildlife Management Institute, XV & 411 pp.

(Received 8 May 1969)

;-N

^ /^r^ k -I .1 ^O NOT CtRCULArrcoMP. zocL.

f|,CLamDr.A<^ei , LIBRARY

B R E V I O R A'^^^

Muse 11 in of CoHiparative Zoology""'

Cambridge, Mass. 24 October, 1969 Number 331

A SCINCOID LIZARD FROM THE CRETACEOUS
AND PALEOCENE OF MONTANA'

Richard Estes

Abstract. Contogenys sloani, n. gen., n. sp., is a scincoid lizard from
the late Cretaceous Hell Creek Formation and middle Paleocene Tongue Ri-
ver Formation of Montana. Its closest relationships seem to be with primitive
members of the Scincidae, and it is tentatively referred to that family.
Contogenys is specialized in possession of short mandibles with the posterior
teeth diminishing little in size relative to anterior ones.

INTRODUCTION

Hoffstetter (1962), in describing known finds of fossil Scincidae,
cited a late Cretaceous record (later described as Sauriscus; Estes,
1964), and noted that no other North American scincid fossils
were known before the late Pliocene. In 1963 I described an early
or middle Miocene occurrence of Eiimeces from Florida, in 1964
Estes and Tihen noted a probable occurrence of Eiimeces in the
late Miocene or early Pliocene of Nebraska, and in 1965, I reidenti-
fied as Eumeces a middle Oligocene skull from Nebraska. A sec-
ond middle Oligocene skull of Eumeces is now known from North
Dakota (Frick Collection, American Museum of Natural History,
FAM 42916). The Cenozoic record of Scincidae in North Ameri-
ca is therefore now much better documented than it was at the
time Hoffstetter's review was written.

Sauriscus, from the late Cretaceous of Wyoming, was referred
to the Scincidae on the rather tenuous grounds that it possessed a
prominent, upturned coronoid process of the dentary (Estes,
1964). Admittedly, while it inay be a scincid, Sauriscus is known
only from dentaries and maxillae, making it difficult to distinguish
from the related Cordylidae and Gerrhosauridae on other than
zoogeographic grounds.

^Fossil vertebrates from the Hell Creek Formation, Montana: Contri-
bution No. 6.

2 BREVIORA No. 331

The presence of still another scincid-like form in the late Creta-
ceous Hell Creek Formation of Montana is therefore of consider-
able interest. As with the Lance Formation specimens, the Hell
Creek fossils are disarticulated, but a fragmentary dentary from
the middle Paleocene of Montana, with postdentary bones in artic-
ulation, allows comparisons with other families to be made with
somewhat greater assurance.

ORDER SAURIA

SUBORDER SCINCOMORPHA

Superfamily Scincoidea

Family Scincidae?

Contogenys sloani, n. gen., n. sp.

Type. MCZ 3681, almost complete left dentary.

Paratypes. MCZ 3682, fragmentary right maxilla; 3683, frag-
mentary left dentary; 3684, twelve fragments of dentaries; 3685,
two parietals. Collected by A. D. Lewis and party, 1964.

Type locality. Bug Creek Anthills, Hell Creek Formation, SW V4
Section 9, T 22 N, R 43 E, McCone County, Montana.

Age. Late Cretaceous.

Etymology. Greek, kontos, short; genys, jaw. The specific
name is for Dr. Robert E. Sloan, who has done most of the col-
lecting in the Bug Creek area.

Referred Specimens. Princeton University (PU) 17035, right
dentary; 17036a, broken left dentary; 17036b, posterior end of
right dentary with associated coronoid, splenial, and parts of sur-
angular and angular. Tongue River Formation, Medicine Rocks,
T 3 N, R 58 E, near Ekalaka, Carter County, Montana; middle
Paleocene.

Diagnosis. A short-jawed scincoid lizard; a prominent labial cor-
onoid process of dentary present; deep lingual coronoid notch
occurring on dentary, posterior to tooth row; blunt, homodont
teeth with squared-off crowns that have faint anterior and posterior
crests; 13-15 dentary teeth; posterior dentary and maxillary teeth
diminishing very little in size and height.

Description. The dentary is relatively short and blunt, and the
Meckelian groove is narrowly open to the symphysis. Dorsoposteri-
orly the dentary turns sharply upward at its tip, forming a small
but prominent coronoid process. The posterior edge of the dentary
curves ventrally and then turns sharply caudad, forming a promi-
nent ventroposterior (or Meckelian) process, the tip of which is

1969 CRETACEOUS SCINCOID LIZARD 3

broken off in the Cretaceous specimens but is present in the
Paleocene specimen PU 17035. The external surface of the den-
tary is smooth and has a row of mental foramina; a depressed
channel is present near the parapet of the jaw, ventral to the teeth.
Labially, between coronoid process and ventroposterior process,
there is a distinct depression for attachment of the adductor muscu-
lature. Lingually a deep facet for the coronoid is impressed on the
dentary, just posterior to the tooth row. There are thirteen teeth
in the complete tooth row preserved on the type, although all but
four of the teeth are broken. The tooth row ends posteriorly with-
out marked reduction in tooth size. There may have been an addi-
tional tooth on the now missing anterior tip. The teeth are strongly
pleurodont and are bordered lingually by a deep, well-defined sulcus
dentalis. The vertical tooth shafts are regularly spaced; they are
quite slender and closely spaced anteriorly, increasing in diameter
and in spacing posteriorly. They project about a third of their total
height above the external parapet of the jaw. The tooth crowns
are blunt, slightly compressed linguolabially, and are essentially
smooth on both lingual and labial sides. Faint vertical anterior and
posterior crests are present in the lingual side of the crowns
(Fig. 1). PU 17035, a right dentary, resembles the Cretaceous
specimens in every way (Fig. 4).

The maxilla is expanded posterolaterally and a triangular wedge
protrudes posterodorsally, which evidendy inserted into the jugal.
The tooth row ends abruptly without marked reduction in size or
height of posterior teeth (Fig. 2).

The referred parietals are short, with widely-divergent supra-
temporal processes. Dorsally, they have a rather coarse, rugose
osteoscutal sculpture. The interparietal scale impression is promi-
nent and raised, but no other epidermal scale markings are present.
The parietal foramen is well developed. Ventrally. temporal
muscle areas are present on each side, but they do not extend onto
the dorsal surface of the bone. A postfrontal or postorbital facet
is present anterolaterally. Posteroventrally, a deep pit for the
parietal ligament is flanked by prominent supraoccipital attachment
areas (Fig. 3).

PU 17036b, a specimen from the middle Paleocene Tongue River
Formation of Montana, shows that the splenial is large and fills
the Meckelian groove, extending posteriorly under the coronoid,
and indicates the nature of the labial sutures and processes of den-
tary and postdentary bones. There is a tongue-like process of the
coronoid separating coronoid process of the dentary from the an-
terior surangular foramen ( Fig. 5 ) . The latter lies entirely on the

4 BREVIORA No. 331

surangular. Lingually the angular reaches anteriorly to about the
midpoint of the surangular "window", and the relatively large
splenial has a large anterior and small posterior mental foramen,
both placed rather far posteriorly.

Discussion. The following combination of characters suggests
that Contogenys is closely related to the scincomorph superfamilies
Scincoidea and Cordyloidea: (1) tooth crowns with weak ridges
on anterior and posterior edges of the lingual side; (2) strongly
divergent posteroventral and coronoid processes of the dentary;
(3) labial coronoid process of the dentary that laps up onto the
anterior edge of the coronoid bone; and (4) prominent labial
tonguelike projection of the coronoid behind coronoid process of
dentary. All Lacertoidea (Teiidae and Lacertidae) have a strong
anterior process of the coronoid overlapping the dentary, unlike
scincoids, cordyloids and Contogenys (Fig. 5).

In labial view, the surangular and angular do not extend anteri-
orly beyond the most posterior tooth, thus resembling the scincids
but differing from the cordyloid condition (Fig. 5). The wedge-
shaped, laterally-projecting posterior border of the maxilla of Con-
togenys resembles the condition in many Scincidae, especially
Scincinae such as Eunieces, in which this wedge fits into a notch
in the jugal. Greer (1969) has placed the Scincinae as the most
primitive of the scincid subfamilies.

The laterally-crested, blunt teeth of Contogenys resemble those
of primitive scincids and gerrhosaurids, although they lack the
lingual striations commonly present in the latter two groups. Cor-
dylids and the Lacertoidea usually have bi- or tricuspid teeth and
relatively long, slender dentaries, unlike Contogenys.

The late Cretaceous Saiiriscus (see p. 1 ) has a long, slender jaw,
bifid teeth, and extension of the notch for surangular and angular
anterior to the termination of the tooth row; in these features it re-
sembles the Cordyloidea more than the Scincoidea, and future
finds may necessitate its removal from the Scincidae. In any case,
there is no resemblance to Contogenys.

The posterior end of the tooth row of Contogenys appears ab-
ruptly truncated because of unreduced size of the posterior teeth.
The microteiid Neiisticurus ecpleopus resembles Contogenys in
this adaptation, but differs significantly in tooth type, closed
Meckelian groove, long labial dentary process of the coronoid, and
other features. Adaptation to a particular diet may have produced
a similar arrangement in the two species, which are certainly
unrelated.

1969 CRETACEOUS SCINCOID LIZARD 5

The parietals of Contogenys superficially resemble those of
primitive Scincinae in having a raised interparietal scute area and
in apparendy having an open upper temporal fenestra. The an-
teroposterior shortness of the bone is a resemblance to the Lacer-
tidae and to some cordyloids, although it can be matched in some
scincids. The ventral epipterygoid flanges of the parietal that occur
in Scincidae are absent in Contogenys.

Scincoidea and Cordyloidea are probably more closely related to
each other than they are to Lacertoidea (Hoffstetter, 1962), and it
is to be expected that ancient members of these families (as well as
modern ones, so far as dentary structures are concerned) would
be difficult to discriminate. Tentatively, I refer Contogenys to the
Scincidae on the basis of the Eumeces-YikQ wedge on the posterior
border of the maxilla and on the general appearance of the teeth.
If the parietals are properly referred, the open upper temporal
fenestra could also be included as a scincid resemblance, although
this primitive condition is also present in the Jurassic cordyloid
Paramacellodus (Hoffstetter, 1967). On the other hand, absence
of the characteristic ventral (epipterygoid) flanges of the parietal
found in aU living scincids casts doubt on the reference to Scinci-
dae. However, lacking more assurance of the association of the
parietals, even a questionable reference to the Scincidae is prefer-
able on both morphological and zoogeographical grounds than is
a suggestion of relationship to cordyloids.

ACKNOWLEDGEMENTS

I thank Allen Greer for helpful comments. Figures la and 2a
were prepared by Miss Tehrie Holden, figures lb, 2b and 3 by
Laszlo Meszoly, figure 4 by Howard Hamman. This research was
supported in part by National Science Foundation Grant GB-7176.

LITERATURE CITED

ESTES, R.

1963. Early Miocene salamanders and lizards from Florida. Quart.
Journ. Florida Acad. Sci., 26: 234-256.

1964. Fossil vertebrates from the late Cretaceous Lance Formation,
Eastern Wyoming, Univ. Calif. Publ. Geol. Sci., 49: 1-180.

1965. Notes on some PaJeocene Lizards. Copeia, 1965, No. 1: 104-
106.

ESTES, R., AND J. TlHEN

1964. Lower vertebrates from the Valentine Formation of Nebraska.
Amer. Midi. Nat. 72: 453-472.

6 BREVIORA No. 331

Greer, A.

In press. A subfamilial classification of scincid lizards. Bull. Mus. Comp.
Zool.

HOFFSTETTER, R.

1962. Revue des recentes acquisitions concernant I'histoire et la systeni-
atique des squamates. Coll. Intern. C.N.R.S., 104: 243-279.

1967. Coup d' oeil sur les sauriens (=lacertiliens) des couches de Pur-
beck (Jurassique superieur d'Angleterre). Coll. Intern. C.N.R.S.,
163: 349-371.

(Received 13 June 1969.)

1969

CRETACEOUS SCINCOID LIZARD

a.

b.

Fig. 1. Contogenys sloani, a, MCZ 3681, type left dentary, labial view,
restorations made from other specimens; h, the same, lingual view; c, MCZ
3683, detail of four teeth from fragmentary left dentary, lingual view, a-b
X 13; c X 27; Hell Creek Formation, Montana, late Cretaceous.

8

BREVIORA

No. 331

Fig. 2. Contogenys sloani, MCZ 3682, posterior portion of right maxilla;
a, labial, b, lingual view; x 13; Hell Creek Formation. Montana, late
Cretaceous.

Fig. 3. Contogenys sloani, MCZ 3685, referred parietal, a, dorsal; b,
ventral view; X 5; Hell Creek Formation, Montana, late Cretaceous.

1969

CRETACEOUS SCINCOID LIZARD

Fig. 4. Contogenys sloani, PU 17035, right dentary, lingual view; X 10;
Tongue River Formation, Montana; middle Paleocene.

Fig. 5. Patterns of labial dentary-postdentary bone articulation in right
mandibles of various lizards, a, Contogenys sloani, PU 17036b, Tongue River
Formation, Montana, middle Paleocene; h. Gerrhosaurus flavigularis, MCZ
50988; c, En nieces longirostris, MCZ 20508; d, Cordyliis cordyliis, MCZ
21570; all X 6.5. Abbreviations: a=angular, c=coronoid, d=dentary,
s=surangular.

^- NA-CC^mb^l^QcJ PO NOT CIRCULAT

I LATE

L.ea.^-""-

B R E V I O K,S'

Mmseiiim of Coimpsirsitive ZpMt>g^'j^o

Cambridge, Mass. 24 October, 1969 Number 332

THE BRAZILIAN TRIASSIC CYNODONT REPTILES
BELESODON AND CHINIQUODON

Alfred Sherwood Romer

Abstract. Skulls of the carnivorous cynodonts Belesodon and Cliiniqiio-
iloii from the Triassic Santa Maria beds of Brazil are described and figured,
as well as mandibles and postcranial materials of the latter genus.

INTRODUCTION

In connection with current studies on the cranial anatomy of
carnivorous cynodonts from the Triassic Chanares Formation of
Argentina, my attention has been recalled to two apparently related
cynodonts, Belesodon and Chiniqiiodon, from the Middle Triassic
Santa Maria Formation of Brazil. These forms were described by
Huene (1944), but described on such imperfect cranial material
that their relationships and much of their structure have remained
in doubt.

Only a single specimen of each was present in Huene's collection.
Of Belesodon, the skull was nearly complete, but somewhat dis-
torted and obviously rather crushed, so that Huene, although fig-
uring it in dorsal and ventral aspects, was unable to give satisfactory
lateral views. Further, its cheek teeth had been lost, so that it was
even suspected at one time of being a gomphodont rather than a
carnivore (Bonaparte, 1963). Of Chiniqiiodon, only a fragmen-
tary and battered skull was found, and although a carnivorous denti-
tion was partially preserved, little of cranial structure can be made
out from Huene's published photographs. Bonaparte (1966) has
described a specimen from the Ischigualasto beds (later in time
than those of Santa Maria), which he believes to be a surviving
Chiniqiiodon, but this specimen is likewise very fragmentary in
nature.

For the two genera. Huene established the family Chiniquo-
dontidae in 1956, but he gave no diagnosis of this new group;
indeed, few facts were available which could be utilized for a family

BREVIORA

No. 332

definition. However, comparison with obviously similar specimens
from the Chanares Formation has helped to clarify the nature of
the Santa Maria forms, as have notes given me by Sr. Bonaparte
resulting from a recent study of Huene's specimens. The general
nature of the two genera is now clear. Belesodon possessed, it is
certain, a dentition similar to that of Chiniquodon, and the two
are closely related.

A major aid to understanding comes from undescribed ma-
terials^ in the Museum of Comparative Zoology, collected by a
Brazilian expedition in 1936, conducted by Dr. T. E. White and

Fig. 1. Dorsal view of the Belesodon skull. MCZ 1533. as preserved.

X V6.

' Preparation of the material here described was made possible by National
Science Foundation Grant GB-50{).

1969 BRAZILIAN IRIASSIC CYNODONTS 3

L. I. Price, in the Santa Maria beds. These include a sicuU ( MCZ
1533) comparable to Huene's Bclesodon type, and two skulls
which appear referable to Chiniquodon (MCZ 3614, 3615).
These skulls were collected south of Candelaria, State of Rio
Grande do Sul, in Santa Maria beds apparently comparable to
those in which Huene had collected at Chiniqua. In this area,
the Harvard expedition collected a considerable quantity of dicyno-
donts, a few cynodonts, and a single archosaur. As at Chiniqua,
and in contrast to the Santa Maria area, no rhynchosaurs were
found. Price has later collected more material from this region, in-
cluding the only South American procolophonid (Price, 1947),
and collaborated with E. H. Colbert in collecting still further Can-
delaria specimens for the American Museum of Natural History.

The skull specimens described here are somewhat distorted, and
furthermore, as in much Santa Maria material, action subsequent
to burial (probably thermal, perhaps due to the superposed Serra
Geral volcanics) has resulted in the presence of numerous breaks
of the bone surface, giving an effect similar to that of "crackle-
ware" pottery (Fig. 1). This has made it difficult to be sure of
the sutural pattern in many instances. In my figures, I have to
some degree "idealized" the specimens, correcting dislocations
and distortions, omitting obviously unnatural breaks and cracks,
and fining in the outlines of various minor deficiencies. This may
have resulted in minor inaccuracies, but I do not believe them
serious, and this procedure yields a better understanding of skull
morphology than would a literal portrayal of damaged detail. Since
I am elsewhere describing much better preserved Chafiares rela-
tives, I shall give here only a very general account of the Santa
Maria material.

BELESODON
(Figs. 1-5)

The MCZ Belesodon skull (No. 1533) measures 224 mm in
length from premaxilla to occipital condyles. It is thus rather com-
parable in size to Huene's type, in which this same measurement
is 258 mm. The general skull proportions (and those of Chiniquo-
don as well) are similar to those of the earlier (and more primi-
tive) cynodonts of the family Thrinaxodontidae. The center of
the orbit is slightly anterior to the mid-length of the skull, whereas
in the type, as figured, it is very slightly posterior. Septomaxillae are
present, but somewhat displaced, and their posterior limits are not
clear; presumably they conformed to the usual cynodont pattern.

BREVIORA

No. 332

The orbital rim is imperfect in the lacrimal region, but there does
not seem to have been development of the somewhat variable
protuberances seen in other chiniquodontids. along the rim an-
terior to the lacrimal duct. Anterior to this point, on the boundary
between lacrimal and maxilla, there is a well-marked pit which
appears not to have been present in the Belesodon type. The sagit-
tal crest is high, and the two parietals are closely appressed, al-
though not fused. There is a narrow cleft between the parietals

Fig. 2. Restored dorsal view of the Belesudon skull, X

Abbrevia-

tions for Figures 2-8: ho, basioccipital; bs, basisphenoid; d. dentary; e,
epipterygoid; /, frontal; fo, fenestra ovalis; /, jugal; //, jugular foramen;
/, lacrimal; m, maxilla; //, nasal; oc\ occipital complex; p, parietal; pap.
paroccipital process; pi. palatine; pm, premaxilla; po. postorbital; pp. post-
parietal; /'/'/, pterygo-paroccipital foramen; pr. prootic; prj . prefrontal; pt.
pterygoid; ptf. postemporal fenestra; sni, septomaxilla; scj. squamosal; t,
tabular.

1969

BRAZILIAN TRIASSIC CYNODONTS

anteriorly, but the presence of a parietal foramen is doubtful. The
condition of the specimen shows clearly that the parietals diverge
posteriorly to take part in the occipital crests, clamped between the
squamosals anteriorly and the postparietal posteriorly. On the oc-
cipital surface, sutures between the occipital complex and the ad-
jacent dermal elements are difficult to determine.

Fig. 3. Restored lateral view of the Belesodon skull. X V2 .

Except for highly developed canines, the teeth are represented
only by their sockets. There were the four incisors commonly
present in cynodonts. The cheek teeth formed a somewhat curved
row, turning outward posteriorly, where the maxilla extends some-
what outward and downward beyond the general line of the cheek.
There appear to have been 12 cheek teeth, which presumably had
the posteriorly curved tips seen in other chiniquodontids. The
secondary palate is greatly elongated beyond the limits seen in non-
chiniquodontid cynodonts. The posterior end of the secondary
palate is somewhat incomplete on the left side; I have restored
it to the condition seen on the right. The maxillary-palatine suture
is obscured, but it is obvious that a great part of the secondary
palate is formed by the palatines. For a short distance at the pos-
terior end, the secondary palatal plate projects freely backward
without being tied in at either side to the primary plate of the
palatine. The posterior end of the vomer is presumably present in
the roof of the choanal region, but its sutural separation from the
pterygoid is obscure. The typical posteriorly-projecting "finger"
of the palatine is present on either side of the primary palate,

BREVIORA

No. 332

Fig. 4. Restored palatal view of the Bclcsodun skull. X Vs.

separating medial and lateral palatal areas of the pterygoid. The
boundaries of the ectopterygoid — presumably present but small
— cannot be determined. On either side of the primary palate, a
distinct longitudinal ridge is present at the line of suture between
palatine and medial palatal ramus of the pterygoid. These ridges
continue backward along the pterygoids, gradually converging pos-
teriorly. A slight medial ridge is present at the line of suture
between the two pterygoids. Ventrally-projecting flanges from the
two pterygoids are highly developed.

The parasphenoid-basisphenoid region appears to have a nor-
mal cynodont structure, but the bone is somewhat imperfect. The
ventral suture between the sphenoid and the occipital complex is

1969

BRAZILIAN TRIASSIC CYNODONTS

obscure. At about the position where this suture would be ex-
pected, somewhat anterior and medial to the jugular foramina, are
a pair of pits which may be foramina. The ventral braincase sur-
face is imperfect lateral and anterior to the fcncstrae ovales. On
the lateral surface of the braincase, relations of the prootic to
adjacent elements are none too clear, but, as is usual in cynodonts,
there is evidence of a groove for a vein lying between the prootic
below and the squamosal and parietal above. The quadrates (as
is all too common in cynodont specimens) are missing.

Fig. 5. Restored occipital view of the Belesodon skull. X V3.

CHINIQUODON
(Figs. 6-9)

As noted above, Chiniquodon was founded on a very incomplete
skull, poorly preserved and poorly illustrated. The Harvard speci-
mens add very considerably to our knowledge of this genus. They
include two skulls (MCZ 3614 and 3615), a pair of mandibles,
and some postcranial materials. Of the two skulls, MCZ 3614
measures 152 mm in length to the occipital condyles; MCZ 3615
is imperfect in the premaxillary region, but appears to be very
closely comparable in size. The type skull would appear to have
been about 140 to 145 mm, if complete. The two Harvard skulls
are thus very close to the type in size and, further, appear to match
the type closely in such features as Huene was able to describe.
The skulls, like that of Belesodon, show a "crackle-ware" pattern
which makes interpretation difficult. As in the case of Belesodon,
I have attempted to present figures of the skull in which the
"crackle-ware" effects are eliminated and in which breaks and

8

BREVIORA

No. 332

effects of crushing and distortion have been eliminated. The skull
roof is rather well preserved in MCZ 3614, but litde can be made
of the palate. In contrast, the roof is not too good in MCZ 3615,
but most of the palatal structure can be readily made out. MCZ
3615 has undergone strong lateral crushing. MCZ 3614, on the
contrary, appears broad and relatively flat, so that dorsoventral
crushing might be suspected. However, this is apparently not the
case, for the braincase at the level of the parietal crest is as high in
MCZ 3614 as in MCZ 3615.

My figures are based on a synthesis of features seen in the two
skulls; the dorsal surface, arches, and occiput conform mainly to
MCZ 3614; the palate and ventral portion of the braincase derive
almost entirely from MCZ 3615. As in Belesodon, the general

Fig. 6. Restored dorsal view of the Chinicjiioilon skull. X %•

1969

BRAZILIAN TRIASSIC CYNODONTS

Fig. 7. Restored lateral view of the Cliinic/iioclon skull. X %.

proportions are comparable to those of thrinaxodonts. In neither
skull are the ascending processes of the premaxillae, separating
the nostrils, preserved; however, their slender distal tips are seen
in MCZ 3614 between the nasals. The narial region is present
only in MCZ 3614, but even here is damaged and the septomaxil-
lae displaced and crushed, so that pattern in the region of the ex-
ternal nares is somewhat doubtful. As in Belesodon, the maxillae
arch far dorsally around the roots of the canines; the canine swell-
ings are not as marked as in Belesodon except posteriorly, where
there is a sharp contraction in cheek width. The posterior end of
the maxilla, beneath the orbit, projects very distinctly below the
general contour of the cheek. The large nasals and the shorter,
triangular frontals, ridged on either side by the postorbitals, com-
pare closely with those of Belesodon. As in that genus, the crests
of the two parietals are separated anteriorly for a short distance,
but details are too imperfect to be sure whether or not a vestigial
parietal foramen was present. The parietals continue backward
conjoined in a high crest. Its summit is thinner than in Belesodon,
and the two bones more closely appressed than in that genus, so
that little trace of a median suture can be made out. The con-
dition of the specimen is such that it cannot be determined how far
the parietals extend posterolaterally into the occipital crests.

Of the circumorbital elements, the prefrontal occupies a mod-
est area above the orbits; its boundaries are none too clear. The
equally small lacrimal is situated below it on the anterior orbital
border; a small but distinct protuberance extends back over the
orbital rim external to the lacrimal foramen. Anterior to this point

10

BREVIORA

No. 332

there is a slight depression in the contour of the face, but this is
much more shallow than the rather deep pit which appears to be
present in Belesodon. The jugal is, as customary, a large element.
Its slender anterior end lies between lacrimal and maxilla just
anterior to the orbit; posteriorly it expands beneath the orbit to
extend most of the length of the zygomatic arch, reinforced ex-
ternally by the squamosal. It forms the lower part of the post-
orbital arch; its suture with the postorbital bone is indistinct. This
last element forms most of this arch. On the dorsal surface it ex-
tends slightly farther forward than in Belesodon, and then with a
dorsal ridge, swings backward to sheath the parietal laterally for
some distance. The squamosal has the typical cynodont pattern of

Fig. 8. Restored palatal view of the Chiiiiquodon skull, X V^-

1969 BRAZILIAN TRIASSIC CYNODONTS 11

two semi-distinct parts. The zygomatic arch ramus constricts pos-
teriorly to join the more medial segment of the bone with a rela-
tively narrow neck far down the side of the skull. The medial
ramus extends from the back end of the sagittal crest as a broad
sheet of bone which, sheathing postparietal and then tabular an-
terolaterally, curves outward and downward along the anterior
surface of the occipital crest. The details in the region of the
quadrate (which is absent) and of the paroccipital process are not
well preserved; clearly seen, however, is the thickened area of the
squamosal which runs downward along the ventromedial margin
of the occipital ramus, lateral to the prootic, to strengthen the inner
end of the quadrate region.

In MCZ 3615 the occiput is strongly compressed laterally;
the proportions are more normal in MCZ 3614, although there is
a slight spread ventrally between the occipital crests. In neither
specimen are details well preserved, and sutures are obscured.

Almost nothing can be made of the palate of MCZ 3614, and
reliance must be placed on MCZ 3615. Even here, the anterior
part of the palate is in none too good condition; however, except
for the palatal exposure of the premaxillae, the general structure
is clear, although the breadth may have been somewhat less than
I have figured it. The secondary palate here is developed to the
same major extent as in Belesodon, with the palatines playing a
major part in it. Details of the primary palate posterior to the
end of the secondary palate are obscure as to sutural relations of
pterygoids, ectopterygoids, palatines, and vomers. There are well-
developed pterygoid flanges, but these are less prominent than in
Belesodon. Posterior to the contraction of the palatal structures
into the basicranial "girder," the ventral braincase structures, of
proper cynodont character, are clearly discernible. The lateral
walls of the braincase are somewhat imperfect in both specimens,
but are rather better preserved in MCZ 3614 . The parietal above,
the squamosal posteriorly, and the pterygoid ventrally, with the
epipterygoid and prootic filling out the side walls of the braincase
area, appear, as far as can be determined, to form a typical cyno-
dont pattern. The distal end of the quadrate ramus of the epi-
pterygoid is unfortunately imperfect.

Of the premaxillary-maxillary dentition, MCZ 3614 shows the
four customary incisors, here of small size, broadly spaced across
the arch formed by the premaxillae. The canines are stout; one,
measuring 20 mm in its exposed portion, is completely preserved.
The cheek teeth are moderately well preserved in MCZ 3615. The
two series begin anteriorly somewhat medial to the canines and

12 BREVIORA No. 332

extend backward, diverging at an angle of about 35^-40^ to each
other, well below the level of the secondary palate; the rows are
nearly straight for most of their extent, but curve somewhat out-
ward posteriorly. On the right side of MCZ 3615 a small anterior
alveolus is followed by eight teeth; on the left, there are eight
teeth, with posteriorly an alveolus in which a ninth tooth was
forming. As in carnivorous cynodonts in general, the teeth are
oval in shape at their bases; subcircular in the case of the small
anterior teeth, elongate anteroposteriorly in the more posterior
members of the series. The crowns are almost all lacking in MCZ
3614, are rather imperfect on the left side of MCZ 3615, but in
fairly good shape on the right side of the latter specimen. The
first three appear to have single cusps. The fourth and fifth teeth
are imperfect, but the fourth appears to have an accessory pos-
terior cusp, and the fifth definitely has a small anterior accessory
cusp as well. On tooth six, the anterior cusp is somewhat medial
but remains small; the posterior one is becoming larger. In teeth
seven to nine the posterior cusp is very prominent, and has di-
vided to give off a secondary posterior cusp near its base. The
anterior cusp persists in teeth seven and eight, and extends pos-
teriorly on the medial surface in cingulum fashion on the former,
it has disappeared in the last tooth. The tips of the main cusps of
teeth five through seven are incomplete, but in teeth eight and
nine, this cusp is complete and is seen to be curved sharply back-
ward.

With the skulls was found a conjoined pair of lower jaws, poorly
preserved, much "crackled," and incomplete posteriorly. The jaws
are of a size appropriate to the skulls. The large dentary is nearly
completely preserved, except for its posterior tip. The splenial is
present, but there are few identifiable remains of other postdentary
elements. The angle of the dentary is sharply marked and thick-
ened ventrally. The symphysis is, as expected, extremely stout.
On the inner surface of the ramus, traces of the meckelian groove
between dentary and splenial are visible anteriorly. Posteriorly,
the outline of the hollow (sharply defined dorsally), which should
house the surangular, is visible. The dentition is poorly preserved.
There appear to have been three incisors in front of the stout
canines. The lower postcanine tooth row forms a nearly straight
line on either side, the two rami diverging at an angle of about
40°. The area for the roots of the cheek teeth forms a very dis-
tinct swelling along the upper part of the inner surface of the
dentary; this swelling — and the tooth row itself — extends to a
point a short distance back of the base of the ascending ramus.

1969 BRAZILIAN TRIASSir CYNODONTS 13

Little can be made of the cheek tooth series except that all were
thin-oval in section and compressed lateromedially, and that the
more posterior ones were, as expected, longer anteroposteriorly
than the smaller anterior ones. There are eight members in the
series.

A collection of postcranial elements of appropriate size (Fig. 9)
was also found with the two skulls. A number of disarticulated
vertebral centra had diameters of 12 to 16 mm, and lengths of 10
to 14 mm. Two scapulae were present, with heights as preserved
of 72 and 81 mm; with one were associated a pair of coracoidal
elements. The shoulder girdle is of a typical cynodont pattern.
Two complete and a third, incomplete, humeri are present. A left
humerus, with a length of 78 mm, is essentially perfectly preserved.
It is of a typical cynodont type, but is relatively short and stout
as compared with that of Diademodon, for example; in that form,
the distal width is but slighdy over half the length, whereas here
this figure is just short of 60 per cent. Associated in the beds
with this material was a tiny humerus, obviously that of a "young-
ster." As preserved, it measures 29 mm in length. The distal
end is not preserved and was presumably poorly ossified, if ossified
at all; even if allowance be made for this situation, the total length
in life would have been but approximately 40 per cent that of
the more adult specimen described.

A disarticulated right ilium is nearly complete, except for the
posterior tip of the blade, but is badly battered and cracked. Its
breadth across the articular areas for pubis and ischium is 34 mm.
The supra-acetabular buttress is highly developed, extending out-
ward some 1 1 mm over the deepest part of the acetabulum. A
small but very distinct pocket is present on the outer surface of the
ilium, close to the acetabular margin, just in front of the buttress.
The greatest length of the iliac blade, as preserved, is 55 mm. Two
ischia are present; the better preserved is illustrated, reversed, in
articulation with the ilium. There is a very distinct posteriorly-
facing surface for the iliotrochantericus.

A left femur, with a length of about 80 mm, is, in general, of
typical cynodont type, slender in build (as compared with the
stout humerus), with a proximal width across the head and greater
trochanter of 32 mm, a shaft diameter of 9 mm. and a distal width
of 23 mm. As typical of the cynodonts. the dorsal surface of the
head is flattened and. indeed, somewhat concave in section.
There is a distinct, if shallow, intertrochanteric fossa, and a dis-
tinct internal trochanter from which a pronounced adductor ridge
extends nearly half the length of the bone. Distally, there is little

14

BREVIORA

No. 332

B

Fig. 9. Postcranial elements found with the Cliinic/iioclon skulls, x %■
A. left ilium and ischium; B. fibula; C. femur, ventral aspect; D. humerus,
ventral aspect; E. tibia, fle.xor aspect; f", scapulocoracoid.

1969 BRAZILIAN TRIASSIC CYNODONTS 15

development of an intercondylar fossa on the dorsal surface. Ven-
trally, the two articular areas for the tibia are very highly de-
veloped and rounded; the articular areas face essentially distally,
indicating a more extended average position of the tibia than in
more primitive forms. Two tibiae and one fibula are present. The
tibiae have slender shafts but expanded termini. The right tibia, the
better preserved of the two, measures 57 mm in length, with
widths proximally of 14 mm, midway of the shaft 7 mm, distally
17 mm. The cnemial crest is sharply developed. A thin ridge,
most prominently developed proximally, is present along the lat-
eral edge of the shaft for about a third of its length, beginning a
short distance below the head of the bone. The fibula preserved
is 59 mm long. The head is broadened but flattened in primitive
fashion; the distal articular surface, stout and convex in shape, is
15 mm wide.

In the Museum of Comparative Zoology collections is a small
cynodont skull, MCZ 3035, collected by Price in the Candelaria
region, measuring 65 mm in length. Little detail can be made out,
but there is evidence of cheek teeth of a recurved, chiniquodontid
type, and the specimen seems surely a juvenile of either Chiniquo-
don or Belesodon.

DISCUSSION

Despite imperfections in the materials, the skulls figured and dis-
cussed above give us a much better picture of the cranial anatomy
of Belesodon and Chiniquodon than could be obtained from
Huene's specimens. It is obvious that these two forms are closely
related, as Huene believed to be the case, and have a number of
distinctive characters in common, notably the remarkable develop-
ment of the secondary palate. Watson and I (1956) and Lehman
(1961) suggested their inclusion in the Cynodontidae. But they
do not show any indication of derivation from typical members of
that family. Their general proportions and some details (such as
dentition) indicate alliance with, and derivation from, the
Thrinaxodontidae, as suggested by Bonaparte (1966), but the
notable advance in the palate seems in itself justification for
Huene's having erected for them the family Chiniquodontidae.

Are Belesodon and Chiniquodon so closely related that they
should be placed in the same genus? The answer is not clear. No
great weight should be given to minor difi'erences between the two
skull types as figured here, since, as I have pointed out, the skulls
on which my illustrations are based are imperfect, and seeming dif-
ferences may be due in part to misinterpretations of the data or
differential crushing. On the whole, I believe it better to maintain
their distinction until further and better materials are available.

16' / BREVIORA No. 332

■y It might further be suggested that Ch'miquodon and Belesodon
/ are merely growth stages of the same form, and that differences
such as the greater number of cheek teeth in Belesodon are merely
a function of age and increase in size. This is possible but, I think,
unlikely. It is probably no coincidence that, in the case of Huene's
materials from Chiniqua and ours from Candelaria, the specimens
found are distinctly members of two size groups. Although it is
known that in cynodonts the length of the dental battery may
increase with additions to the tooth number, the maxillary struc-
ture of Chiniquodon is such that it is difficult to see how several
further teeth could be added posteriorly. Apart from lack of fusion
of elements of the shoulder and pelvic girdles, there is little evi-
dence of immaturity in the known Chiniquodon material. For
example, the humeri of this form, described above, are essentially
completely ossified, with apparently little cartilage at either end.
They seem sufficient (but barely sufficient) to function in an ani-
mal of the proportions indicated by the Chiniquodon skulls; but it
is impossible for such a humerus to have grown to the size neces-
sary for support of an animal of the bulk indicated by the Bele-
sodon skulls.

In sum, I think it best in the present state of our knowledge to
consider Chiniquodon and Belesodon as distinct but closely re-
lated, and to provisionally retain the two as members of sympatric
genera.

LITERATURE CITED

BoNAi'ARii;, .1. F.

1963. la familia Traversodontidae. Acta Geol. Lilloana. 4: 163-194.

1966. Cliiiii(/iiocl(>n Hiiene (Therapsida — Cvnodontia) en el Triasico
de Ischigualasto, Argentina. Acta Geol. Lilloana, 8: 157-169.
HUENE, F. V.

1944. Die fossilen Reptiiien des siidamerikanischen Gondwanalandes.

Munich. 332 pp.
1956. Paliiontologie iind Phylogenie der niederen Tetrapoden. Jena,
716 pp.
Lehman. J. P.

1961. Cynodontia. //(.• Piveteaii. .1.. cd., Traite de Paleontologie. Tome
VI. Vol. 1: 140-191.
Price, L. I.

1947. Urn piocolofonideo do Triassico do Rio Grande do Siil. Serv.
Geol. Min. Brasil. Bol. 122: 9-15.
VVmson. D. M. S.. and a. S. Romfr

1956. A classification of thciapsid reptiles. Bull. Mus. ("omp. /ool..
114: 35-89.

(Received 13 .lime 1969.)

LIBRARY

B R E V I O WA^^

HARVARD

Meseem of Coimpsirsitive Zo^^fSf^..

Cambridge, Mass. 24 October, 1969 Xtmrer 333

THE CHANARES (ARGENTINA) TRIASSIC REPTILE FAUNA

V. A NEW CHINIQUODONTID CYNODONT,

PROBELESODON LEWfS/ — CYNODONT ANCESTRY

Alfred Sherwood Romer

Abstract. A carnivorous cynodont from the Middle Triassic Chafiares
fauna of Argentina is described as Probelesodon lewisi, gen. et sp. nov. This
is an advanced cynodont with an elongate mammal like secondary palate,
related to the chiniquodontids of the Santa Maria Formation of Brazil. The
short face and recurved molars indicate descent from thrinaxodontids rather
than the Cynognathidae. Cynodont ancestry is discussed. There is no evi-
dence of derivation from scaloposauroids and presumably the cynodonts are
of independent descent from the ancestral theriodonts.

INTRODUCTION

As is generally recognized, a marked change took place during
the Triassic in the nature of the terrestrial predaceous vertebrate
fauna. In the early Triassic Karroo beds carnivorous therapsids
are dominant; in the late Triassic, therapsids (except those which
have progressed to the mammalian stage) are practically extinct,
and archosaurs — thecodonts, crocodilians, dinosaurs — have re-
placed thein as the major carnivores. Until recently. Middle Trias-
sic terrestrial vertebrate faunas were almost unknown, but it
would be expected that, at that time, there would have been tran-
sitional conditions, in which predaceous therapsids would be un-
dergoing reduction, and archosaurs, on the other hand, would be
increasing in abundance and diversification.

Our currently increasing knowledge of the excellent series of
Middle Triassic beds of South America shows that this is indeed
the case. The faunas of this age now known from that continent
form a sequence including, in descending order, assemblages pres-
ent in these three formations: Ischigualasto (Argentina), Santa

2 BREVIORA No. 333

Maria (Brazil), and Chafiares (Argentina). In these beds, pre-
daceous cynodonts are undergoing sharp reduction. In the upper-
most of the three faunas, that of Ischigualasto, only a single frag-
mentary specimen of a carnivorous cynodont has so far been re-
covered (Bonaparte, 1966). In the extensive collections made by
Huene (1944) in the Santa Maria beds, there were but two car-
nivorous cynodonts, each represented by a single specimen; further
collections made for Harvard and the American Museum included
only a very few additional individuals. But when we descend to
the lowest of the three faunas, that of the Chanares Formation,
we find that cynodont reduction is relatively less marked. Only
two predaceous members of that group are present, but they are
moderately abundant. Our collections include more than a score
of specimens.

The present work is limited to a description of the cranial anat-
omy of the larger of the two forms present. I plan shortly to pub-
lish on the cranium of the second cynodont, and later, on
postcranial materials of both forms.

The cynodont here described may be formally named and char-
acterized as follows:

Probelesodon lewisi gen. et sp. nov.

Combined generic and specific diagnosis. A chiniquodontid
cynodont. closely related to Belesodon and Chiniquodon, but of
smaller size than the former, the more mature individuals reaching
an average skull length, from snout to condyles, of about 140 mm;
the secondary palate long; generally with nine maxillary cheek
teeth, the "molars" multicuspidate, the main cusp backwardly
curved.

Holotype of the species. No. 64-XI-14-6. La Plata Museum, a
skull and jaws collected from the Chafiares Formation in La Rioja
Province, Argentina, about 10 km east of the point where the Rio
Chanares emerges into the Piano de Talampaya.

The generic name is due to the obviously close relationship,
discussed later, of this form, to Belesodon (and Chiniquodon) of
the somewhat later Santa Maria beds; the specific name is given
in honor of Arnold D. Lewis, chief preparator at the Museum of
Comparative Zoology, who played a major part in the collection
of the Chanares material.

I am indebted to National Science Foundation Grant GB-2454
for aid in collection of the material and Grant GB-8171 for its
preparation and for publication costs.

1969 PR0BELE60D0N LEWISI 3

THE SKUU

(Figs. 1-5)

Of Probelesodon, nine skulls have so far been discovered. Of
these, five representing apparently mature individuals are members
of a single size group, with lengths from premaxillae to condyles
in four of them measuring: 143, 137, 132, and 130 mm (a fifth
member of this group is incomplete posteriorly). A skull which is
incomplete in the occipital region is obviously larger and probably
measured about 155 mm in length. The holotype, on the other
hand, is somewhat smaller, with a basicranial length of 1 1 5 ram.
It is probably somewhat immature; it may be noted that a replace-
ment of canines is taking place. Two "juveniles" have been found.
One measures 72 mm in length; the second lacks the occiput but
was of approximately the same size.

In general proportions, the skull resembles certain of the
thrinaxodontids and differs from such well known forms as Cy-
nognathus and Diademodon in its relatively short muzzle and
broadly flaring zygomatic arches, the width across the arches
reaching 75 to 80 per cent or more of skull length. Anterior to
the external nares the skull terminates in a pointed snout. In cross
section the muzzle is rounded, curving inward ventrally somewhat
toward the tooth rows of the cheeks, although this ventral curva-
ture is not so great as in, for example, the contemporary gompho-
donts. In advanced fashion the sagittal crest is high and thin, as
are the occipital crests. The zygomatic arches are moderately
deep, but the arch and occipital crest connect only by a narrow
waist, in contrast to the broad sweep of the squamosal connecting
the two in many other cynodonts. The postorbital arch is narrow.
The orbits are situated close to the half length of the skull. The
parietal foramen appears to be generally absent; if perhaps present
in some specimens, it is at best vestigial. As in other advanced
cynodonts, the occipital condyle is definitely double. Most notable
of distinctive characters is the great development of the secondary
palate, unmatched in any other known cynodonts except the re-
lated chiniquodontids of Santa Maria.

In external exposure the premaxillae form the area of origin of
the four incisors on either side, and the ventral and medial borders
of the external nares. Conjoined processes of the two premaxillae
extend upward and forward anterior to the nares to form the
pointed snout tip and then, diminishing gradually in size, turn
backward above as slender rods separating the nasals for a con-
siderable distance. A septomaxilla is present, presumably of the

BREVIORA

No. 333

Fig. 1. Piobelesodon lewisi, dorsal view of skull. This and the following
figures of the skull and jaws are based primarily on the type, with supple-
mentary data from other specimens. Figures 1-7 are the size of the type and
approximately % times the size of the largest mature specimens. Abbrevia-
tions for Figures 1-7: a, articular; cm, angular; bo, basioccipital; bs, busi-
sphenoid + parasphenoid; d, dentary; e, epipterygoid; ec, ectopterygoid; /,
frontal; fo, fenestra ovalis; /, jugal; //, jugular foramen; /, lacrimal: /;;.
maxilla /;, nasal; oc, occipital complex; p, parietal; pap. paroccipital process;
pi, palatine; pm, premaxilla; po, postorbital; pp, postparietal; ppi, pterygo-
paroccipital foramen; /;/•, prootic; pra, prearticular; /;/•/, prefrontal; pt,
pterygoid; ptf, posttemporal fenestra; q, quadrate -f quadratojugal; sa.
surangular; sm, septomaxilla: .sy;, splenial; sq, squamosal; t, tabular: \-,
vomer; Vj -)- 3, foramen for trigeminal nerve; vg, groove for a vein.

1969 PROBELESODON LEWIS! 5

typical cynodont type, but was apparently loosely articulated, is
frequently absent in the specimens and is never well preserved.
The maxilla is a large element, extending from the narial region
back along the side of the snout and, curving outward, has a
somewhat thickened terminus, below the orbit, which projects
notably below the level of the jugal arch. This contrasts with the
pattern seen in such cynodonts as Cynognathus and Diademodon,
in which a suborbital projection is formed by the jugal rather than
the maxilla. Behind the naris, the maxilla, in relation to the ob-
viously deep root of the canine, extends, in contact with the
nasal, far up the snout toward the midline; posterior to the canine
region, its margin turns downward again, bordered posteriorly by
the lacrimal and jugal, to terminate at the ventral process men-
tioned above. Several small foramina penetrate the maxilla, and
an especially large one is present between the maxilla and the
lower border of the lacrimal. The region containing the canine
root is prominendy swollen, and the face is expanded ventrally
at the region of the emergence of the canine. The canine swelling
is especially prominent in the larger specimens, and there tends to
develop a clearly marked channel curving around the base of the
canine root.

Of the median roofing elements, the nasals, as in cynodonts
generally, are very long; here they reach back half the length of
the skull roof to terminate between the orbits. They form the
upper margins of the nares; continuing backward between the
maxillae, they are somewhat constricted in width in the region
of the canine roots. They expand laterally to bound the dorsal
margins of the lacrimals, and then constrict again between the
prefrontals to a nearly directly transverse suture with the frontals.
Since this suture is considerably farther posterior than is common
in therapsids, the frontals are reduced in length and area, and
are essentially a pair of triangular wedges, separated suturally from
the postorbitals between and below backwardly converging ridges.
The frontals appear to be exposed laterally within the orbital rim,
between prefrontals and postorbitals. The parietals have as a main
function the formation of a high median sagittal crest. For most of
their length, the two are solidly fused at the narrow tip of the
crest. Anteriorly, they diverge slightly for a short distance. The
parietals descend some distance down the gradually diverging sides
of the sagittal crest, overlapped anteriorly by the postorbitals.
Posterior to the postorbitals the lower margins of the parietals
extend backward along the line of the vein whose course can be
often followed in cynodonts, bounded below by the epipterygoid

6 BREVIORA No. 333

and the prootic region of the braincase. In contrast to conditions
in some other cynodonts, the venous channel and the parietal-
epipterygoid boundary seem to coincide. There is visible here, in
some specimens, an unossified area, presumably cartilage-filled in
life, between parietal above and prootic below. As in cynodonts
generally, the two parietals diverge posteriorly to take part in the
formation of the occipital crests.

Of the series of circumorbital elements, the lacrimal occupies
much of the anterior margin of the orbital rim and extends forward
and upward as in cynodonts generally, to gain a contact with the
nasal as well as having the normal external contacts with the pre-
frontal dorsally, maxilla anteroventrally, and jugal posteriorly at
the lower orbital margin. The bone also forms a portion of the
ossified anterior surface of the orbital socket, between prefrontal
above and jugal below. Ventrally in the orbit, the lacrimal has a
firm connection with the dorsal surface of the palatine. On the
orbital rim, external to the lacrimal foramen, there tends to de-
velop an osseous protuberance, most prominent in the large speci-
mens, the function of which is uncertain. The jugal is, as usual,
highly developed. Dorsally, it forms part of the relatively narrow
postorbital bar; anteriorly, it extends forward below the orbit to
meet the maxilla and lacrimal and internally forms the most
ventral part of the orbital wall. Anterointernally, it gains a con-
tact with the ectopterygoid and pterygoid at the lateral margin of
the palatal plate. As noted above, there is no development of a
suborbital process of the sort seen in Cynognathus, Diademodon,
etc. The main ramus of the jugal extends posteriorly to form a
great part of the substance of the deep zygomatic arch. It reaches
posteriorly and ventrally nearly to the posterior root of the arch,
although sheathed externally over most of its course by the
squamosal. The postorbital forms the upper part of the postorbital
bar, and extends somewhat anteriorly and medially on the skull
roof. Posteriorly on the roof, a pair of flanges from the postorbi-
tals, ridged on their medial margins, extend backward from the
prefrontals, constricting the frontals between them. Below these
dorsal processes, the postorbitals extend backward on either side of
the base of the sagittal crest for some distance, sheathing the
parietals externally. The prefrontal has a modest external exposure
along the upper orbital margin, bounded medially by the nasal
and frontal, anterolaterally by the lacrimal, posteriorly by the
postorbital. Internally, it continues upward the orbital wall
formed more ventrally by lacrimal and jugal and dorsomedialiy
by the frontal.

1969

PROBELESODON LEWISI

Fig. 2. Lateral view of the skull of Probelesodon lewisi.

bs + oc

Fig. 3. Lateral view of the skull of Probelesodon lewisi, with the zygo-
matic arch removed.

The squamosal is one of the largest as well as the most complex
of cranial elements. It consists of two major rami, the two con-
nected by a relatively narrow neck, contrasting with the broader
connection seen in many cynodonts. The external ramus plays a
prominent part in the formation of the zygomatic arch, covering
much of its outer surface and, at the posterior end, forming its
entire thickness. Anteriorly, it ends in a slender external tongue
not far behind the base of the postorbital bar. A short distance
back of this point, it reaches the upper margin and retains this
position for the length of the arch; broken specimens show that

8 BREVIORA No. 333

the upper portion of this squamosal ramus is received in a deep
longitudinal groove in the underlying jugal. The upper margin is
distinctly thickened. Posteriorly, as the arch curves downward and
inward, the upper external surface of this ramus is somewhat
channeled longitudinally, with a mildly concave cross-section.
This channel corresponds to the much more marked channel in
Cynognathus which has been interpreted as an external auditory
meatus. At the posterior portion of the arch, the lower margin of
the squamosal extends downward over the entire width of the jugal
and, extending still further downward, forms a somewhat thickened
flange covering the posterior portion of the mandible. There is no
indication of any specialized surface for articular contact with the
dentary, which lies just internal to this flange.

The occipital ramus of the squamosal forms mainly a thin sheet
of bone which occupies the anterolateral surface of the occipital
crest. At the upper end of the crest, it covers the posterior end
of the parietal, and more distaUy is obviously applied to the an-
terior faces of the postparietal and tabular. On the anterior face
of the occipital crest, the squamosal is separated from the prootic
surface of the otic capsule by an unossified gap, presumably car-
tilage-filled in life, along the line traversed by the vein mentioned
earlier as separating parietal and epipterygoid. Toward the outer
(and lower) part of the occipital crest, this unossified line of
separation increases in breadth. Present here is the anterior open-
ing of the posttemporal fenestra.

In the region of the junction of its two rami, the squamosal
has a complex structure. At the forward curving lower termina-
tion of the occipital ramus, its lateral margin turns ventromedially,
and buttresses the end of the paroccipital process. This area is
somewhat concave in external surface; it is the lower end of the
presumed auditory meatus, and hence the margin of the squamosal
here, together with the adjacent quadrate and articular, may have
afforded support for a tympanum. Anteriorly and laterally to this
area, the squamosal is specialized for support of the quadrate. The
lower margin of the zygomatic arch at its medial end. adjacent to
the flange mentioned earlier, shows the typical pair of notches in
which the quadrate (with the quadratojugal) was loosely inserted.
On the anterior face of the occipital ramus, a thickened area de-
velops at its outer end; this thickened portion descends anterolat-
erally to aid in forming a socket, facing ventrally, in which lay the
main body of the quadrate.

1969

PROBELESODON LEWISI

In posterior view, the occiput is essentially triangular in shape,
with the apex dorsally at the point where the sagittal crest bifur-
cates in formation of the occipital crests, and with a nearly flat
base formed by the lower part of the occipital complex, with the
stout paroccipital processes extending out nearly directly laterally.
The two tall occipital crests diverge at nearly a right angle to
extend backward and outward and, curving downward, terminate
in the area of the jaw articulation. The point of divergence is
situated much farther forward than in most earlier cynodonts. As
noted above, the parietals extend some distance down the occipi-
tal crests. The anterior surface of the crests is formed by the
squamosals; the posterior surface is composed of the postparietal
medially and the paired tabulars laterally. The tabulars extend

Fig. 4. Occipital view of the skull of Probelesodon lewisi.

downward to cover much of the paroccipital processes; the lower
parts of these processes, however, are exposed posteriorly. The
posterior openings of the posttemporal fenestrae lie above the
paroccipitals, within excavations in the ventromedial border of the
tabulars. Sutures between the three dermal elements and the oc-
cipital complex, which occupies the central area of the occiput,
are generally obscure. The occipital elements are fused, without
visible sutures; ventrally the occipitals carry the distinctly paired
condyles and surround the foramen magnum. On either side of
this foramen there are slight swellings which suggest a proatlas
articulation.

The anterior end of the palatal surface is formed by the pre-
maxillae. These bones afford insertion for the "incisor" teeth.
These lie in the form of a widely opened arch. Close to the mid-
line are the long but very narrow incisive foramina, whose boun-
daries appear to be entirely formed by the premaxillae; I see no

10

BREVIORA

No. 333

Fig. 5. Palatal view of the skull of Probelesodoii lewisL

evidence of the presence of the vomer here. From a point back of
the posterior end of the incisive foramina, the sutures between
premaxillae and maxillae run laterally, to become lost in the very
deep pockets for the reception of the large lower canines. Back
of this area stretches the long secondary palate. Approximately
half of this structure is formed by the maxillae. On either side of
the palate the maxillae form a broad base for the insertion of the
cheek tooth series. Anteriorly, the inner margins of the two rows
approach one another somewhat more closely than might be

1969 PROBELESODON LEWISI 11

allowed by muzzle width; posteriorly, the tooth rows curve out-
ward in typical cynodont fashion, so that at the back end of the
dental series the palatal width is about twice that of its anterior
portion. Anteriorly, the tooth bases are close to palatal level;
posteriorly, on either side, the inner margin of the maxilla forms a
distinct shelf, bearing the line of tooth bases well below the level
of the palate. The secondary palate is continued backward to a
remarkable degree by the palatines. The palatine portion of the
secondary palate is somewhat arched and convex ventraUy in
transverse section. Posteriorly, the secondary palatal structure
does not expand laterally; hence the most lateral portions of the
palate here (mainly the ectopterygoid areas) are not involved,
and lie at a somewhat higher level than do the palatines. The
thin posterior margin of the secondary palate is often broken, but
appears to have been somewhat concave in outline as seen from
below. Both maxillary and palatine portions of the secondary
palate show several pairs of small foramina; a particularly large
pair is present at the palatine-ectopterygoid suture.

The secondary palate development here is in strong contrast
to that of all other described therapsids (except other, related,
chiniquodonts ) , in which only the most anterior parts of the pala-
tines are involved. In the extent of the secondary palate, the
chiniquodonts rival mammals.

Beneath the posterior end of the secondary palate can be seen
the posterior end of the vomer. On either side is visible a portion
of the palatine forming part of the primary palate; posteriorly,
there extends a long finger of the palatine overriding the pterygoid.
Lateral to the posterior end of the secondary palate, a small ecto-
pterygoid is persistent, bounded by palatine, maxilla, jugal, and
pterygoid. Medial to the distal end of the maxilla, an extension of
the jugal runs inward and forward to a contact with ectopterygoid
and pterygoid.

There persists a considerable palatal ramus of the pterygoid.
Each pterygoid extends forward medial to the palatine "finger"
mentioned above, to gain contact with the vomer. Lateral to the
palatine there is a second forward extension which gains contact
with the ectopterygoid and, briefly, with the jugal. This region
includes the area of the primitive pterygoid flange. In many
therapsids, notably cynodonts, a ventral process tends to develop
which apparently acts as a guide for the lower jaws. In specimens
of the present species in which these processes are well preserved,
they are more elongate than in any other non-chiniquodont de-
scribed form.

12 BREVIORA No. 333

From the palatine "fingers," ridges run back on either side to
converge medially. Lateral to these ridges, the margins of the
ventral surfaces of the palatal rami of the pterygoids likewise run
back to form ridges parallel to those just described and gradually
converging with them. More laterally, the pterygoids are exposed
in the floor of the orbits and continue backward as a strip of bone
joined dorsally with the epipterygoids. Posteriorly, the pterygoids
fuse in the midline and clamp on either side the anterior end of
the parasphenoid-basisphenoid complex, forming the median ven-
tral "girder" characteristic of advanced cynodonts, which contrasts
with the looser basal connection of palate and braincase present
in more primitive synapsids. On either side of the ventral keel of
this "girder" is a groove, presumably carrying a blood vessel,
bounded externally by a longitudinal ridge. In this ventral region,
sutures between pterygoids, epipterygoids, and basicranial struc-
tures are not clearly seen. It appears, however, that posteriorly
each pterygoid bifurcates, a medial portion sheathing the lateral
surface of the parasphenoid-basisphenoid for a short distance and
the lateral branch swinging outward to accompany the epi-
pterygoid toward the region of the jaw articulation. A thin plate of
bone connecting the two portions forms the floor of the cavity
equivalent to the anterior part of the primitive cranioquadrate
passage. This passage presumably carried the carotid artery,
lateral head vein (if present), and one ramus of nerve VII.

There has been no general agreement as to the relative extent of
epipterygoid and pterygoid in the ramus extending toward and to
the quadrate region. It is generally believed that the pterygoid is
the dominant element here, and that the posterior extent of the
epipterygoid is much restricted. The sutural situation in the
Probelesodon material is obscure, but it seems more reasonable
to conclude that here the quadrate ramus is primarily formed by
the epipterygoid, and that the backward reach of the pterygoid is
limited.

Behind the anterior orbital rim formed by lacrimal and pre-
frontal, there is a major gap in the side wall of the braincase — a
gap filled, of course, in mammals by the orbitosphenoid and ex-
tensions of the adjacent dermal elements. Presumably the optic
nerve, the eye muscle nerves, and nerve Vi, emerged through this
gap. Behind the gap is, as in cynodonts generally, a greatly ex-
panded epipterygoid, lying between parietal and postorbital dor-
sally and pterygoid ventrally. Its anterior boundary is concave as
seen in side view. Posteriorly, it is in contact with the anterior
border of the prootic portion of the ear capsule. This region is

1969 PROBELESODON LEWISI 13

not too well preserved in the material available, but there does
not seem to be any appreciable epipterygoid overlapping of the
prootic, although the suture between the two is an irregular one.
Part way down the line of suture is a foramen which is believed
to have transmitted nerves Vo^^ (plus associated blood vessels).
Ventrally, an extension of the epipterygoid runs backward and
swings outward to the region of the quadrate and squamosal. The
epipterygoid forms the anteroventral wall of the pterygo-paroccipi-
tal foramen, presumably for venous passage.

Although accommodations for housing the quadrate (+ quad-
ratojugal) in the squamosal are present, as described above, this
bone, loosely articulated, tends to drop out when the jaw is dis-
articulated. It is present in the material only in two cases in
which the lower jaws are in place, and in consequence, its struc-
ture cannot be completely made out.

Although, as noted, sutural lines are none too clear in the
region of the median longitudinal bar formed mainly by the
pterygoids, it is obvious that the parasphenoidal rostrum, clamped
more anteriorly between the pterygoids, becomes visible ventrally
in the posterior part of this bar. The anterior end of the para-
sphenoid, possibly with an associated sphenethmoid (presphen-
oid), is surely present internal to the epipterygoids, but I have
not attempted to develop this area of the skull. Behind their point
of ventral emergence, the conjoined parasphenoid and basisphen-
oid expand into a roughly triangular area, with raised edges
and an essentially flat ventral surface carrying a medial ridge.
The anterior portion of this surface is obviously the basisphenoid,
covered ventrally by the parasphenoid; the posterior portion is
part of the occipital complex, but the suture between the two is
generally obscure. On either side of this area is a large jugular
foramen, facing ventrally, and presumably carrying nerves IX-Xl.
I have no certain evidence of hypoglossal foramina. Beyond the lat-
eral rim of the jugular foramen, there is a sharp shift to a laterally
facing surface where there is a well-developed fenestra ovalis.
I am uncertain as to the sutural relations of the opening. Its an-
teroventral boundary appears to be formed by the parasphenoid-
basisphenoid. In well-preserved specimens, the remainder of the
fenestra is bounded by a continuous raised ring of bone, not sub-
divided, which is separated from the adjacent regions of the otico-
occipital complex by distinct grooves. Anterior to the fenestra is
the foramen for nerve VII.

Posterolateral to the fenestra ovalis, there stretches the stout
paroccipital process of the opisthotic. Laterally it ends bluntly. It

14 BREVIORA No. 333

is here buttressed by the squamosal but the tip of the process is
unossified; presumably there was some degree of kineticism here.
As noted above, the posttemporal fenestra passes forward above
the paroccipital process, to emerge anteriorly in a gap between
the squamosal and the more anterior, prootic, aspect of the otic
capsule. Ventroanteriorly, the paroccipital process thins; at its
outer edge it extends downward and forward to reach the posterior
end of the quadrate ramus of the epipterygoid. Continuous with
the paroccipital process, there is seen on the lateral surface of the
braincase (as on the ventral aspect) a forward extension of the
otic capsule, the prootic region. Its thin anteroventral margin
passes forward and medially to form the upper boundary of the
pterygo-paroccipital foramen and then meets, or is covered by,
the lateral extension of the parasphenoid. Laterally, the prootic is
widely exposed. Dorsally and posteriorly, it is separated from the
parietal and squamosal, as noted earlier, by a venous channel;
anteriorly, it is in contact with the posterior margin of the broad
laterally-facing plate of the epipterygoid; ventrally, it continues in
contact with the epipterygoid until the pterygo-paroccipital foramen
is reached.

LOWER JAW

(Figs. 6-7)

The lower jaws are present in three specimens. In all three the
posterior tip is imperfect. The jaw is of the type characteristic of
advanced cynodonts generally. The dentary is dominant, extend-
ing almost the entire length of the ramus. It forms nearly all of
the strong backward-slanting symphysis. The dentary portion
of the lower margin of the ramus is essentially straight. Pos-
teriorly, it terminates at a sharp right angle, as in certain other
advanced forms; from this point the ventral boundary of the
dentary turns sharply upward and then turns backward with a
somewhat thickened lower margin. There is a highly developed
ascending ramus, which dorsally curves backward at a mod-
erately sharp angle; thence the boundary curves forward, down-
ward, and backward to terminate just above, and external to, the
articular region. It is reasonable to believe that the posterior tip
of the dentary was concerned, together with the articular, in jaw
articulation.

On the inner surface of the dentary, a sharp ridge extends for-
ward from the posterior tip of the bone, below which the bone is
excavated for the reception of posterior elements of the jaw; more

1969

PROBELESODON LEWISI

15

anteriorly, the inner surface is grooved for a narrow meckelian
canal which extends from a point just behind the symphysis back
to a point where it is covered by the prearticular. Below, and
partially covering this meckelian groove, is a thin and narrow
splenial, which takes a minor part in the symphysis and extends
backward to a point on the posterior margin of the dentary above
its angular process.

Of other elements of the "normal" reptilian internal surface of
the jaw, there is, in the material, little positive evidence of the
coronoid, which was presumably present in life as a thin flake of
bone on the inner surface of the dentary in the region of the base

Fig. 6. Lateral view of the lower jaw of Probelesodon lewisi.

Fig. 7. Medial aspect of the lower jaw of Probelesodon lewisi.

16 BREVIORA No. 333

of the ascending ramus. Below the assumed position of the coro-
noid is the slender anterior end of the prearticular, which extends
backward, gradually widening, to fuse with the articular. The
surangular is completely covered externally by the dentary. In the
type, the bone is absent, but its position is obvious because there
exists a deep channel for its reception on the inner surface of the
dentary; in a second specimen, the bone is present but cannot be
completely seen. The angular has a modest external exposure.
Internally, the angular is completely covered posteriorly by the
prearticular; more anteriorly, there is a narrow exposure of the bone
below the prearticular, dwindling to a point anteriorly. Ventrally,
just anterior to the point where the dentary covers the bone ex-
ternally, a small ventral prong is the last remnant of the reflected
lamina of the angular, so prominent in the therapsid ancestors. At
the posterior end of the jaw, fused to the prearticular, is the ar-
ticular, essentially an oval structure presenting an articular cup
facing posterodorsally, and with a modestly developed retroarticu-
lar process.

DENTITION

The dentition is of an orthodox cynodont type. There are four
upper incisors in each premaxilla, cylindrical, long, and slender,
and set close together in a somewhat arched transverse row; they
are slanted strongly backward below the pointed snout. Below, on
each side, are three similar incisors directed forward as well as
upward. There are well-developed canines both above and below,
sharply pointed and slightly recurved; the upper ones are some-
what the stronger. In the holotype, the canines are in process
of replacement. Several of the larger specimens have nine upper
cheek teeth; the holotype, presumably somewhat immature, has
eight, with a ninth erupting posteriorly on one side; an "infantile"
specimen has eight. The lower cheek teeth are well seen only in
the holotype and in an "infant." In the better preserved of the
two rami of the type, there are seven teeth, with, apparently, an
eighth developing posteriorly. In the "infant," the lower dentition
is not too well preserved, but only six cheek teeth are seen.

The cheek teeth have single roots which are not deep, as wit-
nessed by the fact that in the fossil specimens, they are frequently
displaced or lost. (In the related Belesodon, for example, all
teeth had been lost in the two known specimens, so that one was
not even sure whether one was dealing with a carnivore or a
gomphodont; in several of the specimens of the Chanares form, a

1969 PROBELESOUON LEWISI 17

considerable portion of the cheek teeth had been lost before
burial.) The teeth are of a common cynodont type, laterally com-
pressed, sharply pointed, and with a tendency for the development
of a fore-and-aft cusp row. As in a number of thrinaxodontids and
related chiniquodontids, there is a strong trend for a backward
curvature of the major cusp. There is some individual variation,
but, in general, in the first two upper teeth — and sometimes in the
third as well — there is only a single cusp, curved backward to a
slight degree. Back of this, in the upper jaw, there is an increasing
development of a posterior accessory cusp, associated with a fore-
and-aft lengthening of the tooth. In the only good specimen
available of the lower jaw dentition, the development of the pos-
terior cusp is initiated on the second tooth present. In the upper
jaw, there tends to be a development of a small anterior cusp,
somewhat toward the inner surface. In the one well-preserved
lower jaw ramus, an anterior cusp is present on teeth four to six.
I have seen no tendency for the development of accessory
"cingulum" cusps, such as might be expected in forms related to
mammal ancestry.

THE FAMILY CHINIQUODONTIDAE

In 1956 Watson and I arranged the carnivorous cynodonts in
three families in morphologically progressive order — Procyno-
suchidae, Thrinaxodontidae, Cynognathidae. At that time, we
included the then-known chiniquodontids in the Cynognathidae,
since they were obviously more advanced than the thrinaxodonts
in such characters as the development of a free angle of the
dentary. In 1956 Huene erected the family Chiniquodontidae
for Belesodon and Chiniquodon, but gave no definition. These
two forms (particularly Chiniquodon) were then very poorly
known; I have elsewhere (1969) given additional data on these
two genera which show that they, plus Probelesodon, form a
natural group, clearly distinct from the Cynognathidae and that
they deserve, as Huene felt, segregation as a separate family of
advanced cynodonts. They share with the better known advanced
cynodonts such features as the great development of the dentary
and corresponding reduction of the other jaw elements. But there
are notable differences from the cynognathid condition. For ex-
ample, they lack the broad posterior connection of zygomatic
arch and occipital crest seen in cynognathids, and have retained
a more primitive situation here. Again, they lack the longer snout

18 BREVIORA No. 333

and proportionately narrow skull of typical cynognathids, and pre-
serve the shorter snout and very broad zygomatic region seen in
thrinaxodontids. It is probable that for chiniquodontid ancestry
we must look to a descent from thrinaxodontids parallel to that of
the cynognathids. SinognatJius (Young, 1959), from the early
Triassic of China, may be antecedent, although the palate is less
developed.

The unique chiniquodontid feature, never attained by known
cynognathids or thrinaxodontids, is the remarkable elongation of
the secondary palate. This is a very advanced character unrivalled
in any other described therapsids, and paralleling the mammalian
condition. We have noted above the fact that the dentary is close
to, if not actually in contact with, the squamosal; but neither in
the present genus nor in the Santa Maria chiniquodonts is there
any development of special squamosal-dentary articular surfaces.
In described chiniquodonts, as in cynodonts generally, there re-
mains a large unossified area in the braincase wall, to be filled in
by orbitosphenoid-presphenoid development before a mammalian
condition is reached. Probelesodon and its allies may be close to
the line of ascent toward mammals, but there is still a considerable
distance to go.

CYNODONT ANCESTRY

Above, we discussed the relationship of Probelesodon to other
cynodonts. It may not be amiss to discuss in addition the un-
settled question of the ancestry of the Cynodontia as a whole.

In 1956 Watson and I proposed that the therapsids be grouped
in two large suborders, the Theriodontia and Anomodontia —
the former term covering the main carnivore groups, such as the
Gorgonopsia, Cynodontia, Therocephalia, and Bauriamorpha; the
latter term, used in a broad sense, for the herbivores — the
herbivorous Dinocephalia, Dromasauria, and Dicynodontia. This
proposal appeared eminently reasonable and has, I believe, been
rather generally accepted in most regards. However, some years
ago (1961) I pointed out one shift which seemed necessary. It
seems certain that the ancestral therapsids, of sphenacodont de-
scent, were carnivores, and hence Watson and I included in the
Theriodontia the earliest and most primitive carnivorous therapsid
group then recognized, the Titanosuchia or carnivorous Dino-
cephalia. But study shows that nearly all the forms usually in-
cluded in the Titanosuchia possess the peculiarly "shouldered"
anterior teeth found in the herbivorous Dinocephalia. It thus

1969 PROBELESODON LEWISI 19

appears that, with this speciaHzation already developed, the titano-
suchians cannot be placed in the Theriodontia as a basal therap-
sid group but must instead be considered as ancestral anomodonts.
Where, then, shall we find truly ancestral therapsids? Such types
known to me in 1961 included only Phthuiosuchus and a few
other poorly known forms from the early Middle Permian of Rus-
sia. In skull structure, Phthinosuchus had attained therapsid
status, but barely so, and is in many regards closely comparable
to the ancestral sphenacodonts. I therefore erected for this genus
and its relatives the suborder Phthinosuchia as a basal therapsid
stock potentially ancestral to such theriodont groups as the
Therocephalia, Gorgonopsia, and Cynodontia, and ancestral,
through the Titanosuchia, to the Anomodontia as well.

At about this time Boonstra, whose work has been concentrated
in great measure on the reptiles of the Tapinocephalus Zone of
South Africa, visited Moscow, familiarized himself with the Middle
Permian faunas of Russia, and, on his return home, published
(1963) an excellent paper on the early phylogeny of therapsids.
Although my work of 1961 was not known to him when he wrote,
he had come to very much the same conclusions that I had as
regards the interrelationships of early therapsids. He recognized
the truly primitive position of Phthuiosuchus and its relatives, and
erected a separate basal group for them. His attention, however,
was concentrated not so much on Phthinosuchus as on Eotitano-
suchus (Chudinov, 1960), description of which was not yet avail-
able when my 1961 paper was prepared, and he hence used the
term Eotitanosuchia rather than Phthinosuchia.

It seems certain, then, that the Cynodontia are of ultimate
phthinosuchian (or eotitanosuchian) derivation. But were they
directly derived, or is their descent to be traced through one or
another of the familiar theriodont groups? In earlier decades, the
Gorgonopsia were rather generally thought to be cynodont an-
cestors. Such ancestry was long favored by Watson (1920, 1951).
In a number of regards, the gorgonopsians represent a primitive
morphological stage antecedent to that of the cynodonts (and, as
a minor point, are the only therapsid group apart from the
cynodonts in which the primitive 2.3.4.5.3 phalangeal formula is
known to have been retained). But there are few indications
among gorgonopsians of any trend toward a cynodont condition;
the gorgonopsians seem to have been, so to speak, "frozen" in a
primitive theriodont morphological pattern, and, in addition, uni-
versally retain such gorgonopsian "trademarks" as the preparietal
bone and a reduced cheek tooth series.

20 BREVIORA No. 333

If the gorgonopsians are excluded, where can we turn? In re-
cent decades the Therocephaha, or rather that advanced series
of therocephalians termed the "scaloposauroids" (which Watson
and I preferred to group with their Triassic descendants, the
Bauriamorpha), have been favored as cynodont ancestors. The
scaloposauroids show various advanced characters. The skull is
lightly built, there is a trend toward a secondary palate, the
dentary is well developed, and so on. These trends lead toward the
advanced condition seen in Bauria; but, it has been suggested, may
there not have been a second advanced line leading to the cyno-
donts? To be sure, most scaloposauroids tend to be long-snouted
forms with a long tooth row, with, in general, little differentiation
of canines, and with the characteristic therocephalian-bauriamor-
phan "trademark" of large palatal vacuities. However, reversal
in such features might have occurred and if transitional forms were
to be found, belief in a scaloposauroid ancestry of the cynodonts
would attain credibility.

Broom, although early disposed to a gorgonopsian ancestry,
later (1938) inchned toward a therocephalian origin. Brink has
been a recent major advocate of cynodont descent from thero-
cephalians, specifically the scaloposauroids (Brink, 1950, 1951).^
Since Brink is the only writer who has attempted to give broad
consideration to the theory of therocephalian ancestry of cyno-
donts, his arguments deserve careful consideration.

Brink's attention was attracted to some small skulls named 5//-
phedestes and Silphedocynodon , which are in general scaloposaurid
in type, but in which postorbital and zygomatic arches are absent.
Whether this absence in the skulls as preserved is due to their
absence in life or to post-mortem damage is not certain. The
presumed absence of arches has, of course, no bearing on possible
cynodont relationships, since these arches (slender in scalopo-
saurids) are highly developed in cynodonts. Brink (1951: 340)
considered that Silphedestes (as well as Silphedocynodon) was
"truly a cynodontid therocephalian, to be more cynodont than
therocephalian and that it should be grouped rather under the
Cynodontia than under the Therocephaha."

^ Unlike Watson and myself. Brink considers the scaloposaurs to pertain
to the Therocephalia proper rather than to the Bauriamorpha. In this section
I have followed him in using "scaloposaur" and "therocephalian" inter-
changeably for the presumed scaloposaurid ancestors of the cynodonts.

1969 PROBELESODON LEWISI 21

In the papers cited above, Brink failed to give specific reasons
for assuming the relationship of cynodonts to these forms (or to
scaloposauroids in general). A further specimen, which Brink
believed strongly confirmed his belief in the scaloposauroid origin
of the cynodonts, was a small skull from the Cistecephalus Zone,
which he prepared by the grinding method and reconstructed, and
named Scalopocynodon (Brink, 1961). It had been damaged
posteriorly before burial. Here, as in the silphedestids, zygomatic
and postorbital arches are absent, but again, whether this is a
natural condition or is due to pre-burial damage is uncertain.
The animal is definitely and clearly a cynodont, as Brink agrees,
although a cynodont of a primitive sort. It is closely comparable
to the procynosuchid cynodonts and differs markedly from scalo-
posauroids in numerous regards, such as the secondary palate, the
dentiticn with cusped "molars" and highly developed canines,
broad alisphenoids, absence of the large palatal fenestrae of
scaloposauroids, a cynodont double condyle, and so forth.

On the other hand, Brink lists 17 features which he believes to
indiciUe the relationship of Scalopocynodon to the Therocephalia.
I shall consider these 17 points in some detail, using Brink's
numeration.

Several points may be immediately ruled out. The absence of
postorbital and zygomatic arches (2, 3) and the "feeble struc-
ture" of the posterior ends of the squamosals (11), if "natural,"
are suggestive of scaloposauroids, particularly the supposedly
archless silphedestids. However, this has nothing to do with cyno-
donts, in which stout arches are universally present. Cited as a
potential scaloposauroid character is the possible entrance of the
frontals into the orbital margin (17); however, conditions here
are uncertain.

Cited as further evidence for scaloposauroid relationships are
various primitive features lost or modified in typical cynodonts
but retained in therocephalians and bauriamorphs. These include:

(1, 5, 7). The presence of an interpterygoid vacuity and the
consequent similarity of the pterygoids of the specimen to those
of therocephalians (p. 144). No interpterygoid vacuity is present
in advanced cynodonts, where pterygoids are firmly fused to the
basicranial axis; such vacuities are present, however, not merely in
scaloposauroids but in all primitive therapsids. The contrast of
the parasphenoidal keel and rostrum to those of cynodonts is,
of course, due to the fact that the parasphenoid-basisphenoid is,
primitively, still free from the pterygoids, rather than fused into
the basicranial bar.

22 BREVIORA No. 333

(4). In contrast to cynodonts, and like scaloposauroids, the
pterygoids are tooth-bearing. This is merely a primitive character,
retained in titanosuchians and gorgonopsians as well as thero-
cephalians.

(6). "The transverse bones contribute rather substantially to
the pterygoid processes."' This is a general primitive therapsid
feature. The area of the "transverse" (ectopterygoid) in the
specimen being considered is about the same as in, for example,
the titanosuchian Titanophoneus and the gorgonopsian Saurocto-
nus ( Romer, 1 96 1 , fig. 11).

(9). The posterior face of the skull had undergone damage
before burial and is not figured by Brink. He states, however, that
it is therocephalian-like in the shallowness of the occiput and in
the fact that the occipital crests are low and not very sharp. This,
however, is a generally primitive character; the impression I get
from his description is that of, for example, Efremov's figure
(1954, fig. 5 1 ) of Phthinosuchus.

(12). "The parietal region is not sharply crested." This is
merely a primitive feature (cf. for example, Phthinosuchus [Efre-
mov, fig. 49]).

(13). "The pineal is situated far back." It does not appear
to be particularly far back, but a posterior position is a primi-
tive synapsid feature (cf. Phthinosuchus, Efremov, fig. 49).

(15, 16). The prefrontals extend well forward and laterally,
preventing a lacrimal-nasal contact. Such a contact, seen in cyno-
donts, is unique among theriodonts; its absence is primitive, not
an exclusive therocephalian feature.

Apart from this series of primitive characters retained in
Scalopocynodon as well as in scaloposauroids, there are three
further points mentioned by Brink as supporting therocephalian
relationship which are of dubious importance. (8). The sella
turcica is shallow in Scalopocynodon , and where known in cyno-
donts is better developed. However, this region is known in but
few therapsids, and the significance of variance here is obscure.

(10). One paroccipital process is preserved in Brink's skull;
he states that it is shorter and slenderer than that of cynodonts.
It appears to be rather shorter proportionately than in cynog-
nathids, but it does not appear to differ notably from the paroccipi-
tal process in other cynodonts for which I have available material
or illustrations.

(14). Brink agrees that the alisphenoid (epipterygoid) is as
broadly developed as in typical cynodonts but without as much
overlap of the prootic, and further, the bone is not as high as is

1969 PROBELESODON LEWISI 23

proper in cynodonts. However, the difference in prootic overlap
does not appear great, and I see no appreciable difference in
alisphenoid height between Scalopocynodon and typical cyno-
donts.

To sum up, such a primitive cynodont as Scalopocynodon shows
a number of archaic therapsid characters also retained in scalopo-
sauroids. But I fail to see that Scalopocynodon furnishes the
slightest evidence that scaloposauroids have anything to do with
cynodont origins. I find it difficult to believe that long-snouted
scaloposaurs with a nearly homodont dentition, large palatal
vacuities and, as far as known, a reduced phalangeal formula of
2.3.3.3.3, should rapidly reverse their evolutionary trends, re-
acquire stout canines, close their palatal vacuities,' re-acquire lost
phalanges, and rather abrupdy take on various other cynodont
characters. It is much more reasonable to believe that the cyno-
donts evolved directly from primitive therapsids of the Phthino-
suchus-Eotitanosuchus type without any close affiliation with other
theriodont groups. The only argument against such a belief is
the negative one that no pre-cynodonts are known from the
Tapinocephalus and Endothiodon zones. Possibly the ancestral
forms were rare or possibly they were absent from the Karroo
before Cistecephalous Zone times — at which time we know that
communications between South Africa and Eurasia were open,
and that primitive cynodonts were already present in the pre-
Ural region of Russia.

REFERENCES CITED

Bonaparte, J. F.

1966. Chiniqiiodon Huene (Therapsida — Cynodontia) en el Triasico
de Ischigualasto, Argentina. ActaGeol. Lilloana, 8: 157-169.

BOONSTRA, L. D.

1963. Early dichotomies in the therapsids. S. Afr. Jour. Sci., 59: 176-
195.

Brink, A. S.

1950. Notes on a second specimen of Homodontosaurus kitchingi. S.
Afr. Jour. Sci., 47: 118-119.

1951. Studies of Karroo reptiles. I. Some small cynodonts. S. Afr.
Jour. Sci.. 47: 338-342.

^ Whaitsiids, it is true, are forms which have lost the palatal fenestrae
typical of therocephalians: note, however, that they retain the typical thero-
cephalian ectopterygoid extending back along the pterygoid flange in the
same position it occupied when the fenestra was present.

24 BREVIORA No. 333

1961. A new type of primitive cynodont. Pal. Africana, 7: 119-154.

Broom, R.

1938. The origin of the cynodonts. Ann. Transvaal Mus., 19: 279-288.

Chudinov, p. K.

1960. [Upper Permian therapsids of Ezhovo location.] Paleont. Zhurn.
Akad. Nauk SSSR. 4: 81-94.

Efremov, I. A.

1954. [Permian vertebrate fauna in USSR.] Trudy Palaeont. Inst.,
Akad. Nauk, 54: 1-416.

HUENE, F. V.

1944. Die fossilen Reptilien des siidamerikanischen Gondwanalandes.

Munich, 332 pp.
1956. Palaontologie und Phylogenie der niederen Tetrapoden. Jena,

716 pp.
ROMER, A. S.

1961. Synapsid evolution and dentition. Internat. Colloq. on the
Evolution of Mammals. Kon. Vlaamse Acad. Wetensch. Lett.
Sch. Kunsten Belgie, Brussels, I: 9-56.

1969. The Brazilian Triassic cynodont reptiles Belesodon and Chini-
cjiiodon. Breviora, Mus. Comp. Zool., No. 332: 1-16.

Watson, D. M. S.

1920. On the Cynodontia. Ann. Mag. Nat. Hist., Ser. 9, 6: 506-524.
1951. Paleontology and Modern Biology. New Haven: Yale Univ.
Press, 216 pp.

Watson, D. M. S., and A. S. Romer

1956. A classification of therapsid reptiles. Bull, Mus. Comp. Zool.,
114: 35-89.

Young, C. C.

1959. Note on the first cynodont from the Sinokannemeyeria-faunas in
Shansi, China. Vert. Palasiat., 3(3): 124-131.

(Received 13 June 1969.)

rVfl-irs- W.:i^cl P.Q NOrCIKCU.ATE

B R E V I O W'¥°

JAN 5 1970
Museiiim of Comparative Zoology

Cambridge, Mass. 30 December, 1969 Numbe^'^:^^^'"^^

POSTLARVAL SUDIS (PISCES: PARALEPIDIDAE)
IN THE ATLANTIC OCEAN*

David L. Shores

Abstract. The paralepidids, SiiJis atrox Rofen 1963 and Siidis hyalina
Rafinesque 1810, are present in midwater collections from the Atlantic Ocean.
Postlarval development is described and illustrated with particular attention
given to head spination. This character is similarly developed in both
species, but it differs in form and degree of head coverage (more extensive
in 5. atrox) providing a basis for species differentiation.

INTRODUCTION

The WHOI midwater trawl collections (Fig. 1) contain numer-
ous postlarval paralepidids referable to the genus Sudis. Exami-
nation of these shows that two species are represented. One, Sudis
hyalina Rafinesque 1810, has been known only from the North
Atlantic and the Mediterranean. The other, referable to Sudis
atrox Rofen 1963, has been known only from the North Pacific.
This note describes the posdarval development of S. atrox, com-
pares this development with that of S. hyalina, and records the
Atlantic distribution of these species.

STUDY MATERIAL

The study material consists of 132 specimens of S. atrox and 32
specimens of 5. hyalina. Collection data is shown in Tables 1 and
2 with the distribution of catches plotted in Figure 1. The speci-
mens were caught by midwater trawls with no opening-closing
mechanism. The majority of the specimens (153 out of 164)
came from tows shallower than 250 meters and show that the two
species can be expected in any collection at this depth or shallower.

=■= Contribution No. 2319 from the Woods Hole Oceanographic Institution.

o»

BREVIORA

No. 334

Accordingly, the "no catch" records in Figure 1 (squares)
show the distribution of tows that were at appropriate depths for
catching the two species but which nevertheless took no specimens.
Figure 1 shows that S. hyalina is a northern species and 5. atrox is
a tropical one, with the two occurring together in the Gulf of Mex-
ico. S. hyalina has also been found in our collections from off the
Rio de la Plata (ca. 32°S) and thus a biantitropical distribution is
suggested. The two specimens of S. atrox collected near 38 °N,
67 °W are probably Gulf Stream waifs.

Siidis atrox was described by Rofen (1963) from a 75 mm SL
Pacific Ocean specimen taken from the stomach of Alepisaurus

m NO CiTCH

• SutfiS OtrOM

X Sua<i hyohro

■ 90TH SPECIES

Figure 1. Atlantic distribution of Siidis atrox Rofen and Sudis hyalina
Rafinesque, based on captures listed in Tables 1 and 2. "No catch" records
are all trawls shallower than 250 m that did not take specimens.

1969 POSTLARVAL SUDIS 3

caught on longline gear at 29°26'N, 134°59'W. Pacific postlarval
specimens were reported by Berry and Perkins (1966) with a
drawing of a 21.5 mm SL specimen. The adult specimens of the
two species differ by vertebral counts (59-60 for S. hyalina vs.
53-54 for 5. atrox) and by the number of lateral line pores (four
to seven per each lateral line section for S. hyalina and three for
5. atrox) (Rofen, 1966). My specimens have been identified by
vertebral counts obtained from X-rays of representative specimens
(60-61 vertebrae for five S. hyalina and 52-54 for 15 S. atrox) and
by the postlarval characters of pigment patterns, degree and type
of spination, and pectoral fin length.

DEVELOPMENTAL CHARACTERS
SUDIS ATROX

The development of S. atrox closely resembles that of its con-
gener, S. hyalina. The pectoral fins are long and spines are present
on the head. The S. atrox head spination is so extensive in develop-
ment and unique in form that it can be used to identify postlarval
specimens. The spines are located on the head as follows:

Preoperculiim. Large, obvious spine at the posteroventral angle,
with serrated edges and hooked processes along its length. Two
large processes at its base, one directed posterodorsally and the
other ventrally. Ventral and posterior edges of the preoperculum
with spine-tipped scalloped flanges. Ventral flange obvious; poste-
rior one small and inconspicuous.

Dorsal surface of head. Two spine-tipped scalloped flanges
above eye, a shorter one directed laterally and a longer one di-
rected dorsally. Paired spine-tipped scalloped flanges medial to
supraocular flanges on dorsal surface of head from tip of snout
posteriorly to a line between anterior margins of orbits.

Mandible. Spine-tipped scalloped flange along side of lower jaw.

Shoulder. Two spines.

These spiny processes are observed in the smallest specimen
(5.5 mm SL) and are most extensively developed in the 10 to
20 mm specimens. They are worn down and indistinct in speci-
mens larger than 30 mm (Table 3).

The development of Sudis atrox is also distinguished by six peri-
toneal sections, by the late development of the elongated pectoral
fin, and by the early development of widespread trunk pigment.

5.5 mm SL (Fig. 2A). Head deep, top involuted with a ridge
on each side. Ridge with a conspicuous spine-tipped scalloped

BREVIORA

No. 334

flange. Spiny flange along ventral edge of preoperculum and
three relatively large spines at the posteroventral preopercular
angle. The middle one of these three spines, which points postero-
ventrally, is truncated. It is apparently broken off; another speci-
men of approximately the same size (6.0 mm SL RHB 1717) has
this middle spine longer and stouter than the other two with an

A-5.5mm

B-9.0mm

C -14.4mm

D- 25.0 mm

Figure 2. Postlarval stages of Sudis atrox Rofen. A. RHB 1107. B.
RHB 1291. C. RHB 1108. D. RHB 1294. Drawn by Martha Howbert.

1969 POSTLARVAL SUDIS 5

antrorse hook ventrally located near the spine tip (shown in Fig.
2 A in broken Hne). Eye compressed, obliquely oriented. Pectoral
fin rounded, short, on short stalked base; rays evident. Dorsal and
anal fin bases barely discernible in embryonic finfold. Peritoneal
pigment extending two-thirds length of trunk from pectoral base to
a point over anterior section of rudimentary anal fin. Anterior
half of this pigment diftuse; posterior half partitioned into three
sections. Dense cluster of small pigment cells on base of preoper-
culum angle spine.

9.0 mm SL (Fig. 2B). Head deep, massive, flat on top, sharply
angular between top and sides. Snout length and head length rela-
tively longer than in 5.5 mm specimen (50.0% vs. 38.0% for
snout length and 46.6% vs. 32.0% for head length). Four canine
teeth on lower jaw. Head studded with spine-tipped scalloped
flanges that are well developed along top of snout, over eye (two
flanges here — one directed laterally and one directed dorsally),
along lower jaw, and on preoperculum ventral edge. Also a large
spine extending back from posteroventral angle of preoperculum,
with serrated upper and lower edges and two prominent antrorse
hooks (Fig. 4A). This spine has a triangular cross section large
at base and decreasing in size out to tip, with the two antrorse
hooks stemming from laterally directed apex or 'keel' of this tri-
angular cross section.

Fins well formed with interspinous rays. Pectoral fins short,
extending only one third of distance to ventral fin base. Six well
formed peritoneal sections going back to anal fin origin. Pigment
patches at jaw tips, scattered cells on lower jaw, small ones on
preoperculum posterior margin and two large stellate chromato-
phores on base of preopercular angle spine. Large chromatophores
spread evenly over cerebral hemispheres, one of the same over
hind brain, and several between the two brain sections.

14.4 mm SL (Fig. 2C). Head form as in 9.0 mm specimen.
Three canines on left side and four on right. Pectoral fins reaching
halfway to ventral bases. Dorsal and anal fins with 10 and 20
rays, respectively. Spines about head as in 9.0 mm specimen, ex-
cepting more spine tips in the flanges (Table 3) and two obvious
spines added to preoperculum ventral edge, apart from spiny flange
already described for that area. Spine deterioration indicated in
dorsally directed supraocular flange in which spine tips are rela-
tively blunt. Spine deterioration suggested in preopercular angle
spine in which the hooks are smaller in relation to the total size
of this spine than is observed in 9.0 mm specimen.

6 BREVIORA No. 334

Pigment about head as in 9.0 mm specimen, excepting thicker
coverage of large chromatophores over brain, with dense arc of
same over anterior edges of cerebral hemispheres. Trunk with
broad grouping of large stellate chromatophores on dorsum be-
tween nape and dorsal fin origin, patch of chromatophores at dorsal
origin, and grouping of three pigment cells over hypural plate.
Six peritoneal sections.

25.0 mm SL (Fig. 2D). General body form essentially as be-
fore with large deep head on relatively short body. Three canines
left and four right. Pectoral fins long, reaching beyond ventral
bases; 15 rays. Pigment cells proximally and medially on rays
1 and 2; patches medially and distally on rays 3-7. Other fin ray
counts: Dorsal — 10; Anal — 21; Ventrals — 9.

Head spination not apparent. Spines in scalloped flanges fewer
(Table 3). Spines, when present, worn down as in supraocular
flanges, where eight spines in lateral flange and thirteen spines in
dorsal flange are reduced to rounded irregularities in what appear
to be only eroded or broken ridges over eye.

Body pigment extensive, evenly spread on dorsum and sides of
same spread over hypural plate. Anterior portion of lateral line
ventrally edged with small pigment cells. (Lateral fine not pig-
mented.) Fin bases with varying degrees of pigmentation: large
light chromatophores on roots of dorsal and anal fins; small dark
cells at base of ventral rays; and large, moderately dark pigment
cells forming vertical line at base of caudal fin. Six peritoneal
sections.

38.0 mm SL (Not figured; RHB 1281). Head and body essen-
tially as before. Head deep, massive. Four canines. Pectoral fin
broken off; 15 rays. Dorsal — 12 rays; Anal — 21 rays; Ventrals
— 9 rays. Vestiges of post larval spination present but very small
and obscure. Primary character is trunk pigment, which is denser
and slightly more extensive than in 25.0 mm specimen. Lower
margin of lateral line pigmented for entire length (no pigment over
lateral line). Vertical bar at caudal base darker, more apparent.
Six peritoneal sections.

SuDis Hyalina

The development of S. hyalina closely parallels that of S. atrox.
The body form is essentially the same, with a relatively short body
and a large head. Spines are present over the eye and on the pre-
operculum lower margin.

1969

POSTLARVAL SUDIS

o

X

a

a

X)

c

o

CO

PC
J"

s
s

3

cr

u

c

cs

>

o

a.

(U

8 BREVIORA No. 334

16.1 mm SL (Fig. 3). Head large (31.6% of SL) but not as
deep and massive as in S. atrox. Trunk short but not as deep and
stout appearing as in 5. atrox. Pectoral fin very long, extending
beyond ventral bases and anus; small pigment cells near bases
of first five rays, larger blotchy cells further out on rays 2-7.

Spination includes eight spine tips on scalloped flange over eye,
six very small spine tips on posterior half of flange running length
of snout, five spine tips on scalloped flange along preoperculum
ventral edge, and a relatively large spine extending posteriorly from
posteroventral angle of preoperculum (Fig. 4B). This latter spine
has a retrorse hook distally (giving spine a bifurcated appearance)
and a second larger retrorse hook at base.

Trunk pigment concentrated in patches, with six patches on
dorsum (three anterior to dorsal fin, one at dorsal origin, one at
dorsal base, and one posterior to dorsal base), one patch midway
on anal fin base, and two over hypural plate above and below
notochord. Eight peritoneal sections. Large stellate chromato-
phores over brain. Small pigment cells on jaws and margins of
preoperculum.

COMPARISON

The distinctive characters of S. atrox are pointed out with refer-
ence to S. hyalina.

1. Spination. Spine-tipped flanges present on both 5. a/ro.y and
S. hyalina on preoperculum ventral edge, over eye, and along pos-
terior part of snout. Additional distinctive spines present on 5.
atrox along lower jaw, anterior part of snout flange, and in a sec-
ond spiny flange over eye. The large spine at preopercular angle
in S. atrox is distinctive for its serrated edges and one or more
antrorse hooks, whereas this spine in S. hyalina is smooth with
only one retrorse hook (Fig. 4).

2. Pectoral fin. Short in early postlarval 5. atrox, reaching ven-
tral origin in three specimens at 16.5 mm (RHB 1289), 17.3 mm
(RHB 1297), and 19.5 mm (RHB 1281). In contrast, the pec-
toral fin of S. hyalina is long, very early in development reaching
ventral origin in an 8.0 mm specimen (RHB 1043) and extending
beyond ventral base and anus in a 9.0 mm specimen (RHB 1047).

3. Pigment. Six peritoneal sections in S. atrox vs. seven to
eight in 5. hyalina. Trunk pigment in S. atrox evenly spread over
trunk surface vs. patch pattern distribution in S. hyalina. 14.4 mm
S. atrox (described in text) — pigment grouping covering middle

1969

POSTLARVAL SUDIS

half of nape to dorsal fin area, plus pigment patch at dorsal fin
origin and three small chromatophores over hypural plate. 16.0
mm S. atrox (RHB 1277) — uniform coverage of dorsum back
to point halfway between dorsal fin and procurrent caudal rays

Figure 4. Preopercular spines. A. Sudis atrox Rofen, 9.0 mm bL, RHB
1291. B. Sudis hyalina Rafinesque, 16.1 mm SL, RHB 1307. Drawn by
Martha Howbert.

10 BREVIORA No. 334

plus hypural pigment. 19.5 mm S. atrox (RHB 1281) — even
coverage of dorsum back to procurrent caudal rays all but con-
nected to and continuous with hypural pigment. (See also pigment
coverage in 25.0 mm S. atrox, Fig. 2D.) In contrast, 5. hyalina
trunk pigment concentrated in nine patches in specimens from
16.1 mm SL (See specimen described and figured above.) to
36 mm SL. Patches clearly defined in 28 mm specimen (RHB
1307), spread out but still recognizable in 33 mm specimen (RHB
1439), and all but blended into uniform coverage in 36 mm speci-
men (RHB 1313).

RELATION OF SUDIS TO MAN

Bites on polypropylene mooring cables serious enough to sever
the individual fibers have been reported off Bermuda and were
connected with fish through tooth fragments left on the cable
(Turner and Prindle, 1965). Siidis hyalina has been identified as
the fish responsible for certain bites in depths of 600 to 800 meters
(Haedrich, 1965). The bites have occurred in pairs on one side of
the cable only and presumably have been made by the long canine
teeth on either side of the lower jaw. The distance between cuts of
a pair range from 30 to 60 mm. Reference to the tooth arrange-
ment of a 405 mm S. hyalina with an 18 mm distance between
dentaries places the size of the cable-biting Sudis at more than a
meter in length. These bites might just as well have been caused
by S. atrox as by S. hyalina. The mandibular canine teeth of both
species are long and strong. The fragments in the rope have ser-
rated edges matching the teeth of adult 5. hyalina. The teeth of
the 75 mm juvenile holotype of 5. atrox have smooth edges (Rofen,
1963) as does a 75 mm specimen of S. hyalina (Rofen, 1966).
Until teeth of larger specimens of S. atrox are obtained, the ques-
tion of the origin of the bites described by Haedrich (1965) re-
mains unanswered.

ACKNOWLEDGEMENTS

My thanks go to Richard H. Backus, James E. Craddock, and
Richard L. Haedrich, Woods Hole Oceanographic Institution, for
their encouragement and critical reviews in preparing this paper.
The work was supported by NSF Grant GB-7108 (R. H. Backus,
principal investigator).

1969 POSTLARVAL SUDIS 1 1

LITERATURE CITED

Backus, R. H.. J. E. Craddock, R. L. Haedrich, and D. L. Shores

1969. Mesopelagic fishes and thermal fronts in the western Sargasso
Sea. Marine Biology 3 (2): 87-106.

Berry, F. H., and H. C. Perkins

1966. Survey of pelagic fishes of the California current area. Fishery
Bull. Fish and Wildl. Serv.. 65 (3) : 664-665.

Haedrich, R. L.

1965. Identification of a deep-sea mooring-cable biter. Deep-Sea
Res., 12: 773-776.

Rofen, R. R.

1963. Diagnosis of new genera and species of alepisauroid fishes of
the Family Paralepididae. Aquatica, No. 2: 5-7.

1966. Family Paralepididae. In: Fishes of the Western North Atlantic.
Mem. Sears Found. Mar. Res., 1 (5): 448-459.

Turner. H. J., Jr., and B. Prindle

1965. Some characteristics of "fishbite" damage on deep-sea mooring
lines. Limnol. and Oceanogr., 10, supplement: R259-R264.
November.

(Received 20 August 1969.)

12 BREVIORA No. 334

TABLE 1 — STUDY MATERIAL
Sudis atrox Rofen, 1963

Size

No.

Range

Ma.x.

Coll.

of

(SL)

Depth

No.

Specs.

(mm)

Position

Date

(m)

581

1

8

19°21'N,

66°20'W

2 Mar.

54

90

962

1

12

5°24'N,

39°55'W

13 Feb.

63

860

967

1

8

0°26'N.

32°06'W

18 Feb.

63

60

972

3

20-21

0°03'N,

25°00'W

23 Feb.

63

87

981

1

9

5°42'S,

32°25'W

1 Mar.

63

120

1101

2

25-38

I3°18'N,

60°05'W

13 June

65

187

1105

1

15

15°42'N.

64°13"W

14 June

65

230

1107

4

6-12

19°52'N,

69"46'W

16 June 65

140

1108

1

14

19°55'N,

70°04'W

16 June

65

525

1200

1

14

16°13'N,

59°09'W

1 1 Feb.

66

450

1202

4

10-11

13°00'N,

49°38'W

14 Feb.

66

220

1216

1

9

1°12'N.

44°39'W

1 1 Apr.

66

410

1218

1

21

9°12'N,

47°03'W

16 Apr.

66

360

1222

2

8-22

13°55'N.

57°00'W

30 Apr.

66

300

1253

4

10-17

16°38'N.

64°27'W

25 May

66

133

1258

2

28-35

13°32'N,

7r24'W

27 May

66

210

1261

1

24

13°04'N.

73°12'W

28 May

66

300

1263

3

14-43

12°58'N,

73°34'W'

29 May

66

120

1274

8

8-12

13°07'N,

78°23'W

5 June

66

109

1277

21

10-17

16°12'N,

78°00'W

6 June 66

198

1281

2

20-38

17°08'N.

79°35'W

6 June

66

205

1282

5

11-18

17°07'N,

79°32'W

7 June

66

185

1289

4

17-35

21°11'N.

85°12'W

10 June

66

170

1290

2

12-15

2ri7'N,

85°22'W

11 June

66

124

1291

2

9-12

22°54'N.

9r36'W

12 June

66

57

1294

3

15-25

20°48'N,

95°48'W

17 June

66

100

1297

13

15-33

23°13'N.

94°50'W

19 June

66

128

1298

3

19-22

23°55'N.

94°00'W

19 June

66

370

1307

1

13

27°0rN,

90°02'W

22 June

66

95

1310

1

14

26°12'N.

87'54'W

23 June

66

2150*

1423

1

29

12°2I'S,

3r04'W

1 Mar.

67

90

1425

3

5-10

16°07'S,

29°59'W

3 Mar.

67

85

1427

2

15-18

18°10'S,

29°40'W

5 Mar.

67

140

1428

4

6-8

18°21'S.

29°39'W

5 Mar.

67

75

1432

2

10-11

23°08'S.

32°22'W

9 Mar.

67

110

1505

2

10-10

37°40'N.

66°50'W

25 Aug.

67

105

1713

6

8-11

25°06'N.

67°45'W

1 Dec.

68

125

1716

1

15

23°24'N.

67"30'W

1 Dec.

68

140

1717

6

6-10

23°00'N,

67"29'W

2 Dec.

68

30

1718

3

6-28

22°51'N,

67°30'W

2 Dec.

68

210

1969

POSTLARVAL SUUIS

13

6 Dec. 68 65

8 Dec. 68 87

1727 I 12 26°46'N. 67"32'W

1736 1 13 ZS^l-VN. 67 24-W
R/ V 'Anton Rniiin' Cr. 19

Sta. 813 1 45 1 r'06'N, 7S21'W 7 Aug. 66 2250^=

^Oblique tow

The specimens in collections 1107 and 1108 were the first Atlantic speci-
mens of 5. airo.x to be reported (Backus et al., 1969).

TABLE 2 — STUDY MATERIAL
Sudis hyalina Rafinesque, 1810

Size

No.

Range

Max.

Coll.

of

(SL)

Depth

No.

Specs.

(nun)

Position

Date

(ni)

861

1

28

41°40'N.

61°57'W

15 Sept.

62

64

866

1

11

39°20'N,

66°45'W

18 Sept.

62

68

867

2

17-20

39°03'N,

67°18'W

19 Sept.

62

110

1022

1

16

42°35'N,

45°56'W

10 Sept.

64

50

1043

1

8

39°28'N,

31°00'W

25 Sept.

64

35

1047

1

9

39°25'N,

36°56'W

27 Sept.

64

52

1050

1

11

39°31'N,

43°33'W

29 Sept.

64

53

1263

1

16

12°58'N,

73°34'W

29 May

66

120

1290

2

13-14

21°17'N,

85°22'W

1 1 June

66

124

1307

10

16-48

27°01'N.

90°02'W

22 June

66

95

1313

1

36

23°55'N,

83°12'W

25 June

66

145

1315

1

12

25°46'N,

79°47'W

26 June

66

71

1438

1

30

32°53'S,

46°06'W

16 Mar.

67

110

1439

2

21-33

32°56'S,

46°12'W

16 Mar.

67

40

1508

1

34

36°27'N,

67°18'W

26 Aug.

67

190

1706

1

13

30°10'N,

67°32'W

28 Nov.

68

217

1727

1

7

26°46'N,

67°32'W

6 Dec.

68

65

1737

1

14

28°45'N,

67°26'W

8 Dec.

68

135

Captain Bill III', C

;r. 65-1

Sta. MWT-

1 2

14-18

38°11'N,

69°47'W

11 May

65

90

14

BREVIORA

No. 334

O
to

"■■A

■^
CO

D
W)

C^

-4— »

1/3

>

■*— »

(/3
O

CO

OQ
<

c

3
O
O

QJ

c
'5-

V3
C

'-4— »

O

(U

•4— >

T3

c
c

13
>

(D

c
"5-

175

2
CO

(N

"a

o
a

V.

ex.
S
03

to

UJ

1 1^

o\

ov

vo

5j -;:;
0

to
O

a-

_c

'q-

an

to

c

o

w-1

vC

HJ

c

'a.

en

73
U

■4— <
03

u

0

u

u

k.

D.

<u

in

c«

^

^

a

r

0

1/1

•a

%

0

X)

(N

(N

U

CT\

•4-1

c«

c

(U

<N

c

m

in

c

en

^

C/1

0

c«

vn

■0

X)

c

u

u.

_c

0

^

"5-

c

in

^

en

0

c3

m

-0

X)

c

u

k-

C

0

^

'Sh
«5

c

tn

^

tn

0

rt

73

X

00

c

U

u

c

0

en

00

<s

q
as

1)
c

'S.

en

T3

X

tn en
TO oj
X en

T3
C

nj en

en x; .^

o S

O c«
J= X

•<:1-
■^

O c

S -a

w

,

s

<A

en

3

_>^

0X3

>

rt

en

^—

C

tu

0

«

ffi

C<0 So

u

en

euDx:

w

c

■*~t

0

c

c«

%

c

'5-

E

en

en OX)

fN

3

OX) f

^ !>,> c

•- C CIh

^ C O

T3 <=

O
u

u

c

s

-0

0

X

c
0

X

u

3

c

X

Q-

"^

00

c3

00

DO NOT CIRCULATE

B R E V I O R "A*"''

JAN 5 1970

Mmseiiiinii of Comparsitive ZooIoJ^ard

JJMNEBSlTfi

Cambridge, Mass. 30 December, 1969 Number 335

TWO NEW LATE CRETACEOUS FISHES
FROM MONTANA AND WYOMING'

Richard Estes

Abstract. Two new fossil fishes from the late Cretaceous Hell Creek
Formation of Montana are described, based on disarticulated material. One
of these fishes is also present in the late Cretaceous Lance Formation of
Wyoming.

Palaeolabriis montanensis, n. sp., is the first North American record of a
genus previously recorded only from the late Paleocene or early Eocene
Dormaal deposits in Belgium. Originally described as a labrid teleost, new
material from the Hell Creek Formation suggests that it may be a specialized
amioid holostean. However, premaxillae and vertebrae of teleostean type
may be referable to Palaeolabriis; if the association is correct, it could pos-
sibly be a very primitive teleost.

Coriops amnicoliis, n. gen., n. sp., is referred to the teleost family Albu-
lidae. Originally known only from the Lance Formation of Wyoming, it is
now also recorded from the Hell Creek Formation in Montana.

INTRODUCTION

The disarticulated remains of two distinctive fossil fishes have
been obtained from a collection of lower vertebrate fossils from
Bug Creek Anthills, McCone County, Montana. The fossils occur
in the late Cretaceous Hell Creek Formation (Sloan and Van Va-
len, 1965).

easier (1967) described Palaeolabriis dormaalensis from the
late Paleocene or early Eocene of Belgium, basing his description
on what he believed to be pharyngeal tooth plates. These plates
are interpreted here as vomers, and similar elements in the Bug
Creek Anthills sample are supplemented by a number of other
disarticulated elements probably referable to this genus, indicating
the presence of Palaeolabrus in the Hell Creek Formation.

^ Fossil vertebrates from the late Cretaceous Hell Creek Formation, Mon-
tana: Contribution No. 7.

2 BREVIORA No. 335

In 1964, I referred basibranchial and parasphenoid tooth plates
from the Lance Formation of Wyoming to an undescribed genus
of Albulidae. Subsequent work has revealed the presence of this
fish in the Hell Creek Formation of Montana, and study of a related
albuloid family, Phyllodontidae (Estes, 1969), indicates that these
dentitions are distinctive enough to warrant a name.

Morphological comparisons given below indicate that the two
groups of disarticulated elements noted above belong to different
taxa. The Lance Formation form, here referred to a new genus
Coriops, is much rarer at Bug Creek than in the Lance Formation,
and is represented by (in general) smaller specimens; conversely,
Palaeolabrus does not occur in the Lance Formation localities.

ORDER AMIIFORMES?

Family Palaeolabridae, n. fam.

lype species: Palaeolabrus dormaalensis Cosier 1967

Palaeolabrus montanensis, n. sp.

Type. MCZ 9343, tooth plate, probably vomerine.

Paratypes. MCZ 9342, vomerine (or anterior dermopalatine?)
tooth plates; 9339, 9341, 9344, right dentary fragments; 9380,
posterior dermopalatines; 9347, 9352, entopterygoids(?); 9345,
9351, first (symphysial) coronoids; 9348, coronoid fragments;
9452, prearticular; 9346, 9353, right and left premaxillae; 9350,
questionably-referred vertebrae.

Type locality. Bug Creek Anthills, west half of section 9, T 22
N, R 17 W, McCone County, Montana; collected by A. D. Lewis
and party, 1964.

Etymology. Latin, montanensis, from Montana.

Diagnosis. Distinguished from the only other known species.
Palaeolabrus dormaalensis, from the late Paleocene or early Eocene
of Belgium, by its generally larger size and uniformly styliform
teeth lacking any development of an apical point.

Description. Presumed vomerine plates subtriangular, entirely
covered with styliform crushing teeth ventrally, each with a tiny
enameloid tip that is flattened and slightly depressed below the
crown surface (PI. 3d); dorsally, assumed anterior and lateral
borders thickened for articulation with (presumably) premaxiUa
and maxilla or dermopalatine (PI. la-f).

Posterior dermopalatine (ectopterygoid) strongly twisted, wid-
ened at anterior end, tapering posteriorly in both known specimens;
teeth styliform as on vomers, larger anterolaterally, becoming much

1969 LATE CRETACEOUS FISHES 3

smaller posteromedially; both specimens show extensive wear and
both are broken at both anterior and posterior ends (Pi. li-j).

Presumed entopterygoids with a single row of marginal (medial)
tooth bases; laterally, teeth appear only as scattered, isolated tooth
bases except posterolaterally, where they become grouped in re-
gion of presumed contact with posterior dermopalatine (ectoptery-
goid). Dorsally a ridged, flattened surface present for presumed
contact with braincase bones; anterior end of bone pointed, rapidly
expanding posteriorly to a flat sheet; aU specimens broken (PI.
2a-b).

Anterior coronoid (first or symphysial) with marked symphysial
expansion; several rows of styliform teeth like those of above ele-
ments present; posterior coronoids with deeply concave articulating
surfaces; lingual surface of posterior coronoids uniformly convex
and covered with the same styliform teeth as on palatal elements
(PI. Ig-h)

Prearticular represented by only a fragment of the posterior end,
showing semilunar border at edge of mandibular fossa, and a small
process curving dorsally toward the dentary; teeth as on palatal
elements.

Dentary massive, with flattened coronoid articulation surface on
medial side, and open area anteriorly for symphysial expansion of
first coronoid; a strong anele formed in the dentary about at level
of fifteenth tooth; posteriorly, dentary truncate, with a very small
opening for vascular and nervous structures; dentary tooth row
single, teeth acrodont; only tooth bases preserved, the latter with
wrinkled, irregular borders as on palatal elements; MCZ 9344 with
nineteen tooth bases in the complete row; ventrally, dentary weakly
sculptured anteriorly, prominent sculpture begins at level of angle
of dentary; sculpture deeply pitted; laterally, a deep fossa for liga-
mentary attachment of maxiUa present near end of tooth row; a
lateral row of sensory canal pores present; tiny "coronoid" process
of dentary present behind last tooth in row (PI. 2d-e, 3a-b).

Referred premaxillae with weak exterior sculpture; anterior end
of bone expanded, roughened, flattened medially, and somewhat
laterally-directed; on MCZ 9346, twenty tooth bases present in
a single row, with one extra medial tooth base anteriorly; on
MCZ 9353, two tooth bases present in the medial row; posterior
end of premaxilla excavated dorsally, elongated, and tapering to a
point posteriorly (PI. 2c).

Questionably-referred vertebrae deeply amphicoelous; promi-
nent excavations laterally for the arch bases; neural arch processes
projecting above centrum (PI. 3e-f).

4 BREVIORA No. 335

Generic reference. Casier (1967) described Palaeolabrus dor-
maalensis from the Dormaal deposits in Belgium, of latest Paleo-
cene or earliest Eocene age. The type specimen is a bone that
Casier believed to be of pharyngeal origin; he therefore referred
Palaeolabrus to the Labridae. The Hell Creek Formation elements
are here interpreted as vomers or perhaps anterior dermopalatines
(see below), and they closely resemble those of the Belgian form
(cf. PI. 1, a-b and c-f). P. dormaalensis is smaller than P. nionta-
nensis, and its teeth are also more pointed. This may be only a
size-related difference not of specific significance, but some Bug
Creek specimens almost as small as the Belgian form do not have
pointed tooth crowns. The strongly-ridged anterior and assumed
lateral borders of these bones are very similar in the two forms
and leave little doubt of the proper generic reference of the Mon-
tana form.

Association of elements referred to Palaeolabrus montanensis.
Association of these presumed vomerine (or anterior dermopala-
tine) tooth plates with the posterior dermopalatines, coronoids,
and prearticular is clear; the teeth on these elements are all of
identical styliform appearance, with a small, slightly depressed cap
of enameloid material at the tip (PI. 3d). MCZ 9345, a well-pre-
served first (symphysial) coronoid, fits the dentary MCZ 9344
almost perfectly; this seems to confirm association of these two
bones, as does the similar wrinkled appearance of the tooth bases
of both specimens. The wrinkled tooth bases are also present on
the premaxillae and presumed entopterygoids.

Affinities of Palaeolabrus. The unquestionable mutual associa-
tion of the elements referred here to vomers, posterior dermopala-
tines, and coronoids makes it clear that Palaeolabrus does not be-
long to the advanced teleost family Labridae. The presence of
well-developed toothed coronoids and paired vomers precludes
inclusion in any known teleost group, although a very primitive
teleost might well have them. The general appearance of the den-
tary itself (discussed further below) and the fact that the actual
symphysis lies on the coronoid are conditions suggestive of the
amioids, and these similarities are also present in Amia itself. In
the dentary the general shape, short tooth row, open symphysial
region, enclosed sensory canals, and recessed labial notch near
the "coronoid process" (to receive the posterior end of the max-
illa) are resemblances to amioids, especially the Amiidae. The
strong angulation of the dentary is common in short-jawed "holo-
steans" but is also compatible with the above resemblances to

1969 LATE CRETACEOUS FISHES 5

amiids. Tooth histology of Palaeolahrus and its contemporary
Aiuia jragosa (also from the Hell Creek Formation; Estes and
Berberian. 1969) indicates general similarity. The specializations
often occurring in the orthodentine tissues of teleosts are absent;
only simple orthodentinal tubules are present.

The bones of the palate also show amioid resemblances. The
posterior dermopalatines are relatively much shorter than those of
Amia, but the curvature, or torsion, and the presence of a strong
lateral ridge dorsally, are resemblances to the Recent genus. Some
primitive teleosts, such as elopids, also show this condition, how-
ever. The presumed entopterygoids resemble those of Amia in hav-
ing a narrow anterior end, and a flattened, ridged area dorsafly for
attachment of braincase bones (or cartilages). They differ, how-
ever, in having only a single, curved, medial row of teeth, whereas
the entopterygoids of Amia have a large patch or shagreen of teeth
on this bone. This tooth patch in Amia does, however, have a
curved medial row of enlarged teeth that suggests similarity to the
Hell Creek form. The presumed entopterygoid may have been, in
life, more denticulate in Palaeolabrus than it seems, for palatal
teeth are often only irregularly fused to the underlying bones in
Amia, and in macerated specimens a patchy distribution of teeth
may result.

The type specimens of the two species of Palaeolabrus are clearly
paired bones, as Casier (1967) realized. In considering the possi-
bility of an amioid affinity for Palaeolabrus, my first interpretation
of these bones was that they were a fused group of posterior supe-
rior pharyngeal tooth patches like that of Amia, which has a similar,
subtriangular shape (Nelson, 1969, pi. 82, fig. 1). In Amia, how-
ever, these patches are formed of a number of discrete, rounded
groups of teeth connected by cartilage, and a small, rounded
branchial bone with a prominent dorsal process is incorporated
into this patch. No such structures or process is suggested in the
type elements of Palaeolabrus. The dorsal appearance of the tooth
plate is smooth, rather like that of the anterior dermopalatines of
Albula. On the other hand, while the long posterior process of the
Amia vomer is not present on the presumed vomer of Palaeolabrus,
the enlarged anterior and presumed lateral ridges of this element
are suggestive of the premaxillary and anterior dermopalatine at-
tachment surfaces of the vomer of the Recent genus (PI. Ic, e).
Tentatively, then, this element is considered a vomer, but it could
possibly be an anterior dermopalatine.

6 BREVIORA No. 335

Several unique characters and some features dissimilar to amiids
now require comment. The premaxillae, referred here to Palaeo-
labrus on the basis of similarity of tooth bases to those of the den-
tary, are completely unlike those of Amia, and strongly resemble
those of such primitive teleosts as Elops and (to lesser degree)
Megalops in posterior elongation and absence of the enlarged pos-
terodorsal process of this bone that occurs in Amia and other "holo-
steans." An elopid does occur in the Hell Creek Formation and
Lance Formation fauna, and is represented by a few scales, ver-
tebrae, dentaries, and fragments of sculptured skull bones. These
elements, aside from their sculpture, suggest a Megalops-\\\iQ form
(Estes, 1964; Estes, Berberian, and Meszoely, 1969). The denta-
ries show that this elopid, while not identifiable as to species, was
probably prognathous as in other elopids, and that the jaw margins
were covered with several rows of teeth unlike those of Palaeo-
lahrus. There is little chance, then, that the premaxillae referred
here to Palaeolabrus belong to this elopid, quite aside from the rea-
sons noted above for associating them with Palaeolabrus.

The vertebrae tentatively referred here to Palaeolabrus are also
unlike those of amiids or most other "holosteans." They most
closely resemble teleost vertebrae, although some amioids, e. g.
Oenoscopus (Saint-Seine, 1949, pi. 18c), have similar centra.
Known Bug Creek Anthills teleosts were relatively too small to have
vertebrae of the size referred here to Palaeolabrus. It is, of course,
possible that a large teleost, other than the elopid noted above, is
present and is (so far) unrepresented on the basis of cranial ele-
ments. An alternative position is that Palaeolabrus is like some
other advanced "holosteans" in having a number of progressive,
teleostean resemblances; Amia itself is quite advanced toward the
teleostean level. An extension of this view is that Palaeolabrus is
a very primitive teleost, perhaps at the pholidophoroid level, a sug-
gestion I owe to Colin Patterson (in litt., 1969).

Cretaceous freshwater fishes are poorly known and most of them
are represented only by disarticulated elements. Since most of their
marine relatives are preserved as articulated, flattened specimens,
it is rarely possible to make detailed comparisons of individual cra-
nial elements, and identification of the freshwater material is there-
fore often difficult. If the association of elements suggested above
for Palaeolabrus is correct, then it is an interesting and significant
fish, showing some striking advances toward the teleosts in verte-
bral and premaxillary structure. I do not imply, however, that the
Palaeolabrus material indicates relationship of primitive teleosts

1969 LATE CRETACEOUS FISHES 7

and amioids. While the general resemblances of Palaeolahrus seem
to be amioid. there are too many differences from the Amiidae,
from other known amioid families, and from known teleosts, to
place it in any known family. Palaeolahrus should therefore be
placed in its own family, Palaeolabridae. n. fam., for which the
definition is the same as for the genus. Based on the amioid re-
semblances of the elements that are clearly referable to Palaeo-
lahrus, I include it tentatively in the Order Amiiformes. This
questionable reference is difficult to maintain if the elopid-like pre-
maxilla and the teleost-like vertebrae are properly referred (as I
believe that they are), but inclusion in any known teleost group on
the basis of the known material is much less defensible.

ORDER ELOPIFORMES

SUBORDER ALBULOIDEI

Family Albulidae

Cor/ops amn'icolus, n. gen., n. sp.

Type. American Museum of Natural History (AM) 9317, para-
sphenoid with attached tooth plate; collected by M. C. McKenna
and party, 1960.

Paratypes. AM 9319, four basibranchials and five parasphenoids
with attached tooth plates; AM 9301, anterior end of parasphenoid,
showing anterodorsal process; AM 9320, pterygoid fragment; other
University of California Museum of Paleontology (UC) specimens
cited by Estes (1964:49).

Type locality. UC loc. V5620, Lance Formation, Lance Creek,
Niobrara County, Wyoming (Clemens, 1963).

Etymology. Greek, korios, bug (from the fact that specimens of
this fish occur at Bug Creek Anthills); tops, a small fish; Latin,
amnicolus, a river dweller.

Referred specimens. MCZ 9340, five broken parasphenoids with
attached tooth plates. Hell Creek Formation, Bug Creek Anthills,
west half of section 9, T 22 N, R 17 W, McCone County, Montana,
collected by A. D. Lewis and party, 1964. AM 9318, basibranchial
with attached tooth plate, and AM 9316, ?pterygoid tooth plate
fragment. Lance Formation, UC loc. V5711, Niobrara County,
Wyoming, collected by M. C. McKenna and party, 1960.

Diagnosis. An albulid with teeth fused to their attachment areas,
basibranchial with well-defined branchial arch attachment areas;
parasphenoid with very smaU otic articulation areas; parasphenoid
tooth plates with anteroposterior channel rather than arched as in
Alhula.

8 BREVIORA No. 335

Description. Parasphenoid bone compact (PI. 4a), not arched
anteroposteriorly; bone surface coarsely vascularized dorsally; a
prominent anterodorsal process present, broken on all specimens
in varying degree except in AM 9301; strong "parethmoid" articu-
lation areas present anterolaterally; strong interorbital septum ar-
ticulation area present on midline; posterolateral surfaces of bone
prominently concave; posteriorly, paired slots present posterolater-
ally for (probably) prootic, apparently no articulation area for
basioccipital; teeth styliform, bluntly-pointed, without enameloid
tips; tooth-bearing surface of parasphenoid anteroposteriorly-chan-
nelled, concave. Basibranchial tooth plate with teeth similar to
those of parasphenoid (PI. 3c), bone convex from side-to-side
with essentially straight occlusal border; basibranchial tooth plate
teardrop-shaped in small specimens, with the point at the anterior
end, larger specimens becoming more symmetrically oval; ventrally,
hyoid arch attachment areas well separated; anteriorly a channel
present, probably for basihyal plate, followed by paired antero-
posterior ridges that probably clasped the basihyal; more
posteriorly, an expanded, shallow channel present, probably for
basibranchial I; posterior to this a smooth, expanded, and flattened
area occurs that probably covered more posterior basibranchial
elements (PI. 4b).

Discussion. No name was originally proposed by Estes (1964)
for this fish because of difficulty in comparing specimens of isolated
tooth plates and scales, with entire fishes. More recent study
(Estes, 1969) has shown that most records of fossil albuloid fishes
are based on tooth plates, and that criteria exist for distinguishing
the various types regardless of lack of association with whole fishes.
The plates described here can be distinguished easily from other
albuloid dental plates by the articulation pattern of braincase or
palate bones, or hyoid attachment scars, and by the presence of
fused teeth.

The bone articulation patterns on parasphenoid and basibran-
chial are similar to those oiAlbula, especially in the case of the para-
sphenoid. The major difference from Albula is the absence of
extensive otic or basioccipital articulation areas (cf. Estes, 1969,
figs. 5, 6g); probably these bones (in part) clasped the sides of the
parasphenoid, which is more excavated in this region than in that
of Albula. While the basibranchial scars are identifiable with fair
certainty, there are small differences from those of Albula and the
phyllodontids (Estes, 1969, fig. 2, cf. Pi. 4b, this paper).

1969 LATE CRETACEOUS FISHES 9

The pattern of articulation surfaces on both parasphenoid and
basibranchial bones suggests reference to the Albulidae. The ex-
istence of only two very closely related genera of living albulids
makes it difficult to say whether or not the differences signify more
than generic separation from the Recent forms, but 1 believe that
the evident similarities justify allocation to the Albulidae at present.

Little tooth replacement is evident in the dentition of Coriops;
this feature and its fused teeth separate it clearly from the related
phyllodontids. Although many specimens of Coriops are present
in the Lance Formation sample, only two fragments of pterygoid
tooth plates have been identified. They show torsion of the tooth
plate as in Albula, but no other distinctive features are visible. The
presence of pterygoids separates Coriops from the phyllodontids,
and supports allocation to the albulids.

ACKNOWLEDGEMENTS

I am grateful to Colin Patterson, David Bardack, and Gareth
Nelson for helpful comments on the possible relationships of Palae-
olabrus. Mr. Fred Maynard prepared the photographs. This re-
search was supported in part by NSF grant GB-7 1 76.

LITERATURE CITED

Casier, E.

1967. Le Landenien de Dormaal (Brabant) et sa faune ichthyologique.
Mem. Inst. Roy. Sci. Nat. Belgique, 156: 1-66.
Clemens, W.

1963. Fossil mammals of the type Lance Formation, Wyoming.
Part I. Introduction and Multituberculata. Univ. Calif. Publ.
Geol. Sci., 48: 1-105.

ESTES, R.

1964. Fossil vertebrates from the late Cretaceous Lance Formation,
Eastern Wyoming. Univ. Calif. Publ. Geol. Sci., 49: 1-180.

1969. Studies on fossil phyllodont fishes: Interrelationships and evo-
lution in the Phyllodontidae (Albuloidei). Copeia, 1969, no.
2: 317-331.

ESTES, R., AND P. BERBERIAN

1969. The status of Amia (=Kindleia) fragosa (Jordan), a Cretaceous
amiid fish, with notes on related European forms. Mus. Comp.
Zool., Harvard Univ., Breviora no. 329: 1-14.
EsTEs, R., P. Berberian, and C. Meszoelv
In press. Lower vertebrates from the late Cretaceous Hell Creek Forma-
tion, McCone County, Montana. Mus. Comp. Zool., Harvard
Univ., Breviora no. 337: 1-33.

10 BREVIORA No. 335

Nelson, G.

1969. Gill arches and the phylogeny of fishes, with notes on the clas-
sification of vertebrates. Bull. Amer. Mus. Nat. Hist., 141:
479-552.
Saint-Seine, P. de

1949. Les poissons des calcaires lithographiques de Cerin (Ain).
Nouv. Arch. Mus. Hist. Nat. Lyon, 2: vii + 357 pp.
Sloan, R. and L. Van Valen

1965. Cretaceous mammals from Montana. Science, 148: 220-227.

(Received 22 August 1969.)

1969

LATE CRETACEOUS FISHES

11

METRIC

TTTT

PLATE 1

a — /', Palaeolahriis dormaalensis Casier, 1967, right vomer (?), Paleogene,
Belgium, an Institut Royal de Sciences Naturelles de Belgique specimen not
figured in original description, specimen broken posteriorly, c — /, P.
montanensis, n. sp., late Cretaceous, Hell Creek Formation, Montana, c —
d, dorsal and ventral views of type left vomer (?), MCZ 9343; e — /, the
same, paratype right vomer (?), MCZ 9342; g — h, lateral and medial views
of right symphysial coronoid, MCZ 9345; / — /, dorsal and ventral views of
right posterior dermopalatine, MCZ 9380. Arrows = assumed anterior (a)
and lateral directions; all X 2.

12

BREVIORA

No. 335

1969 LATE CRETACEOUS FISHES 13

V) ra u (Ai

o ^ ^ 3 X

O C (ij

ca 'I- o - •"

> „ .« >f ^

- ^ o ^ ^

> C c«

° S 3 s s

O o ^ <u

c c o « ^

■^3 -i-p ~a iS _-

1-^ C/5 ON t|_l ^

< ^ ^ « i s

S a: 0 -r^ :£ <

- ^> ^ 3

§ I E o^

s « - =^ ?;

^ « ^ ^ S

<U „r "O O '

^

'•3

-J"

■♦-»
B

d.

0)

OJ

'^

«3

E

4-»

c

M-l
0

ON

N

•0

■4— »

en

.2

cj"

(/I

.2i

U

u.
«^^

§

'>

0

S

.0

(U

0
04

1^ =

2 <u

0

c
>

c
■a

u

0

5

(U

0

T3
C

a

c

03

"0

_r

0

%)

C«

G3

T3

^

&

u

^

&

.Si

u

_u

=2

>

03

>

14

BREVIORA

No. 335

METRIC 1

PLATE 3

o, h, d, Palaeolabriis montanensis, n. sp., late Cretaceous, Hell Creek For-
mation, Montana, a, ventral and b, dorsal views of composite complete
right dentary, anterior end = MCZ 9344, posterior end = MCZ 9341; d,
much enlarged view of unworn teeth from type left vomer (?), MCZ 9343.
c, Coriops amnicoliis, n. gen., n. sp., late Cretaceous, Lance Formation,
Wyoming. Much enlarged view of unworn teeth from basibranchial tooth
plate, AM 9319. In a — b, specimen of posterior end of dentary reduced
slightly to fit anterior end, the latter X 2, c — d,Y. 15, e — f X 2.

1969

LATE CRETACEOUS FISHES

15

PLATE 4

Coriops amnicoliis, n. gen., n. sp., late Cretaceous, Lance Formation, Wyo-
ming, a, ventral (occlusal), dorsal, and right lateral views of type para-
sphenoid tooth plate, AM 9317, anterodorsal process restored in lateral view
from AM 9301, locality V5620. b (left), dorsal and (right), ventral views
of four basibranchial tooth plates to show proportional changes with in-
creasing size, top three specimens = AM 9319, V5620, bottom specimen =
AM 9318, V5711; all x 2. Dotted lines = interpreted branchial arch at-
tachment areas, BHP = basihyal plate, BH =r basihyal, Bl = first basi-
branchial, PB = posterior basibranchial elements. Posterior end of bottom
hree specimens broken, anterior to the right.

IM^'-

B R E V I O R™A--°"

IMiuseiuinti or L^oimpsirsitive //Oology

HARVARD

Cambridge, Mass. 30 December. 1969 Number UtWlVERSITYi

THE STATUS OF THE EAST AFRICAN ELEPHANT
ARCHIDISKODON EXOPTATUS" DIETRICH 1942

Vincent J. Maglio

Abstract. The syntype collection of "Arcliidiskodon cxoptcitus" Dietrich
1942 from LaetoHl, east Africa, is reexamined in the light of more recent
material from other African localities. The collection is found to he com-
posed of two distinct taxa — one referable to Elepluis rccki Dietrich 1916
and the other to a primitive species of Lo.xodoiita. In the interest of nomen-
clatorial stability, an E. rccki specimen is chosen as the lectotype of "A.
cxoplatiis,' thus reducing that name to the junior synonymy of E. rccki. It
is concluded that two faunal horizons are represented in the Laetolil area
as suggested by earlier workers, one correlating with both Kanapoi and Yel-
low Sands at the base of the Omo sequence, and the other correlating with
the later Omo beds, possibly antedating Bed I Olduvai by a short time
interval.

INTRODUCTION

Since its description by Dietrich in 1942, the name "Arcliidisko-
don exoptatiis" has been applied to several different taxa by differ-
ent workers. As a result, the original concept of the species is often
misunderstood. Recent concepts have varied from a rather primi-
tive, low-crowned form conspecific with the Upper Siwalik A. plan-
ijroiis (Arambourg, 1947) to a considerably more progressive,
higher-crowned form either ancestral to E. recki (Cooke. 1960) or
synonymous with it as an early stage (Leakey, 1965). Because
of the growing significance of the Elephantidae for purposes of
correlation, especially in the Plio-Pleistocene of Africa, it is im-
portant to establish firmly the status of this taxon in order to avoid
further confusion.

The syntype collection of "Arcliidiskodon exoptatus" was col-
lected by Kohl-Larsen during his 1938-1939 expedition to the
southern Serengeti in what was then Tanganyika Territory. The
fossil localities cover an extensive area south of Olduvai Gorge in
the Vogel River area of the Serengeti Plain, just north of Lake

2 BREVIORA No. 336

Eyasi. The fossil-bearing Laetolil beds are exposed in the drainage
valleys of five river systems — Vogelfluss, Gadjingero, Deturi.
Oldogom, and Marambu. The yellow-grey Laetolil tuffs are locally
interrupted by a basalt (Kent, 1941), which may represent a fair
interval of time. Based on Hopwood's ( 1936) analysis of the Lae-
tolil fauna, Kent suggested two distinct faunal horizons — one
more or less contemporary with Beds Lll at Olduvai, and the other
somewhat earlier in the Lower Pleistocene.

Many of the fossils are yellowish white in color and chalky in
preservation. Other specimens are brown to black in color and
are more highly mineralized. Dietrich considered this difference
in preservation, along with morphological differences, as evidence
indicating two faunal zones — the "old fauna" and "younger
fauna" — thus supporting Kent's earlier view. The yellowish
white specimens are the older, the black the younger. He consid-
ered the older fossils as products of redeposition, being mixed with
material of a considerably later age. Despite this, however. Dietrich
believed the entire collection of Proboscidea to be uniform and to
represent a single, variable species. Whereas Hopwood (1936)
and Kent (1941) recognized two elephantid species in this mate-
rial — Palaeoloxodon recki and Elephas aff . planifroiis — Dietrich
considered the Laetolil collection to be a single new species re-
lated to the "E. planifrons-E. nieridioiuilis" group, but represent-
ing a distinct African branch. As I will show below, this view of
a single species derived from E. planijrons cannot be supported
on present evidence. Hopwood's original analysis was essentially
correct.

The syntype collection on which Dietrich founded his species
consists of 108 molar fragments, nearly all of which are fragmen-
tary or severely worn. Most of these are too incomplete for mean-
ingful diagnosis. Among the 108 specimens the following were
identified by Dietrich: 12M', 9Mi, 5M-, 12M. 4M"'. 24Mo. The
remainder of the collection consisted of milk molars. A reexami-
nation of this collection in the light of the now abundant com-
parative material from other east African localities shows that a
large number of these determinations were incorrect. The mixing
of two distinct taxa as well as the misidentification of individual
specimens as to their serial position in the tooth row resulted in
a specific diagnosis which had little objective relatioaship to any
real taxon. The reasons for this confusion lie not in Dietrich's
analysis of the collection, which generally was excellent, but pri-
marily in the fragmentary nature of the material and the lack of

969

"ARCHIDISKODON EXOPTATUS'

adequate comparative collections at tiie time he wrote.

The type collection, housed in the Institute for Paleontology,
Humboldt University, and six specimens in the British Museum
(Natural History) collected several years earlier by Dr. L.S.B.
Leakey, have been examined. The material conclusively shows
the presence of two taxa, which may be distinguished on morpho-
logical grounds and which differ in preservation. These two forms

155
150

145
140
135
130

125

120

II 5

X 110

uj 105

I

2 100

O 95
a:
o
90

85

80

75

70

65

60

55

50

mm

u

to (D C> C^ ^ iO
— — cJ oJ OJ oJ

CO O cvj ^ VO
OJ lO lO rO ro

00 O ro e
ro *f V £

ENAMEL THICKNESS

Figure 1. Distribution of first, second, and third permanent molars in
the syntype collection of "Archidiskodon exoptatiis" with respect to enamel
thickness and crown height. Group A and B as discussed in text. For each:
thin line, observed range; short crossbar, mean; solid rectangle, ± one
standard deviation from the mean; open rectangle, 98% confidence interval
of the mean.

4 BREVIORA No. 336

support the contention that two faunal horizons are present. These
can be tentatively correlated with other east African deposits for
which acceptable radiometric age determinations are now available
(Maglio, in press, a).

The following analysis is based on 37 of the better specimens in
the Kohl-Larsen collection and the six British Museum specimens.
In Fig. 1 the first, second, and third permanent molars are plot-
ted on a scatter diagram, with average enamel thickness as the
abscissa and maximum crown height as the ordinate. Two groups,
A and B, separate out on both morphological and metrical criteria.
These two groups may be distinguished as follows:

Group A Group B

Enamel thin and folded Enamel thicker and smooth
Height/width index for per- Height/width index for per-
manent molars 120-200 manent molars 85-1 10
1 1-13 plates on M3 9-1 1 plates on M3
Weak to absent anterior si- Prominent anterior sinus
nus

Similar criteria were used to separate the milk molars. In Tables
1 and 2 the specimens in each group are listed and comparative
measurements given. Group A is inseparable on all criteria from
Elephas recki Dietrich 1916, and group B is very close to a new
primitive species of Loxodonta from east Africa (Maglio, in press,
b).

GROUP A
(Plates I-III)

A total of 19 specimens is included in this group. Unless other-
wise stated, all specimens are in the Institute for Paleontology,
Humboldt University of Berlin.

M'\ (PI. Ill, 7). Two fragmentary specimens of the right side
from Vogelfluss. The fragments are too incomplete to allow deter-
mination of the total number of plates. It is typical of all the speci-
mens in this group, as it is also for the earlier stages of E. recki,
that the enamel is only very weakly folded near the apex of the
crown but is much more strongly folded toward the base. One spe-
cimen (Vo. 11-13/1.39) is in an early stage of wear; the worn
plates show rather large, complexly folded posterior loops with
smooth to weakly folded enamel on the remainder of the enamel

1969 "ARCHIDISKODON EXOPTATUS" 5

figure. The lamellar frequency^ is rather low (3.8). but this frag-
ment is the anterior portion of an M-">; this value is probably lower
than the average value for the entire molar. The second specimen
(Vo. 330 K.L. 18 9-10. 38) is broken across the basal portion
of the crown, revealing a low cross-section of the plates. The me-
dian folds are reduced to a slight central widening of the plates,
indicating fusion of the anterior and posterior columns into the
plate faces. The enamel is strongly folded around the entire sur-
face of the plate. Based on these two incomplete specimens, M'^
may be characterized as follows:

P (no. of plates) =?; L (maximum length) =:=?; W (width) =:
77-94 mm; H (height) = 119-121 mm; H/W (100 X height/
width index)= 128-155; LF (lamellar frequency)= 3.8-4.3;
ET (enamel thickness) r= 2.3-3.4 mm.

M;j. (PI. I, 1-2; PL II, 4-6). Six incomplete molars from Vogel-
fluss and Gadjingero. The most complete specimens are Z. 94.96
and K.L. -10- 13.39, both of the right side. The former bears the
last nine plates with the posterior root system underlying the last
eight. From the position of the posterior and intermediate root
bases, this molar must have had at least three and more likely four
additional anterior plates, making a total of 12-13. As in M-', the
enamel is little folded near the apex but is much more so toward
the base. Thus, no. Z. 94.96, which has only the first five preserved
plates in wear, has nearly smooth enamel except on the first and
most worn plate, which is just beginning to show weak folding in
the median portion of the enamel figure. G.K.-18 is a well worn
specimen with columns that are nearly completely fused into the
plate faces. The enamel is coarsely folded, especially in the median
portion of the enamel figures. Anterior columns are generally
small or lacking entirely, whereas the posterior ones are free at the
apex. The latter form prominent loops or sinuses on the enamel
figures of moderately worn plates, but these disappear in later
stages of wear as the columns become fused into the plates toward
the base. M^ may be characterized as follows:

P=12-13; L=approx. 250-300 mm; W=:70-88 mm; H=100-
147 mm; HW= 122-198; LF=4.3-5.4; ET=2.6-3.9 mm.

' The average number of plates in a distance of 10 cm measured parallel
to the crown base.

6 BREVIORA No. 336

M~. One incomplete left molar from Garussi, a tributary of Vog-
elfluss (no number). Although lacking its lingual half, this speci-
men is complete as to length and has nine plates plus a strong
posterior heel. The enamel is moderately folded and there is a
small but prominent posterior sinus. It is too worn for determi-
nation of the crown height.

P=9; L=116 mm; W=:?; H=?; H/W=?; LF=6.0; ET=

3.1-3.8 mm.

M2. A single specimen (G.K. 1/39) from Garussi. Only the
anterior three plates are preserved. From what remains, this molar
appears to have been similar to M-.

P=:?; L=?; W=77 mm; H=?; H/W=?; LF=5.2; ET=2.4-

2.8 mm.

Ml. One incomplete specimen (Vo. Aa) with the first five plates
preserved. The crown is too worn for a determination of its height.

P=?; L=?; W=62.1 mm; H=?; H/W=?; LF=5.3; ET=:

2.7-2.9 mm.

M^. (PI. I, 3). One partial specimen (Vo. A) from Vogelfluss.
The last five plates and a strong posterior heel are preserved. The
wear figures form a narrow loxodont pattern, without strong median
sinuses but with angular median expansions. The enamel is
strongly folded. The specimen is too worn for a determination of
the crown height.

P=?; L=?; W=66 mm; H=?; H/W=?; LF=5.1; ET=2.1-

3.0 mm.

dM"*. One incomplete specimen (BM L. 171 OS). Six plates
are preserved and it is probable that this represents the total num-
ber for this tooth. The crown is short and broad with thin, coarsely
folded enamel and only slight median expansions; there are no
true sinuses.

P=6; L=106 mm; W=82 mm; H=?; H/W=:?; LF=:5.7;

ET=:1.6-2.2 mm.

dM4. Two incomplete teeth (G.K. 2/39 and Vo. 313 2b(7.73))
from Vogelfluss and Garussi. These specimens are too fragmen-
tary to allow a determination of the crown height or the number
of plates. From what remains, we may characterize this molar
type as follows:

P=?; L^?; W=56-59 mm; H=?; H/W=?; LF=5.9-7.3;

ET=: 1.7-2.4 mm.

dM;,. (PI. Ill, 8-9). One complete specimen (BM M- 14942)
from Vogelfluss. There are six plates and a very strong heel. The
apices of the plates are divided into numerous small digitations

1969 ARCHIDISKODON EXOPTATUS 7

with remnants of a slightly deeper median cleft showing in very
early stages of wear. Weak median sinuses are seen on moderately
worn plates and the enamel is thin and strongly folded. A strong
anterior root supports the first two plates; a small internal inter-
mediate root supports portions of both the third and fourth plates,
and the posterior root supports plates 4-6 and the heel.

P=6; L=73.5 mm; W=i37 mm; Hr=31 mm; H/W=85; LF=

8.1; ET=1. 3-1.7 mm.

dM-. (PI. Ill, 10-11). Two complete specimens (Z.60 and
Z.68) from Vogelfluss and Garussi. Z.68 is unworn and bears
three plates as well as a strong two-cusped anterior ridge. A pos-
terior heel consists of seven small columns closely appressed to the
last true plate. The tooth is narrow anteriorly but broadens con-
siderably at the second plate and is widest at the third. A stout
root supports the last two plates and the heel; a second root sup-
ports the anterior ridge and first true plate.

P=3; L=25-27 mm; W=22 mm; H=13-18 mm; H/W=58-80;

ET=1.1 mm.

dMo. (PI. Ill, 12-13). One complete specimen (Z.62) from
Garussi. There are four plates plus a strong posterior heel. As in
dM2, the tooth broadens posteriorly but not nearly as much. The
first two plates have only three digitations, the third has five, and
the fourth, eight. Two roots are present as in dM-.

P:=4; L=26.5 mm; Wr=18 mm; H=14 mm; H/W=77; ET=?.

This assemblage of molars as a whole compares well with ma-
terial from other east African localities, such as Olduvai Beds I-II,
and the upper part of the Omo sequence, which are referable to
different stages of Elephas recki. As discussed elsewhere (Maglio,
in press, a, and Cooke and Coryndon, in press), the most primitive
stage (stage 1 of Maglio) attributable to E. recki occurs at Kika-
gati, Uganda (Hopwood, 1939), a deposit which appears to be
equivalent to the upper Kaiso beds (H.B.S. Cooke, pers. comm.).
This form, originally referred to "Archidiskodon griqua" by Hop-
wood, has approximately 13 plates on the M3, a lamellar frequency
of 4-5, and a height/width index of about 120-135. The enamel is
relatively thick (2.8-3.3 mm) and smooth, lacking the character-
istic folding of later stages of this species. Large anterior and
posterior sinuses are formed with wear due to the presence of
median columns fused for the most part to the surfaces of the
plates. The posterior columns may be free at their apices, and are
generally lower in height than the associated plates. As a result,
the sinuses do not appear in the enamel figure until intermediate
stages of wear.

8 BREVIORA No. 336

Successively more progressive stages of E. recki are found in
the later Omo beds, Beds I-II Olduvai, and in Bed IV Olduvai.
The later Omo stage (stage 2 of Maglio, op. cit.) has, on the aver-
age, slightly thicker enamel than does the Laetolil material, and is
proportionately slightly lower crowned. The lamellar frequency is
greater (5-6) and the number of plates appears to have been
slightly higher. The worn enamel figures show little or no develop-
ment of an anterior sinus, but a persistent fused posterior column
is present as in the Laetolil material. The enamel is only weakly
folded in the median portion of the plates.

The Olduvai Bed I-II form (stage 3) is somewhat more pro-
gressive than the Omo and Laetolil form, having generally thinner,
more highly folded enamel and reduced, irregular sinuses.

The Laetolil assemblage as a whole would apear to be closest to
the later Omo population of Elephas recki. Based on the elephants,
at least part of the Laetolil fossiliferous beds may be correlated with
this part of the east African sequence.

GROUP B
(Plates IV-VI)

Twenty-four specimens are included here as follows:

M"^. (PI. IV, 14-15). Three incomplete specimens (BM
M- 154 16, G.K. V, and Vo. 70) from Vogelfluss and Garussi. The
most complete specimen (BM M-15416) bears the last seven
plates and probably had no more than nine or ten when complete.
Only the last plate permits a measure of the crown height (68.0
mm), but the maximum height must have been 15-20 mm greater.
The enamel is thick and not folded. Anterior and posterior sinuses
are present on the plates in intermediate stages of wear. The plates
are well spaced — there are only 3.5 in 10 cm. G.K. V is unworn
but has been sectioned at about the middle of its height. Except
for slight coarse folding of the enamel in the median part of the
wear figure, the enamel is smooth as in the previous specimen.

P=?9-10; L=?; W=:76-85 mm; H=68-83 mm; H/W=96-109;

LF=3.5-4.3; ET=3. 3-4.3 mm.

Mg. Two incomplete specimens (BM LS 9VI35 and Vo.
9-10.3B). The height/ width index of the one unworn specimen
(BM LS 9VI35) is very low, and it is likely that the average height
for this molar type was somewhat greater. Other characters are as
in M^.

P=?; L=?; W=87-94 mm; H=79 mm; H/W=85; LF=3.4-

4.0; ET=3.0-3.7 mm.

1969 "archidiskodon exoptatus" 9

M-. One specimen (G.K. 2.39II) from Garussi. The last six
plates are preserved and the last three are unworn. The enamel is
thick and unfolded.

P=?; L=?; W=80 mm; H=12 mm; H/Wnr91; LF=4.9;

ET=i3.4-3.6 mm.

Mo. (PI. IV, 16; PI. V, 18-21). Five incomplete specimens
from Vogelfluss and Garussi. The most complete specimen (Vo.
9/10.38) bears the last seven plates with the posterior root system
supporting the last five. A strong anterior root underlies the first
one and one-half plates suggesting that the tooth is essentially
complete except for a probable anterior ridge and perhaps one
additional plate. The plates are thin toward their apices but
broaden rapidly toward the base. The wear figures are widest in
the midline and have rounded anterior and posterior sinuses. The
enamel is thick and not folded. In another specimen (Vo. N), the
sinuses are larger and the enamel is somewhat wavy, although not
folded as in E. recki. The posterior columns may be free for part
of their height, as in no. 5882, becoming fused with the plate face
toward the base. The transverse valleys between the plates are
broadly open and U-shaped. As in other molars in this group, the
crown height is roughly equivalent to its width. This molar type
is characterized as follows:

P=7-8; L=approx. 200 mm; W=81-88 mm; Hr=80-87 mm;

H/W=91-106; LF=4.2-5.0; ET=3. 1-4.5 mm.

Mj. (PL VI, 22-23). Five specimens from Vogelfluss and
Garussi. Two specimens (5828 and 5824) are complete but well
worn, and probably represent the left and right tooth of the same
individual. Both have the anterior plates worn down to the root,
but from the position of the anterior root, it is clear that the total
number of plates was seven. The last plate is only slightly worn
and offers a means of estimating the maximum crown height (about
80 mm). The enamel is thick and smooth. Prominent sinuses are
present on several of the worn plates. The enamel figure indicates
the presence of a weak median cleft on the upper half of the an-
terior four or five plates. A strongly backward-curving anterior
root supports the first one and one-half plates; an equally strong
and curved intermediate root underlies the lingual half of plates 3
and 4. The posterior root system supports the last three plates.

P=7; L=: 155-165 mm; W=69-88 mm; H=67-80 mm; H/W=

96-99; LF=4.6-5.1; ET=2.5-4.0 mm.

dM4. (PI. IV, 17). Two nearly complete specimens (Vo. 330
(7.78) and 5827) from Vogelfluss. Both specimens have five

10 BREVIORA No. 336

plates as well as a small posterior heel. Prominent anterior and
posterior columns are fused into the plates for their entire height
and with wear form sharp sinuses. Though slightly wavy, the enam-
el is essentially smooth and rather thick.

P=5; L=126 mm; W=51-58 mm; H=?; H/W=:?; LF=5.0-

5.9; ET=2.0-3.2 mm.

dM-'''. Two specimens from Vogelfluss and Garussi. No. 5818
is complete, with five plates, an anterior ridge, and a posterior heel.
The enamel is weakly folded around the entire surface of the
plates. There are weak median loops on the enamel figure in early
stages of wear, but these become more prominent with increased
wear. The second specimen (5830) has larger sinuses and a deep
median cleft on the first two plates.

P=5; L=ll mm; W=39-41 mm; H=32-42 mm; H/W=78-

92; LF=8. 1-8.3; ETr=l. 2-2.0 mm.

dMo. (PI. VI, 24-25). Two specimens (5883 and 5886) from
Vogelfluss and Garussi. Six plates are preceded by a small anterior
ridge. The enamel is weakly but very coarsely folded and the wear
figure is very irregular. Small median swellings on the anterior
and posterior faces of the plates mark the position of the fused
columns. The tooth is proportionately wider than the correspond-
ing tooth of E. recki and has one less plate. An anterior root sup-
ports the first one and one-half plates, and the posterior root sys-
tem supports the last four.

P=6; L=54-71 mm; Wrr35-37 mm; H=31 mm; H/W=84;

LF=9.3-9.5; ET=1.5-2.0 mm.

dMo. (PI. VI, 26-27). Two complete specimens (5837M and
5837G) from Marambu and Garussi. This is smaller and pro-
portionately less elongated than the corresponding tooth of E. recki.
There are only three plates present with a small anterior ridge and
a posterior heel. The plates have only three to four digitations in
contrast to the 7-8 of E. recki. Unlike the condition in the latter
species, there is only a single root, constricted vertically into an
anterior and posterior portion.

P=3; L=19-22 mm; W=15-16 mm; H=14-16 mm; H/W=

86-102; ET=1. 2 mm.

The twenty-four specimens in the present group B certainly
represent a form considerably more primitive than that of group A
in every trait that can be used to characterize the evolution of
molars in elephants. In Table 3, ranges of measurements are sum-
marized for the six molars of each group. The available measure-
ments and the morphology of specimens in group B are close to

1969 "ARCHIDISKODON EXOPTATUS" 11

Kanapoi species "C" of Maglio (in press, a) and to those of Main-
mutluis ajricanavus. The major differences between these latter
two species lie in the skull. A poorly preserved skull from north
Africa (Arambourg, in press) has been referred to M. ajricanavus
and demonstrates the Manimutluis affinities of this species. How-
ever, a skull and skeleton of Kanapoi species "C" (Maglio, in
press, b) demonstrates its ancestral relationship to Loxodonta afri-
cana. Though very similar in dentition, these two fossil species
can be distinguished on teeth alone when a suitable sample is avail-
able; there are 1-2 fewer plates in the Kanapoi species, the plates do
not taper toward the apex as markedly as in M. ajricanavus, and
the median sinuses are generally larger. Molars from lower Kaiso,
Yellow Sands (Omo), the Chemeron beds, and Kanam are also
referable to Kanapoi species "C." The Laetolil elephant here re-
ferred to group B appears also to belong here.

DISCUSSION

As for the status of "Arc/udiskodon exoptatus," it is clear that
the name encompasses two distinct taxa, one {E. recki Dietrich
1916) with priority. It is my opinion that the name "A. exoptatus"
should be suppressed for the following reasons: 1) The concept of
the species as originally intended has been confused in the litera-
ture to the point where recent workers cannot be certain of the
proper diagnosis. Even if adequately limited to one good taxon,
the name would still invoke confusion in the minds of some work-
ers who must deal with the past literature. 2) With the availability
of the excellent and abundant new material from Kanapoi and
other localities, it is unwise to maintain the fragmentary material
from Laetolil as the type collection of any species. Though part of
this collection is probably conspecific with the Kanapoi species of
Loxodonta, identity with this taxon (or any other) can not be cer-
tain on present evidence. Where possible, fossil species should be
founded on the most adequate material available.

Since Dietrich did not select a type specimen, I, as first revisor,
select as the lectotype of ''A. exoptatus" IPUB no. Z. 94-96, a right
Mo of group A. Thus. Arc/udiskodon exoptatus Dietrich 1942 be-
comes a junior synonym of Elephas recki Dietrich 1916. The pres-
ent group B is then referred to Loxodonta sp., pending description
of the new species from Kanapoi.

12 BREVIORA No. 336

CONCLUSIONS

The occurrence of two species at Laetolil tends to confirm
earlier suggestions that the Laetolil fauna represents two distinct
horizons. One, containing a stage 2 E. recki, correlates best with
the later Omo beds, and may antedate Bed I Olduvai, but only by
a relatively short interval of time. The second and earlier fauna
with a primitive species of Loxodonta seems to correlate best with
Kanapoi, Yellow Sands, Chemeron, and Kanam. Both species
occur at the Vogelfluss and Garussi exposures, but only E. recki
has been recorded from the Gadjingero exposures. The significance
of this is uncertain. The drainage of the Gadjingero lies to the
north of both Vogelfluss and Garussi and generally at a higher
altitude. It is possible that it includes only the upper levels of the
Laetolil beds. Whether several distinct levels are involved as pro-
posed by Kent ( 1941 ), or whether we are dealing with redeposition
and mixing of two faunas at a single horizon as suggested by Diet-
rich (1942) is not certain on present evidence.

ACKNOWLEDGEMENTS

I wish to express my gratitude to Dr. Karl-Heintz Fischer of the
Institut fiir Palaontologie u. Museum, the Humboldt University of
Berlin for permission to examine and photograph the syntype col-
lection of "A. exoptatus," and to Dr. Anthony SutcHffe of the
Department of Vertebrate Paleontology, British Museum (Natural
History) for permission to study relevant specimens in the collec-
tions of that institution. I am grateful to Professor Camille Aram-
bourg for allowing me to examine unpublished material in his col-
lection. Thanks are also due to Professor H. B. C. Cooke and Mrs.
Shirley Ccryndon for allowing me to see their unpublished manu-
script on the Kaiso beds and for permission to study specimens
discussed in that work. Professors Bryan Patterson and H. B. S.
Cooke kindly read the manuscript and offered helpful comments.

The research was supported in part by NSF Grant nos. GP-1 188
and GA-425 to Professor B. Patterson and by an Evolutionary
Biology Training Grant to the Department of Biology, Harvard
University, NSF Grant no. BG-7346 (Reed C. RoUins, principal
investigator).

1969 "ARCHIDISKODON EXOPTATUS" 13

LITERATURE CITED

Arambourg, C.

1947. Contribution a Tetude geologique et paleontologique du bassin
du lac Rudolfe et de la basse vallee de I'Omo. Deuxieme partie
Paleontologie. Mission Scient. Omo 1932-1933, 1, Geol.-
Anthrop. : 232-562.

"Le vertebres du Villafranchien de I'Afrique du Nord." Arch.
Mus. Natl. Hist. Nat.. Paris, 1969.
Cooke, H. B. S.

1960. Further revision of the fossil Elephantidae of southern Africa.
Palaeontologia Africana, 7: 46-58.
Cooke, H. B. S. and S. Coryndon
In press Fossil mammals from the Kaiso formation and other related
deposits in Uganda. Fossil Vertebrates of Africa, 2.
Dietrich, W. O.

1916. Elephiis antiquus recki n. f. aus dem Diluvium Deutsch-Ost-

afrikas. I. Arch. Biontogolie, 4(1) : 1-80.
1942. Altestquartare Saugetiere aus der siidlichen Serengeti, Deutsch-
Ostafrika. Palaeontographica, 94 (A) : 43-133.
HopwooD, A. T.

1936. New and little-known mammals from the Pleistocene of Kenya
Colony and Tanganyika Territory. I. Ann. Mag. Nat. Hist.,
17 (102) : 636-641.
1939. The mammalian fossils. //;.- O'Brien, The prehistory of Uganda
Protectorate. Cambridge, pp. 308-316.
Kent, P. E.

1941. The recent history and Pleistocene deposits of the plateau
north of Lake Eyasi, Tanganyika. Geol. Mag., 78 (3) : 173-
184.
1965. Olduvai Gorge 1951-1961, Vol. 1. London: Cambridge Univ.
Press, 118 pp.
Maglio, V. J.
In press a. Early Elephantidae of Africa and a tentative correlation of

African Plio-Pleistocene deposits. Nature, (London), 1969.
In press b. Four new species of Elephantidae from the Plio-Pleistocene of
northwestern Kenya. Breviora, Mus. Comp. Zool.

(Received 19 August 1969.)

ABBREVIATIONS

BM — British Museum (Natural History).

IPUB — Institut fur Palaontologie u. Museum, der Humboldt Uni-
versitat zu Berlin.

14 BREVIORA No. 336

W D- ^^— T^u-iONO"^. o\^ooooa\or^^TtTt^-■
• ' 5 L^ P m' r<-i r^i r^i r<-i r^' r^. r<i m ri ri rn rj (^i ri — i >— ■ i
u .- tUgv, mq<Nvqr-_Ov-)--Tti~--— ;\Dr^a\<^ l-^ I

cr «

(L)

V. en

i s

i-J <<-] Tj- m "^ -^ ■^ Tt >/-, vo i/-, 1/-, >o >/". >/~i r-~ oc

^

C3 E ^^^ „„„„

•f^ E ^ o ON >/-; I |~~: Tf d IT-; I I I I I I I vD r-^ (N Tf

■^ c
o

c

rj — — I Tj- r) O O

V* "^ ^"^ w ri r; rj ^1 CI ffl c) « M eo CI « C) lO « C) M

0'« C-ONt^OOOOOOvCI^OC r-^^VOt>0>/-i'/'ir'^(N(N —

c^o ^+ + + + + + + + + + + + + + 4-

■ - . , c •-- a\ oo\d r^ r J -^ r<^ v-i — f<-i ov "/o' •— \o rn >/-i r--^ vd

4^ oj r-oor-~0'/'iOoot~~-o\r-r-aNOOooTj-r-~-r4<NeN

-C^.-r (N — (N — r^— .

O

CO :; ^^

< <: '33 cih

o ^ -J - + + + + + + +

-'=^- - -+ + + + + + + +^-+ +

on
C
OJ

■5 1

?i 03 ««««««-) CI— ( ^ ^J ^J ^J ^^ ^> ^5 ^>

OSf-*!) t,u,t- — ^u — ^ — — —

OJ

c

(/2

3

c ^"g

<U SI y; "O OO

D- U E

Ov

(D O <U

1/5

d o = C =5^" S^ '^ <^ t^

y^-£: _00-.00 ...QOr-'TO^

t/3 (U 0) ' 3

S< ^ ^ S 2: — ^ £: i — i "^ — £: ^ £: m °- '^ m °- '^ "^

4

1969

ARCHIDISKODON EXOPTATUS

15

^^-^mrJ'ot^vo o\>'^'*^o\0(~~-o«or4 000ooc>

HP Tf T}-' -^ r<^ r<-) r<-) i r<-i ■^ tJ- Tt (^ -^ r<-) rn m r<-) r<i (vj (vg _' (vj ^j
, , ^ r I I I I I I I I I I I " I I I

PL,

Tf r'^ -^ rW '^ ■^ V-i T^ T^ Tf Tt Tt T*' >/". ■^' ■^' >/-]>/-, OC 00 OS ON

\0 Co ON >0
ON ON O 00

ON

MO </-i — r^ I ON ^

O O ON ON ON On

I 00 rj ^
I r- ON 00

O 00

ffii

O O i/"i >— '

1 O H O r<-i

O oo ri On

00 \o 00 r~-

1 ri

1 l^l^dfN

00 °° 00 00

O >0

O NO
00 NO

O lO -H

ri ri — J

r*^ Tf r<-i

r^oO'^ONi/iNDsonoonvrivcfn

ri>o\d'<4'NDONrn-^-^r~-\Dt^rioN

oooor~-ONoor~-ooooooooooooooND

' "^ « -^ •m rt c^ c>

>^ 00 O O O — ; f^ O

-^ r~-^* ON t--^ <o <o vd

CN

CO
<

c
o

PQ

Oh

O

O

C

_ + ++++ + ++

c^— 'f^<^!<^p-Hr4 lOo — oo . ,_-,._-,,,

'-' ,- '^" Tt w-> ri r-^ — ." — ; — I o od 1/^ lo' \^(^d>^^^^-^^a\

crJoNOmOOoNON ^r--i/-)v^ vor--ONrjr-ND(~^>on'— •

+ + + +

— ;^00TfOOr4O<or)

+ +- + ++ +

+

:+++ ++++++ +■

<u

TT (/3

II

C/3 O

c
,< on

'■ « CO « M c-i

m n y: « to ?i ci ■?! ci ri ri

S2SSSiSSSSSSSSSSSTIJ"^"0-aTD'5"OT3

^ z

Ut-L. — — L.U— — t-— 1-

_ ^

5 CQ

t-- — ON

ffl = ^
r^j On •

d'^2

00

o . o

r)

Or,

6g

O ly-, ^,

■^oo^^--

•^ So

'~^r--ooo<^NDr^r~-

^ri — moooor»^r<^

ooooJrrirJ<^oooooooooooooo

^ V^i »/^l V-l l/"i lO «"l lO

^ oo c' oo oo <- ri I
^ ir-, ,^ i/-i i/-i f^ — .

2 k— I tJJ ' — ■ '

C/D • N , > ^ < . I , < . . . I . . , I . , , I . I ^ I . I , .-~, I . I . . . .— ^ 1-^

16 BREVIORA No. 336

TABLE 3

Comparative measurements of Elephas recki and Loxodonta sp.

from Laetolil.

No. of

ET

M'

plates

H/W

LF

(mm)

N

Elephas recki

?

128-155

3.8-4.3

2.3-3.4

2

Loxodonta sp.

9-10

96-109

3.5-4.3

3.3-4.3

3

M3

Elephas recki

12-13

122-198

4.3-5.4

2.6-3.9

6

Loxodonta sp.

?

84

3.4-4.0

3.0-3.7

2

M-"

Elephas recki

9

6.0

3.1-3.8

1

Loxodonta sp.

7

91

4.9

3.4-3.6

1

M,

Elephas recki

—

5.2

2.4-2.8

1

Loxodonta sp.

7-8

91-106

4.2-5.0

3.1-4.5

5

M*

Elephas recki

—

5.3

2.7-2.9

1

Loxodonta sp.

—

0

M,

Elephas recki

—

5.1

2.1-3.0

1

Loxodonta sp.

7

96-99

4.6-5.1

2.5-4.0

5

dM*

Elephas recki

6

5.7'

1.6-2.2

1

Loxodonta sp.

0

dMi

Elephas recki

5.9-7.3

1.7-2.4

2

Loxodonta sp.

5

5.0-5.9

2.0-3.2

2

dM'

Elephas recki

— .

—

—

0

Loxodonta sp.

5

78-92

8.1-8.3

1.2-2.0

2

dM3

Elephas recki

6

85

8.1

1.3-1.7

1

Loxodonta sp.

6

84

9.3-9.5

1.5-2.0

2

dM^'

Elephas recki

3

58-80

1.1

2

Loxodonta sp.

—

—

—

—

0

dM2

Elephas recki

4

77

1

Loxodonta sp.

3

86-102

1.2

2

PLATES

18

BREVIORA

No. 336

'^>

? .

i.

\i

i»-^t?

' "^f

"J-

."^ %

-^■^''

Plate I

1 . FJcphas reiki. IPUB Z. 94.96, r.M.-,; occlusal view. X 2/5.

2. Eiephas rccki. IPUB Z. 94.96. r.M.; lingual view. X 2/5.

3. Eiephas recki. IPUB Vo. A, r.M,; occlusal view. X 1/2.

1969

'archidiskodon exoptatus

19

Plate II

4. Elephas recki, IPUB K. L. 10-13.3.39, r.Mo; occlusal view, x 2/5.

5. Elephas recki, IPUB K. L. 10-13.3.39, yM,; lingual view X 2/5.

6. Elephas recki, IPUB G. K.-18. I.M.; occlusal view. X 1/2.

20

BREVIORA

No. 336

13

10 v4r

11

1I'>^

v^lll^lffi*,^^^

Plate III

7. Elephas recki, IPUB 330 (K. L. 18/9-10.38), r.M'; anterior view, x 1/2.

8. Elephas recki, BM M-14942, l.dMs; occlusal view. X 2/3.

9. Elephas recki, BM M- 14941, l.dM:3; lingual view. X 2/3.

10. E/ep/jfl.? m A/, IPUB Vo.Z. 68, l.dM-; occlusal view. X 1.

11. Elephas recki, IPUB Vo. Z.68, l.dM'; lingual view, x 1.

12. £/£?/j/;fl5 /TcA/, IPUB Z. 62. r.dM-; occlusal view. X 1.

13. Elephas recki, IPUB Z. 62. r.dM.; buccal view. X 1.

I

1969

ARCHIDISKODON EXOPTATUS

21

15

Plate IV

14. Loxodoma sp., IPUB G.K. V, r.M'; sectioned surface. X 1/2.

15. Loxodonta sp., IPUB G.K. V, r.M'; lingual view, x 1/2.

16. Loxodonta sp., IPUB Vo. C, Mo. X 1/2.

17. Loxodonta sp., IPUB Vo. 330 (7.78), r.dM4; occlusal view, x 1/2.

22

BREVIORA

No. 336

18

1969 "ARCHIDISKODON EXOPTATUS" 23

Plate V.

18. Loxodonta sp., IPUB 9/ 10-38, r.Mo; occlusal view. X 2/5.

19. Loxodonta sp., IPUB 9/10-38; r.Mo; lingual view. X 2/5.

20. Loxodonta sp., IPUB 5882, I.M.; occlusal view. X 1/2.

21. Loxodonta sp., IPUB 5882, l.M.; lingual view. X 1/2.

24

BREVIORA

No. 336

1969 "archidiskodon exoptatus" 25

Plate VI

22. Loxodonta sp.. I PUB 5828. l.M,; occlusal view. X 1/2.

23. Loxodonta sp., IPUB 5828, l.M,; lingual view. X 1/2.

24. Loxodonta ip., IPUB 5883, l.dM.; occlusal view, x 2/3.

25. Loxodonta sp., IPUB 5883, l.dM,,; lingual view, x 2/3.

26. Loxodonta sp., IPUB 5837M, r.dMj; occlusal view. X 1.

27. Loxodonta sp., IPUB 5837M, r.dM^; buccal view. X 1-

DO NOT CfcB^.UisC.cn0MP. ^O-'-
lo^l ' ■ LIBRARY.

^^W3 1970

B IR E V I O R^4,»o

UNlVERSlTYi vtr

Moseenti of Cojmparative Zoology

Cambridge, Mass. 30 December, 1969 Number 337

LOWER VERTEBRATES FROM THE LATE CRETACEOUS
HELL CREEK FORMATION, McCONE COUNTY,

MONTANA'

Richard Estes, Paul Berberian, and Charles A. M. Meszoely^

Abstract. Fifty-five lower vertebrate species are recorded from the Up-
per Cretaceous Bug Creeic Anthills locality. Hell Creek Formation, McCone
Countv. Montana. This faunal list closely resembles that from the Lance
Formation of Wyoming, also of late Cretaceous age (Estes, 1964). In
spite of this general similarity, a number of species are present at Bug Creek
Anthills that are absent from the Lance Formation sample: Cf. Paralhula
c«.v('/ ( Albuloidei: Phyllodontidae ). Falaeolahnis montaiwnsis ( Amiiformes?;
Palaeolabridae), Proampliiuimi cictacea (Caudata: Amphiumidae), an un-
identified discoglossid frog, an undescribed baenid turtle, Adociis sp.
(Testudinata; Dermatemydidae), Peneteiiis aquiloniiis^ (Sauria; Teiidae),
and an unidentified boid snake.

Both Bug Creek Anthills and the Lance localities are in the upper third
of Hell Creek and Lance Formation sections, respectively. A similar flood-
plain-riparian association of freshwater and terrestrial forms occurs in both
areas, including hybodontid, pristid, and dasyatid sharks: sturgeons and
paddlefish: amiid, albulid, and sciaenid fishes; six genera of salamanders;
four tienera of frogs: seven "enera of turtles: eleven genera of lizards; two
genera of snakes: two genera of crocodilians: and eight genera of dinosaurs.
Slight climatic, ecological, and geographical difl"erences are probably respon-
sible for the minor differences between Lance and Hell Creek faunas: that
from the Bug Creek Anthills locality is not notably more advanced in char-
acter than that from the Lance.

INTRODUCTION

In the summers of 1962, 1963, and 1964, Robert E. Sloan and
his associates at the University of Minnesota collected a series of
samples of vertebrate fossils at and near Bug Creek. McCone
County, Montana, in the late Cretaceous Hell Creek Formation.
The richest of these localities, Bug Creek Anthills, has yielded an
extraordinary sample of disarticulated fossil vertebrate material

' Fossil vertebrates from the late Cretaceous Hell Creek Formation. Mon-
tana: Contribution No. 8.

-Department of Biology. Northeastern University, Boston. Mass. 02115.

2 BREVIORA No. 337

(Sloan and Van Valen, 1965; Estes, 1965, 1969d-h; Estes and
Berbcrian. 1969). In the next paper in this series (Estes and
Berberian, in press), paleoecological techniques developed by
Shotwell (1955, 1958) will be ap'plied to the MCZ sample from
Bug Creek Anthills; we here provide only an annotated faunal list
of the lower vertebrates at this locality (for collections used, see
Acknowledgments). A few taxa present locally in localities at
or above the Bug Creek Anthills level (noted in the text) are also
included, and a complete list of known Hell Creek Formation
vertebrates will be included in Estes and Berberian (in press).

Since many thousands of specimens have been recovered, the
individual bones and their frequency of occurrence are not listed
in the systematic section. Minimum numbers of individuals and
relative abundance of species are given in Table 1 ; similar data
are also given for the Lance Formation, for comparison.

ANNOTATED LIST OF LOWER VERTEBRATES

The faunal list from Bug Creek Anthills is very similar to that
described from the Lance Formation of Wyoming (Estes, 1964),
as shown in Table 1, and a similar floodplain-riparian environ-
ment of deposition and rather similar climatic conditions are indi-
cated. The two local faunas show the same, relative abundance of
most forms, thus indicating, to a certain degree, that the samples
probably reflect actual abundance. Therefore, only those com-
ments supplementing or modifying conclusions reached bv Estes
(1964) appear here. A detailed paleoecological analysis and
comparison of the Bug Creek Anthills and Lance local faunas
will appear in Estes and Berberian (in press).

TABLE 1

Minimum number of individuals and relative abundance of lower

vertebrates in two localities,
lower vertebrates Lance Formation Hell Creek Formation

( V562() ) ( Bug Creek Anthills )

MNl

RA

MN1--

RA

Loin hiclioii

4

1.0

1

.3

MyU'clapliiis

.2

1

.3

S(/ii(ilirliiiici

.2

X

X

l.siliyrhizd

.2

X

X

Acipcii.scr

.2

7

1.8

Pitlcopscpliiinis

1*

.2

3

.5

"Prolciniiti"

_2

1

.3

A mill

103

21.0

94

19.4

liclonosioiniis

5

1.0

1

.3

1969 LATE CRETACEOUS LOWER VERTEBRATES 3

Lepisosteus 7 1.4 25 5.5

C orlops 26 4.5 3 .5

cj. Pamlhiila X X 1 .3

Pahwolabnis X X 7 1.6

Elopidae 1 .2 1 .3

Platacodon 42 8.2 2 .4

Hahrosaiirus 73 14.1 9 1.8

Opisthotnton 41 8.0 107 22.9

Scapherpeton 22 4.2 61 14.4

Lisserpeton 1* .2 72 12.8

Prodesmodon 11 2.2 2 .4

Proampliiuma X X 2 .4

Scotiophryne 1* .2 6 1.2

r/. Barbourula 1 .2 1 .3

other frogs 12 2.4 3 .5

Adocits X X 1 .3

Basilemys 1 .2 X X

Compsemys 1 .2 1 .3

Trionyx 1 .2 1 .3

Eiibaena 1 .2 3 .5

other baenids 1 .2 4 .8

Emydinae 1 .2 3 .5

Brachychampsa 1 .2 1 .3

Leidyosiichiis 1 .2 1 .3

Champsosaiirus 1 .2 2 .4

dinosaurs 8 1.6 7 1.8

Peneteiiis X X 1 .3

Chamops 15 3.0 3 .5

Meniscognathus 15 3.0 X X

Leptochamops 15 3.0 1 .3

Haptosphenus 3 .5 1 .3

Contogenys X X 4 .8

Saurisciis 4 .8 X X

Exostiniis 4 .7 3 .5

c/. Gerrhonotus 3 .6 X X

Pancelosaunis 20 4.0 15 3.2

Parasaniwa 36 7.0 1 .3

Paraderma 2 .4 1 .3

Palaeosaniwa 1 .2 1 .3

Colpodontosaurus 2 .4 3 .5

Litakis 1 .2 X X

Cutty sarkus 13 2.5 2 .4

Coniophis 1 .2 1 .3

boid snake XX 1 .3

?Pterosauria 1 ^ X X

Total 509 100.0 472 100.0

* Additions since Estes, 1964, Table 5.
**Only MCZ sample used in this calculation.

4 BREVIORA No. 337

CLASS CHONDRICHTHYES
ORDER SELACHII

Family Hybcdontidae
Lonchidion selachos Estes, 1964

MCZ 9330, teeth and cephalic spines, resemble those figured
by Estes (1964. figs. 1, 2d, 3a-b, d) from the Lance Formation
of Wyoming . None of the tricuspid teeth with squatinoid roots
interpreted by Estes (1964, figs. 2a-c, 3c) as anterior or sym-
physial teeth occur in the Bug Creek sample. Patterson (1966.
p. 331), in describing Jurassic and Cretaceous English species of
this genus, suggested that these tricuspid teeth belonged either to
Squatirhina americana or to an undescribed taxon. The preserva-
tion and appearance of the tricuspid Lance teeth indicate that they
belong to a squatinoid other than Squatirhina. Patterson (1966,
p. 326) has referred tricuspid teeth (presumed symphysial teeth)
to his Jurassic species L. heterodon, but no root structure is known.
As he says (1966, p. 331), it is possible that Lonchidion, like
Heterodontus, developed anterior tricuspid teeth with squatinoid
roots, but we agree with him that the Lance tricuspid teeth do
not belong to Lonchidion.

ORDER BATOIDEA

Family Pristidae
Ischyrhiza avonicola Estes, 1964

Only one specimen of this species is represented; it is American
Museum of Natural History 9330, from the Harbicht Hill local-
ity, which is about the same stratigraphic level as Bug Creek Ant-
hills and about twenty miles away. Both Bug Creek Anthills and
Harbicht Hill are termed transitional or ". . . of Paleocene as-
pect . . ." by Sloan and Van Valen (1965).

The specimen is 5.0 mm long, its crown flattened and keeled
anteroposteriorly. The base is extensively and evenly eroded; the
etched appearance suggests that it may have been of coprolitic
origin (Fig. la-b).

Ischyrhiza avonicola is recorded from the Cretaceous and Paleo-
cene of Texas by Slaughter and Steiner (1968), who note that it
seems distinct from larger species of Ischyrhiza and Onchosaurus
and is therefore unlikely to represent young of such larger forms.
They have also noted a possible Turonian-Paleocene size increase
in /. avonicola. The Paleocene Kinkaid Formation specimen fig-
ured by Slaughter and Steiner (1968, fig. 3m) is about 3.5 mm

1969

LATE CRETACEOUS LOWER VERTEBRATES

Fig. I. Ischyrhiza aronicola, AMNH 9330, rostral tooth; o, dorsal,
and h, anterior views; Harbicht Hill, Hell Creek Formation, McCone Coun-
ty. Montana; X 8.

long, which is about maximum size of Lance Formation speci-
mens (Estes, 1964. p. 14). They state (1968, p. 237) that their
Turonian Eagle Ford Formation specimens are smaller than those
from the Lance Formation and the one they figure (1968. fig. 3k;)
is about 1.8 mm long, not significantly smaller than the minimum
2.0 mm of Lance specimens. Their figure 3k-m, offered as a
temporal sequence of size change, thus represents approximately
the size variation occurring within Lance Formation specimens,
although they state ( 1968, p. 236) that the ". . . size of the teeth
in the Lance collections is very constant . . ." Since they do not
give size range for any of their samples, it is difficult to know in
what part of the observed range their figured specimens fall. How-
ever, the Harbicht Hill specimen is 5.0 mm in total length, sig-
nificantly exceeding the known range of Lance specimens. The
crown of the Harbicht Hill specimen is longer relative to the base
than in other specimens of this species, but this condition is partly
the result of erosion. In other features it does not differ from
/. avonicola and can be referred to it without much doubt. A late
Paleocene specimen from Wyoming ( Shotgun member. Fort Union
Formation, specimens collected by Craig Wood) is 3.9 mm long.
This specimen and that from Harbicht Hill may provide substan-
tiation for Slaughter and Steiner's suggestion of temporal size in-
crease in /. avonicola, and may also strengthen the 'Taleocene
aspect" of the Harbicht Hill locality suggested by Sloan and Van
Valen (1965).

6 BREVIORA No. 337

Family Dasyatidae
Myledaphus bi parti tus Cope, 1876

MCZ 9331, isolated teeth and dermal denticles, closely resem-
ble specimens figured by Estes (1964, figs. 7, 8d). As indicated by
Estes, Myledaphus is close to the African late Cretaceous Para-
palaeobates. Specimens of the latter figured by Stromer and
Weiler (1930, pi. II. fig. 15; incorrectly" cited by Estes, 1964,
p. 18, as 1913) show a crown pattern coarser than but essentially
identical to that of Myledaphus, having the same transverse ridge
separating anterior and posterior grinding surfaces. In view of the
great similarities between teeth of Parapalaeobates and Myleda-
phus, no more than specific difference between them is justified.
Although histological comparison has not yet been made, we doubt
that differences supporting generic separation could be found; we
therefore suggest ( 1 ) that Parapalaeobates be considered a junior
synonym of Myledaphus, thus extending the range of the latter
genus to the Upper Cretaceous of North Africa, and (2) that
M. pygmaeus be separated from M. bipartitus by its relatively
coarser sculpture and more flattened, rounded crown. In spite of
the name, the teeth of M. pygmaeus are as large as the largest speci-
mens of M. bipartitus.

CLASS OSTEICHTHYES
ORDER ACIPENSERIFORMES

Family Acipenseridae
Acipenser eruciferus Cope, 1 876

MCZ 9433-9442, 9448-9450. Fragmentary pectoral spines,
cleithra, clavicles, supracleithra, pterotics, dermosphenotics, supra-
orbitals, (?) dermosupraoccipitals, (?) parietals, hyoid elements,
ventral precaudal scute, and dermal scutes have been recovered;
terminology of Gregory (1933) for the dermal bones is followed.
The dermal sculpture of specimens that possess it is varied; pustu-
lar, ridge-and-valley, and pitted types occur, depending on element
and size, but a latticelike pattern occurs on most specimens. The
sculpture is fine in small specimens, much coarser in larger ones.
Many Acipenser-Wke lateral scutes occur (Estes, 1964, fig. 11a;
Lambe, 1902, pi. 21). Such scutes also appear in living Scaphi-
rhynchus, although scute sculpture is less coarse than in Acipenser,
at least in our comparative material. Many tiny, comb like scutes
like those covering the body of Recent examples of both genera

1969

LATE CRETACEOUS LOWER VERTEBRATES

occur also (Fig. 2c); these are vertically elongated rather than
expanded and flattened like polyodontid scutes, but resemble them
to some degree (Fig. 2j-k).

Fig. 2. a, Acipenser erucifenis, restored right supracleithrum, MCZ 9450,
X 1; ^, Recent Acipenser sp., the same, X 0.5; c, A. eriiciferus, dermal
scute, MCZ 9442; d, the same, left clavicle, MCZ 9448, x 1.8; e. Recent
Acipenser sp., the same, X 0.5; f-k, Paleopsephiiriis wilsoni, f, rostral(?)
scute, MCZ 9446, X 1-8; g, outer, and h, dorsal views of referred dermal
denticle, MCZ 9445, X 3; /, dorsal view of left maxilla, MCZ 9332, X 3;
/, outer, and k, inner views of two dermal denticles, MCZ 9445, X 3; o,
c-d, f-k from Bug Creek Anthills, Hell Creek Formation, McCone County,
Montana. Sculpture pattern indicated by coarse stipple on a-b, d-e.

8 BREVIORA No. 337

The shoulder girdle elements resemble those of Acipenser more
than those of Scaphirliyiw/ius. The cleithrum is a powerful element
with a complexly curved attachment area for the pectoral spine
and fin. The area of dermal bone separating adductor attachment
from pectoral spine articulation is of relatively greater anteroposte-
rior extent than it is in Recent Acipenser, and this condition seems
to be true of the plate-like ventral expansion as well. In other
respects the specimens match those of Recent Acipenser.

The clavicle closely resembles that of Recent Acipenser (Fig.
2d).

The supracleithrum differs from that of Acipenser in having the
dermal ridge closer to the unsculptured suprascapular-extrascapu-
lar articulation surface ( Fig. 2a); one (or both) of the latter two
bones evidenlly extended over relatively more of the supracleith-
rum than it did in Recent Acipenser.

Pectoral spine fragments are as in Lance Formation specimens:
coarsely grooved longitudinally, expanded proximally, and sub-
triangular in cross-section. Cross-section of all fin spines is com-
pressed as in Oldman Formation A. alhertensis (see Estes, 1964,
fig. lib, and pp. 21-22).

Preserved hyoid bones are mostly ceratohyals. In robust devel-
opment and widely-flared ends, they closely resemble those of
Acipenser. The hollow, unossified ends show numerous transverse
growth rings.

Dermal roofing bones are all fragmentary and generally undiag-
nostic. Parietals, dermosupraoccipitals, and operculum are re-
ferred on general resemblance to these same rather featureless
bones in Acipenser, and some may be misidentified. Dermosphe-
notic and supraorbital fragments are characteristic but are too
broken for discussion. The pterotics have a posteroventral (hyo-
mandibular-intercalary) ridge that ends centrally rather than con-
tinuing two-thirds of the way forward as in living Acipenser.

The type of Acipenser eruciferus is a fragment of dermal roofing
bone from the Lance Formation that Cope (1876) beheved to be a
lungfish tooth plate. Estes (1964) referred similar Lance Forma-
tion material to Acipenser, as A. eruciferus (Cope), although the
type is a nomen vanwn. No nomenclatorial difficulties should be
encountered unless articulated material should show that more than
one species of Acipenser is present in the Lance Formation and
that they are indistinguishable on the basis of dermal scute sculp-
ture. In identifying material from other deposits, use of the name

1969 LATE CRETACEOUS LOWER VERTEBRATES 9

A. eniciferus is less assured, yet the close temporal and geogra-
phical association of Lance and Hell Creek Formations lends some
security to the reference. We prefer to continue using this name
since the only other described species from the midcontinent is
A. alherteusis (Lambe, 1902), itself a iiomen nudum based only
on dermal scutes from the Campanian Oldman Formation of Al-
berta that are inseparable from those of A . erucijerus.

The Hell Creek specimens referred to Rhineastes sp. by Brown
(1907, p. 842) are actually referable to Acipenser erucijerus (Es-
tes, 1964, p. 21).

Protoscaphirliync/ius squamosus Wilimovsky, 1956

This poorly known acipenserid was described from the same
locality in the Fort Peck region that produced Paleopsephurus wil-
soiii (Wilimovsky, 1956). It is thus high in the Hell Creek Forma-
tion section, perhaps at Harbicht Hill level or higher (Sloan and
Van Valen, 1965, fig. 1). No specimens referable to this form
have been identified at Bug Creek Anthills.

Family Polyodontidae
Paleopsephurus wilsoni MacAlpin, 1 947

MCZ 9332, 9443-9447, maxillae, pterotics, frontal, rostral der-
mal scutes, and dermal denticles have been identified. The maxil-
lae are distinctive bones, being flat externally and all bearing a
prominent ectopterygoid process grooved dorsomedially (Fig. 2i).
All five specimens are broken at both ends but direct comparison
with the type specimen makes the identification clear. MacAlpin
(1947) noted that the ectopterygoid process was present as a thin
process in Psepluirus and that a similar process was also present
in Saurichthys and Elonichthys. Woodward (1895, pi. 1, fig. 3)
figures an almost exacdy similar process in Chondrosteus that was
not noted by MacAlpin.

Frontal and pterotic match comparable elements in Paleopse-
phurus. Dermal sculpture on the frontal is much more prominent
than on the pterotic. Both degrees of sculpture development occur
on various parts of the type specimen and the variation is probably
not significant.

Some dermal scutes are elongated, stellate, coarsely ridged, and
thickened dorsally (Fig. 2f). We interpret these as rostral scutes
because of their similarity to those of Recent paddlefish; they differ
from the latter in being relatively less stellate, more robust, and

10 BREVIORA No. 337

more anteroposteriorly elongated. These differences may indicate
a rostrum proportioned more as in Polyodon than in Psephurus.

Trunk dermal denticles differ from those figured by MacAlpin
in being flat-topped, and in having an attachment process almost
at right angles to the body of the scute (Fig. 2j-k). Scute surfaces
are coarsely striated and grooved, and the posterior borders are
denticulated. One larger, less expanded scute (Fig 2g-h) may be
from near the shoulder girdle. Among scutes figured by MacAlpin
( 1 947 ) , those of Crossopholis are most like the Bug Creek speci-
mens, but polyodontid dermal scute variability is great, and the
total range is not known for Paleopsephurus, so that the variation
may not be significant.

ORDER ASPIDORHYNCHIFORMES

Family Aspidorhynchidae
Belonostomus longirostris (Lambe, 1902)

MCZ 9333, a fragmentary predentary and a skull fragment, re-
semble Lance Formation specimens (Estes, 1964, p. 22).

ORDER AMIIFORMES

Family Amiidae
Amia fragosa (Jordan, 1927)

Large numbers of specimens of this species have been recovered
(MCZ 9286-9293, 9390-9432), and have been reviewed by Estes
and Berberian (1969), who substantiated Janot's (1967) sugges-
tion that Kindleia is a synonym of Amia.

Amia ("Protamia") sp.

Two glassy, translucent tooth tips (MCZ 9334) are similar to
those of Lance Formation specimens. No vertebrae or large skull
elements occur in the large Bug Creek sample. Janot (1967) has
suggested that Protamia, like Kindleia, is a synonym of Amia.

The presence of both small and large amiids is common in many
late Mesozoic and early Cenozoic deposits both in Europe and
in North America (Estes, 1964); careful study of this material
would be of considerable systematic and zoogeographic interest.

ORDER AMIIFORMES?

Family Palaeolabridae
Palaeolahrus montanensis Estes, 1 969h

This fish was described by Estes (1969h) on the basis of palatal
tooth plates, skull and mandibular elements, and vertebrae. The

1969 LATE CRETACEOUS LOWER VERTEBRATES 11

type species is from the late Paleocene or early Eocene Dormaal
deposits in Belgium (Casier, 1967).

ORDER LEPISOSTEIFORMES

Family Lepisosteidae
Lepisosteus occidentalis Leidy, 1856

Almost all bones of skull and skeleton are present (MCZ 9354-
9389).

Estes (1964) gave a restoration of this species, which is a wide-
snouted form closely related to the living alligator gar L. spatula.
The quadratojugal, found at Bug Creek but not in the Lance For-
mation sample, does not differ fundamentally from the same bone
in Lepisosteus spatula.

The circumorbital series (incomplete in the Lance Formation
sample) has been identified and shows no significant differences
from that of Recent Lepisosteus.

ORDER ELOPIFORMES
SUBORDER ELOPOIDEI

Family Elopidae

Estes (1964, p. 48, fig. 23) described large teleost scales having
a granular apical sculpture as resembling those of the late Cretace-
ous elopid Dinelops, from the English Chalk. Casier (1966, p. 133,
pi. 13. fig. 3) described similar scales from the Eocene London
Clay and referred them to an unknown elopid. Two scale frag-
ments (MCZ 9338) are identical to the Lance Formation speci-
mens but do not suggest any further clues as to the generic identity
of the fish.

SUBORDER ALBULOIDEI

Family Phyllodontidae
cj. Paralbula casei Estes, 1969a

MCZ 9335 consists of three teeth, two of which are superim-
posed phyllodont teeth; the other is a single tooth with the charac-
teristic coarse tooth pattern of Pflra//)«/a ca-sd (Estes, 1969a). The
directly successional rather than the overlapped arrangement of the
teeth indicates a phyllodontine rather than a paralbuline replace-
ment (Estes, 1969a), but an occasional directly successional pair
of teeth may occur in the latter group. The coarse surface sculpture
and appearance of the isolated tooth are as in the type of P. casei
(Estes, 1969a, fig. 3). P. casei is known from the Campanian of

12 BREVIORA No. 337

Wyoming, Maestrichtian of New Jersey, and Eocene of England;
this record constitutes the highest stratigraphic occurrence of the
species in the Western Interior region of North America.

Family Albulidae
Coriops amnicolus Estes, 1969h

These albuloid parasphenoid and basibranchial tooth plates
(MCZ 9340) differ from those of Alhula in having the teeth fused
to the plates and in lacking otic articulation surfaces on the para-
sphenoid. Coriops also occurs in the Lance Formation (Estes,
1969h).

ORDER PERCIFORMES

Family Sciaenidae
Platacodon nanus Marsh, 1889

MCZ 9336, two fifth ceratobranchials; MCZ 9337, referred den-
taries, maxillae, vertebrae, and other bones.

The fifth ceratobranchials are paired and have the characteristic
tooth arrangement of this species as indicated by attachment areas
(Estes, 1964, p. 51, fig. 25) although no teeth are preserved on
these specimens. Referred skull elements, vertebrae, and spiny
rays may belong to this species or may include other perciform
types as well.

Teleostei incertae sedis

Numerous vertebrae and miscellaneous fragments of skull bones
may indicate teleosts other than the five noted above, but they are
not distinctive enough for identification.

CLASS AMPHIBIA

ORDER CAUDATA

SUBORDER AMBYSTOMATOIDEA

Family Scapherpetontidae
Scapherpeton tectum Co^q, 1876

MCZ 3673. Vertebrae, atlantes, dentaries, maxillae, parietals,
postdentary "compound" bones, and exoccipitals are preserved.
Estes (1969b) referred both this genus and Lisserpeton to the
Ambystomatoidea on the basis of fused postdentary "compound"
bones, including fused prearticular and angular.

1969 LATE CRETACEOUS LOWER VERTEBRATES 13

Lisserpeton hairdi Es,tes. 1965

MCZ 3674, 3677-3679. Vertebrae, atlantes. dentaries, maxil-
lae, parietals, postdentary "compound" bones, and exoccipitals
are preserved. Figures and descriptions of this relative of Scaph-
erpeton are given in Estes (1965 ) .

Family Prosirenidae
Prodesmodon copei Estes, 1964

MCZ 3652, trunk vertebrae, are rare at Bug Creek Anthills,
much rarer than in the Lance Formation. Reasons for referring
this genus to the Prosirenidae are given in Estes ( 1969c) .

Family Batrachosauroididae
Opisthotriton kayi Auflfenberg, 1961

MCZ 3676. Vertebrae, atlantes, dentaries, exoccipitals. pari-
etals, premaxillae. and maxillae have been identified. Reasons for
referring this genus to the Batrachosauroididae are given in Estes
(1969b).

SUBORDER SALAMANDROIDEA?

Family Amphiumidae
Proaniphiiima cretacea Estes, 1969e

This earliest amphiumid, represented only by vertebrae, has been
described by Estes (1969e) from Bug Creek Anthills. It is prob-
ably ancestral to the Recent Amphiuma.

SUBORDER MEANTES

Family Sirenidae
Habrosaurus dilatus Gilmore, 1928

MCZ 3675. Vertebrae, atlantes, and dentaries are present, and
are similar to Lance Formation specimens.

ORDER SALIENTIA

Family Discoglossidae
Scotiophryne pustulosa Estes, 1969f

This frog was described by Estes (1969f) on the basis of ilia,
humeri, maxillae, and squamosals; the Recent Eurasian genus
Bombina appears to be its closest relative.

14 BREVIORA No. 337

cf . Barbourula sp.

MCZ 3653, a single right ilium, has a relatively large acetabular
fossa with a prominent flare of the anterior border and a promi-
nent iliac symphysis, as noted by Estes (1964, p. 55, fig. 56) in
the Lance Formation specimens. Material described but not named
by Hecht and Hoffstetter (1962) indicates that a European early
Oligocene discoglossid also displays these characteristics. The
European ilia are similar to Lance and Bug Creek specimens and
may perhaps be placed in the same genus, which will be described
in a paper now in preparation by Hecht and Hoffstetter.

Undescribed genus and species

Distinctive humeri known from only two specimens (MCZ
3654) have raised olecranon scars of rather limited proximal ex-
tent that are sharply demarcated by concave areas on either side.
The humeral ball is relatively large and projects in lateral view.
The medial epicondyle is blunt and is broken on both specimens;
the lateral epicondyle is essentially undeveloped (Fig. 3).

These unusual humeri are referred to the Discoglossidae on the
basis of the blunt, truncated medial epicondyle, the large and
projecting ball, and the limited extent of the olecranon scar; these
conditions resemble those of Alytes, an essentially fossorial dis-
coglossid. Somewhat similar conditions also prevail in R/iino-
phrynus, also a burrower, although the detailed resemblance is
with Alytes. Humeri referred to Scotiophryne are quite different
(Estes, 1969f, fig. 3), and because of both the dissimilarity of the
humeri and the very different adaptations in the Recent genera,
it seems unlikely that the Barbourula-likQ ilium and these Alytes-
like humeri are from the same taxon. The Bug Creek humeri are
distinctive, but we can offer no further comment on their rela-
tionships at this time.

Family Pelobatidae?
Eopelobates sp. ?

Material questionably referred to the Pelobatidae by Estes
(1964, figs. 30, 31c-e) and other specimens placed "near Hy-
lidae?" (1964, fig. 31a-b) are perhaps referable to a primitive
species of Eopelobates on the basis of squamosal shape and sculp-
ture similarities to a new Eocene pelobatid from North America
(Estes, 1970, in press). Bug Creek material (MCZ 3655) includes
several fragmentary maxillae, one fragmentary squamosal, and

1969

LATE CRETACEOUS LOWER VERTEBRATES

15

Fig. 3. Right humeri of undescribed genus and species of discoglossid;
a-c, dorsal, ventral and anterior (outline) views of MCZ 3654a; cl-f, the
same, MCZ 3654b; Bug Creek Anthills, Hell Creek Formation, McCone
County, Montana; X 8; cross-hatching indicates broken surface.

three ilia, aU similar to the Lance Formation specimens. No fur-
ther comment on these specimens is possible at this time.

Other Frog Remains

( 1 ) Two maxillary fragments that lack external sculpture, have
posterior processes indicating the presence of a complete maxillary
arcade, and have teeth set in a relatively deep sulcus dentalis; (2)
two humeri with medial, symmetrical olecranon scars that are of
a type common in many of the advanced frog families; (3) one
procoelous vertebra and two other badly worn and abraded frog
vertebrae that are the only axial frog elements in the Bug Creek

16 BREVIORA No. 337

sample; (4) a heavily worn and broken ilium with a well-defined
dorsal crest that is of a type seen in several frog families; (5) three
postdentary lower jaw fragments, two proximal radioulna frag-
ments, and a badly broken tibiofibula that are distinctively anuran
but have no special identifying features.

ORDER TESTUDINATA

Family Baenidae
Eubaena cephalica Hay, 1 908

A nearly complete skull, fractured on the supraoccipital crest
and missing the right maxillary area, is present, along with isolated
cranial and mandibular elements (MCZ 3510-3512, 3519, 3530).
This is the second known skull referable to this species, and it
closely resembles the type from the Lance Formation of Wyoming.
The skull referred to E. cephalica by Estes (1964, p. 97) belongs
to undescribed genus 1 noted below. A revision of the baenid
turtles is now being made by Eugene GafTney, and further comment
on these specimens will appear there.

Undescribed genus and species 1

Disarticulated skull elements (MCZ 3514) and shell fragments
are referable to a distinctive new baenid genus, and will be de-
scribed in a forthcoming study by Eugene Gaffney.

Undescribed genus and species 2

Another new genus of baenid turtle occurs in the Bug Creek
sample and is represented only by isolated maxillae (MCZ 3515-
3518). It differs from both Eubaena and the other undescribed
baenid in having an exceptionally wide maxillary triturating sur-
face without a ridge. These specimens will also be described by
Eugene Gaffney, on the basis of more complete material from the
Paleocene of Wyoming.

Family Dermatemydidae
Compsemys victa Leidy, 1856

MCZ 3671, numerous shell fragments, have the characteristic
sculpture pattern of many closely-set. flat-topped pustulae seen in
all described species of this genus. C. victa occurs in the Lance
Formation of Wyoming, and the Bug Creek specimens probably
belong to this species.

1969 LATE CRETACEOUS LOWER VERTEBRATES 17

Adociis sp.

The Bug Creek specimens referred here all have a very fine
sculpture (about eight to ten pits per cm) as in some species of
Adocus (Gilmore, 1919). Brown (1907; p. 842) originally iden-
tified Hell Creek specimens as A. lineolatus, but Gilmore indicated
(1919), p. 25 and other papers) that specific identification can-
not be determined by sculpture pattern. In the late Cretaceous,
two types of Adocus sculpture occur, however: a very fine type
with eight to ten pits per cm, and a more well-defined type with
six to seven pits. All Hell Creek Formation specimens are in the
former group suggesting that these two sculpture types are not
simply variants from diflferent shell areas. Lance Formation speci-
mens are clearly referable to Basilemys (Estes, 1964), having
coarse sculpture with only three to four pits per cm.

Because too few specimens of Adocus are known, the validity of
the described species cannot be assessed at this time, but it is safe
to say that more are named than is justified by the material.
Basilemys and Adocus are probably quite closely related (as com-
pared with other fossil so-called dermatemydids). Adocus is lim-
ited to the Maestrichtian and occurs on the East Coast as well
as in the Western Interior. It is primitive in having inframarginal
shields and unexpanded pectoral shields. Basilemys, which ex-
tends through both Campanian and Maestrichtian stages, has not
gone as far as Adocus in neural and suprapygal reduction. The
two genera seem to be distinct (Table 2) and probably had a
common ancestor in pre-Campanian time.

TABLE 2
Comparison of characters separating Adocus and Basilemys

Adocus Basilemys

1. Inframarginals large, extend- 1. Tiny axillary and inguinal in-
ing across bridge. f ramarginals (except B.

nobilis).

2. A single suprapygal; posterior 2. Two or three suprapygals;
neurals reduced. posterior neurals unreduced

(except B. nobilis?)

3. Posterior marginal shields 3. Posterior marginal shields
large, elongated anteropos- narrow, elongated medio-
teriorly. laterally.

4. Plastral lobes rounded. 4. Plastral lobes tend to be

acute.

18 BREVIORA No. 337

5. Pectoral shields little ex- 5. Pectoral shields greatly ex-
panded medially. panded medially.

6. Sculpture relatively smooth 6. Sculpture relatively rough and
and fine, about six to ten pits coarse; about three to four
per cm. pits per cm.

7. Carapace length 480-670 mm. 7. Carapace length 690-940 mm.

Family Testudinidae

Subfamily Emydinae?

Unidentified genus and species

Fragmentary eighth cervical vertebrae (MCZ 3567-3568, 3573)
have a double concave articulation surface posteriorly. Numerous
fragments of peripherals (MCZ 3656) and costals (MCZ 3657)
show deeply impressed shield sulci and marked changes in eleva-
tion between the shield areas.

Double, concave articulation surfaces on the eighth cervical
vertebra are known only in the Testudinidae (Wilhams, 1950).
The fragments of carapace and plastron closely resemble the shells
of pond turtles, especially Pseudefnys. The questionable family
reference given by Estes (1964, p. 99) to similar specimens from
the Lance Formation is confirmed by the distinctive cervical verte-
brae present here. This is the earliest record of the family, other-
wise not known before the Eocene Echmatemys; it will be discussed
further in a study in preparation.

Family Trionychidae
Trionyx sp.

A partial left hypoplastron (MCZ 3658) and many costal and
neural fragments (MCZ 3672) all bear a characteristic trionychid
sculpture. The hypoplastron shows that the plastron was reduced,
indicating that the specimen cannot be referred to the line of fossil
trionychids often designated as Plastomemis. The presence or ab-
sence of a prenuchal bone cannot be demonstrated in this material,
but recent work (Webb, 1962) indicates that the presence of a
prenuchal is insufiicient ground for separating the genus Aspi-
deretes. It thus seems best to refer this material to Trionyx (s. 1.)
as well as the material referred to Aspideretes heecheri by Estes
(1964).

1969 LATE CRETACEOUS LOWER VERTEBRATES 19

CLASS REPTILTA
ORDER EOSUCHIA

Family Chanipsosauridae
Cliaiupsosauriis sp.

MCZ 365 1 , ten vertebrae, a few tooth crowns, and three ribs are
present. The specimens are clearly referable to this genus but are
specifically indeterminable. Champsosaiirus was evidently rare at
Bug Creek Anthills, and is represented only by small individuals.

ORDER SAURIA
SUBORDER SCINCOMORPHA

Family Teiidae
Chamops seguis Marsh, 1 892

MCZ 3659, dentary fragments, fragment of left maxilla, four
unnumbered tooth-bearing fragments; MCZ 3660, fragmentary
parietals. These specimens were originally cited by Estes (1964,
p. 108) as possibly forming a new species of Chamops because
Bug Creek specimens available at that time all seemed to show
less bulbous tooth bases than did Lance Formation fossils. Sub-
sequently collected Bug Creek specimens do not bear out this
distinction. Chamops most closely resembles the Recent South
American species Callopistes maculatus (Estes, 1969d).

Leptochamops demiciilatus (Gilmore, 1928)

MCZ 3661, two maxillae and a few tooth-bearing fragments,
are poorly preserved but appear to belong to this species.

Haptosphenus placodon Estes, 1964

MCZ 3686, fused right dentary and splenial and a coronoid
with adhering parts of fused dentary and surangular are preserved.
The dentary shows the short, heavy jaw; short, rather Chamops-
like teeth; and fused postdentary bones with visible suture lines
as in Lance specimens. Haptosphenus is aberrant in the fusion
of jaw elements; we do not know of a comparable situation in liz-
ards. Nevertheless the closest resemblances of Haptosphenus
seem to be with Chamops, differing from the latter both in bone
fusion and in having almost acrodont teeth; it may be an aber-
rant teiid derived from a Chamops-VikQ ancestor.

20 BREVIORA No. 337

Peneteius aquilonius Estes, 1969d

This lizard was described by Estes (1969d); it shows resem-
blances both to the Recent Teiiis and Dicrodon as well as to the
late Cretaceous Polyglyphanodon. The holotype is the unique
specimen.

Family Scincidae ?
Contogenys sloaniEstQS, 1969g

This lizard was described by Estes (1969g). Contogenys re-
sembles members of the Scincidae; although it is not clearly refer-
able to that family, it is certainly a member of the Scincoidea.

Family Anguidae
Pancelosaurus piger (Gilmore, 1928)

This species was described in detail and removed from Pelto-
saurus by Meszoely (1970), who, in his revision of the Anguidae,
placed Pancelosaurus at the base of the subfamily Anguinae rather
than in the Gerrhonotinae as suggested by Estes (1964).

Family Xenosauridae
Exostinus lancensis Gilmore, 1928

A number of dentary and maxillary fragments and a referred
frontal are present. The jaw elements (MCZ 3662a) do not differ
significantly from Lance Formation specimens of this species. The
frontal (MCZ 3662b) may be referable to E. lancensis on the basis
of dermal sculpture pattern. It is eroded, and because of this and
its small size, it does not display a sculpture pattern as well de-
veloped as that on the larger Lance Formation parietal referred
by Estes (1964, pi. 3). If properly referred, frontals were paired
in E. lancensis, a condition that, while different from that in later
species of Exostinus and from Xenosaurus itself, is not a surprising
one in view of its Cretaceous age.

Diploglossa incertae sedis
Colpodontosaurus cracens Estes, 1964

Well-worn fragments of dentaries and maxillae (MCZ 3663)
lack teeth except in one specimen.

Estes (1964, p. 127) placed Colpodontosaurus as Diploglossa
incertae sedis on the basis of a tiny free ventral border of the in-
tramandibular septum, the presumed lack of jaw hinge, and the

1969 LATE CRETACEOUS LOWER VERTEBRATES 21

absence of basal fluting on teeth. Teeth on the type specimen
have been broken since the original figure (1964, fig. 60) was
made, and these fresh break surfaces show an irregular, almost
fluted appearance; under high magnification, several faint grooves
occur on the teeth of UCMP 49938, a maxilla from the Lance
Formation.

Although a tiny free ventral border is present on the intramandi-
bular septum, it is little different from that of Varanus and
Parasaniwa. Reexamination of the type of Colpodontosaiirus in-
dicates that the presence or absence of a jaw hinge in this specimen
cannot be determined, although the dentary seems to have a rela-
tively greater posterior projection than it does in Parasaniwa.

On the maxiUary fragments, the posterior end shows elonga-
tion of the posterior external mental foramina as in Parasaniwa
and varanids. The dorsal border of the bone is elongated and
gendy sloping as in Parasaniwa and the anguids.

Absence of sculpture, delicate construction, condition of intra-
mandibular septum, and elongated mental foramina are all as in
varanids, and the last two characters show parasaniwid resem-
blances as well. The essential absence of basal infolding of teeth,
and the long, slender, posterior process of the maxilla are char-
acters suggesting diploglossans. The latter character is completely
unlike that in Varanus, Saniwa and parasaniwids in indicating that
the tooth row extended posteriorly well under the orbit.

Although new interpretation and new specimens have shown
some varanid and parasaniwid characters, there are enough de-
tailed differences to maintain Colpodontosaurus as Diploglossa
incertae sedis.

Family Parasaniwidae
Parasaniwa wyoniingensis Gilmore, 1928

A few dentary and maxillary fragments and a parietal (MCZ
3664) show the characteristic simple infolding of tooth bases,
fused intramandibular septum and sculptured skull roof of topo-
typic material from the Lance Formation.

Paraderma bogertiEstes, 1964

A fragmentary left maxilla, an isolated tooth with adherent jaw
fragment and two referred vertebrae (MCZ 3687) do not differ
from Lance Formation specimens.

22 BREVIORA No. 337

INFRAORDER PLATYNOTA

Family Varanidae
Palaeosaniwa, cf. P. canadensis GilmorQ, 1928

Only a single large varanid vertebra occurs in the Bug Creek
sample (MCZ 3665), and it differs from Lance Formation speci-
mens in having less well-developed zygosphenes. The Bug Creek
specimen is about the size of the type (from the Campanian
Oldman Formation of Canada), and has convex lateral borders of
the centrum, as do Eocene Saniwa and both Oldman Formation
and Lance Formation Palaeosaniwa.

ORDER SAURIA?

Family incertae sedis
Cuttysarkus mcnallyi Estes, 1964

Three dentaries (MCZ 3666) show no significant differences
from the Lance Formation specimens. The systematic position of
this genus is even less clear now than when discussed by Estes
(1964). Nearly forty dentaries are known from the Lance For-
mation type area, but no maxillary fragments of this distinctive
animal were discovered.

A number of people have commented on relationships of Cutty-
sarkus, and have offered possibilities spanning all lower vertebrate
Classes. One colleague suggested that the name was based on
arthropod mandibles, yet we must reject his contribution by noting
that Cuttysarkus jaws are composed of characteristic, cellular
vertebrate bone. One of the most reasonable possibilities came
from C. Wilson Kerfoot, who suggested that there were many
similarities between the dentaries of Cuttysarkus and the sala-
mander Opisthotriton. Absence of a sulcus dentalis, lack of defi-
nition of Meckelian groove, straight posterior border of the dentary
and posteroventral depression lingually are all indicative of sala-
mander dentary structure, although each of these features can be
matched in one or another lizard group. The teeth are not pedi-
cellate but this condition can be matched in the salamanders Pro-
desmodon and Habrosaurus. The absence of maxillae could be
interpreted as evidence that Cuttysarkus was a larval salamander.
Yet all Lance and Hell Creek Formation salamanders are distinc-
tive, and most are essentially of "larval" or paedomorphic type
(Estes, 1964). Cuttysarkus ]2lvj?, are too distinctive themselves, in
any case, to be from larvae of any of the salamanders present. In
the Lance Formation, each known type of salamander vertebrae

1969 LATE CRETACEOUS LOWER VERTEBRATES 23

is matched by distinctive skull elements consonant with the verte-
brae in size, morphology, state of preservation, and frequency of
occurrence. Mandibular elements of Proamphiiima (Estes, 1969e)
are unknown, but there is no resemblance of Cuttysarkus to
Am phi II ma in maxillary structure. Cuttysarkus may not be a liz-
ard, but we retain it there for the present.

ORDER SERPENTES
SUPERFAMILY BOOIDEA

Family Aniliidae
Coniophis precedens Marsh, 1892

Only a few vertebrae are present (MCZ 3667) and indicate no
difference from Lance Formation specimens (Estes, 1964; Hecht,
1959). A pair of coossified frontals (MCZ 3668) is also referred;
no comparisons are made at this time pending study of other Coni-
ophis material by Hecht.

Family Boidae

Subfamily Erycinae ?

Unidentified genus and species

A single vertebra (MCZ 3669) is poorly preserved and broken
but is of interest in demonstrating the presence of a second species
of snake from the Cretaceous of North America (Fig. 4). Centrum
length (CL, Auffenberg, 1963, fig. 3) is 8.8 mm, about the size
of some Eocene Boavus vertebrae but twice as large as the largest
known Coniophis specimen. In having a con-dylar ball with
rather sessile edges, sessile haemal carina and paradiapophyses, and
a compact centrum shape, this specimen resembles vertebrae of
the subfamily Erycinae (sensu Hoffstetter, 1955) rather than those
of the Boinae. Erycines have previously not been reported before
the Eocene (Hecht, 1959). Without characters of the neural
arch, this specimen cannot be identified more specifically.

ORDER CROCODILIA
SUBORDER EUSUCHIA

Family Crocodylidae

Subfamily Crocodylinae

Leidyosiichus sternbergi Gilmore, 1910

Many crocodile teeth (MCZ 3648) occur in the Bug Creek
sample and, on the basis of shape and general appearance, are sim-
ilar to those of the type specimen and referred Lance Forma-
tion material (Estes, 1964).

24

BREVIORA

No. 337

Fig. 4. Vertebra of boid snake, MCZ 3669; a, anterior, b, lateral, c, pos-
terior, and d, ventral views; Bug Creek Anthills, Hell Creek Formation,
McCone County, Montana; X 4.

Subfamily Alligatorinae
Brachychampsa montana Gilmore, 1911

Bulbous, low-crowned teeth (MCZ 3650), scutes, fragmentary
limb, vertebral and skull bones are relatively common. The teeth
are relatively smaller than those of the type specimen, which is
also from the Hell Creek Formation of Montana. Some of the
skull and skeletal fragments referred here probably belong to
Leidyosuchus.

ORDER SAURISCHIA

SUBORDER THEROPODA

INFRAORDER COELUROSAURIA

Family Coeluridae ?
Unidentified genus and species

These delicate teeth (MCZ 3694) are serrated only on their
posterior borders and are rare at Bug Creek Anthills. Estes (1964)

1969 LATE CRETACEOUS LOWER VERTEBRATES 25

noted that similar Lance teeth with fine serrations resembled those
of Velociraptor, but in shape and size they resemble those of other
coelurids as well, including the Oldman Formation Chirostenotes
and the Triassic Coelophysis. Generic identifications cannot be
based on teeth of this sort (PI. lb).

Paronychodon lacustris Cope, 1876

A few teeth (MCZ 3645) of this peculiar type occur in the Bug
Creek sample (PI. Id). Paronychodon teeth have been figured
several times, most recently by Russell (1935, pi. 2, fig. 8). The
flattened lingual side may indicate an anterior tooth; other re-
ferred specimens with the same coarse striations (PI. le) lack
the flattened side and may be from the posterior part of the tooth
row. The Bug Creek specimens appear to be unworn, and are
unserrated; Lance Formation and Judith River Formation speci-
mens may be either serrated or unserrated.

Theropoda? incertae sedis

A few teeth (MCZ 3680) of the straight-sided type figured by
Estes (1964, fig. 69b) occur in the Bug Creek material.

INFRAORDER DEINONYCHOSAURIA ?

Family Dromaeosauridae ?

These short-crowned, sharply recurved teeth have about 40
serrations per 5 mm anteriorly, 30 posteriorly. They are of the
same tooth type as those described by Cope as Laelaps, from the
Judith River Formation of Montana. The Bug Creek specimens
(MCZ 3695, PI. IC) are the smaflest teeth of this type that we
have seen, although several Lance Formation specimens approach
them in size.

Laelaps is often included in Dryptosaurus, and Lance Forma-
tion specimens of the Laelaps type were questionably referred to
Dryptosaurus by Estes (1964). The recent revision of Colbert
and RusseU (1969) suggests that Laelaps might better be included
with the dromaeosaurs. Generic reference of this type of teeth,
however, is difficult; the teeth from Bug Creek Anthills do not
fall within any of the ranges of tooth serration number outlined
by Colbert and Russell (1969, pp. 39-40). Sloan (1969, pers.
comm.) found Gorgosaurus in other Bug Creek localities (Bug
Creek West, Harbicht Hill), but its teeth are larger than any of
the above.

26 BREVIORA No. 337

ORDER ORNITHISCHIA
SUBORDER ORNITHOPODA

Family Hypsilophodontidae
Thescelosaurus neglectus Gilmore, 1913

A few teeth (MCZ 3649) resemble teeth of this species (see
Sternberg, 1940, p. 483, figs. 1-8).

Family Pachycephalosauridae ?

Several teeth (MCZ 3729) are obtuse, with little development of
a cingulum (PI. lA), and may belong to this family.

Family Hadrosauridae
Anatosaiinis sp.

These teeth (MCZ 3646) are relatively common (for dinosaur
teeth) in the Bug Creek sample and, in general, are somewhat
smaller than most specimens from the Lance Formation. Most
of the specimens appear to be heavily-worn teeth shed in replace-
ment.

SUBORDER CERATOPSIA

Family Ceratopsidae
Triceratops sp.

As for Anatosaiirus, the teeth referred here (MCZ 3647) are
relatively common, smaller in general than most Lance Formation
specimens, and are heavily worn, shed in replacement.

CONCLUSIONS

The striking feature of the lower vertebrate faunal list from Bug
Creek Anthills (BCA) described here is its similarity to that from
localities in the Lance Formation of Wyoming, especially Uni-
versity of California locality V5620 (Estes, 1964), as shown in
Table 1. Mammals are excluded so that direct comparison of the
lower vertebrates can be made. The taxonomic similarity is sup-
plemented by a broad similarity of relative abundance of genera in
the two localities; holosteans, salamanders, and lizards are the
most common groups at both sites. Rather significant differences
in the relative abundance of individual species occur, however, and
may be ecological in origin; these differences will be discussed in
more detail in Estes and Berberian (in press), and the mammalian

1969 LATE CRETACEOUS LOWER VERTEBRATES 27

species included, but some general comments may be made now.

The diminished abundance, diversity and specimen size of the
sharks at Bug Creek Anthills relative to those of V5620 probably
indicates that BCA had reduced access to marine conditions.
Turdes and lizards, also common at V5620, are again poorer in
diversity, abundance, and preservation at BCA. Since the pre-
sumably more active, free-swimming salamanders and bony fishes
are diverse, abundant, and well preserved at BCA, it is probable
that a riparian habitat supporting the lizards and turtles was less
accessible. The relative number of specimens and the specimen
size of dinosaur material is less at BCA than at V5620; whether
this condition is the result of depositional environment or is a re-
flection of the imminent extinction of the group is unknown, but
the latter possibility is the more probable. Although further an-
alysis may disprove this, we suggest that the vertebrate fossils at
BCA were deposited in the larger, more open waterways of the
floodplain, and that the site of deposition was farther from the
adjacent shores, than it was at V5620. Sloan and Van Valen
(1965) came to rather similar conclusions based on analysis of the
mammalian fauna, and have expressed the faunal differences
among the various Hell Creek Formation localities in terms of the
proximal and distal community concept of Shotwell ( 1955).

Some mammals from Bug Creek Anthills are different from those
of the Lance Formation localities and indicate a "Paleocene as-
pect," according to Sloan and Van Valen (1965). As noted
above, the non-mammalian fauna of BCA differs little from that
of the Lance. The additional taxa at BCA may be grouped into
three categories: (1) Unique records, (2) Forms previously
known only from the Paleogene, and (3) Records of taxa already
known from both Cretaceous and Tertiary deposits but not occur-
ring in the Lance Formation. The unique forms include the bold
snake, the teiid lizard Peneteius, and the amphiumid salamander
Proamphiuma. As these animals are representatives of living fam-
ilies and are related to modern genera, with the possible exception
of the boid, it might be said that they are representative of an
"advanced" element in the faunule. Since their stratigraphic ranges
are unknown, however, this group does not specifically indicate
a "Paleocene aspect." Only two taxa compose the second group.
Palaeolabrus, a fish otherwise known only from a single Paleogene
locality in Belgium, is a poor indicator of the Paleocene affinity
of the BCA lower vertebrate fauna. One of the undescribed
baenid turtles is known from middle Paleocene specimens from

28 BREVIORA No. 337

Wyoming; this is its first Cretaceous record. Again, it is a rare
form and a poor indicator of Paleocene relationships. The third
group includes the fish cf. Paralbula, the frog Scotiophryne, and the
turtle Adociis. These taxa are known from other Cretaceous and
Paleocene deposits in North America, although they do not occur
in the Lance Formation, and thus do not indicate "a Paleocene
aspect" for the fauna. Adocits, apparently an advanced deriva-
tive of the late Cretaceous Basilcinys, is most similar to specimens
of Adociis from the Arapahoe Formation of Colorado, which,
like the Hell Creek Formation, is of late Cretaceous age (Weimer,
1960, fig. 2).

We therefore conclude that the faunal differences between Bug
Creek Anthills and V5620 localities indicate mainly minor eco-
logical differences, and that there is little evidence for a significant
difference in age or faunal type between the two sites. The geo-
graphic position of the Hell Creek localities, the unique character
of the mammalian fauna, and the presence of lower vertebrates
absent in the Lance Formation, probably indicate that, during
late Cretaceous time, the Bug Creek Anthills fauna was more closely
associated with a northern, perhaps climatically more temperate,
aspect of the floodplain fauna of the Western Interior of North
America.

Connections to a marine environment were less available at BCA
than at V5620 as indicated by the reduced shark fauna. Since the
Bearpaw Sea was regressing southeastward at latest Cretaceous
time (Weimer, 1960), it may have been closer to V5620 than to
BCA. It is also possible that the difference in access to marine
conditions was only the result of local conditions, and that geo-
graphically, the actual sea was approximately equidistant from
both locafities; no evidence as to the actual geographic position
of the strandline is available for latest Cretaceous time, however.

ACKNOWLEDGEMENTS

It is a pleasure to acknowledge our gratitude to Robert Sloan,
whose enthusiastic support of our eft'orts and generous contribu-
tion of fossil material to many museums and universities has made
this study possible. In addition to the MCZ collection listed and
described in the body of this paper, we have had access to other
collections, material from which has proved useful in assessment
of variation although no additional taxa were present in them.
These included unnumbered specimens from the University of
Minnesota and the Saint Paul Science Museum, and catalogued

1969 LATE CRETACEOUS LOWER VERTEBRATES 29

material as follows: Princeton University nos. 20468, 20554-
20569, 20571-20573, 20806-20807; University of Kansas nos.
12453-12477; American Museum of Natural History nos. 8123-
8129, 8131, 8140-8157, 9321-9329. Thomas Rich kindly sent a
small collection from the University of California (Berkeley) as
well as his personal collection.

Welcome discussion and help on various matters have come
from Colin Patterson on Lonchidion, Cecile Poplin on amiids. Max
Hecht on frogs, Allen Greer on scincid lizards, C. Wilson Kerfoot
on Ciittysarkus, Mary Mickevich on dental histology of fishes.
Dale Russell on dinosaur teeth, and Eugene Gaffney on baenid
turdes. We are grateful to Donald Baird and Robert Sloan for
helpful comments on the manuscript.

Figures 1-4 are by Laszlo Meszoly; the photographs are by
Fred Maynard. This research was supported in part by National
Science Foundation grants GB-1683, GB-4303, and GB-7176 to
the senior author.

LITERATURE CITED

AUFFENBERG, W.

1961. A new genus of fossil salamander from North America. Amer.

Midi. Nat., 66: 456-465.
1963. The fossil snakes of Florida. Tulane Studies Zool., 10: 131-
216.
Brown. B.

1907. The Hell Creek beds of the Upper Cretaceous of Montana:
Their relation to contiguous deposits, with faunal and floral
lists and a discussion of their correlation. Bull. Amer. Mus.
Nat. Hist. 23: 823-845.
Casier, E.

1966. Faune ichthyologique du London Clay. British Museum (Nat.
Hist.), Text -|- atlas; xiv -f 496 pp.

1967. Le Landenien de Dormaal (Brabant) et sa faune ichthyolo-
gique. Mem. Inst. Roy. Sci. Nat. Belgique, 156: 1-66.

Colbert, E., and D. Russell

1969. The small Cretaceous dinosaur Dromaeosaiirits. Amer. Mus.
Nat. Hist., Novit. No. 2380: 1-49.
Cope, E.

1876. On some extinct reptiles and batrachians from the Judith River
and Fox Hills beds of Montana. Proc. Acad. Nat. Sci. Phiia.,
1876: 340-359.
ESTES, R.

1964. Fossil vertebrates from the late Cretaceous Lance Formation,
eastern Wyoming. Univ. Calif. Publ. Dept. Geol. Sci., 49: 1-
180.

30 BRE\aoRA No. 337

1965. A new fossil salamander from Montana and Wyoming. Copeia,
1965, No. 1:90-95.

1969a. Studies on fossil phyllodont fishes: Interrelationships and evo-
lution in the Phyllodontidae (Albuloidei). Copeia, 1969, No.
2:317-331.

1969b. The Batrachosauroididae and Scapherpetontidae, late Creta-
ceous and early Cenozoic salamanders. Copeia, 1969, No. 2:
225-234.

1969c. Prosirenidae, a new family of fossil salamanders. Nature (Lon-
don). 224: 87-88.

1969d. Relationships of two Cretaceous lizards (Sauria, Teiidae).
Mus. Comp. Zool., Harvard Univ., Breviora 317: 1-8.

1969e. The fossil record of amphiumid salamanders. Mus. Comp.
Zool., Harvard Univ., Breviora 322: 1-11.

1969f. A new fossil discoglossid frog from Montana and Wyoming.
Mus. Comp. Zool.. Harvard Univ., Breviora 328: 1-7.

1969g. A scincoid lizard from the Cretaceous and Paleocene of Mon-
tana. Mus. Comp. Zool., Harvard Univ., Breviora 331: 1-9.

1969h. Two new late Cretaceous fishes from Montana and Wyoming.
Mus. Comp. Zool., Harvard Univ., Breviora 335: 1-15.

1970. New fossil pelobatid frogs and a review of the genus Eopelo-
bates. Bull. Mus. Comp. Zool., Harvard Univ. In Press.

ESTES, R., AND P. BERBERIAN

1969. Amia (zi^Kindleio) fragosa (Jordan), a Cretaceous amiid fish,

with notes on related forms. Mus. Comp. Zool., Harvard

Univ., Breviora 329: 1-14.
In press. Paleoecology of a late Cretaceous vertebrate community

from Montana. Mus. Comp. Zool., Harvard Univ., Breviora

343.

GiLMORE, C.

1910. Leidyosuchiis sternbergii, a new species of crocodile from the
Ceratops beds of Wyoming. Proc. U. S. Nation. Mus., 38:
485-502.

1911. A new fossil alligator from the Hell Creek beds of Montana.
Proc. U. S. Nation. Mus., 41 : 297-302 .

1913. A new dinosaur from the Lance Formation of Wyoming.

Smithson. Misc. Coll., 61 : 1-5.
1919. Reptilian faunas of the Torrejon, Puerco, and underlying Up-
per Cretaceous formations of San Juan County, New Mexico.
U. S. Geol. Surv. Profess. Pap. 119: 1-68.
1928. Fossil lizards of North America. Mem. Nat. Acad. Sci., 22:
ix + 201 pp.
Gregory, W.

1933. Fish skulls: a study of the evolution of natural mechanisms.
Trans. Amer. Philos. Soc, 23: 75-481.
Hay, O.

1908. The fossil turtles of North America. Carnegie Inst. Washing-
ton, Publ. 75: iv + 568 pp.

1969 LATE CRETACEOUS LOWER VERTEBRATES 31

Hecht, M.

1959. Reptiles and Amphibians. //;: P. McGrew, The geology and
paleontology of the Elk Mountain and Tabernacle Butte r.rea,
Wyoming. Bull. Amer. Mus. Nat. Hist., 119: 121-176.
Hecht, M., and R. Hoffstetter

1962. Note preliminaire sur les amphibiens et les squamates du Lan-
denien superieur et du Tongrien de Belgique. Inst. Roy. Sci.
Nat. Belgique, 38: 1-30.
Hoffstetter, R.

1955. Squamates de type moderne. //;: J. Piveteau, Traite de Paleon-
tolcgie. Paris: Masson et Cie., 5: 606-662.

Janot, C.

1967. A propos des amiides actuels et fossiles. Colloq. Intern. C. N.
R. S., 163: 139-153.
Jordan, D.

1927. Kindleio, a new genus of cichlid fishes from the Upper Cretace-
ous of Alberta. Canad. Field Nat., 41: 145-147.
Lambe, L.

1902. New genera and species from the Belly River series (mid-Cre-
taceous). Contr. Canad. Paleon., 3: 25-81.
Langston, W.

1956. The shell oi Basilemys vcirialosa (Cope). Ann. Rept. Nat. Mus.
Canada, 142: 155-165.

Leidy, J.

1856. Notice of remains of extinct reptiles and fishes, discovered by
Dr. F. V. Hayden in the Bad Lands of the Judith River, Ne-
braska Territory. Proc. Acad. Nat. Sci. Philadelphia, 1856:
72-73.
MacAlpin, a.

1947. Paleopsephiints wilsoni. a new polyodontid fish from the Up-
per Cretaceous of Montana, with a discussion of allied fish, liv-
ing and fossil. Cont. Mus. Paleon.. Univ. Michigan, 6: 167-
234.
Marsh, O.

1889. Discovery of Cretaceous Mammalia, part 2. Amer. Jour. Sci.,

Ser. 3,38: 177-180.
1892. Notice of new reptiles from the Laramie Formation. Amer.
Jour. Sci.. Ser. 3, 43: 449-453.
Meszoely, C.

1970. North American fossil anguid lizards. Bull. Mus. Comp. Zool.,
Harvard Univ.. in press.
Patterson. C.

1966. British Wealden Sharks. Bull. British Museum (Nat. Hist.),
Geol., 11: 283-350.
Russell, L.

1935. Fauna of the Upper Milk River beds, Southern Alberta. Trans.
Roy. Soc. Canada, Ser. 3, 29: 115-127.

32 BREVIORA No. 337

Shotwell, J.

1955. An approach to the paleoecology of mammals. Ecology, 36:
327-337.

1958. Intercommunity relationships in Hemphillian (mid-Pliocene)
mammals. Ecology, 39: 271-282.
Slaughter, B., and M. Steiner

1968. Notes on rostral teeth of ganopristine sawfishes, with special
reference to Texas material. Jour. Paleon., 42: 233-239.
Sloan, R., and L. Van Valen

1965. Cretaceous mammals from Montana. Science, 148: 220-227.
Sternberg, C.

1940. Thescelosaiinis edmontonensis, n. sp., and classification of the
Hypsilophodontidae. Jour. Paleon., 14: 481-494.
Stromer, E., and W. Weiler

1930. Ergebnisse der Forschungsreisen Prof. E. Stromers in den
Wuesten Aegyptens. VI: Beschreibung von Wirbeltier-Resten
aus dem nubischen Sandsteine Oberaegyptens und aus aegypti-
schen Phosphaten nebst Bemerkungen ueber die Geologic der
Umgegend von Mahamid in Oberaegypten. Abh. Bay. Akad.
Wiss. Munchen, (N. F.) 7: 1-42.
Webb, R.

1962. North American Recent soft-shelled turtles (family Triony-
chidae). Univ. Kansas Publ. Mus. Nat. Hist., 13: 429-611.
Weimer, R.

1960. Upper Cretaceous stratigraphy. Rocky Mountain Area. Bull.
Amer. Assoc. Petrol. Geol., 44: 1-20.
WiLIMOVSKY, N.

1956. Protoscaphirhynchiis sqiiamosiis, a new sturgeon from the Up-
per Cretaceous of Montana. Jour. Paleon. 30: 1205-1208.

Williams, E.

1950. Variation and selection in the cervical central articulations of
living turtles. Bull. Amer. Mus. Nat. Hist.. 94: 51 1-561.
Woodward, A.

1895. Catalogue of the Fossil fishes in the British Museum, Part 3.
London, xliii + 544 pp.

(Received 22 August 1969.)

1969

LATE CRETACEOUS LOWER VERTEBRATES

33

A

B

D

E

\

Plate I. Dinosaur teeth. A. ?Pachycephalosauridae, MCZ 3729; B, ?Coe-
luridae, MCZ 3694; C, ?Diomaeosauridae, MCZ 3695; D, Paronycliodon
lacuslris, MCZ 3645; E, cf. Paronycliodon laciistris, MCZ 3645; Bug Creek
Anthills, Hell Creek Formation. McCone County, Montana; scale in mm.

DO NOT CiRCULATF.

, ^ IX ^ ^ r 3m^ r'Jjtj ^^-U£. COMP. ZOCI

■ LIBRARY.

B R E V I O' R A ,970

Meseuiiiii of Comparative Zoolog^ERsiTYJ

Cambridge, Mass. 30 December. 1969 Number 338

A NEW FROG OF THE GENUS DISCODELES
(RANIDAE) FROM GUADALCANAL ISLAND

Walter C. Brown'

and
T. Preston Webster

Abstract. A fourth species of Solomon Island's Discodeles, D. nialii-
kiiihi (Ranidae), is described from 28 specimens collected near Malukuna,
Guadalcanal, Solomon Islands. Among the species collected are three (H\-
Ici littea. a new species of Batraclxylodes, and Platymuntis myersi) previous-
ly unknown from Guadalcanal.

INTRODUCTION

Boulenger (1918a. 1920). in his treatment of the probable
evolutionary lines within the large genus Rana, noted that the hy-
laranid and discodelid subgenera shared one characteristic, horizon-
tal groove associated with the dilated disks at the tips of the toes
and sometimes the fingers, but differed in the characteristics, struc-
ture of the omosternum. and degree of union of the outer meta-
tarsals. He also (1918b) pointed out the close affinities of the
genera Cornufer and Platymantis (both are currently included in
one genus, Cornufer, by many authors) with Discodeles, particu-
larly in the presence of a horizontal groove on the toe disks and in
the union of the outer metatarsals throughout most or all of their
length. Noble ( 193 1 . p. 521 fi"), recognizing the subfamily Cornu-
ferinae in the Ranidae, was obligated to raise Discodeles and Hy-
larana to generic rank, since Rana (sensii stricto) was retained in
the subfamily Raninae. He noted the close relationships of the gen-
era Ceratobatrachus, Discodeles, Cornufer and Platymantis in

' Stanford University. Stanford. California, and Menlo College. Menlo
Park, California

2 BREVIORA No. 338

terms of the structural features discussed above and their geo-
graphical unity. He also called attention to the presence of a dis-
tinct, fleshy, median papilla on the tongue in Discodeles and its
absence in Coninfer, a character somewhat difficult to use in many
preserved specimens. Noble further surmised that direct develop-
ment was characteristic of the species in this group of genera. This
suggestion was based on knowledge of direct development of Dis-
codeles opisthodon and Cornujer guentheri (op. cit.. p. 64). This
has been substantiated since for a number of species of Platyman-
tis (includes Cornujer), see Alcala, 1962.

Prior to 1968. dating from the time of the descriptions of the
three species of Discodeles ( Boulenger, 1884), collections in the
archipelago failed to reveal any further species. These explora-
tions, however, did indicate that the three species were rather wide-
ly dispersed within the Solomons (Table 1). Therefore the dis-
covery in 1968 by one of us (Webster) of a population in the
mountains of Guadalcanal which represents a very distinct species
was somewhat of a surprise.

The populations which have been found in New Britain and the
Admiralty Islands are seemingly closely related to one or the other
of the three species originally recognized. Rcma ventricosa Vogt,
1912, (= vogti, Hediger, 1934) is very similar to and may be
conspecific with D. opisthodon (Brown, 1952, pp. 36-37; Zweifel.
1960, pp. 4-7); and Rana bufoniformis cognata Hediger (1934)
is synonomized with D. guppyi (Zweifel. 1960. p. 4).

Discodeles malukuna sp. nov.

Holotype. MCZ 79462, a mature male, collected at Malukuna
area, elevation about 2500 feet. Guadalcanal Island, July 2, 1968.
by T. Preston Webster.

Parcitypes. MCZ 79463-79489. from the same area as the holo-
type.

Diagnosis. A relatively small Discodeles, at least for a sample
of males as compared to other known species, largest available male
measuring 58.3 mm in snout-vent length; head broad relative to
snout-vent length (Fig. 2); eye relatively large, slightly less than,
to about equal to, length of snout; first finger longer than second
or fourth when adpressed; tips of fingers not dilated, rounded, lack-
ing a terminal circummarginal groove; first and fifth toe about one-
fourth webbed, web reaching the distal edge of the tubercle of the
inner toe and failing to reach or barely reaching the distal tubercle

1969

A NEW DISCODELES

TABLE 1

Distribution of the species of Dlscodeles In the Solomon Islands.

Bou

salnviU

e

New Georgia

Guadalcanal

San Cristobal |

Group

Group

Group

Group 1

3

f*^

m

(a

^

^

<

3

u

(0 (9

c

va

(« c

<U

V)

>

oo

&o

)-•

rt O

U

J£

n

0

J3 Oj

o <0

•o «

p

C X

ta

o

W >^

(U

o ^

4)

4) ^^

1

Species
Discodeles

a

to

tb

cn

z

o

>

K

U£

z

o:

o

u

uj

X

malukuna

Dlscodeles

X

X

X

X

X

X

X

X

X

X

X

X

X

buf oniformis

Discodeles

X

X

X

X

X

X

X

X

X

X

guppyl

Discodeles

X

X

X

X

X

X

X

X

opisthodon

of the outer toe; second and third toes about one-half webbed; out-
er metatarsals not firmly united in the distal fourth, or less, of their
length; venter grayish brown to blackish, usually with numerous
white spots.

Description. A relatively small Discodeles as compared to
known species; the snout-vent length, about 43 to 58 mm for 11
males, is unknown for adult females (the largest female, measur-
ing 63.2 mm in snout- vent length, is possibly almost mature as
judged by the fact that the oviducts are just beginning to show
shallow convolutions); habitus tapering from the head to the groin;
head breadth 115 to 123 per cent of head length for 11 mature
males, 89 to 99 per cent of tibia length, and 44 to 49 per cent of
snout-vent length for the same group of males; diameter of eye 91
to 102 per cent of length of snout; diamenter of tympanum 39 to 55
per cent of diameter of eye; interorbital distance 17 to 21 percent
of the head length for 1 1 mature males; upper jaw round, pointed,
protruding; loreals moderately oblique and concave; fingers with-
out webs; first finger longer than the second or fourth when ad-
pressed; tips of fingers bluntly rounded, undilated (Fig. 4), without
a circummarginal groove; subarticular tubercles large, prominent
but not pointed; a few large, faint, palmar tubercles; metatarsal
tubercles low, rounded, outer broadly oval, inner more elongate;
toes with moderate webs, strongest between second — third and
third — fourth; web reaching the distal edge of the subarticular
tubercle of the inner toe and not or barely reaching the distal tu-
bercle of the outer toe (Fig. 5); tips of toes rounded, moderately
dilated and depressed, the ventral part of the disk separated from

4 BREVIORA No. 338

the dorsal by a circummarginal groove; subarticular tubercles
large and prominent but not sharply protruding; outer metatarsal
tubercles small and round, inner moderately narrow and long (Fig.
5); skin of dorsum, lateral surfaces, and upper surfaces of the
limbs smooth except for a pair of folds between the postorbital
and axillary region; large, flat tubercles on posterior and ventral
surface of thighs; venter faintly granular posteriorly in some in-
stances.

Color (in preservative). Grayish brown to blackish brown on
the dorsum; lateral surfaces and limbs lighter grayish to grayish
brown; upper loreal region, borders of eyelids, upper tympanum
and edges of urn-shaped folds (especially at posterior ends) black;
lips with two prominent dark bars; a black anal patch and hind
limbs marked by narrow, dark transverse bands; venter grayish to
blackish brown with numerous light spots.

Measurements of holotype (in mm). Snout-vent length 58.3;
length of head to posterior edge of tympanum 22.3; breadth of
head 25.9; diameter of eye 7.7; diameter of tympanum 3.7; length
of snout 8.2; interorbital distance 4.0; length of hind limb 92.0;
length of tibia 28.5; length of third finger to proximal edge of basal
tubercle 7.9; length of first finger to base of tubercle 5.6.

Etymology. The species name is that of the type locality in the
central mountains of Guadalcanal.

Remarks. Of the four known species of the genus Discodeles, all
present in the Solomon Islands, this is the most readily distin-
guished because of the extremely reduced webbing between the toes,
the lack of warty tubercles on the dorsum and upper surfaces of the
hind limbs ( these are present in some degree even in D. giippyi, the
least tuberculate of the other three species), the urn-shaped pattern
formed by the pair of folds on the anterior part of the dorsum, and
the conspicuous black markings.

The intuitive evaluation, head broader than for other known
species of the genus (particularly for larger, mature specimens), is
borne out by plotting head breadth against snout-vent length (Fig.
2), although this ratio is only slightly greater than for bujonijor-
mis. Moreover, dilTerential growth patterns for the four species are
evident in several proportions: (1) head breadth relative to tibia
length, Figure 3; (2) interorbital distance relative to head length,
Figure 1.

In such characters as the lack of dorsal tubercles, the reduced
webbing of the toes (intermediate between that of Platy mantis
guppyi and Discodeles opisthodon), and the anterior, urn-shaped
folds, D. malukuna is somewhat intermediate between the other

1969 A NEW DISCODELES 5

species of Discodeles and those species of Platymantis exhibiting
the greatest degree of webbing between the toes. This appears to
further substantiate the probable close relationship between Dis-
codeles and Platymantis postulated by Noble ( 193 1, p. 523) .

Previously seven species of amphibians were known from Guad-
alcanal (Brown, 1952). These included:

Hyla thesaurensis

Batrachylodes vertebralis

Ceratobatrachiis gnentheri

Discodeles guppyi

Platymantis guentheri

Platymantis papuensis weberi

Rana papua novaebritanniae
In addition to Discodeles malukuna, Webster's collection from
the mountainous area around Malukuna includes specimens of
three other previously unrecorded species: Hyla lutea, Batrachy-
lodes new species/ and Platymantis myersi.

ACKNOWLEDGMENTS

Thanks are due to Mr. G. F. C. Dennis and his son Tony and to
the people of Malkuna. Partial support in the field was provided
by NSF grant GY 4555 to T. P. Webster and by NSF 6944 to Er-
nest E. Williams. Illustrations were prepared by Walter Zawojski,
Stanford University.

LITERATURE CITED

Alcala, a. C.

1962. Breeding behavior and early development of frogs of Negros,
Philippine Islands. Copeia, 1962: 679-726.

BOULENGER, G. A.

1884. Diagnosis of new reptiles and batrachians from the Solomon
Islands, collected and presented to the British Museum by H.
B. Guppy, Esq. Proc. Zool. Soc. London, 1884: 210-213.

1918a. On the Papuan, Melanesian, and North Australian species of
the genus Rana. Ann. Mag. Nat. Hist., Ser. 9, 1 : 236-242.

1918b. Remarks on the batrachian genera Cor/;/^/fr-Tschudi, Platy-
mantis GiJnther, Simomantis, g.n., and Staiirois Cope. Ann.
Mag. Nat. Hist., Ser. 9, 1: 372-375.

^ Brown (in press) : described on Bougainville material.

6 BREVIORA No. 338

1920. A monograph of the South Asian, Papuan, Melanesian, and
Australian frogs of the genus Rana. Rec. Indian Mus., 20: 1-
226.
Brown, W. C.

1952. The amphibians of the Solomon Islands. Bull. Mus. Comp.
Zool., 107: 1-64. pis. 1-8.
Hediger, H.

1934. Beitrag zur Herpetologie und Zoogeographie Neu Britanniens.
Zool. Jahrb., Abt. Syst., 65: 441-582.
Noble, G. K.

1931. Biology of the amphibia. New York: McGraw-Hill Co., xiii -|-
577 pp.

VOGT, T.

1912. Beitrag zur Reptilien und Amphibienfauna der Siidseeinseln.
Sitzber. Ges. Naturf. Freunde, Berlin, Jahrg. 1912: 1-13.

ZWEIFEL, R. G.

1960. Results of the 1958-1959 Gilliard New Britain Expedition 3.
Notes on the frogs of New Britain. Amer. Mus. Novitates,
2023: 1-27.

(Received 30 September 1969.)

10 r-

bJ

I-

10

Q

^6

OQ

Fig.

0

o malukuna

o opisthodon

• guppyi

♦ bufoniformis

1

0 15 20 25 30 35 40 45 50 55

HEAD LENGTH

Fig. 1. Differences in interorbital distance relative to head leng.h for four
species of Discoileles. (Measurements in mm.)

1969

A NEW DISCODELES

60 r-

50

40 —

<

UJ

cr
m 30

Q

<

UJ

I 20

10

Fig 2

o malukuna

o opisthodon

• guppyi

♦ bufoniformis

30 40 50 60 70 80 90

SNOUT-VENT LENGTH

100

110

120

Fig. 2. Differences in head breadth relative to snout-vent length for four
species oi Discodeles. (Measurements in mm.)

70 i-

60

X 50

h-
O
<

u

<

30

20

10

Fig. 3

o malukuna

o opisthodon

• guppyi

♦ bufoniformis

20

30

40

50 60

TIBIA LENGTH

70

80

90

Fig.
species

3. Differences in head breadth relative to tibia length for four
of Discodeles. (Measurements in mm.)

8

BREVIORA

No. 338

■ /

"\

.■^^

Fig. 4. Discodeles malukuna: inferior view of hand.

1969

A NEW DISCODELES

f

f

.-'''

■

/

/

.•^

y

imA <ta. f^^nA^tt^-i. i£ici^r*nHtf

Fig. 5. Discodeles malukima: inferior view of foot.

DO NOT CIWCULATF.

LIBRARY.

BREVIORAo 1970

MiaseeiM of Comparative Zoolq^VERsiTY]

Cambridge, Mass. 30 December. 1969 Number 339

THE GENUS STENOONOPS (ARANEAE,
OONOPIDAE) IN PANAMA AND THE WEST INDIES

Arthur M. Chickering

Abstract. A total of seventeen species of the genus Stenoonops are rec-
ognized in this paper. Eleven species are described as new. Six of these are
described from Jamaica, W. 1.. two from Puerto Rico, two from St. Thomas,
U. S. Virgin Islands, and one from Antigua, B. W. I. Scapliioides reducta
Bryant. 1942 from St. Croix, U. S. V. I., is placed in Stenoonops.

This is the fifth paper in the series planned for publication on
the various genera in the Family Oonopidae in Central America
and the West Indies.

Once more I take this opportunity to express my appreciation for
the continued aid and encouragement extended to me by the staff
of the Museum of Comparative Zoology, Harvard University.
Grants GB-1801 and GB-5013 from the National Science Founda-
tion have made it possible for me to carry on extensive collecting
activities in Panama, Costa Rica, Florida, and the West Indies dur-
ing the years 1963-1968 and to continue my studies in the Museum
of Comparative Zoology during that period. My acknowledge-
ments are gratefully extended to the following persons for the loan
of types and other specimens to aid me in more coinpletely under-
standing the genus under consideration: Dr. J. G, Sheals and Mr.
D. J. Clark of the Department of Zoology, British Museum (Nat-
ural History); Dr. \VilHs J. Gertsch. recently retired froin the
American Museum of Natural History, New York City, and also
Mr. Wilton Ivie, Research Fellow in the American Museum of
Natural History.

The types of Stenoonops phonetiis sp. nov. will be deposited in
the American Museum of Natural History. All other types de-
scribed in this paper together with my entire collection of the genus
Stenoonops will be deposited in the Museum of Comparative Zo-
ology, Harvard University.

2 BREVIORA No. 339

Genus Stenoonops Simon, 1891

The type species is Stenoonops scabriculus Simon by monotypy,
based upon a male from St. Vincent, B. W. I., which I have had on
loan for study. Since the recognition of the genus in 1891, only
three additional new species have been recognized and assigned to
this genus in the region at present under study. One additional
species from St. Croix, U. S. Virgin Islands, was recognized as new
by Miss Bryant and assigned to a new genus but is now included
in Stenoonops. One species has been known only from females;
one species has been known from both sexes; all others have been
known only from males.

In addition to the general family characteristics, the features of
the genus Stenoonops as observed in this study may be stated as
follows: total length varies from about 1.1 mm to 2.1 mm. The
general shape and height of the carapace differs considerably in
the various species included in this paper. The ventral margin of
the carapace is often plainly serrated. The eyes are six in number,
arranged in two rows in a fairly compact group; the ALE are
nearly always at least slightly the largest of the six and the distance
between them varies considerably; the posterior row is typically
somewhat recurved but occasionally is straight or gently procurved.
The chelicerae, maxillae and lip usually appear to be without con-
spicuous modifications; a notable exception is furnished by S. in-
solitus sp. nov. The legs are moderately long and slender without
great differences in length, without true spines but with many hairs
and bristles. Trichobothria are present but their exact placement
and specific dift'erences among species have not been accurately
determined. The male palp is simple and without special modifica-
tions except in the case of the tarsus, which exhibits a high degree
of variation among the recognized species under study at the pres-
ent time. In species like 5. portoricensis and S. iusolitus sp. nov.,
the distal palpal tarsal features are conspicuously characteristic,
but in numerous other species, these structures are so obscure and
so difficult to differentiate that errors are always confronting the
investigator. Jets of semen extending from the distal end of the
tarsal bulb are often quite conspicuous, sometimes deceptive and,
perhaps, lead to erroneous conclusions regarding the extent and
shape of the embolus. I think I have more difiicultics in making
decisions at this point than in any other genus thus far studied in
the family Oonopidae. The abdomen is completely lacking a dorsal
scutum but scuta are more or less well defined at the anterior end
of the venter. Typically, there is a fairly well-defined epigastric

1969 STENOONOPS SPIDERS 3

scutum reaching forward to meet the pedicel and a narrow ventral
scutum, often overlooked in descriptions, just behind the genital
groove. I think I have usually been able to discern these scuta
even when they lack the usual yellowish or brownish coloration.
The epigynal areas in females are also often very difficult to dis-
cern and to represent in drawings. In these cases reliance has
been placed upon characteristics of the eyes, features of the cara-
pace, sternum, etc. The spinnerets appear to be without special
modifications; no chitinous sclerite has been observed at the base
of the spinnerets as is so often observed in other genera in the fam-
ily. A pair of dark-colored subsurface structures typically appear
on the venter shortly anterior to the bases of the anterior spinner-
ets, but they are lacking in some species.

Although uncertainties still persist to trouble the author of this
paper, it seems at this time that I must recognize a total of seven-
teen species of the genus Stenoonops from the region under consid-
eration. This list may be given as follows: Stenoonops cletus sp.
nov.; Stenoonops diniotus sp. nov.; Stenoonops econotiis sp. nov.;
Stenoonops halatus sp. nov.; Stenoonops hojfi sp. nov.; Stenoonops
insolitiis sp. nov.; Stenoonops lucradus sp. nov.; Stenoonops maca-
bus sp. nov.; Stenoonops minutus Chamberlin and Ivie; Stenoonops
nitens Bryant; Stenoonops noctucus sp. nov.; Stenoonops padiscus
sp. nov.; Stenoonops petrunkevitchi Chickering; Stenoonops pho-
netus sp. nov.; Stenoonops portoricensis Petrunkevitch; Stenoonops
reductus (Bryant); Stenoonops scabriculus Simon. Only one spe-
cies is known from each of the following: Panama; St. Vincent,
B. W. I.; and Antigua, B. W. I. Three species are known from
Puerto Rico and six are known from Jamaica, W. I. Two species
are known from St. Croix, U. S. V. I.; three are known from St.
Thomas, U. S. V. I.; and four have been identified from St. John,
U. S. V. I. Three males and two females from Virgin Gorda, Brit-
ish Virgin Islands, are at present assigned to S. nitens Bryant with
some uncertainty. A single female from St. Lucia, B. W. I., is also,
with some uncertainty, assigned to S. nitens Bryant. One species
has been recorded from both North and South Bimini, Bahama
Islands. Either because of ocular anomalies or the obscurity of
features necessary for definite identification, a few specimens from
the region under study have been left unassigned to any species.

Stenoonops cletus sp. nov.
Figures 1-6

Holotype. The male holotype is from Jamaica, W. I., St. Cath-
erine Parish, three miles east of May Pen, November 22, 1957.

4 BREVIORA No. 339

The name of the species is an arbitrary combination of letters.

Description. Total length 1.28 mm, exclusive of the slightly ex-
tended spinnerets. Carapace 0.6 mm long; 0.46 mm wide oppo-
site posterior border of second coxae where it is widest; 0.3 mm
tall; rises somewhat just posterior to PME and then gendy arches
along midline to beginning of posterior declivity with tallest point
shortly before the beginning of the declivity (Fig. 1 of paratype);
with no median thoracic groove or pit; surface finely granular;
sparsely covered by short, dark hair; ventral margin very finely
serrated. Eyes: six in two rows as usual; in a compact group (Fig.
2); viewed from above, posterior row definitely recurved and oc-
cupies about three-fourths of width of carapace at that level. Ratio
of eyes ALE : PME : PLE=7 : 6 : 6. All eyes somewhat irregular
in outline; long diameters always used for measurements. ALE
separated from one another by about one-seventh of their diam-
eter; separated from PME by about one-eighth of their diameter
and from PLE by about one-third of the diameter of the latter.
PME contiguous for nearly one-third of their circumference; tan-
gent to PLE. Posterior row wider than anterior row in ratio of
about 17 : 12. Only a moderate amount of black pigment in ocu-
lar region. Height of clypeus equal to about two-thirds of the
diameter of ALE; clypeus somewhat convex and provided with
six stiff bristles, two of these very long, others short. Chelicerae:
vertical; essentially parallel; tapering gradually from base to tip;
about 0.11 mm long; fang regularly curved and slender; no teeth
observed along fang groove in dissected specimen; apparently with-
out basal boss. Maxillae and lip essentially typical of the genus.
Sternum: somewhat convex; finely granular; only moderately lobed
and grooved; only a little longer than wide between second and
third coxae; posterior end apparently somewhat bilobed and just
opposite bases of fourth coxae which are separated by nearly 1.2
times their width. Legs: 4123 in order of length; true spines lack-
ing but some bristles are spiniform. Palp: short; simple except
for the tarsus which is normally inflated and inconspicuously termi-
nated (Figs. 3-4); right palpal tarsus with a conspicuous jet of
semen extended from embolus. Abdomen: ovoid; widest in the
middle; 0.43 mm tall; quite typical of the genus in general; epi-
gastric and ventral scuta fairly well oudined but division between
the two very obscure; the pair of subsurface structures just anterior
to spinnerets only dimly oudined. Color in. alcohol: carapace a
clear yellowish brown; sternum somewhat lighter; legs yellowish
with some variations; abdomen yellowish white with some irides-
cence; scuta somewhat darker yellowish.

1969 STENOONOPS SPIDERS 5

Female paratype. Total length 1.54 mm. Carapace 0.56 mm
long; 0.44 mm wide; 0.26 mm tall; otherwise essentially as in male
(Fig. 5). Eyes: in nearly all essential features, these seem to be
in agreement with those of the male; ALE are somewhat further
apart than in the male. Chelicerae, maxillae, lip, and sternum all
nearly as described for the male holotype. Legs: 4123 in order
of length as in male; palp probably without a claw; trichobothria
observed but exact placement not determined. Abdomen: 1.04
mm long; regularly ovoid in shape; scuta and epigynal area essen-
tially as shown in Figure 6.

Diagnosis. This species seems to be closely related to Stenoon-
ops nitens Bryant. It differs from that species chiefly in the fea-
tures of the male palp, the sternum, and the shape of the carapace.

Records. The described female paratype was taken with the
male holotype. Male and female paratypes are in my collection
from several localities in the following named parishes in Jamaica,
W. I.: Trelawney; St. Andrew; St. Thomas; and St. Catherine.

Stenoonops econotus sp. nov.
Figures 7-9

Holotype. The female holotype is from Puerto Rico, Mayagiiez,
University farm north of University campus, January 31, 1964.
The name of the species is an arbitrary combination of letters.

Description. Total length 1.47 mm exclusive of the somewhat
extended spinnerets; including these organs total length is 1.54 mm.
Carapace 0.56 mm long; 0.4 mm wide opposite posterior borders
of second coxae where it is widest; about 0.2 mm tall (Figs. 7-8);
plainly serrated along ventral margins; gently arched along mid-
line from PME to beginning of steep posterior declivity opposite
third coxae; unevenly granular along lateral and posterior surfaces;
no median thoracic groove or pit present. Eyes: six as usual in a
compact group; occupy five-sixths of width of carapace at that level;
posterior row gently recurved, viewed from above and wider
than anterior row in ratio of about 16 : 13. Ratio of eyes
ALE : PME : PLE = nearly 6:5:5. ALE separated from
one another only by a broad line; separated from PME and PLE
by nearly the same distance. PME contiguous to one another and
barely separated from PLE. Height of clypeus equal to width of
somewhat oval ALE. Chelicerae, maxillae, and lip apparently
typical of females of the genus. Sternum: longer than wide in
ratio of about 4 : 3'; widest just behind second coxae at level of
a lateral lobe; lobes and grooves well developed; posterior end

6 BREVIORA No. 339

pointed and extended just opposite bases of fourth coxae which
are separated by nearly 1.3 times their width. Legs: 4123 in order
of length; otherwise as usual in the genus. Abdomen: ovoid and
quite typical of the genus in general; the typical dark, ventral, sub-
surface spots barely visible just anterior to spinnerets; epigastric
and ventral scuta well outlined but occupy only about half the
width of the venter at that level (Fig. 9); epigynal area obscurely
distinctive. Color in alcohol: carapace a dark brown with irregu-
lar, darker, radiating, lateral and posterior streaks caused by con-
centrations of granulations; only a moderate amount of black
pigment in ocular area; sternum a lighter yellowish brown; legs
and mouth parts yellowish with variations; abdomen yellowish on
dorsum with an irregular white, transverse band about two-thirds
from base and a pair of white areas at anterior end; venter generally
yellowish with scuta a light yellowish brown.

Diagnosis. This species appears to be somewhat closely related
to Stenoonops nitens Bryant, but it differs from that species in re-
spect to color pattern, scuta, and relative position of eyes.

Record. The male is unknown and there are no paratypes.

Stenoonops nitens Bryant
Figures 10-13

Stenoonops nitens Bryant, 1942: 321, figs. 5, 10. The holotype male from
St. Croix, U. S. Virgin Islands, is in the Museum of Comparative Zoology,
Harvard University, examined.

Numerous specimens of this species have been collected on St.
Croix, and a male and a female have been selected for study; a
brief report follows :

Male. Total length exclusive of the extended spinnerets and
chelicerae 1.14 mm; including these structures total length is 1.26
mm. Carapace 0.53 mm long; 0.4 mm wide; nearly 0.18 mm tall
(Fig. 10). Ratio of eyes ALE : PME : PLE = nearly 5 : 4.5 : 4.
ALE separated from one another by nearly one-fourth of their
diameter (Fig. 11). The clypeus is quite porrect and with height
about equal to diameter of ALE. The scuta are fairly well oudined
and closely similar to those of female but are less well delineated.
The male palp is very obscure. The holotype was described as
having "two very slender points at tip." I believe that these
"points" are simply the lateral boundaries of the embolus, which
extends as a very short tube that varies in appearance in different
individuals and in different positions in the same individual (Fig.
12).

1969 STENOONOPS SPIDERS 7

Female. Total length exclusive of the spinnerets and cheUcerae
1.21 mm; including these structures total length is about 1.29 mm.
Carapace nearly the same as in male. Figure 13 represents the
abdominal scuta and epigynal area of the female.

The specimens now assigned to this species include a group re-
garded for a considerable length of time as another new species.
At present they are regarded as what might be considered a variety
by some taxonomists. My collection now contains many specimens
of both sexes from the following localities: St. Croix, U. S. Virgin
Islands, vicinity of Frederiksted, March, 1964 and September,
1966; several localities on St. Thomas, U. S. Virgin Islands, Feb-
ruary, 1964 and July-August, 1966; St. John, U. S. Virgin Islands,
March 1964 and September, 1966. Two females collected near
Mayagiiez, Puerto Rico, January, 1964, may possibly belong to this
species. This suggests the probability that careful collecting in the
West Indies will greatly extend the known territorial range of this
species. A female collected in Costa Rica on Mt. Irazu on July
27, 1965, has been regarded as a possible member of this species
but is now left unplaced.

Stenoonops portoricensis Petrunkevitch
Figures 14-18

Stenoonops portoricensis Petrunkevitch, 1929: 72, figs. 61-63. The holo-
type male from Toa Alta, Puerto Rico, W. I., is in the American Museum
of Natural History, New York City, examined. Roewer, 1942: 280; Bonnet,
1958: 4156.

I did not collect the male of this species during my period of
field work in Puerto Rico in 1964. The male holotype is the only
male specimen known at present and, unfortunately, it is in poor
condition for study. The right palp, figured by Dr. Petrunkevitch,
is missing, and I have found it impossible to see the parts in the
left palp as originally described. More careful collecting is needed
to add to our knowledge of the genus in this island. The female
described here was for a time regarded as representing a new spe-
cies but, following observations on the holotype and awaiting
more complete knowledge regarding the species, I have decided
to treat it as a female of Stenoonops portoricensis.

Female. Total length 1.63 mm, including the extended spinner-
ets; excluding the spinnerets, the total length is about 1.5 mm. Car-
apace 0.6 mm long; about 0.5 mm wide opposite second coxae
where it is widest; about 0.31 mm tall; considerably raised behind

8 BREVIORA No. 339

PME and then nearly level to beginning of steep posterior declivity;
regularly rounded along ventral margin; closely similar to cara-
pace of male holotype (Figs. 14, 18). Eyes: essentially as indi-
cated for male in original description of holotype; ratio of eyes
ALE : PME : PLE = nearly 6:6:5 (Fig. 15). Clypeus ob-
scured by tuft of bristles. Mouth parts and legs appear essentially
as described for the holotype male except for the palps; these parts
with many short hairs but no true spines. Sternum: with many
short, black hairs obscuring faintly indicated lobulations. Abdo-
men also with many short, black hairs as described for the male
(Fig. 16); with a weakly outlined epigastric scutum occupying less
than half the width of the venter and a very narrow ventral scutum;
epigynal area very indistinct (Fig. 17).

Records. The described female is from the University farm north
of the University campus, Mayagiiez, Puerto Rico, January 25,
1964. Two females collected in the vicinity of Mayagiiez have not
yet been placed satisfactorily in any known species. Several speci-
mens in the collection of the Museum of Comparative Zoology,
tentatively assigned to this species, have been reassigned to other
genera or left unplaced because of uncertainties.

Stenoonops dimotus sp. nov.
Figures 19-20

Holotype. The female holotype is from Jamaica, W. I., St.
Andrew, Ferry, near Red Hills Road, October 28, 1957. The
name of the species is an arbitrary combination of letters.

Description. Total length 2.09 mm, exclusive of the somewhat
extended chelicerae; including these structures, total length is 2.14
mm. Carapace about 0.81 mm long; 0.57 mm wide opposite poste-
rior border of second coxae where it is widest; 0.29 mm tall; ven-
tral margin finely serrated; otherwise essentially typical of females
of the genus (Fig. 19). Eyes: six in a compact group; outlines
obscure and irregular. Viewed from above, posterior row definitely
recurved and longer than anterior row in ratio of about 17 : 13;
ALE conspicuously silvery, PME less so; posterior row occupies
about three-fourths of width of carapace at that level; PME ap-
parently tilted outward so that outlines are very unusual in appear-
ance (Fig. 20). Ratio of eyes ALE : PME : PLE = nearly 6 :
5:5. ALE separated from one another by about their radius,
and from PME and PLE by slightly more than half their radius.
PME contiguous to one another and narrowly separated from PLE.
Only a small amount of black pigment in ocular area; area usually

1969 STENOONOPS SPIDERS 9

black is here reddish brown. Height of clypeus equal to slightly
more than the diameter of ALE; clypeus with several stiff, black
bristles and quite porrect. CheHcerae, maxillae, and lip apparently
quite typical of females of the genus. Sternum: plainly lobed in
fairly typical manner; longer than wide in ratio of nearly 4:3;
widest just behind second coxae where a lobe extends laterally; with
numerous incurved bristles; posterior end pointed and not quite
reaching proximal end of fourth coxae, which are separated by
nearly their width. Legs: 4123 in order of length; only slight dif-
ference in lengths of fourth and first legs; many hairs, bristles and
several trichobothria observed but no true spines. Abdomen: typi-
cal in general of females of the genus; scuta very obscure with
outlines difficult to follow; ventral scutum hardly more than a
transverse line; obscurity of epigynal area prevents any clear under-
standing of the characters of this region. Color in alcohol: cara-
pace light yellowish brown, darkened somewhat along lateral sides
and margins; with almost no black pigment in ocular region;
sternum nearly like carapace; first two pairs of legs only a little
lighter than carapace; third and fourth legs somewhat more yellow-
ish; abdomen nearly clear white throughout with many short, dark
hairs.

Diagnosis. This species appears to be closely related to Ste-
noonops portoricensis Petrunkevitch and, apparendy, belongs to
the group containing Stenoouops macabus sp. nov. and several
other species. The eyes appear to be unique, however, and the
ventral scutum is unusually narrow.

Records. The male is unknown and there are no paratypes.

Stenoonops insolitus sp. nov.
Figures 21-26

Holotype. The male holotype is from Jamaica, W. L, St. Cath-
erine Parish, three miles east of Old Harbour, October 21, 1957.
The name of the species is a Latin adjective meaning unusual,
chosen because of the unusual anatomical features.

Description. Total length 1.56 mm. Carapace 0.68 mm long;
0.48 mm wide opposite second coxae where it is widest; 0.26 mm
tall; rises slightly behind PME and then is nearly level along mid-
line to beginning of posterior declivity where there are rough gran-
ulations; surface in general finely granular, ventral margin very
finely serrated but visible only when viewed from above; otherwise
essentially typical of the genus (Fig. 21). Eyes: six as usual in the
genus. Ratio of eyes ALE : PME : PLE = nearly 6.5 : 5.5 : 5.

10 BREVIORA No. 339

Posterior row gently recurved; occupies about two-thirds of width
of carapace at that level; outlines of eyes somewhat obscure. ALE
separated from one another by nearly one-third of their diameter
(Fig. 22); nearly contiguous to PLE and separated from PME by
a broad line; PME contiguous for about one-third of their circum-
ference and barely separated from PLE. Height of clypeus nearly
equal to two-thirds of the diameter of ALE. Chelicerae and lip
essentially typical of the genus. Maxillae: general form shown in
Figure 23; distal end of each provided with a pair of very long and
conspicuous spines not seen in any other species known to me.
Sternum: longer than wide in ratio of about 3:2; not extended
between fourth coxae, which are separated by somewhat more than
their width; with marginal lobes and grooves moderately well devel-
oped; third coxae subglobose, others more elongated. Legs: 4123
in order of length; no true spines observed but spiniform bristles
are numerous; trichobothria observed on tibiae and metatarsi.
Palp: essential features shown in Figures 24-26. Distal end of
tarsus appears divided; minor differences in the form of the distal
end of the palpal tarsi have been noted among the paratypes. Ab-
domen: ovoid; typical of the genus in general; surface with numer-
ous short, dark hairs; the usual scuta are present but barely visible.
Color in alcohol: carapace and sternum a clear yellowish brown;
legs and mouth parts somewhat lighter; abdomen almost white.

Female paratype. Total length 2.07 mm. Carapace 0.77 mm
long; 0.55 mm wide opposite second coxae where it is widest;
about 0.28 mm tall; otherwise essentially as in male. Eyes: ratio
of eyes almost exactly as in male. Spiniform brisdes on clypeus
conspicuous. Posterior row of eyes occupies about two-thirds of
width of carapace at that level; otherwise essentially as in male.
Chelicerae, maxillae, and lip as usual in females of the genus;
there are no indications of the conspicuous maxillary spines pos-
sessed by the male. Sternum : features somewhat more conspicuous
than in male; slightly more than two-thirds as wide as long; bristles
conspicuous on lobes and between bases of fourth coxae, which are
separated by somewhat more than their width; grooves between
marginal lobes quite conspicuous. Legs: 4123 in order of length;
essentially as in male holotype. Abdomen: essentially as in male
except for the scuta and epigynal area, which are also very poorly
oudined. Color in alcohol: essentially as in male; the paired dark
spots on the venter just anterior to the first pair of spinnerets, so
frequently seen in species of this genus, were not seen in the holo-
type male but are present here as three pairs of faintly outlined

1969 STENOONOPS SPIDERS 11

dots somewhat irregularly placed and are completely lacking in
some paratypes.

Diagnosis. This species appears to belong to the group of spe-
cies represented by Stenoonops portoricensis Petrunkevitch and
Stenoonops petnmkevitchi Chickering, but the distinctive features
of the male palp and, especially, the maxillary spines of the male
definitely establish it as a new species.

Records. The described female paratype was taken with the
holotype male. Two male paratypes were also taken with the
holotype. Eight males and five females are also in the collection
from Jamaica, W. I., as follows: St. Ann Parish near Moneague,
May 20, 1956 (C. C. Hoff); St. Andrew Parish, Stony Hill, Octo-
ber 18, 1957; St. Catherine Parish, Guanaboa Vale, December 4,
1957; St. Catherine Parish, Evarton, November 29, 1957, and
Ferry, June 19, 1954.

Stenoonops lucradus sp. nov.
Figures 27-28

Holotype. The male holotype is from St. Thomas, U. S. Virgin
Islands, August 25, 1966, in the vicinity of Charlotte Amalie from
hay and weed debris. The name of the species is an arbitrary com-
bination of letters.

Description. Total length exclusive of the extended spinnerets
1.45 mm; including the spinnerets total length is about 1.55 mm.
Carapace about 0.59 mm long (slightly overlapped by abdomen);
nearly 0.45 mm wide opposite second coxae where it is widest;
0.24 mm tall; gently arched from PME along midline to a slightly
raised portion just before beginning of steep posterior declivity;
first half of declivity very steep, second half gradually descending;
a tuft of hairs accentuates the top of the declivity. Eyes: six as
usual in a compact group; very little difference in long axes of the
three pairs of eyes but some differences in shape have been noted;
posterior row slightly recurved (Fig. 27). ALE separated from
one another by about two-fifths of their diameter; separated from
PME by about three-tenths of their diameter and from PLE by
one-fifth of their diameter. PME contiguous as usual and separated
from PLE only by a line. With several bristles and a moderate
amount of black pigment in ocular area. Clypeus with numerous
bristies and with height about equal to the diameter of ALE.
Chelicerae, maxillae, and lip obscure and difficult to observe but
apparently typical of the genus and without special modifications.

12 BREVIORA No. 339

Sternum: moderately convex; moderately lobed and only slightly
grooved; with posterior end unusually lobed just before bases of
fourth coxae, which are separated by about their width; margins
with stiff, black, incurved bristles; first coxae somewhat elongated,
others nearly subglobose. Legs: 4123 in order of length; no true
spines observed; many hairs and brisdes present. Palp: only tarsus
inflated; with a very distinctive, twisted embolus (Fig. 28). Abdo-
men: typical of the genus; with scuta and genital area barely dis-
cernible. Color in alcohol: carapace a medium yellowish brown,
lighter along median region and darker along lateral sides; with
numerous black hairs; ocular region with a moderate amount of
black pigment; sternum nearly the same as the carapace; legs
somewhat lighter; abdomen nearly white with many short, black
hairs; with no evidence of the dark spots on the venter just anterior
to the spinnerets.

Female paratype. Total length 1.46 mm, exclusive of the ex-
tended spinnerets; including the spinnerets total length is 1.6 mm.
Carapace about 0.61 mm long (slightly overlapped by abdomen);
nearly 0.46 mm wide opposite second coxae where it is widest;
about 0.28 mm tall; gently arched from iust behind PME to begin-
ning of posterior declivity nearly opposite third coxae; first half of
declivity steep, lower half a very gradual descent to posterior bor-
der. Eyes: quite different in appearance frorrt those of male; seen
from above, posterior row very slightly procurved but almost
straight; posterior row occupies about two-thirds of width of cara-
pace at that level and is wider than anterior row in ratio of nearly
3 : 2. Ratio of eyes ALE : PME : PLE == nearly 5:5:4. PME
tilted outward so they appear very narrow when viewed from above;
their appearance is similar to that of the PME of Stenoonops dimo-
tus sp. nov. from Jamaica. ALE separated from one another by a
little less than their radius and separated from PME by nearly the
same distance; separated from PLE by a little less than this dis-
tance. PME contiguous as usual and nearly contiguous to PLE at
one point. Height of clypeus nearly equal to diameter of ALE.
Chelicerae, maxillae, lip, and sternum essentially as in male. Legs:
4123 in order of length but with only small differences among them.
Abdomen: almost exactly as in male with scuta hardly discernible;
epigynal area without observable distinctive features. Color in al-
cohol: also almost exactly as in male holotype.

Diagnosis. This is another species that appears to be closely re-
lated to Stenoonops petrimkevitchi Chickering and Stenoonops
portoricensis Petrunkevitch. The features of the palpal tarsus are
quite distinctive and definitely serve to establish it as a new species.

1969 STENOONOPS SPIDERS 13

Records. The described female is from St. John, U. S. Virgin
Islands, July 23, 1966. One male was taken with the female.

Stenoonops macabus sp. nov.
Figures 29-30

Holotype. The male holotype is from Jamaica, W. I., St. Cath-
erine Parish, three miles east of May Pen, November 22, 1957.
The name of the species is an arbitrary combination of letters.

Description. Total length 1.78 mm. Carapace 0.66 mm long;
0.5 mm wide opposite second coxae where it is widest; 0.22 mm
tall; slightly raised just behind PME and then nearly level along
midline to beginning of steep posterior declivity; with no median
thoracic groove or pit; surface very finely granulate and corrugated;
quite typical of the genus in general. Eyes: six in two rows as usual
in the genus. Ratio of eyes ALE : PME : PLE = nearly 11:9:7
(outlines somewhat obscure). Posterior row occupies about two-
thirds of width of carapace at that level and is moderately recurved,
viewed from above (Fig. 29). ALE separated from one another by
about their radius, from PME by about half their radius and from
PLE by about one-fifth of their diameter. PME contiguous for
about one-fourth of their circumference and nearly tangent at one
point to PLE. Posterior row of eyes wider than anterior row in
ratio of about 6:5. The clypeus bears a row of long, stiff bristles;
height of clypeus nearly equal to three-fourths of long axis of ALE.
Chelicerae as usual in the genus. Maxillae long, slender, slightly
convergent. Lip essentially typical of the genus; reaches nearly to
middle of maxillae. Sternum: essentially typical of the genus; with
lobes and grooves along margins moderately well developed;
bluntly terminated shortly before bases of fourth coxae which are
separated by a little less than their width. Legs: 4123 in order of
length; spines are lacking as usual; otherwise typical of the genus.
Palp: all segments except tarsus simple and without special modi-
fications; tarsus inflated and extended as a long, curved embolus
(Fig. 30) not seen elsewhere in the genus. Abdomen: 1.1 mm
long; 0.59 mm wide; ovoid in general; spinnerets as usual in the
genus; scuta are very poorly outlined. Color in alcohol: carapace
with the usual yellowish brown color; sternum and legs somewhat
lighter; abdomen very light yellowish white with numerous, small,
irregular, white, glistening flecks.

Diagnosis. This species also appears to belong to the group of
species represented by Stenoonops portoricensis Petrunkevitch, but
the palp is very distinctive and plainly establishes it as a new species.

14 BREVIORA No. 339

Records. The female is unknown and there are no male para-
types.

Stenoonops petrunkevitchi Chickering

Figures 31-33

Stenoonops petrunkevitchi Chickering, 1951: 241, figs. 29-30. The male
holotype from Barro Colorado Island, Panama Canal Zone, July, 1938, is
in the Museum of Comparative Zoology.

Female. There is no certainty that the female described here be-
longs in this species. It seems reasonable, however, to make this
assumption because of the close resemblance to the male and be-
cause no other species have been reported from this region as far
as I have been able to learn. Total length 1.37 mm. Carapace 0.61
mm long; 0.44 mm wide opposite posterior borders of second coxae
where it is widest; 0.20 mm tall; somewhat raised just behind PME
and then level along midline to beginning of fairly steep posterior
declivity opposite anterior border of third coxae; surface finely
granular; dorsal surface with a double row of dark hairs all directed
toward the midline (Figs. 31-33). Eyes: six in a compact group;
apparently very near to those of male but outlines are indistinct.
Chelicerae essentially as in male. Maxillae: nearly parallel; some-
what widened distally; about twice as long as lip. Lip: about as
wide distally as at base; sternal suture appears nearly straight.
Sternum: essentially as in male; third and fourth coxae nearly
globose; first and second somewhat elongated. Legs: essentially
as in male. Abdomen: essentially as in male; scuta barely visible;
epigynal area without distinctive features.

Records. The described female is from Barro Colorado Island,
Panama Canal Zone, August 16, 1954. Since the male holotype
was collected in July, 1938, the following specimens have been
added to the collection: two males taken on Barro Colorado Island,
February 8, 1958; two males taken in the Panama Canal Zone For-
est Preserve, January 6, 1958. One male collected by Dr. A. M.
Nadler on Barro Colorado Island, April, 1953, is also referred to
this species. Apparently this species is not abundant in regions of
Panama where I have collected. What is probably an immature fe-
male from the Canal Zone Forest Preserve, July 23, 1950, may rep-
resent a new species.

Stenoonops hoffi sp. nov.
Figures 34-42

Holotype. The male holotype is from Jamaica, W. I., St. Thomas
Parish, Morant Point, May 6^ 1956. Collected by Dr. C. C. Hoff

1969 STENOONOPS SPIDERS 15

from axils of thatch palm. The species is named after the collector.
Description. Total length 1.14 mm including extended spinner-
ets; excluding the spinnerets, total length is 1.06 mm. Carapace
0.51 mm long; 0.38 mm wide opposite second coxae where it is
widest; 0.18 mm tall; sharply narrowed opposite PLE; gently
arched along midline from PME to beginning of steep posterior
declivity; broad median region from PME to beginning of posterior
declivity smooth and shining; lateral and posterior surfaces irregu-
larly granular; with ventral margin finely serrated; few hairs or
bristles observed except in ocular region where there are several
stiff brisdes (Figs. 34-35). Eyes: six as usual in a compact
group; viewed from above, posterior row moderately recurved and
wider than anterior row in ratio of about 5:4; posterior row occu-
pies nearly the entire width of carapace at that level. Ratio of ALE
: PME : PLE = 6 : 5 : 4.5. ALE bright silvery; separated from
one another by nearly one-fourth of their long axis; only slightly
oval; separated from PME by about one-sixth of their long axis
and from PLE by a line; with considerable black pigment in ocular
area. Clypeus somewhat porrect; with several spiniform bristles;
height equal to about three-fourths of the long axis of ALE (Fig.
36). Chelicerae: vertical; parallel; apparendy without special
modifications. Maxillae: long, slender, convergent and touching
beyond lip. Lip: much wider than long; otherwise typical of the
genus. Sternum: quite convex; surface smooth through narrow me-
dian region but elsewhere irregularly granulate; quite plainly lobed
and grooved along lateral margins; longer than wide in ratio of
about 5:4; widest opposite interval between second and third
coxae but nearly as wide between first and second coxae; third and
fourth coxae subglobose, others somewhat elongated; posterior end
blundy terminated but apparendy slighdy bilobed; extended be-
tween bases of fourth coxae which are separated by nearly 1 .5 times
their width. Legs: 4123 in order of length; no true spines ob-
served. Palp: tarsus inflated as usual with terminal features diffi-
cult to observe but apparendy distinctive (Figs. 37-40); femur,
patella, and tibia simple and without special modificadons. Abdo-
men: regularly ovoid; spinnerets somewhat extended beyond tip
of abdomen; about two-thirds as wide as long; epigastric and nar-
row ventral scuta moderately well outlined (Fig. 41). Color in
alcohol: carapace brown with irregular streaks on lateral and poste-
rior surfaces caused by the granuladons previously referred to;
sternum somewhat lighter brown; legs and mouth parts yellowish
with variations; coxae darker; abdomen yellowish and with con-
siderable iridescence; scuta somewhat darker.

16 BREVIORA No. 339

Female paratype. Total length, exclusive of the slightly extended
spinnerets 1.21 mm. Carapace 0.54 mm long; 0.4 mm wide; nearly
0.19 mm tall. Otherwise the briefly described female paratype ap-
pears to be essentially almost identical to the male in nearly all
features. Even the abdominal scuta and epigynal area differ only
slightly from corresponding regions of the male (Fig. 42).

Diagnosis. This species seems to be closely related to Stenoonops
minutus Chamberlin and Ivie, now known from Florida and South
Bimini, Bahama Islands. It differs from that species in respect to
color pattern and the minute and obscure features of the male pal-
pal tarsus.

Records. The described female paratype is from Jamaica, W. I.,
St. Thomas Parish, Morant Point. Oct. 14, 1957; taken from palm
debris on the ground. Two male paratypes were taken with the
holotype. Two female paratypes are in the collection from Tre-
lawney Parish, fourteen miles east of Falmouth, May 15, 1956
(C. C. Hoff), and one additional female is from St. Catherine
Parish, Guanaboa Vale, December 4, 1957.

Stenoonops minutus Chamberlain and Ivie

Stenoonops minutus Chamberlin and Ivie, 1935: 8, figs. 1-2, 14. The
male holotype from Tampa, Florida, U. S. A., is now on semipermanent
loan from the Museum of the University of Utah in the American Museum
of Natural History, New York City, N. Y., examined. Roewer, 1942: 280;
Bonnet, 1958: 4156; Chickering, 1969 (in press).

Several specimens of both sexes on loan from the American Mu-
seum of Natural History have been assigned to this species. All of
these were collected on South Bimini, Bahama Islands, B. W. I.,
as follows: May, 1951 (W. J. Gertsch and M. A. Cazier); June,
1951 (M. A. Cazier and C. and P. Vaurie); July, 1951 (C. and
P. Vaurie); April, 1952 (E. Mayr); December, 1952 and March,
1953 (A. M. Nadler). For some time these specimens were re-
garded as representing a new species and were described as such.
Following a comparison with specimens of Stenoonops minutus
Chamberlin and Ivie from Florida, they have now been assigned
to that species. Minor differences have been noted but they are
now regarded as well within the range of normal variation within
a species.

Stenoonops phonetus sp. nov.
Figures 43-46

Holotype. The male holotype is from Puerto Rico, Cidra, Treas-
ure Island, February 26-27, 1955; collected by Dr. A. M. Nadler.

1969 STENOONOPS SPIDERS 17

It will be deposited in the American Museum of Natural History,
New York, N. Y. The name of the species is an arbitrary combina-
tion of letters.

Description. Total length 1.19 mm. Carapace 0.55 mm long;
nearly 0.39 mm wide opposite second coxae where it is widest;
nearly 0.20 mm tall; raised considerably iust behind PME and then
nearly level to beginning of steep posterior declivity opposite an-
terior borders of third coxae (Figs. 43-44); without a median
thoracic groove or pit; central regions smooth with few hairs; lateral
and posterior surfaces irregularly granulate; ventral margins plainly
serrated. Eyes: six in two rows as usual; in a compact group nearly
as broad as the carapace at that level; posterior row gendy re-
curved and wider than anterior row in ratio of nearly 7 : 6. Ratio
of eyes ALE : PME : PLE = 12 : 11 : 10. PME considerably
longer than wide; some irregularities in outlines of eyes noted. ALE
separated from one another by nearly one-third of their diameter;
separated from PME only by a line and contiguous to PLE; PME
contiguous for nearly one-third of their circumference and con-
tiguous to PLE for only a short distance. Clypeus considerably
porrect; height equal to nearly two-thirds of the long diameter of
ALE; with several quite long, slender spines. Chelicerae, maxillae,
and lip quite typical of the genus as far as observed. Sternum: quite
conspicuously lobed and grooved; longer than wide in ratio of
nearly 4:3; widest just behind second coxae where a lobe extends
laterally; posterior end terminates just opposite bases of fourth
coxae which are separated by somewhat more than their width;
surface very finely granulate. Legs: 41 =23 in order of length; true
spines are lacking as usual. Palp: typical of several species in the
genus with features very inconspicuous (Fig. 45); a jet of semen
is very conspicuous, however. Abdomen: quite typical of the genus;
with epigastric and ventral scuta fairly well outlined and essentially
as in female; with internal sacs just anterior to spinnerets on the
venter showing through fairly clearly. Color in alcohol: carapace
a medium brown with irregular darker patches along lateral and
posterior surfaces caused by granulations; only a moderate amount
of black pigment in ocular area; sternum a somewhat lighter yellow-
ish brown; legs and mouth parts yellowish with little variation;
palpal tarsus white.

Female paratype. Total length 1.29 mm. Carapace nearly 0.55
mm long; nearly 0.40 mm wide just behind second coxae where it is
widest; general shape and other features essentially as in male; with
ventral border serrated as in male. Eyes: essentially as in male but
occupying a little less of the width of the carapace; ALE separated

18 BREVIORA No. 339

a little further from PME than in the male; otherwise spaced essen-
tially as in that sex; clypeus equally porrect and fully as tall as in
male. Chelicerae: very finely rugulose in front; otherwise essen-
tially as in male. Maxillae and lip also essentially as in male. Ster-
num: longer than wide in ratio of about 17 : 14; somewhat less
conspicuously grooved and lobed than in male; extends to just
opposite bases of fourth coxae which are separated by nearly five-
fourths of their width. Legs essentially as in male. Abdomen:
essentially typical of females of the genus; epigastric and ventral
scuta moderately well outlined (Fig. 46); paired dark sacs ante-
rior to spinnerets fairly well outlined. Color in alcohol essentially
as in male.

Diagnosis. This species seems to be closely related to Stenoonops
minutus Chamberlin and Ivie now known from Florida and South
Bimini, Bahama Islands, and for a time was included with that spe-
cies. A careful comparison, however, seems to show that it differs
from that species significantly with respect to the following features:
relative position of the eyes; appearance of the clypeus; shape of
the carapace; and, especially, in respect to the conspicuously ser-
rated ventral border of the carapace (Figs. 43-44).

Records. The described female paratype is from Mayagiiez,
Puerto Rico, January 19, 1955. In addition to the holotype and
the described female paratype, I have had the following specimens
for study: one male taken with the described female paratype; one
female from Rio Piedras, March 14, 1959; two males from Rio
Piedras, Humacao Co., March 2, 1955. All of these are from
Puerto Rico and all were collected by Dr. A. M. Nadler. All
specimens mentioned here will be deposited in the American Mu-
seum of Natural History, New York, N. Y.

Stenoonops scabriculus Simon
Figures 47-48

Stenoonops scabriculus Simon, 1891: 565. The holotype male from St.
Vincent, B. W. I., is in the British Museum (Natural History), examined.
Simon, 1892: 447; 1893: 296; Petrunkevitch, 1911: 129; 1928: 87; Roewer,
1942: 280; Bonnet, 1958: 4156.

The male from St. Vincent, B. W. I., is the type species of the
genus. The original description was very briefly given in Latin and
was not accompanied by figures. A male, believed to be the holo-
type, was loaned for study from the British Museum (Natural
History). The date accompanying this species, however, was given
as 94-10-17.

1969 STENOONOPS SPIDERS 19

Male holotype. Total length 1.39 mm, exclusive of the somewhat
extended spinnerets and chelicerae; including these parts total
length is 1.51 mm. Carapace 0.65 mm long; 0.43 mm wide; 0.2
mm tall; slightly arched along midline from PME to beginning of
moderately steep posterior declivity; surface finely granular; with
no indication of a median thoracic groove or pit; ventral margin
finely serrated. Eyes: posterior row occupies nearly three-fifths
of width of carapace at that level; only slightly wider than anterior
row; moderately recurved (Fig. 47). Ratio of eyes ALE : PME :
PLE r= 6 : 5.75 : 5. ALE separated from one another by about
one-third of their long axis; barely separated from PME and PLE.
PME contiguous to one another for one-third of their circumfer-
ence; separated from PLE by a line. Height of clypeus equal to
slightly more than one-half the long axis of ALE. Chelicerae:
nearly vertical; slightly extended; essentially parallel; with no spe-
cial modifications observed. Maxillae: probably divided distally
but fragility of specimen prevents close examination. Lip: hidden
by numerous hairs but apparently without special modifications.
Sternum: moderately lobed and grooved; surface finely granular;
longer than wide in ratio of about 9:7; widest opposite interval
between second and third coxae where well-developed lobes extend
laterally; posterior end narrowed and bluntly rounded just opposite
bases of fourth coxae which are separated by a Httle more than five-
fourths of their width. Legs: 4123 in order of length; with few
bristles and with no spines observed. Palp: femur of moderate
length, both patella and tibia short; essential features of the tarsus
shown in Figure 48; tarsus with numerous broad hairs on the cym-
bium and the usual dorsal pad of short, fine hairs. Abdomen: ovoid;
0.81 mm long; 0.5 mm wide a little behind the middle; spinnerets
as usual in the genus; epigastric and ventral scuta hardly visible.
Color in alcohol: carapace and sternum yellowish brown; only a
small amount of black pigment in ocular area; legs and mouth parts
lighter than carapace and sternum with variations; abdomen very
light yellowish.

Records. Simon reported this species from Venezuela as well
as from St. Vincent, B. W. I. I failed to collect the species during
my visit to this island in 1966 and, as far as I have been able to
determine, it has not been reported since the holotype was taken.

Stenoonops halatus sp. nov.
Figures 49-52

Holotype. The female holotype is from Antigua, B. W. L, Fig
Tree Hill, near Old Road, August, 1967. Collected by Mrs. Elsa

20 BREVIORA No. 339

Sabath. The name of the species is an arbitrary combination of
letters.

Description. Total length 1.62 mm exclusive of the slightly
extended spinnerets (Fig. 49); including the spinnerets total length
is 1.67 mm. Carapace about 0.61 mm long; about 0.5 mm wide
opposite posterior border of second coxae where it is widest; about
0.27 mm tall; raised slightly just behind PME and continues to
rise somewhat to beginning of steep posterior declivity (Fig. 50)
nearly opposite third coxae; surface finely granular and with few
hairs. Eyes: six as usual in a compact group; posterior row mod-
erately recurved and occupying slightly more than three-fourths
of width of carapace at that level and wider than anterior row in
ratio of about 6:5. Ratio of eyes ALE : PME : PLE = about
8:7: 6.5. ALE separated from one another by somewhat less
than one-half their radius; barely separated from PLE and sep-
arated from PME by only a line. PME contiguous for about one-
fourth of their circumference and nearly contiguous to PLE for
only a short distance; with a moderate amount of black pigment
in ocular area. Clypeus quite porrect and with height nearly equal
to radius of ALE; with several spiniform bristles (Fig. 51). Cheli-
cerae, maxillae, and lip essentially typical of females of the genus.
Sternum: obscured by curled, fragile legs; apparently quite typical
of the genus; longer than wide in ratio of about 19 : 15; quite dis-
tinctly lobed and grooved; slightly bilobed at posterior end, which
is just opposite the middle of the fourth coxae, which are separated
by slighdy less than 1.5 times their width. Legs: 42=13 in order
of length; no true spines observed; all legs moderately long and
slender. Palp with many fine hairs and bristles. Abdomen: ovoid;
longer than wide in ratio of nearly 4:3; epigastric and ventral scuta
quite clear and typical of the genus (Fig. 52). Spinnerets also
quite typical of the genus. Color in alcohol: carapace a light yel-
lowish brown with somewhat darker areas corresponding to uneven
distribution of granulations; sternum nearly like carapace; legs
and mouth parts yellowish with some variations; abdomen nearly
white in general, with faint reticulations; scuta light yellowish
brown with variations; just anterior to the base of the anterior spin-
nerets a pair of darker, irregular spots occur as usual in the genus;
there are also many short, black hairs on this part of the body.

Diagnosis. This species seems to be most closely related to
Stenoonops reductiis (Bryant) but it differs considerably from that
species with respect to the shape of the carapace, features of the
sternum, and characteristics of the scuta and epigynal area.

Records. The male is unknown and there is no female paratype.

1969 STENOONOPS SPIDERS 21

Stenoonops reductus (Bryant)
Figures 53-58

Scaphioides reducta Bryant, 1942: 327, pi. 1, figs. 6, 8. The female holo-
type from St. Croix, U. S. Virgin Islands, is in the Museum of Comparative
Zoology. Harvard University, Cambridge, Mass., examined.

Miss Bryant apparently regarded this species as a close relative
of the species placed in the genus Scaphiella Simon and erected a
new genus for it. I am obliged to consider the species as belonging
to the genus Stenoonops Simon and am placing it here on the basis
of the appearance of the male palp, the ventral scuta, eyes,
sternum, and general form of the whole body. A male from St.
Croix, U. S. V. I., has been selected for rather detailed description
as given below.

Male. Total length 1.56 mm exclusive of the somewhat ex-
tended spinnerets; including the spinnerets total length is 1.61 mm.
Carapace 0.64 mm long; 0.5 mm wide opposite second coxae
where it is widest; 0.26 mm tall; gently arched from PME to begin-
ning of steep posterior declivity; surface notably granular which
gives the ventral margin a finely serrated appearance; with no
evidence of a median fovea or groove (Figs. 53-54). Eyes: six as
usual in a compact group; posterior row gently recurved and occu-
pying fully two-thirds of width of carapace at level of posterior
border of PLE. Ratio of eyes ALE : PME : PLE = 12 : 11 : 10.
ALE narrowly separated from one another and separated from
PME and PLE by about one-seventh of their long axis. PME
contiguous for nearly one-third of their circumference and con-
tiguous to PLE for a short distance. Clypeus quite porrect; height
of clypeus nearly equal to long axis of ALE. Chelicerae, maxillae,
and lip essentially as described for the holotype female; with no
special modifications observed. Sternum: convex; only a little
longer than wide just behind second coxae where it is widest; con-
siderably narrowed in front; surface granular with sharply accentu-
ated marginal lobes and grooves; a rounded lobe at posterior end
between bases of fourth coxae, which are separated by about seven-
fifths of their width; third and fourth coxae subglobose; first and
second coxae slightly more elongated. Legs: 41=23 in order of
length; numerous short hairs on all legs but no spines observed.
Palp: femur of moderate length, about twice as long as patella;
both patella and tibia short; tarsus with distinctive features (Figs.
55-56). Considerable variation has been noted among the avail-
able males in respect to length of terminal palpal tarsal structures.

22 BREVIORA No. 339

Abdomen: ovoid in general; 0.88 mm long exclusive of the some-
what extended spinnerets; 0.61 mm wide near middle; epigastric
and narrow ventral scuta are well developed (Fig. 57) and chitin-
ized essentially as described for the female holotype except that
both scuta are closely pressed together. Color in alcohol: carapace
a rich medium brown somewhat lighter along median region and
with a moderate amount of black pigment in ocular area; sternum
nearly Hke carapace; mouth parts and legs yellowish with varia-
tions; abdomen nearly white with scuta clearly delineated and col-
ored nearly like sternum. Epigynal area of female somewhat dis-
tinctive (Fig. 58).

Records. The described male is from St. Croix, U. S. Virgin
Islands, Frederiksted, March 23, 1964. Numerous specimens of
both sexes are in the collection from several localities in the vicinity
of Frederiksted and King's Hill, March, 1964, and September,
1966. Two males and a female were taken on St. John, U. S. Vir-
gin Islands, March, 1964.

Stenoonops noctucus sp. nov.
Figures 59-63

Holotype. The male holotype is from St. Thomas, U. S. Virgin
Islands, taken from hay and weed debris on roadside in outskirts
of Charlotte Amalie, February 9, 1964. The name of the species
is an arbitrary combination of letters.

Description. Total length exclusive of the slightly extended
chelicerae and spinnerets 1.19 mm; including these structures total
length is 1.28 mm. Carapace 0.5 mm long; 0.38 mm wide just
behind second coxae where it is widest; 0.14 mm tall (slightly
lower than usual); nearly level along midline from PME to be-
ginning of moderately steep posterior declivity; surface smooth
and with no evidence of a median thoracic groove or pit. Eyes:
unusual in the genus; six in a compact group; posterior row slightly
procurved and occupying nearly three-fifths of width of carapace
at that level. Ratio of eyes ALE : PME : PLE = nearly 4.5 : 5 : 4.
Oudine of eyes somewhat irregular and difficult to observe espe-
cially with respect to PME (long axes used for measurements).
ALE separated from one another by nearly three-eighths of their
long axis; separated from PME by about three-sixteenths of their
long axis and from PLE by only a line. PME unusually elongated
(Fig. 59) and contiguous for nearly their long axis and contiguous
to PLE for about one-fourth of the long axis of the latter; some
asymmetry of eyes noted. Clypeus very narrow; apparently its

1969 STENOONOPS SPIDERS 23

height is not more than one-fourth of the long axis of ALE. Cheli-
cerae: slender, vertical; of moderate length. Maxillae: slighdy con-
vergent; about twice as long as lip, which is about as long as wide
at base. Sternum: moderately raised and then nearly flat; longer
than wide in ratio of about 7 : 6; widest between second and third
coxae where a lobe extends laterally and bears a cluster of stiff,
curved bristles (Fig. 60); not otherwise conspicuously lobed or
grooved; surface smooth with numerous short hairs or bristles gen-
erally curved inward; bluntly rounded posterior end extends just
to base of fourth coxae, which are separated by about five-fourths
of their width; third coxae globose, fourth subglobose, others some-
what elongated. Legs: 4123 in order of length; with many stiff
hairs and spiniform bristles but no true spines. Palp: femur some-
what elongated; patella and tibia short; tarsus inflated and with
very obscure but quite distinctive features (Figs. 61-63). Abdo-
men: elongate ovoid; spinnerets of moderate length; epigastric and
ventral scuta only weakly indicated; with many short, dark, stiff
hairs. Color in alcohol: carapace and sternum a light yellowish
brown; legs and mouth parts somewhat lighter; abdomen with sev-
eral irregular, purplish spots of different sizes against a nearly white
dorsal background; venter nearly white with a pair of yellowish
spots just anterior to the spinnerets.

Female paratype. Total length 1.34 mm. Carapace 0.55 mm
long; 0.41 mm wide; 0.15 mm tall. Eyes essentially as in male.
Chelicerae, maxillae, and lip also essentially as in male with minor
differences considered unnoteworthy. Palp with many stiff, con-
spicuous bristles. Sternum essentially as in male with the conspic-
uous lateral lobes extended between second and third coxae. Legs
also essentially as in male. Abdomen: with epigastric and ventral
scuta faintly visible; epigynal area indistinguishable. Color in al-
cohol: cephalothorax and appendages as in male; abdomen with
the purplish spots somewhat more clearly shown than in male; in
the anterior third of the dorsum are two pairs of somewhat irregu-
lar purplish spots and in the posterior third of the dorsum there is
a pair of larger purplish spots; on the venter just anterior to the
spinnerets there is a pair of somewhat oval, brownish spots; else-
where the abdomen is a very light yellowish.

Diagnosis. This species appears to exhibit many of the usual
generic characters of Stenoonops but does not seem to be closely
related to any other species in the genus as I have come to know
it. The eyes, features of the sternum, and color pattern all seem
to indicate that this species represents a new group to add to the

24 BREVIORA No. 339

genus. In the past it might very well have been assigned to a new
genus by some taxonomists.

Records. The described female paratype, together with two ad-
ditional females, were all taken from hay and weed litter on the
roadside close to the spot that yielded the male holotype on Feb-
ruary 9, 1964. One female was taken on St. John, U. S. Virgin
Islands, Cruz Bay, March 1, 1964.

Stenoonops padiscus sp. nov.
Figures 64-68

Holotype. The male holotype is from Jamaica, W. I., St. Cath-
erine, three miles east of May Pen, November 22, 1957. The name
of the species is an arbitrary combination of letters.

Description. Total length about 1.25 mm (the holotype is very
fragile and somewhat dismembered) exclusive of the extended
spinnerets and chelicerae; including these structures, total length
is about 1.36 mm. Carapace 0.52 mm long; 0.37 mm wide just
behind second coxae where it is widest; about 0.19 mm tall; ventral
margin somewhat irregular but not definitely serrated as in some
species; continues from PME nearly level to beginning of very
gradually descending posterior declivity; surface very finely granu-
late; hairs along medial region convergent toward midline. Eyes:
six in a compact group; posterior row occupies about six-elevenths
of width of carapace at that level; posterior row somewhat pro-
curved, measured by posterior borders and viewed from above.
Ratio of eyes ALE : PME : PLE = nearly 5:5:4 (long di-
ameters used for measurements). PME much narrowed (Fig. 64);
only about half as wide as long. ALE separated from one another
by about three-tenths of their diameter; barely separated from PLE
and separated from PME by less than one-fifth of their diameter.
PME contiguous for fully one-third of their circumference; barely
separated from PLE. Posterior row of eyes only a little wider than
anterior row. Height of clypeus somewhat less than radius of ALE.
Chelicerae, maxillae, and lip apparently typical of the genus. Ster-
num: widest near middle where a lateral lobe extends between sec-
ond and third coxae; only longer than wide at this point in ratio
of about 6:5; with practically no marginal grooves; with numer-
ous bristles especially along margins; sternal suture at base of lip
nearly straight; posterior end extended between bases of fourth
coxae which are separated by nearly five-fourths of their width.
Legs: 4123 in order of length; no spines observed; trichobothria

1969 STENOONOPS SPIDERS 25

observed but exact number and placement not determined. Palp:
only tarsus with special modifications (Figs. 65-67). Abdomen:
essentially typical of the genus except that the scuta and genital
region are very obscurely outlined and not suitable for illustration.
Color in alcohol: carapace, sternum, legs, and mouth parts a pale
yellowish with variations; abdomen nearly white with variations.

Female paratype. Although no females were taken with either
male assigned to this species, it seems reasonable to assume that
this specimen represents the female of the species because of its
close resemblance to the holotype. Total length 1.43 mm exclusive
of the somewhat extended chelicerae and spinnerets; including
these structures, total length is about 1.54 mm. Carapace about
0.57 mm long; 0.42 mm wide opposite posterior border of second
coxae where it is widest; nearly 0.21 mm tall; slightly arched along
midline from PME to beginning of moderately steep posterior
declivity nearly opposite interval between second and third coxae;
no serrations observed along ventral border. Eyes: essentially as
shown in Figure 68; very similar to those of male; all eyes quite
silvery. Chelicerae, maxillae, and lip essentially as usual in the
genus. Sternum: longer than wide in ratio of about 4:3; widest
opposite interval between second and third coxae where a conspic-
uous lobe extends laterally as in male; moderately lobed but with-
out grooves; with many incurved bristles especially along margins;
sternal suture straight; posterior end extended between fourth
coxae which are separated by about 1.2 times their width. Legs:
essentially as in male. Abdomen: ovoid as usual; generally typical
of the genus except that the scuta and epigynal area are barely
discernible and without distinctive features. Color in alcohol:
essentially as in male except that the abdomen is yellowish white
with a faint darker reticulation more evident on dorsum than on
venter.

Diagnosis. In my judgment, this species is another somewhat
aberrant member of the genus, closely related to Stenoonops noc-
tucus sp. nov. but somewhat remotely related to other species that
I have had an opportunity to study.

Records. The described female paratype is from Jamaica, W. I.,
St. Thomas, Morant Point, Maumee Bay, October 14 ,1957. In
addition to these two specimens, my collection contains the follow-
ing specimens, all from Jamaica, W. I. : one male from St. Andrew
Parish, Newcastle Road, five miles from Kingston, May 10, 1950
(C. C. Hoff); three females from St. Catherine Parish, two miles
east of Ferry, May 17, 1956 (C. C. Hoff); St. Andrew Parish,
Ferry, one mile west of Red Hills Road, Oct. 6, 1957; St. Thomas
Parish, Morant Point, May 6, 1956 (C. C. Hoff).

26 BREVIORA No. 339

BIBLIOGRAPHY

Bonnet, Pierre

1958. Bibliographia Araneorum. Toulouse. Vol. 2(4).
Bryant, Elizabeth

1942. Notes on the spiders of the Virgin Islands. Bull. Mus. Comp.
Zool., 89:317-363.
Chamberlin, R. v., and Wilton Ivie

1935. Miscellaneous new American spiders. Bull. Univ. Utah, 26,
(4): 1-79.

Chickering, Arthur M.

1951. The Oonopidae of Panama. Bull. Mus. Comp. Zool. 106(5):

207-245.
1969. The Family Oonopidae (Araneae) in Florida. Psyche, in press.
COMSTOCK, J. H.

1940. The Spider Book. Revised and edited by W. J. Gertsch. New
York: Doubleday, Doran & Company, Inc.
Petrunkevitch, Alexander

1911. A synonymic index-catalogue of spiders of North, Central,

South America, etc. Bull. Amer. Mus. Nat. Hist., 29: 1-809.
1928. Systema Araneorum. Trans. Connecticut Acad. Arts and Sci.,

29: 1-270.
1939. Catalogue of American Spiders. Pt. 1. Trans. Connecticut
Acad. Arts and Sci., 33: 135-338.

ROEWER, C. FR.

1942. Katalog der Araneae. 1: 1-1040. Bremen.

Simon, E.

1891. On the spiders of the Island of St. Vincent. Pt. 1. Proc. Zool.
Soc. London, Nov. 17, 1891: 549-575.

1892. Histoire naturelle des Araignees. Deuxieme fidition.
1903. 2 vols. Paris: Librarie Encyclopedique de Roret.

(Received 30 September 1969.)

Index to species of Stenoonops

cletus 5 nit ens 9

diinotus \2 noctucus 34

econotus 7 padisciis 37

halatus 30 petrunkevitchi 21

hoffi 22 phonetiis 25

insolitus 14 portoricensis 10

lucradus 16 reductus 32

macabus 19 scabriculus 28
minutus 24

1969

STENOONOPS SPIDERS

27

Figures 1-6. Stenoonops cletiis sp. nov. Fig. 1, Carapace of male; right
lateral view. Fig. 2, Carapace of male from above. Fig. 3, Left palpal tibia
and tarsus; retrolateral view. Fig. 4, Distal end of palpal tarsus; nearly ven-
tral view. Fig. 5, Carapace of female; right lateral view. Fig. 6, Epigastric
and ventral scuta of female from below.

28

BREVIORA

No. 339

Figs. 7-9. Stenoonops econotus sp. nov. Figs. 7-8, Carapace of female
holotype from above and in left lateral view, respectively. Fig. 9, Ab-
dominal scuta and epigynal area of holotype from below. Figs. 10-13.
Stenoonops nitens Bryant. Fig. 10, Carapace of male; right lateral view.
Fig. 11, Eyes of male from above. Fig. 12, Left palp of male; retrolateral
view. Fig. 13, Scuta and epigynal area from below.

1969

STENOONOPS SPIDERS

29

Figs. 14-18. Stenoonops portoricensis Petrunkevitch. Fig. 14, Carapace
of female; right lateral side. Fig. 15, Eyes of female from above. Fig. 16,
Abdomen of female from above. Fig. 17, Epigynal area of female. Fig.
18, Carapace of male holotype; right lateral side.

30

BREVIORA

No. 339

Figs. 19-20. Stenoonops dimotus sp. nov. Fig. 19, Carapace of holotype
from above. Fig. 20, Eyes of holotype from above. Figs. 21-26. Stenoonops
insolitus sp. nov. Fig. 21, Carapace of male holotype right lateral view^.
Fig. 22, Eyes of holotype from above. Fig. 23, Right maxilla from dis-
sected male paratype. Figs. 24-25, Left palp of holotype; retrolateral and
nearly ventral views, respectively. Fig. 26, Distal end of palpal tarsus; more
enlarged and turned to show division. Figs. 27-28. Stenoonops lucradus sp.
nov. Fig. 27, Carapace of holotype male from above. Fig. 28, Left palp
of male holotype; prolateral view.

1969

STENOONOPS SPIDERS

31

Figs. 29-30. Stenoonops macabus sp. nov. Fig. 29, Eyes of male holotype
from above. Fig. 30, Left palp of male holotype; prolateral view. Figs. 31-
33. Stenoonops petninkevitchi Chickering. Figs. 31-32, Carapace of female
from above and right lateral side, respectively. Fig. 33, Carapace of male;
right lateral side.

32

BREVIORA

No. 339

Figs. 34-42. Stenoonops hoffi sp. nov. Figs. 34-35, Carapace of holotype
male from above and right lateral view, respectively. Fig. 36, ALE of
holotype from in front. Fig. 37, Left palp of holotype; retrolateral view.
Figs. 38-39, Left palp of paratype male; prolateral and nearly ventral views,
respectively. Fig. 40, Distal end of left palpal tarsus of paratype male;
nearly dorsal view. Fig. 41, Abdominal scuta of paratype male from below.
Fig. 42, Abdominal scuta of described female paratype from below.

1969

STENOONOPS SPIDERS

33

Figures 43-46. Stenoonops phonetus sp. nov. Figs. 43-44, Carapace of
male holotype from above and from left lateral side, respectively. Fig. 45,
Left palpal tarsus; nearly dorsal view. Fig. 46, Epigastric and ventral scuta
of female from below. Figs. 47-48. Stenoonops scabriculus Simon. Fig.
47, Eyes of holotype from above. Fig. 48, Left male palpal tarsus; retro-
lateral view.

34

BREVIORA

No. 339

Figs. 49-52. Stenoonops halatiis sp. nov. Fig. 49, Body of female holotype
from above. Fig. 50, Carapace of holotype; left lateral view. Fig. 51, Eyes
of holotype from above. Fig. 52, Abdominal scuta and epigynal area of
holotype from below. Figs. 53-58. Stenoonops reductus (Bryant). Fig.
53, Carapace of male from above. Fig. 54, Carapace of male; right lateral
view; lower magnification. Fig. 55, Patella, tibia, and tarsus of left male
palp; prolateral view. Fig. 56, Left palpal tarsus from below. Fig. 57-58,
Epigastric and ventral scuta of male and female, respectively, from below.

1969

STENOONOPS SPIDERS

35

Figs. 5y-63. Slenoonops nuctiaiis sp. nov. Fig. 59, Eyes of male holotype
from above. Fig. 60, Sternum of male holotype from below. Fig. 61. Left
palp of male: prolateral view. Fig. 62, Left palpal tarsus; nearly ventral
view. Fig. 63. Left palpal tarsus; nearly dorsal view. Figs. 64-68. Stc-
nooiiops [uiclisciis sp. nov. Fig. 64, Eyes of male from above. Figs. 65-
66, Left palpal tarsus of male holotype; prolateral and ventral views,
respectively. Fig. 67. Left palp of male paratype; dorso-prolaterai view.
Fig. 68, Carapace of described female paratype.

^>0 '

DO NOT CIRCULA'

B R E V I O »

I.lUSPa^MP. ZOCL
LIBRARY

Meseum of CoiMpsirsitive Zoold/^^'S 1970

Cambridge, Mass. 30 December, 1969 NumberB^O^VARD
DNIVERSlTYj

DISTRIBUTION OF THE CENTROLOPHID FISH

SCHEDOPHILUS PEMARCO, WITH NOTES

ON ITS BIOLOGY'

Richard L. Haedrlch^ arrd Fernando Cervlgon^

Abstract. Schedophiliis pemarco is a non-schooling, somewhat rare fish
that occurs along the West African coast from 19N to 17°S on the bottom
at depths of 50-250 m. In two different years, a solitary specimen has been
taken in the southeast Caribbean; these specimens were probably wafted
from West Africa via the North Equatorial Current. S. pemarco can be
distinguished from other Schedophiliis by the anal count of III 16-19 and
the irregular horizontal stripes on the sides. AUometry is not particularly
marked between 80 and 240 mm SL. Fishes of 200 mm SL could be sexed,
but were immature. Food items may include jellyfish, hyperiid amphipods,
euphausiids, and small fish.

Schedophiliis pemarco (Poll, 1959) is a near-shore fish of the
tropical coast of West Africa. The type material, comprising four
specimens, was from near the mouth of the Congo River {cci. 6°S)
and from near Porto Amboin {ca. 11°S). Blache (1962) reported
the species from the Gulf of Guinea, but gave no particulars as to
the specimens on which the report was based. During the Guinean
Trawling Survey (GTS), at least 14 more specimens were taken
from oiT the mouth of the Kunene River {ca. 17°S) to Sierra Leone
{ca. 8°N). This series and additional specimens from the "Walther
Herwig" give a better impression of the species' distribution, and
provide some information on the allometric growth and food habits
of Schedophilus pemarco.

Although essentially a fish of the eastern tropical Atlantic,

1 Contribution no. 2386 from the Woods Hole Oceanographic Institution.

- Woods Hole Oceanographic Institution, Woods Hole, Mass., and Mu-
seum of Comparative Zoology, Harvard.

3 Estacion de Investigaciones Marinas de Margarita, Nueva Esparta,
Venezuela.

BREVIORA

No. 340

Schedophilus pemarco can range far from this area. This note re-
ports two specimens, caught in different years, from off Venezuela
in the southeastern Caribbean Sea.

Schedophilus pemarco (Fig. 1) is one of the most easily dis-
tinguished members of the genus. The low anal finray count
(111 16-19) is the best diagnostic character. All other Schedophihis
have at least 20 anal finrays, and one has as many as 40. The pat-
tern of thin irregular horizontal stripes on the sides is another
good distinguishing feature. Only S. griseoUneatus (Norman) has
a similar pattern, but this western South Atlantic species has 31-33
dorsal finrays in comparison to the 23-26 found in S. pemarco.

JRfl«05.S
if.

Fig. 1. Schcdopliihis pcnuirco (Poll), an 89-mm specimen from the Carib-
bean, 10°53.5'N 6I°()().6'W. Museum of Comparative Zoology specimen.

Figure 2 shows the extent of allometric growth in Schedophihts
pemarco. It is nowhere particularly marked, and, by the time the
fish has reached a length of approximately 200 mm SL, the rela-
tive proportions are more or less stabilized and no longer change
with growth. This is very generally the case in stromateoids, the
most dramatic changes occurring early. In S. medusophagiis
(Cocco), for example, the greatest changes take place between 10
and 80 mm SL (Haedrich 1967: 47). Clearly, smaller specimens
of 5. pemarco, when found, may be expected to differ morpho-
metrically from the ranges displayed by the larger specimens here
reported.

1969

SCHEDOPHILUS PEMARCO

Q

5

Uj
o

Schedophilus pemarco

n

20

40 60 eo

100

130

140

160

leo

200

:m

'

'

'

'

'

30

■

SNOUT

•

•

•

/ 1^

•
•

•

4

30

'

111,

'

'

'

25

•
EYE DIAMETER

•

••

^

*!•

• •

• •

•

%

45

•

40

*

•

•

•

•

• •
•

•

'

35

LENGTH OF UPPER JAW

45

1

'

1

1

'

40
35

INTERORQtTAL WIOTH

•

•
t

•
•

•
•
•

«

•

•

,

1 t 1

,

—1 —

,

1

^

1

20 40 60 80 100 170 140 160 180 200 220

STANDARD LENGTH in MILLIMETERS

Uj

§

Uj
(0

S
o
Q:
Uj
Q.

45

40
35

10 CO <o ao 100 no uo wo i«o va 120

•

LENGTH OF HEAD ^ ^ * ' * i

35

30

25

- X

• • • • m %

LENGTH OF PECTORAL FIN * "• *

25
20

• • ••

LENGTH OF PELVIC FIN ,* * * -

50
45
40

X

• • •• .

MEOORSAL DISTANCE •*,*•* * f

70
65
60

PREANAL DISTANCE * ♦* ", | • * '

60
55
SO

45

> •

. • • •

MAXIMUM DEPTH , * ^
t 1 1 < t t 1 1 1 1 I

— 1 1 1 1 1 1 1 — ~i r~

DEPTH OF CAUDAL PEDUNCLE

.J I L_

20 40 60 60 100 120 140 160 ISO 200 220

STANDARD LENGTH in MILLIMETERS

Fig. 2. Scatter diagram of the allometric growth in Schedophilus pemarco.
West African, x Caribbean.

The leveling off of the allometric growth curve may correspond
approximately to the onset of maturity. Specimens of 200 mm SL
could be sexed, though the gonads even in these were quite small
and poorly developed. None of the specimens examined were ma-
ture. It is not likely that the failure to find fully developed gonads
is merely a function of time of the year, since specimens in the
material studied came from both spring and fall.

Stomachs were often empty. The most common food item en-
countered was the amorphous, soupy remains of jellyfish. In
two stomachs, there were hyperiid amphipods about 5 mm long.
Euphausiid remains were found on one occasion, and, in the
stomach of a 152-mm fish, there was a 60-mm carangid fish, prob-
ably Trachurus.

All specimens were taken in bottom trawls. Figure 3 shows the
depth of water in which each specimen was taken. Most of the
catches (>60%) occurred in water shallower than 55 m. Plotted

BREVIORA

No. 340

against depth-of-capture in Figure 3 is the size of each specimen.
The plot suggests that as Schedophilus pemarco grows larger, it
seeks deeper water. All the specimens larger than 200 mm SL
were taken at depths of 150 m or greater. This pattern of behavior
is very commonly seen in the stromateoids, although in most cases
the smaller fishes are pelagic, not bottom-living.

to
Q:

to
Uj

5:

CL

s

I

O
I

s

&

Q

0

'

1

•

■

r-

X

•

• •
•

••

. '^

•

100

•

•

•

200

-

•

-

300

-

•

-

400

-

•
• •

•

•

500

-

I

•

-

60

120

160

200

240

280

STANDARD LENGTH in MILLIMETERS

Fig. 3. Depth distribution of Schedophilus pemarco. Each point repre-
sents one specimen. All data from bottom trawls; when a depth range is
indicated, only the mid-point is plotted. • West African, x Caribbean.

The bottom trawls used were of various sizes. The smallest was
about 10 m across, the largest about 45 m. Those used by GTS
were roughly 25 m. Despite these large trawls, most catches
(>80%) were of single fish. The greatest number from any one
trawl was three. Schedophilus pemarco is thus probably a non-
schooling, somewhat rare, fish. The species was, for example, not
abundant enough to enter into the analysis of species assemblages
based on the 480 trawls of the GTS survey (Fager and Longhurst,
1968).

1969

SCHEDOPHILUS PEMARCO

Figure 4 shows the horizontal distribution of Schedophilus
pemarco. One of the northernmost records, at about 19°N, is
from Cervigon (1960:69,74), where the fish was referred to as
Miipits sp. aff. M. ovalis. The records along the African coast fall
effectively within the 75 °F isothere of Hutchins and Scharff
(1947). Outside this region, the surface temperature is always
colder than 75 °F (24°C). Seasonal cooling, of course, reduces
the surface temperature somewhat on each end of this distribution,
but, for the most part, the temperature is always warmer than
18°C (Hutchins and Scharff, 1947). The 200-m temperature
across this region, more or less stable the year around, ranges
from about 10° to 15°C (Sverdrup et al, 1942).

The two records of Schedophilus pemarco from the Caribbean,
reported here for the first time, are of some interest. There can be
no doubt as to the identification. The two fish almost certainly
originated in Africa, and it is not hard to imagine the pathway that
could be used, although the time needed seems somewhat long.
Unpublished drift bottle data on file at the Woods Hole Oceano-
graphic Institution suggests that the journey from the Mauretanian
region to northeastern South America or the Lesser Antilles via
the North Equatorial Current would take something of the order
of six months. The Caribbean specimen illustrated in Figure 1 is
one of the smallest S. pemarco known. Smaller specimens, as in

30°

20°

Fig 4. Geographical distribution of Schedophilus pemarco. The arrows
indicate the general path of the North Equatorial Current.

6 BREVIORA No. 340

other Schedophilus, may be pelagic and may associate with float-
ing objects. If this is the case, as is Hkely, these fish would be
prime candidates for transport.

Schedophilus pemarco does not seem to have established popu-
lations in the New World. It was not encountered, for example, in
the extensive trawl surveys conducted during 1957-1959 on the
continental shelf of British Guiana (Lowe, 1962), and continuous
surveys both off the Guianas and in the southeastern Caribbean by
Venezuela during 1962-1968 have taken only the specimens re-
corded here. Nonetheless, the taking of two specimens in each of
two different years indicates that immigration from West Africa
may occur with some regularity.

Transport and establishment are, of course, two quite different
things. In comparison to the West African shore fish fauna, the
fauna of the New World tropics is very rich indeed, and, in such
a situation, one would expect the successful colonists to move from
west to east. The data support this expectation. Briggs (1967)
finds that no essentially West African fishes have been successful
in crossing the Atlantic, but that over 100 species have established
themselves in the opposite direction. Movement from west to east
need not be an upstream movement, for there are zoogeographi-
cally important countercurrents in the equatorial region (Schel-
tema, 1968).

Description. Schedophilus pemarco is a deep-bodied, rather
firm-fleshed fish. The profile of the head slopes steeply to the blunt
snout. The head and nape lack scales. The preopercle is beset
with 12-19 prominent spines. The large eye is centrally located,
roughly one-half its diameter or more below the dorsal profile.
The large mouth is inclined somewhat downward; the end of the
maxilla extends to under the center of the eye or beyond. The
pectoral fin is broad and rounded. The pelvic fins, inserted directly
below the pectorals, are long and often reach beyond the anus.
The color in preservative is grayish brown, with blue-gray-tinged
horizontal lines running all along the sides. The fins are gray or
blackish, darker than the body. The pelvics are particularly black.
The head is uniformly dark with very dark, thin opercles, and the
eye is bluish. In the summary of meristics which follows, the two
numbers in parentheses are the figures for the two Caribbean
specimens 88 and 162 mm SL respectively: in % SL — head 33-
41 (41, 33), pectoral length 24-31 (30, 24), pelvic length 18-27
(24, 20), predorsal distance 38-47 (47, 38), preanal distance
57-66 (63, 60), maximum depth 43-55 (55, 44), least depth
caudal peduncle 10-13 (12, 11); in % of head — snout 25-30

1969

SCHEDOPHILUS PEMARCO

(26, 28), eye diameter 21-29 (29, 23), length upper jaw 39-43
(40, 39), interorbital width 32-42 (33, 35); counts — dorsal V-
VII 23-26 (VI 24, VI 24), anal III 16-19 (III 16, III 17),
pectoral 19-22 (21, 20), gill-rakers 5-7 + 1 + 13-16, most
commonly 7 + 1 + 15 (7 + 1 -f- 15, 7 + 1 + 15), vertebrae
always 10 -|- 15. Count frequencies for fin elements are given in
Table 1.

Table 1. Count frequencies of fin elements in Schedophilus pemarco.

Anal spines are III.

Dorsal spines

Number of specimens

Dorsal finrays

Number of specimens

Anal finrays

Number of specimens

Pectoral finrays

Number of specimens

V

VI

VII

5

11

2

23

24

25

26

3

12

3

1

16

17

18

19

1

11

7

1

19

20

21

22

11

specimens used in this study, unless otherwise noted, are de-
posited in the Museum of Comparative Zoology, Harvard. Dis-
position of the others is in the Institut fiir Seefischerei, Hamburg
(ISH); Museu de Historia Natural La Salle, Caracas (MHNC);
or Tropical Atlantic Biological Laboratory, Bureau of Commercial
Fisheries, Miami (TABL). Material examined: GTS — 1 spec.
90 mm SL, Geronimo 2-246, 9 Sept 63, 4°31'S 10°53'E, 300 m
(TABL); 1 spec. 153 mm SL, Thierry 1-2, 21 Sept 63, 5°06'N
1°05'W, 20 m; 1 spec. 152 mm SL, La Rafale 18-4, 30 Oct 63,
4°35'N 8°25.5'W, 50 m; 1 spec. 189 mm SL, La Rafale 16-8,
4 Nov 63, 5°17'N 9°55'W, 400 m; 1 spec. 139 mm SL, La Rafale
15-2, 5 Nov 63, 5°55.5'N 10°12'W, 30 m; 1 spec. 119 mm SL,
La Rafale 11-3, 15 Nov 63, 7°18.5'N 12°4rW, 40 m; 3 spec.
188-217 mm SL, La Rafale 11-8, 16 Nov 63, 7°12'N 12°46'W,
350-500 m; 3 spec. 123-165 mm SL, La Rafale 10-3, 17 Nov 63,
7°32'N 13°2rW, 40 m; 1 spec. 103 mm SL, La Rafale 9-4, 19
Nov 63, 7°53'N 13°58'W, 50 m; 1 spec. 119 mm SL, Thierry 5,
31 Mar 64, 5°13'N 4°56'E, 70 m. Other West African — 3 spec.
237-242 mm SL, Walther Herwig 143/164, 20°3rN 17°42'W,
200 m (ISH); 1 spec. 180 mm SL, Undaunted 68-262, 24 Mar 68,
17°02'S 11°40'E, 55 m, 40' trawl (TABL). Caribbean — 1 spec.
88 mm SL, 31 July 67, 10°53.5'N 61°00.6'W, 32-38 m; 1 spec.
162 mm SL, 7 Apr 68, 10°51'N 64°irW, 33 m (MHNC).

8 BREVIORA No. 340

ACKNOWLEDGMENTS

The bulk of the specimens used in this study were made avail-
able by William J. Richards, then at the Bureau of Commercial
Fisheries Biological Laboratory, Washington; Bruce B. Collette,
Bureau of Commercial Fisheries Ichthyological Laboratory, Wash-
ington; and Leslie W. Knapp, Smithsonian Oceanographic Sorting
Center, Washington. Gerhard Krefft, Institut fiir Seefischerei,
Hamburg, provided the "Walther Herwig" specimens. C. Godfrey
Day provided unpublished data on drift bottles in the tropical
Atlantic. Michael H. Horn and James E. Craddock have kindly
read the manuscript. Financial support came from National Sci-
ence Foundation grant GB-7108 to the Woods Hole Oceano-
graphic Institution.

LITERATURE CITED

Blache. J.

1962. Liste des poissons signales dans I'Atlantique Tropico-Oriental
Sud - du Cap des Palmes (4° Lat. N) a Mossamedes (15°
Lat. S) (Province Guineo-Equatoriale). Trav. Centre Ocean-
ogr. de Point-Noire, O.R.S.T.O.M. 1962. Pp. 13-106.
Briggs. John C.

1967. Relationship of the tropical shelf regions. Stud. Trop. Ocean-
ogr. Miami, 5: 569-578.

Cervigon, F.

1960. Peces recogidos en el curso de las companas realizadas a bordo
del "Costa Canaria" desde cabo Bojador a Guinea Portuguesa
(Africa occidental) y consideraciones sobre su distribucion.
Inv. Pesq. XVII: 33-107.

Fager, E. W., and a. R. Longhurst

1968. Recurrent group analysis of species assemblages of demersal
fish in the Gulf of Guinea. J. Fish. Res. Bd., Canada, 25 (7):
1405-1421.

Haedrich, Richard L.

1967. The stromateoid fishes: systematics and a classification. Bull.
Mus. Comp. Zool., Harvard, 135 (2): 31-139.

HUTCHINS, L. W., AND M. SCHARFF

1947. Maximum and minimum monthly mean sea surface tempera-
tures charted from the "World Atlas of Sea Surface Tempera-
tures". Sears Found., J. Mar. Res. 6 (3): 264-268.
Lowe, Rosemary H.

1962. The fishes of the British Guiana continental shelf, Atlantic coast
of South America, with notes on their natural history. J. Linn.
Soc. London, Zoology, XLIV (301): 669-700.

1969 SCHEDOPHILUS PEMARCO 9

Poll, M.

1959. Poissons V - Teleostcens acanthopterygiens (deuxicme partie).
Res. Sci. Exped. Oceanogr. Beige Cote Afr. Atlantic Sud, Vol.
4, 3B: 1-417.

SCHELTEMA, RUDOLF S.

1968. Dispersal of larvae by equatorial ocean currents and its impor-
tance to the zoogeography of shoal-water tropical species.
Nature (London)r217 (5134): 1159-1162.
SvERDRUP, H. U., M. W. Johnson, and R. H. Fleming

1942. The Oceans. Englewood Cliffs, N.J.: Prentice-Hall, Inc. 1082
pp.

(Received 1 October 1969).

DO NO I ^-irtv^^-;""-

V

'Nil. Of-, I / , MUS. COMP. ZOOL

^mbr,^Jq^7 LIBRARY

B R E V I O R91A

HARVARD

Miiaseiuiimi of Cojmpsirative u^YRj^hrjt'v

Cambkidce, Mass. 15 Jam ak\, 1970 Numhkk 341

FOUR NEW SPECIES OF ELEPHANTIDAE FROM THE
PLIO-PLEISTOCENE OF NORTHWESTERN KENYA

Vincent J. Maglio

Abstract. Four new species and one new genus of fossil elephants are
described from the Plio-Pleistocene of northwestern Kenya. Stci^otctra-
Ih'IocIoii orhus sp. nov. is more progressive and elephantinelike than its
north African relative, S. syrticiis, hut it could not have been ancestral to
the earliest Elephantinae. Stciiotetrahclodon is placed in the family Ele-
phantidae as a distinct subfamily, the Stegotetrabelodontinae. Prinielepluis
ooinphotlicraidcs gen. et sp. nov. is described as the earliest known genus
of the Elephantinae; it is morphologically and stratigraphically suitable
as the basal genus from which later elephants could have been derived.
Loxodonta adiiitrora sp. nov. is distinguishable from Mamiuiithiis afri-
canaviis. in skull characters primarily, but also in dentition. It is con-
sidered to be the earliest known member of the Loxodonta lineage and
may have been directly ancestral to the living L. africana. Elepluis ekoren-
sis sp. nov. is clearly related to, but is more primitive than, E. recki in
both molar and skull morphology and is placed at the base of that lineage.

INTRODUCTION

In 1965. while working in Miocene deposits in Turkana District,
northwestern Kenya, an expedition of the Museum of Comparative
Zoology, under the direction of Professor Bryan Patterson, discov-
ered fossil-bearing sediments of Plio-Pleistocene age well exposed
in the drainage of the Kanapoi, a dry wash tributary to the Kakurio
river (Patterson, 1966). The locality is at 36" 04' E and 2' 19' N,
38 miles west and 9 miles south of Teleki's volcano at the southern
end of Lake Rudolf (Fig. 1 ). Clastic sediments predominate with
pyroclastics interbedded throughout the sequence. The total meas-
ured thickness is over 200 feet but the vertical distribution of fos-
sils indicates no major faunal change, though some new elements
appear near the top of the section. Potassium/Argon age deter-
minations of 2.9 ± 0.3 (Patterson, 1966), 2.7 ± 0.2 (Patterson
ei ill., in MS), and 2.5 :t; 0.2 million years (Patterson and
Howells, 1967) from an overlying basalt are questionable on

2 BREVIORA No. 341

faunal evidence. Correlation based on Proboscidea (Maglio. in
press) and Suidae (H.B.S. Cooke, pers. conim.) indicate near con-
temporanicty with Yellow Sands (Mursi Formation) at the base of
the Omo sequence, for which the K/Ar date of 4.05 ± 0.2 million
years is available (Howell, 1968).

The Kanapoi fauna is rich in vertebrate remains that include:
cf. Australopithecus (Patterson and Howells, 1967); Parapapio
jonesi (Patterson, 1968); Lepus sp.; Hystrix sp.; Tatera sp.; En-
hydriodon sp. nov. (Kurten, pers. comm.); Hyaena namaquensis;
Deinotherium hozasi; Anancus sp.; Loxodonta adaurora sp. nov.
(this paper); Elephas ekorensis sp. nov. (this paper); Stylohip-
parion sp.; Ceratotheriuni sp.; Nyanzachocrus, (2) spp. nov.;
Notochoerus cf. capensis; N. cf. A', eiillus (suid identifications by
H.B.S. Cooke); Hippopotamidae, sp. nov. (S. Coryndon, pers.
comm.); Okapia sp.; Giraffa sp.; Trui^elaphus sp.; other Bovidae;
Crocodylus sp.; Euthecodon sp.; Podocneniis sp.; Testudo sp. cf.
T. aminon; Trionichydae indet., (Chelonia identifications by R. C.
Wood) ; numerous remains of fish and molluscs.

Overlying the Kanapoi basalt is a small area of poorly exposed
sediment at Ekora from which some vertebrate remains have been
recovered. The locality is at 36' 1 1' E and 2" 31' N, 17 miles
NNE of Kanapoi. The Proboscidea are the best preserved and
most numerous fossils from these sediments. The tentative faunal
list includes: Anancus sp.; Loxodonta adaurora sp. nov.; Elephas
ekorensis sp. nov.; Simopithccus sp. (Patterson, pers. comm.); and
Ceratotherium sp.

While exploring for further outcrops of the Kanapoi beds in
1967, a museum expedition worked a richly fossiliferous, thick
sequence of sediments at Lothagam Hill (Patterson et al., in MS).
The exposures cover about three square miles and are situated at
36° 04' E and 2° 53' N, some 40 miles north of Kanapoi and three
miles southwest of the Kerio delta on the southwestern side of
Lake Rudolf. Two fossiliferous levels of coarser, red, fluviatile or
deltaic sediments are separated by fine-grained lake beds probably
representing a significant interval of time. On faunal evidence the
upper fiuviatile beds correlate with those of Kanapoi, but the lower
beds, of which up to 1500 feet are exposed, appear to be consider-
ably earlier. Though a K/Ar date of 3.71 ± 0.23 million years
has been obtained for a basaltic sill between the lower fiuviatile and
the fine upper lake beds (Patterson, pers. comm.). this may repre-
sent no more than an upper limit for the age of the entire sedi-
mentary sequence in the Lothagam area. Ihe actual age of the

1970

NEW PLIO-PLEISTOCENE ELEPHANTIDAE

lower fluviatile beds would appear to be closer to 5.0-5.5 million
years on faunal grounds (Maglio, in press). The fauna has not
yet been studied in detail, but it includes: Australopithecus sp. cf.
A. ajrkanus (Patterson and Howells, in preparation); Macaca sp.;
Anomaluridae, nov.; Hyaenidae indet.; Felinae indet.; Machairo-
dontinae indet.; ?Civettictis sp.; Orycteropididae, gen. et sp. nov.

Figure 1. Generalized map of the southern Lake Rudolf area, north-
western Kenya, showing locations of the Kanapoi, Ekora and Lothagam
deposits.

4 BREVIORA No. 341

(Patterson, pers. comm.); Deinotheriiim sp.; Gomphotheriidae
indet.; Stegotetrahelodon orbus sp. nov. (this papsr); Primelephas
go)nphotlieroides gen. et sp. nov. (this paper); Stylohipparion sp.;
Rhinocerotidae, Brachypotherinae, nov.; Nyanzachoerus, (3) spp.
nov. (Cooke, pers. comm.); Hippopotamidae, (2) spp. nov.
(Coryndon, pers. comm.); Giraffa sp.; Tragelaphus cf. T. nakuae;
Gazella sp.; Redunca aflf. ancystrocera; other Bovidae (bovid iden-
tifications by A. W. Gentry); Crocodylus sp.; Euthecodon sp.;
Podocnemis, (2) spp. nov. (R. C. Wood, pers. comm.); Testiido
sp. cf. T. amnion; Trionychidae indet.; Squamata indet.; Pythoninae
indet.; fish and mollusc remains.

Lothagam, Kanapoi, and Ekora have yielded numerous remains
of Proboscidea of the families Gomphotheriidae and Elephantidae.
Four new taxa of the latter group are represented in the collection
and a detailed description of them is in preparation. Because of the
importance of elephants for purposes of relative correlation, and
because of the numerous, more fragmentary elephant remains turn-
ing up at other African localities, it seems desirable to present at
this time a preliminary diagnosis of these new taxa to serve as a
basis for comparison of less complete material. All specimens are
at present in the Museum of Comparative Zoology but are the
property of the National Museum, Nairobi. Specimen numbers
given are catalogued in the National Museum, Centre for Pre-
history and Palaeontology, Nairobi.

ELEPHANTIDAE

STEGOTETRABELODONTINAE *

Stegotetrahelodon Petrocchi 1941

Stegotetrahelodon, Petrocchi 1941:110.
Stegolophodon, Petrocchi, 1943:123.

Type Species. Stegotetrahelodon syrticus Petrocchi, 1941.

Emended Diagnosis. M3 with six to seven transverse plates
superficially divided by vertical grooves; strong median cleft not
extending to the base of the crown; complete enamel loops forming
with moderate wear; prominent isolated accessory columns present

* Aguirre (1969) suggested that the genus Stegotetrahelodon ". . . may
be a subfamily, Stegotetrabelodontinae . . v ," presumably of the Gom-
photheriidae, though this is vague. However, no formal action to estab-
lish this subfamily was taken at that time. In this paper (and in press)
I include this group as a subfamily of the Elephantidae.

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 5

behind the plates; crown height less than the width; enamel 4-7 mm
in thickness, not folded; transverse valleys between plates open but
V-shaped in longitudinal section; plates strongly triangular in
longitudinal s:ction; lamellar frequency, 2.5-3. Mandible with
long, massive symphysis bearing very long incisors.

Distribution. Late Pliocene, Sahabi beds, Libya; lower fluviatile
beds, Lothagam Hill, Kenya.

Stegotetrabelodon orbits'^' sp. nov.
(Pis. I-II)

Type. PAL. LOTH. 354, complete left lower jaw with unworn
left M.,, worn left M.,, and left and right lower incisor in situ, in-
complete left and right M-\ complete right M- and skeletal frag-
ments.

Horizon and Locality. Late Pliocene, lower fluviatile beds,
Lothagam Hill, Kenya.

Hypodigm. Type and PAL. LOTH. 359, complete left and right
M.,, left M'' and right M- from a single individual. PAL. LOTH.
347, worn left M^^ PAL.LOTH. 349, complete right M.,. PAL.
LOTH. 352, complete left M,. PAL.LOTH. 355, partial'left and
right M.,. PAL.LOTH. 360,' worn left M-^ PAL.LOTH. 367,
complete right M^ PAL.LOTH. 374, partial M'^ and skeletal
fragments. PAL.LOTH. 342, complete left M., and partial right
M^,. PAL.LOTH. 350, complete left and right M... PAL. LOTH.
36'6, complete kft M-. PAL.LOTH. 344^ partial left dM^ and
dM^. PAL.LOTH. 365, complete right dM.,.

Referred Material. PAL.LOTH. 368, left and right mandibular
incisor and right premaxillary incisor. PAL.LOTH. 369, right
humerus. PAL.LOTH. 370, right femur, left humerus, left and
right fibulae and partial pelvis.

Diagnosis. About 12 per cent smaller than S. syrticus. Man-
dibular incisors short, forming about 38 per cent of the total jaw
length. Free columns persisting only behind the first two plates on
M3; relative crown height of M3 13 per cent greater than in S.
syrticus. Lamellar formula:

M3 6X^ M2 5X, dM4 6^ dM3 ?_
TX~ 6X ? 3

L. orfez/5-=childless. In allusion to the terminal position of this species
and its apparent non-ancestry to later species of Elephantidae.

6 BREVIORA No. 341

Discussion. In a series of papers, Petrocchi (1941, 1943, 1952,
1953) described a genus of Proboscidea, Stegotetrabelodon, from
the late Pliocene beds of Sahabi, Libya. There can be little doubt
as to the validity of this genus, but its significance as a morpholog-
ical intermediate between the Gomphotheriidae and the Ele-
phantidae has only recently been stressed (Aguirre, 1969; Magho,
in press). Petrocchi described two species of Stegotetrabelodon,
S. syrticus and S. "lybicus," and one of Stegolophodon, S. "saha-
bianus," from the Sahabi beds. The latter two species were each
based on a single molar. The type of S. "lybicus" is an unworn
lower M3; comparison with the type mandible of 5. syrticus is
difficult owing to excessive wear in the latter. Nevertheless, the
only differences between them are those expected to occur with
wear. On present evidence the two must be considered conspecific.
The type of Stegolophodon "sahabianus" is an incomplete M''.
Undoubtedly Petrocchi was struck by the progressive aspect of
this specimen as compared with the lower molar of S. "lybicus"
(the upper molars on the type skull of S. syrticus are severely
worn). However, associated upper and lower molars of the new
species from Lothagam described here show that the upper M3 was
more progressive in structure than the lower, a common feature in
Proboscidea generally. In size, the type molar of S. "sahabianus"
is larger than the few known specimens of Stegotetrabelodon from
Sahabi, but well within the expected range of size variation as ex-
hibited for molars of the Lothagam species. Without more sub-
stantial evidence to the contrary, Stegolophodon "sahabianus"
should also be considered as a synonym of Stegotetrabelodon
syrticus.

The mandibular tusks of S. syrticus are about two-thirds as long
as the premaxillary tusks. They are very long and slender, and
form 57 per cent of the total jaw length in contrast to only 38 per
cent for S. orbus. The type mandible of the latter is a young adult,
however, and adult size of the mandibular incisors may have been
proportionately somewhat greater. These tusks in S. syrticus are
elliptical in cross-section, being compressed laterally. The lower
tusks of S. orbus are also elliptical as are the small lower tusks of
the dwarfed species, Elephas celebensis (Hooijer, 1954). The
mandibular tusks of S. syrticus are nearly three times longer than
in S. orbus, though similar in diameter.

The third molars are large and massive in 5. syrticus; they are
proportionately broader than in the Lothagam species (Table 1),
but the crown height is about the same, resulting in a lower H/W

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 7

index (100 X height width). A free column is present behind
each molar plate in the Sahabi species, but in S. orbiis all but the
anterior two columns have become incorporated into the plates and
are present in the form of small vertical ridges fused into the pos-
terior faces of the plates. These may be absent on the more
posterior plates. On molars of 5. syrticus, a deep cleft divides the
plates longitudinally; weaker, but still strong lateral clefts are
present so that the apex of each plate is divided into four rather
widely separated digitations that fuse with wear. In S. orbus, the
lateral clefts are very weak and shallow, and all clefts are tightly
compressed. Some of the apical digitations tend to be further sub-
divided into two or more smaller units, each separated by a weak
cleft. The crown apex may thus have up to six or seven subdivi-
sions. Upper molars of S. orbus are more progressive than the
lowers in that the plates are more consolidated, with a weaker
median cleft. There are seven plates on the lower and six on the
upper M3 in both the Sahabi and the Lothagam species.

The second and first true molars of S. orbus are somewhat more
progressive in appearance than M3. The plates are more consoli-
dated, and apical digitations are separated only by weak and com-
pressed clefts. The deeper median cleft perisits on the anterior
two or three plates until intermediate stages of wear, but a com-
plete enamel figure is formed on the remaining plates in early wear.

The enamel is very thick and smooth. Cement is abundant, but
it usually does not fill the valleys completely nor does it usually
invest the apex and sides of the molar plates.

The mandible is long anteroposteriorly in both species, but in S.
syrticus the symphysis is 70 per cent longer than in S. orbus, though
there is no difference in symphyseal width. Proximal tusk separa-
tion is nearly twice as great in S. syrticus as in the Lothagam
species.

The systematic position of Stegotetrabelodon is clearly transi-
tional between the Gomphotheriidae and the Elephantidae. Yet in
spite of certain conspicuous gomphothere characters, such as long
mandibular incisors and a prominent median cleft on M3, the
molars are in general more elephant-like than gomphothere-like.
In this genus, the inner and outer cones of the ancestral gompho-
there molar are obliterated by median compression and fusion to
form very platelike structures. The gomphothere trefoil pattern is
gone and the old median folds that formed them persist as the free
columns in the transverse valleys. Median swellings toward the
base of the plates are all that remain of the anterior trefoil, whereas

8 BREVIORA No. 341

the posterior one first became a prominent isolated column, and in
the more advanced Elephantinae was fused into the plate face.
These fused columns are responsible for the development of median
"sinuses" on the wear figures of many elephant molars. Stegotetra-
belodon is best placed in the family Elephantidae as a primitive
subfamily.

With respect to reduction of mandibular incisors, increased rela-
tive crown height, fusion of isolated columns into the molar plates,
and reduced division of the crown through compression and shal-
lowing of clefts between the apical digitations, 5. orbus sp. nov.
would appear to represent a more progressive species than S.
syrticus. Though very close in many particulars to the earliest
Elephantinae, 5. orbus was probably not ancestral to any known
species of elephant, and occurs in the same deposits as the earliest
species of Elephantinae (see below).

ELEPHANTINAE
Primelephas\ gen. nov.

Type Species. Primelephas gomphotheroides sp. nov.

Diagnosis. Molars very low crowned, the height being one-
half to three-fourths of the width; the median cleft lacking, but
strong grooves superficially dividing the plates into prominent col-
umns; plates wear as complete enamel loops; plates triangular in
longitudinal section, wider toward the root; transverse valleys open
to the base, but strongly V-shaped in cross-section; enamel 3-6 mm
in thickness, not folded; lamellar frequency, 3-4. Cement abun-
dant but usually not filling valleys completely. Mandible with very
small but prominent incisors in some individuals, possibly males
only; symphysis short. Lamellar formula:

M3 IX, M2 5X, Ml ? .
SX "6~ "5^

Distribution. Late Pliocene, lower fluviatile beds, Lothagam
Hill, Kenya; Kolinga and ?Koula, Tchad; Nyawiega, Kaiso beds,
Uganda.

Included Species. Stegodon korotorensis Coppens, 1965.

Discussion. This genus is distinguished from the more primitive
Stegotetrabelodon in having much reduced mandibular incisors and
a short symphysis. The molars are elephantine in structure, lack-
ing median clefts and having stronger plate development. In later

tL. primus = first, and L. elephas = elephant.

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE

S

o

^-*

)>.

?^

en

C

O

ts

O

•>.»

«)

-Q

Q

V.

■•-.

^

■*^

O

>

00

o

^-.

c

S

^

d

s

on

_c

cn

s

*v

-C)

c/5

o

c

'o

o

c

"^

' ^^

o

1-1

*^^

cd

■iJ

,

-Cj

3

>

^

jn

O

;i^

c

■*-*

"5

S

a

LLI '4J

(U

4=;

O

1- o

C/2

c

cd

C

^

<U

en

■^

i^

(D

^

Ui

S-(

-•^

3

O

o

CiO

t*-i

Cm

c3

s

to

111

to

(U

^

03

CO

CLt3

g

c

o

C3

U

i^ ^

(50 w

^ -^

Q! (N (N

(Nj r-.' en

I I I

O m 00

CN uS <>i

CO
CM

q

ci

'-'

CNl

o

i^

X

T

00

1

1

vO

00

q

vO

o

rn

^O

o

r-

-o q o 00 in

CN ^ CN)' CN o

vO 1^ r~- 00 (N

00 vO

r-

q

CN

tn

1

o

1

X

1

o

t
in

1

t~~

q

d

in

CN

in

00

T— 1

CN

T— «

^ o
r- so

c\

in

1

q
rn

X

j

in

1

in

1

o

'^

o

^

00

00

T— <

^

CM

en

CN]

<N

q

SO

in

CN

en
in

O

en
in

so

oo'

cn

d

00

(N

1

so
1

1

X

cn

o

1

00

1

00

I

1

en

1

q

r--

r~-

1

o

1

CN

en

in

CN

oo'

in

o

u

CO

O

c

«1

(U

c

^

o

J=

■4-*

^^

x:

<u

W)

s

c

rt

<u

C

u

ttJ

" ■£ >< r. 2

c *-• Vf, (u <" t- ^3

'^ ^ r- r- ^ '««=-r;S"'

^ o E E ■ - -g. S S S .1

c3

J Z S S I

10 BREVIORA No. 341

genera of Elephantinae, the mandibular incisors disappear external-
ly though they were probably present as tooth buds in symphyseal
crypts in the early species of Loxodonta and Elephas (see below).
In one case, mandibular incisors reappeared, possibly through pae-
domorphosis, in the dwarfed species Elephas celebensis Hooijer.
The crown height is lower than in the most primitive species of
later genera, and the number of plates is lower also. This genus
appears structurally and chronologically to have been a member of
the basal group from which later Elephantinae were derived.

In later genera the crown height is greater and the plates are not
strongly triangular in longitudinal section; the anterior and pos-
terior faces of each plate are nearly parallel. The transverse valleys
between plates are broadly open and U-shaped, and the enamel is
considerably thmner. In all but the most primitive known species
of Marmniithiis, the number of plates for each molar is greater than
in Primelephas.

Primelephas gomphotheroides* sp. nov.
(Pis. III-IV)

Type. PAL.LOTH. 351, associated left and right M'\ left
M., and fragmentary palate.

Horizon and Locality. Late Pliocene, lower fluviatile beds,
Lothagam Hill, Kenya.

Hypodigm. Type and PAL.LOTH. 358, left and right M-, right
Mo, left and right M,, and mandibular symphysis. PAL.LOTH.
363, partial right M-' and mandibular symphysis. PAL.LOTH.
375, incomplete right Mj.

Referred Material. PAL.LOTH. 364, fragmentary M,. PAL.
LOTH. 371, molar fragments, calcaneum, astragalus, incomplete
fibula and an occipital condyle. PAL.LOTH. 376, molar plate.

Diagnosis. Lower crowned than P. korotorensis (Coppens),
crown height for M3 60-65 per cent of crown width at base. Other
characters as for the genus.

Discussion. Until recently, the most primitive known species
of the Elephantinae was Mammuthus siibplanifrons, a poorly
known species from the older gravels of the Vaal River of South
Africa, Kanam, and the Chemeron beds in east Africa, and the
Chiwondo beds in Malawi (see Maglio, in press). Available data
suggest that this species occurred in Africa earlier than the first

* gomphotheroides = gomphothere-like.

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 11

appearances of elephants in Europe and Asia (M. meridionalis and
E. planijrons respectively), and it has been considered close to the
ancestral stock of all later elephants. Recent study (Maglio, in
press, and in preparation) has shown that sub planijrons is on the
Mamnuithus lineage and therefore did not occupy the more broadly
ancestral position once allocated to it.

The stratigraphic position of Primelephas gomphotheroides is
lower in the Pliocene section of Africa than is M. siibplcmijrons, or
in fact any other known member of this subfamily. Morphologi-
cally, it is generalized enough to have served as the basis for later
specializations that were to characterize the three later genera of
elephants (Elephas, Loxodonta, and Mammuthiis) . The molars
of P. gomphotheroides have a strong vertical ridge near the center
of the plate faces, except on the first or second plates where the
vertical ridges are free and columnar at their apices. With wear,
these ridges form the median loops or sinuses so typical of early
elephants. The number of plates is fewer and their relative height
lower than in any other species of Elephantinae (Table 2). The
enamel is thick and unfolded as in other early species. Transverse
valleys between the molar plates are V-shaped in longitudinal sec-
tion, but not compressed at the base as in the stegodonts. In all
later species, including M. siibpkmifrons, the valleys are broadly
open and U-shaped at the base; this serves to maintain the efficiency
of the tooth as a grinding surface into late stages of wear by keep-
ing the enamel ridges separated by a cement interval down to the
base of the crown.

The symphysis of the jaw is much shorter than in Stegotetrabelo-
don, even though small incisors are present. As a result, the root
of the incisor passes into the horizontal ramus. These incisors are
no longer closely appressed in the midline of the symphysis as in
Stegotetrabelodon , but lie on either side of the "spout" formed by
the downturned symphysis. They extend along the ventral wall of
the horizontal ramus anterior and somewhat beneath the front por-
tion of the mandibular canal. There is some indication of variation
in size for these incisors; in two very fragmentary mandibles there
apparently were no external tusks but only a pair of deep cavities
lateral to the symphysis and connected to the mandibular canal by
a short channel. These two specimens may represent juvenile in-
dividuals, but it is also possible that incisor size or even their pres-
ence or absence was a sexual character, being larger in the males.

Specimens referable or probably referable to Primelephas are
known from other African localities. Several molar fragments

12 BREVIORA No. 341

from Nyawiega in the Kaiso sequence indicate the presence of a
closely related form. Although tentatively referring this material
to M. siibplcmijrons (Cooke and Coryndon, in press), H. B. S.
Cooke independently recognized the primitive nature and generic
distinctness of some of the specimens, but was unable to diagnose
this new form without better material. All of the Kaiso material
referred by Cooke and Coryndon to M. subplanijrons is the same
and should be referred to Primelephas. Although additional data
may show this form to be specifically distinct, on present evidence
it is best referred to Primelphas gomphotheroides.

A small fragment was described by Coppens (1965) from Tchad
as Stegodon korotorensis sp. nov. This specimen is very incom-
plete, consisting of the posterior three plates of a lower M3, but it
is clearly not a true Stegodon. The plates are divided apically into
only a few well-separated digitations, and, though V-shaped in
longitudinal section, the transverse valleys are not compressed at
the base as they are in stegodonts. The only other known speci-
men, an equally incomplete upper M3, is similar to the type. This
form is very close to the Lothagam and Kaiso specimens and
should be included in the genus Primelephas. These two specimens
of P. korotorensis are distinguishable from P. gomphotheroides in
being proportionately higher crowned, but otherwise they are too
fragmentary for diagnosis of the species, and more detailed com-
parisons with material from other localities is not possible. On
present evidence we must recognize two species of Primelephas —
P. gomphotheroides and P. korotorensis.

Loxodonta F. Cuvier 1825
Type Species. Loxodonta africana (Blumenbach) 1797.
Loxodonta adaurora] sp. nov.
(Pis. V-VI)

Elephas cf. meridionalis Nesti, Maclnnes, 1942: 92.
Archidiskodon exoptatus Dietrich (in part), Dietrich, 1942: 72. tt

+ L. adaiirora^iai dawn,
ttin a review of the syntype collection of "Archidiskodon exoptatus," I
(1969) have shown that two distinct taxa are present — one is referred
to Elephas recki Dietrich 1916, and is represented by the better speci-
mens in the collection; the other taxon is a more primitive species.
The specimens of the latter taxon are insufficient for the establishment
of a species and a recki specimen was chosen as the lectotype of ''A.
exoptatus." The non-recki material in the syntype collection from
Laetolil is here referred to L. adaitrora sp. nov.

970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 13

«J

»-.

o

>•*

o

:^

o

-«

cC

s

o

OJ

'^

3

</^

13

c

>

(U

C

T3

a

^

4-^

f-N

«

-^

D.

J

-4—*

■*— »

Q.

>

c

cd

■■4—

o

^"^

OJ

r<-)

ir

■4— »

'r-

S

r-

o

7^

o

,5

.u

-1

j=

^

o

'^

■^

J

~

LU

'X

^

-1 c

CO ■-'

C

.£■

<i I/)

P

'C

■^

^ c

■^

-~

>

?~

c

o

D.

"J;

(U

c

3

~

5

I/)

OD

.;:;

d.

o;

c^

V)

(U

4.^

ly^

r-

(U

_aj

4—*

QJ

c

^—

>

(U

C3

-4— »

Co

D/J

o

■^

C3

^

c

D.

^ Sj

cd

E

'V

■^

_u

O

u

c
o

<

_^»

^

c

Q
<

u

H

O

I— I

<

^ O

<

H
O

01

^

0

1

0

ro'

' in

00

0

\0

o f^ ' Ln c r^

IT)

'f 00

so

^

0

^"

0

00

1

r-J

X

CN

0

^ — '

^

cS

r-~

0

u~,

O)

ro

lO '^

:§ ^ j. -■ ^

<N

6 ^ I S S ^

CN M un

00 lT)

c-'

oi

0

LT,

CJ

yi

c

-J-.

OJ

OJ

^

c

cr

C/5

^^

<U

aj

a

1-

X

tj-

c^

IE

u

"H.

T3

,^

—

C

£

a>

c

E

Ti

0

~3

.-^

^

OJ

c

"Z.

d

^

1J

■^-

UJ

^

z:

X

X

14 BREVIORA No. 341

Ek'pluis cf. plaiiifrons Falconer and Caiitley, Aiambourg, 1947: 114.
"Elephas . . . close to E. exoptcitus", Patterson, 1966: 5.

Type. PAL.KANAP. 385, a nearly complete skeleton, the skull
and jaws bearing partially worn last molars in situ.

Horizon and Locality. Plio-Pleistocene, Kanapoi beds, Kenya.

Hypodigm. Type and PAL. LOTH. 353, skull with left M--M%
right M-' and left premaxillary incisor. PAL.KANAP. 383, left
and right M-\ worn left M"' and incomplete left and right M.,.
PAL.KANAP. 390, complete right M-^ PAL.KANAP. 40^7, com-
plete right M... PAL.EKA. 423, palate with left and right M^.
PAL.KANAP.' 386, left M-. PAL.KANAP. 389, incomplete right
M-. PAL.KANAP. 403, partial ramus with incomplete left M.,.
PAL.KANAP. 406, jaw ramus with r.M,-M,. PAL.KANAP. 381",
jaw with left and right M,. PAL.KANAP. 391, right dM^.
PAL.KANAP. 411, right dM^ left and right dM.,. PALrKANAP.
382, right dM''. PAL.KANAP. 396, right dM-='. PAL.KANAP.
392, right dM-. PAL.KANAP. 400, lef^t dM,.

Localities for Hypodi'gin. All from the Kanapoi beds except
353, which is from the upper tluviatile beds at Lothagam Hill, and
423, from Ekora.

Diagnosis. Molars low crowned, the height equal to or less than
the width; enamel thickness 3-5 mm and not folded; very large
anterior and posterior enamel folds forming prominent loops or
sinuses with wear. Plates thick and well separated, the lamellar
frequency being about 2.5-4. Lateral sides of plates lacking the
strong tapering as in M. ajricanavus. Skull like that of L. africana
but with very large and somewhat flaring premaxillae, long frontal,
and prominent fronto-parietal ridges lateral to external naris.
Parietals and occipitals not expanded. Occipital condyles high and
projecting. Tusks long and gently curved in a single plane, not
twisted as in Maininuthns. Mandible with vestigial incisive cavi-
ties. Lamellar formula:

M3 8-10, M2 7^ Ml ^ dM4 5^ dM3 5^ dM2 3^

10-11 ? 6 ? 5 3

Discussion. Until now there has been little direct evidence con-
cerning the early stages in the evolution of Lo.xodonta africana.
Coppens (1965) placed the north African species Manuinit/ius
africanavns on the Lo.xodonta lineage, but recent study and new
material (Arambourg, pcrs. comm.) show that africanavns was not
on that line.

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 15

Other early elephants such as Elephas planijrons from the Villa-
franchian of Asia, and M. nwrhUomiUs from deposits of similar age
in Europe, were already dilTcrentiated along different lines and
were unrelated to the living African elephant. The African Plio-
Pleistocene species, M. subplanijrons, though poorly known, has
Mamnnithiis affinities and also appears to have been unrelated to
Loxodonta. From all available evidence, Loxodonta adauroni sp.
nov. is the only known species certainly on the Loxodonta lineage,
and extends this line back to the earliest Pleistocene or latest Plio-
cene in Africa.

The earliest known representatives of the three Pleistocene
genera of Elephantinae, Loxodonta, Mammiithus, and Elephas,
were not yet greatly differentiated with respect to dentition, though
distinct on skull characters. This fact has resulted in the confusion
of these lines when only teeth were available for study. In L.
adawora, M. africanavus, and E. planijrons, the molars are all
rather low crowned, being about as high as they are wide; the
enamel is thick and unfolded with a more or less prominent median
fold or loop caused by the partial fusion of intravalley columns into
the face of the plates; and the number of plates is similar, about
9-12 for M3 (Table 3). Nevertheless, certain differences occur in
the molars that distinguish populations of these three species. In
M. africanavus the sides of the plates taper strongly toward the
apex of the crown, and the molar tends to be proportionately nar-
rower than in either L. adaurora or E. planijrons. The lamellar
frequency for M. ajricanavus is somewhat greater than for L. ad-
aurora, the latter being more primitive in this respect. Finally, the
relative crown height for M3 of ajricanavus is nearly 20 per cent
greater than for L. adaurora. The lamellar frequency, crown
height, and other parameters for the known specimens of E. plani-
jrons from the Siwalik Hills show a considerably greater degree of
variation than for other well known species. Yet there exists every
intermediate stage between the extremes for each measurable para-
meter. It seems likely that in E. planijrons we are dealing with an
evolutionary series of populations for which stratigraphic control is
lacking.

Far more important for the determination of generic criteria in
these early elephants are the skulls (Figs. 2 and 3), which are very
different in the three lines and clearly establish relationships with
later spegies. The cranial morphology for L. adaurora and E.
planijrons are each known from several specimens. The skull of
M. subplanijrons is unknown and that of M. ajricanavus is known

16

BREVIORA

No. 341

only from a poorly preserved referred skull recently described by
Professor Arambourg. Though this description is not yet published,
photographs kindly supplied by Professor Arambourg show that
this species is on the Mammuthiis line and that it was close to, but

<_ /

B'

Figure 2. Diagrammatic representation of skull shape for two species
of Loxodonta. A, Loxodonta odaiirora sp. nov. B, Loxocionta africana.
For each: left, anterior view; right, left-lateral view. A is based on PAL.
LOTH. 353 (solid line) and PAL.KANAP. 385 (broken line). Not to scale.

1970

NEW PLIO-PLEISTOCENE ELEPHANTIDAE

17

more primitive than, M. meridionalis. The following cranial com-
parisons are made with the later species representing the Mam-
muthiis type.

A'

Figure 3. Diagrammatic representation of skull shape for early species of
Mammiithiis and Elephas. A, Mainmiithus meridionalis. B, Elephas
planifrons. For each: left, anterior view; right, left-lateral view. Recon-
structions based on several specimens: A, Geol. Inst., Univ. of Florence
Nos. 1049, 1051 and 1054. B, British Museum (N.H.) No. 3060 and
Panjab Univ. specimen, no number.

18 BREVIORA No. 341

Elephas planifrons, (Fig. 3B).

Although tusks are lacking from the four known skulls of this species,
isolated incisors probably referable to it from the Siwalik Hills and an
associated tusk with planifrons dentition from Bethlehem (Hooijer,
1958) indicate that the tusks were gently curved in a single plane, and
probably not very large. The fronto-parietal surface is flat and broad,
and separated laterally from the temporal fossa by a sharp ridge
formed where the parietal bone makes an acute angle between the two
surfaces. This ridge is more sharply defined in later species of Elephas,
and in the living species it is very prominent along the lateral margins
of the forehead. The parietals and occipitals are slightly expanded with
a median-sagittal depression typical of this genus. The temporal con-
striction is slight, resulting in a very broad fronto-parietal surface.
The occipital condyles are situated high on the head, about one-third
down from the dorsal surface of the occiput. The external naris is
very small and not downturned at the sides. The tusk sockets appear
to have been moderately separated and the premaxillae are nearly paral-
lel to the fronto-parietal plane. The latter is inclined at about 55 to 60
degrees to the vertical axis of the skull measured perpendicular to the
palate. As a result, the tusks are directed forward and downward. The
skull is greatly compressed parallel to the facial axis.

Mammuthns meridionalis. (Fig. 3A).

The tusks in Mammiithus are typically massive and spirally twisted.
The tusks on the skull of M. africanaviis (Arambourg, pers. comm.)
and a tusk associated with a molar of M. siibplanifrons (Meiring, 1955),
have this structure. The fronto-parietal surface is strongly concave
dorsoventrally and flat to slightly convex in the lateral direction. A
prominent but rounded angle of the parietal separates the fronto-parietal
surface from the temporal fossa. The parietals and occipitals are ex-
panded dorsally so that the occipital condyles lie closer to the level of
the palate than to the top of the occiput. The external naris is large and
laterally elongated; it is slightly downturned at the sides. The tusk
sockets are closely spaced proximally. but diverge distally. The pre-
maxillae are nearly parallel to the fronto-parietal plane, and the latter
forms an angle of about 50 to 55 degrees with the vertical axis of the skull.
As in E. planifrons, the tusks are directed forward and downward. As
in Elephas, the skull is strongly compressed parallel to the facial axis.

Loxodonta adanrora sp. nov. (Fig. 2A)

The tusks are massive, but only gently curved upward, and are in a
single plane. As in L. africana, the fronto-parietal surface is slightly
convex in both the dorsoventral and lateral directions, but the frontal is
proportionately longer, resembling Elephas more in this respect. As a
result, the orbits are lower on the head than in L. africana. The forehead
curves laterally into the temporal fossae without a sharp, angular border.
The parietals are not expanded, nor is the occipital region which, like
the living African species, is nearly flat posteriorly. The highest point of

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 19

the parietals is in the mid-sagittal line; there is no median depression.
The temporal constriction is very slight, as in the recent species. The
occipital condyles are low on the skull. The external naris is large and
weakly downturned at the sides. Tusk sockets are widely separated and
flaring distally; these are far more massive than in the living species.
The premaxillae and maxillae are fused in the median line for their
entire length and form an angle of about 20 degrees to the fronto-
parietal plane. In L. africana this angle is about 15-20 degrees. The
fronto-parietal plane is inclined at about 45 degrees to the vertical axis
of the skull, and is therefore more nearly vertical in orientation than in
E. planifrons or M. meridionalis. The skull is only slightly foreshortened
in the anteroposterior direction and the skull is more nearly rounded in
shape than it is in either Mammuthus or Elephas.

In Loxodonta adawora, a pair of elongated cavities near the
symphysis suggests the presence of vestigial incisive buds, but these
apparently never developed into external tusks. These cavities
communicate posteriorly with the mandibular canal by a narrow
channel. Such structures have also been observed in two man-
dibles of E. planifrons in the Indian Museum. In those early ele-
phants with external mandibular tusks, such as E. celebensis and
P. gomphotheroides, a narrow channel passes forward from the
anterior end of the mandibular canal into the enlarged alveolus for
the incisor. Vestigial incisive cavities have not been observed in
more progressive fossil species or in the two hving species.

Despite the numerous references to E. planifrons in Africa
(e.g. Joleaud, 1928; Romer, 1928; Kent, 1941; Maclnnes, 1942;
Arambourg, 1947, 1948a, and 1948b, etc.), that species does not
appear to have occurred outside of Asia.* Specimens so referred
in Africa probably belong either to M. africanavus or to L. adau-
rora. The former appears to have been confined to north Africa,
but the data are inconclusive. Material probably referable to
L. adawora has been recovered from the Kaiso formation (Cooke
and Coryndon, in press), Kanam, Yellow Sands (Omo), Laetolil
(Maglio, 1969), the Chemeron beds, and the Chiwondo beds
(Mawby, in preparation). (See appendix for a discussion of speci-
mens from the Vaal River).

Morphologically, L. adawora is distinct from M. africanavus,
with which it is often confused on molar evidence alone. It appears
to have been the dominant elephant during the late Pliocene and

* Specimens called E. planifrons from Europe belong to early stages of
M. meridionalis.

20 BREVIORA No. 341

early Pleistocene of east Africa but was quickly replaced by
Elephas recki, which became the dominant species until middle
Pleistocene times.

All evidence indicates a close relationship between L. adaurora
and Primelephas gomphotheroides. Thus, Loxodonta appears to
have differentiated very early in the history of the subfamily.

Elephas Linnaeus 1758
Type Species. Elephas maximus Linnaeus 1758.
Elephas ekorensis'^ sp. nov.
(PI. VII)

Type. PAL.EKA. 424, left and right M"*.

Horizon and Locality. Plio-Pleistocene, Ekora beds, Ekora,
Kenya.

Hypodigm. Type and PAL.KANAP. 387, partial right M^ and
right M-. PAL.KANAP. 395, partial left M-'. PAL.EKA. 420,
mandible with incomplete left M^. PAL.KANAP. 412, incomplete
right M-.

Referred Material. PAL.EKA. 422, skull with eroded palate,
lacking teeth.

Localities for Hypodigm. 420 and 422 from the type locality;
395, 387 and 412 from the top of the Kanapoi beds.

Diagnosis. Molars with crown height 10-25 per cent greater
than width; M3 broader anteriorly, becoming very narrow pos-
teriorly; anterior and posterior sinuses very prominent; enamel
3-4 mm thick and very weakly folded near the bases of the plates.
Plates well separated with a lamellar frequency of about 3.5-4.
Eleven plates plus a strong heel on M-^ Skull compressed parallel
to the fronto-parietal plane; parietals slightly expanded with a mid-
sagittal depression; fronto-parietal surface strongly inclined to the
vertical axis of the skull; external naris very large; strong fronto-
parietal flanges forming a sharp angle between the forehead and
temporal fossae; tusk sockets parallel, widely separated and
parallel to the facial axis.

Discussion. The molars of Elephas ekorensis sp. nov. are
clearly distinguishable from those of Loxodonta adaurora, with
which they occur both in the upper part of Kanapoi and at Ekora.
The plates in the former are thinner and more closely spaced, and
the enamel is weakly folded on moderately worn plates. The

* ekorensis = the elephant from Ekora.

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 21

^.

q

q

c<s

q

OC)

en

^'

o

UO

M

c<-i

LT)

iri

^M

,

^^

^

2

s

1
o

1

q

1

CO

1

q

1

q

1

^

(^i

m"

CN

iri

r-

H-.

00

r-

r~

s

^^

Q

"^

a.

^_

.^

q

CN

o

crs'

Cv

Ci.

<rN

CN

in

(N

T— t

eg

o

s^

•^

tN

lO

lT)

^^

^^

T— t

'— •

i

Q

s

1

q
o

00
CN

1
o

CN

1

1
q

c

C7\

1

o

00

00

CM

Q

c:

v:

c

q

q

en

iri

SO

CN

^ — '

u-i

O

Tf

00

t^

to

CN

Tt

Tj-'

T 1

T— 1

T— 1

1

s

1

1

q

CO

1
en

1

o

(M_

1

1

o
o

i.

00

00

CN

CN

1

§

"Q

"S<

-c

O

(N

-*«.«

to

r-

o

o

q

C^

00

o

en

O

r-

O
^

5

s

o

iH

O

CN

T

CN]

T

CN
1

en

n

u-i

tij

^

^

CN

m'

en

00

Tt-'

CN

LU

O

t5

c;^.

00

00

_l
CO

<

m

-o

^

H

,

u

>i

>

1 — 1

,— ,

q

a

o

G

o

oo

CN

_

in

en

O

en

C/2

^

.

CO

m

^'

en

T— 1

T— 1

T— •

C^

r-*

a.

s

1

1

1

1

1

1

1

c

5

V5

^-

q

CN

o

m

^-

o

<D

.^

a

r--'

en

CN

^~'

NO

o

00

B

^

^
p

CN

CN

CN

^

a

i.

5

3

^

a

c/i

^

K3

^

a

CN

CN

CN

6

c^i

o

00

CN

o

CN

o

o

f*-

!^

"1

m

^'

en

T 1

,_^

,_^

c:^

(D

s:

s

1

1

1
in

00

1

NO

1

1
o

_>

Q

•-J

00

rn

ni

en

d

CN

§

(N
CN

o

CN

s

o

U

u

CO

CO

c

aj

<u

C

3

^

cr

OT

4— »

2

c«

"a

X

JZ

"aS

^

"«

4->

c

s

E

o

■5

^

_1

c

o
Z

^

^

22 BREVIORA No. 341

crown height is greater than the width and M3 narrows posteriorly,
unlike that tooth in L. adaiirora. The skulls of the two species
are very different. In E. ekorensis (Fig. 4A), the premaxillae do
not flare as in the Kanapoi species, the parietals are expanded
laterally and dorsally with a median-sagittal depression as in
E. planijrons (Fig. 4D), so that the brain case is widest in this
region and not in the auditory region as in L. adaurora. The skull
is compressed as in E. planijrons and E. recki, and contrasts with
the more rounded skull of L. adaurora. The fronto-parietal axis
of E. ekorensis is strongly inclined to the vertical axis of the skull,
again in contrast to the condition in Loxodonta. The skull differs
from that of E. planijrons in having more prominent, widely
separated tusk sockets and a very large external naris.

As discussed elsewhere (Maglio, in press), the earliest recog-
nizable stage of E. recki (stage 1 of 4 stages there recognized)
occurs at Kikagati, Uganda. This form is close to E. ekorensis,
but differs significantly in having proportionately thinner enamel,
a greater lamellar frequency, and a proportionately higher crown.
There also appears to have been a greater number of plates on M3.
Abundant elephant remains, including skull material, from the
White and Brown Sands localities at Omo probably represent
stage 1 recki, but this material has not yet been studied in detail.
The cranial remains are similar to the skulls of stage 2 recki
described by Arambourg (1947) from the type Omo area (Cop-
pens, pers. comm.), and in the following discussion comparisons
of E. ekorensis are made with the latter material.

The skull of E. ekorensis differs from that of stage 2 recki (and
probably stage 1 as well), though the differences are in degree
only (Fig. 4A and B). The facial compression, expanded parie-
tals, strong frontal flanges, parallel tusk sockets, and inclination
of the facial axis to the vertical axis of the skull are characteristic
features of the two species, but in E. ekorensis these are only
weakly developed in comparison with E. recki. The fronto-
parietal region in the Ekora species is still relatively flat; the
anterior expansion of the parietals is not yet evident. In recki,
the fronto-parietal area is concave anteroposteriorly because of
the forward expansion of the parietal vault. The tusk sockets of
recki are more closely spaced than in E. ekorensis and the ex-
ternal naris is larger and strongly downturned at the sides.

In both cranial and dental characteristics, E. ekorensis is con-
siderably less progressive than E. recki. The Ekora sediments
appear to be earlier than Tuff B at Omo, which is the earliest
recorded occurrence of E. recki. An ancestral-descendent rela-
tionship between these two species is indicated by present evidence.

1970

NEW PLIO-PLEISTOCENE ELEPHANTIDAE

23

..— U-

Figure 4. Diagrammatic comparison of skull shape for species of Elephas.
A. Elephas ekorensis sp. nov. B, Elephas recki. C. Elephas luaxinuis. D.
Elephas planifrons. Anterior view. Reconstruction in dotted lines for A is
conjectural. Not to scale.

24 BREVIORA No. 341

CO

o -s

5^ . -

to

§:cf

<
o
<

-2d '^ -,

_^ o ex

UJ rn UJ

DQ S =t:
<

eg

^ Q <N eg

■ 5 I I

Q, .< ' ^ CO

O i)

(u « o ^

3

§ w t

(U

St3

(U 03

^

*^

u

CI,
S

o
U

o

1

t

C>

1

q

1

SO

m'

C^

CJ

M

C

c«

(U

u

3

o

u

^j

Ic

<4-

rt

*^

U

a

__

rt

r"!

<u

«^-

c

E

S

o
6

1>

c

03

-J

UJ

J

Z

(N

o

in

CN

CO

CO

o

ro

CO

iri

,—1

00

,—1

^-^

00

1

1

1

1

1

CM

LO

q

(N

LO

q

q

CNJ

lO

■^^

o

lO

o

Ln

00

rO

in

q

1

00
1

T)-

CN
[

CM

m

1

1
q

1

q

1

(M

I

_

1

00

iri

o

CO

u-i

00

c<

o

CO

CM

q

U-)

^_

q

CO

iri

CN

CO
CO

-H

i

1 <^'
1

iri

■^

00
1

^-

'—

c3

1

' 00

1

1

1

00

1

CN

^'

■^

d

■^'

o

-^

00

o

rj

M

*— •

T— ■

-«

«J

^

o

■^

q

■o

00

co'

o

<r^

r-

1.M

d

00

CN

O

I

(N

r^

1
in

I

00

CN

00

iri

CN

CN

O

CO

r-"

o

1 — ■

^«

1

1

1

1

1

NO

1

^'

in'

T^

CTs

o

»—

X
u

•o

c

£ ^ >

^ X X

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 25

CONCLUSIONS

The abundant new proboscidean material now available from
the Plio-Pleistocene deposits at Lothagam, Kanapoi, and Ekora
helps significantly to bridge previous gaps in our knowledge of the
origin and early stages in the evolution of the Elephantidae. This
material is all the more important because it allows a means of
relative correlation based on elephants that has been shown to be
of great potential value, even for poorly known localities (Maglio,
in press). The transition from gomphothere to elephant now
seems established, and the time and place of the origin of the
Elephantidae is reasonably determined as the later Pliocene of
Africa. Morphological evidence strongly favors a monophyletic
origin for this group.

APPENDIX

A number of elephant species based on incomplete type speci-
mens have long caused confusion in the literature because they are
inadequate to provide the kind of diagnosis necessary for fossil
species. The following described taxa may be conspecific with
Loxodonta adaiirora or possibly with Mammuthiis africanavus,
but should not be given priority because of inadequacy of the
types which are, in most cases, the only known specimens. The
localities are vague and the stratigraphic relationships are essen-
tially unknown.

Archidiskodon vanalpheni Dart, 1929. A single incomplete left
M^ from the Middle Terrace of the Vaal River, South Africa.
The stratigraphic relationships are uncertain, and it is not pos-
sible to determine the total number of plates with confidence,
though there may have been about ten. The plates are broad
and well separated; the specimen resembles M'^ of L. adanrora.
Archidiskodon milletti Dart, 1929. An incomplete left M' from
the Middle Terrace of the Vaal River. The plates are some-
what thinner than in A. vanalpheni, but not significantly. As
with that form, the specimen is insufficient for adequate com-
parison with other material, without skull evidence.
Archidiskodon loxodontoides Dart, 1929. A single fragmentary
left M^ from the Middle Terrace of the Vaal River. From what
remains this specimen appears to be close to the previous two
forms, but specific diagnosis is not possible.
Loxodonta griqua Houghton, 1922. Several very fragmentary
specimens from the Vaal River gravels, horizon unknown. This
material is too incomplete to allow confident specific comparison
with any other specimen. It was made the type of a new genus,
Metarchidiskodon, by Osborn (1934).

26 BREVIORA No. 341

From all available evidence, the above material appears to be
close to east African material here referred to Loxodonta adaurora,
and to specimens of M. ajricanaviis, but specific identity with
either can not be proved because of the incomplete morphological
and stratigraphic data. The South African specimens have a
greater lateral taper to the plates than do those of L. adaurora and
in this respect are closer to M. ajricanaviis. The types are inade-
quate for specific diagnosis and in view of the excellent and com-
plete material from other localities, they should not be used as
the basis for specific comparison with material for which strati-
graphic data and associated faunas are known.

The names A. vanalpheni, A. milletli, A. loxodontoides, and
L. griqiia must be considered nomina dubia, and the names applied
only to the type specimens.

ACKNOWLEDGEMENTS

I am grateful to Professor Bryan Patterson for permission to
study the proboscidean remains collected during expeditions of the
Museum of Comparative Zoology. The fossil material was sent
to Harvard University with the kind cooperation of the Ministry of
Natural Resources of the Kenya Government and of the National
Museum. Comparative material was made available through the
courtesy of Dr. C. Arambourg, Museum of Natural History, Paris;
Dr. F. Azzabi, Museo Libico di Storia Naturale. Tripoli; Dr. A.
Berzi, Museo di Geologia e Paleontologia, Universita di Firenze;
Dr. D. A. Hooijer, Rijksmuseum of Natural History, Leiden;
Mr. R. E. Leakey, National Museum, Nairobi; Dr. L. S. B. Leakey,
Centre for Prehistory and Paleontology, Nairobi; Prof. J. P.
Lehman, Museum of Natural History, Paris; Dr. R. Liversidge,
McGregor Memorial Museum, Kimberley; Dr. K. H. Fischer, In-
stitute for Paleontology, Humboldt University, East Berlin; Dr.
L C. Pande, Department of Geology, University of Chandigarh;
Dr. Petronio, Department of Geology, Cita Universitaria, Rome;
Dr. M. V. S. Sastri, Indian Museum, Calcutta; Dr. A. Sutclifl'e,
British Museum (Natural History); and Dr. R. Tedford, The
American Museum of Natural History. I have benefited greatly
from discussions with Drs. E. Aguirre, C. Arambourg, H. B. S.
Cooke, Y. Coppens, Mrs. S. Coryndon, and Dr. D. A. Hooijer.
Prof. Bryan Patterson read the manuscript and ofl'ered helpful
comments.

The field work and research was supported in part by grants
from the National Geographic Society and the Museum of Com-
parative Zoology to the author, by NSF Grants no. GP-1188 and

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 27

GA-425 to Prof. Patterson, and by an Evolutionary Biology Train-
ing Grant to the Department of Biology, Harvard University, NSF
Grant no. BG-7346 (Reed C. Rollins, principal investigator).

LITERATURE CITED

Aguirre, E.

1969. Evolutionary history of the elephant. Science, 164 (3886):
1366-1376.
Arambourg, C.

1947. Contribution a Tctude geologique et paleontologique du bassin
du lac Rudolf et de la basse vallee de I'Omo. II, Paleontologie.
Miss. Scient. Omo, 1932-1933, 1, Geol.-Anthrop.: 232-562.
1948a. Les vertebres fossiles des formations des plateaux Constan-
tinois. (Note preliminaire). Bull. Soc. Hist. Nat. Afr. N.,
Alger, 38: 45-48.
1948b. Les mammiferes pleistocenes d'Afrique. Bull. Soc. Geol. Fr.,
Paris, (5) 17: 301-310.
Cooke, H. B. S.

1960. Further revision of the fossil Elephantidae of southern Africa.
Palaeontologia Africana, 7: 46-58.
Cooke, H. B. S., and S. Coryndon

(in press). Fossil mammals from the Kaiso formation and other related
deposits in Uganda. Fossil vertebrates of Africa, vol. 2, (1969).

COPPENS, Y.

1965. Les proboscidiens du Tchad, leur contribution a la chronologic
du Quaternaire Africain. Actes du Ve Congr. Panafricain de
Prehist. et I'Etude du Quaternaire, 1963: 331-387.
Dart, R. A.

1929. Mammoths and other fossil elephants of the Vaal and Limpopo
watersheds. S. African Jour. Sci., Johannesburg, 26: 698-731.
Dietrich, W. O.

1942. Altestquartare Saugetiere aus der siidlichen Serengeti, Deutsch-
Ostafrika. Paleontographica 94 (A): 43-133.
Houghton, S. H.

1922. A note on some fossils from the Vaal River gravels. Trans.
Geol. Soc. S. Africa, Johannesburg, 24: 11-16.
HooiJER, D. A.

1954. Pleistocene vertebrates from Celebes. XL Molars and a
tusked mandible of Archidiskodon celebensis Hooijer. Zool.
Med. Museum, Leiden, 33: 103-120.
1958. An early Pleistocene mammalian fauna from Bethlehem. Bull.
Brit. Mus. (Nat. Hist.), London, 3 (8): 267-292.
Howell, F. C.

1968. Omo research expedition. Nature (London), 219 (5154):
567-572.

28 BREVIORA No. 341

JOLEAUD, L.

1928. Elephants et dinotheriums pliocenes de I'fithiopie: contribution
a Tctiide paleogcographique des proboscidiens africains. XIV
Int. Geol. Congr., Madrid, 3: 1001-1007.
Kent, P. E.

1941. The recent history and Pleistocene deposits of the plateau
north of Lake Eyasi, Tanganyika. Geol. Mag., London, 78:
173-184.

MacInnes, D. G.

1942. Miocene and post-Miocene Proboscidea from east Africa. Trans.
Zool. Soc. London, 25: 33-106.

Maglio, V. J.

1969. The status of the east African elephant " Archidiskodon exop-
tatus" Dietrich 1942. Breviora, no. 336: 1-25.
In press. Early Elephantidae of Africa and a tentative correlation of
African Plio-Pleistocene deposits. Nature (London), 1969.
Meiring, a. J. D.

1955. Fossil proboscidean teeth and ulna from Virginia, O.F.S. Re-
searches Nas. Mus. Bloemfontein, 1 (8): 187-202.

OSBORN, H. F.

1934. Primitive Archidiskodon and Palaeoloxodon of South Africa.
Amer. Mus. Novit., no. 741: 1-15.
Patterson, B.

1966. A nevi' locality for early Pleistocene fossils in northwestern
Kenya. Nature (London), 212 (5062): 577-581.

1968. The extinct baboon, Parapapio jonesi, in the early Pleistocene
of northwestern Kenya. Breviora, no. 282: 1-4.
Patterson, B., and W. W. Howells

1967. Hominid humeral fragment from early Pleistocene of north-
western Kenya. Science, 156 (3771): 65-66.

Petrocchi, C.

1941. I giacimento fossilifero di Sahabi. Boll. Soc. Geol. Italiana, 60
(1): 107-114.

1943. I giacimento fossilifero di Sahabi. Coll. Scient. Docum. a Cura
(Min. A. I., IX), Vol. 12.

1952. Notizie generali sul giacimento fossilifero di Sahabi Storia
degli Scavi-Resultati. Rendiconti Accad. Naz. Dei XL, Ser. 4,
3 (75 della fondazione): 1-33.
1953-1954. I proboscidati di Sahabi. Rendiconti Accad. Naz. Dei XL,
Ser. 4, 4-5 (76-77 della fondazione): 1-74.

ROMER, A. S.

1928. Pleistocene mammals of Algeria. Fauna of the paleolithic sta-
tion of Mechta-el-Arbi. Bull. Logan Mus., Beloit, Wisconsin.
1: 80-163.

(Received 23 September 1969.)

PLATES

30

BREVIORA

No. 341

PLATE I

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 31

PLATE I

1. Stegotetnibeloiloii orhiis sp. nov. PAL. LOTH. 3.59, r.M.,. Occlusal
view. X V3.

2. Stegotetrabelodon orbiis sp. nov. PAL.LOTH. 359, r.M.,. Buccal
view. X Vs.

3. Stegotetruhelodon orhiis sp. nov. PAL.LOTH. 359, I.M-'. Occlusal
view. X V3.

4. Stegotetrahc'iodoii orhiis sp. nov. PAL.LOTH. 359, I.M-"-. Buccal view.
X Vs.

32

BREVIORA

No. 341

PLATE II

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 33

PLATE II

5. Stegotetrahelodon orbits sp. nov. Type. PAL. LOTH. 354, mandible
with M.,-M . Left-lateral view. Approx. X 1/10.

6. Stegotetrahelodon orbits sp. nov. Type. PAL. LOTH. 354, mandible
with M.,-M . Dorsal view. Approx. X V4.

34

BREVIORA

No. 341

10

PLATE III

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 35

PLATE III

7. Primelephas gomphotheroides gen. et sp. nov. Type. PAL.LOTH. 351,
l.M . Occlusal view. X Vs.

8. Primelephas gomphotheroides gen. et sp. nov. Type. PAL.LOTH. 351,
l.M.,. Lingual view. X Vs.

9. Primelephas gomphotheroides gen. et sp. nov. Type. PAL.LOTH. 351,
l.M'^. Occlusal view. X 1/3.

10. Primelephas gomphotheroides gen. et sp. nov. Type. PAL.LOTH.
351, l.M^. Lingual view. X 1/3.

36

BREVIORA

No. 341

PLATE IV

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 37

PLATE IV

11. Primelephas goniphotheroides gen. et sp. nov. PAL. LOTH. 358, frag-
mentary symphysis showing ventral border of incisive alveolus (arrow).
Anterior view. Approx. X V2.

12. Primelephas gomphotheroides gen. et sp. nov. PAL. LOTH. 358, sym-
physis with partial left incisor in place. Left-lateral view. Approx. X V2.

13. Primelephas gomphotheroides gen. et sp. nov. PAL. LOTH. 376,
isolated molar plate, ?M3. Anterior view. X V2.

38

BREVIORA

No. 341

PLATE V

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 39

PLATE V

14. Loxodonta adaurora sp. nov. Type. PAL.KANAP. 385, mandible.
Left-lateral view. Approx. X 1/7.

15. Loxodonta adaurora sp. nov. Type. PAL.KANAP. 385, r.M .
Occlusal view. X 1/3.

16. Loxodonta adaurora sp. nov. Type. PAL.KANAP. 385, l.M^.
Occlusal view. X Vi.

40

BREVIORA

No. 341

PLATE VI

1970 NEW PLIO-PLEISTOCENE ELEPHANTIDAE 41

PLATE VI

17. Loxodonta adaurora sp. nov. PAL. LOTH. 353, skull. Left-lateral
view. Approx. X 1/9.

18. Loxodonta adaurora sp. nov. Type. PAL.KANAP. 385, skull.
Ventral view. Approx. X 1/11.

42

BREVIORA

No. 341

PLATE VII

970 Ni:\v I'l i()-i'i.i;isrc)ti-.Ni: ilipiian i idai; 43

PLATE VII

19. Elcphas ckorcnsis sp. nov. Type. PAI..EKA. 424, r.M'\ Occlusal
view. X V^ ■

20. Elepluis c'korcnsi.s sp. nov. Type. PAL.EKA. 424, r.M"'. Lingual
view. V iA.

21. Elcphas ckorcnsis sp. nov. PAL.EKA. 422. skull. Anterior view.
Approx. X 1/10.

^> i i •

V5

/VW-L.£:an.lov^,ci^e1 DO NOT ClRC^J'.'^^^O^lP. zoc

' LIBRARY

J * «f o

B R E V I O R „.„„„„

UNlVERSIXYi

Mmseiiini of Comparative Zoology

Cambridge. Mass. ;:{0 J.axuary, 1970 Number 342

REDESCRIPTION OF POMACENTRUS OTOPHORUS

POEY 1860, A VALID SPECIES FROM THE

CARIBBEAN (PISCES: POMACENTRIDAE)

Robert W. Toppi

Abstract. Pomacentrus otophorus Poey 1860 is restored to valid status
and redescribed on the basis of material from Panama and Jamaica. It is
distinguished from its western Atlantic congeners by a black opercular
"earspot," soft vertical fin tips of bright yellow-orange, and the combina-
tion of short head, small eye, deep suborbital, absence of ocelli, and strong
suborbital and preopercular serrations.

P. otophorus is unique among American damselfishes in that the adults
are excluded from waters of sustained high salinity. This aspect of their
ecology may have a bearing on the dissimilarity between Panama and
Jamaica populations, as well as on the implied phyletic relationship between
the Pomacentridae and Cichlidae.

INTRODUCTION

In late 1967 my attention was drawn to an unrecognized dam-
selfish living in the Chagres River Estuary on the Caribbean coast
of Panama. During subsequent months more than 60 specimens
were collected from this and other Panama localities. In June,
1968, 10 additional specimens were obtained from Jamaica.

Upon review of nominal Caribbean species I found that this
material seemed to agree in detail with the description of Poma-
centrus otophorus Poey 1860. Examination of the holotype,
USNM 4710, confirms the identity and thereby restores the species
to valid status.

Poey's original description was drawn from a single 1 30 mm TL
specimen and served as a model for his generic treatment. Jordan
and Evermann's (1898: 1555) account was essentially an abridg-
ment of Poey's original description.

1 Present address: Florida Department of Natural Resources, Marine
Research Laboratory, St. Petersburg, Florida; Contribution No. 131.

2 BREVIORA No. 342

Rivas (1960) reviewed the Florida and western Bahamas spe-
cies of Pomacentrus, but was unable to locate the holotype of
P. otophorus at the Museum of Comparative Zoology (a deposi-
tory of Poey's types). On the basis of Poey's unfigured descrip-
tion of its size, coloration, anterior profile, elongation of vertical
fins, and number of pectoral rays, Rivas judged it to have been a
large adult male of P. leucostictus Miiller & Troschel 1 848, and
accordingly synonymized the two. Subsequent works have fol-
lowed this synonymy by omitting P. otophorus from their lists.

METHODS

Panama specimens were obtained with Chem Fish Collector, a
"bio-aquatic stupefier"; collecting was supplemented by diving
observations. Methods for making counts and measurements fol-
low those of Rivas (1960). In Table 1, positive or negative al-
lometry resulting from ontogenetic change is indicated by (-)-)
or (— ). Color descriptions are based on 35 mm Kodachrome
transparencies of fresh Panama material. Specimens have been
placed in the following collections: USNM (United States National
Museum), MCZ (Museum of Comparative Zoology), ANSP
(Academy of Natural Sciences of Philadelphia), FMNH (Field
Museum of Natural History, Chicago), UMML (Institute of Ma-
rine Sciences, University of Miami), FSBC (Florida [State] Board
of Conservation, now the Florida Department of Natural Re-
sources, St. Petersburg).

MATERIALS EXAMINED

Pomacentrus otophorus: USNM 4710 (holotype), 130 mm TL,
Cuba, 1860. ANSP 109645, 102 mm. Port Antonio, Jamaica.
E. C. Raney and D. P. de Sylva, 13 Oct. 1967. USNM 204607,
(5) 39-88 mm, Port Antonio, 13 Oct. 1967. MCZ 46672, (4)
76-93 mm, Frenchman's Cove, Jamaica (11.9 %o sahnity), H.
Albrecht, 3 April 1968. MCZ 42737, (3) 32-62 mm, mouth
of Chagres R., Canal Zone, Panama, I. and R. W. Rubinoff,
9 June 1962. MCZ 46671, (4) 44-95 mm, mouth of Chagres R.,
I. and R. W. Rubinoff, 13 Aug. 1966. USNM 204606, (1^4) 11-
83 mm, mouth of Chagres R. (0 %o salinity), 24 Oct. 1967.
FSBC 5596, 51 mm, Chagres R., Tarpon Club Boat Basin (fresh
water), 24 June 1968. UMML 25066, (8) 14-56 mm, Panama
Canal, Atlantic Third Locks (21.5 %o salinity), 4 Nov. 1967.
FSBC 5597, (2) 16-32 mm, cleared and stained, Panama Canal,
Atlantic Mine Docks (28.1 %o salinity), 4 Nov. 1967. FSBC

1970 REDESCRIPTION OF P. OTOPHORUS 3

5598, (25) 13-79 mm, Panama Canal, Atlantic Mine Docks,
4 Nov. 1967. USNM 204608, (3) 40-64 mm, Panama Canal, At-
lantic Mine Docks, 4 Nov. 1967. FSBC 5599, 49 mm, Chagres R.,
1/2 km from mouth (0 %o salinity), 24 Oct. 1967. F^NH
69519, (9) 11-92 mm, mouth of Chagres R., (31.5 %o salinity),
13 March 1968.

Pomacentrus leiicostictiis: MCZ 25152, 70 mm, "Central Amer-
ica." MCZ 42728, (2) 53-62 mm, Galeta Pt., Canal Zone, Pan-
ama, I. and R. W. Rubinoff, 6 June 1961. MCZ 42731, 15 mm,
Galeta Pt., I. and R. W. Rubinoff, 26 April 1962. MCZ 42732,
(2) 20-50 mm, Galeta Pt., I. and R. W. Rubinoff, 26 April 1962.
MCZ 42738, IS mm, Galeta Pt., I. and R. W. Rubinoff, 20 June
1962. FSBC 5600, 69 mm, Galeta Pt., 7 Oct. 1967. 7 specimens,
15-70 mm, Devil's Beach, Ft. Sherman, Canal Zone, Panama, 12
Nov. 1967.

Pomacentrus fuscus: 10 specimens, 38-64 mm, Galeta Pt., Canal
Zone, Panama, 25 July 1966.

Cichlasoma tnacuUcauda: MCZ 46673, 104 mm, Chagres R.,
Tarpon Club Boat Basin (fresh water). Canal Zone, Panama, 24
June 1968. FSBC 5157, 62 mm, cleared and stained, Tarpon
Club Boat Basin, 24 June 1968.

Pomacentrus otophorus Poey

Figure 1.

Pomacentrus otophorus Poey, 1860: 188 (original description, Cuba);

1868: 326 (coloration, Cuba)
Eupomaccntrus otophorus. Jordan and Evermann, 1898: 1555 (description

after Poey, Cuba). Jordan, Evermann and Clark, 1930: 414 (listed,

Cuba).

Diagnosis

A euryhaline pomacentrid with short head, small eye, and deep
suborbital. Dorsal fin XII (rarely XIII), 13-14 (rarely 12); anal
II, 12 (rarely II, 11); pectoral rays 18 (occasionally 17 or 19).
Suborbitals 4 and 5 and vertical margin of preopercle strongly
serrate. Adults dark brown to blue-black; operculum with a
black "earspot" at upper angle (ofop/2on/5=bearing an ear); dis-
tal one-third of soft dorsal, anal, and caudal fins bright yellow-
orange; pectoral fin either unpigmented or tinged with orange
except for a dark blotch at base. Juveniles blue-black with tur-
quoise-blue stripes and spots; caudal fin and distal portions of soft
vertical fins yellow, transparent at tips. OceUi never present.

4 BREVIORA No. 342

Description

A shallow-bodied species (depth 43-49% in SL), with dorsal
outline more strongly rounded than ventral; anterior profile
strongly convex. Caudal lobes rounded. Soft dorsal and anal fins
rounded, both extending well beyond caudal fin base. Pelvics
produced or not.

Dorsal and pelvic fins inserted well forward. Head short (29.7-
33.3% in SL); orbital diameter small (7.5-11.5% in SL); sub-
orbital deep (3.2-5.5% in SL). Table 1 lists proportional mea-
surements for specimens greater than 39 mm.

Scales large, ctenoid; vertical fins with scale sheaths extending
onto interradial membranes about halfway to fin margins. Pec-
toral fin with small scales covering base. Pelvic fin naked, but
with a strong ensiform scale at each fin axil and another between
the fins. Cheeks scaled. Jaws, preorbital region, and suborbitals
1-3 naked; scale of suborbital 4 embedded when present. Lateral
line terminating below base of first soft dorsal ray. Teeth uniserial,
close-set, entire, similar in both jaws.

Fourth and fifth suborbital ossicles bearing 0 to 7 (mean, 2.4)
moderate to strong posteriorly directed serrae, usually present in
specimens 45-50 mm or larger. Vertical limb of preopercle with

5 to 15 (mean, 10.7) strong serrae; angle smoothly rounded.
Dorsal spines 12, 1 of 60 with 13; dorsal rays 13 or 14, 2 of 60
with 12; anal rays 12, 2 of 60 with 11; pectoral rays 17-19. Body
scales 26-29; tube-bearing scales in lateral line 16-19. Gill rakers
on lower limb (including raker at angle), 8-10; on upper limb, 2-3.
Trunk vertebrae 11; caudal, 15. Cheek scales 3-4 between sub-
orbital edge and angle of preopercle. Distribution of meristic
characters is shown in Table 2.

Coloration

Typical of many pomacentrids, P. otophoriis changes color with
ontogeny, tending toward a less conspicuous adult.

Small juveniles (<15 mm) of blue-black ground color, with
turquoise-blue markings arranged as follows (Fig. 2): 1) a con-
spicuous arc along dorsal one-third of corneal perimeter, colinear
with a postorbital stripe, continuing as a row of spots along scale
row just above lateral line, extending to posterior dorsal fin base;
2) successively less distinct rows immediately below lateral line,
each row extending caudad from operculum, but failing to reach
peduncle; 3) prominent markings on scales bordering dorsal fin

1970 REDESCRIPTION OF P. OTOPHORUS 5

base, with spots extending onto scales of interradial membranes;
4) a pair of variably broken lines on head and nape extending
from premaxillary groove posteriad between orbits and converging
at dorsal fin origin; 5) several markings arranged circumorbitally,
including a conspicuous line on vertical limb of preopercle. Pec-
toral fin either transparent or tinged with orange, except at base.
Pelvic fin pigmented, sometimes transparent or orange near tip.
Posterior tip of opercle transparent. Caudal yellow, becoming
transparent near tips of rays. Ocelli absent.

Larger juveniles with reduced turquoise-blue markings, which
by 30 mm appear as small circular blue spots on scales above
lateral line, on operculum, around eye, and occasionally else-
where. Preopercular markings and those below lateral line now
barely visible; paired lines on head reduced in breadth. Opercu-
lum developing a blue-black "earspot" resembling that of the pri-
mary freshwater centrarchid fishes of North America, and shared
by a few western Pacific pomacentrid species.

Subadults and adults with head and body dark brown to sooty
blue-black. Centers of scales black, producing vertical striations.
Blue markings indistinct or absent. Distal one-third to two-thirds
of soft vertical fins yellow-orange, this proportion generally de-
creasing with size; dark ground color encroaching well distad
along interradial membranes. Pectoral fin transparent or tinged
with orange, except for a black inverted triangle at base. Earspot
now prominent, blue-black to black. Opercular spine and fleshy
flap unpigmented. Sexual dichromism not apparent in preserved
specimens.

Most Panama specimens with a pair of dusky lines on caudal
fin, generally present by 35 mm, often obscured by extensive body
pigmentation in large fish; their extent and prominence variable,
but occurring in a constant position relative to caudal skeleton:
counting ventrad on third hypural, ventral line occupies membrane
between third and fourth lepidotrichs; counting dorsad on fourth
hypural, dorsal line lies between fourth and fifth lepidotrichs.
These lines not present on Jamaica specimens.

Comparisons

Pomacentnis otophonis differs from all other western Atlantic
congeners by its distinctive "earspot." Fresh adults are quickly
distinguished by their bright yellow-orange fin tips, juveniles by
their bright turquoise-blue markings. Additional differences in-
clude the small number of dorsal and anal rays, and the combina-
tion of short head, small eye, and deep suborbital.

6 BREVIORA No. 342

Although P. otophorus closely resembles P. leiicostictus in gen-
eral physiognomy, preserved material is readily distinguished by
the suborbital and preopercular serrations and the persistent ear-
spot. Subadults are further separated by the lack of oceUi. Com-
parative morphometric and meristic data for P. leucostictiis from
Panama is included in Tables 1 and 2.

Geographical Variation

Although comparative material is limited, differences between
Panama and Jamaica collections are readily apparent. Caudal
lines, for example, present on most Panama specimens greater than
35 mm, are lacking in similar material from Jamaica. Ranges of
measurements and counts for specimens from the two populations
are included in Tables 1 and 2. Using "t" tests on material of
comparable size, significant differences (P=0.01, d. f. ^28) can
be recognized in snout length (t=4.78) and interorbital width
(t^4.79). Figure 3 shows separation of the two populations on
the basis of these two characters.

To evaluate the taxonomic status of the two populations, inter-
and intraspecific differences among closely related species were
assessed, following suggestions of Mayr (1969: 197):

1) Degree of difference among closely related sympatric spe-
cies was evaluated by comparing Panama collections of Pomacen-
trus otophorus, P. leiicostictus, and P. fiiscus.

2) Degree of difference between geographically separated popu-
lations of P. leiicostictus and of P. fusciis were evaluated by com-
paring Panama data with those of Rivas (1960) for similar mate-
rial from Florida-western Bahamas.

The coefficient of difference, CD, defined as the difference be-
tween means divided by the sum of standard deviations, was deter-
mined for the same 10 sets of characters in each comparison.
Sums of the 10 coefficients, SCD, for the various combinations are:

Sympatric species

P. otophorus vs. P. fusciis 13.91

P. otophorus vs. P. leiicostictus 8.56

P. leiicostictus vs. P. fuscus 7.41

Widespread populations

P. fuscus (Panama vs. Florida) 4.47

P. otophorus (Panama vs. Jamaica) 3.89

P. leiicostictus (Panama vs. Florida) 2.46

1970 REDESCRIPTION OF P. OTOPHORUS 7

The observed differences between Panama and Jamaica popula-
tions of P. otopJioriis are consistent with those of widespread popu-
lations of related species, but do not approach those of interspe-
cific rank. The two populations are, on this basis, judged to be
conspecific.

Ecology and Behavior

Recognized species of Pomacentrus in the western Atlantic are,
as adults, typically obligate coral reef dwellers, and are thus prin-
cipally stenohaline. Juveniles of some, however, regularly inhabit
lagoons and rock pools, where they are exposed to fluctuating
salinities. Representatives of the genera Abudefduf and Nexilarius
are still more broadly tolerant of environmental extremes, espe-
cially as juveniles, and can be experimentally maintained in fresh
water for indefinite periods.

P. otophorus is unique among western Atlantic pomacentrids
since the adults are excluded from waters of sustained high salin-
ity. Jamaica specimens were collected in salinity of 11.9 %o,
whereas Panama specimens were collected in salinities of 0 to
31.5 %o. The species is common in the Atlantic end of the
Panama Canal and adjacent brackish waters, where salinities of
21.5 to 28.1 %o were recorded, and in the Chagres River Estu-
ary, where salinities ranged over the full regime of 0 to 31.5 %o.
Specimens have also been collected from the river proper, and
from the Tarpon Club Boat Basin adjacent to Gatun Dam spill-
way, over 10 km upstream. Figure 4 shows Panama collection
localities.

In waters of sustained high salinity (> 30 %o) its absence is
notable; not a single specimen has been collected in coral reef
habitats during some eight years of sampling by the Smithsonian
Tropical Research Institute.

In Panama, P. otophorus was collected from waters subject to
occasional high turbidity and turbulence. In the Panama Canal
and Chagres estuary, substrate is of moderately sedimented rubble;
in the Third Locks, rubble and mud. The Tarpon Club Boat
Basin is lacustrine, of good transparency, with dense growths of
phanerogams along the shore and heavy algal growth on the docks
and pihngs. A single 51 mm specimen of P. otophorus was col-
lected there on 24 June 1968, along with the following associates,
now in MCZ and FSBC collections:

1 Gambusia episcopi, 35 mm

2 Cichlasoma maculicauda, 62-104 mm

8 BREVIORA No. 342

12 Eiicinostomus sp., 35-58 mm

3 Diapteius rhombeus, 30-43 mm

4 Centropomus peclinatus, 88-1 10 mm
20 Gobionelliis sp., 22-55 mm

8 Microeleotris sp., 17-40 mm
1 Eleotris isthmensis, 22 mm

P. otophorus is exclusively herbivorous throughout life, as far
as is known. The only animal material present in gut samples was
small invertebrates probably ingested incidental to grazing.

In Panama, reproductive activity was not observed. In Jamaica,
the color pattern of courting males apparently involves a concen-
tration of dark pigment in the anterior parts, while the posterior
body and vertical fins remain yellow. Nest-guarding males retain
the yellow in parts of each scale (Alan Emery, personal communi-
cation, 27 June 1968).

Distribution

Although the degree of difference between Panama and Jamaica
populations of P. otophorus is consistent with that demonstrated
between Panama and Florida populations of related species, only
about one-half the latitudinal distance is involved, suggesting that
intraspecific heterogeneity in P. otophorus may exceed that of its
Caribbean congeners. There is, indeed, ample ecological basis for
such an expectation:

The exclusion of P. otophorus from waters of persistently high
salinity may effectively give rise to geographically isolated demes
along the Caribbean coast of Central America. If its reproductive
activities include the nest building and parental care typical of
other damselfishes, genetic exchange between demes is probably
Hmited. Their Caribbean congeners, however, have no extrinsic
salinity barriers to continuous populations along the coast. This
greater freedom of gene flow can be expected to result in more
panmictic populations.

Relationships

A close relationship has long been suspected between the pan-
tropical marine Pomacentridae and the secondary freshwater Cich-
lidac. The two share a number of anatomical similarities (e. g.,
united lower pharyngeals) and have been united into the order
Chromides by Jordan and Evermann (1898: 1511) and Jordan
(1923: 218). They are mutually unique among the Perciformes

1970 REDESCRIPTION OF P. OTOPHORUS 9

in having but a single pair of external nostrils. Despite their
divergence of habitat, the two families also show close similarities
in reproduction and behavior (Wickler, 1967).

In Panama, one specimen of P. otophorus (FSBC 5596) was
taken in company with two specimens of the cichlid fish, Cichla-
soma maciiUcauda; gut analyses showed that both species had been
feeding on filamentous green algae. This ecological overlap
strengthens the implication of close phyletic proximity for the
two famiUes.

ACKNOWLEDGMENTS

I am particularly indebted to Ira and Roberta W. Rubinoff for
their generosity and assistance throughout this study, and to Alan
R. Emery for providing specimens from Jamaica and for reading
the manuscript. I thank Stanley H. Weitzman, Robert H. Gibbs,
Jr., and Robert Kanazawa for providing information on the holo-
type, and Mrs. Myvanwy Dick for the loan of specimens. I am
also grateful to Giles W. Mead, C. Richard Robins, and Dale S.
Beaumariage for critically reading the manuscript. The financial
support of a Smithsonian Predoctoral Internship is gratefully
acknowledged.

LITERATURE CITED

Jordan, David S.

1923. A classification of fishes including families and genera as far
as known. Stanford Univ. Publ., Univ. Ser., Biol. Sci., 3(2):
77-243.
Jordan, David S., and B. W. Evermann

1898. The fishes of North and Middle America. Bull. U. S. Nat.
Mus., 47(2): 1241-2183.
Jordan, David S., B. W. Evermann, and H. W. Clark

1930. Checklist of the fishes and fishlike vertebrates of North and
Middle America north of the northern boundary of Venezuela
and Colombia. Rep. U. S. Comm. Fish., 1928, Pt. 2, App. 10.
670 pp.
Mayr, Ernst

1969. Principles of Systematic Zoology. New York. McGraw-Hill,
Inc. 428 pp.
PoEY, Felipe

1860. Poissons de Cuba, especes nouvelles. Mem. Hist. Nat. Isia de

Cuba, 2(49): 115-336.
1868. Synopsis piscium cubensium, Pt. 2. Anal. Soc. Espan. Hist.
Nat., 5: 131-218.

10 BREVIORA No, 342

RivAs, Luis R.

1960. The fishes of the genus Pomacentriis in Florida and the western
Bahamas. Quart. Jour. Florida Acad. Sci., 23(2): 130-162.
WiCKLER, Wolfgang

1967. Vergleich des Ablaichverhaltens einiger paarbildender sowie
nicht-paarbildender Pomacentriden und Cichliden (Pisces:
Perciformes). Zeitschr. Tierpsychol., 24(4): 457-470.

(Received 29 October 1969.)

1970 REDESCRIPTION OF P. OTOPHORUS 11

LU

_l
CD
<

6n

/-*\

VO

o\

CT\

00

0\

I--

oo

r--

rn

oo

O

ON

_

NO

NO

oo

^

4J

>/1

r~-

00

ON

o

00

00

ON

r«^

oo

v-t

"^

r^

<N

r^i

^"-^

S

C<~t

VO

m

m

•^

(N

m

^

rn

03

s

'^

rl

^^

On

NO

(N

^

r^

o

r~-

NO

O

C

o

as

o

CO

-^

■o

ri

»— 1

r~~

r-

r-~

ITl

r^

NO

m

NO

S

c

Oh

1

m

1

"^

r^

ON

1

OS

1

ro

On

1

^

1

oi

1

1

m

■■*"

C3

r^

»— «

>o

NO

^^

00

rr^

NO

>o

^

■o

o

r<-)

(N

»— 1

00

^
s

Oi

ro

\o

r~-

r~-

O

r^

00

00

(N

00

(^)

^

NO

ON

oo

o

0^

m

^

m

ro

'^

*— t

(N

(N

r«^

CO

■***

o

■*3

6?

to

^-v

C

^

00

1

o

ri

t~~

NO

NO

oo

<N

NO

'^

m

m

ON

1^

o

r-

oo

o\

00

^^

o

oo

OO

'^

ON

>0

lO

00

^H

ON

^^

O

x^

^^

m

VD

ro

CO

'^

^H

n

CO

m

m

s

o

■^

^

.^

•■■*i

03

fN

>o

r~-

1^

f.

NO

>o

m

00

r^

oo

m

ON

_o

<u

o

^H

^H

o\

r~,

^

^-H

>o

NO

00

NO

o

m

m

'^

cC

b

s

c

03

1^

a\

'si-

o

0\

On

ON
1

>A

1
00

as

1

lO

1
ON

r<-i

O

oo

NO

E

a!

m

r^

t^

NO

ON

0\

oo

t^

m

00

ro

'^

>0

r-

"o

oo

1-^

tn

VO

m

fN

'^

'"'

rN

(N

m

(N

ci

_o

to

"cS

S

^

- — >

<N

~H

ON

(N

>0

NO

^

NO

NO

T-,

00

f^J

r«-)

</~i

ON

in

4>

o

r-~

1^

^

t^

O

On

oo

OO

^

O

>o

NO

OO

o

r^j

(N

_>

:^

fN

r<-i

NO

r<-.

ro

r^

^

^

r-i

m

Tt

r«~,

-s:

•t^

o

^ — ■

"^

03

r^

NO

ro

-*

ON

00

•^

00

00

rj

'^

r)

c

II

o

s

lU

>/-,

On

o

00

O

irl

oo

*-H

o

oo

NO

ON

ro

NO

M-

o

c

Pi

a\

r<-j

1

o

ON

00

ON

1

NO

1

"^

1
ON

NO

1

00

1

1

1

NO

<^

^-N

a^

a

^

lO

lO

NO

ON

oo

t^

t^

m

ON

rn

lO

NO

r-

o

00

^

1

Cu

m

NO

f<-l

fN

^

I— 1

CN

M

^

CN

1

f^

5j

. »v

O

>.

!o

■4-*

>)

o

K

«

«<

s

•^

<u

f-

a.

o

rt

'^

r^

ON

'^

00

_

t^

r<1

m

ON

^^

r~-

o

rj

^^

m

NO

13

"S.

B

Xi

s

O

lO

00

ON

o

00

r^

^

On

m

NO

NO

0\

ON

00

C-h- 1

o

3

1—1

r<-i

r-

m

tN

1—)

'^

t— 1

fS

<N

en

(N

O

X

u

Z

C/5

C/3

•*~i

a;

D

^

C/3

o

o

Oh

c«

II

+

o B

4- I Q<

E' ^^ I I ^-v ' >^ (U ,

_ _ X)

o Vr « X "Si ^ ""^ f^o >- "^ -^ y O CJJ - _

•a .1=; o >- >> ^ 2 "C •:::
uoa.a,cuIc/520oo.5oauPHCu<P

12

BREVIORA

No. 342

CM

LU

_l
DQ
<

rt

lU

^

S

s

■♦-N4

"O

'»*-

•*.^

u

l-n

tJi

o

c

;>

05

s

a;

<i>

ftn'

\o

•a

'"^

O

o

I>1

(/3
(U

■4— »

o

a
i-i

to

•c

c

o

(Z)

-i-i

3

c^

X5

Ui

m

wo

oo

L-H

o
Q

>/^

o

m

m

>o

i-H

r-H

*-H

'^

'^

VO

T-H

*^

■^H

rs|

(N

^

vo

»— 1

T-^

<s

m

00

(S

i-H

-

O

O

>0

"O

^^

,-^

03

E

03
C

oi

s

C
03
0.

03

E

03
1— >

C3
Oh

" — '

t^

^

^

■^

:^

J^

O

O

^

-i^

■-^

o

^

l^

Ij

o

o

o

0

^

Q^

a,'

Cl]

<N ^ ^

O

Z

c

03

q

q

U

CO

fN

<^

S

1— 1

»—)

»— <

rvi

^

c

"— '

rj

y—t

0!

,_,

^^

<N

ec!

1— (

^

Tf

^-H

cn

00

i^H

<s

r-~

o

m

T"H

'^

r—l

^ ,

(S

d

as

o

ir>

^

r— (

a

a

03

E

E

03
C
03

03
C

E

a
1—1

Oh

^^

2

03

5

^

■»^

;-!

Cj

1^

V

"3

O

o

B

"5.

"5.

o

<

o

o

0

-S

cC

=c

a^

c

03

q

q

m

CJ

00

o6

00

s

"

a

as

as

c

-^

00

— <

03

r~-

—1

00

oi

T-H

c^

Tj-

>o

oo

o

o

o

-si-

T-H

r-

so

6

Z

o

o

■/->

>r^

^_^

^^

03

E

03

03

E

C3
C
03
Oh

'c3

c

C/5

E

u

^^

^ '

2

>i

in

03

ik

.O

O

O

2

'S5

o

-^

O

(U

Cl.

Cl.

O

CL

0

o

■»**»

■•^

-S

o

o

0^

a;

cC

c

03

q

'^

en

aj

od

00

ON

S

*— I

^-H

CT\

as

o

c

^^

^^

r)

03

VD

r-^

00

Pi

r-H

O

so

OS

O

1-H

>o

r^

00

r— *

CO

-'t

(N

r-

VO

1-H

VO

r— (

6
Z

a\

O

<ri

^

1-H

^-^

on

OS

03

o!

E

o

E

03

C

a

E

C3

03
C
C3

u

Oh

^-1

' — '

c

\_^

U^

5

^

5

5

»—
O

03

t^

O

2

o

rt

Ci.

Cl.

o

J

0

o

0

■S

a;

cC

a^

c

03
0)

c

03

as

00
CM

so

o
Z

o

>>

•a
o

1970

REDESCRIPTION OF P. OTOPHORUS

13

u
O

a

a
a
a

CO

GO

at

u

o

3
O

a

00

o

c

Oh

2

14

BREVIORA

No. 342

u

CO

o

Z

o

S

o
H

a

-a

c

a

•^^
a

><
w

c
o

CO

>
3

So

So

a

0,

(U

u
D
60

1970

REDESCRIPTION OF P. OTOPHORUS

15

1

^^^

=3 1

cz qj
V. ^

I 1

~ — p

•1

1

<o

n basis of n
errors on eit

■■

L

#

-

f

-

O

horu.
anda

■

^ o

-

s otop
d 2 st

—

©N

f

•

• -

an

\

•

-

\

...

c .2

\

N«| 0

o -

evia

■•*

\l

~l\ ~"

• -

100

tion of

lations of
standard d

p

\

\
\

•-

$-1
o

o

3

_.

o
o

o

^<

-

a,

d Jamaica
Range, m
St separatio

o

o

ol

—

bo

o

-

o

o
o

_

inama
il wid
ine of

oo

cd cd

—

o O

" i3 ;:=
^ IS 11

o

O Panam
# Jamaic

—

C/2

. Separation of

ngth and interor

are shown. AB

m (u e

■^ <a

aj ^ (U

"

= 1 e

.S" c '^

_J

1 III

1 o

1 1 1 1 1

. 11 . .

1.1.

1 1

O
oo

tU t« O

o

o
o

s?

§

■-3 '35

(is J^ uouJodoad) q;piM jB^iqaoaaiui

16

BREVIORA

No. 342

Figure 4. Panama collection localities: 1. Mouth of Chagres R., 2.
Chagres R., 1/2 km from mouth, 3. Chagres R., Tarpon Club Boat Basin,
4. Panama Canal, Atlantic Mine Docks, 5. Panama Canal, Atlantic Third
Locks.

' DO NOT CiRCUlU^^COMP. zoc:

LIBRARY

B R E V I O R^'A^^o

HARVARD
ITYi

Mmseiim of Comparative Zool

Cambridge. Mass. :M) January, 1970 Number 343

PALEOECOLOGY OF A LATE CRETACEOUS
VERTEBRATE COMMUNITY FROM MONTANA'

Richard Estes and Paul Berberian

Abstract. At least 78 species representing five classes of vertebrates
occur at the Bug Creek Anthills locality, in the Hell Creek Formation,
McCone County, Montana. The fauna closely resembles that from the
Lance Formation of Wyoming, also of late Cretaceous age.

Quantitative paleoecological techniques originally devised for late Ceno-
zoic mammals required only slight modification to prove useful when applied
to the late Cretaceous sample. The resulting analysis shows that at Bug
Creek Anthills, fishes, aquatic salamanders, and mammals are more abun-
dant than the amphibious or terrestrial turtles, dinosaurs, and (especially)
lizards, when compared with these groups in the Lance Formation fauna.
While the latter was probably deposited within the general environment of
a swamp forest with relatively small watercourses, the Bug Creek Anthills
fauna seems to have been laid down in the relatively deeper waters of
major rivers issuing from those lowland swamps. Gar and bowfin fish, and
elongated salamanders, are dominant species in the aquatic community
nearest to the site of deposition; some of the mammals seem to have been
part of a nearby riverbank community.

Excluding birds, 73 genera occur in the Cretaceous sample while 128
genera occur in its closest modern analogue, the Recent fauna of southern
Louisiana. If this comparison is valid, the fossil sample may represent
about two-thirds of the generic diversity present on the subtropical Cre-
taceous floodplain.

INTRODUCTION

Recent work by Dr. Robert Sloan and his associates has resulted
in recovery of a large collection of vertebrate fossils from a series
of late Cretaceous localities in the upper one-third of the Hell
Creek Formation, McCone County, Montana (Sloan and Van

1 Fossil vertebrates from the late Cretaceous Hell Creek Formation,
Montana: Contribution No. 9.

2 BREVIORA No. 343

Valen, 1965). The lower vertebrates from this part of the section
have been described in a series of papers listed by Estes, Berberian,
and Meszoely (1969). The fossil sample is from a quarry, and
the specimens were carefully concentrated by the use of washing
and screening techniques (McKenna, 1962). All fossil material
was removed and an attempt was made to identify each specimen;
it is thus possible to make comparison with the similarly-collected
sample from the Lance Formation of Wyoming (Estes, 1964), and
to apply techniques of paleoecological analysis developed by Shot-
well (1955). A substantial faunal similarity between the two
samples indicates a similar environment and inferred climate; es-
sential contemporaneity of the two localities further enhances the
paleoecological significance of such a comparison.

Bug Creek Anthills (BCA) is the richest of the Hell Creek lo-
calities on and near Bug Creek that were discovered by Sloan. A
collection from this locality made by a field party of the Museum of
Comparative Zoology with the aid of Sloan and his field crew was
used for the numerical analysis discussed below. The total faunal
list was determined by examination of collections made by a num-
ber of institutions (Estes, et al., 1969). BCA is a quarry site that
occurs in a cross-bedded stream channel within the upper one-third
of the Hell Creek Formation, about 80 feet below the coal bed
used locally as the arbitrary mapping boundary between the
Cretaceous and Tertiary (Sloan and Van Valen, 1965).

FAUNAL LIST

This list includes only species from BCA and other localities in
the section on Bug Creek; those not present in BCA are followed
by locality data in brackets. The list of mammals has been pro-
vided by Sloan, with additions by W. A. Clemens, Jr. (1969, pers.
comm.). With the further addition of the unidentifiable ornithom-
imid material cited by Osborn (1916), the ankylosaurid Ankylo-
saunis magniventris (Brown, 1908), and the panoplosaurid teeth
noted by Brown (1907: 842) this becomes the most complete list
of valid Hell Creek Formation vertebrates available. Lamna sp.
was also cited as present by Brown (1907: 842) but the single
specimen mentioned is no longer available to substantiate the
identification.

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 3

VERTEBRATES FROM THE BUG CREEK LOCAL FAUNAS,
HELL CREEK FORMATION

CLASS CHONDRICHTHYES
ORDER SELACHII

Family Hybodontidae

Lonchidion selachos Estes

ORDER BATOIDEA

Family Pristidae

Ischyrhiza avonicola Estes [Harbicht Hill]
Family Dasyatidae

Myledaphus bipartitus Cope

CLASS OSTEICHTHYES

INFRACLASS CHONDROSTEI
ORDER ACIPENSERIFORMES

Family Acipenseridae

Acipenser erucijerus Cope

Protoscaphirhynchus squamosus Wilimovsky [Rock

Creek, SEVi Sect. 36, T23N, R43E]
Family Polyodontidae

Paleopsephurus wilsoni MacAlpin

INFRACLASS HOLOSTEI

ORDER ASPIDORHYNCHIFORMES

Family Aspidorhynchidae

Belonostomus longirostris (Lambe)

ORDER AMIIFORMES

Family Amiidae

Amia jragosa (Jordan)
Amia (=zProtamia) sp.

ORDER AMIIFORMES?

Family Palaeolabridae

Palaeolabrus montanensis Estes

ORDER LEPISOSTEIFORJV4ES

Family Lepisosteidae

Lepisosteus occidentalis Leidy

INFRACLASS TELEOSTEI
ORDER ELOPIFORMES

Family Elopidae

unidentified genus and species

BREVIORA No. 343

Family Phyllodontidae

cf. Paralbula casei Estes

Family Albulidae

Coriops amnicolus Estes

ORDER PERCIFORMES

Family Sciaenidae

Platacodon nanus Marsh

TELEOSTEI incertae sedis

CLASS AMPHIBIA

ORDER CAUDATA

SUBORDER AMBYSTOMATOIDEA

Family Scapherpetontidae

Scapherpeton tectum Cope
Lisserpeton bairdi Estes

Family Prosirenidae

Prodesmodon copei Estes

Family Batrachosauroididae

Opisthotriton kciyi Auffenberg

SUBORDER SALAMANDROIDEA?

Family Amphiumidae

Proamphiuma cretacica Estes

SUBORDER MEANTES

Family Sirenidae

Habrosaurus dilatus Gilmore

ORDER SALIENTIA

Family Discoglossidae

Scotiophryne pustulosa Estes

cf. Barbourula sp.

undescribed genus and species
Family Pelobatidae?

Eopelobatesl sp.

CLASS REPTILIA

ORDER EOSUCHIA

Family Champsosauridae
Champsosaurus sp.

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY

ORDER TESTUDINATA

Family Baenidae

Eiihaena cephalica Hay

undescribed genus and species 1

undescribed genus and species 2
Family Dermatemydidae

Compsemys victa Leidy

Adocus sp.
Family Testudinidae
Subfamily Emydinae?

unidentified genus and species
Family Trionychidae

Trionyx sp.

ORDER SAURIA

SUBORDER SCINCOMORPHA

Family Teiidae

Chamops segnis Marsh

Leptochamops denticulatus (Gilmore)

Haptosphenus placodon Estes

Peneteiiis aqiiilonius Estes
Family Scincidae?

Contogenys sloani Estes

SUBORDER ANGUIMORPHA

Infraorder Diploglossa
Family Xenosauridae

Exostinus lancensis Gilmore
Family Anguidae

Pancelosaurus piger (Gilmore)
Family Parasaniwidae

Parasaniwa wyomingensis Gilmore

Paraderma bogerti Estes
Diploglossa incertae sedis

Colpodontosaurus cracens Estes
Infraorder Platynota
Family Varanidae

Palaeosaniwa, cf. P. canadensis Gilmore
Sauria? incertae sedis

Family incertae sedis

Cuttysarkus mcnallyi Estes

BREVIORA No. 343

ORDER SERPENTES

Superfamily Booidea
Family Aniliidae

Coniophis precedens Marsh
Family Boidae

unidentified genus and species

ORDER CROCODILIA
SUBORDER EUSUCHIA

Family Crocodilidae
Subfamily Crocodilinae

Leidyosiichiis sternbergi Gilmore

Thoracosaurus neocesoriensis (DeKay) [SW'/4

Sect. 28, T24N, R44E]
Subfamily Alligatorinae

Brachychampsa montana Gilmore

ORDER SAURISCHIA

SUBORDER THEROPODA
INFRAORDER COELUROSAURIA

Family Coeluridae?

unidentified genus and species
Paronychodon lacustris Cope

THEROPODA? incertae sedis

undescribed genus and species

INFRAORDER DEINONYCHOSAURIA

Family Dromaeosauridae?

unidentified genus and species

INFRAORDER CARNOSAURIA

Family Tyrannosauridae

Gorgosaurus lancensis Gilmore? [Bug Creek West;

Harbicht Hill]

Tyrannosauriis rex Osbom [NEi/t Sect. 7, T22N,

R43E]

ORDER ORNITHISCHIA

SUBORDER ORNITHOPODA

Family Hypsilophodontidae

Thescelosaurus neglectus Gilmore
Family Pachycephalosauridae?

unidentified genus and species
Family Hadrosauridae

Anatosaurus copei Lull and Wright

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 7

SUBORDER CERATOPSIA

Family Ceratopsidae

Triceratops prorsus Marsh

CLASS AVES

ORDER CHARADRIIFORMES

Family Cimolopterygidae
Cimolopteryx sp.

CLASS MAMMALIA

SUBCLASS ALLOTHERIA

ORDER MULTITUBERCULATA

Superfamily Ptilodontoidea
Family Ectypodidae

Cimexomys minor Sloan and Van Valen

Mesodma formosa (Marsh)

Mesodma thompsoni Clemens
Family Ptilodontidae

Cimolodon nitidus Marsh

Essonodon browni Simpson
Superfamily Taeniolabidoidea
Family Cimolomyidae

Cimolomys gracilis Marsh

Meniscoessus borealis Simpson
Eucosmodontidae

Stygimys kuszmauli Sloan and Van Valen
Family Taeniolabididae

Catopsalis joyneri Sloan and Van Valen

SUBCLASS METATHERIA
ORDER MARSUPIALIA

Family Didelphidae

Didelphodon vorax Marsh
Alphadon marshi Simpson
A. wilsoni Lillegraven
Pediomys elegans Marsh
P. krejcii Clemens
P. hatcheri (Osborn)
P. cooki Clemens
P. florencae Clemens
Glasbius intricatus Clemens

BREVIORA No. 343

SUBCLASS THERIA
INFRACLASS EUTHERIA
ORDER DELTATHERIDIA

Family Palaeoryctidae

Cimolestes incisus Marsh

C. magnus Clemens and Russell

Procerberus jormicarum Sloan and Van Valen

ORDER INSECTIVORA

Family Leptictidae

Gypsonictops hypoconus Simpson
G. illuminatus Lillegraven?

ORDER PRIMATES

Family Paromomyidae

Purgatorius ceratops Van Valen and Sloan [Har-
bicht Hill]

ORDER CONDYLARTHRA

Family Arctocyonidae

Protimgulatiim donnae Sloan and Van Valen
P. spp. [Bug Creek West; Harbicht HiU]

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 9

INTERPRETATION OF THE FAUNA
Introduction

The Bug Creek Anthills local fauna from the Hell Creek Forma-
tion differs only in detail from that of the Lance Formation (Estes,
1964).^ There are few Lance species that do not also occur in
the BCA sample, but the latter includes some additional forms of
special interest as well as some significant differences in relative
abundance (Table 1) from Lance forms (see also Estes et al.,
1969: table 1).

Because time of deposition of the upper parts of both the Lance
Formation and the Hell Creek Formation was essentially synchron-
ous, because their lithologies are similar and indicate similar
coastal floodplain environments of deposition, because both local
faunas bear a similar relationship to late Cretaceous mid-contin-
ental seaways, because their vertebrate faunas are very similar, and
because the two samples were carefully collected using the same
techniques, an unusual opportunity is provided for comparing
these samples.

The Bug Creek sample was carefully collected by the use of
washing and screening techniques, all fossil material was retained,
and an attempt was made to identify each fossil fragment; it was
therefore possible to apply the community analysis method devised
by Shotwell (1955, 1958). Since only a few aspects of Shotwell's
method were applied to the Lance Formation sample by Estes
(1964), we have used American Museum of Natural History
Lance Formation samples from V5620 and V5711 to make some
additional comparisons for this study.

The time spent in sorting and identifying the thousands of
specimens used in this study (including collections from all institu-
tions) has been the major factor delaying completion of this series
of papers. However, we agree with ShotweU (1963: 9, in reply
to Wilson, 1960: 9) that careful sorting, however time-consuming,
is a procedure necessary to any quantitative study and that the
time involved in this activity is not a valid objection to his method.
Shotwell developed his method for analysis of late Cenozoic mam-
malian faunules; it is our intent to test its utility when applied to
late Cretaceous lower vertebrate samples. Voorhies (1969) has
offered some cogent criticisms of Shotwell's method. Some of these

1 Comparisons with the Lance Formation refer only to the two major
localities V5620 and V5711 (Clemens, 1963).

10

BREVIORA

No. 343

UJ

_i

CQ
<

I-

a

6

u
O

u

^ (50

^ s

S

o

O

(U

o
C

C -

^>

cd O
> fc/J

C4-I
(/5 '■(->

on K3

0)

(U Cl,

B
o
o

s

o

O
a

C

c3

C

3
JD

d
o

<S

5

X

■v.*

Q<

a

c

o

li.

!U

■^

O

^

G

>-~;

''^i
■<

o

(J

■♦»*

-Nh O OJ

O u-l O <N

l^' O -o en

CO CO r~
o

<N

CO CN

C<-) lT)

c^i

00 00 -o t^ lO 00 r--

CN

LT)
CO

o '-^ o o Tt-_ •^^ ^_
' LT) ' CN CN od r--'

c^i

LO

CO so O

LO O

o so o o o ■^

t^

CN

00

^ o

LO

CN

oo

fsj so -si- r-.

CO T^

SO
SO
00

CN
O

o o

CO O

CO so
00

OsoOOOOOOOOOOOO

ooLnr^r~-ooor-or~ocNCNt^

■^T-.CNCNCNJ.^CN-^CN

o o ^

O CO Tj-
r-i CN t^

so CO r-. ^

— o

'^ -^ CN

SO CN

C\

r^

o

r-

CN

^— •

CN CN

O ^

(^1 •^ (Nj U-) -*
" CO ' cc^

O (^I CO CO O SO
v-<' " ■rt Tj-' oo'

[~^ CO T^ "^ ^

CN

LO CO

CN r-

o

U

u

c
o

3 03

-a -a ^

00

3
u

3

x:

CO

D

00

3

3
X)

C

a,

03

(u o

o

■§ °

< CQ

■*t! fr^ C^

00

n.

o

4J O . •

3
;-.

X)

"o

u a. W

O

c
o
■o
o
u

03
rt

c c

o o

•r OJ

i: a

■s x:

c« n.

'5. «

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 11

[~-, I ^^ ^^ ^^ ^^ ^^ ^^ ^ ^ ^ (Si I (V] (N) CS C<^ I "^ ^. t^.

'—(N I ^^y-iT-l-^-^T-lT-ly-iT—CO I LDinUICvq I TJ-Tj-fSl

Ln r~- in o t~- r^ vo

1 iOcncs(NcOTt-'<^CMTf0000

iri sd r-^ 00 .^ V <N

' '^^ un T}-" ^ T-< T-i ^ ^' V CN)

V

r-- 00 r~- I ^

r-i T— I ' CM

Cv

CO o

CO r~- o SO

1 cv CO r--

r-

■O -^ M- LO CO 00 CO

1 1^ 00 CN

1 '^

m in

CO ^ \0 (N
iri

' oo' u-i Tt

^
^

CM ■^ T— 00 CN 00

' ^ so

' so'

o

o

1 "=>

o

o

1 o

'^

CNJ

O)

1 (N

Csj

CO

1 CO

>

■^

1 — 1

"-"

■<-'

T— *

■^

OloOOOOO I oooooooo

[--r^OOsOsOC^ I OCsCACnCsOOcO

uoo"^(N'-'OGir) I r-~o>-oi-OLnoO'-''-<'^o I cN'-'Cs

I^-^T-icO l^^t-iOOCO LOiriT-HcO '-'I

■^rOCOCsCNsOCg I (NOgrfsocOfMCNCOCslCNTj- | r^CMvO

CNCMCNIsO^^cOt— c I T— t-HcO-^cOr-iT— cCMt^'^rO

I Ifiriili tllll lllllUll

12

BREVIORA

No. 343

eo

^

^

-CS

ftj

a

?

V.

O

li<

J^

«-N

o

:^

<3

:§

^s

CO

.0

^

<3

U

5

1-

0

u,

•ie

u

S?

k^

0

:^

"^

ijj

^

tci

&?

•*■*

CO

•s

:^

c

^

^

<u

%>

K.

0

^

60

aj

s

cq

DC

^

^

CN CM CO rl CO

o o

■^ 00

co'

CO CM VD

ON r- tn

O vO CM
CM CO

10

CO

CM 00 ■<-i

Cn CO 10 10 o

-*" cm' t^ t^ »-<

00 CM "^
CO

' CO T-i CM CT^" 00 T^ CM t~-

■>-■ CO 10 un •.-•

in (M ^

r~- -^ 0

CO (N T^ ■^ CO

^" ^ ■

cm' vO

00

CI, Oh Ch !1h

LO SO 00
VO NO
CM

O CN 10 CO

so Tt

00000

CO CO CO CO CO

lO CO <N CM

000
cO 10 >0
^ CM Cs]

CN 00

O^^CUCLiD-iCLiOO'^OOOOCOOO

Looooor-ooi^ooo

cOOCLiD-iCl-iCLi'^^'OCMr--sO

crs CM

r^CMCMCM'^ I COCMCM I CM^OOCLhOh&hD-i

CM

uo >-• -^ T-i CO

o

CM

00 O Oh Oh Oh Oh

Tt T-H

co_ o '— _ CM a CO

' ^' T-'

CM CO 00 '-' CO CM
CM

^Q

'S

a

(U

B

-o

f3 03

■3

3

a
(/I
o

c
o
•a

lis

o .ti
U J

CO

2, .2

D O

O O

w 3

" 2

C3

e

« O

PQ

E o b

t«

o "S "^

So c -5 o •

CO

3
J? en

O o

c
o

O

c
o

o <u

CO

o

•H|lliJl|'gl.-i:|-|

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 13

O ld *— ' ^— '
oi ' ' "

o
o

^ un

00

CO (L,

r-. CL, 1

1 '^. ^.

VO '

' vd o

:3: cu
o

t-~ e^

o o

^ 00

CO

■^ CS

CN

T— 1 T-H

vO

00

I~~-

o
o

in CL,

^ o

o

o ^

<L)

c

o (u
•a xi

3 rt

b ^ J 6 u

00 O ^ O 1)

c« u X '-' r;

^>> . £ « 13 2 t: o
O u oa H eu G, H

Ml

o

o

4:

s

o

S

o
vo

o

a.
o

«l .^

>

lU

•s

ca
x>

CO

CO

O

>> ii

c

a
^^ o

I— X

Z i2

C «J 3

O on CO

= - - 2 >>

E ^

>^ c) CO

c«

IS

lU

or)

3

CO

(L>

"H.
S

ca
t«

N
U

s

a
u

E

(L>

C3 o .5

CC C3 C«

3
T3

>

•5

c

c ^
dJ .-

E i:
c

^

o

a.

C '^

c

3 C "O

C 3 " "

C 03

3
C

T3

3 <D >- ii

.S — c -S

6" u 3 to

»- C (U

C3 «i

c

CO O

.X II II II II

■5 5 < ^ W

> 5 P^ ^ Z

X)
X

<

a

CO

c
o

c

CO

C
<u

E

o

(L>

CO
CO
(D
C

X D

2 E
o

u

CO

3

i

CO

3

T3
1)

CJ

dJ

u
O

CJ

>

c« tj

03

•T3

(U

t-i

(U

>

o
o
<u

o a
c •-

■s W
3 i,

.^ >> CO

- •- &
O

5 ^ M

ti ^ •-

03

w -S T3

CO D 3
"^ - CJ

■:5 '^

CU CJ

i:i c« a

c
o

g> U 03 03

U

u

14 BREVIORA No. 343

are noted below (pp. 21-22) with a discussion of their relevance
to our study. ShotwelFs terminology is used throughout this
paper, and the terms are defined below (pp. 19-20).

General Comparisons
Fishes

Excepting Squatirhina americana, all Lance Formation sharks
are present in the Bug Creek localities, although they are much
rarer there. Ischyrhiza avonicola is represented only by a single
tooth (from the Harbicht Hill locality), and Lonchidion selachos
is represented only at Bug Creek Anthills (BCA) and by very few
specimens. The best represented elasmobranch present is the ray
Myledaphus bipartitus; this is perhaps a parallel for the fact that
rays and skates are more common today in fresh waters than are
other chondrichthyans. The low overall abundance of these forms
at Bug Creek Anthills probably indicates a restriction of marine
access relative to that of the major Lance Formation localities
(Estes, et al., 1969: 28).

Bony fishes are represented by a more complete series of ele-
ments at Bug Creek Anthills than in the Lance sample; also, the
state of preservation of the bone is generally better and there is
less rolled and waterwom material. As a percentage of total
minimum number of individuals, including mammals (Shotwell,
1955, etc.), bony fishes are less abundant than in the V5620
Lance sample (BCA 20%, Lance 21%). All Lance fishes also
occur in the Bug Creek local faunas, and Protoscaphirhynchus,
Palaeolabrus and cf. Paralbula are additional bony fish present at
Bug Creek. The c/. Paralbula specimens do not necessarily indi-
cate proximity of the nearshore marine environments in which
this fish is usually found; more probably these few, poorly pre-
served specimens were transported into the Bug Creek area by
some other contemporaneous animal.

In combination, the bony fishes (at least the primitive ones)
indicate resemblance to the primitive fish fauna of the Mississippi
River drainage today, especially in its lower reaches near the Gulf
Coast (Estes, 1964: 166). Acipenser, Polyodon, Lepisosteus, and
Amia all occur now in that area (See Table p. 23), as well as
in the Wyoming and Montana fossil samples.

Amphibians

As for the bony fishes, the BCA salamanders show greater di-
versity than those of the Lance local fauna; all species found in

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 15

the latter are present and Proamphiuma occurs as well. Skull
elements are somewhat better represented than in the Lance
sample and most of the salamander bones give less indication of
transportation wear. On the other hand, few limb elements of
salamanders are found in the Bug Creek localities, in contrast to
their relative frequency in the Lance sample. By analogy with
Recent salamanders, the neotenic skull elements and vertebral
form of the fossil salamanders indicate that all genera were fully
aquatic (Estes, 1964), possibly excepting Prodesmodon.

In comparing relative abundance of Bug Creek species with
that from the Lance Formation (Tables 1, 2; Figs. 1, 2), a sig-
nificant increase for Lisserpeton and Opisthotriton is apparent, as
is an equally significant and much reduced frequency of Habro-
saurus. Although rarity of Prodesmodon makes comparison less
meaningful, there seems to be a significant decrease in abundance.
Scapherpeton shows an increase of a less striking magnitude. The
total relative abundance (Table 1) of salamanders at Bug Creek
Anthills (35.5%) is significantly greater than at V5620 in the
Lance Formation (21.4%). The two quarries indicate great
similarity in the manner of their deposition and this difference
probably has ecological significance, as suggested below.

Sirenid and amphiumid salamanders resembling the fossil forms
still live in the southeast United States today, while the Cretaceous
frogs indicate relationships to now extinct or rehct groups: dis-
coglossids are now Palearctic and Oriental, and the Bug Creek
forms show relationship to genera from both of these areas.
With the possible exception of the Alytes-Yikt form, all Bug Creek

TABLE 2

Comparison of minimum number of individuals and relative
abundance of salamanders in two late Cretaceous locaUties.

V-5620,

Bug Creek

Anthills,

Lance

Formation

Hell Creek

Formation

MNI

RA

MNI

RA

Habrosaurus

71

49

9

4

Scapherpeton

22

15

61

24

Lisserpeton

1

1

72

28

Opisthotriton

41

28

107

42

Prodesmodon

11

7

2

1

Proamphiuma

0

0

2

1

146 100% 253 100%

16 BREVIORA No. 343

frogs app3ar to have been aquatic or semiaquatic, judging from
their modern counterparts. The frogs are similar to those from
the Lance Formation in being primarily discoglossid and in having
genera in common (c/. Barbourula, Scotiophryne, Eopelobates?
sp.).

Turtles

All Lance turtles identified by Estes (1964) are represented at
Bug Creek, except that Adocus seems to have replaced its close
relative Basilemys, and undescribed baenid 2 does not occur in the
Lance local fauna.

The baenids are the most abundant of the turtles; unfortunately
they have no modem relatives and their ecological requirements
are unknown. Both the baenids and the emydine-like testudinid
were probably amphibious-aquatic types.

The other turtles are fragmentary and poorly represented. The
trionychid material indicates permanent waters, either large rivers
or streams, or large quiet bodies of water with soft mud or sandy
bottoms.

Champsosaurs

Champsosaurs were aquatic, fish-eating reptiles that perhaps
filled niches analogous to those of the Recent garpike or the gavial.

Lizards

Diversity of the lizard fauna from the Bug Creek Anthills local-
ity is about the same as in the Lance Formation. The Lance forms
Meniscognathus , Litakis, cf. Gerrhonotus, and Sauriscus do not
occur, but these genera are very rare in the Lance Formation lo-
calities. In comparison with the latter, relative abundance of
lizards at Bug Creek is strikingly reduced and the material is
poorly preserved. Contogenys replaces Sauriscus as the scincoid
in the fauna, and the highly specialized teiid Peneteius makes its
first appearance. Only the anguid Pancelosaurus (Meszoely,
1970) is as abundant as it was in the Lance sample (approxi-
mately 2-3% of total individuals represented in the sample), sug-
gesting that it may have been amphibious or aquatic.

Snakes

The oldest known North American snake, Coniophis (Lance
Formation, Wyoming), occurs in a number of other North Ameri-
can Upper Cretaceous, Paleocene, and Eocene deposits. With

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 17

the exception of Helagras from the Puerco and Torrejon Forma-
tions, no other fossil snaice has been known to occur in North
America until the Eocene.

The presence of a boid snake at Bug Creek Anthills is indicated
by a single vertebra. This snake may have been related to the
burrowing (or cryptic) Oligocene to Recent subfamily Erycinae.

Crocodilians

Presence of both crocodiles and alligators in the Bug Creek local
fauna recalls the situation in subtropical Horida today, in which
both Alligator and Crocodylus are represented. Jepsen (1963:
680) has suggested that crocodihans occupied ecological niches in
the past that are filled today by large terrestrial mammahan carni-
vores. Analysis of stomach contents of Recent crocodihans (Kel-
logg, 1929) reveals that only about 6% of their food is mammal-
ian. If Kellogg's study can be used as a criterion, it is hkely that,
as today, mammals did not form a significant part of the diet of
crocodilians in the past. Since the primary food of large mammal-
ian carnivores today is other mammals, the niche relationships of
crocodilians and large terrestrial mammalian carnivores are not
necessarily comparable. Assuming the food habits of crocodihans
to be conservative, it would be more logical to assume that large
mammalian carnivores (themselves preying on mammals) evolved
with their prey. This assumption can be correlated with, and
receives support from, the late Cretaceous and early Cenozoic ap-
pearance of the significant predators of mammals: snakes, large
varanid lizards, some birds, and of course, the carnivorous mam-
mals themselves.

Dinosaurs

Although similar in diversity to those from the Lance Forma-
tion, the dinosaurs at BCA are represented by a lower relative
weight of specimens (see below, and Table 4). Tyrannosaurus is
not present at BCA, and has not been found in the upper 100 feet
of the HeU Creek Formation. We have used the name Gorgo-
saurus rather than Dryptosaurus here (c/. Estes, 1964) although
firm generic identification cannot be made on the basis of isolated
teeth. Sloan (1969, pers. comm.) states that all Hell Creek
Formation ceratopsians are referable to Triceratops prorsus, and
all hadrosaurians to Anatosaurus copei.

18 BREVIORA No. 343

Methods of Analysis

Paleoecological analysis of any large sample of a diverse fossil
vertebrate assemblage may yield useful information about past
inter- and intra-community relationships, and contribute to an un-
derstanding of community development through geologic time.
Such analysis must take into consideration factors that elucidate
taphonomic changes (Olson, 1966) between the original com-
munity structure (life assemblage) and the death assemblage that
the fossil sample represents. Some death assemblages preserved
in or near their natural habitats (e.g., those of pond deposits), may
differ little from their counterparts in life. Most fossil vertebrate
deposits, however, result from deposition by moving water and
thus may have their original composition modified by a variety of
sedimentary, biological, and ecological factors. The weight dif-
ferences between compact and spongy bone tissues, the size of
bony elements, the amount of preservable tissue available in an
individual, and the relative fragihty of the skeletal tissues may
interact with current speed to modify the original composition of
the life assemblage. Sedimentary sorting may thus produce a
sample radically different from the original life composition. Prox-
imity of the life zone of the animals to the eventual area of depo-
sition of the fossils is also a factor modifying the assemblage,
and, as Olson points out, deposits formed by moving water may
"tap a variety of life zone along the course of the stream and its
tributaries."

Shotwell (1955, 1958, 1963) devised a quantitative method by
which environmental reconstruction of some late Cenozoic mam-
mal communities could be made. Many late Cenozoic vertebrate
fossil samples are predominantly mammalian. This is the result of
collecting bias in some samples, but other factors may apply.
For instance, deossification has progressed through time in many
lower vertebrates, resulting in fewer, more delicate bony elements
in their descendants. Also, the relatively coarse sediments in
which many late Cenozoic vertebrate samples occur imply rapid
currents of deposition and a high rate of erosion of delicate ele-
ments.^ Another factor is ecological; temperature and humidity
have decreased in North America and Europe since the middle

1 Bulk collecting methods may produce a large lower vertebrate fauna;
thus Wilson (1968) collected a Pliocene fauna with 39 lower vertebrate
and 34 mammalian species. Also, Shotwell (1969, in litt.) indicates that
his recently collected samples "are primarily lower vertebrates, at least
those occurring in fine-grained sediments [emphasis supplied] ."

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 19

Ccnozoic, and areas that were once humid, subtropical, and
tropical are now temperate and more arid. Many lower verte-
brates have retreated south in response to these changes, remaining
with their temperature optima and evolving quite slowly or be-
coming extinct; many temperate niches thus became available for
capture by the rapidly evolving mammals.

The community analysis method devised by Shotwell (1955) is
summarized in his 1958 article. We will not recapitulate his
method or findings here, but the major precepts of his analysis will,
we hope, emerge in context below.

The first parameter obtained from a community analysis utilizing
the Shotwell method is the minimum number of individuals (here-
after referred to as MNI). The MNI itself, reflecting the fewest
number of individuals required to produce the fossil elements, is
a rough indication of abundance, but it is more useful as a percent
of the total MNI of the sample: this is the relative abundance
(RA). The relative abundance alone is a ready means of com-
parison of samples, and one that is not difficult to calculate; Estes
(1964) used it to show the constancy of vertebrate abundance in
numerous Lance Formation localities, Wilson (1960: 9) sug-
gested that a percentage "derived from numbers of specimens, with
perhaps some subjective adjustment for other considerations" could
be as useful as Shotwell's method. To us, percentages calculated
from the actual numbers of individuals present have a clear bio-
logical advantage over those derived merely from numbers of
specimens, especially for lower vertebrates. Bones of the latter
are often common, particularly if an extensive dermal skeleton is
present. Sedimentary or erosional factors, such as those oper-
ating postdepositionally in "blowout" accumulations, may selec-
tively act on bones, and the number of elements, especially resistant
ones, is thus much less hkely to reflect actual abundance. Never-
theless, "accuracy" of the Shotwell or the Wilson method is not
demonstrable for a fossil sample. Confidence in either method
can only be suggested by the use of theoretical models or collec-
tion of material from modern sediments and comparison with
existing abundance; these data are, in general, difficult to obtain
and not yet available (Van Valen, 1964: 109).

As Shotwell has pointed out (1955: 331) different mammal
groups vary widely in the number of elements that can be con-
tributed by one individual. This difficulty is magnified when
lower vertebrates are considered. Some species are wholly or in
part cartilaginous or show strong ontogenetic variation in ossifica-
tion. Others may enrich the sample (and perhaps give a false
impression of abundance) by contributing mainly shed teeth from

20 BREVIORA No. 343

an essentially continuous replacement process. In addition, be-
cause of the great antiquity of the fauna, some animals may con-
tribute more bones than their Recent relatives because of decrease
of ossification in, and loss of elements by, the modem forms.
Therefore Shotwell's corrected number of specimens (CNS) (see
below), a correction necessary because of group variation, is
almost certainly a more variable factor as we have used it than
it is for late Cenozoic mammals. We have given the estimated
number of contributing elements (ENE) for the various taxa,
basing it on the number of preservable, identifiable bones in mod-
ern relatives or in known complete fossil material. These esti-
mates are necessarily approximations but are probably correct
within the context used here. When Shotwell (1958: 273) made
corrections in his 1955 figures, he noted that "the results are httle
difl'erent from those presented earlier," and we have made calcu-
lations with similar results. The estimates vary from animal to
animal for reasons noted at the beginning of this section. Thus
for the ray Myledaphus we include a figure only for teeth and
precaudal vertebral column as the only recognizable, preservable
structures, yet even this figure approaches five hundred, whereas
for mammals, in which most bones are identifiable, the total figure
is only about one hundred and ninety.

The total number of specimens for the species (NS) and the
estimated number of elements (ENE) yields a corrected number
of specimens (CNS) when formed as a ratio:

CNS = number of specimens X 100

estimated number of elements

(Shotwell, 1958: 272-273). The relative completeness (RC) for
each species is then determined according to Shotwell (1955: 332).
This is the corrected number of elements for each individual:

RC = CNS

MNI

The RC is the parameter that is used to determine the closeness of
an animal to the site of deposition and thus can be used as a
measure of whether the animal belongs to a "proximal" or "distal"
community (Shotwell, 1955: 330). In a wind-deflation locality
("blowout") the RC is less vahd because erosion causes delicate
elements to become unrecognizable in a short period of time. In
such blowouts, the RC may become a measure of animals pos-
sessing the most numerous durable elements (Estes, 1964: 151-
152). This is not the case with Bug Creek Anthills, however,
which is a quarry.

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 21

Table 1 gives MNI, RA, NS, ENE, CNS, and RC for each
species identified in the MCZ sample from Bug Creek Anthills;
MNI and RA from Lance locaHty V5620 are inckided for com-
parison. A few rare taxa are not included since they were not
recovered in the MCZ sample (cf. pp. 3-8).

Finally, it is necessary to discuss the criticisms of Shotwell's
method offered by Voorhies (1969), in order to justify our use of
it here. Voorhies discussed taphonomic processes that produced
an accumulation of Pliocene vertebrate fossils in Nebraska, and
emphasized the lack of sedimentary data used by Shotwell in his
analysis. Shotwell assumed that attritional mortality produced the
quarry accumulations in Oregon on which he developed his
method; he could thereby interpret the minimum numbers of in-
dividuals as being representative of actual abundance in the Plio-
cene. Voorhies, on the other hand, suggested (1969: 54) that
stream sorting of bones caused many of the differences between
quarries that Shotwell ascribed to community differences. The
fact that the Pliocene quarry accumulation in Nebraska was the
result of mass mortality (probably winterkill) seems to have gen-
erated many of Voorhies' criticisms of Shotwell's method.

In spite of this and other objections raised by Voorhies, we feel
that the minimum number of individuals recovered at Bug Creek
Anthills and at the Lance Formation localities is, in general, rep-
resentative of former life abundance. Our reasons for this belief
are hsted below but are specific to the localities studied here; each
fossil accumulation must be judged separately. Techniques such
as those proposed by Shotwell must be apphed to a variety of
fossil accumulations of diverse ages before the overall utihty of the
method for environmental reconstruction can be determined.

1. Estes (1964: 151-158) demonstrated the essentially uni-
form relative abundance of individuals in Lance Formation
localities, and Estes et al. (1969; this paper, pp. 10-13)
have shown that Lance relative abundance is similar to that
at Bug Creek Anthills. In addition, the large BCA collec-
tions in various museums do not differ materially in relative

abundance inter se.

2. The Lance and Hell Creek samples are composed of rela-
tively small animals, in contrast to those studied by Voorhies
and Shotwell. The common animals in the Cretaceous
samples are represented by almost all of the principal skele-
tal elements. Although sedimentary sorting has modified
relative numbers of some bones, their varied shapes and

22 BREVIORA No. 343

sizes seem to indicate that sedimentary factors have not been
the only influence on accumulation. With rare exceptions,
selective preservation of uniform, durable elements (Estes,
1964: 157) does not appear to have modified relative abun-
dance significantly either at V5620 or at BCA.

3. Abrasion of bony elements of members of the aquatic
"proximal community" is significantly less than that for
members of the various "distal communities." As in Voor-
hies' sample (1969: 57), a range of fresh and abraded ex-
amples of each bone can be demonstrated at BCA, reflecting,
probably, contemporaneous reworking of channel deposits
as noted by Estes (1964: 159) for the Lance Formation
sample,

4. Bones of larger animals such as crocodilians, large turtles,
and dinosaurs are reduced in numbers by sedimentary phe-
nomena at BCA; the slow current speeds failed to transport
the large bones. Except for baenid turtles, however, all
BCA material of these groups is more abraded and frag-
mented than that from Lance Creek, where a similar paucity
of large elements also occurred. This is one factor leading
us to suggest that BCA was deposited farther from a riparian
habitat than was V5620.

5. While larvae or young are poorly represented at both Lance
and Bug Creek localities, a wide range of body size does
occur. The greater number of small rather than of large
individuals of many common species seems to indicate a
natural proportion rather than a sedimentary reworking,
since element size of small individuals is well below the
median specimen size.

6. The large size of both Lance and Hell Creek samples re-
inforces the above conclusions.

Admittedly the above factors are subjective or hard to quantify.
Nevertheless, they seem to suggest that attritional rather than mass
mortality was operating to produce the Lance and Hell Creek
samples, and that the relative abundance of fossils can be assumed
to bear a reasonably close relationship to actual life abundance.

Completeness of the Sample

Shotwell (1955: 329) compared diversity of his Phocene mam-
mal faunas with that of the Recent fauna of the United States as a
whole as well as with average figures for some Recent local faunas.

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 23

His comparison indicated a reasonable similarity in diversity and
he used this as an indication that his quarry methods had sampled
the Pliocene fauna rather well (about 85%, if the comparison is
justified), although a better comparison might have been made
with a fauna ecologically and climatically more similar to his fossil
fauna, perhaps from an area in Africa. Ecological interpretation
of the vertebrate fauna from the Lance Formation (Estes, 1964)
apphes equally well to that from the Hell Creek Formation, with
some differences in detail discussed elsewhere in this paper. From
both faunal and climatic standpoints, the closest modern analogue
of the two fossil faunas can be found in the lower Mississippi
River drainage and in similar areas on the Gulf Coast of the United
States. Since the Mississippi drainage is related to the former
path of some of the major epicontinental seaways in North Amer-
ica during the Cretaceous, the similarity of the fossil and Recent
faunas is not surprising.

TABLE

3

Generic

Diversity

of Cretaceous and Recent Faunas*

Lance genera

Bug Creek genera

Recent genera

chondrichthyans

4

3

3

chondrosteans

2

3

3

holosteans

4

5

2

teleosteans

3

4

34

(total fishes)

(13)

(15)

(42)

amphibians

salamanders

5

6

10

frogs

4

4

5

(total amphibians)

(9)

(10)

(15)

reptiles

turtles

6

7

11

eosuchians

1

1

0

lizards

14

12

5

snakes

1

2

18

crocodilians

2

2

1

dinosaurs

10

8

0

(total reptiles)

(34)

(32)

(35)

birds

5

1

149

mammals

14

16

36

total genera

75

74

277

total genera remo

ving birds

70

73

128

*The Recent fauna is that of southeastern Louisiana, the delta region of
the Mississippi River (Blair, et al., 1968; Lowery, 1960; Bishop, 1947).

24 BREVIORA No. 343

Table 3 compares the number of vertebrate genera for the two
Cretaceous localities with that for the Recent Mississippi drainage
area in southeastern Louisiana. Birds are included, although they
must be removed in the final analysis (Shotwell, 1955: 328) in
order to make the results more comparable. Table 3 shows several
obvious major differences from the modem fauna: (1 ) relative ab-
sence of teleosts, snakes, and birds, (2) presence of dinosaurs, and
(3) primitive nature of the mammals. Teleosts present in the mod-
ern fauna are mainly Clupeiformes, Cypriniformes, and Perci-
formes. The former group is not represented at Bug Creek but was
common in contemporaneous seas. They would be difficult to recog-
nize on the basis of disarticulated remains and may occasionally
have penetrated into the freshwaters of the Cretaceous as they do
today. Cypriniformes are not known before the Eocene. Perci-
formes first occur in the Cretaceous (Estes, 1964; Estes, et al.,
1969) but they are not common until the Eocene. Relative lack
of diversity of these two orders in the Cretaceous samples is thus
probably in part an evolutionary rather than a sampling phe-
nomenon.

The dinosaurs contribute notably to the Cretaceous fauna but
are of course absent from the Recent fauna, where their niches as
large herbivores and carnivores have been filled primarily by Artio-
dactyla and Carnivora.

The relative absence of snakes from the Cretaceous fauna is
probably of evolutionary origin; most Recent and Cenozoic snake
diversity occurs within the Colubridae, a family not yet known
before the Miocene.

Only one bird is known from the Bug Creek Anthills site; Brod-
korb (1963) has noted the presence of five genera in the very
similar Lance local fauna. Even so, the presence of approximately
150 genera of birds in the Recent fauna of southeastern Louisiana
forms the greatest contrast of fossil and Recent faunas. The habits
of birds and the dehcacy of their bones are probably partial reasons
for the relatively small number of known late Cretaceous birds.
However, as Brodkorb points out (1963: 70), the known Creta-
ceous forms (while well-distinguished) indicate that the characters
of the modern orders are merging, and again much of this dis-
crepancy between modern and fossil fauna must be an evolutionary
one.

The difference in mammalian fauna seems also to be of evolu-
tionary origin, since the Cretaceous fauna belongs to six primitive
orders. The overall discrepancies between Recent and fossil groups

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 25

are easily seen, and when the difference in evolutionary develop-
ment in fislies, snakes and birds is taken into account, the fossil
sample may represent as much as two-thirds of the vertebrate spe-
cies that might have lived on the Cretaceous mid-continental flood-
plain. As Fox (1968) has noted, the diversity known in the fossil
habitat does not necessarily reflect the total diversity that was
actually present in a group during the late Cretaceous; other habi-
tats not represented in the fossil record undoubtedly included re-
lated forms.

Relative Weight

Shotwell (1955, 1958, 1963) did not discuss the possibility of
using figures based on weight of fossil specimens. It is clear that
the vagaries of preservation and mineralization preclude using
weight of fossils as an indication of biomass, however useful this
item of information might be. Nevertheless, we have obtained
some useful results from carefully qualified use of the relative
weight (RW) of specimens for major groups of lower vertebrates
and some other fossil material from both Lance and Bug Creek
localities (Table 4). Specific gravity of the material and relative
size of the animals from both V5620 and BCA are essentially
identical; use of weight therefore seems justified.

The information on relative weight that is of particular interest
is based on material that cannot be used to express the minimum
number of individuals easily (if at all). For instance, floral mate-
rials are not common in washed and screened material. Seeds may
occur, however, often as durable casts of the inside of the seed
coat. In the Lance Formation locaUty V5620, such seed casts
are common and may indicate closeness of that locaUty to stream-
bank or riverbank, a conclusion also supported by the high rela-
tive weight of turtle material. Twice as much unidentifiable mate-
rial occurs at V5711 than at V5620, emphasizing the extent of
erosion in a blowout locality. The slightly higher RW of copro-
lites at V5711 is probably also caused by resistance to erosion of
these hard pebbleUke objects. At BCA, coprolites are rare, and
invertebrates (except for a few traces) absent; this is probably
because of deposition of the fossils farther away from the shore.

The abundance of bony fish material at BCA is clearly shown
by the weight data. Although the relative abundance of Bug
Creek bony fishes is slightly less overall than at Lance localities,
the large amount of identifiable, unworn material at BCA makes
it clear that this locality was quite close to the natural habitat of

26

BREVIORA

No. 343

(especially) Lepisosfeus and Amia. Estimated body size of these
two common fishes was similar in the two localities, so that this
factor does not influence the weight data.

The difi:erence in amount of dinosaur material is quantified by
the weight data. Although tooth size of some BCA dinosaurs is
slightly smaller than in the Lance localities, the difference is not
enough to affect the relative weight significantly. In part, the re-
duction in RW of dinosaur material at Bug Creek, when compared
to that of the Lance localities, may represent reduced frequency of
dinosaurs in the presumed deeper, more open channels represented
in this area of Hell Creek deposition. Nevertheless, the reduction
in numbers of dinosaur individuals in the late Cretaceous that ac-
companied their eventual extinction probably was the most im-
portant factor in reduction of the relative weight of dinosaur fossils.

TABLE 4

Comparison of Relative Abundance by Weight of Major Groups

in Lance Formation (Wyoming: UC Localities V5711 and

V5620, AM Coll.) and Hell Creek Formation (Montana;

Bug Creek Anthills, MCZ Collection)

Bug Creek

V5711

V5620

Anthills

Relative

Relative

Relative

Abundance

Abundance

Abundance

by weight (%)

by weight (%)

by weight (%)

sharks

0.25%

0.647o

0.137o

bony fishes

4.10

6.33

39.35

salamanders

1.22

1.58

3.80

frogs

0.01

0.04

0.07

turtles

8.23

35.56

8.12

crocodiles

3.53

6.79

3.32

dinosaurs

10.90

14.56

0.70

snakes

0.0 1

0.00

0.01

lizards

0.18

0.04

0.46

other groups

0.05

0.90

5.58

unidentifiable material

69.61

30.24

38.60

invertebrates

0.38

0.90

0.00

coprolites

1.16

0.94

0.23

seeds

0.00

1.51

0.01

99.02%

100.03%

100.68%

Note: Calculations do not include mammals or gar scales because no
weight figures were available for the Lance material.

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 27

Community Analysis
Relative Abundance

Figures 1 and 2 and Table 5 express the relative abundance
(RA) of dominant genera and major groups at BCA and Lance
locality V5620.

The faunal analysis diagrams (Figs. 1, 2) also include mam-
mals, and indicate that the increased diversity of the latter at BCA
in comparison with that at V5620 is accompanied by an increased
relative abundance as well: RA at V5620 = 26% ; RA at BCA =
33%. Lower vertebrates show greater abundance of some species
at BCA than at V5620 (Fig. 1). Amia, Scapherpeton, Opistho-
triton, Lisserpeton, and Pancelosaiirus together account for about
50% of the MNI at BCA; the same taxa form only 27% at
V5620. In the major groups (Fig. 2), bony fishes and sala-
manders account for about 50% of the MNI in both locahties
with mammals and lizards comprising about 45% of the re-
mainder. At BCA, the relative proportions of these four groups
are difi'erent from those at V5620: salamanders and mammals
(rather than fishes and lizards) are the dominant groups, totaling
about 65% of the MNI. There is a substantial complementary
reduction in lizards, and (to lesser degree) fishes, from V5620
to BCA.

TABLE 5

Relative Abundance of Major Groups of Lower Vertebrates in
Three Late Cretaceous Localities

Hell Creek Formation
Bug Creek Anthills

sharks 2 1 1

bony fishes 38 25 30

salamanders 30 49 52

frogs 3 1 2

turtles 1 1 3

lizards 23 21 8

dinosaurs 2 1 2

other 11 2

Lance

Formation

V5620

V5711

2

1

38

25

30

49

3

1

1

1

23

21

2

1

1

1

100%

100%

'70

These overall differences have some ecological significance, we
believe. As Estes et al. (1969) note, the Bug Creek Anthills
fossils indicate deposition in more open waterways, more lowland

28

BREVIORA

No. 343

PLATACOOON

V5620

Figure 1. Relative abundance of dominant lower vertebrate genera in
two late Cretaceous localities: a, Bug Creek Anthills; b, Lance
locality V5620.

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 29

MARSUPIALS

ELASMOBRANCHS

V5620

Figure 2. Relative abundance of major groups of vertebrates in two late
Cretaceous localities: a, Bug Creek Anthills; b. Lance locality
V5620.

30 BREVIORA No. 343

habitat, and greater distance from the terrestrial, streambank habi-
tat than do those from the Lance Formation. This difference,
however, is small; the overall ecological similarity between the two
local faunas, as demonstrated by the faunal list, is beyond doubt.
More specific reference to the difference in habitats represented
will appear below.

Relative Completeness

The relative completeness (RC) of members of the fauna was
used by Shotwell (1955, etc.) to indicate proximity of taxa to the
site of deposition, and by extension, he used this parameter to
delimit "proximal" and "distal" communities. Shotwell (1955:
332) arbitrarily set the limit of the proximal community at the
average RC (= number of specimens per individual) for the total
sample. RC and RA for the Bug Creek Anthills sample is shown
in Figure 3 and Table 1 . At BCA this method for distinguishing
proximal and distal communities does not work satisfactorily; the
complexity of the fauna is not resolved by so simple a distinction.
The most proximal community at BCA is, however, clearly aquatic,
and includes the following dominant species:

Lepisosteus occidentalis — alligator gar

Scapherpeton tectum — aquatic (elongated?) salamander

Amia jragosa — bowfin

Opisthotriton kayi — aquatic, elongated salamander

Lisserpeton bairdi — aquatic (elongated?) salamander
Other, less common members of this aquatic community included
sturgeons, the discoglossid frog Scotiophryne, the sirenid sala-
mander Habrosaiirus, the presumed amphibious lizard Pancelo-
sauriis, and the baenid turtles. A proximal terrestrial (riparian)
community can also be postulated to include most of the multi-
tuberculate mammals, and the two common eutherians. The high
frequency of the multituberculate Mesodma seems to indicate that
it may have overrun the soft, marshy margins of the rivers as small
rodents and some other small mammals do today. Mesodma and
the other multituberculates seem, from their dentition, to have been
seed or seed-coat eaters. The primitive ungulate Protungidatum
was probably also herbivorous, adapted for browsing on marshy
vegetation. Some of the mammals may have been amphibious, as
occurs today in marshy habitats in various parts of the world.

Of the total of approximately 78 species, 42 belong to the "dis-
tal community" as delimited by ShotwelFs method. Considerable

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 31

taxonomic diversity occurs among the "distal" group; sharks, bony
fishes, salamanders, frogs, lizards, dinosaurs, crocodilians, birds and
mammals are all included. Many (rather than one) distal com-
munities are certainly represented, some aquatic and some terres-
trial. Most of the lizards, the dinosaurs (excepting Anatosaiinis),
and 1 3 of the 20 species of mammals probably were part of a ter-
restrial, streambank-riverbank community; Coriops and Plataco-
don, Londiidion, CJiampsosaunis, and Trionyx seem to be the
nucleus of a distal aquatic community perhaps inhabiting local
ponded areas in streams, as indicated by their greater abundance
in Lance Formation localities than at BCA. Amphibious species
such as the emydine turtle, the dermatemydids, the crocodilians,
and the remaining frogs inhabited the ecotone between aquatic and
terrestrial habitats. Finally, a group of fishes of essentially marine
affinities (the sharks, the elopid, cj. Paralbula, and Belonostomus)
seem to represent a group of diadromous forms occasionally mi-
grating into the major streams and rivers. Unfortunately, RC
figures are not available for the Lance localities. Nevertheless, a
few Lance species whose relative completeness can be subjectively

Figure 3. Relative completeness and relative abundance of vertebrates
from Bug Creek Anthills. RC is given along the radius; RA
in degrees of arc. The "proximal community" as indicated by
the average RC (=4.9) is delimited by the arrovt'ed circum-
ferential line. Members of groups A - E are indicated by letters
in parentheses following the name of taxa in Table 1.

32 BREViORA No. 343

estimated reinforce some of the above conclusions. These include
Coriops, Platacodon, and Hahrosaurus; their apparently greater RC
at V5620 than at BCA tends to confirm our suggestion (based
essentially on relative abundance) that they were stream-dwellers
rather than river-dwellers.

Some substantiation for the postulated habitat differences be-
tween BCA and V5620 comes from the nature of the sediments
themselves. While we cannot treat this subject in detail here, the
relatively greater amount of clay-sized material in the sandstone
at BCA indicates more sluggish waters and deposition nearer the
open areas of the waterways than was the case at V5620, where
sandy streambanks appear to have been the site of deposition of
the fossils.

Among recent ecological situations, Lance locahty V5620 prob-
ably represents a wooded swamp habitat, with small to medium-
sized watercourses and some ponding. The equivalent for Bug
Creek Anthills, however, seems to be a major waterway issuing
from such a wooded swamp: similar habitats were available in or
near both localities, but terrestrial and riparian habitats were
restricted at BCA, which may have been nearer the delta region
of one of the major rivers emptying into the remnants of the late
Cretaceous sea.

The flora of the Hell Creek Formation was studied by Shoe-
maker (1967) and Norton and Hall (1969). As for the sedi-
ments, floral analysis cannot be detailed here. However, it is
worth noting that the "Transition Flora" found in the upper part
of the Hell Creek Formation by Norton and Hall begins at a level
slightly above that of Bug Creek Anthills, while the majority of
the species of the "Upper Cretaceous Flora" disappear at a level
somewhat below the locality. Floral materials are poorly repre-
sented in the sandstones of the lower half of the upper part of the
Hell Creek Formation, and therefore the point at which the cli-
matic deterioration postulated by Hall and Norton (1969) begins
is difficult to determine. Similarity of the Bug Creek Anthifls
fauna to that of the Lance Formation, in spite of the differences in
the mammalian fauna noted by Sloan and Van Valen (1965),
seems to indicate that a subtropical climate (as for the Lance
Formation) still prevailed in this area of Montana in Bug Creek
Anthills time.

1970 PALEOECOLOGY CRETACEOUS VERTEBRATE COMMUNITY 33

ACKNOWLEDGEMENTS

Dr. Robert E. Sloan has enthusiastically supported our efforts,
and has generously allowed the Museum of Comparative Zoology
to collect the fossil sample used in the quantitative study. He has
also contributed much valuable information about geological set-
ting, details of occurrence, and collecting methods. We are very
grateful for his help. Mr. Arnold Lewis deserves special thanks
for providing first-hand knowledge of collection of the MCZ
sample. Dr. Malcolm McKenna kindly contributed the American
Museum of Natural History Lance Formation sample used for
comparison.

Dr. Charles Meszoely made preliminary analyses of a number
of Hell Creek Formation collections from the American Museum
of Natural History, University of Cahfornia, University of Kansas,
and Saint Paul Science Museum. These were helpful in docu-
menting the stability of relative abundance in the MCZ collection.
Mr. Berberian prepared most of the raw data and preliminary
analysis of the MCZ sample.

We thank Donald Baird, W. A. Clemens, Jr., Walter Coombs,
Bryan Patterson, Dale Russell, J. A. Shotwell, and R. E. Sloan for
helpful suggestions.

The figures are by Mr. Laszlo Meszoly. This research was sup-
ported in part by National Science Foundation grants GB-1683,
GB-4303, and GB-7176, to the senior author.

LITERATURE CITED

Bishop, S.

1947. Handbook of Salamanders. New York: Comstock Publ. Co.,
XIV + 555 pp.
Blair, W., Blair, P. Brodkorb, F. Cagle, and G. Moore

1968. Vertebrates of the United States. 2d ed. New York: McGraw-
Hill Book Co., Inc., XII + 819 pp.
Brown, B.

1907. The Hell Creek beds of the Upper Cretaceous of Montana:
Their relation to contiguous deposits, with faunal and floral
lists and a discussion of their correlation. Bull. Amer. Mus.
Nat. Hist., 23: 823-845.

1908. The Ankylosauridae, a new family of armored dinosaurs from
the Upper Cretaceous. Bull. Amer. Mus. Nat. Hist., 24: 187-
201.

Brodkorb, P.

1963. Birds from the Upper Cretaceous of Wyoming. Proc. XIII
Intern. Ornith. Congr., pp. 55-70.

34 BREVIORA No. 343

Clemens, W.

1963. Fossil mammals of the type Lance Formation, Wyoming.
Part I. Introduction of Multituberculata. Univ. Calif. Publ.
Geol. Sci., 48: 1-105.

ESTES, R.

1964. Fossil vertebrates from the late Cretaceous Lance Formation,
Eastern Wyoming. Univ. Calif. Publ. Geol. Sci., 49: 1-180.

EsTES, R., P. Berberian, and C. Meszoely

1969. Lower vertebrates from the late Cretaceous Hell Creek Forma-
tion, McCone County, Montana. Mus. Comp. Zool., Harvard
Univ., Breviora, no. 337: 1-33.

Fox, R.

1968. Studies of late Cretaceous vertebrates. IL Generic diversity
among multituberculates. System. Zool. 17: 339-342.

Hall, J., and N. Norton

1969. Palynological evidence of floristic change across the Creta-
ceous-Tertiary boundary in eastern Montana. Palaeogeog.,
Palaeoclimatol., PalaeoecoL, 3: 121-131.

Jepsen, G.

1963. Eocene vertebrates, coprolites, and plants in the Golden Valley
Formation of western North Dakota. Geol. Soc. Amer. Bull.,
74: 673-684.
Kellogg, R.

1929. The habits and economic importance of alligators. U. S. Dept.
Agric. Tech. Bull., 147: 1-36.

LOWERY, G.

1960. Louisiana Birds. 2d ed., Louisiana State Univ. Press, XXIX
+ 556 pp.
McKenna, M.

1962. Collecting small fossils by washing and screening. Curator,
5: 221-235.
Meszoely, C.

1970. North American fossil anguid lizards. Bull. Mus. Comp. Zool.,
Harvard Univ., 139.

Norton, N., and J. Hall

1969. Palynology of the Upper Cretaceous and Lower Tertiary in

the type locality of the Hell Creek Formation, Montana,

U.S.A. Palaeontogr., ser. B, 125: 1-64.
Olson, E.

1966. Community evolution and the origin of mammals. Ecology,
47: 291-302.

OSBORN, H.

1916. Skeletal adaptations of Ornitholestes, Striithiomimiis, Tyran-
nosanriis. Bull. Amer. Mus. Nat. Hist., 35: 733-771.
Shoemaker, R.

1967. Fossil leaves of the Hell Creek and Tullock formations of
eastern Montana. Palaeontogr., ser. B, 119: 54-75.

1970 PAT FOFCOrOCV rRFTACF.OUS VF.RTEBRATE COMMUNITY 35
SllOTWELL. J.

1^55. An approach to the paleoecology of mammals. Ecology, 36:

327-337.
195S. Intercommunity relationships in Hemphillinn (miil-Phocenc)

mammals. Ecology. 39: 271-282.
l'-)63. Pliocene mammalian communities of the Juntnra hasin. Trans.

Amer. Philos. Soc, 53: 7-21.
Sloan. R.. .\nd L. Van Vai.fn

1965. Cretaceous mammals from Montana. Science, 14!J: 220-227.
Van Valen, L.

1964. Relative abundance of species in some fossil mammal faunas.

Amer. Nat.. 98: 109-116.

VOORHIES, M.

1969. Taphonomy and population dynamics of an early Pliocene
vertebrate fauna, Knox County, Nebraska. Contrib. Geol.
Univ. Wyoming, Spec. Pap. 1: 1-69.
Wilson, R.

1960. Early Miocene rodents and insectivores from northeastern
Colorado. Univ. Kansas Pal. Contrib.. 7: 1-92.

(Received 4 December 1969.)

H ^-C L-^r^WJ^C^ DO NOT CkcQLATF

BREVIORA

MUS. COMP. ZOOL.

Mmseiiiiiii of Comparsitive Zoolog^RApY

CAMBRrooE, Mass. 27 February, 1970 Numb^/\^4 ^ ^Qy^

THE CHANARES (ARGENTINA) TRIASSIC Harvard

REPTILE FAUNA University

VI. A CHINIQUODONTID CYNODONT WITH AN
INCIPIENT SQUAMOSAL-DENTARY JAW ARTICULATION

Alfred Sherwood Romer

Abstract. The cranial anatomy of a small carnivorous chiniquodontid
cynodont from the Middle Triassic Chanares Formation of Argentina is
described and figured; this new form is named Probainognathiis jenseni. In
addition to other advanced characters, Probainognathiis has a definite jaw
articulation, accessory to that of quadrate-articular, between squamosal and
dentary. It is not improbable that the Chiniquodontidae lie close to and,
perhaps, on the main evolutionary line leading to mammals.

In a previous paper in this series (Romer, 1969b), I have
described, as Probelesodon lewisi, a predaceous cynodont of the
family Chiniquodontidae, and pointed out features which suggest
that this Middle Triassic family, including the only known carniv-
orous cynodonts of that age, may well represent the therapsid line
leading toward mammals. I here describe a smaller contemporary
chiniquodontid which is still further advanced in the presence of an
incipient squamosal-dentary jaw articulation. This form may be
formally described as:

PrOBAINOGNATHUS jenseni, gen. ET SP. NOV.

Holotype. La Plata Museum 1964-X1-14-7, a skull and jaws.
Collected from the Chafiares Formation in La Rioja Province, Ar-
gentina, about 3 km north of the point where the Rio Chanares
debouches into the Campo de Talampaya.

Combined generic and specific characters. A small chiniquo-
dontid cynodont, mature specimens having an average basal skull
length of about 71 mm. Seven cheek teeth of the type common in
carnivorous cynodonts, the cheek teeth generally showing marked
evidence of wear. The squamosal has a shallow but well-defined
cavity for dentary and surangular articulation.

2 BREVIORA No. 344

The generic name has reference to the advanced nature of the
jaw articulation. The specific name is in honor of James A. Jen-
sen, formerly of the Harvard staff, and now in charge of geologic
and paieontologic collections at Brigham Young University; he
served brilliantly as a collector on both Harvard expeditions to
Argentina.

Collection and preparation of this and other Chanares materials
were supported by National Science Foundation grants GB-4615
and 8171.

The skull of this small carnivore (Figs. 1-3, 5, 6) is similar in
nearly all regards to that of other members of the Middle Triassic
family Chiniquodontidae, including the contemporary, newly des-
cribed genus Probelesodon. Since I have described the skull of
that form in some detail, Probainognathus may be somewhat more
briefly treated by comparison with it.

The size is relatively small; the average skull length (measured
from the base of the premaxillary tooth border to the condyles) is,
in a series of twelve skulls, 71 mm. The skull proportions are
those of chiniquodonts in general, inherited presumably from the
Thrinaxodontidae. The skull is proportionately broad, with a nar-
row and relatively short face and a greatly expanded zygomatic
region. The width across the zygomatic arches in Probelesodon is
about 75 per cent of skull length, here nearly 80 per cent. A pro-
portionate difference from Probelesodon lies in the deeper otico-
occipital region, mainly related to greater depth of the parietals,
and a higher occipital crest. In correlation with smaller size, the
orbits are relatively larger (by about 20 per cent) than in Probele-
sodon; they appear to be directed somewhat more dorsally than in
that genus. In Probelesodon the centers of the orbits are some-
what back of the half-length of the skull; here they are somewhat
anterior to this point. As in other chiniquodontids, there is a
nubbin of bone projecting into the orbit from its anterior edge ex-
ternal to the opening of the lacrimal duct. The sagittal crest ex-
tends further posteriorly than in Probelesodon before cleaving into
the paired occipital crests, giving a straighter posterior margin to
the skull, as seen in dorsal view, and making the occipital region
below the crests more concave. The braincase is relatively broader
in the parietal region.

The general pattern of the dermal roofing elements ( Figs. 1,2)
is similar to that in other chiniquodonts. The snout, as in other
members of the family, projects well forward of the premaxillary
tooth row; the processes of the premaxillae, ascending over the

1970

PROBAINOGNATHUS JENSENI

Fig. 1. Dorsal view of the skull of Probainognathiis. This and Figs.
2-3, 5-8 are composites, about the size of the holotype. Abbreviations for
Figs. 1-3 and 5-8: a, articular; an, angular; bo, basioccipital; hs, basisphen-
oid and parasphenoid; c, coronoid; d, dentary; e, epipterygoid; /, frontal; fo,
fenestra ovalis; /, jugal; jf, jugular foramen; /, lacrimal; m, maxilla; n,
nasal; oc, occipital complex; p, parietal; pap, paroccipital process; pi, pala-
tine; pm, premaxilla; po, postorbital; pp, postparietal; ppf, pterygoparocci-
pital foramen; pr, prootic; pra, prearticular; prf, prefrontal; pt, pterygoid;
ptf, posttemporal fenestra; q, quadrate plus quadratojugal; sa, surangular;
sm, septomaxilla; sp, splenial; sq, squamosal; st, stapes; /, tabular; v, vomer;
v., „, foramen for trigeminal nerve.

V 2+3

Fig. 2. Lateral view of the Probainognathiis skull.

4 BREVIORA No. 344

external nares to be inserted posteriorly between the nasals, are so
slender that the snout region has been broken off in many of the
specimens. With snout loss, there is little available evidence as to
the development of the septomaxilla. In the Brazilian chiniquo-
dontids (Romer, 1969a), this element is highly developed; here
the only remains preserved consist of a thin layer of bone lying
along the posteroventral margin of the external nares, separated in
part from the maxilla by a small foramen.

The maxilla is slightly swollen in the area overlying the root of
the canine, but this swelUng is much less marked than in Probele-
sodon. The suborbital arch swings upward from the line of the
tooth-row much more markedly than in Probelesodon, in correla-
tion with the more upward facing of the orbit here. The sub-
orbital, postorbital, and zygomatic arch regions are similar to those
in other chiniquodontids except that (partially in correlation with
occipital depth) the zygomatic arch swings down posteriorly more
sharply than is generally the case; notable is a marked ventral ex-
tension of the squamosal at the posterior end of the arch. As in
other chiniquodontids, the channel in the squamosal believed to be
for the external auditory meatus is less developed than in cyno-
gnathids and gomphodonts generally.

As in other chiniquodonts, the palatine participates to a promi-
nent degree in the formation of the secondary palate (Fig. 3), ex-
tending back in this structure to the level of the back end of the
tooth-row. Since, however, the tooth-row is here relatively short,
the palatine contributes somewhat less than half the length of the
secondary palate, whereas, in Probelesodon the palatine portion is
somewhat longer than that contributed by the maxillae. The two
tooth-rows curve only slightly outward posteriorly, in contrast with
a much greater lateral curvature in Probelesodon, and hence the
palate is relatively narrow. In correlation, presumably, with this
constriction, the outer margins of the palatal surface, just inside the
posterior cheek-teeth on the maxillae, are somewhat excavated to
give room for the posterior lower molars when the jaws are closed.
Beneath the secondary palate posteriorly, the roof of the primary
palate is more highly arched than in Probelesodon. As in other
chiniquodontids, the palatines send posteriorly long slender proc-
esses along ridges converging backward from either lateral mar-
gin of the secondary palate. Presumably a small ectopterygoid was
present, but I have not been able to determine its boundaries satis-
factorily. The posterior end of the vomer is visible above and
posterior to the end of the secondary palate. Pterygoid flanges are

1970

PROBAINOGNATHUS JENSENI

Fig. 3. Palatal view of the Probainognathiis skull.

well developed, but are not as elongate as in Probelesodon. Pos-
terior to these flanges the two pterygoids constrict and embrace the
base of the braincase for some distance, as well as running out for
a short distance along the flange of the palatal structure extending
to the quadrate region. The main substance of this flange, how-
ever, is formed by the epipterygoid.

The general region of the jaw articulation is of interest (Fig. 4).
The stapes is absent in most specimens. In one, the basal portion
is still present, inserted in the fenestra ovalis. The shape of the
part preserved indicates that this ossicle was rather surely of the
type seen in cynodonts generally, essentially a slender quadrilateral,
with a large foramen, extending out to a contact with the quadrate.
The squamosal, as in other chiniquodontids and in thrinaxodontids,
is distinctly divided into two rami, occipital and zygomatic, with a
narrow neck between the two. The occipital ramus forms the
outer side of the posterior end of the sagittal crest and descends
along the anterior surface of the occipital crest. Near the point of
union of the two rami are two deep notches for the quadrate and
quadratojugal. In other described chiniquodonts, these notches
are visible only ventrally; here, they are open dorsally as well,
maldng the quadrate even freer than is usual in cynodonts. The

BREVIORA

No. 344

Fig. 4. External, internal, and ventral views of the region of the jaw
articulation of Probainognatlins; a, articular; an, angular; c, incipient
"glenoid fossa" on squamosal; cond, occipital condyle; ep, epipterygoid; /,
jugal; if, jugular foramen; per, periotic; pra, prearticular; q, quadrate; san,
surangular; sq, squamosal; st, stapes.

quadrate (with which a small quadratojugal is united), is inserted
in these two notches; it is, in our specimens, always loosened and
pulled anteriorly somewhat out of position, and has been lost in a
number of cases. The articular face of the quadrate is convex;
that of the quadratojugal is concave, apparently for reception of the
projecting posterior end of the surangular. Just anterior and dor-
sal to the quadrate is the lower end of the pronounced ventral
flange of the zygomatic ramus of the squamosal. On its inner sur-
face, close to the anterior end of the notch for the quadratojugal,
there is a shallow but well-outlined oval socket, obviously for an
additional area of lower jaw articulation. The nature of this
articulation is discussed below.

The anterior portion of the braincase proper has not been ex-
plored. (I hope at a later date to serial-section a skull of this
form.) There is, as in cynodonts generally (Fig. 5), a large area

1970

PROBAINOGNATHUS JENSENI

Opposite the orbit in which the braincase wall is unossified. Pos-
terior to this, as in other advanced cynodonts, there is an expanded
epipterygoid which, ventrally, sends an extension back to a quad-
rate contact. As noted earlier, the parietals are somewhat greater
in vertical extent than in Probelesodon, and shghtly more expanded
transversely. The basicranial and periotic regions are quite similar
to those seen in Probelesodon. However, in several specimens the
outer end of the paroccipital process is incompletely ossified, leaving
a gap of somewhat variable shape and proportions between the
process and the buttress formed laterally to it by the squamosal.

Fig. 5. Occipital view of the P robainognathus skull.

On the occiput (Fig. 5), the unusual posterior extent of the
sagittal crest has resulted in the upper part of the occipital face of
the skull being rather more concave than in Probelesodon. Sutures
between tabulars and postparietal with the fused occipital elements
are difficult to determine. The posttemporal fenestrae are small.

V2 + 3

Fig. 6. Lateral view of the braincase region of Probainognathus; the
zygomatic arch removed.

8 BREVIORA No. 344

The lower jaw (Figs. 7-8) is, in general, quite similar to that
described for Probelesodon. The dentary is very large. There is a
high coronoid process, thickened along its anterior margin, which
rises, in the position of jaw occlusion, nearly to the level of the
sagittal crest and curves backward above nearly to the level of the
jaw articulation. From the tip of this process, the posterior margin
of the dentary curves forward and downward to extend posteriorly
toward the articular area. This posterior projection of the dentary
is little thickened, and there is no expansion in the form of a con-
dyle. This process is incompletely preserved in most specimens,
and the relation of the dentary to the other jaw bones posteriorly is
frequently obscure. In the better preserved specimens, however, it
seems clear that this posterior process was appHed to the dorso-
lateral surface of the surangular, terminating just short of the pos-
terior articular expansion of the latter bone. From its relative
position, it is clear that the terminus of the dentary, together with
the surangular, fitted into the shallow cavity on the squamosal
external to it, and that we have here the first step in the develop-
ment of the future mammahan squamosal-dentary jaw articulation.

Fig. 7. Lateral view of the jaw of Probainognathus.

The jaw symphysis, formed almost entirely by the dentary, is a
broadly fused one. Behind the symphysis the lower margin of the
dentary extends backward in a nearly straight fine to a sharply
marked angle. This lower dentary margin is considerably thick-
ened, with a rounded ventral margin. From the angle, the pos-
terior margin turns nearly vertically upward and then curves back-
ward, leaving the posterior part of the angular and, to some degree,
the surangular exposed externally below it. About opposite the

1970

PROBAINOGNATHUS JENSENI

posterior end of the tooth-row the dentary develops on its inner
surface a horizontal ridge which extends back to the posterior end
of the bone; below this thickened ridge is a lengthwise excavation
in which is lodged the longitudinal "rod" of conjoined bones which
support the articular.

Fig. 8. Medial view of the jaw of Probainognathiis.

The splenial is a long, but thin and slender, element. At its
anterior end it takes a small part in the jaw symphysis; it is gen-
erally imperfectly preserved, but appears to continue backward,
below the anterior end of the angular, to terminate at a point above
the angle of the dentary. In other chiniquodontid genera I have
found no trace of a coronoid bone, although it surely existed in
vestigial form as a thin internal flake of bone. In Probainognathus,
as well, it is absent in most instances. In one specimen, however,
the bone is present, but is imperfectly seen (the jaws are in tight
occlusion), and I am not sure of its precise outline. As in cyno-
donts generally, its function seems to be the provision of a surface
(presumably covered with stout epithelium) to play against the
ventral process of the pterygoid during jaw movement.

Disregarding for the moment the incipient dentary-squamosal
contact, articulation with the skull is mediated by a moderately ex-
panded area at the posterior tip of the jaw, which is convex dor-
sally where it fits into the concavity of the quadratojugal, somewhat
concave ventrally to meet the quadrate proper. The lower portion
of this articular area is presumably formed by the articular bone.
Despite the absence of a suture, it is probable that the upper part of
the articulation is formed by the posterior end of the surangular,

10 BREVIORA No. 344

thus continuing to this advanced stage the quadratojugal-surangular
articulation seen m various more primitive reptiles, and even in
labyrinthodonts.

The functioning of this relatively small articular structure as a
jaw support would be impossible were it not firmly bound to the
"rod" of bone mentioned above, which runs forward more than
half the jaw length in the longitudinal recess in the dentary. This
structure consists of angular, surangular, and prearticular, all ap-
parently firmly joined together and, in the case of surangular and
prearticular, united without suture to the articular. The reflected
flange of the angular, already small in other advanced cynodonts, is
here reduced to a tiny process preserved in only one specimen; it
lies far forward of the articular region, above the angle of the
dentary.

As in other chiniquodontids, there are four upper and three
lower incisors in each ramus; these are slender, and subcircular in
section. The upper incisors are essentially vertical in position; the
lowers are mildly procumbent. Canines are moderately developed;
as preserved they are nearly straight, not sharply tipped, and are
relatively broader anteroposteriorly than in transverse section.
There are seven cheek teeth (Fig. 9) above and below in all well-
preserved specimens. Their pattern appears to be that common in
all carnivorous cynodonts — thin from side to side, but elongate

Q-

Fig. 9. Cheek teeth (shown as of tlie left side) as preserved in various
specimens of Probainognathiis. First four rovv's, upper teeth; bottom row,
lower teeth, X21/2.

1970 PROBAINOGNATHUS JENSENI 11

anteroposterioiiy, with a prominent cusp at about midlength and
two or three accessory cusps anterior and posterior to it. Since in
most cases the lower jaws, when preserved, are in occlusion, little
can be seen of the lower cheek teeth but, as far as can be observed,
they appear to be similar to the uppers. A number of cheek denti-
tions are shown in Figure 9. It appears probable that in all cheek
teeth there is, as usual in cynodonts, an accessory cusp both anter-
ior and posterior to the principal one; in addition, however, there
appears to have been a second posterior cusp in most cases. How-
ever, accurate description of the cusp pattern is made difficult be-
cause of a condition unusual in cynodonts. In most specimens, the
teeth show evidence of heavy wear, leaving the tooth with a chisel-
hke distal shearing edge. Mr. Vincent Maglio has kindly X-rayed
the upper cheek-tooth region in one specimen. As is the case,
where known, in other cynodonts, all the teeth are persistently
single-rooted. I have made no attempt to study tooth succession.

DISCUSSION

Probainognathus is obviously a member of the family Chiniquo-
dontidae, resembling the other described members of this Middle
Triassic group — Chiniquodou, Belesodon, Probelesodon — in
major cranial characters. As I have recently pointed out (Romer,
1969a), the chiniquodontids are quite generalized cynodonts, read-
ily derivable from the thrinaxodontids of the early Triassic, but
more advanced in certain features. Most notable of these is the
high development of the secondary palate to an essentially mam-
malian stage, this structure extending back to the level of the pos-
terior end of the maxillary tooth-row, with a considerable involve-
ment of the palatine. Further, as I have commented in discussion
of Probelesodon, the dentary in chiniquodontids, in general, extends
backward to such a degree that it must have been functionally
involved in jaw articulation, although in previously described
genera there was no contact of this bone with the squamosal.

Probainognathus differs from the contemporary Probelesodon in
certain regards. It is much smaller in size. The cheek tooth-row
is shorter (and, in correlation, the involvement of the palatine bone
in the formation of the secondary palate is less). In the dentition,
the "molars" tend to develop a fourth posterior member of the
linear series of cusps. Uniquely, there is a strong trend for the
wearing down of the cheek teeth to a chisel edge, with obUteration
of the cusps.

12 BREVIORA No. 344

Most important of differences, however, in Probainognathus is
the development of a definite, if slight, articulation of dentary with
squamosal. As described above, the latter bone curved very sharply
downward over the region of the jaw articulation, and on its inner
surface here shows a shallow but definitely outlined cup for articu-
lation with the posterior end of the dentary. There is no develop-
ment of a condyle of mammalian type on the dentary; the articular
surface is merely the flat outer surface of the slender posterior ex-
tremity of the bone, slightly thickened in a ridge above the bar of
bone (angular and surangular and prearticular) which braces the
articular anteriorly. This dentary process would be too weak to be
of functional significance were it not closely applied and supported
by the underlying surangular.

In advanced therapsids generally, we see stages in which the
dentary has increased to become the major jaw element, carrying
not merely the dentition but also most of the jaw musculature.
The jaw is, however, inefficient in that the dentary itself takes no
part in articulation with the skull, and must rely on reduced pos-
terior elements of the jaw to connect the dentary with the articular
which, in turn, is connected with the major skull structure only
through the feeble quadrate. An improved connection between
dentary and skull would be highly desirable. Dr. A. W. Crompton
has pointed out to me a primitive "essay" toward remedying this
situation. In Trirhachodon (in which the dentary does not reach
as far back as in chiniquodontids) there is a distinctive boss on the
ventral border of the squamosal in the region of the jaw articula-
tion. This lies close to the posterodorsal margin of the surangular
(here still a fairly stout bone), thus supplementing the rather feeble
articular-quadrate jaw joint (Crompton, 1963). A similar squa-
mosal boss is present in the traversodont Massetognalhus, and
presumably performs a similar strengthening function. In the long
run, however, the type of shift in the articulation initiated in Pro-
bainognathus was to be of much greater phylogenetic importance.

Most diagnostic criteria by which reptiles and mammals may be
contrasted have to do with matters of function, or of "soft anat-
omy." One major osteological differentiation, useful to the pale-
ontologist, has to do with jaw articulation. In reptiles, the joint
lies between quadrate and articular; in mammals, a new joint has
developed between squamosal and dentary, with subsequent loss
from the lower jaw of the half-dozen elements (apart from the
dentary) which were primitively present — some lost entirely,
others incorporated into the middle ear.

1970 PROBAINOGNATHUS JENSENI 13

Until recent years the reptile-mammal contrast in jaw articula-
tion was clear-cut paleontologically. Early Mesozoic mammals
are, for the most part, very inadequately known; but, in many
cases, lower jaws are present in which the dentary bears a condyle
indicating the presence of a squamosal-dentary articulation. And
while, on the reptilian side of the picture, dentary and squamosal
may closely approach one another, no actual contact between the
two has previously been reported in cynodonts — the group from
which most, if not all, mammals appear to have originated.

In the last dozen years or so the situation had become somewhat
clouded. In some early mammals certain of the "supernumerary"
jaw elements appear to have survived in reduced fashion, and in at
least some early forms, notably docodontids, a quadrate-articular
connection persisted accessory to a major squamosal-dentary joint
(Kermack and Mussett, 1958a,b; Kermack, 1965). Crompton
(1958, 1963) described, as Diarthrognathiis a small late Triassic
form of uncertain phylogenetic position (probably an aberrant off-
shoot from primitive cynodonts) which surely has nothing to do
with mammal origins, but has paralleled the mammals in having
acquired a contact of squamosal and dentary.

These discoveries resulted in a plethora of literature concerning
the origin, nature, and systematics of the Mammalia which I need
not discuss in detail. Currently, ideas of polyphyletic origins of
groups are highly fashionable. In early mammals, dentitions and
dentary bones (which generally are all the material available) show
considerable variation, and it has been suggested that mammals
took origin from reptiles in anywhere from four to (improbably)
nine discrete lines (Simpson, 1959). Under such hypotheses the
class Mammaha is not a natural assemblage — "a grade, not a
clade" — and if it is attempted to restore phyletic unity to the group,
it might be necessary either to restrict the class to the higher mam-
mals, the Theria, whose unity not even the most ardent of poly-
phyletic advocates have (as yet) questioned, or, at the other ex-
treme, to extend the term Mammalia downward to include part or
all of the Therapsida, or even the entire subclass Synapsida.

I see no reason to adopt either of these extremes and, further, am
skeptical as to whether any major degree of polyphyly is involved
in mammal origin. It is axiomatic that if all stages in the evolution
of one group from another were known, it would be impossible to
make a definitive separation between the two. In attempting to
draw a line between reptiles and mammals, all the changes in-
volved should be considered, rather than using a single feature as

14 BREVIORA No. 344

arbitrarily definitive. If, for example, we were to define a mammal
(as frequently advocated) as a form in which an articulation be-
tween squamosal and dentary had been attained, Probainognathus
would be, necessarily, separated from its obvious close relatives in
the family Chiniquodontidae and removed, not merely to another
family or order, but to an entirely different class of vertebrates!
This would be absurd. The reptile-mammal boundary should be
established not on a single character, but on a consideration of the
sum total of the features by which a fully developed mammalian
type is distinguished from its reptihan forebears (Simpson, 1959;
Romer, 1965). Far more important than initiation of a shift in
jaw articulation are changes in: cheek-tooth succession (and
roots); the auditory region; and, most especially, growth of the
brain to mammalian pattern and size, with concomitant reorgani-
zation of braincase structure. If (as one may hope) adequate
knowledge of c3niodonts and primitive mammals of the late Tri-
assic presently becomes available, the next generation of research
workers will surely be troubled in establishing a boundary; happily
for us today, the gaps in our knowledge are such that we can main-
tain the position that all well-known therapsids lie definitely below
a proper point of class cleavage.

As said above, I see little reason to believe that there has been
any great degree of polyphyly in mammalian origins (except pos-
sibly in the case of a few groups in which dental evidence is
obscure: monotremes, multituberculates, and the problematical
little haramyids). If we survey the Reptilia for mammal ancestors
it seems clear that these ancestors lie among the theriodont therap-
sids. Further, as far as our present rather considerable knowledge
of theriodonts goes, it is only to the carnivorous Cynodontia that
we can turn with confidence for mammalian forebears. Among
the Mesozoic mammals, we find interesting differences in cusp
patterns in the cheek teeth and in the presence or absence of an
angle on the dentary; but there is, as yet, nothing known of these
groups which would debar derivation of most if not all of them
from typical cynodonts with an enlarged dentary and cheek teeth
with a simple fore-and-aft pattern of cusps. We need not claim
that all mammals have descended from a single species of advanced
cynodont; but there need not, I think, have been any notable de-
gree of polyphyly concerned.

Cynodont cranial-structure has often been compared, with profit,
with that of mammals. Such comparisons have been based, of
necessity, on the structure of Lower Triassic cynodonts, in default

1970 PROBAINOGNATHUS JENSENI 15

of any adequate knowledge of later types. With accumulating
knowledge of the Middle Triassic family Chiniquodontidae, such
comparisons may now be made starting from a somewhat more
advanced cynodont level. A major difficulty in any study of this
sort hes in the fact that because of our inadequate knowledge of
cranial structure of early mammals, our comparisons are in gen-
eral made, compulsorily, with advanced therian types. We are
now aware that many early mammals still retained traces of the
reptiUan pattern in jaw structure, and the same was surely true of
the skull as well. It is to be hoped that work on Triassic mammal
material now in hand, but not as yet described, will yield data of
interest in transitional features.

Surely the greatest advance made by mammals over their reptil-
ian ancestors hes in the reorganization and enlargement of the
brain, with a necessarily concomitant restructuring of the brain
cavity. Some degree of change in this area appears to have been
occurring among cynodonts, as is seen in the olfactory bulb-turbi-
nal arrangement in Diademodon and Nythosaunis (Watson, 1913).
But even in chiniquodontids there is little evidence of advance in
brain size. As in other cynodonts, the endocranial cavity was still
small. The epiptei-ygoid is already expanded as an alisphenoid, but
the involvement, as seen in therian mammals, of orbitosphenoid and
frontal in the sheathing of the expanded brain, was still to come.
It is reasonable to believe that mammalian brain evolution pro-
gressed slowly during the course of the Mesozoic, and when ade-
quate material becomes available one may expect that mammals of
the late Triassic (and even Jurassic) may show a brain much less
expanded, and perhaps less perfectly ensheathed in bone than in
Cretaceous and later therians (Simpson, 1927).

The otic region is one in which change has been initiated in
chiniquodontids, but is far from completed. It is reasonable to
believe that the eardrum lay close to the region of the jaw articula-
tion in cynodonts (Parrington, 1955). The stapes is in proper
position and, presumably, in contact with the quadrate. This ele-
ment (with the conjoined vestige of the quadratojugal) is already
small in size and (as shown by its frequent loss in specimens) is but
loosely socketed in the squamosal, and seems readily available for
transformation into an incus once the shift in the jaw joint had
occurred. The articular, as the future malleus, is also small in
size and, of course, in proper articulation with the "incus," i.e.,
quadrate. Here, however, a considerable change must be made to

16 BREviORA No. 344

reach mammalian conditions, for closely connected with the articu-
lar is a bar of bone formed by prearticular, surangular, and angular
(without which the cynodont articular would be ineffective in jaw
articulation). This structure could not be reduced until the new
dentary-squamosal joint was well established. As the evidence of
marsupials and early placentals shows, construction of an auditor}'
bulla for ossicle protection and tympanum support was a late and
variable development. There is a puzzhng point here. The ento-
tympanic is a mammalian neomorph, but the Y-shaped embryonic
tympanic bone of marsupials has been not unreasonably compared
with the reptilian angular, one branch of the Y thought comparable
to the reflected lamina of the therapsid angular (Watson, 1951, fig.
67, etc.). However, in advanced cynodonts this lamina has been
reduced almost to the vanisliing point and lies far forward from the
presumed position of the eardrum.

The replacement of the old jaw articulation by the new has al-
ways been assumed to have been a gradual process rather than a
"great leap forward," and the condition in Probainognathus is in
agreement with this assumption. In advanced cynodonts generally,
and in chiniquodontids especially, the dentary has enlarged to as-
sume much of its future mammalian size and pattern; and, as I
have earlier noted (Romer, 1969b), in chiniquodontids generally,
it is so close to the squamosal near the region of the reptilian artic-
ulation that it was surely a functional element in jaw suspension.
In Probainognathus, actual articulation is initiated; further de-
velopment of a glenoid socket and expansion of the posterior tip of
the dentary to condylar form during the early stages of mammaUan
development can be readily envisaged.

In still other regards than those mentioned, the chiniquodontid
skull can be rather closely compared with that of a generalized
mammal. A number of changes in dermal bone pattern are
needed to attain advanced, therian, mammaUan structure; it is to be
expected that certain reptilian features were retained in the skull of
early mammals. In the narial region the ascending internarial
processes of the premaxillae (already very slender in Probainogna-
thus) disappear in mammals while, on the other hand, the pre-
maxillae push upward back of the nares, with elimination of the
septomaxillae. The maxilla, already enlarged in cynodonts, pushes
farther backward and upward, with reduction of its neighbors and
elimination of the prefrontal. The postorbital bar, already slender
in chiniquodontids, disappears with elimination of the postorbital

1970 PROBAINOGNATHUS JENSENI 17

element. The parietal foramen, much reduced in cynodonts gen-
erally, has disappeared in chiniquodontids. In this family the sagit-
tal and occipital crests already have the mammalian pattern, as
does the occiput. The condyles are of the double, mammalian
type; the four occipital elements are already fused in mammalian
fashion, and the postparietal and tabulars, which have united with
the occipital bone in mammals, are at least partially fused with it in
chiniquodonts. In the palatal surface, the anterior region is already
close to the mammaUan condition in chiniquodontids. The sec-
ondai7 palate is already developed to mammalian status. Further
palatal changes required are: elimination of the ectopterygoid,
already quite small in chiniquodontids; reduction of the extent of
the pterygoid posteriorly; and reduction (with elimination of the
quadrate-articular jaw joint) of the posterior extension of the
epipterygoid toward the articular region.

In sum, apart from expansion and elaboration of the brain cap-
sule and changes (already initiated in chiniquodontids) in the
auditory and jaw joint region, modifications needed to transform
a chiniquodontid skull into that of a primitive mammal are not
too numerous and not of major magnitude. I see nothing in the
structural pattern of such a form as Probainognathus which need
debar it from a position directly antecedent to a primitive mammal.
At present, chiniquodontids are known only from South America,
but it is not improbable that they will be found in the future to
have a wider distribution. It is not unreasonable to believe that in
this family we have a group of direct mammalian ancestors.

BIBLIOGRAPHY

Crompton, a. W.

1958. The cranial morphology of a new genus and species of ictido-

sauran. Proc. Zool. Soc. London, 130: 183-216.
1963. On the lower jaw of Diarthrognathus and the origin of the
mammalian lower jaw. Proc. Zool. Soc. London, 140: 697-
753.
Kermack, K. a.

1965. The origin of mammals. Sci. J., Sept. 1965: 66-72.
Kermack, K. A., and F. Musset.

1958a. The jaw articulation in Mesozoic mammals. Proc. XV Int.

Congr. Zool., Sec. V (8): 442-443.
1958b. The jaw articulation of the Docodonta and the classification of
Mesozoic mammals. Proc. Roy. Soc. London, ser. B, 148:
204-215.

18 BREVIORA No. 344

Parrington, F. R.

1955. On the cranial anatomy of some gorgonopsids and the synapsid
middle ear. Proc. Zool. Soc. London, 125: 1-40.
ROMER, A. S.

1965. Possible polyphylety of the vertebrate classes. Zool. Jb. Syst.,
92: 143-156.

1969a. The Brazilian Triassic cynodont reptiles Belesodon and Chini-
quodon. Breviora, Mus. Comp. Zool., No. 332: 1-16.

1969b. The Chanares (Argentina) Triassic reptile fauna. V. A new
chmiqiiodontid cynodont, Prohelesodun lewisi — cynodont an-
cestry. Breviora, Mus. Comp. Zool., No. 333: 1-24.
Simpson, G. G.

1927. Mesozoic Mammalia. IX. The brain of Jurassic mammals.
Amer. J. Sci., ser. 5, 14: 259-268.

1959. Mesozoic mammals and the polyphyletic origin of mammals.
Evolution, 13: 405-414.
Watson, D. M. S.

1913. Further notes on the skull, brain, and organs of special sense in
Diodemodon. Ann. Mag. Nat. Hist., ser. 8. 12: 217-228.

1951. Paleontology and Modern Biology. New Haven, Yale Univ.
Press, 216 pp.

(Received 3 December, 1969.)

'^'^'^^ DO NOT CiivCULATE

B R E V I O R ^Z'^r-

Mmseunti of Comparative Zoology A igyg

Cambridge, Mass. 27 February, 1970 Number ^RVARD

tmtVERsiTY

LOWER TRIASSIC (SCYTHIAN)
AMMONOIDS FROM NEPAL

Bernhard Kummel

Abstract: The Austrian Dhaula Himal Expedition to Nepal of 1963
reported a thin unit of limestone of Lower Triassic (Scythian) age con-
formable on Late Permian limestone. The Triassic units contain two am-
monoid faunas, the lower comparable to the "Meekoceras" beds of the
Himalayas and the second, containing Anasibirites, is of mid-Scythian age.
The lowest of these faunas is separated from fossiliferous Permian beds
by one meter of unfossiliferous limestone.

As a member of the Austrian Dhaula Himal Expedition of
1963. sponsored by the Austrian Himalayan Society, Dr. Gerhard
Fuchs (1964. 1967. 1968) studied and mapped a large area of
the Nepal Himalayas in the region of Dhaulagiri. An interesting
result of this study was the discovery of late Permian and early
Triassic strata in apparent conformity. Fuchs (1967: 18) sum-
marized the character of the Lower Triassic strata as follows:
"The Lower Trias forms a thin (10-25 m) easily recognizable
band. At the base there is a bed of dark limestone 1.5-2.5 m .thick
with ferruginous weathering. The lowermost part of this bed
is still Permian [Collection 108] 0-0.8 m above base, while in the
upper part the first ammonites are found 1.5-2.5 m above base.
Grey shales with a few limestone layers, then light, thin-bedded,
dense limestones and dark nodular limestone succeed.

"In the northern parts of Dolpo the basal bed is rich in Fe and
Mn showing a violet to brown colour. Here ammonites are the
only fossils while in the S also lamellibranchiata are found. The
fauna as well as the character of the sediments indicate bathyal or
even abyssal deposition, showing that the sea had deepened rapidly
at the beginning of the Mesozoic. The Lower Trias of Nepal is
similar to Painkhanda and Spiti. but differs from eastern Kumaon.
The latter seems to have been deposited near the southern shore
of the Tethys."

BREVIORA

No. 345

In June, 1967. I was able to examine in Vienna Dr. Fuchs'
collections from these Lower Triassic formations. The preserva-
tion is in general very bad and most specimens are distorted and
elliptical in shape. From this assemblage the best preserved speci-
mens of five collections were made available for study. A summary
of the locality and horizon data on these collections is given in
Table 1.

Collection
Number

106

107

96

95

110

TABLE I
Fossil localities of Lower Triassic of Nepal

Description

Five meters of light gray and blue-gray, platy limestone 14
meters above last bed with Permian fauna. At Kar, 3.7
kilometers WNW of Tukot (Barbung Khola).

A two-meter bed of thin platy, light gray limestone directly
overlying highest fossiliferous Permian limestone in same
section as collection 106.

3.5 meters of gray platy limestone 17.2 meters above last
bed with Permian fauna. Upper course of the valley E of
Terang (Barbung Khola).

Two meters of light gray, platy-layered limestone one meter
above last bed with Permian fauna, same section as
collection 96.

Seven meters of platy limestone with gray, clay-slate inter-
beds, upper Scythian, 4.3 m above base of measured section.
6 km N of Barbung village on the N running ridge (Bar-
bung Khola W side).

Two distinct Scythian ammonoid zones are represented in these
Nepal faunas. The uppermost, represented by collections 106 and
1 10, belong to the Anasibirites Subzone of the Owenites Zone and
is mid-Scythian in age. Collection 106 contains only Anasibirites
kiui^icuuts. Collection 110 contains Anasibirites /<ini>ianiis and
Hemiprionites typus. Both these genera are very characteristic of
the Anasibirites Subzone. Faunas of this subzone are known
from Kotal-e-Tera. Afghanistan (Kummel and Erben, 1968), the
Upper Ceratite limestone of the Salt Range (Waagen, 1895; Kum-
mel, 1966), from Byans in the Himalayas (v. Krafft and Diener,

1970 LOWER TRIASSIC AM MONOIDS NEPAL 3

1909), from Timor (Welter, 1922), South China (Chao, 1959),
western Australia (identihed on basis of photographs sent by B. E.
Balnie), from Japan (data summarized by Bando, 1964a). Pri-
morye region (Zakharov, 1968), and western United States
(Mathews'^ 1929; Smith. 1932; Kummcl. 1954).

The second Scythian horizon is that represented by collections
95, 96, and 107 and is summarized as follows:

Collection 95

"Ophiceras" cf. obtiisoaniiulatmu Diener
Prionolobus sp. indet. 11
Koninckites sp. indet.

Collection 96

Prionolobus cf. Ungtiense v. KrafTt

Collection 107

Gyronites jreqiieus Waagen

Prionolobus sp. indet. I

Flemingites sp. indet. I

Flemingites sp. indet. II

Koninckites sp. indet.

Koninckites cf. kraffti Spath

Proptycliites sp. indet.

Anakashmirites sp. indet.

"Ophiceras' cf. obtusoangulatum Diener

The genera and species of these three collections are charac-
teristic forms of the Lower Ceratite limestone and the Ceratite
marls of the Salt Range and the so-called "Meekoceras" beds of
Spiti, Painkhanda, and Byans of the Himalayas. That is, these
faunas are younger than those of the lowest Scythian Otoceras-
Ophiceras Zone and older than those of the Owenites Zone. Spath
(1934: 27) recognized a number of zones within two divisions
(Gyronitan and Flemingitan ) for this part of the geologic column.
I am in the process of revising the ammonoids of these horizons
from the Salt Range and the Himalayas. It is premature to have a
new zonal nomenclature, but extensive changes are indicated. For
the purpose of the present discussion the most important thing
about these three faunas is that they are not of the lowest Scythian
Otoceras-Ophiceras Zone. These fossiliferous beds are separated
by approximately one meter of unfossiliferous limestone from a
horizon that has yielded a varied Permian fauna. The brachiopods
of this Permian fauna have been described by Waterhouse (1966).

4 BREVIORA No. 345

who concluded they were correlative with the Upper Permian
Cyclolobus Zone. The corals from these Permian faunas have
been reported on by Fliigel (1964, 1966). The latest summary
of the Permo-Triassic formations of the Salt Range (Kummel and
Teichert, 1970, eds.) concludes that most probably one or pos-
sibly two stages are missing at the top of the Permian sequence
in that area. It would appear that the Permian of Nepal likewise
does not include the youngest beds of this system. The one meter
bed of unfossiliferous limestone that separates the Permian faunas
from collections 95, 96, and 107 may include the Otoceras-Ophi-
ceras Zone of lowest Scythian age, but this remains an uncertainty
until fossils are found.

I am particularly grateful to Dr. Fuchs for giving me the oppor-
tunity to describe these specimens. Dr. R. Sieber of the Geologische
Bundesanstalt was also most helpful in the arrangements of this
loan.

SYSTEMATIC PALEONTOLOGY

Class CEPHALOPODA Cuvier, 1797
Subclass AMMONOIDEA Zittel, 1884

Family Gyronitidae Waagen, 1895

Genus Gyronites Waagen, 1895

Type species, Gyronites frequens Waagen, 1895

Gyronites sp. indet.

Plate 3, figures 9, 10

A small, elliptical specimen of approximately 26 mm in diameter
and with an umbilical diameter of 8 mm. The conch is com-
pressed, with gently arched lateral areas. The venter is broad and
flat.

This genus is the predominant member of the fauna of the
Lower Ceratite limestone in the Salt Range of West Pakistan. The
specimen recorded here differs from the typical Salt Range species
in being slightly more compressed. This difference, however, may
be merely a reflection of the preservation. The generic assign-
ment, however, does appear to be correct.

Occurrence. — Collection 107

Repository. — Geologische Bundesanstalt, Vienna.

1970 LOWER TRIASSIC AMMONOIDS NEPAL 5

Genus Prionolohiis Waagcn, 1 H95
Type species, Prionolohiis atavus Waagen, 1895; lectotype
chosen by Spath, 1934, p. 96

Prionolohiis cf. liiii^tiense v. Krafft

Plate 2, figure 4

Meekoceras lingtiense v. Krafft, /'/; v. Krafft and Diener, 1909, p. 25, pi.

2, fig. 1; Diener, 1915, p. 193.
Prionolohiis lingtiense, — Spath, 1934, p. 102.

A single, distorted, elliptical specimen of approximately 38 mm
in diameter; the umbilicus measures about 9 mm in diameter. The
lateral areas are broadly convex, ending at an angular ventral
shoulder. The venter is broad and arched. No suture is visible.
Among the many species from the Lower Triassic of the Salt
Range and the Himalayas, this specimen is most similar to P. ling-
tiense V. Krafft. This similarity is most pronounced in the broad,
arched venter, and it is on this character that it diflfers from the
specimens described here as Prionolobus sp. indet. I and II.

Occurrence. — Collection 96

Repository. — Geologische Bundesanstalt, Vienna.

Prionolobus sp. indet. I

Plate 2, figure 1

A single specimen of approximately 41 mm in diameter and an
umbilical diameter of 13 mm. The lateral areas are very gently
arched and the venter is narrow and flat. The suture is vaguely
visible, showing two lateral lobes and a short auxiliary series. The
poor preservation does not allow a specific determination, but
assignment to this genus appears fairly secure.

Occurrence. — Collection 107

Repository. — Geologische Bundesanstalt, Vienna.

Prionolobus sp. indet. II

Plate 1, figure 4

A highly distorted, elliptical specimen whose most distinctive
feature is a narrow flat venter. This specimen is clearly not con-
specific with that recorded here (PI. 1, fig. 1) as Prionolobus sp.

6 BREVIORA No. 345

indet. I. The primary difference is the greater umbilical diameter
of this species and the greater convexity of the lateral areas. The
extreme elliptical distortion of the specimen precludes specific
identification.

Occurrence. — Collection 95

Repository. — Geologische Bundesanstalt. Vienna.

Family Flemingitidae Hyatt. 1900

Genus Flemingites Waagen. 1892

Type species, Ceratites iiemingianus de Koninck. 1863

Flemingites sp. indet. I

Plate 2, figure 2

A distorted, elliptical specimen of approximately 70 mm in
diameter. The umbilicus measures about 31 mm in diameter, the
adoral whorls 25 mm in height and 15 mm in width. The lateral
areas are broadly convex and converge on a narrowly rounded
venter. No shell or surface markings are preserved. The suture is
incompletely preserved but does not show two lateral lobes and a
short auxiliary series.

The poor preservation of this specimen does not allow specific
comparison. Among the variety of forms assigned to Flemingites,
this specimen bears most resemblance to Flemingites glaber
(Waagen, 1895: 188, pi. 11, figs. 2a-d).

Occurrence. - — Collection 107

Repository. — Geologische Bundesanstalt, Vienna.

Flemingites sp. indet. II

Collection 107 has a portion of phragmocone that measures
about 80 mm in length, with a whorl height of about 40 mm. The
sutures are well preserved and typically flemingitid in aspect. The
shape of the whorl section and pattern of the suture suggests a
strong similarity to Flemingites compressus Waagen (1895: 202,
pi. 15, fig. 1; pi. 16, fig. 1). The Nepal specimen, however, shows
no ribbing of any kind, but this may be the result of preservation.

Occurrence. — Collection 107

Repository. — Geologische Bundesanstalt Vienna.

1970 LOWER TRIASSIC AMMONOIDS NEPAL 7

Family Paranoritidae Spath, 1930

Genus Koninckites Waagen, 1895

Type species, Koninckites vetustus Waagen, 1895

Koninckites sp. indet.

Plate 1, figures 2, 5

This genus is represented by two poorly preserved, distorted,
elliptical specimens. The first (Pi. 1, fig. 2) is approximately
68 mm in diameter and the second (PI. 1, fig. 5) about 73 mm.
The umbilici have diameters of about 8 mm and 9 mm respec-
tively. The flanks are gently arched and the venter narrowly
rounded. The suture, though somewhat weathered, clearly shows
the two prominent lateral lobes and the developed auxiliary series.
The pattern of the suture is similar to that of species of this genus.

The genus Koninckites is especially well known from Lower
Scythian strata of the Salt Range and the Himalayas. In the Salt
Range the genus is a conspicuous member of the fauna from the
Lower Ceratite limestone of Waagen (1895), which is the basal
part of the Mittiwali Member of the Mianwali Formation of Kum-
mel ( 1966). In the Himalayas species of this genus occur in the
so-called "Meekoceras" beds.

Occurrence. — Collection 107

Repository. — Geologische Bundesanstalt, Vienna.

Koninckites cf. kraffti Spath
Plate 1, figure 3

Meekoceras varaha Diener, in, Krafft and Diener, 1909, p. 17, pi. 2, figs.

4a-d, 2, 3, 5, 6v.; pi. 14, figs. 7, 8; Diener, 1915, p. 195.
Koninckites kraffti Spath, 1930, p. 28; Spath. 1934, p. 155, fig. 43c.

One of the better preserved and least distorted of the specimens
can be allied to Koninckites kraffti. The specimen has a diameter
of approximately 43 mm and is all phragmocone. The umbilicus
is small, lateral areas arched, and venter rounded. The suture,
though weathered, has two lateral lobes and a well-developed
auxiliary series.

Occurrence. — Collection 107

Repository. — Geologische Bundesanstalt, Vienna.

S BREVIORA No. 345

Family Proptychitidae Waagen, 1895

Genus Proptychites Waagen, 1892

Type species, Ceratites lawrencianus de Koninck, 1863

Proptychites sp. indet.

Plate 1, figure 1

This identification is the least satisfactory of this small assem-
blage of fossils. The specimen, like most of those in the collection,
is distorted and elliptical and no suture is preserved. The assign-
ment to the genus Proptychites is suggested by the rounded venter
and steep umbilical walls. The whorl section is not very inflated,
but this \ attribute to a slight crushing of the specimen. The above
characters fit well with those of a young specimen of Proptychites
but not with those of any other genus of this approximate horizon.

Occurrence. — Collection 107

Repository. — Geologische Bundesanstalt, Vienna.

Family Kashmiritidae Spath, 1934

Genus AnakasJunirites Spath, 1930

Type species, Danubites nivalis Diener, 1897

Anakashmirites sp. indet.

Plate 2, figure 7

A distorted, elliptical, highly weathered specimen. It is an
evolute form with a whorl whose height is only slightly more than
the width. The venter is broad and rounded. Along a small part
of the whorl two prominent radial ribs are preserved. The generic
assignment of this form appears fairly secure, but the extremely
poor preservation does not allow any chance of specific com-
parisons.

Occurrence. — Collection 107

Repository. — Geologische Bundesanstalt, Vienna.

Family Prionitidae Hyatt. 1900
Genus Heiniprionites Spath, 1929
Type species, Goniodiscus typus Waagen, 1895.

Hemiprionites typus (Waagen)

Plate 3, figures 6-8

Gonoidiscus typus Waagen, 1895. p. 128, pi. 9. figs. 7, 8a, b, 9, c, 10;
Diener, 1915, p. 135; Mathews, 1929, p. 31, pi. 5, figs. 12-21.

1970 LOWER TRIASSIC AMMONOIDS NEPAL 9

Lccanitcs mliocnsi.s Yehara, 192(1. p. 40, pi. 13. figs. 6. 7.
Meekoceras (Gyronites) sawataniim Yehara. 1926. p. 39. pi. 13. figs. 3-5.
Ophiceras tahoensis, — Yehara, 1928. p. 161, pi, 15, figs. 4, 4a, 4b, 5, 7.
Meekoceras sawataniim. — Yehara. 1928, p. 159, pi. 13, figs. 6, 7, 7a, 8,

8a, 9.
Meekoceras katoi Yehara. 1928, p. 157, pi. 15, figs. 3, 3a.
Meekoceras morianum Yehara, 1928, p. 155, pi. 15, figs. 2, 2a, 3, 4, 5.
Meekoceras kiiharaiuim Yehara, 1928, p. 155, pi. 14, figs. 5, 5a.
Goaiodisciis anicricaniis Mathews, 1929, p. 32, pi. 5, figs. 22-27.
Goniodisciis walcotti Mathews, 1929, p. 32, pi. 6, figs. 1-5.
Goniodisciis iitahensis Mathews, 1929, p. 33, pi. 6, figs. 29-31.
Goniodisciis shiiiuardi Mathews, 1929, p. 33, pi. 6, figs. 11-14.
Goniodisciis slocoini Mathews. 1929, p. 34. pi. 6. figs. 15-17.
Goniodisciis ornatiis Mathews, 1929, p. 34, pi. 6, figs. 6-10.
Goniodisciis biitleri Mathews, 1929, p. 35, pi. 6, figs. 18-21.
Heiniprionites />•/>//*, — Spath, 1929, p. 270; Spath, 1934, p. 330, fig. 114.
Anasibirites (Goniodisciis) ornatiis, — Smith, 1932, p. 75, pi. 80, figs.

11,12.
Anasibirites (Goniodisciis) sinithi. — Smith, 1932. p. 76, pi. 80, figs. 13-15
Anasibirites (Goniodisciis) typiis, — Smith, 1932, p. 76, pi. 31, figs. 11, 12.
Anasibirites (Goniodisciis) iitahensis, — Smith, 1932, p. 77, pi. 80, figs.

9, 10.
Meekoceras shikokiiense Shimizu and Jimbo, 1933, p. 15.
Heiniprionites americaniis, — Spath, 1934, p. 333.
Heiniprionites M'olcotti, — ^ Spath, 1934, p. 334.
Heiniprionites iitahensis, — Spath, 1934, p. 334.
Heiniprionites ornatiis, — Spath, 1934, p. 335.
Heiniprionites katoi, — Bando, 1964a. p. 87, pi. 1. figs. 4. 9a-b, 12a-c;

Bando, 1964b, p. 7; Bando, 1966, p. 8.
Heiniprionites tahoensis, — Bando, 1964a, p. 88, pi. 1, figs. 7a-b, 8, pi. 2,

figs. 4a-b, 18; Bando, 1964b, p. 7; Bando, 1966, p. 8.
Heiniprionites morianiis, — Bando, 1964a. p. 89, pi. 2. figs, la-c, 2, 9a-b:

Bando, 1964b, p. 7; Bando, 1966, p. 8.
Hemiprionites kiiharanits, — Bando, 1964a, p. 90, pi. 2, figs. 12a-b; Bando,

1964b, p. 7; Bando, 1966, p. 8.
Hemiprionites kuharaniis iyoniis Bando, 1964a, p. 91. pi. 1, figs. 14, 18,

pi. 2, figs. 15a-b; Bando, 1964b, p. 7; Bando, 1966, p. 8.
Heiniprionites sawataniis, — Bando. 1964a, p. 92, pi. 1. figs. 5a. b, pi. 2,

figs. 8a-b, 13a-b; Bando, 1964b, p. 7; Bando, 1966, p. 8.
Hemiprionites shikokuensis, — Bando, 1964a. p. 93. pi. 1. figs. 13, 20,

22, pi. 2, figs. lOa-c: Bando 1964b, p. 7; Bando. 1966, p. 8.
Hemiprionites shimizni Bando, 1964c, p. 335, pi. 49, figs. 1, 2.
Hemiprionites sp. Bando, 1964c, p. 336, pi. 49, fig. 4.

This spscies is represented by three small, generally poorly pre-
served specimens. The conch, however, is very distinctive. It is
compressed, involute, with a flattened venter and angular ventral

10 BREVIORA No. 345

shoulders. This species was first described by Waagen (1895) on
specimens from the Upper Ceratite limestone (= upper part of
Mittiwali Member of Mianwali Formation, Kummel, 1966) at
Chhidru, Salt Range, West Pakistan. The four specimens avail-
able to Waagen (1895) are very poorly preserved. The lectotype
(Waagen, 1895, pi. 9, figs. 8a, b, c) is a phragmocone of approxi-
mately 29.0 mm in diameter with only one side of the conch and
the venter preserved. Because of the poor preservation, the serra-
tions on the ventral shoulders and the ribs crossing the venter, as
indicated in Waagen's illustration, are doubtful. Only part of the
suture is visible and this is highly weathered. One of the paralecto-
types (Waagen, 1895, pi. 9, fig. 7) is an extremely poorly pre-
served specimen in which the only clear feature preserved is the
flattened venter. The other paralectotype (Waagen, 1895, pi. 9,
fig. 10) is incorrectly represented. The ornament shown in
Waagen's illustrations is nothing more than faint growth lines;
likewise the specimen bears no serrations on the ventral shoulders.

1 have available approximately thirty topotype specimens of this
species. Though the preservation leaves much to be desired it is
clear that there is much variation in the width of the whorls, width
of the fiat venter, and umbilical diameter. The Nepal specimens
recorded here are directly comparable to a number of the speci-
mens in my Salt Range collection.

Hemiprionites is almost always found in association with Anasi-
birites. The latter genus is a ribbed form which shows fantastic
intraspecific variation (Kummel and Erben, 1968). A large num-
ber of species has been established for the genus Anasibirites. De-
tailed examination of large collections from Afghanistan and Timor
clearly demonstrates the wide range of intraspecific variation. The
collection of topotype specimens from Chhidru, Salt Range, West
Pakistan, clearly shows that Hemiprionites typus is also a species
displaying a significant degree of variation in conch parameters.
A large number of species has been established on the basis
of small and fragmentary specimens from Japan by Yehara ( 1926,
1928), Shimizu and Jimbo (1933), and Bando (1964a). All
these species fall well within the range of variation found in Hemi-
prionites typus from the Salt Range. I can find no basis for sepa-
rating them from the Salt Range species. The new species, Hemi-
prionites dunajcnsis Zakharov (1968) from the Primorye Region,
appears likewise to be nearly identical with H. typus. I can find
no basis for separating this form from the type species. The
Anasibirites fauna of Fort Douglas, Utah, has yielded a well-
preserved fauna including Hemiprionites. Mathews (1929), who

1970 LOWER TRIASSIC AMMONOIDS NEPAL 11

made the original study, and Smith ( 1932), who was first revisor,
established a large number of new species of Hemiprionites, at
the same time assigning some of their specimens to the type
species. Here again analysis of topotype material shows that all
these species are based on characteristics that are highly variable.
The fossil beds from which these specimens are derived strongly
suggest a unified, interbreeding population. I find no criteria
within these forms to separate them from Hemiprionites typus.

Though Hemiprionites is geographically fairly widespread there
are relatively few species. Within Tethys, in addition to Hemi-
prionites typus, there is H. hun;^eri Kummel (in Kummel and
Erben, 1968) from the Anasibirites Subzone at Kotal-e-Tera,
Afghanistan. This species is distinctive in the evolute nature of
the conch and the closely spaced transverse, thin ribs on the
venter. I find the Timor Hemiprionites timorensis somewhat of an
enigma. On the one hand, the general conch morphology of this
species is very similar to that of H. typus, on the other, Spath
(1934: 332) writes of transitional forms to Anasibirites multi-
formis. Hopefully a larger sample of specimens of this species will
become available. The Spitsbergen Hemiprionites garwoodi Spath
(1934: 336) is another distinctive species in its irregular lateral
ribs.

Occurrence. — Collection 1 10

Repository. — Geologische Bundesanstalt, Vienna.

Family Sibiritidae Mojsisovics, 1896
Genus Anasibirites Mojsisovics, 1896
Type species, Sibirites kingianus Waagen, 1895

Anasibirites kingianus (Waagen)

Plate 3. figures 1-5

Sibirites kingianus Waagen. 1895. p. 108. pi. 8, figs. la-c. 2a-c.

Anasibirites sp. Bando, 1964c. p. 337, pi. 49, figs. 3a-c.

Anasibirites kingianus, — Kummel and Erben. 1968, p. 135. pi. 20, figs.

6, 7, pi. 22, figs. 12-17, pi. 23, figs. 1-18.
Anasibirites multiformis Nakazawa and Bando, 1968, p. 96. pi. 5, figs. 2-5.
Anasibirites nevolini Zakharov. 1968. p. 131, pi. 25. figs. 4-5.

This genus and species was extensively reviewed by Kummel
and Erben (1968) in their discussion of a rich mid-Scythian fauna
from Kotal-e-Tera, Afghanistan. The genus is widely distributed
and in many places has yielded an abundance of well-preserved

12 BREVIORA No. 345

specimens. The characteristic ornamentation of the genus has
given the typologist a field day in the estabhshment of new species.
Analysis of the large collection from Kotal-e-Tera, Afghanistan,
has shown that the pattern of ornamentation is highly variable and
gradational in nearly every respect. The conclusion was that there
is only one species of Auasibirites. The specimens recorded here
from Nepal are typical forms, differing in no way from previously
described and illustrated specimens considered to be Anasihirites
ki}igicmus. This genus and species is now known from a number
of localities in Tethys and the circum-Pacific region. It is well
represented at Kotal-e-Tera, Afghanistan, as mentioned above; in
the Upper Ceratite limestone of the Salt Range, West Pakistan; in
the Himalayas of northern India; in South China; Timor; western
Australia; Japan; the Primorye Region; and in western United
States.

Occurrence. — Collection 106 (PI. 3, fig. 1), collection 110
(PI. 3, figs. 2-5).

Repository. — Geologische Bundesanstalt, Vienna,

Genus Uncertain
"Ophiceras"? cf. obtusoangulatum Diener

Plate 2, figures 3, 5, 6

Ophiceras ohtiisoangiilatum Diener, in v. Krafft and Diener, 1909, p. 82,
pi. 37, fig. 6.

There are two specimens in the collection that are very similar
to Diener's species from the "Meekoceras" beds at Lilang in the
Himalayas. The species was established on the basis of a single
specimen. I cannot accept Diener's assignment of his species to
the genus Ophiceras, primarily on the tabulate nature of the venter.
At the same time there are no other Lower Scythian genera to
which this form can be assigned. It is most probable that this
form represents a new genus, but on the basis of the material
available it would be unwise to establish a new genus at this time.

The Nepal specimens are both distorted and elliptical with the
inner whorls especially poorly preserved. The conches are widely
evolute, the whorls compressed, converging toward a narrow, sub-
tabulate venter bordered by subangular shoulders. The suture con-
sists of a lar^e first lateral lobe, a smaller second lateral lobe, and
a short auxiliary series on the umbilical shoulder and wall.

Occurrence. — Collection 95 (PI. 2, fig. 3), collection 107
(PI. 2, figs. 5, 6).

Repository. — Geologische Bundesanstalt, Vienna.

970

LOWER TRIASSIC AM MONOIDS NKPAL

13

REFERENCES

Bando, Y.

1964a. The Triassic stratigraphy and ammonite fauna of Japan. Sci.

Rep. Tohoku Univ. Sendai, Ser. 2 (Geol.), 36(1): 1-137.
1964b. Contribution to the Lower Triassic biostratigraphy of Japan.

Mem. Fac. Liberal Arts Ed., Kagawa Univ., Pt. 2, No. J 28:

1-14
1964c. On some Lower and Middle Triassic ammonoids from Japan.

Trans. Proc. See. Japan, N.S., No. 56: 332-344.
1966. A note on the Triassic ammonoids of Japan. Mem. Fac.

Liberal Arts Ed., Kagawa Univ., Pt. 2, No. 138: 1-19.
Chao, Kingkoo

1959. Lower Triassic ammonoids from Western Kwangsi, China.

Palaeont. Sinica, 145: 1-355.

DiENER, C.

1897. Himalayan fossils. The Cephalopoda of the Lower Trias.

Mem. Geol. Surv. India, Palaeont. Indica, Ser. 15, 2(1): 1-191.
1915. Fossilium Catalogus. I Animalia. Pt. 8, Cephalopoda

Triadica. Berlin: W. Junk, 369 pp.

Korallenfaunen aus dem Palaozoikum West-Nepals. Ver-
handl. Geol. Bundesanst., 1: 15-16.

Palaozoische Korallen aus der Tibetischen Zone von Dolpo
(Nepal). Sonderb., Jahrb. Geol. Bundesanst., 12: 101-119.

Beitrag zur Kenntnis des Palaozoikums und Mesozoikums der
Tibetischen Zone in Dolpo (Nepal-Himalaya). Verhandl.
Geol. Bundesanst., 1: 6-15.

Zum Bau des Himalaya. Denkschr. Osterreich. Akad.
Wiss., 113: 1-211.

The geological history of the Himalayas. 23rd Int. Geol.
Congr. Prague, 3: 161-174.

Flugel,H.
1964.

1966.

FucHs, G.
1964.

1967.
1968.

Hyatt, A.
1900.

Cephalopoda. /// Zittel, K.A. v.. Textbook of Paleontology
(trans, and ed. by C. R. Eastman), London: Macmillan, 1:

502-604.

DE KONINCK. L.

1863. Descriptions of some fossils from India, discovered by

Dr. A. Fleming, of Edinburgh. Quart. Jour. Geol. Soc.

London, 19(1): 1-19.
Kraft, A. v., and C. Diener

1909. Himalayan fossils. Lower Triassic Cephalopoda from Spiti,

Malla Johar, and Byans. Mem. Geol. Surv. India, Palaeont.

Indica, Ser. 15, 6(1): 1-186.

14 BREVIORA No. 345

KUMMEL, B.

1954. Triassic stratigraphy of southeastern Idaho and adjacent

areas. Prof. Pap. U.S. Geol. Surv., 254 H: 165-194.
1966. The Lower Triassic formations of the Salt Range and Trans-
Indus Ranges. West Pakistan. Bull. Mus. Comp. Zool.,
134(10): 361-429.
KuMMEL, B.. and H. K. Erben

1968. Lower and Middle Triassic cephalopods from Afghanistan.
Palaeontographica, 129(A): 95-148.
KuMMEL, B., and C. Teichert

1970. Stratigraphy and paleontology of the Permian-Triassic boun-
dary beds. Salt Range and Trans-Indus Ranges. West
Pakistan. //; Kummel. B., and C. Teichert, eds.. Stratigraphic
Boundary Problems: Permian and Triassic of West Pakistan.
Spec. Publ. Univ. Kansas Dept. Geol., 4:00-00.
Mathews, A. A. L.

1929. The Lower Triassic cephalopod fauna of the Fort Douglas
area. Utah. Mem. Walker Mus., 1: 1-46.
Nakazawa, K., and Y. Bando

1968. Lower and Middle Triassic ammonites from Portuguese Timor
(Palaeontological study of Portuguese Timor, 4). Mem.
Fac. Sci. Kyoto Univ.. Ser. Geol. Min.. 34(2): 83-113.
Shimizu, S., and N. Jimbo

1933. On the Triassic ammonites from Tao, Uonashi-mura (Shiro-
kawa-cho), Higashiuwa-gun, Ehime Prefecture. Chikyu (The
Globe), 19(1): 10-31.

Smith. J. P.

1932. Lower Triassic ammonoids of North America. Prof. Pap.
U.S. Geol. Surv., 167: 1-199.
Spath, L. F.

1929. Corrections of cephalopod nomenclature. The Naturalist
(London), No. 871: 269-271.

1930. The Eo-Triassic invertebrate fauna of East Greenland. Medd.
cm Gr0nland, 83: 1-90.

1934. Catalogue of the fossil Cephalopoda in the British Museum
(Natural History). Part IV, The Ammonoidea of the Trias.
London, pp. 1-521.

Waagen, W.

1892. Vorlaufige Mittheilung itber die Ablagerungen der Trias in

der Salt-Range (Punjab). Jahrb. K.K. Geol. Reichsanst.,

42(2): 1-10.
1895. Salt Range fossils. Fossils from the Ceratite Formation.

Pt. I, Pisces-Ammonoidea. Mem. Geol. Surv. India, Palaeont.

Indica. Ser. 13, 2: 1-323.
Waterhouse, J. B.

1966. Lower Carboniferous and Upper Permian brachiopods from

Nepal. Jahrb. Geol. Bundesanst., 12: 5-95.

970

LOWER TRIASSIC AMMONOIDS NEPAL

15

Welter, O. A.

1922.

Yehara, S.
1926.

1928.

Zakharov.
1968.

Die Ammoniten der Untcren Trias von Timor. Paliiont.
Timor, E. Schweizerbart'sche Verlag., Stuttgart: 11(19):
83-154.

On the Lower Triassic ammonites from Nomura. lyo Province.
Jour. Geol. Soc. Tokyo. 32(386): 37-40.
The Lower Triassic cephalopod and bivalve fauna of Shikoku.
Trans. Jap. Jour. Geol. Geog.. 5: 135-172.
Yu. D.
Biostratigrafiya i ammonoidei nizhnego Triasa yuzhnogo
Primorya. ( Biostratigraphy and ammonoids of the Lower
Triassic from the southern Primorye. ) Akad. Nauk SSSK.
Sibirskoe Otdel.. Parnevostochnyi filial, Geol. Inst., 175 pp.

(Received 4 December 1969.)

16

BREVIORA

No. 345

X

%

^S'*^ fK^ ' ^ Y*" " * ,

1970 LOWER TRIASSIC AMMONOIDS NEPAL 17

EXPLANATION OF PLATE 1

Fig. 1. Proptychites sp. indet. I, from collection 107, X 1

Fig. 2. Koninckites sp. indet.. from collection 107, x I

Fig. 3. Koninckites cf. kraffti Spath, from collection 107, x 1

Fig. 4. Prionolohiis sp. indet. II, from collection 95, x 1

Fig. 5. Koninckites sp. indet., from collection 95. x I

AH specimens are from Lower Triassic of Nepal and are deposited in
the Geologische Bundesanstalt, Vienna.

18

BREVIORA

No. 345

970 LOWER TRIASSIC AMMONOIDS NEPAL 19

EXPLANATION OF PLATE 2

Fig. 1. Prionolohus sp. indet. I, from collection 107, X 1

Fig. 2. Flemingites sp. indet. I, from collection 107, X 1

Fig. 3. "Ophiceras" cf. ohtitsoangiilatum Diener, in v. Kraflft and
Diener (1909) from collection 95, X 1

Fig. 4. Prionolobiis cf. lingtiense v. Krafft, in v. Krafft and Diener
(1909) from collection 96, xl

Fig. 5, 6. "Ophiceras" cf. obtusoangiilatiim Diener, in v. Krafft and
Diener (1909) from collection 107, x 1

Fig. 7. Anakashmirites sp. indet., from collection 107, X 1

All specimens are from Lower Triassic of Nepal and are deposited in
the Geologische Bundesanstalt, Vienna.

20

BREVIORA

No. 345

:?>^

<S*\

, ^'^■^r>.'. ■^.,

i? 7

^ >

^'"^■- i. ri^-.-^

10

970 LOWER TRIASSIC AMMONOIDS NEPAL 21

EXPLANATION OF PLATE 3

Figs. 1-5. Anasihirites kingianiis (Waagen). Fig. 1 from collection 106,
XI. Figs. 2-5 from collection 110, Fig. 2, X 2, Figs. 3-5, X 1

Figs. 6-8. Hemiprionites sp. indet., from collection 110, x 1-5

Figs. 9- 10. Gyronites typiis, from collection 107, X 1.5

All specimens are from Lower Triassic of Nepal and are deposited in
the Geologische Bundesanstalt, Vienna.

\ -^ WW «-, w LIBRARY

APR *^ lQ7n

B R E V I O rJkI

UNiVERSITYj

Meseem of Comparative Zoology

Cambridge, Mass. 3 April, 1970 Number 346

NEW FROGS OF THE GENUS BATRACHYLODES
(RANIDAE) FROM THE SOLOMON ISLANDS

Walter C. Brown^ and Fred Parker^

Abstract. Prior to 1963, only two lowland species of the genus Batrachy-
lodes (Ranidae) were recognized from the Solomons archipelago. In the
past few years intensive sampling of the populations of frogs in the moun-
tain areas of Bougainville has revealed the presence of eight species of
this endemic genus. Brief note is also made of their known geographic
distribution in the archipelago as well as their zonal distribution in the
mountains.

INTRODUCTION

In this paper, the second based on the extensive collections of
amphibians obtained by the junior author on Bougainville Island,
Solomon Islands, new species of the endemic ranid genus Batrachy-
lodes are described.

The genus was set up by Boulenger (1887, p. 337) to accom-
modate a single species, vertebralis. Stemfeld (1921, p. 435)
described two frogs from Buka Island which he regarded as micro-
hylids. He named these Sphenophryne wolfi and Chaparina fred-
erici, each based upon a single specimen. Mertens (1929, p. 266)
sank both in the synonomy of Batrachylodes vertebralis. Brown
and Myers (1949, p. 379) described B. trossulus from Bougain-
ville Island.

The collections made by the junior author on Bougainville Island
during the period 1961-63 and 1966 contain samples of eight
species. Re-examination of the types of Chaperina frederici and
Sphenophryne wolfi confirms the fact that the former is synony-
mous with B. vertebralis (northern Bougainville population). How-
ever, our present large series indicates that B. wolfi is distinct from

1 Division of Systematic Biology, Stanford University, and Menlo Col-
lege, Menlo Park, California; 2 Australian Government Service.

2 BREVIORA No. 346

B. vertebraUs. Thus, three species of this genus have been pre-
viously recognized and described. The other five species are new.

All collections previously assigned to B. vertebraUs from various
islands in the Solomons Archipelago should now be re-examined
in terms of our present knowledge. Some of these specimens, usu-
ally representing very small samples, may actually belong to some
other of the large-disked species.

Measurements, determined to the nearest 0.1 mm, with a Helios
dial caliper, are based on preserved specimens. Snout-vent length
is the distance from the tip of the snout to the vent with the speci-
men held flat, ventral surface down; the tibia length is the length
of the bone of the lower leg; the head length is the distance from
the tip of the snout to the posterior edge of the tympanum; head
breadth is measured at the angle of the jaws; diameter of the eye is
from the anterior to the posterior edge of the socket; first and
fourth finger lengths from the tip to the base of the subarticular
tubercle; third and second finger lengths from the tip to the base of
the second articular tubercle; interorbital distance is breadth of the
bone between the eyes. Methods of determining other measure-
ments such as breadth of disks are probably not subject to much
variation.

Specimens in the collections of various institutions are identified
as follows: Museum of Comparative Zoology, Harvard University
(MCZ); Division of Systematic Zoology, Stanford University
(SU) ; British Museum of Natural History (BMNH); Senckenberg
Museum, Frankfurt am Main (SMF); and, California Academy of
Sciences (CAS).

DESCRIPTIONS OF SPECIES

The genus can, arbitrarily, be conveniently divided into two
groups of species: (1) Those with broadly dilated finger disks (ex-
cept for the inner finger), the breadth of the disk of the third finger
is very rarely less than 35 per cent the length of the third finger as
measured to the base of the second subarticular tubercle, much
more broadly dilated than the largest toe disk; (2) Those species
in which the tips of the fingers are not or but moderately dilated,
the breadth of the disk of the third finger very rarely more than
one-third the length of the third finger, less than to about equal to
the breadth of the largest toe disk. The first group, with the ex-
ception of B. vertebraUs, are all relatively large species; the second
group are small species.

1970 NEW BATRACHYLODES 3

Small-disked species
Batrachylodes trossulus Brown and Myers

Batrachylodes trossulus Brown and Myers, 1949, Journ. Washington
Acad. Sci., 39: 379-80; Torokina, Bougainville I. (type in United States
National Museum).

Definition. A small Batrachylodes, snout-vent length for 19 ma-
ture males ranging from 18.1 to 20.8 mm (Table 1), and for 20
mature females from 18.1 to 22.7 mm (a female 17.3 mm in snout-
vent length is immature) ; tips of fingers scarcely to slightly dilated,
ventral part delimited by a circummarginal groove; first finger
shorter than second which is slightly shorter than to equal to fourth
(Fig. 1 ) ; breadth of disk of third finger about 15 to 25 per cent of
length of third finger; disks of toes larger than those of fingers, toes
with minute webs at base; hindlimbs relatively long, length of tibia
about 48 to 55 per cent of snout-vent length; breadth of head
about 62 to 73 per cent of length of tibia (Table 2). More than
100 specimens were examined, including MCZ 74824-79.

Color. A blackish band with a smooth dorsal margin extending
from the tip of the snout through the eye and ear to the groin or
nearly to the groin, becoming narrower on the posterior part of the
trunk, under surface of head, throat and usually pectoral region
dusky to brown or blackish; dorsum grayish cream to reddish or
brownish gray, more or less uniform, or with darker markings,
most frequently a small to moderate, dark chevron in the region of
the forelimbs.

Distribution. Known from both northern and southern collecting
sites of Bougainville Island, as well as the type locahty, Torokina,
on the central west coast, and possibly Choiseul Island (see Brown
and Myers, 1949). It is recorded at elevations from near sea level
to about 4000 feet.

Batrachylodes minutus sp. nov.

Holotype. MCZ 41391, an adult female, collected by Fred
Parker in the Aresi area, at 2000 feet elevation, on Bougainville
Island, Solomon Islands, 14 April 1963.

Paratypes. (all collected by Fred Parker on Bougainville Is-
land) MCZ 41390, from the Aresi area, at 2000 feet elevation, 14
April 1963; MCZ 55534, from Melilup, between 2500 and 3500
feet elevation, collected 14 May 1966; MCZ 55037-49, 55521-28,
55533-40, 55546, 62020-23, SU 23989, 23991-92, from Mutahi,
between 2200 and 3500 feet elevation, collected between 9 and

4 BREVIORA No. 346

20 May 1966. MCZ 36924 (9 September 1962), MCZ 41386
(25 December 1962), and SU 21852 (30 December 1962), from
Pipekei area, between 1000 and 2000 feet elevation.

Diagnosis. A small Batrachylodes, measuring about 16 to 19.5
mm in snout-vent length for 20 mature specimens available; disks
of fingers not or scarcely dilated (Fig. 2, left), ventral part rather
vaguely delimited by a shallow circummarginal groove; first finger
usually shorter than but occasionally nearly equal to second, which
is equal to or slightly longer than fourth when adpressed; disks of
toes larger than those of fingers; toes with minute webs at base;
tibia usually less than 48 per cent of snout-vent length; breadth of
head about 70 to 78 per cent of length of tibia; without broad,
blackish band on lateral surfaces, under surface of head and
throat more or less dusky to brown or blackish.

Description. Snout-vent length of 16 mature females 16.7 to
19.3 mm (a female measuring 16.3 is apparently immature judging
by the fact that the oviducts are straight and undilated), of seven
mature males 15.8 to 17.7 mm (Table 1); hindlimbs moderate,
length of tibia ranging from about 43 to 53 per cent of snout-vent
length; breadth of head slightly less to slightly greater than length
of head, and from 70 to 85 per cent of length of tibia (Table 2);
eye moderately large, its diameter slighdy less to shghtly greater
than length of snout, and about 31 to 48 per, cent of breadth of
head; tympanum distinct, about 50 per cent of diameter of eye;
canthus rostrahs broadly rounded; loreal region nearly vertical;
tongue without notch posteriorly; a fold dorsal and posterior to
tympanum; fingers slender, without webs, tips scarcely expanded
(Fig. 2) ; ventral pad rather vaguely dehmited by a circummarginal
groove, breadth of disk of third finger about 14 to 24 per cent of
length of third finger (Table 2); first finger usually shorter than
second which is longer than or about equal to fourth when ad-
pressed; hindhmbs moderately long; subarticular tubercles large,
elliptical, scarcely protruding; inner and outer metacarpal tubercles
relatively large and flattened; length of tibia about 43 to 52 per
cent of snout-vent length; toes webbed at base; tips of toes moder-
ately dilated; breadth of disk of third toe about 1.38 to 2.33 times
breadth of the disk of third finger; subarticular tubercles large,
broadly oval, and not strongly protruding; inner metatarsal tubercle
large, broad, reaching nearly to subarticular tubercle of first toe;
outer metatarsal tubercle low and round, or scarcely evident.

Color (in preservative). Dorsum variable, ground color light
greenish olive, greyish olive, or brownish; sometimes with variable

1970 NEW BATRACHYLODES 5

dark or blackish blotches and flecks, forming an irregular inter-
orbital patch and a crude, hourglass pattern, more or less complete,
on the dorsum between the forelimbs and the groin; a narrow to
moderate, often much interrupted, blackish band from the snout to
the region of the tympanum or the forelimbs usually evident; an-
terior edge of thighs, posterior edge of metatarsal area, and sole of
foot also usually marked with a dusky to blackish, irregularly mar-
gined band; upper surface of hindlimbs with three more-or-less dis-
tinct, broad, dark, transverse bands, blackish, irregular crescent at
the anal region; a dusky patch on the posterior surface of the distal
part of the thigh and proximal tibia. Under surface of head,
throat, and pectoral region heavily marked with dark flecks; ab-
dominal region lighter.

Measurements of type. Snout-vent length 17.6 mm; length of
head (to posterior edge of tympanum) 6.2 mm; breadth of head
5.9 mm; diameter of eye 2.5 mm; diameter of tympanum 1.2 mm;
length of snout 2.2 mm; length of tibia 8.4 mm; length of third
finger to base of proximal subarticular tubercle 2.1 mm; breadth
of disk of third finger 0.4 mm; breadth of disk of third toe 0.8 mm.

Distribution. Known from collecting sites on both southern and
northern BougainviUe Island at altitudes from about 1000 to 3000
feet.

Juveniles. MCZ 41387, a juvenile measuring 8 mm in snout-
vent length, is not included in the paratypic series, but is tentatively
assigned to this species. In general color pattern, it is less distinct
from B. trossulus than are the adults with which it was taken.

Comparisons. B. minutus, in size and relative lack of dilation of
the finger disks, is most similar to B. trossulus. It difi'ers in its
smaller size at maturity (Table 1); the slightly broader head rela-
tive to the snout-vent length, the greater length of the tibia relative
to the breadth of the head, the shorter third finger relative to the
breadth of the head (Table 2); and noted features of the color
pattern, especially the greatly reduced and less conspicuous dark
brown to blackish bands on the sides of the head and body.

Batrachylodes mediodiscus sp. nov.

Holotype. MCZ 41589, an adult male, collected by Fred Parker
in the Pipekei area, at 2000 feet elevation, on Bougainville Island,
Solomon Islands, 29 December 1962.

Paratypes. (all collected by Fred Parker on Bougainville Is-
land) MCZ 44056 (5 August 1963), 41682-83 (13 AprU 1963),
and SU 21812 (1 May 1963), Aresi area, between 600 and 1500

6 BREVIORA No. 346

feet; MCZ 38066, 41650, and BMNH 1964-690 (22 January
1963), 38255 (26 February 1963), MCZ 41600-01 (31 Decem-
ber 1962), MCZ 41692-93 (16 April 1963), 41694-95 (1 May
1963), 42295, MCZ 42438 (28 June 1963), MCZ 44148 (28
July 1963), 44162-63 (7 August 1963), 44174-75 (15 Septem-
ber 1963), Kunua area, foothills, between 100 and 1000 feet ele-
vation; MCZ 74101-04, Lake Loloru, between 4000 and 4300 feet
elevation, collected 31 March 1963; MCZ 53111, 53124, 53127-
29, 53131-34, 53136-38, 53141-45, 53151-53, 53156-57, 53159-
66, 53168-74, 53178-94, 53196, 53198-205, 53208-27, 53229-
45, 53247-59, 54777-82, 55016-79, 55523-25, 55541-42, 55547,
71082-453, 73272-77, 74105, 74470-722, 74732-823, 76679-80,
79653, Mutahi, between 1800 and 3900 feet elevation, collected
between 9 and 20 May 1966; MCZ 55005-6, Pamauita, at 1200
feet elevation, collected 23 March 1966; MCZ 41584-88, 41593,
41599, SU 21814, and SMF 60016, Pipekei, between 2000 and
3200 feet elevation, coUected between 29 and 31 December 1962;
SU 21815, Pipekei, foothills, 500 feet elevation, collected Decem-
ber 1962; MCZ 55552-53, Topanas, collected 8 May 1966.

Diagnosis. A Batrachylodes of moderate size, snout-vent length
of mature specimens ranging from about 21 to 25 mm in our pres-
ent sample (Table 1); disks of fingers and toes usually about
equally dilated; breadth of disk of third toe about 90 to 125 per
cent of breadth of disk of third finger; breadth of disk of third
finger 23 to 34 per cent of length of third finger, and about 35 to
60 per cent of diameter of tympanum.

Description. Snout-vent length of 20 mature females 21.7 to
27.0 mm (a female measuring 20 mm in snout-vent length is im-
mature), of 22 mature males 21.6 to 25.8 mm (Table 1); hind-
limbs relatively long, length of tibia ranging from 48 to 57 per cent
of snout-vent length for 20 specimens; breadth of head slightly
less than to slightly greater than its length and about 66 to 79 per
cent of length of tibia; length of snout greater than diameter of eye;
eye moderate, its diameter 27 to 39 per cent of breadth of head;
tympanum distinct, its diameter about 60 to 70 per cent of di-
ameter of eye; canthus rostralis rounded; loreal region only mod-
erately oblique, not markedly concave; tongue with a shallow notch
posteriorly or entire; a low to moderate fold dorsal and posterior to
tympanum; fingers moderately slender, without webs, tips ex-
panded into moderate disks; breadth of disk of third finger about
23 to 35 per cent of length of third finger, usually slightly less than
breadth of disk of third toe (Table 2), and about 35 to 60 per cent

1970 NEW BATRACHYLODES 7

of diameter of tympanum; ventral pad of finger disks separated
from dorsal part by a circummarginal groove; first finger shorter
than second which is somewhat shorter than to about equal to the
fourth when adpressed; subarticular tubercles moderately large but
not strongly protruding; outer metacarpal tubercle round; inner
broadly oval; median tubercle large (Fig. 2); tips of toes expanded
into moderate disks, usually slightly larger than those of fingers,
breadth of disk of third toe 90 to 125 per cent of breadth of disk
of third finger for 20 specimens (Table 2); ventral part of disk
separated from dorsal part by a circummarginal groove; subarticu-
lar tubercles moderate, not strongly protruding; inner metatarsal
tubercle narrow, elongate; outer small, round.

Skin of dorsum usually with a few short, narrow, raised folds
which are associated with the dark pigmented spots; venter rela-
tively smooth except for the small white tubercles of chin and
throat of males; posterior thighs granular.

Color (in preservative) . Dorsal ground color variable, ranging
from greyish tan, through greyish olive to dark brown; a small
blackish spot on either side of the middorsal line in the axillar re-
gion, often more elaborate, suggesting a "V" or spread "H," least
obvious in the more melanistic individuals; usually a dark blotch
or spots on the dorsum at the anterior end of the pelvic girdle; a
dusky to black line from the snout posterior to the tympanic region
or, at times, continued on the flanks; anterior and posterior thighs
and anal region with irregular dusky to blackish markings; two
dark transverse bands more or less evident on the thighs; venter
uniformly fight, or variously mottled with brown, especially on
under surface of head and throat.

Measurements of type. Snout- vent length 23.6 mm; length of
head to posterior edge of tympanum 9.0 mm; breadth of head 9.1
mm; diameter of eye 3.0 mm; diameter of tympanum 2.1 mm;
length of snout 3.5 mm; length of hindfimb 37.0 mm; length of
tibia 11.7 mm; length of third finger 3.0 mm; diameter of disk of
third finger 0.9 mm; diameter of disk of third toe 1.0 mm.

Distribution. In addition to the large type-series from both
northern and southern sites of Bougainville Island, eleven speci-
mens MCZ 41238-43, 41671-72, 55007; and SU 21811, 21813
from Buka Island are referred to this species. The known altitu-
dinal range extends from the coastal plains to about 4000 feet.

Comparisons. B. mediodiscus is larger in size at maturity than
B. mimitus or B. trossulus, and the disks of the fingers (except the
first) are more broadly dilated, being about equal to those of the
toes.

8 BREVIORA No. 346

Large-disked species

The large-disked species are much less easily distinguished on
the basis of conspicuous key-characters than are the small-disked
species. In some instances, it is difficult to assign individual speci-
mens, especially juveniles, to a given species, as is in some instances
true for other frogs, for example, some species of Platymantis.
However, such species are frequently distinguishable in the field
on the basis of habits, behavior, and/or voice characteristics. Ade-
quate samples may then usually be separated on the basis of a
combination of morphological characters. The large-disked species
of Batrachylodes are such a group. These species include B. verte-
bralis (the type species), B. wolfi, and three previously unde-
scribed. These are considered in the followmg pages.

Batrachylodes wolfi (Sternfeld)

Sphenophryne wolfi Sternfeld, 1921, Abhand. Senckenberg. Naturf. Ges.,
36, p. 435: Buka I., Solomon Islands (type in Senckenberg Museum).
Batrachylodes vertebralis (part), Mertens, 1929, Zool. Anz., 80, p. 266.

This is one of the large species of this genus, as measured by
snout-vent length at maturity (Table 1). Although individual
specimens of this species and B. vertebralis, B. elegans, or B. mon-
tanus might be confused, representative samples are readily distin-
guished on the basis of a combination of several characteristics, as
indicated in the comparisons. Batrachylodes wolfi is, therefore,
resurrected from the synonomy of Batrachylodes vertebralis.

Definition. A large Batrachylodes, snout-vent length of 20 ma-
ture males 25.2 to 30.6 mm, of 24 mature females 28.2 to 31.7 mm
(Table 1) (a female measuring 26.5 mm snout-vent is immature,
judging by the straight, undilated oviducts); tips of fingers broadly
dilated (Fig. 3), much broader than tips of toes, breadth of disk of
third finger about 38 to 55 per cent of length of third finger,
breadth of disk of first finger 50 to 67 per cent of breadth of disk
of third finger; first finger much shorter than second which is about
equal to fourth when adpressed (77 to 88 per cent as measured);
length of third finger about 35 to 45 per cent of length of head;
disks of toes less broadly dilated than those of finger; toes with very
minute webs at base; length of tibia usually less than 45 per cent of
snout-vent length; head relatively broad, breadth greater than 80
per cent of length of tibia; interorbital breadth about 116 to 150
per cent of length of third finger, and 73 to 92 per cent of length
of fourth finger (Table 3). More than 100 specimens were exam-
ined, including MCZ 54135-325.

1970 NEW BATRACHYLODES 9

Color. Dorsum greenish (olive green in life), occasionally dark
brown for melanistic specimens, generally uniform or nearly so; a
moderate, blackish band (in life), its breadth about one-third to
one-half the depth of the loreal, extending from the tip of the
snout through the eye and the tympanum and remaining more or
less uniformly broad until reaching the post-axillary region where
it may disappear or fade gradually toward the groin; dark trans-
verse bars on hindlimbs usually absent; venter creamy white in
typical color pattern, brownish in melanistic specimens.

Distribution. Presently recorded from both northern and south-
ern sites on Bougainville (elevation from near sea level to about
4000 feet), and from Buka Island.

Comparisions. Based on the Bougainville samples, B. wolfi is
slightly larger at maturity, both mean and maximum snout-vent
length, for both males and females, than B. vertebralis, and smaller
than B. elegans, B. montanus, and B. gigas (Table 1). B. wolfi
exhibits less variation in color for the large sample (500+ speci-
mens), and more uniform dorsal coloration of individual speci-
mens than other species of this large-disked group, with the excep-
tion of some examples of B. elegans. The dorsal ground color of
B. wolfi is also, except for some melanistic examples, more yellow-
ish or olive green than is characteristic of other species.

The disk of the first finger appears more broadly dilated on the
average than the first finger disks of other species, with the excep-
tion of B. gigas. This is also supported in the ratio of the breadth
of the first finger disk to the breadth of the third finger disk
(Table 3). The mean ratio of the interorbital distance to the length
of the fingers is greater than the mean ratio for the other species,
with the exception of the population referred to B. vertebralis from
northern Bougainville.

For B. wolfi, the undersurface of the legs is uniformly light, not
flecked or blotched with brown, as in B. elegans, and the iris is
reddish in life, in contrast to the silver-grey iris of B. elegans.
B. wolfi may be further distinguished from B. gigas by the breadth
of the finger disks relative to length of fingers, for example, the
breadth of the third finger disk to the length of the third finger
(Table 3); and from B. montanus by the much shorter third
finger relative to the length of the head (Table 3).

Batrachylodes vertebralis Boulenger

Batrachylodes vertebralis Boulenger, 1887, Proc. Zool. Soc. London, 1887,
p. 337, pi. xxviii, fig. 3; Fauro I., Solomon I. (type in British Museum).

10 BREVIORA No. 346

Chaperina frederici Stemfeld, 1921, Abhand. Senckenberg. Naturf. Ges.,
36, p. 435; Buka I., Solomon Islands (type in Senckenberg Museum).

Fauro Island, the locality for the unique specimen upon which
B. vertebralis was based, is located at the southern end of Bougain-
ville Island, from which it is separated by a distance of 9 km or
slightly less. A small island lying in the passage divides it into two
channels and further reduces the actual overwater distance by
nearly half. A sample from the population of this species occurring
on southern Bougainville Island is in reasonable agreement with
the type, including size at maturity. On the contrary, no individual
of a sample of more than 100 specimens from the northeastern
part of Bougainville is as large as the type. Boulenger (1887,
p. 337) gives the snout-vent length of the holotype, a mature
female, as 30 mm; our measurement, using a vernier caliper (see
methods), is 28.1 mm. The largest female in our present sample
from southern Bougainville is 29.5 mm.

The following definition of the species is based upon a sample of
specimens from the southern end of Bougainville, including MCZ
73800-73871.

Definition. A moderate-sized to large Batrachylodes, snout-vent
length ranges from 25.4 to 29.5 mm for 5 mature females, and
21.4 to 28.6 mm for 18 males (Table 1) (several females between
23.0 and 24.0 mm in snout- vent length are not yet mature, judging
by the straight, undilated oviducts); breadth of head 75 to 97 per
cent of length of tibia; interorbital breadth 57 to 61 per cent of
length of fourth finger, 47 to 54 per cent of length of third finger;
tips of fingers, except for the first, broadly dilated (Fig. 3, right),
much broader than tips of toes; breadth of disk of third finger
about 33 to 43 per cent of length of third finger and 73 to 94 per
cent of length of first finger; disk of first finger scarcely dilated,
about 40 to 53 per cent of breadth of disk of third finger; first
finger much shorter than second, its length being 70 to 80 per cent
the length of the latter (as measured); the second slightly shorter
than or equal to fourth when adpressed; disks of toes much less
broadly dilated than those of fingers; breadth of disk of third toe
71 to 87 per cent of breadth of disk of second finger; toes with
minute webs at base; length of tibia 43 to 52 per cent of snout-
vent length; head breadth usually less than 80 per cent of length of
tibia (Table 3).

Color. Dorsum grayish cream to tan to brownish, darker mark-
ings in various patterns, occasionally with a light vertebral strip;
a dark line or narrow band with irregular dorsal margins or a series

1970 NEW BATRACHYLODES 11

of blotches from the tip of the snout through the eye and the dorsal
part of the tympanum, with a deep indentation anterior to the fore-
limbs, dark transverse bands of the hindiimbs; usually brownish
mottled, at least anteriorly.

Variation. The sample of the population from northern Bou-
gainville which is referred to B. vertebralis differs from the sample
of population of southern Bougainville in a number of character-
istics: the generally smaller size at maturity for both males and
females (Table 1); the greater frequency of lighter shades of color
(grayish) in the dorsal ground color; and, the somewhat broader
interorbital relative to length of the fingers, length of head, and
length of snout (Table 4). The call, as observed in the field, ap-
pears identical. The noted differences are sufficiently great that
these two populations may well prove to be distinct subspecies.
However, since we do not have information regarding the charac-
teristics of samples from the more central areas of Bougainville,
we have chosen, for the present, to hold in abeyance a decision as
to the taxonomic importance of the observed differences.

Distribution. B. vertebralis is recorded from numerous islands
throughout the archipelago (Brown, 1952), but in view of the
number of morphologically similar species which are now known,
the collections upon which the records are based should be rede-
termined. The population in southern Bougainville ranges from
sea level to about 4000 feet; the population on northern Bougain-
ville ranges from sea level to about 3000 or 3500 feet.

Comparisons. Differences in size relative to the other large-
disked species and several characteristics which distinguish B. ver-
tebralis from B. wolfi have been noted in the previous section.
B. vertebralis may be further distinguished: from B. gigas on the
basis of the smaller third toe disk relative to breadth of the disk of
second or third fingers; from B. montanus by the greater breadth
of interorbital relative to the breadth of the disk of third finger and
the larger tympanum relative to the third finger disk; and, from
B. elegans by the usually smaller tympanum relative to the third
finger disk (Tables 3 and 4).

Batrachylodes gigas sp. nov.

Holotype. MCZ 73764, an adult female, collected by Fred
Parker in the Lake Loloru area, at 4300 feet elevation, on Bou-
gainville Island, Solomon Islands, 31 March 1966.

12 BREVIORA No. 346

Paratypes. (all collected by Fred Parker on Bougainville Is-
land) MCZ 62018-19, 73765, and CAS 117442 from Lake Lo-
loru, 4300 feet elevation, collected 31 March 1966.

Diagnosis. A large Batrachylodes, snout-vent length of 42.5 to
46 mm for three mature females (Table 1 ) ; breadth of disk of first
finger usually more than 50 per cent of breadth of disk of third
finger; breadth of disk of third finger less than 40 per cent of length
of third finger, 65 to 75 per cent of the length of first finger, and
55 to 75 per cent of interorbital breadth; breadth of disk of third
toe about 90 to 95 per cent of the breadth of disk of second finger
and 65 to 85 of breadth of disk of third finger; diameter of tym-
panum about 80 to 110 per cent of breadth of disk of third finger;
interorbital distance relatively narrow, about 95 to 100 per cent of
length of first finger, 30 to 40 per cent of length of third finger and
20 to 22 per cent of length of head; first finger shorter than second,
which is much shorter than fourth when adpressed; and, an irregu-
larly margined dark or blackish band or blotches on anterior thigh,
inner surface of tibia, and lower surface of tarsus.

Description. Snout-vent length of three mature females 42.7 to
46 mm (two females 34 mm in snout- vent length are immature, as
judged by the straight, undilated oviducts); hindlimbs moderately
long, length of tibia 44 to 52 per cent of snout- vent length for five
specimens; breadth of head slightly less than to shghtly greater
than its length, and about 80 to 97 per cent of length of tibia; eye
moderate, diameter of the eye about 80 to 100 per cent of length
of snout; tympanum distinct, its diameter about 46 to 55 per cent
of diameter of eye; interorbital distance about 20 to 22 per cent of
head length, less than 55 per cent of length of third finger, and
usually slightly less than length of first finger (Table 3); canthus
rostrahs rounded; loreal region moderately oblique, only slightly
concave; tongue entire, truncate posteriorly or with a minute notch
posteriorly; a fold dorsal and posterior to the tympanum; fingers
without webs, the tips broadly dilated except for that of first;
breadth of disk of first finger about 50 to 62 per cent of breadth of
disk of third finger; breadth of disk of third finger about 31 to 39
per cent of length of third finger; and about 80 to 110 per cent of
diameter of tympanum; ventral pad of disks separated from dorsal
part by a circummarginal groove; first finger shorter than second,
which is shorter than fourth when adpressed; subarticular tubercles
moderately large but not strongly protruding; inner metacarpal
tubercle small, low, and rounded; middle one broadly oval; outer
one elongate (Fig. 1); tips of toes expanded into moderate disks,

1970 NEW BATRACHYLODES 13

smaller than those of fingers, breadth of disk of third toe about 65
to 85 per cent of breadth of disk of third finger, and 90 to 95 per
cent of breadth of disk of second finger (Table 3); ventral part of
disk separated from dorsal part by a circummarginal groove; sub-
articular tubercles moderately large, rather low; inner metatarsal
tubercle elongate, oval; outer small, round; toes with a very shallow
web at the base, except between the first and second.

Skin of dorsum and venter smooth; posterior surface of thighs
with low, flattish granules.

Color (in preservative). Dorsum variable, from light grayish
oUve or light brownish ohve to salmon red; a dark interorbital bar
and a dark "W" just posterior to the head are usually distinct;
lateral surface of head through eye and tympanum and upper lat-
eral surface of anterior part of body marked by a blackish band
which tapers and fades posteriorly; lower lateral surfaces and
anterior part of venter or entire venter densely marked by dark
flecks; posterior surface of lower arm and both anterior and pos-
terior surface of thighs with irregularly margined dark longitudinal
bands; moderate, dark, transverse bands on hindlimbs may be
strongly or only faintly marked.

Measurements of type. Snout- vent length 46.0 mm; length of
head to posterior edge of tympanum 16.6 mm; breadth of head
17.6 mm; diameter of eye 5.5 mm; diameter of tympanum 2.6 mm;
length of snout 6.7 mm; interorbital distance 3.75 mm; length of
hindlimb 69.4 mm; length of tibia 24.0 mm; length of third finger
7.1 mm; diameter of disk of third finger 2.8 mm; diameter of disk
of third toe 1.9 mm; length of first finger 3.75 mm; first finger disk
1.55 mm; length of second finger 4.55 mm; second finger disk 2.2
mm; length of fourth finger 6.3 mm; fourth finger disk 2.55 mm.

Distribution. Known only from the type locality in the moun-
tains near Lake Loloru area at the southern end of Bougainville
Island at altitudes of about 4000 to 5000 feet.

Comparisons. This is the largest known member of the genus.
The smallest of the three mature females (preserved state) mea-
sures 42.7 mm and the largest 46 mm in snout-vent length. Males
of this species are unknown. As compared to the other three
broad-disked species, the disks of the fingers are slightly less di-
lated; the breadth of the disk of the third finger is less than 75 per
cent of the length of the first finger (usually greater than 80 per
cent for the other large-disked species), and less than 40 per cent
of the length of the third finger length (usually greater than 40
per cent for the rest of the species other than vertebralis from the

14 BREVIORA No. 346

southern part of Bougainville). The disk of the third toe is almost
as large as the disk of the second finger- 90 to 95 per cent (usu-
ally less than 90 per cent for the other large-disked species). The
interorbital distance relative to the length of the fingers is usually
less than for other large-disked species with the exception of
B. montanus. The breadth of the disk of the first finger is about
50 to 60 per cent of the breadth of the disk of the third finger. In
this ratio gigas resembles wolfi, whereas the other large-disked
species usually show this ratio as less than 50 per cent.

Batrachylodes elegans sp. nov.

Holotype. MCZ 54559, an adult female, collected by Fred
Parker at Mutahi, between 2200 and 3200 feet elevation, on Bou-
gainville Island, Solomon Islands, 16 May 1966.

Paratypes. (all collected by Fred Parker on Bougainville Is-
land) MCZ 54333, 54337, 54339, 54341, 54345-47, 54353,
54358, 54361, 54363, 54365-66, 54514-58, 54560-65, 54615-
758, 55186-89, 73298-385, 73519-644, 78822, 79075-461,
79654-79, and CAS 117444-65, from Mutahi, between 1800 and
3600 feet elevation, collected between 10 and 20 May 1966; MCZ
54566-67, 54569-72, 54574-75, 54577-80, 54582, 54584-94,
54596-614, 55190-94, 73633-41, from Melilup, between 3000 and
4000 feet elevation, collected between 12 and 20 May 1966; MCZ
55195-6 from Ramazon River at Melilup, 19 May 1966.

Diagnosis. A moderately large Batrachylodes, snout-vent length
of mature males measuring about 25 to 32 mm and of mature
females 26 to 35 mm (Table 1); dorsal ground color relatively
uniform brownish or variously patterned, a dark brownish, dorsally
even-margined lateral band on the side of the head, narrowing on
the flanks; ratio of interorbital breadth to length of first finger about
1.2 to 1.6, to length of third finger about 0.53 to 0.76, and to length
of fourth finger 0.61 to 0.77; disk of first finger scarcely dilated,
its diameter 41 to 52 per cent of diameter of third finger disk;
diameter of third finger disk usually greater than diameter of
tympanum (94 to 130 per cent); diameter of third toe disk 79
to 100 per cent of diameter of second finger disk and 61 to 81
per cent of diameter of third finger disk.

Description. Snout-vent length of 24 females 26.7 to 34.6 mm,
of 30 males 24.8 to 32.0 mm (however, several females between
26 and 28 mm in snout-vent are immature, as judged by the
straight, undilated oviducts); hindlimbs moderately long, the length
of the tibia about 42 to 50 per cent of the snout-vent length;

1970 NEW BATRACHYLODES 15

breadth of head slightly less to slightly greater than length of head,
74 to 87 p>er cent of length of tibia; eye moderate, its diameter less
than length of snout; tympanum distinct, its diameter 49 to 55 per
cent of diameter of eye and 94 to 120 per cent of breadth of disk
of third finger (Table 3); canthus rostralis rounded; loreal region
moderately oblique, slightly concave; tongue with a shallow-
rounded or V-shaped notch posteriorly; a low, moderate fold dorsal
and posterior to tympanum; fingers without webs; tips of fingers,
except for first finger, broadly dilated (the dilation of the disks and
the relative lengths of the finger do not differ greatly from the con-
dition exhibited by B. veitebralis; Fig. 3, right); diameter of
disk of third finger 38 to 48 per cent of length of third finger, ven-
tral pad of disks separated from the dorsal part by a circummargi-
nal groove; first finger shorter than the second which is somewhat
shorter than the fourth when adpressed (71 to 84 per cent as
measured); subarticular tubercles moderate, oval to round, not
strongly protruding; inner metacarpal tubercle large, longer than
broad; outer metacarpal tubercle similar but smaller; median tuber-
cle large, oval; toes webbed at base; tips of toes expanded into
moderate disks; diameter of disk of third toe about 80 to 100 per
cent of diameter of disk of second finger, 61 to 81 per cent of
diameter of disk of third finger (Table 3) ; ventral part of disk sep-
arated from dorsal part by a circummarginal groove; subarticular
tubercles moderate in size, somewhat more protruding than those
of fingers; inner metatarsal tubercle elongate, narrow, oval; outer
one round; solar area smooth.

Skin of dorsum and belly is smooth. The posterior thighs are
marked by a pattern of low, flattish granules.

Color (in preservative). Dorsal ground color grayish to dark
brown, relatively uniform or with varied pattern of fighter or
darker markings; a dark brown band with relatively smooth dorsal
margin extending from the snout through the eye and tympanum
and posteriorly on to the flank. This band usually covers the entire
loreal region. Many specimens exhibit a whitish line or series of
white spots along the upper margin of the dark, lateral band and
variable white spotting of the limbs and dorsum. The undersurface
of the hindlimbs and the anterior venter are lightly or heavily
speckled with brown.

Measurements of type. Snout-vent length 34 mm; length of head
12.2 mm; breadth of head 13.1 mm; diameter of eye 4.2 mm; di-
ameter of tympanum 2.2 mm; length of snout 4.8 mm; length of
hindlimb 51.7 mm; length of tibia 15.8 mm; length of third finger

16 BREVIORA No. 346

5.6 mm; diameter of disk of third finger 2.25 mm; diameter of disk
of third toe 1.5 mm.

Distribution. Known from the mountains of northern Bougain-
ville Island at elevations from 2000 to 4000 feet.

Comparisons. In size at maturity B. elegans is only slightly
larger than B. wolfi, on the average, and females somewhat smaller
than those of B. montanus. Other characteristics which distinguish
B. elegans from B. vertebralis, B. wolfi, and B. gigas have been
discussed in previous sections. B. elegans is further diflferentiated
from B. montanus by the greater breadth of the interorbital rela-
tive to the length of the fingers; the larger tympanum relative to the
breadth of the third finger disk, and the greater breadth of the
third toe disk relative to the breadth of the second or third finger
disks.

Batrachylodes montanus sp. nov.

Holotype. MCZ 55009, an adult female, collected by Fred
Parker at Melilup, between 3500 and 4000 feet elevation, on Bou-
gainville Island, Solomon Islands, 19 May 1966.

Paratypes. (all collected by Fred Parker on Bougainville Is-
land) MCZ 55010 (31 May 1966), 73762-3 (31 March 1966),
from Lake Loloru, between 4000 and 4300 feet elevation; MCZ
55011, 55013-15, from Melilup, between 3500 and 4000 feet ele-
vation, collected 19 May 1966; MCZ 55529-32, and CAS 1 17443,
from Mutahi, between 3000 and 3600 feet elevation, collected 18
May 1966.

Diagnosis. A large Batrachylodes, snout-vent length of mature
males ranging from about 27 to 35 mm, of mature females 39 to
about 40 mm for the limited sample available (Table 1); diameter
of tympanum less than (70 to 91 per cent) breadth of third finger
disk; disks of fingers, except for first finger, much more broadly
dilated than those of toes, breadth of disk of third toe about 48 to
67 per cent of breadth of disk of third finger, and about 60 to 80
per cent of breadth of disk of second finger; breadth of disk of first
finger usually less than 50 per cent of breadth of disk of third
finger; interorbital breadth about 40 to 50 per cent of length of
third finger, 50 to 60 per cent of length of fourth finger, and 90 to
110 per cent of length of first finger, breadth of disk of third finger
about 70 to 100 per cent of interorbital breadth (Table 3); dorsal
ground color gray to brown; brownish band on side of head nar-
row, irregularly margined or broken, usually not extending pos-
terior to the forelimbs.

1970 NEW BATRACHYLODES 17

Description. Snout-vent length of five mature females 38.2 to
39.4 mm, and eight mature males 26.7 to 35 mm (a female mea-
suring 29.4 mm in snout-vent length is immature); hindlimb mod-
erate, length of the tibia from 42 to 47 per cent of the snout-vent
length; breadth of head usually slightly greater than length of head,
and about 81 to 100 per cent of length of tibia; eye moderate, its
diameter 28 to 35 per cent of breadth of head, about equal to or
somewhat less than length of snout; tympanum distinct, its di-
ameter about one-third to one-half diameter of eye and 69 to 91
per cent of diameter of disk of third finger (Table 3); canthus
rostralis broadly rounded; loreal region moderately oblique, not or
but slightly concave; a low fold may or may not be evident dorsal
and posterior to the tympanum; fingers without webs; tips, except
for first finger, broadly dilated (the dilation of the disks and the
relative lengths of the fingers do not differ markedly from the
condition exhibited by B. vertebralis; see Fig. 3, right); breadth of
disk of third finger about 35 to 43 per cent of its length, about 70
to 100 per cent of interorbital breadth and 80 to 97 per cent of
length of first finger (Table 3); ventral pad of finger disks sepa-
rated from dorsal part by a circummarginal groove; first finger
shorter than second, which is shorter than fourth when adpressed
(about 71 to 81 per cent, as measured); subarticular tubercles
moderate, low; inner and middle metacarpal tubercles broadly
oval; outer ones narrow elongate; hindlimbs moderately long; toes
with minute webs at base; tips of toes dilated into moderate disks;
diameter of disk of third toe about 60 to 80 per cent of diameter of
second finger disk and about 48 to 67 per cent of diameter of third
finger disk (Table 3); ventral part of disk separated from dorsal
part by a circummarginal groove; subarticular tubercles moderate,
rounded, not strongly protruding; inner metatarsal tubercle large,
about twice as long as broad; outer small, round.

Skin of dorsum and venter smooth; posterior surface of thighs
with low, rather inconspicuous granules.

Color {in preservative). Dorsal ground color grayish tan to
brown, mottled, or frequently with a long, dark hourglass marking
or dark chevron in region of forelimbs, along with dark blotches
posteriorly; the dark band on the side of the head is narrow, irregu-
larly margined, terminating at the tympanic region, or fading grad-
ually on the flanks; venter marked by scattered, small dark flecks.

Measurements of type. Snout-vent length 38.6 mm; length of
head 13.6 mm; breadth of head 14.6 mm; diameter of eye 4.6 mm;
length of snout 5.2 mm; interorbital breadth 3.2 mm; diameter of

18 BREVIORA No. 346

tympanum 2.1 mm; length of hindlimb 58.3 mm; length of tibia
17.7 mm; length of third finger 7.3 mm; diameter of disk of third
finger 2.6 mm; diameter of disk of third toe 1.6 mm.

Distribution. Known from higher elevations; about 3000 to 4000
feet in the mountains at the northern and southern ends of Bou-
gainville Island.

Comparisons. Females of B. montanus, based on our limited
sample, apparently attain larger size at maturity than any of the
other species with the exception of B. gigas. It differs from
B. wolfi, B. gigas, and B. elegans in some features of color pattern
(see those species). It may be distinguished from B. elegans, B.
vertebralis, and B. wolfi in the greater length of the fingers relative
to the interorbital distance, and the smaller tympanum relative to
the breadth of the third finger disk. This species also differs from
B. gigas in the greater breadth of the third toe disk relative to the
breadth of the second or third finger disks.

Distributional Pattern

The genus Batrachylodes is widely distributed in the Solomons
Archipelago, from Buka Island in the north to Santa Ana Island in
the south, but surprisingly is unknown from New Ireland, the
southernmost island in the Bismarks, about 100 miles from Buka,
and is thus apparently endemic to the Solomorts.

The evidence as to altitudinal distribution of the genus on the
large, northern island of Bougainville, obtained in the present
study, indicates that the two large species, B. gigas and B. mon-
tanus, occur only above 3000 feet, the approximate lower limit of
the submontane rainforest. B. elegans has a somewhat broader
montane range, being recorded from elevations as low as about
2000 feet. B. minutus has been recorded only from elevations be-
tween 1000 and 3000 feet. The species B. wolfi and B. trossulus
have been recorded from sea level to about 4000 feet, as has
B. vertebralis in southern Bougainville Island. The population of
vertebralis from the north end of Bougainville is apparently lim-
ited to elevations below 3000 to 3500 feet. Also, it should be noted
that B. gigas is known only from the mountains at the southern end
of Bougainville, and B. elegans is known only from the mountains
at the north end. This interesting distributional pattern for the
genus on Bougainville suggests that thorough exploration of the
mountain habitats of other of the Solomons islands may well dis-
cover other isolated, montane species.

1970 NEW BATRACHYLODES 19

KEY TO THE SPECIES OF BATRACHYLODES
IN THE SOLOMON ISLANDS

As previously noted, the several large-disked species of this
genus, though readily distinguished in the field on the basis of
voice characteristics and at times living-color features which dis-
appear in preservative, do not lend themselves to ready identifica-
tion of individuals by means of a typical dichotomous key. It is
hoped, however, that the following key, based on a combination of
characters amounting almost to a diagnosis, may serve to distin-
guish even small samples of adult individuals. Differences in char-
acteristics of juveniles are not yet clear.

1. Tips of fingers only slightly to moderately dilated; breadth of
disk of third toe usually greater than to about equal to breadth
of disk of third finger; breadth of disk of third finger usually
less than 30 per cent (occasionally about one-third) the length

of third finger 2

Tips of fingers moderately to broadly dilated; breadth of disk
of third toe usually less than 75 per cent (rarely up to 85 per
cent) of breadth of disk of third finger; breadth of disk of third
finger greater than 30 per cent of length of third finger 4

2. Size at maturity about 21 to 27 mm in snout- vent length (60
specimens); breadth of disk of third toe about 90 to 125 per

cent of breadth of disk of third finger B. mediodiscus

Size at maturity about 15 to 23 mm in snout-vent length (62
specimens); breadth of disk of third toe about 140 to 180
per cent of breadth of disk of third finger 3

3. Size at maturity 15.8 to 19.3 mm in snout-vent length (22
specimens); breadth of head more than 70 per cent of tibia
length (Table 2); lateral surfaces without extensive dark band
from tip of snout to near groin and on outer surface of thigh (if
present, much reduced), usually with broad, dark, transverse

bands on hindlimbs B. mimitus

Size at maturity 18.1 to 22.7 mm in snout-vent length (40
specimens); breadth of head rarely more than 70 (usually
about 65) per cent of tibia length (Table 2); lateral surfaces
with prominent blackish band from tip of snout to near groin
and on outer surface of thigh; dark, transverse bands on hind-
limbs lacking or very faint B. trossulus

4. Mature females attaining more than 40 mm in snout-vent
length; breadth of disk of third finger about 30 to 40 per cent
of length of third finger and usually less than 75 per cent of

20 BREVIORA No. 346

length of first finger; interorbital distance about 95 to 100 per
cent of length of first finger; uneven-margined, blackish band
or large blotches on anterior thigh, inner surface of tibia and

under surface of tarsus B. gigas

Mature females not attaining more than 40 mm in snout-vent
length; breadth of disk of third finger 35 to 55 per cent of
length of third finger and usually more than 75 per cent of
length of first finger; blackish bands absent from anterior thigh,
inner surface of tibia, and under surface of tarsus 5

5. Snout-vent length of mature females between 30 to 40 mm;
interorbital distance less than 60 per cent of length of fourth
finger and 90 to 110 per cent of length of first finger; tym-
panum usually less than 90 per cent of breadth of third finger
disk; breadth of third toe disk 48 to 65 per cent of breadth of

third finger disk B. montamis

Snout-vent length of mature females between 24 and 35 mm;
interorbital distance usually greater than 60 per cent of length
of fourth finger, and usually greater than 110 per cent of
length of first finger; breadth of third toe disk 55 to 85 per
cent of breadth of third finger disk 6

6. Snout-vent length of 40 mature males about 21 to 28 mm, of
30 mature females 24 to 29 mm; dark band on side of head
usually confined to upper loreal region, rather uneven-mar-
gined dorsally, terminating at point anterior to forelimbs or, if
extending on to the flank, usually with a break or deep indenta-
tion just anterior to the forelimb; breadth of head about 70 to
80 per cent of length of tibia (only for population from north
ern Bougainville Island); interorbital breadth less than 60 per
cent of length of third finger, and usually less than 120 per
cent of length of first finger (only for population from northern

Bougainville Island) B. vertebralis

Snout-vent length of 50 mature males 25 to 32 mm, of 44 ma-
ture females 27 to 32 mm; dark band on side of head usually
even-margined dorsally (often covering most of loreal region
for B. elegans), and usually without break or indentation an-
terior to foreUmb; breadth of head about 75 to 95 per cent of
length of tibia; interorbital breadth usually greater than 60 per
cent of length of third finger and usually greater than 130 per
cent of length of first finger 7

7. Dorsal ground color greenish to olive (dark brown in melanis-
tic individuals), relatively uniform; lateral band blackish (at
least in life); interorbital distance 73 to 92 per cent of length

1970 NEW BATRACHYLODES 21

of fourth finger length of first finger 55 to 67 per cent of length
of fourth finger; disk of first finger distinctly dilated and
breadth of its disk usually greater than 50 per cent of breadth

of third finger disk B. wolfi

Dorsal ground color tan or brown, variously marked with
darker or lighter patterns; lateral band dark brown; interorbital
distance about 60 to 77 per cent of length of fourth finger;
length of first finger 48 to 60 per cent of length of fourth
finger; disk of first finger scarcely dilated and breadth of its
disk usually less than 50 per cent of breadth of third finger
disk B. elegans

ACKNOWLEDGEMENTS

We are indebted to Dr. Ernest Williams, Museum of Compara-
tive Zoology, for suggesting that we review this genus. We are also
indebted to Dr. Alice Grandison, British Museum of Natural His-
tory, for the opportunity of examining the type of Batrachylodes
vertebralis, and to Dr. Konrad Klemmer, Senckenberg Museum,
for the opportunity of examining the type of Sphenophryne wolfi
{= Batrachylodes wolfi). This study is part of the senior author's
program concerned with the herpetofaunas of the Pacific Islands.
The program has been supported by National Science Foundation
grants No. GB-409 and No. GB-4156; the collection and curating
of the material was partly supported by National Science Founda-
tion grant GB-6944 to Ernest E. Williams. Illustrations were pre-
pared by Mr. Walter Zawojski, Stanford University.

LITERATURE CITED

BOULENGER, G. A.

1887. Second contribution to the herpetology of the Solomon Islands.
Proc. Zool. Soc. London, 1887: 333-338, pi. xxvii.
Brown, Walter C.

1952. Amphibians of the Solomon Islands. Bull. Mus. Comp. Zool.,
107: 1-64, pis. 1-8.
Brown, Walter C, and Myers, George S.

1949. A new frog of the genus Batrachylodes from the Solomon Is-
lands, lour. Washington Acad. Sci., 39(11): 379-380.
Mertens, Robert.

1929. Zur Synonymie der Froschgattung Batrachylodes Boulenger.
Zool. Anz., 80: 266-268.
Sternfeld, R.

1921. Zur Tiergeographic Papuasiens und der pazifischen Inselwelt.
Abhand. Senckenberg. Naturf. Ges., 36: 375-436, pi. xxxi.

(Received 18 November 1969.)

22

BREVIORA

No. 346

TABLE 1

Size at maturity of the species of Batrachylodes

known from the Solomon Islands

(R=:range, M=mean, N=:number)

Batrachylodes
minutus

Snout-vent length at maturity

Male Female

R=15.8-17.7 R=16.7-19.3
M=16.9 M=17.8

N=6 N = 16

Batrachylodes

R = 18.1-20.8

R=18.1-22.7

trossulus

M = 19.4

M=20.7

N=r20

N=20

Batrachylodes

R=21.6-25.8

R=2 1.7-27.0

mediodiscus

M=23.7

M=24.2

N=22

N=20

Batrachylodes

R=25.2-30.6

R=28.2-31.9

wolfi

M=27.7

M=29.9

N=:20

N=r20

Batrachylodes

R=2 1.6-26.7

R=24.1-27.6

vertebralis

M=24.3

M=26.06

(north end of Bougainville)

N=22,

N=20

Batrachylodes

R=2 1.4-28.6

R=25.4-29.5

vertebralis

M=26.08

M=27.36

(south end of Bougainville)

N=18

N=5

Batrachylodes

R=24.8-32.4

R=26.7-34.6

elegans

M=29.1

M=30.5

N=30

N=24

Batrachylodes

R =

R =42.7-46.0

gigas

M =

Mr=44.47

N=

Nr=3

Batrachylodes

R=26.7-35.4

R=38.2-39.4

montanus

M=32.6

M = 38.6

N=8

N=5

1970

NEW BATRACHYLODES

23

t3
03

•-J

W-l

*n

_<

v%

ra

o\

00

r-

r«-

o

o

o

r^

rn

00

Tf

■^ (S

o\

\o

•^

>o

VO Tt

VO

t^

o

^

VO

o

vo r-~

o

d
1 1

d

tN

d

d

r4

d d

fN

o<SZ oi^Z a:^z

CM

LU

_l
CO

<

-« c

i: c

« II

° c

■§ S

^ II

l-H
O

fcjj
c

in
<L)
O

C

lU II

■-I J

(D 0^

_ ^

>s: s:

«5

o

<s

(S

00

>r>

t~-

CO

-"I-

fn

d

o

d

00 00

6 t^

Os >0

— '^

Tt On

On <N

r^ m

VO

m r^

o

<^ '^ O

d d

d d

(N

O O fN

II II

li

II II

II

II II II

«^ S Z

oiSZ

f^ S Z

en
ro

rJ

00 m
m 00

VO

<o

00

rj

CM

t-H

r4

1

o

fN 0\

o r-

Tt- t-;

o

"=^9 0

.— 1 -^

(N

O '-I (N

o^SZ o^SZ a:sz

C t3

.2 M

k-l ^^

§ ^
.S -^

-iC

tc

V,

■^

•*-^

c

0

<c.

^

■^

J-*

i.

^

'^

a

jb

•>->

o

(^

"^ O)

oo

m 00

rf On

'^ '^. \o

o o —

o

CM

d

d

00
00

o

CO
(N

<N

d

d

<N

d

d

c^SZ o^SZ oi^Z

J9

e<9

to ss

^

<U

S a

^

■^ ^

"S M

o »

o 5

/5r ^

■hyl
utll

ji

« c

y '?

o "a

? •=

? o

<3 <4)

^ s

^ -

i 5

a

a

Q

«5

«5

oq

24

BREVIORA

No. 346

«5

o
o

00

o

as —
o o --

II II II

Pi S z

\D

o

d

rj

^H

oo

\o

"*

>n

<N

d

d

—

oi S Z

o
o

\£> oo

d d
II II

CO

LU

_J
GO
<

h-

CD E

° II

'o

° II

d *^

O

Q^

C/3

O

C 1/1

Vh 1/1

■-3 ^

c y3

o

p —
G

5

_^

o

1)

^

"3

"^

.-^

>,

c

>

J3

C

■«^

!!j

3

Zj

a

^

O

o

CQ

■yi

CQ

to

-s: Oo
^ 'Si

a

CQ

l>5

"::• s:

to

CQ

to

as
so
00

d
d

o
00

o

o^S Z

o

as m

CT\ as
d d

a! S Z

Oi S Z

0

p

-^

(N

>.

«S

VO

t-~

00

0

d

d

r~4

0^ S Z

ri

d

d

ON

d

0

CM

f^ s z

00

o O >/-i

c^S Z

o^S Z

r-

^

00

0

0

0

"/-i

»— '

^

r^

0

d

d

(N

oi^ Z

o

VO

00

o o

SO t~~

d d

<N O
00 t^

d d

o

oi S Z

o

oi S Z

0

0

00

— I

d

d

1

00

00
00
00

0

1

— 0

0

00 vD
m vD

0

d

d

d d

n '

d d

(N

(^ s z

r»^

m

00

d

r-

0

VO

ir^

VO

0

d

d

(N

c^ S Z

CL,

^

2

'"O

8

<M

■*-»

0

"O

Si

u

■^-»

r*^

"O

c4

tM

U

0

IM

QQ

•^ is

1" Si
^ c

°^

Si a

^^

u o
(1

pq

•5

O

u O

03

<u O
u

PQ

•o c

.Q ._. .^

to t;
-- o

0)

OQ

2 w

a

1970

NEW BATRACHYLODES

25

«o

<N

ro

00

VO

«s

rn

»-H

t^

«o

»— 1

VO

0\

(<1

o

^-H

d

d

o

Ov

cs

o

4

\o

<N

00

Cs

VO

<N

00

o

(N

oo

V)

o\

r^

00

r»i

Tf

■«1-

f<^

ON

q

m

^

>o

(N

V£)

t^

^4

(N

r<-)

r4

d

d

1— t

d

t-H

1-H

d

d

d

d

■—I

d

d

»::s z

»iS z

«:§ z

f^s z

0^:^ z

■*

ON

VO

r-»

00

rf

ON

«o

>n

^

(N

cn

d

rj

d

^— )

o

ON 00

^

t-^ >o

-i t~-

O >/-)

00 (N

r~-

VO

r- <N

'- TT

en NO

■<^ >/-i

o

q

o

>0 NO

o

00 ON

o

m m

o

d d

(N

l-H

>— 1

CM

d d

fS

d d

<N

d d

r^

p^s z

p^s z

o^S z

f^S z

f^S z

ON

q

VO

^— (

d

i-H

d

^

d

■4 — H

VO r-.

-^ 00
ON ON

tT VO

r-- oo

(^ NO
00 ON

00 m
On <N

d d

«n

d d

>o

d d

«o

d d

<o

d d

<o

f^S z

f^s z

p^s z

p^s z

p^s z

(S

r<l

VO

VO

r-

o

t~

t^

ON

r-

M

CO

o

>n

d

l-H

o

On -h

1

VO ^

Ch

00 ■'t

(N Tj-

VO

o

rt 00

<^

>o

O m

U-i VO

o

l-H

Tf

o

NO VO

o

ON

q

o

m fo

o

d d

<N

T-H

'-H

CN

d d

(S

d

i-H

«s

d d

<s

"^ s z

p^s z

p<s z

o^S Z

p^s z

<o

w-i

«r>

(S

l-H

O

00

r«.

O

On

ts

ro

d

"O

d

»-H

d

as '^

1

<S CO

VO •'^

VO ON

r- in

VO

■*

f<^ O

o o

O ro

V-l NO

o

l-H

rn

O

r^ oo

o

r^ ON

o

r«i m

o

d d

(N

l-H

'^

tS

d d

(N

d d

es

d d

(S

f^S z

f^s z

p<S z

f^S z

p^s z

«3 ri

•a
x>

u

t4_i OX)

o o

C en

N-l

J3

Ih

O

Ih

(U

o

00

c

J2

tC

to

•4—1
C/3

a

'T-l

u

-si

jai

iH

on

H)

Li

<4-l

g

^

60

C

tM

u

o

3

tJ-H

tC

o

VkH

<4-l

o

•4— »

CO

C

-4— >

a
a
a

S

o

C3

•a

Ih
<4H

Ih

u

s

4)

>>

o

x:

-a

c

^

C3

(U

O

ca

(U

<u

.C

4>

tl

Ih

H-l

i5

03

(5

CQ

26

0

00

o

O 00

- 5

«5

^ 2 F> ^

O "

^-^^

en

oo

Ik.

00

3
O
CQ

O *r\

d d

o
r4

BREVIORA

00

vo

vo

r4

ON

«n

o

d

\C

C5

1-^

vo

00

r-

o

VD

r--

00

rj

■^

Tt

(N

d

d

d

d

^—

|i

II

T

II

II

II

OL

S

z

ai

S

z

■^

tn

Tl-

ro

Ov

«o

d

O

r^

"O

4

o

m

<N1

fs

r~

r-~

00

~

rf

'^

o

d

d

rsi

d

d

tN

No.

346

d

0.356-
0.398

13

f^ s z

oi S Z

o^S Z

0^ S Z

-*
o

6 c

^ 00

fn m o

d d <^

II II II

<^S z

■o

c
o
u

CO
LU

_J

GQ
<

I-

^

SO

r-

<o

■*

-^

SO
1^

■*
r-

On

d

d

d

d

~ s

rn

t-~

00

00

6

00

vo

r~

00

^

</-i

o

vo

^^

Tj-

vo

t~-

vo

vo

>/1

v-i

o

m

rn

c

d

d

>n

d

d

V-i

d

d

r4

d

d

>o

II

II

II

II

II

II

II

1

II

II

1

II

Oi

2

z

«

S

z

oi

S

z

Pi

S

z

65

«N

vo

vo

S

■^

>o

cs

r»«

■^

00

o

*r\

■*

d

-^

9

d
1

00

<N

6

<N

r~-

00

^-.H

00

.— »

00

</-!

<N

•-^

vo

■*

^l

t^

t^

o

t~

Os

o

■*

•t

o

m

TJ-

o

i: ^

d

d

<s

d

d

(S

d

d

(N

d

d

fN

II

II

II

1

1

II

II

II

II

II

1

II

(^

S

z

oi

s

Z

05

S

z

Ctf

S

z

Co

«s

l-H

vo

2

00

■^

^N

""a-

■X:

00

q

vo

"o

d

T^

1

o

00

d

o

d

00

V-,

■f^ ^

r-

r^

m

(^

o

vo

vo

1—

53 *

r~

00

o

r-

00

o

</-!

>/-l

o

ro

^

O

:*

d

d

<N

d

d

<s

d

d

<s

d

d

fN

II

II

II

II

II

II

II

II

II

II

II

II

Pi

s

z

OiJ

S

Z

C^

S

z

aJ

S

z

u.

<u

M

^4—

o

C

<— •

•o

:/;

c

/^

ri

o

•—

^

^

(n

4>

•3

00

c

<4-l

(C

o

•o

.c

L.

•t-t

m

TJ

03

>4H

<U

o

I-.

oa

% OX)

O c
-S "^

00 1:2
c —
u

^

C/3

u

U

•3

00

U-<

c

o

5C

J3

"O

(S

<4-l

u

O

u

PQ

o

(U O

u

QQ

CO

•o

CO

a

o

tC

j:

m

c3

«<-i

<U

O

u

OQ

<*H 00

o c
x:

C m

u

.970

NEW BATRACHYLODES

27

r-

ON

VC

<o

f^

00

r—

r-

fi

TT

Tf

ON

o

d

d

d

<^ <N

4

ON

fN

•o

00

<N

fN 0\

fN

^

\C

00

00

VD

■* '^

<N

■*

•^

<N

en

ro

<N

r~

00

r4

d d
II II

ll

d
II

d

II

T\

d

II

d
II

T

d
II

d
1

1 1

«S

z

Qi

2

z

Qi

S

z

ai

s

z

«s

00

t^

■<t

r-«

00

a\

o\

■t

^

m

00

d

d

d

d

O v-i

o

</-!

vi

CT\

>H

^M

0\ ro

o

VO

>o

VO

^

CO

en T^r

o

^

'^

o

rn

m

o

r-

00

o

d d

<N

d

d

(N

d

d

n

d

d

(N

^^ ■Z, Pi ^Z Pi^ Z.

f^S z

<N

■/-i

^

0\
00
00

d

d

d

d

4 —

4

oo

o

4 o\
\c 00
en m

m o
r~ 00

c? d

Wl

d

d

fN

d <di »-i

d d

V-l

p^SZ »::sz f^SZ

f^S z

>o

00

■*

00

(n

o

o

vo

■*

>o

Tt

00

o

d

d

O

o\

m

6

NO

4

</-!

o

vo

r^

o

(N

NO

en

r-

rf

o

m

■*

o

-*

^

o

m

m

o

r-

00

o

d

d

<N

d

d

fS

d

d

fs

d

d

n

c^l^Z ft^SZ oiSZ

=^S Z

\o

00

ON

t^

••a-

vo

as

rj-

■t

■^

m

ON

d

d

d

O

"o

t~~

"A

V£>

00

00

>A

m

</n

ON

t-~

fN

<N

t^

fN

t--

m

m

o

m

Tf

o

m

m

o

00

00

o

d

d

(S

d

d

CN

d

d

tN

d

d

CN

<«:^z p< ^z «2z

o^s z

u.

(U

VM

IM

o

C

c

^

M

•a

C

m

u

^

c
u

•a

.2

•2
a

> to

— c

O hJ

c

<4H

o

J3

•a

•o

CS

c^

<L>

U

J3

u

m

c

> 1-1

' Ofl
3 C
O ii

n

IZl

<4-l
O

^

T3

■S

c3

<a

(U

(U

x;

u

BQ

j2 .2

28

BREVIORA

No. 346

LU

_J
CO

<
I-

O oj
O t3

o

lvuq.io.idiui S d °

UlSudi jiious

jvjiqjojsjuj

T3

C3

jaSiii/ pA£

/o ljl8u9J

.laSiiif IS I

invi

/o ijfSiid']

\

ca

«-

0)

1 ./<?gj/// ;/;/7

•sS

E

1 /o ifiSu^-j

So

C

' AdSlllj IS[

iz

/O l{l8U3J

1

cd (u

d

1

C/3

-2 2

f .laSiiij. nip

E

\ fO IjlSUBJ

'^ ^

II

nces
sout

Hipvauq
i jDjiqjoJdiui

^ TS

aJ

1

0^ S

b,

p

fce =«

i JdSlllj pj£

v-

; /o if}8u9j

pi

i nipvauq

jDiiqJOJSWI

so

oo

vo r-; o

O O <N

■^ 00

Tj- ON

^ ^ O

O O fN

o

m O

/O Udl3lUVlQ

ijfpoauq
IDiiquoJOiui

o

O O (N

as

r<-) OS

^ f^ O
O O (N

CO

•^ ^ o

O O (S

00

r-) oo

oo 00

d

d

o

ON
Ti-
ts

o

O

d

d

o

VO

o

V-1

d

O

d

O

00

o

Tf

o

Tt-

d

d

O

o

00

Ti-

Tl- >o Q

o o «s

ON
00
SO

d

"^ ^ o

O O fN

o

ON
00

Tl-

d

00

d

o

«N

ON
SO

o

00
NO

d

NO

d

O

^^'^

^%'^

-Q :- o -J

O 53 I" C
'^ -Q J2 03

y i: O 3
5 ^ c o

«5

5 fe <;;

o

»

•a

c

'>

<u

c

JJ

rt

*-•
3
O

(/5

3
O

s_^

PQ

1970

NEW BATRACHYLODES

29

Fig. 1. (to contrast large disked and small disked species) Left: Batrachy-
lodes trossuhis, inferior view of hand. Right: B. gigas, inferior view of hand.

30

BREVIORA

No. 346

Fig. 2. Left: Batrochylodes minutiis, inferior view of hand. Right: B.
mediodiscus, inferior view of hand.

970

NEW BATRACHYLODES

31

Fig. 3. Left: Batrachylodes wolfi, inferior view of hand. Right: B. verte-
bralis, inferior view of hand.

S-Nfl-^t3.-(o.:i.el DO NOT ci^t^u^;2?!^''- ^°°^

B R E V I O 'R A

HARVARD

Miiseitanti of Comparative ^oofS'J^

Cambridge, Mass. 3 April, 1970 Number 347

DESCRIPTION, OSTEOLOGY AND RELATIONSHIPS OF

THE AMAZONIAN CYPRINODONT FISH FLUVIPHYLAX

PYGMAEUS (MYERS AND CARVALHO)

Tyson R. Roberts^

Abstract. The minute Amazonian cyprinodontid Fluviphylax pygmaeus
(Myers and Carvalho) is redescribed and its osteology worked out based on
material obtained by the Expedigao Permanente da Amazonia (EPA). Its
relationships apparently lie either with the Procatopodinae (a subfamily
otherwise restricted to Africa) or with the Fundulinae, and not with the
Rivulinae. A new subfamily, Fluviphylacinae, is proposed for it. The very
smallest egg-laying and live-bearing cyprinodonts occur in Amazonia,
whereas the largest ones are found in habitats geographically or ecologically
isolated from rich fish faunas.

ACKNOWLEDGEMENTS

I wish to thank Prof. George S. Myers, Division of Systematic
Biology, Stanford University, and Dr. Neal R. Foster, Department
of Limnology, Academy of Natural Sciences of Philadelphia, for
critically reviewing the manuscript of this paper and offering help-
ful suggestions.

INTRODUCTION

This paper deals with an almost unknown but widely distributed
Amazonian cyprinodontid fish of minute size, Fluviphylax pyg-
maeus (Myers and Carvalho). Prof. George S. Myers of Stanford
University kindly supplied the following information about its origi-
nal discovery. In 1942-44, when Myers was working at the Museu
Nacional in Rio de Janeiro in collaboration with Sr. Antenor Leitao
de Carvalho, Naturahsta of the Museu, a man named Alexandre
Parko was collecting for the Museu in Amazonia. Although pri-
marily concerned with insects, Parko obtained a few vertebrates,
including the type of a remarkable toad, Bujo dapsilus Myers and

1 Assistant Curator of Fishes, Museum of Comparative Zoology, Harvard
University, Cambridge, Massachusetts 02138.

2 BREVIORA No. 347

Carvalho (1945a), from Benjamin Constant, and the type of a
strange, leaf-nosed lizard, Anolis phyllorhinus Myers and Carvalho
(1945b), from Borba on the lower Rio Madeira.

The Parko material from Borba also included specimens of a
remarkable new genus and species of minute cyprinodontid with
enormous eyes, obviously closely related neither to Rivulus nor to
any other previously described South American member of the
family. Myers already had been aware that a cyprinodontid of this
type existed in Amazonas; as a student under the late Prof. Carl H.
Eigenmann he had found one partially squashed and otherwise
poorly preserved specimen from the lower Rio Amazonas while
unpacking fishes collected by the late Carl Ternetz in 1923-25.
This specimen was in no condition for description and seems even-
tually to have become lost. Carvalho and Myers planned to de-
scribe the Parko specimens, but Myers had to leave Rio before this
was done, bringing with him to Stanford only three specimens.
Later, when Carvalho was at Stanford on a Guggenheim Fellow-
ship, some more work was done on the fish, and Carvalho prepared
a drawing of it. Unfortunately the drawing and data were lost in
the late 1950's.

While preparing a resume of cyprinodontid subfamilies, Myers
(1955) gave a brief diagnosis of Parko's tiny Amazonian fish,
which now may be reproduced:

One new fish that aquarists will probably see before too long
is the very tiny Potamophylax pygmaeus Myers and Carvalho,
from the Middle Amazon. It is one of the smallest of all killies,
and its place in the above classification is not clear. The strange
characteristics, as worked out by Mr. Carvalho and me, include
enormous eyes, a dorsal fin set entirely behind the anal fin, high-
set pectoral fins and a deep "pocket" beside the latter fin. It
represents a hitherto unknown genus and species and it may be
related to the African lampeyes.
Although no type specimens were designated, the description satis-
fied the rules of zoological nomenclature in effect at the time it was
published, and therefore is nomenclaturally valid. Whitley (1965:
25) pointed out that Potamophylax Myers and Carvalho is pre-
occupied by Potamophylax Wallengren, 1891 (a neuropteran in-
sect), and proposed the substitute name Fluviphylax, with P. pyg-
maeus as type species. The species apparently never has been im-
ported as an aquarium fish and nothing further has been reported
about it.

Since the types are few in number and only in fair condition, it

1970 AMAZONIAN CYPRINODONT FISH 3

was highly desirable to obtain more material. In 1968 I partici-
pated in field work in Amazonas as a member of the Expedic^ao
Permanente da Amazonia. EPA is a continuing, cooperative efTort
among the Museu Goeldi in Belem, the Instituto Nacional de Pes-
quisas da Amazonia (INPA) in Manaus, and the Museu de Zool-
ogia of the Universidade de Sao Paulo (MZUSP)/ financed by the
Fundagao de Amparo a Pesquisa of the state of Sao Paulo. I wish
to thank Dr. P. E. Vanzolini, Director of the Museu de Zoologia,
for arranging my participation. September 20 through November 5
we worked on the Rio Solimoes between Manaus and Santo Antonio
do Iga (the Middle Amazon). The scientific party consisted of
Heraldo A. Britski (leader), Julio Cesar de Garavello, Silvia
Gerken, Naercio A. Menezes and myself. Our primary purpose
was to collect fishes and we were fortunate in obtaining Fliiviphylax
at several localities. The extensive collections from this trip are
being sorted at the Museu de Zoologia. Additional Fliiviphylax
were located in the material obtained by EPA on the Lower Ama-
zon in 1967 (its first year in operation). With the kind permission
of Heraldo Britski, who is in charge of the fish collections, I
brought two lots to the United States for this study. Finally, it is of
considerable interest to note that the Museum of Comparative
Zoology has a few small lots of Fluviphylax obtained by the Thayer
Expedition under Louis Agassiz in 1866. In all probabiUty Gar-
man had the Agassiz specimens in hand when he wrote his mono-
graph, "The Cyprinodonts" (Garman, 1895), but passed them
over as the young of Rivulus.

Since they are liable to be confused, care has been taken to dis-
tinguish Rivulus from Fluviphylax in the following generic charac-
terization. The easiest distinctions to use involve eye size, nature of
orbital rim, number of suborbital scale rows, width of head, condi-
tion of anterior nostril, jaw structure, and relative positions of
pectoral and pelvic fins.

Genus FLUVIPHYLAX Whitley

Potamophylax Myers and Carvalho, in Myers, 1955: 7 (original description:
type species P. pygmaeus Myers and Carvalho, by original designation;
name preoccupied by Potamophylax Wallengren, 1891, a neuropteran
insect) .

1 Formerly Departamento de Zoologia of the Secretaria da Agricultura of
the state of Sao Paulo. The name was changed in June, 1969, when it be-
came part of the Universidade de Sao Paulo.

4 BREVIORA No. 347

Fluviphylax Whitley, 1965: 25 (replacement name for Potamophylax Myers
and Carvalho, and therefore taking same monotype).

Characteristics. Tiny egg-laying cyprinodonts, widespread in
Amazonas, typically found in still backwaters. Largest specimen
available for measurement 18.5 mm; females as small as 14.5 mm
with ripe ovaries.^ Body moderately elongate, sub-cylindrical in
cross-section, sides not compressed. Body axis straight. Head
about as wide as deep (head considerably wider than deep in
Rivulus). Eye extremely large, nearly as deep as head, its dorsal
margin either flush with dorsal contour of head or barely below it,
and largely exposed dorsally (supraorbital portion of frontal bone
only partially covering eye); orbital rim free (eye of moderate size,
covered dorsaUy, and with attached orbital rim in Rivulus). Pre-
orbital space narrow, about .25 diameter of eye. Interspace be-
tween eye and upper angle of preopercle extremely narrow; pre-
opercular segment of cephalic sensory canal system very close to
posterior margin of eye. A single row of scales extending just be-
low eye. Below this suborbital scale row the interopercle can be
lifted sideways to reveal a deep furrow in the skin separating genio-
hyoideus muscles from jaw suspension (there are at least two or
three suborbital scale rows in Rivulus, and the area between genio-
hyoideus muscles and jaw suspension is smooth and scaled over).
Snout short, about .5 eye diameter, its dorsal contour with a slight
concavity anterior to eye. Anterior nasal opening a pore flush with
snout (tubular in Rivulus). All segments of cephahc sensory canal
usually open, but a few specimens have one or more of the seg-
ments closed. Some scales of midlateral series with pit organs.
Mouth small, superior, the small upturned jaws set at an angle of
about 45 degrees (mouth less superior in Rivulus, with a very char-
acteristic, deep, right-angled notch or recess between nasal and
lacrimal bones into which rictal membranes fit). Jaw teeth conical,
in two irregular rows in both jaws, teeth of outer and inner rows of
about the same size. Tongue free from floor of mouth for most of
its length; rounded in cross-section near the tip, posteriorly a some-
what flattened triangle in cross-section (with apex downward and
lying between left and right m. geniohyoidei) (in Rivulus tongue
attached to floor of mouth except for its anterior third, flattened in
cross-section throughout its free portion, and broadly rounded or
blunt at tip). No teeth on palate (prevomer usually bearing teeth
in Rivulus) . Intestine in anterior third of body forming one and a

1 Standard lengths only are recorded in this paper.

1970 AMAZONIAN CYPRINODONT FISH 5

half vertical loops at an angle of about 30 degrees to body axis
(anterior portion of looping to left and posterior portion to right
side of body); straight in remaining two-thirds of body cavity.
Stomachs of numerous specimens contain minute insect larvae.
Gill rakers smooth, slender shafts, widely set, ten on lower and two
or three on upper limb of first gill arch. Pseudobranch an ovoid,
bilobate, white, glandlike mass without filaments about .5 mm high
in a 16-mm specimen, covered by a transparent membrane. Origin
of gill opening about .25 eye diameter in front of pectoral fin base,
on a level with or slightly below origin of uppermost pectoral ray.
Gill membranes free from isthmus, united to each other at a point
on a vertical with posterior margin of eye. Branchiostegal rays
usually five on either side.

Dorsal fin small, with five or six rays, its origin distinctly poste-
rior to a vertical line through base of last anal ray (dorsal fin set
equally far back in some species of Rivulus, but usually with more
rays). Anal fin much larger than dorsal, with eight or nine rays.
Anal fin pterygiophores slender except for the first, which is mod-
erately expanded. Caudal fin rounded. Pectoral fins high-set, up-
permost pectoral ray originating distinctly above lateral midfine of
body, and extending well beyond base of pelvic fin (uppermost
pectoral ray inserted below midline in Rivulus, and pectoral fin
failing to reach pelvic base) , Internal to base of pectoral fin a deep
pocket, the entrance to which is guarded by delicate tissues easily
ruptured (similar pockets present in all cyprinodonts). Pelvic fins
separate, not connected by membrane, with six rays. Pelvic inser-
tion equidistant between origin of lowermost pectoral ray and anal
fin origin.

In mature males dorsal, pelvic, and anal rays slightly to moder-
ately elongated, and dorsal, pelvic, anal and caudal fins with a
barred pattern (strongest on anal and caudal). Shape of caudal fin
unaffected by sexual dimorphism. Examination of scales and fins
with direct and transmitted light failed to reveal contact organs.
Ripe females contain 30-40 eggs, the largest of which are some-
what over .1 mm in diameter, and their abdomens are slightly dis-
tended (best seen viewing specimen from above). Teasing the sur-
face of eggs isolated from the ovary reveals elongate chorionic
filaments which apparently originate from a restricted portion of
the chorion, but no short, spinelike projections (short, spinelike
projections are distributed over the entire chorionic surface in
various atherinoids, Oryzias and Horaichthys) . Ovary single, with-
out indication of division between right and left halves. Males with
cloacal-genital area unswollen, a small, distinctly tubular opening

6 BREVioRA No. 347

to sperm duct slightly in advance of base of first anal ray. Some
females with cloacal-genital area considerably swollen, an oviducal
pouch opening in front of first anal ray (the membranes of this
pouch do not encroach upon first anal ray, nor do they extend
along side of anal fin base, as happens in some Fundulinae).

Fluviphylax pygmaeus (Myers and Carvalho)
Figures 1 and 2

Potcimophylax pygmaeus Myers and Carvalho, in Myers, 1955: 7 (original
description; type locality "Middle Amazon"=Borba, lower Rio Ma-
deira; types in Museu Nacional, Rio de Janeiro; three paratypes at
Stanford).

Fluviphylax pygmaeus Whitley, 1965: 25 (name only).

Material Studied. SU 50196, three paratypes 14.9-16.1 mm, Rio
Madeira at Borba, collected in 1943 by A. Parko; MZUSP 6245,
13 specimens 11.5-18.5 mm, "lago above Manaus," margem es-
querda Rio Negro, Amazonas, Brasil, collected by EPA April 28,
1967; MZUSP 5590, 50 specimens 9.9-17.0 mm, Lago Parauacui,
near Oriximina, Para, Brasil, February 9, 1967, EPA; MCZ 41367,
five specimens 12.9-14.4 mm, Lago Hyanuary, Amazonas, Brasil,
1865-66, Thayer Expedition; MCZ 6262, five specimens 10.3-
10.5 mm, Obidos, Para, Brasil, 1865-66, Thayer Expedition; MCZ
6265, one specimen 12.9 mm Lago Alexo,^ Amazonas, Brasil,
1865-66, Thayer Expedition,

1 In response to my inquiry, Prof. Myers provided the following informa-
tion about this locality:

Many Thayer Expedition fishes came from Lago Alexo. The name may
be misspelt, as are some other Thayer localities. It may have been
"Lagoa Aleixo." However, there is no such place listed, spelt either way,
in the 4 volumes of Pinto's "Diccionario Geographico," including the sev-
eral little addenda at the ends of the volumes. "Journey in Brazil," the
narrative of the Thayer Expedition by Louis and Mrs. Agassiz, has no
index and does not give a complete itinerary of the separate trips made by
expedition members in the Amazon. However, on p. 294, it records the
return of Thayer to Manaus on December 6, 1865. Going backward in
the book, the last previous mention of Thayer is on p. 282, where it says
that he and Bourget were sent (from Manaus) to "Lake Cudajas, to be
gone ten days," this being recorded under the date of November 20. That
is, unless two separate small trips were involved, which is doubtful,
Thayer started for "Lake Cudajas" on November 20 and "returned from
Lago Alexo" 16 days later. "Lago Alexo" may be the same as "Lake
Cudajas," the correct name of which Thayer could have found to be
"Lago Alexo" once he got there. In any event, Lago Alexo cannot be far
from Manaus.

1970

AMAZONIAN CYPRINODONT FISH

>v;

V

\^^

8 BREVIORA No. 347

Morphometric data has not been assembled because superficial
examination of the specimens reveals no proportional variations of
consequence and because accurate measurements of the small
specimens are difficult. The proportions given in the following
description were obtained by measurements on a few specimens or
by stepping off distances with a pair of dividers and are meant to
provide a general impression. Counts of fin rays and scales made
on many of the specimens reveal little variation; pectoral, dorsal
and anal fin counts vary by only a ray or two.

Body proportions. Greatest depth of body 5.0. Least depth of
caudal peduncle 9.0. Head 4.0. Eye 2.5 in head. Bony inter-
orbital space slightly greater than eye diameter. Snout about .5 eye
diameter. Width of mouth about .75 interorbital space. Dorsal fin
about same size as pelvic fin, length of dorsal fin base about equal
to length of base of first five anal rays. Anal fin far larger than
dorsal fin, and slightly larger than pectoral fin. Caudal fin distinctly
larger than anal fin. Length of caudal peduncle about equal to anal
fin base.

Fin rays. Dorsal fin with five or six rays; anal fin usually with
eight rays, sometimes nine; in alizarin specimens first two and last
two rays of dorsal and anal fins simple, other rays branched. Pec-
toral fin with 10 or 11 rays, pelvic fin with six rays, caudal fin usu-
ally with 17 rays (counts on alizarin preparations reveal eight prin-
cipal, five upper and four lower procurrent caudal fin rays). Rays
in all fins without unusual specializations.

Squamation. Scales uniformly cycloid, 25 or 26 in midlateral
series. Anteriorly three, sHghtly posterior to anal origin two, and
on caudal peduncle one, rows of scales between midlateral and
dorsomedian scale rows. About 18 or 19 scales in median pre-
dorsal series from occiput to origin of dorsal fin. Nine scales
around caudal peduncle. Seven scales in diagonal series from in
front of dorsal fin origin to middle of anal fin base. About five
scales along anal fin base. Fins themselves scaleless except for two
rows of normal-sized scales extending beyond hypural fan onto
proximal portion of caudal fin. Scales on either side of genital re-
gion unenlarged and unspecialized.

I tentatively interpret the squamation on the dorsum of the head
in Fluviphylax (Fig. 2) as typically consisting of a single h scale,
a single g scale (which overlaps the e scales in some specimens but
not in others), two e scales, two d scales, and a single a scale. Per-
haps owing to the exceptionally large eyes, the a scale appears to be
further forward than usual. Some specimens appear to have two

1970

AMAZONIAN CYPRINODONT FISH

scales in the "a" position. Some specimens have a small scale an-
terior to the h scale. There are a pair of scales on either side of the
head over the supraorbital portion of the frontal bone. This is a
simphfied pattern similar to that probably found in many unrelated

Fig. 2. Scale pattern on dorsum of head in 15.1 -mm specimen.

forms in which the pattern has been secondarily reduced, including
Epiplatys sexjasciatus, young Nothobranchius (or Aphyosemiorf!)
sjoestedti, and perhaps young Pantanodon podoxys (Whitehead,
1962, figs. 12 and 13).

Coloration. The MZUSP lots (including several not Hsted above)
I have examined vary from those in which the specimens are rela-
tively dark and have markings that stand out to others in which the
specimens are quite pale and the markings are washed out. I am
fairly confident, however, that all represent the same species. The
coloration is sexually dimorphic. In females in all lots there is a
more or less sharp demarcation between the relatively dark sides
and the light belly, while the fins are clear or slightly dusky, but
never have distinct markings. In males, on the other hand, the
body is more uniformly colored, the belly about the same color as
the sides; the sides usually bear a series of hazy, lozenge-shaped,
vertical bars, from two or three restricted to the pectoral region to
a dozen or so for the length of the body; the dorsal, anal and caudal
fins bear concentrations of melanophores (see Fig. 1).

Scant observations were made on coloration in live Fluviphylax
during the EPA fieldwork in which I participated, since all were

10 BREVIORA No. 347

taken during rotenone collecting and were seen but briefly before
they had to be preserved. I distinctly recall that they lack the shim-
mering bluish-greenish reflections on the sides possessed by African
Procatopodinae such as Poropanchax rancureli and Aplocheilich-
thys shiotzi. The most striking feature of their coloration is a sil-
very sheen, presumably guanin, on the entire dorsal surface of the
eyes. As a little Fluviphylax swims along just below the water's
surface, two brilliant silvery dots (surprisingly large considering
the size of the fish) are seen moving along.

Sex ratios. In some MZUSP lots almost all intermediate-sized
specimens bear eggs, the smallest specimens are quite evidently
immature, and only a few of the largest individuals are males. This
suggests they might be protogynous hermaphrodites. In other lots,
however, nearly half of the specimens, including ones as small as
12.5-14 mm, are males.

ECOLOGY

Fluviphylax has been collected along large stretches of the Lower
and Middle Amazon, and presumably occurs in suitable habitats
throughout the Amazon basin. I remember catching a single speci-
men near the mouth of the Rio Jauaperi, well up on the Rio Negro,
but the specimen was lost before it could be preserved. Fluviphylax
lives in stifl, shaded, black- or brown-tinted, shallow backwaters,
not in the main rivers or even in small, flowing streams. Most
fishes in such places are small species. The principal predaceous
kinds are likely to be a pair of "dwarf" species of Acestrorhynchus,
A. nasutus, and A. minimus, and occasionally Hoplias, Hoplery-
thriuus, or Potamorrhaphis. In EPA collections Fluviphylax also is
associated with Crenuchus, Curimatopsis, several species of Hemi-
grammus and Hyphessobrycon, Iguanodectes, Hypopomus, a mi-
nute species of Trichomycterus, two or three species of Apisto-
gramma, Cichlasoma jestivum, two tiny species of Microphilypnus,
Poecilia (Pamphorichthys) minor, and other smafl species. Rivulus
also occurs together with Fluviphylax, but my impression is that
Fluviphylax tends to stay in more open water than Rivulus, which
during the dry season is Hkely to be left in small pools that are
drying up. Rivulus frequently occurs singly in water only inches
deep and probably spends long periods relatively motionless,
whereas Fluviphylax is usually over deeper water and possibly
maintains small schools near the surface. Stomach contents of
numerous specimens contain minute insect larvae or adults (no
other food items observed).

1970 AMAZONIAN CYPRINODONT FISH 11

Fluviphylax is evidently the smallest known oviparous cyprino-
dont. In all sizable samples obtained by EPA (representing the
months of February-April and September-October) small speci-
mens predominate, despite probable sampling errors favoring
larger specimens. This indicates that reproduction occurs through-
out much of the year, perhaps year-round. If so, the bulk of the
population probably is sexually immature at all times, and I would
not be surprised if the average size of an individual in these popu-
lations falls between 9 and 1 1 mm. If we can judge from other
small oviparous cyprinodonts, hatchlings are probably between 4
and 6 mm.

OSTEOLOGY
Figures 3-13 (Reduced to same scale.)

This account of the osteology of Fluviphylax is based on four
specimens (11.5, 14.4, 16.5 and 17.0 mm) from the MZUSP 5590
and two specimens (17.8 and 18.5 mm) from MZUSP 6245. The
figures are based on the larger two of the four specimens from
MZUSP 5590. All statements which follow were based on observa-
tions verified in at least two or three specimens. The 11.5 mm
specimen is well ossified; differences noted between it and larger
specimens are 1) failure of nasal bones to ossify; 2) less develop-
ment of anterolateral parasphenoid wings; and 3) presence of a
sixth branchiostegal ray on left side. No osteological differences
attributable to secondary sexual dimorphism were detected.

Braincase (Figs. 3-5). Braincase typically cyprinodontoid in
appearance (c/. numerous figures of cyprinodont crania in dorsal
view in Rosen and Bailey, 1963). Cranium very slightly tapered
anteriorly. Supraorbital processes of frontal well developed;
frontals with a concavity behind supraorbital process for reception
of dermosphenotic, which lies over sphenotic and snugly against
frontal. Mesethmoid a thin ossification folded back on itself at its
anterior margin to form a double lamina open posteriorly. Meseth-
moid not contacting prefrontals. Prevomer, parietals, and basi-
sphenoid bones absent. Anterolateral parasphenoid wings moder-
ately developed, not contacting pterosphenoids to form posterior
myodomes (in cyprinodonts anterolateral parasphenoid processes
play the role corresponding to that of the basisphenoid in other fish
groups) . Dorsal border of foramen magnum formed exclusively by
exoccipital bones. Supraoccipital bone typically cyprinodontoid
in form, Exoccipitals and epiotics without posteriorly-directed
processes or wings. Supraoccipital with a close-set pair of short

12

BREVIORA

No. 347

parasphenoid

mesethmoid

prefrontal

dermosphenotic

sphenotic

pterotic

supraoccipita

Fig. 3. Dorsal view of cranium, 16.5-mm specimen (all osteological
figures based on specimens from MZUSP 5590).

spines extending straight back near dorsal midline. Basicranium
with a basioccipital and two exoccipital condyles (Fig. 5). Ventral
surfaces of pterotic and sphenotic bones with well-developed sock-
ets for articulation of hyomandibular bone (Fig. 4).

Facial bones (Figs. 6, 7). Nasal bone small and weakly ossified.
Lacrimal bone weakly ossified, its shape seemingly determined
solely by its function as a bony trough for preorbital segment of
cephalic sensory canal system and its relation to anterior margin of
eye. Bony trough in upper arm of preopercle for postorbital seg-
ment of cephahc canal system extremely close to hind border of
eye. Bony troughs for segments of cephalic canal system formed

1970

AMAZONIAN CYPRINODONT FISH

13

mesethmoid

prefrontal

parasphenoid

pterosphenoid
dermosphenotic

sphenotic

sphenotic socket
for hyomandibular

pterotic socket
for hyomandibular

pterotic

basioccipital' ■-■ - — . ^ exoccipital

Fig. 4. Ventral view of cranium, 16.5-mm specimen.

only by lacrimal, preopercle and dermosphenotic bones. The very
large eye is apparently unassociated with any radical modifications
or displacements of bones. Subopercle very large, forming entire
posterior margin of gill cover. Upper arm of subopercle termi-
nating in a weak flange forming dorsal margin of an externally
visible concavity in opercular margin just below dorsal origin of
gill slit. Lower arm of subopercle with a strong, dorsally-directed
flange lying between opercle and interopercle. Interopercle rery
elongate.

Jaws and jaw suspension (Figs. 6, 7). The jaws do not differ
from what one would expect in a small, surface- or near-sur;ace-
feeding, insectivorous cyprinodont. Premaxillaries moderately pro-

14 BREVIORA No. 347

exoccipital supraoccipital

/ epiotic

'' , L— ^^=^^:r---J— _^ / pterosDhenoid

supracleithrum ,„,,„„„ „,

foramen magnum basioccipital , ,. , .

pterotic socket
for hyomandibular

Fig. 5. Occipital view of cranium, 17.0-mm specimen. Pterosphenoid
visible because frontal and dermosphenotic bones have been removed.
Viewed from angle to right of fish.

tractile, ascending premaxillary processes broad-based and short,
with rounded posterior margins. Maxillaries slender, with well-
developed mesial and palatine processes. Mesial processes liga-
mentously attached to ventral surface of ascending premaxillary
processes; back and forth movement of premaxillaries accompanied
by equal movement of mesial maxillary processes, and when trans-
mitted to distal portion of maxillaries (beyond palatine processes)
these rotate on their axes while remaining in more or less the same
vertical plane. An 1 1.5-mm specimen has about 26 teeth in upper
jaw and 22 in lower; a 16.5-mm specimen, about 35-40 teeth in
both jaws.

Hyomandibular bone with well-developed, separate joints for
articulation of pterotic and sphenotic bones. Palatine, ectoptery-
goid, and entopterygoid intimately united (for discussion of these
eletients in cyprinodonts see Rosen, 1964: 232, fig. 10). Sym-
plectic bone elongate. Quadrate with a slender, posteriorly-directed
extension. Metapterygoid absent.

Hyoid and pharyngeal arches (Figs. 8, 9). The hyoid arch has
the following peculiarities: 1) hypohyal of each side a single, un-
divided element; 2) interhyal absent; 3) branchiostegal rays, usu-
ally five on each side, articulating with ceratohyal; 4) urohyal with
a st:ong, dorsally-directed process at its anterior end; and 5) epi-
hyal and ceratohyal not rigidly joined by a dorsal bony ridge.

1970

AMAZONIAN CYPRINODON T MSH

15

mesethmoid

dentary

parasphenoid

lacrimal

prefrontal

premaxillary

maxillary

palatine

nasal

Fig. 6. Dorsal view of jaws and anterior bones of head, 16.5-mm
specimen: head tihed back.

maxillary
premaxillary / ^^^^^^.^^

ectopterygoid

joint with joint with-

sptienotic pterotic

dentary
coronomeckelian

angular \ quadrate ^ subopercle

interopercle

Fig. 7. Internal view of jaws, jaw suspension and opercular elements,

16.5-mm specimen.

b

16

BREVIORA

No. 347

Branchiostegal rays usually simple, but in 17.8-mm specimen an-
teriormost ray on one side with a hooklike, medially-directed proc-
ess midway on its length.

glossohya

ceratohyal ^pihyal

Fig. 8. External view of Jiyoid arch and urohyal (separated), 16.5-mm
specimen.

hypobranchials

glossohyal
hypohyal

basibranchials
ceratohyal

ceratobranchials

second
pharyngobranchial

epibranchials

ihyal

third and fourth
pharyngobranchials

Infrapharyngeal

Fig. 9. Dorsal view of hyoid and pharyngeal arches, with dorsal ele-
ments shown on left side only, 16.5-mm specimen. Third hypobranchial
on right side not ossified in this specimen; ossified in other specimens.

1970

AMAZONIAN CYPRINODONT FISH

17

Pharyngeal arches with two elongate ossified basibranchials. First
pharyngobranchial absent. Second pharyngobranchial intimately
associated with but separate from upper pharyngeal plate, toothless
in some specimens, in others bearing four to six conical teeth. Up-
per pharyngeals, presumably consisting of united third and fourth
pharyngobranchials, with a dense patch of conical teeth. Lower
pharyngeals of left and right sides not in contact, with two or three
irregular rows of conical teeth.

Fig. 10. Frontal view of first vertebra, 16.5-mm specimen.

\^epipleural

rib of second
vertebra

irst rib?

coracoid

Fig. 11. External view of pectoral girdle, 16.5-mm specimen; first and
second ribs drawn separately; in normal position they lie directly behind
the radial bones, or behind radial bones and articulations with pectoral rays.

18 BREVIORA No. 347

Vertebral column. Vertebrae 26 excluding hypural, 12 abdomi-
nal and 14 caudal or 13 abdominal and 13 caudal. First vertebra
(Fig. 10) with complete neural arch, three condylar facets, and
associated pair of epipleural intermuscular bones, without trans-
verse processes or ribs. A small element associated with the distal
end of the rib of the second vertebra is probably the true first rib
(Fig. 11). This element is characteristic of cyprinodontoids; in
Oryzias melastigma it is unusually long. Rosen and Bailey (1963:
31) interpreted what is evidently the same element in Poeciliidae
as a secondary postcleithrum. In Fluviphylax it lies in the body
wall, and is separated from the pectoral girdle by the deep pocket
internal to the pectoral fin base. Remaining abdominal vertebrae
with large, stout transverse processes to which are attached proxi-
mal tip of ribs, and simple epipleurals. In some specimens epi-
pleurals fail to develop in association with posterior abdominal
vertebrae. Caudal vertebrae lacking epipleurals. Haemal canals
moderately large.

Pectoral girdle (Fig. 11). Posttemporal bone forked, upper
limb articulating with epiotic, lower with exoccipital (intercalar
absent). Supracleithrum small, scalehke, interposed between post-
temporal and cleithrum. Scapular foramen present. Cleithrum
rather slender. Lower limb of coracoid slender, its anterior end
extending to near anteroventral tip of cleithrum; space between
cleithrum and coracoid large. Postcleithra absent. Pectoral radials
four. Pectoral splint present, consisting of two very short, bony
splints tightly adherent to dorsal surface of ray halves of uppermost
pectoral ray.

Pelvic girdle (Fig. 12). Distal tip of fifth rib ligamentously at-
tached to dorsal surface of median pelvic projection. Posteriorly
directed bony process of pelvic girdle very short, lacking in some
specimens.

,rv

PBP

Fig. 12. Ventral view of left pelvic girdle and fin rays, 17.0-mm speci-
men. PBP = posteriorly directed bony process.

1970

AMAZONIAN CYPRINODONT FISH

19

Caudal skeleton (Fig. 13). Hypural fan fused into a solid unit,
even in 11.5 — mm specimen, to which six of the eight principal
caudal rays attach. Epural and haemal spines of hypural centrum
each bearing a single principal ray.

Fig. 13. Last three vertebrae and caudal skeleton, 11.5-mm specimen.
Tiny separate element just above hypural centrum not evident in other
specimens.

RELATIONSHIPS

The osteological characteristics of Fluviphylax are such that it
does not fall automatically in any of the groups into which Sethi
(1960) divided the oviparous cyprinodonts of the world (it is ob-
viously unrelated to Pantanodontinae [Whitehead, 1962; Rosen,
1965] which Sethi did not consider), but they indicate that its rela-
tionships either lie with Procatopodinae or else with Fundulinae,
and not with Rivuhnae. In order to facilitate comparisons these
characters of Fluviphylax are listed here (the order is similar to
that employed by Sethi).

1. Mesethmoid a bony, double lamina.

2. Prevomer absent.

3. Nasal bones reduced.

4. Dermosphenotic prominent.

5. Parietals absent.

6. Anterolateral parasphenoid processes short, not contacting pterosphe-
noids.

7. Foramen magnum bordered exclusively by exoccipitals and basi-
occipital.

8. Exoccipital and basioccipital condyles well developed.

9. Transverse processes of first vertebra meeting dorsally, not articulating
with occiput.

10. No distinct gap between first and second vertebrae.

1 1. Transverse processes of precaudal vertebrae stout and elongate.

20 BREVIORA No. 347

12. Epiotic processes absent.

13. Metapterygoids absent.

14. Teeth strictly conical, disposed in two irregular rows on both jaws.

15. Premaxillary processes short and broad, extending over mesial maxil-
lary processes.

16. Mesial maxillary processes large and ligamentously attached to ventral
surface of ascending premaxillary processes.

17. Outer maxillary processes weakly developed.

18. Hypohyals consisting of a single element on each side.

19. Branchiostegal rays normally five on each side.

20. Third and fourth pharyngobranchials united.

21. Infrapharyngeals separated.

22. Two ossified basibranchials.

23. Posttemporal bone forked.

24. Supracleithrum present.

25. Postcleithra absent.

26. First rib greatly reduced, associated with distal tip of rib of second
vertebra.

27. Large space between cleithrum and coracoid.

28. Posteriorly-directed, bony pelvic process short or lacking.

29. Haemal canals moderately large.

30. Hypural fan a solid piece.

31. Vertebrae 26.

Comparison with Rivulinae. Differences between Fluviphylax
and Rivulinae are numerous and fundamental, the most important
being that Rivulinae have a cartilaginous mesethmoid, toothed
prevomer, metapterygoid, dorsal and ventral hypohyals on each
side, separated third and fourth pharyngobranchials, and a hypural
fan divided in two or even (in Aplocheilus) three parts. It should
be noted that our knowledge of the osteology of Rivulus is based
on a single species {R. bondi Schultz). Although there are many
species of Rivulus, the genus is very constant in its characters
(Myers, 1927: 119) and since Sethi found relatively few differ-
ences between R. bondi and other genera of Rivulinae it is reason-
able to expect only relatively minor differences among various
species of Rivulus. A number of non-osteological differences be-
tween Rivulus and Fluviphylax were pointed out above in the gen-
eric characterization of Fluviphylax.

Fluviphylax is a very distinct genus but is notably lacking in
bizarre or "unexpected" speciahzations or distinctive osteological
characters that would set it apart readily from all other cyprinodont
groups. Most of its osteological characters occur in many or even in
most cyprinodont groups, apart from Rivulinae, and thus are of little
value in assessing phylogenetic relationships. Regarding character
31, 26 is near the lowest number of vertebrae in cyprinodonts.

1970 AMAZONIAN CYPRINODONT FISH 21

Diminutive representatives of any group are likely to have reduced
vertebral counts approaching this number. Since an overwhelming
proportion of these characters occur in Procatopodinae and Fun-
dulinae, in contrast to other groups, the relationships of Fluviphy-
lax may lie with one of these two subfamilies.

Comparison with Procatopodinae. All Procatopodinae in which
the osteology has been studied agree with Fluviphylax in characters
4-13, 17, 18, 20-22, 24, and 26-30, and one or more species also
agree in characters 2, 14, 19, 23 and 25. In Poropanchax ran-
ciireli the prevomer is absent and the posttemporal is forked about
as in Fluviphylax (personal observation). Perhaps the most im-
portant difference between Procatopodinae and Fluviphylax is that
Procatopodinae have a cartilagmous mesethmoid and in none of the
Procatopodinae is the mouth position quite as in Fluviphylax. On
the other hand, the elevated pectoral fin position is virtually identi-
cal in Fluviphylax and Procatopodinae. In Poropanchax rancureli
the three uppermost pectoral rays articulate with the scapula and
the portions of the cleithra that form the posterior waU of the gill
chambers are almost vertical, whereas in Fluviphylax only the up-
permost pectoral ray articulates with the scapula and the posterior
wall of the gill chambers is oblique. These observations suggest
that the elevation of the pectorals in Fluviphylax and Procatopo-
dinae may have been achieved in different ways. It should be noted
that the pectoral fins are extremely elevated in the remarkable, tiny
cyprinodontoid Horaichthys, presumably related to Oryzias (see
Kulkarni, 1940: 379-423, and especiaUy figs. 2 and 18). In
Horaichthys, incidentally, the dorsal fin is very small and its posi-
tion is more posterior than in any other cyprinodontoid. This sug-
gests that the high-set pectorals, small, posteriorly-placed dorsals
and reduced body size in many procatopodins and in Fluviphylax
are functionally correlated characters.

Comparison with Fimdulinae. AU Fundulinae in which the os-
teology has been worked out agree with Fluviphylax in characters
1, 4, 7-11, 13, 15, 16, 18, 20-24, 26 and 28-30, and one or more
species also agree in characters 6, 12, 14, 17, 19 and 25. The
most important differences between FunduUnae and Fluviphylax
appear to be that Fundulinae have a prevomer and parietals, and
low-set pectoral fins with a restricted space between coracoid and
cleithrum. Pectoral fin position and amount of space between cora-
coid and cleithrum may weU be functionally correlated characters.
The prevomer and probably the parietals have been lost in several
cyprinodont lines, so their absence is not wholly unexpected in a

22 BREVIORA No. 347

form so tiny as Fluviphylax. Most adult Fundulinae have poste-
riorly-directed epiotic processes but these develop late and are
sometimes unformed in specimens the size of Fluviphylax. Of
described Fundulinae, perhaps Fimdulus notatus bears the greatest
superficial resemblance to Fluviphylax. A fish derived from
F. notatus but smaller (and with fewer vertebrae), with enlarged
eyes, open instead of tubular cephalic laterosensory canals, a con-
siderably shortened snout (involving loss of prevomer, reduction
of nasal and lacrimal bones, and shortening of ascending premaxil-
lary processes), shghtly elevated pectoral fins (with an enlarged
space between coracoid and cleithrum), and a more posteriorly
placed dorsal fin (with fewer rays) would be exceedingly hke
Fluviphylax.

Whatever its relationships may be, Fluviphylax does not belong
in any of the recognized subfamilies. Therefore a new subfamily is
proposed for it.

FLUVIPHYLACINAE new subfamily

A subfamily of very small, oviparous cyprinodonts with excep-
tionally large eyes, free orbital rim, dorsal fin placed far posteriorly,
pectoral fin set high, tips of pectoral fin rays extending beyond base
of pelvic fins. The main osteological characters of the subfamily
are fisted and discussed above. The subfamily, perhaps ecologically
equivalent to some of the African Procatopodinae, is known only
from the Amazon basin.

DISCUSSION

Fluviphylax is closest phenetically to Procatopodinae, and per-
haps it is also closest to them phyletically, but relationship with
Fundulinae is too strong a possibility to be ignored. At least until
the phyletic relationships of Fluviphylax are clearer it is best to
recognize it as forming a monotypic subfamily, the Fluviphy-
lacinae. In considering the relationships of Fluviphylax, I have
refrained on purpose from zoogeographic speculation. In this in-
stance zoogeographic aspects are so controversial they can only
serve to confuse matters until relationships are clearly understood.
That Procatopodinae are almost exclusively restricted to conti-
nental Africa is insufficient grounds to disquaUfy them as possible
relatives of Fluviphylax. In this connection, I have tended to think
of Fluviphylax as relatively recently derived simply because of its
tiny size, a conclusion which may be quite erroneous.

1970 AMAZONIAN CYPRINODONT FISH 23

Fluviphylax illustrates a feature of cyprinodont distribution that
should be elaborated, namely that the tiniest forms occur in the
midst of the world's richest fresh water fish fauna (Amazonas),
whereas the relatively giant forms are found in situations geo-
graphically or ecologically isolated from rich fish faunas. The only
other described oviparous cyprinodont (if it is truly an oviparous
cyprinodont, and adult) as small as Fluviphylax is Huhbsichthys
laurae Schultz, represented by a single 14-mm specimen from
"Pampan, Estado de Trujillo, Venezuela, probably Rio Motatan
drainage" (Schultz, 1949: 96-97). The smallest viviparous cy-
prinodont, Poecilia (Pamphorichthys) minor (Garman), is Ama-
zonian. As stated by Garman (1895: 92), "the length of the
males is about seven-tenths of an inch and that of the females about
eight" and "females of three-fourths of an inch in total length
contain fully developed embryos." I have examined a 14.5-mm
specimen collected by EPA in Lago Parauacui, near Oriximina,
state of Para, in which the gonopodium is fully mature and serrae
are formed on the second ray in the pelvic fins of both sides (cf.
Hubbs and Hubbs, 1945: 295-296, fig. 2). The types were ob-
tained by the Thayer Expedition near Villa Bella, roughly a hun-
dred miles up the main Amazon River from Obidos and about the
same distance from Oriximina (for a map showing the route of
the Thayer Expedition, as well as the routes of Spix and Martins,
Castelnau, Natterer and Schomburgk, see Eigenmann, 1917, pi. 1).

Myers (1966: 769-771) pointed out that poeciliids apparently
radiated into forms of diverse ecology and size in "the Ostario-
physan vacuum" of Central America before Cichlidae arrived there.
The advent of cichhds may have cut down this diversity, especially
in the larger, stream-inhabiting poeciliids, "leaving the many
smaller poeciliids, as in North and South America, only in what
might be termed peripheral habitats." Pamphorichthys provides a
beautiful example of a tiny poeciliid in a cichlid- and characid-rich
environment.

Regarding the largest cyprinodonts, the Orestiidae occur only in
Andean lakes and associated rivers and streams in Bolivia, Peru,
and northern Chile at altitudes between eight and 16 thousand feet
(Eigenmann and Allen, 1942). Anableps, Belonesox, and the
largest species of Fimdulus and of Poecilia all are brackish-water
forms generally restricted to estuarine or near-estuarine habitats or
are euryhahne and tend to occur either in fresh water where the
fresh water fish fauna is relatively poor (e.g., Belonesox) or else in
marine and semi-marine habitats where the marine fish fauna is

24 BREVIORA No. 347

relatively poor (e.g., most of the largest species of Fundulus). I
suspect that Belonesox would be unable to coexist with the preda-
tory characoids Hoplias or Ctenolucius , and that this may have
helped shape the southern boundaries of its range. Orestiidae may
be unable to withstand the introduction of gamefish such as trout
into their mountain retreats.

During fieldwork in Ghana I gained the impression that the
largest Rivulinae there, particularly Epiplatys, usually occur m
isolated situations such as in small streams above waterfalls or
along the margins of swampy areas, where they are sometimes the
only fishes present. I suspect that the largest Amazonian Rivulinae
tend to be similarly isolated. The largest West African procatopo-
din, Aplocheilichthys spilauchena, generally is restricted to estua-
rine or near-estuarine habitats and the lowermost reaches of rivers.
In a recent account of the Procatopodmae of Ghana, Scheel (1968:
277) did not mention the presence of this species there. It is
abundant near the mouths of the Volta and Tano rivers and prob-
ably occurs in suitable habitats along the entire coast of Ghana.
It is the only member of the Procatopodinae to occur on both sides
of the Dahomey Gap, a fact explained by its essentially estuarine
distribution. The species has been recorded from Senegal to lower
Congo (Boulenger, 1915: 62), and from the island of Fernando
Poo (Scheel, 1968: 278). It is apparently the only procatopodin
to have reached a locahty separated by the sea from continental
Africa. It may have reached there in the not-too-distant past when
the island probably was connected to the mainland, some sixty
miles distant. The Procatopodinae, incidentally, may qualify as
primary freshwater fishes, but we need more physio-ecological data
to determine this.

Hoedeman (1956-61) worked out supposed homologies be-
tween scales in the dorsocephalic squamation in a large number of
cyprinodont species and used this information as a basis for sug-
gesting relationships at various taxonomic levels. Species and
species-groups do have characteristic scale patterns, and these may
be useful in determining relationships between forms in which the
patterns have remained relatively unchanged. But the difficulties
of interpretation increase proportionately as the number of scales
changes or the patterns are radically different. The fundamental
problem involves homologies. In two of the three major criteria
for determining whether structures are homologous, morphology
and embryology, the head scales presumably are all identical.
That is, they have the same histological structure, and so far as we

1970 AMAZONIAN CYPRINODONT FISH 25

know, arise from anlage which are indistinguishable one from
another.

Hoedeman frequently explained reductions in number of head
scales as being due to "fusions." Surely decreases and increases
in the number of head scales correspond to the formation of fewer
or more scale anlage, not to fusion or the analogous process im-
plied, "splitting" of scales. Change in scale number probably is
accompanied by partial or total rearrangement of spacing between
scale anlage in an otherwise essentially uniform field with scale-
forming potential. If this is so, is it not erroneous to apply homolo-
gies as Hoedeman has done?

The sole criterion Hoedeman used to determine homologies is
that of relationships to other structures. He assumed that the scale
overlying the pineal organ is homologous in all cyprinodonts
(Hoedeman, 1958a: 23). He did not point out any anatomical
connection between the pineal gland and this scale, however, and
the relationship would seem to depend entirely on the superficial
characteristic of position. Using this scale, the "a scale," for a
starting point, Hoedeman extrapolated the rest of his homologies
according to the spatial relationships of the remaining scales to it
and to each other. This is clearly a precarious operation. Granted
that all cyprinodonts develop dorsocephalic squamation, one
would expect to find a scale more or less directly over the pineal
organ in every species. In fact, the exact position of the scale over
the pineal organ varies considerably, and sometimes there are two
scales side by side more or less directly above it. In Rivulinae, the
scale in the "a" position usually is overlapped by all of the scales
neighboring it, whereas in various species belonging to other sub-
families it is the topmost scale (Hoedeman, 1958a)!

On the other hand, some of the relationships between scales
pointed out by Hoedeman are remarkably constant; furthermore,
within groups the scales tend to remain in the same relationship to
segments of the cephalic sensory canal system. In many cyprino-
donts (c/. Rosen and Mendelson, 1960; Whitehead, 1962, fig. 12)
what I would interpret as the "g scale" is unpaired and has a seg-
ment of the supraorbital cephalic sensory canal on either side of
it; occasionally it is the anteriormost scale. Immediately behind it
is a pair of "e scales," which are frequently notched at their
posterolateral borders where they are approached by the dorsal-
most portion of the posterior infraorbital canal. Behind these
e scales is a median scale (sometimes paired), the a scale, near the
posterior margin of the eye. In cyprinodonts with such a pattern

26 BREVIORA No. 347

perhaps we may speak in terms of homologies between scales in
corresponding positions in different species.

As I have pointed out in another connection (Roberts, 1967:
251-252) reduction or loss of structures, which is to be expected
in small species, frequently leads to convergent characters. In
Pantanodon podoxys (Whitehead, 1962: 125-126, fig. 13), as in
Fluviphylax, the dorsocephalic squamation is so reduced that it is
of no help in assessing relationships. In the interpretation White-
head hesitantly decided upon, Pantanodon has a large h scale, two
g scales, two e scales, and a single a scale. According to this ar-
rangement, not only is there a pair of g scales, but the a scale is
further posterior than in any other cyprinodont! If homologies can
be drawn between the scales in Pantanodon and other cyprino-
donts, one might re-interpret Whitehead's two e scales as a pair of
a scales, his pair of g scales as e scales (which are usually paired),
and his h scale as a ^ scale (which is usually single). When this is
done, it turns out that the dorsocephalic squamation in young
Pantanodon podoxys is rather similar to that in Fluviphylax.

The remarkable silvery sheen on the dorsum of the eyes of
Fluviphylax deserves further comment. A silvery sheen on the
dorsal surface of the eyes is characteristic of certain African lamp-
eyes (Procatopodinae) and also, surprisingly, of Oryzias mela-
stigma; a similar color feature occurs in the little atherinoid
Pseudomugil signatus, also known as the Australian Blue-eye (per-
sonal communication from N. R. Foster). An analogous silvery
color, brilliantly visible when the fishes are viewed from above as
they swim near the surface, is the shining occipital spot in Aplo-
cheilichthys spilauchena, Aplocheilus panchax, A. lineatus, some
species of Epiplatys, Fundulus olivaceous, F. notti (the Starhead
topminnow), and F. notatus (personal communication from N. R.
Foster and G. S. Myers). Miehe (1911) suggested that the occipi-
tal spot in Aplocheilus panchax facilitates capture of the insects it
preys upon by luring them closer to the water surface. A more
basic function may be to reflect harmful radiations so that they do
not pass through the brain or eyes, as the case may be. In Fluvi-
phylax, which generally inhabits shaded places, the eyes' silvery
mantle may lessen the need for visual accommodation and mini-
mize the dazzling eftect of bright light as the fish swims into a sun-
fit area.

1970 AMAZONIAN CYPRINODONT FISH 27

LITERATURE CITED

BOULENGF.R, G. A.

1915. Catalogue of the fresh-water fishes of Africa in the British
Museum (Natural History). 3, pp. xii + 526.

ElGENMANN, C. H.

1917. The American Characidae. Mem. Mus. Comp. Zool. (Harvard),
43, pt. 1, pp. 1-102.

ElGENMANN, C. H., and W. R. Allen

1942. Fishes of Western South America. University of Kentucky,
XV -f- 494 pp.

Garman, S.

1895. The cyprinodonts. Mem. Mus. Comp. Zool., 19 (1): 1-179.
HOEDEMAN, J. J.

1956. Die bisher beschriebenen Formen und Arten der Gattung
Rivulus Poey. Aquarien und Terrarien, 3 (7): 199-202.

1958a. The frontal scalation pattern in some groups of toothcarps
(Pisces-Cyprinodontiformes). Bull. Aquatic Biol., 1 (3): 23-32.

1958b. Rivulid fishes of the Antilles. In Studies on the Fauna of
Curasao and other Caribbean Islands, 8: 112-126.

1959. Rivulid fishes of Suriname and other Guyanas. In Studies on
the Fauna of Suriname and other Guyanas, 3: 44-98.

1961. Studies on cyprinodontiform fishes. 10. On the probable evolu-
tion of the frontal scalation patterns. Bull. Aquatic Biol., 2
(18): 82-92.

HuBBS, C. L., and L. Hubbs

1945. Bilateral asymmetry and bilateral variation in fishes. Pap. Mich.
Acad. Sci., Arts Lett., 30: 229-231.

KULKARNI, C. V.

1940. On the systematic position, structural modifications, bionomics
and development of a remarkable new family of cyprinodont
fishes from the province of Bombay. Rec. Indian Mus. Calcutta,
42: 379-423.

MlEHE, H.

1911. Das Silberfeld des Haplochilus panchax und seine Reakiion auf
Licht. Abhandl. Siichs. Akad. Wiss. (math. phys. Kl.). 32:
362-375.

Myers, G. S.

1927. An analysis of the genera of neotropical killifishes allied to

Rivulus. Ann. Mag. Natur. Hist., ser. 9, 19: 115-129.
1955. Notes on the classification and names of cyprinodont fishes.

Tropical Fish Magazine, 4 (4): 7.
1966. Derivation of the freshwater fish fauna of Central America.

Copeia, 1966 (4): 766-773.

28 BREVIORA No. 347

Myers, G. S., and A. L. Carvalho

1945a. Notes on some new little-known Brazilian amphibians, with an
examination of the history of the Plata salamander Ensatina
platensis. Bol. Mus. Nac, Rio de Janeiro, Zool., 35: 1-24.

1945b. A strange new leaf-nosed lizard of the genus Anolis from
Amazonia. Bol. Mus. Nac, Rio de Janeiro, Zool., 43: 1-14.

Roberts, T. R.

1967. Virilia, a new genus of sexually dimorphic characid fishes from
West Africa, with remarks on characoids having an incomplete
lateral line. Stanford Ichthyol. Bull., 8 (4): 251-259.

Rosen, D. E.

1964. The relationships and taxonomic position of the halfbeaks,
killifishes, silveroides and their relatives. Bull. Amer. Mus.
Natur. Hist., 127, art. 5: 219-267.

1965. Oryzias madagascariensis Arnoult redescribed and assigned to
the East African fish genus Pantanodon (Atheriniformes,
Cyprinodontoidei). Amer. Mus. Novitates, no. 2240: 1-10.

Rosen, D. E., and R. M. Bailey

1963. The poeciliid fishes (Cyprinodontiformes), their structure,
zoogeography, and systematics. Bull. Amer. Mus. Natur. Hist.,
126, art. 1: 1-176.

Rosen, D. E., and J. R. Mendelson

1960. The sensory canals of the head in poeciliid fishes (Cyprino-
dontiformes), with reference to dentitional types. Copeia, 1960
(3): 203-210.

SCHEEL, J. J.

1968. Description of a new species of Procatopodinae (Cyprinodonti-
dae, Pisces) from Ghana, with remarks on the frontal patterns
of scales and of neuromasts in West African procatopodine
species. Rev. Zool. Bot. Afr., 78 (3-4): 277-283.

SCHULTZ, L. p.

1949. A further contribution to the ichthyology of Venezuela. Proc.
U. S. Nat. Mus., 99 (3235): 1-211.

Sethi, R.

1960. Osteology and phylogeny of oviparous cyprinodont fishes.
Doctoral thesis, University of Florida. University Microfilms,
Ann Arbor, Michigan, 275 pp.

Whitehead, P. J. P.

1962. The Pantanodontinae, edentulous toothcarps from East Africa.
Bull. Brit. Mus. (Natur. Hist.), Zool., 9 (3): 103-137.

Whitley, G. P.

1965. Some fish genera scrutinized. Proc. Roy. Zool. Soc. New South
Wales, 1964-65: 25-26.

(Received 28 January 1970.)

Nf^^Cls DO NOT CIRCULATE

MUS. COMP. ZOOL.
LIBRARY

B R E V I O R A970

^ « HARVARD

MiiseiiiiitM of Coimparatave ZoEDilUag^riTYi

Cambridge. Mass. 15 May 1970 Number 348

THE RELATIONSHIPS OF THE SKINKS REFERRED TO

THE GENUS DASIA

Allen E. Greer

Abstract. The time honored lygosomine genus Dasia is divided into
three genera (Dasia, Lamprolepis and Apterygodon) primarily on the basis
of differences in skull osteology and external morphology. Dasia (6 species)
and Lamprolepis (3 species) seem to have evolved independently from a
primitive Mahiiya-WkQ ancestor, whereas Apterygodon ( 1 species) seems to
have evolved from a basically Dasia-Wke stock. All three taxa appear to
have had southeast Asia as their area of origin.

When Boulenger (1887) published his third volume of the
Catalogue of the Lizards in the British Museum, he placed three
species in the subgenus Keneuxia. In the two major revisions of
lygosomine skinks since Boulenger (M. A. Smith, 1937; Mittle-
man, 1952), these three species plus other more recently de-
scribed species have been grouped together in the genus Dasia.
Yet in spite of the long historical precedent of treating these
species as a generic entity, there are no characters that will serve
to unite the species in a common genus, distinct from other skink
genera. On the contrary, there are several characters by which
the species can be assigned to three distinct genera, viz.: Dasia
(6 species). Apterygodon (1 species), and Lamprolepis (3 spe-
cies).' Curiously enough, the type species of the three genera are
the three species that comprised Boulenger's (1887) subgenus
Kerwuxia.

In Table 1 I have listed, roughly in order of what seems to me
to be their taxonoinic importance, the characters that readily serve
to distintzuish the three cenera.

1 The names of the two genera beside Dasia are available from the pri-
mary synonymy of Dasia.

BREVIORA

No. 34S

UJ

_i

CQ
<

I-

50

c

St)

CO

o
<u

a
o

60

>

U3

15

c

(U

J2

o

IS
u

o

o

c

60

4J ^ S
1^ S c

u

(U

a

a.

5.

5
--J

S

Q.

O ^

60 :/;

4J O

0 2
■a

• - (U

1 S

-2 .2
a. E

•a
S

— —

.2 6).

u iZ

<u >

a ^

Cl- o

60

c 2i

^ E

o

60

1- 60

CL. a

o

60

<«-l

SO

w

o

D.

E
E

■2

ami of

1).

3

>- 60

1

c

'E

C

>>

_c

^

<

_C

60

_2

u

y;

"^

y5

.

"5

yj

Q.

1)

T3

D

' ■

O

"^

l»

>.

~"

O

X

C

'H/j

C

60

■•"^

-a

,

^

c
o

Cw

.r.

U-

\r.

"5

^«

V

r3

rt

if
O

60
>.

1)

60

5

C

c

c

a

ijl

'-

u

a.

-y.

D.

— ■

-r-.

O

J^

■o

'.^-

>^

^

y

73.

'5

60

O
2

^

o

«

•a

C3

u

o

y5
C

O

2

O

>
15

C
,-3

o

60

c

C3
X

y5

"o

yj

60

C3

X

g

60

C
3

>>

E

c

O
u
o

o

D.

.^

O
>■

o

D.

O

c

X

X

X

o
c

Q.

u
O

c
o

•■*-

Cw

c

^

JH

O

y?

X

r-.

7Z

D.

y;

C

X

d

60

O

E

X

y?

(4—

o

5

o

^

c«

cS

c
o

4J

u

c

V5

o

4— »

X

A

—

OJ

J^

c

S

D.

X)

15

_o

u

6/j

C

o

>-

x:

<D

a.

"2

o

'y; ^"

O <^.

a.

>

4J

'5

60

X

^ 6/j

b ^

X

?3

yi

-o
"d

60

u

a.

E
J-.

y:

.:4

c

X

'J-.

u-

o

-a

"D ^

^

>>

y;

~

■ - o

60 >

<

>

O

iT

1
c

X

>>

E

c

y:

6/j

.

^

u

1-
o

^-^

«

-o

ri

eJj

D.

O

O
X

OJ
■*-•

C

>>

03

E

60

E

"2
'S

;::^

Ij

X

•a

OJ

60

TD

o

c

1/3

O

60

X

x

y

o

X

D.

."2

D.

>.

C

ID
OJ

o

60
>.

5 E
t "

Z a.

o .E

6/; .-3

c -o

D.

«.

n.

--3

L.

x:

U

o

•^

_c

c

c

UJ

^^-

"3

y?

y)
X
:3

o

X

«

:-.

SJ

O

<U

'~

X

i

yi
13

C
O
X

D.

\n

O

^^

O

E

-«— 1

X

«

E

y

o
c

o

X

o

o

_2

'5

o
E

CO

y;
O

1>
c

'n

a

a.

O

'^

«

w

■y.

a

y^

<L>

C3

X

:^

y

u

3

D.

OJ

, :

^—

y

C

o

>.

a>

■yi

L.

lU

■X.

<u

'S5

w

id

D.

a

a.

y5

4J

c

— '

E

o

^'

o

>,

o

o_

>.

X

_„

"y?

•T3
O
X

■^

.?

O

X

u

c

*-•

"3

u
O
"y

a

ii y;

"«

O

d

a

J.

^

Z

£ 2

O X ~

o

O

y5

O
Xi

C
u

l-(

o

c

(U

yi

c

X i4_

CS

a.

y5

W

«

;-

c«

a

2

OJ

■y

.5

y:

>~,

"5

—

w

w

u

y

D.

c

15

E

o

o

yi

o

o

C

c

«

■^j

D.

v^

.5

OJ

Q

L_

!J

:g

X

.~

Q.

OJ

y

c

'■J

^

c

^

<;

—

O

1970 SKINKS REFF.RRFD K) Nil (iHNUS DASIA 3

Dasia Gray

Dasia Gray, 1839. Ann. Mag. Nal. Hist., vol. 2. p. 3.'^5 (Type species:

Dasici olivacfd Ciray. 1839, by monotypy).
Liotiopis Fitzinger. 1843, Syslema Reptilium, p. 22 (Type species: Eiipiepes

ernestii Diimcril and Bibron, 1839 = Dasia olivacea Gray, 1839, by

monotypy).
Tliccony.x Annandale, 1906, Spolia Zeylanica, vol. 3, p. 191 (Type species:

Euprcpcs luiliaiuis Haly and Nevill. 1887, by monotypy; not Thccanyx

Gray, 1845).

Diagnosis. Of all the skinks with supranasal scales (Apterygo-
don, Dasia, Emoia, Eugongylus, Eumecia, Lamprolepis, Mahuya,
Macroscincus, "Otosaurus," and Riopa) only Dasia, Eumecia,
Mabuya (17 of 30 species examined), and Riopa (3^of 13 spe-
cies examined) have pterygoid teeth. Eumecia is a genus encom-
passing one or two species with greatly reduced appendages and
digits and is easily distinguishable from the strong-limbed, penta-
dactyl Dasia. Dasia differs from Riopa and Mabuya in palatal
characters: in Dasia the palatal rami of the pterygoids meet medi-
ally, whereas they are separated by an interpterygoid vacuity in
Mabuya and by posteriorly projecting medial processes from the
palatines in Riopa ( Fig. 1 ) .

Referred species. Griffini Taylor, 1915; haliana Haly and Nevill,
1887; moultoni Barbour and Noble, 1916; olivacea Gray, 1839;
semicincta Peters, 1867; subcaerulea Boulenger, 1891.

There may be more species of Dasia than those listed here, for
Taylor ( 1963) believes that "more than a single species" has been
confused under the name olivacea.

Distribution. Extreme southwestern India {subcaerulea); Cey-
lon (haliana); southern Thailand, Cambodia, Laos and Vietnam
south of IS'" N Lat., Con Son, the Andaman and Nicobar Islands,
Borneo and Java (olivacea); Palawan Islands (griffini); Philip-
pines (semicincta) . Moultoni occurs on Borneo with olivacea in
the only known case of sympatry between any two species in the
genus. The distribution is disjunct; there are no species between
southern India and Ceylon and southeast Asia.-

' In an earlier paper (Greer, 1967b) 1 said that only 1 (houringi) of
these 13 species had pterygoid teeth. This is not correct, for a re-examina-
tion of this material shows two additional species (afer and piiticiata) with
pterygoid teeth.

- This particular disjunct distribution is relatively well known in other
reptiles and vertebrates. See the symposium on the zoogeography of this
distribution in the Proc. Nat. Inst. Sci. India, vol. 15, no. 8. 1949.

4

BREVIORA

No. 348

Lamprolepis
smaragdina

Riopa bowringi

Mabuya longicaudata

Figure 1. Ventral view of the secondary palate in Mabuya longicau-
data, a primitive lygosomine, and representatives of three genera — Dasia.
Lamprolepis and Riopa — thought to have evolved independently from an
ancestor much like this primitive lygosomine. Not drawn to scale. Abbre-
viations: E, ectopterygoid; P. palatine; PT. pterygoid.

970

SKINKS REFERRED TO THE GENUS DASIA

Although it is evident tiom the distribution outlined above, 1
should perhaps emphasize that Dasia, although quite widespread
in southeast Asia and the western end of the Indo-Australian
Arehipelago, is an almost classic example of a taxon whose east-
ernmost distributional limits extend to, and thereby help to define,
Wallace's Line.

Species taxonomy. Several factors make it ditticult to know
where the species boundaries should be drawn in this genus. First,
with the exception of moultoni and olivacea on Borneo, all the
species are allopatric. Second, with the exception of the Ceylonese
haliami, the species are very similar in several classically important
scale counts, and third, two of the species (moultoni and sub-
caeriilea) are only known from single specimens. In view of these
difficulties it may be worthwhile for me to express my opinions of
the species taxonomy of Dasia.

Taylor (1922) proposed that grifftni (Palawan) and semicincta
(Philippines) be reduced to the rank of subspecies of the wide-
spread olivacea of southeast Asia. Although the three forms are
similar in several important scale counts, the following features
serve to separate them and, to me, suggest specific status for each.

olivacea

Supranasals separated

Black bands broader
than light bands in
young

Postorbital bone
present

griffim

Supranasals meet

Black bands about
equal to light bands
in young

semicincta

Supranasals meet

Black bands broader
than light bands in
young (Fig. 2)

Postorbital bone
absent

I have been able to examine only a single skull of two of the
species, but as I have never found the presence or absence of the
postorbital to vary intraspecifically in skinks, 1 do not hesitate to
accept it as a reflection of specific difference until more material is
available to prove otherwise. Unfortunately I have not been able
to see a skull of griffini.

The type and only known specimen of moultoni is a young indi-
vidual that differs from the only other Dasia occurring with it on
Borneo, i.e., olivacea, in having the supranasals meeting medially
instead of being separated and in having the black bands narrower,
instead of broader, than the light bands. Moultoni is most similar
to griffini on Palawan, which has the black bands equal in width

6 BREVIORA No. 348

to the light bands and which shows, as a variation, 3 of the 4
supraoculars in contact with the frontal, as is the case in the type
of mouhoni. Other Dasia and some i^rijfini have but 2 of the 4
supraoculars in contact with the frontal.

Only two specimens of the south Indian siibcaenilea are known,
and the similarity between this species and the southeast Asian
oUvacea is reputed to be remarkable (M. A. Smith. 1935. 1949).
Apparently the most noticeable difference between the two species
is the presence of two black streaks on the neck of subcaenilea
that are never present in olivacea. The great geographic distance
between the two species is no assurance that other, as yet undis-
covered, difference might exist between the two species, for as
mentioned above (see footnote on page 3), the relationships be-
tween many vertebrates in the two separate areas of southern
Jndia-Ceylon and southeast Asia are surprisingly close. There are,
however, only a few examples of conspecific relationship between
nontlying animals in the two areas, and this fact, along with the
slight differences that do exist between the two forms, tend to make
me retain subcaenilea and olivacea as distinct species.

The Ceylonese haliana is readily distinguished from other Dasia
on the basis of a lower number of longitudinal scale rows at mid-
body (24 instead of 26-30) and in the persistence of the juvenile
banding into adulthood.

Reproduction. Information on the mode of reproduction is
available for only two of the six species of Dasia and both are ovi-
parous. M. A. Smith ( 1935) says that olivacea is oviparous and
lays ''six eggs at a time." but Cantor ( 1 847 ) found 1 I eggs in one
female. Deraniyagala (1953) reports that a captive haliana laid
"two soft-shelled eggs."

Field notes. As far as is known all the species of Dasia are
arboreal to semiarboreal.

According to M. A. Smith ( 1935) and Taylor (1922), olivacea
is arboreal and subarboreal in its habits and. in addition, shows "a
preference for small islands" (M. A. Smith. 1935). Hendrickson
( 1966) found that on Palau Tioman. olivacea occurred mainly on
the coastal plain where it replaced the upland Mabiiya multi-
jasciata, but it was not as common in its knvland habitat as was
M . niultifascidta in the upland forest.

Mr. James P. Bacon. Jr.. who has observed olivacea from plat-
forms in the forests of Sarawak, informs me (letter of 15 July
1968) that he and his group saw ten specimens, all of which were
above 50 feet from the forest floor, and with respect to sites and

1970 SKINKS REFERRED TO THE GENUS DASIA 7

types of activity were similar to Aptcryi^odon villains (see below).
There was, however, one major difTerence between these two ar-
boreal skinks: "In comparison to vittalus, which scampers around
like a terrestrial Mahuya or Sphenoiuorphiis, olivacea moves with
almost painful deliberateness."

Taylor (1915, 1922) has the following comments on Dasia
semicincta, a close relative of olivacea: "Specimens of this species
were collected from the tops of large felled trees. It is a retiring
species and is never observed on the ground," and "it is retiring
and is seldom observed. It is arboreal in habit, feeding for the
most part in the trees. Usually the animals inhabit trees which are
densely covered with vines."

W. C. Brown (letter of 20 June 1968) has very kindly sum-
marized for me his observations on Dasia griffiiii, a close relative
of both olivacea and semicincta: "The specimens of griffini were
collected in dipterocarp forest between 200 and 1000 feet eleva-
tion in the Thumb Peak area of Palawan. Most were taken in
bark crevices or beneath loose bark on both living and dead trees.
Two I were | collected on rocks on the river bank. . . . This is not
an abundant species."

All we know of the ecology of snbcaenilea is that the type and
second known specimen were collected at 1 1 00 and 5900 feet
altitude respectively. M. A. Smith (1949) thought that the sub-
digital lamellae of the species indicated that it had "arboreal
habits."

Dasia haliana is considered by Deraniyagala (1953) to be a
"rare arboreal species restricted to the low country" of Ceylon.
The same author mentions two individuals being captured from a
hole and a cavity of a tree, respectively, and "according to Haly
it lives on the tops of high trees" (Smith, 1935 ).

Deraniyagala (1953) provides another important clue to the
behavior of haliana in his comment, "the gait of this arboreal spe-
cies is much slower than in other Ceylon skinks even when climb-
ing a tree." This comment calls to mind Bacon's statement of the
"almost painful deliberateness" with which olivacea moves about
the trees. One wonders if the slow movements of these two species
may not be characteristic of all the species of Dasia in contrast to
the more usual, quicker motions of the species of the other two
genera recognized here.

Juvenile color patterns. As far as is known, all juvenile Dasia
are strikingly marked with light and dark crossbands. The bands
persist in adult haliana but disappear with age in the other species
for which information is available.

8

BREVIORA

No. 348

Taylor's (1922) account of a juvenile semicincta is the most
vivid available for any species in the genus: "The body above is
glossy coal black with a series of brilliant orange yellow bars
from the tip of the snout to the end of the tail. . . . The legs and
digits I are I barred with canary yellow; the regenerated tail is brick
red; below, on j the | body the bars widen and lose themselves in
the immaculate canary yellow of the abdomen." There is a total
of eight to nine yellow bars on the neck and body of this specimen
described by Taylor. Another juvenile with ten light crossbands
on the neck and bodv is shown in Figure 2.

u

FiGURi 2. f^orsal view of a juvenile (snout-vent length ^ 57 mm) and
adult (snout-vcnl length = 117 mm) Dasia sciiiicinctu from the Philippine
Islands showing the light and dark crossbands in the juvenile and their
absence in the adult.

1970 SKINKS REFERRM) K) TlIF GFNUS DASIA 9

According to M. A. Smith ( 1935), "young \<>livacea\ arc black
above, with narrow silvery or yellowish, rather irregular transverse
crossbars, from II to 14 in number, on the neck and body."

A late juvenile or subadult i^rif]mi (snout-vent length — 67 mm)
in the Museum of Comparative Zoology (MCZ) may indicate
something of the juvenile coloration for this species. This speci-
men is light brownish above with 14 dark brown crossbars, en-
compassing I 3 light interspaces, from the nape to the base of the
tail.

The type and only known specimen of luoiiltoiu is in the MCZ.
It is clearly a juvenile (snout-vent length = 45 mm) and is pre-
dominately yellowish brown with 14 narrow dark brown cross-
bands, separated by 1 3 wider light interspaces, from the nape to
the base of the tail. The crossbands extend laterally to, but not
across, the venter.

In luiUana "the number of black crossbands | in both juveniles
and adults | varies from five to six from the shoulders to above
I the I hips. . . . The width of the black bands alters with age. In
the young they are equal to or only slightly narrower than the light
interspaces. . . ." (Deraniyagala, 1953). Annandale (1906)
figures a juvenile haliana in which there are seven dark crossbands,
encompassing six light interspaces, from the nape to the base of
the tail. According to Deraniyagala (1953) the "young | are |
pink with black crossbands."

The two known specimens of suhcaerulea arc 57 and 70 mm in
snout-vent length, but both lack any trace of crossbarring. It
would seem, therefore, that either this species has no crossbanded
pattern in the juvenile or the only known specimens are too old to
display them.

The most interesting question about these crossbanded patterns
is, of course, their function. Their function, however, is almost
surely related to the ecology and behavior of their bearer, and as
we know precious little of these matters in the adults and nothing
in juveniles, we have very few facts with which to work.

Striking crossbanded patterns such as those seen in Dasia
are really quite unusual in skinks, whether adult or juvenile. In
fact, there are only about six other species among the approxi-
mately 800-(- species of skinks that have an even closely com-
parable pattern, viz.: the scincines Scincopiis jasciatus, Scinciis
scinciis (some individuals), and Scelotes splendidus, and the lygo-
somines Sphenomorphus fasciolatus, Tiliqua occipitalis (some in-
dividuals), and T. genardi (especially pronounced in juveniles).

10 BREVIORA No. 348

Other morphological evidence argues against a common ancestry
for Dasia and any of these other species with crossbanded pat-
terns, so their similarity in this regard is truly convergent.

The species of this list do share one ecological similarity, i.e.,
they are all ground dwellers that occasionally penetrate the loose
substrate, but this is not a particularly unique ecological situation
for skinks and it carries no ready explanation of the adaptiveness
of the crossbanded pattern in such a situation.

Perhaps the juvenile color patterns are mimetic, taking advan-
tage of either a specific but as yet unknown aposematic model or
a general abstract model recognizable to predators gaining their
experience with several different specific models. It could also be
that the young occupy the same general habitat as the parents, and
the juvenile color pattern acts as a signal to inhibit feeding attacks
by the adult.

Relationships with other genera. In any group of skinks it is
reasonable to assume that pterygoid teeth and supranasal scales
are primitive features which may be lost in more advanced spe-
cies. In possessing these two characters, therefore, Dasia is a
primitive genus. On the basis of the secondary palate, however,
Dasia is advanced, for the close medial apposition of the palatal
rami of the pterygoids may be reasonably thought of as the ex-
treme development of the uniquely scincid secondary palate
(Greer, 1970).

The only living lygosomines that are like Dasia in retaining both
supranasals and pterygoid teeth are Riopa, Mabiiya, and Eumecia.
Of these three genera, Mabuya is the best candidate for the closest
ancestral relative of Dasia.

Eumecia is an African genus with one or two species. The taxon
is very similar to Mabuya in its skull morphology and is distin-
guishable from this genus solely by its attenuate body form and
reduced appendages and digits. Eumecia is clearly derived from
Mabuya (Greer, 1967b), and its diagnostic specializations make
it an unlikely ancestor for any other living skink.

If one combines all the primitive features of the various species
of Riopa (exclusive of Eugongyhis) . the similarities with Dasia
are impressive; supranasal scales, scaly lower eyelid, pterygoid
teeth and oviparity. There is, however, an important difference in
the relationships of the bones of the secondary palate, for in Riopa
the pterygoids are separated by close-fitting, posteriorly-projecting
medial processes from the palatines (Fig. 1 ), whereas in Dasia the

1970 SKINKS REFERRED TO THE GENUS DASIA 11

pterygoids meet medially and are not separated by palatine pro-
cesses. In both genera the secondary palate, initiated by the medial
apposition of the palatine bones, is extended further posteriorly by
the partial (Riopa) or complete {Dasia) incorporation of the
pterygoids. A posterior extension of the secondary palate is un-
doubtedly an advanced condition in lygosomines, but there is only
a functional similarity between Riopa and Dasia in this regard, for
the way in which the complete secondary palate is extended pos-
teriorly, is fundamentally different in both taxa.

In addition to the palatal differences between Riopa and Dasia
which preclude any close phylogenetic relationship, there are other
important differences between the two groups. Riopa, for instance,
lacks the crossbanded patterns in the young that are characteristic
of Dasia, and it is also an essentially terrestrial taxon in contrast to
the arboreal Dasia.

Mabuya is quite unlike both Dasia and Riopa in that the palatal
rami of the pterygoids are widely separated with their medial edges
smoothly diverging posteriorly (Fig. 1). This palate is primitive
for lygosomines in the sense that the pterygoids are in an "un-
committed" state and could theoretically be incorporated into the
secondary palate in a number of ways. In fact, it is this primitive
palatal condition, along with the presence of supranasal scales,
scaly eyelids, pterygoid teeth, well-developed pentadactyl limbs, a
full complement of distinct head scales, and oviparity that makes
Mabuya (especially the southeast Asian species in which all these
characters occur) a very likely candidate for the ultimate ancestor
of any lygosomine skink. In this light, it is easy to regard Dasia
as being derived from a primitive Mabuya-\\k.Q ancestor by the
medial apposition and ultimate meeting of the palatal rami of the
pterygoids along the midline.^

Dasia is similar to Mabuya in regards other than the sharing of
the primitive lygosomine characters mentioned above. The body
scales of Dasia (primarily the posterior body scales in adults) are
keeled, as are the body scales of all the species of Mabuya except
those eight or nine species (exclusive of M. atlantica) inhabiting
the New World; these have smooth scales. Also, unlike Riopa and
Eumecia, which are largely terrestrial and cryptic in their habits.

1 Similarly, one can conceive of the evolution of Riopa from a primitive
Mabuya through the medial apposition of the palatal rami of the pterygoids
concurrently with the posterior extension of medial palatine processes
(Fig. 1).

12 BREVIORA No. 348

many species of Mabiiya are partly arboreal in their habits. Such
a propensity for arboreality could have provided the behavioral
background on which to build the morphological adaptations to
arboreality in Dasia.

To summarize this section we may outline the changes that oc-
curred in the evolution of Dasia from a primitive Mahuya-Wko.
ancestor (also see Fig. 6).

1 . The apposition of the palatal rami of the pterygoids along
the midline of the secondary palate until they met and formed a
posterior extension to the secondary palate.

2. The evolution of a striking crossbanded pattern in the
young.

3. The evolution of a greater degree of arboreality.

4. The development of a slower gait.

Zoogeography. It is interesting to note that Dasia occupies the
same general southeast Asian area occupied by the species that
seem to be its closest ancestral relatives, i.e.. the primitive species
of Mabuya. In fact, like Dasia, Mabuya is, with only one or two
exceptions, also confined to the area west of Wallace's Line. It
looks, therefore, as if the evolution of Dasia from a Mabuya-WkQ
ancestor, and perhaps even the early evolution of Mabuya itself,
is a phenomenon of the tropical parts of the southeastern Oriental
Region and has had little or nothing to do with the area east of
Wallace's Line, i.e., Wallacea and the Australian Region.

Apterygodon Edeling

Apterygodon Edeling, 1864, Nedel. Tijdschr. Dierk.. vol. 2, p. 201 (Type
species: Apterxiiodon vittatnm Edeling, 1864. by monotypy).

Diagnosis. Along with Dasia, Apterygodon differs from all other
skinks with supranasal scales in having the palatal rami of the
pterygoids meeting medially along the midline of the palate. Ap-
terygodon, however, differs from Dasia in lacking pterygoid teeth
and in having a small fingerlike process of the ectopterygoid bone
that projects anteriorly toward the palatine and partially excludes
the palatal ramus of the pterygoid from a position on the infra-
orbital vacuity (Fig. 3).

Referred species. Only the type species is placed in the genus.

Distribution. Borneo. There is also a single specimen in the
British Museum (Natural History) which is questionably from
Ceram. This locality is almost certainly in error, however, for,
with the exception of the Ceram record, the species has never been
recorded from anywhere other than Borneo.

970

SKINKS REFERRED TO Till- GENUS DASIA

13

Tropidophorus beccari

Sphenomorphus
solomonis

Figure 3. Ventral view of the secondary palate in Aptciyi^'odon and
representatives of two other genera, SphcnomorpJuis and Tropidophorus.
with species displaying similar palates. 1 he ectopterygoid process has prob-
ably evolved independently in each of these three groups. The palatal bones
of Apterygodon are partially disarticulated. Not drawn to scale. Abbrevia-

14 BREVIORA No. 348

Mode of reproduction. A female with a snout-vent length of
71 mm, collected in the Baram District of Sarawak, contains two
eggs in the right oviduct and one in the left oviduct. The three
eggs are surrounded by a fairly thick shell which indicates that the
species is probably oviparous.

Field notes. By letter, Mr. James P. Bacon, Jr. has very gen-
erously summarized for me his observations on vittatus which
were made in the course of platform observations on the altitudinal
stratification of the reptiles in the Sarawak forest. Below, I have
paraphrased the information on this species contained in Mr.
Bacon's letter of 15 July 1968.

In one year of platform work from 36 to 110 feet above the
forest floor there were 1 69 observations of vittatus. On the basis
of number of observations, vittatus was the most frequently en-
countered lizard; a Draco was the second most frequently en-
countered species with 32 observations. Vittatus occurs between
0-120-]- feet. The majority of the observations were made be-
tween 40 and 120 feet, although this range may reflect platform
distribution more than it does lizard distribution. The observations
made so far indicate that vittatus is the dominant canopy lizard in
Sarawak, at least in the upper canopy (A and B strata).

Vittatus is a sunloving skink and moves about on trunks and
branches of all kinds during the day. It spends its time exploring
bark crevices and the interiors of epiphytes and in basking. The
lizards are capable of short (half a foot or less) jumps between
adjacent branches and vines.

Dr. Robert Inger has also given me his brief but valuable im-
pressions about vittatus in the field. He says in a letter of 28 June
1968 that "vittatus is a sun lover. On rare occasions vittatus will
descend to the ground, but only at the base of large trees."

Variation. Undoubtedly because of its canopy-dwelling habits
(see "Field notes" section for the genus), Apterygodon vittatus is
a little known skink. The Museum of Comparative Zoology is
fortunate in having 15 specimens, and it is worthwhile to record
the variation in several taxonomically useful characters for these
specimens. All the specimens are from Sarawak.

TTie prefrontal scales are separated in all specimens, but two
individuals have the scales of the snout broken up into a number
of small, asymmetrical scales. Nuchal scales are lacking except
in two individuals, which each have a single pair. There are
always 4 supraoculars with 2 (in 1 8 of the 30 possible cases) or 3
(in the remaining 12 cases) supraoculars in contact with the

1970 SKINKS REFERRED TO THE GENUS DASIA 15

frontal. The fifth supraocular is beneath the eye except in t)ne of
the 30 possible cases where it is the fourth. There are 28-30 longi-
tudinal scale rows at midbody, and 16-22 (avg. ^ 18.2) sub-
digital lamellae beneath the fourth toe. All 15 specimens are
adults, and the largest two are each 72 mm in snout-vent length.

Relationships with other genera. Except for the absence of
pterygoid teeth and the presence of an ectopterygoid process,
Apterygodon and Dasia are very similar in several important skull
characters. In both taxa there is a small postorbital bone, the
palatal rami of the pterygoids are similar in shape in both taxa
and meet medially, and there are 9 teeth on the premaxillae. The
skull differences between the two genera do not really preclude the
idea of the close relationship, for, as will be shown below, an
ectopterygoid process has probably evolved numerous times in
lygosomines, and the loss of pterygoid teeth is a familiar feature
of lygosomine evolution.

Two aspects of the external morphology of the two genera make
the idea of their close relationships even more compelling. The
more remarkable of these is the numerous fine, approximately
longitudinal striations on the dorsal and lateral body scales and
the scales of the appendages. These striations were first noted by
M. A. Smith (1935) in Mabuya longicaudata and were later re-
discovered by Taylor and Elbel (1958) in Dasia olivacea and
Mabuya longicaudata. An extensive examination of other lygo-
somines has revealed the striations only in Dasia semicincta, D.
griffini, D. moultoni (the only species of Dasia available for exami-
nation), Apterygodon, and Mabuya tytleri. The striations, there-
fore, seem to be confined to Dasia, Apterygodon, and two species
of southeast Asian Mabuya, i.e., those Mabuya which are most
similar to the probable ancestor of Dasia. It appears then that
the striations are another indication of the close relationships
within this group of skinks. Unfortunately, the function of these
striations is completely unknown.

The second external character that aligns Dasia and Aptery-
godon is a pair of enlarged "heel scales" on the rear foot in the
males of both taxa. These scales in males are not only consider-
ably larger than the surrounding scales, but they also seem to have
an inner texture (viewed through the semitransparent epidermis)
which distinguishes them from nearby scales and which belies a
glandular function. In females this pair of scales is only slightly
larger than the surrounding scales, and they appear to lack the
textural differences that distinguish the scales from their fellows in

16 BREVIORA No. 348

males. Paired enlarged heel scales are also evident in some
Mabuya, but only in a few species (most notably M. longicaiuiata)
are they larger (and then only slightly) in males than in females,
and they are never differentiated as in male Dasia and Aptery-
godon (Fig. 4). The heel scales of Mabuya are probably homolo-
gous with thoss of Dasia and Apterygodon, but only in this latter
group have they become greatly dift'erentiated (to serve a glandular
function?) in males.

Because we know nothing about the juvenile color patterns of
Apierygodon, we have no idea how the genus compares with
Dasia in this important regard. It seems unlikely, however, that a
skink with an adult color pattern of anterior light longitudinal
stripes on a dark ground color would have a juvenile color pat-
tern consisting of light and dark crossbands, but admittedly my
feeling on this may stem more from a sense of "taste" than from
biological logic.

Aside from the differences in adult and possibly juvenile color
pattern, the only other nonskeletal difference of any significance
between Apierygodon and Dasia is in the gait, for, as mentioned
above (see "Field notes" section under Dasia), Apierygodon has a
faster, more commonly "skink-like" gait than does Dasia.

In summary, the differences between the two taxa are important
but do not override the similarities on which the idea of their close
relationship rests. I think, however, the differences do serve to
justify the generic separation of the two taxa. In attempting to
interpret these similarities and differences phylogenetically, we can
hypothesize that Apierygodon and Dasia once shared a common
ancestor that was decidedly Dasia-Wke except for the slower gait
(and cross-banded juvenile color pattern?), and from this ances-
tor, Apierygodon evolved on the one hand through the loss of the
pterygoid teeth and the independent evolution of an ectopterygoid
process, and Dasia evolved on the other hand through a line that
subsequently acquired a slower gait (Fig. 6).

Since this view of the phylogeny of Apierygodon and Dasia im-
plies that the ectopterygoid process in Apierygodon has evolved
independently of its evolution in other lygosomines, perhaps a
word should be said about the significance of this process in lygo-
somines in general. It is true that the process, especially when it
is well developed to the point of forming a firm articulation with
the palatine bone, correlates well with other characters, which,
taken together, serve to delimit large natural groups (Greer,
1967a). That this process, however, is not indicative of one and

970

SKINKS RI-FI.RRFD TO THE GENUS HASIA

17

E

CD

O

c

?3

X

c

r-
C

^ is

To

OQ

(LI ;.

o r-

y: ^

8?

< E

'^ r-

¥ ■£

^ c

-r (L>

c
u

>

18

BREVIORA

No. 348

only one major line of lygosomine evolution, i.e., that it has
evolved more than once, is suggested by two lines of evidence.
First, the process is variably developed in its anterior extent in
species that are clearly closely related, e.g., Hemiergis and the
Australian alpha Leiolopisma, and second, even those major
groups which show a consistently strongly developed process ar-
ticulating with the palatine often have very little else in common
and, in some cases, are quite distantly separated geographically
(Table 2). Thus not only has the process evolved independently
in Apterygodon, but it has also done so in several other lygosomine
lines as well. Again, like so many other skull characters that seem
to have some taxonomic importance, we have no information
about the function of the ectopterygoid process in these skinks.
Presumably it may have something to do with skull kinetics, which
in turn may have something to do with feeding habits. In any
event, once the function of the process is known, its evolutionary
significance and. therefore, its taxonomic significance should be-
come clearer.

Zoogeography. The Bornean distribution of Apterygodon fits
with the hypothesis of the close relationship of Apterygodon and
Dasia, for Borneo is in that area of the Oriental Region where
Dasia seems to have had much of its evolutionary history.

TABLE 2

Lygosomine taxa which are apparently only distantly related but
which as groups show strongly developed ectopterygoid processes.

Group

1.) Ctenotiis, Egernia (most species).
Tiliqiia, Coriicici, Tribolonotiis

2.) Leiolopisma virens. L. flavipes,
L. prehensicaiicla, L. semoni

3.) Tropiclophonis (some species)

4.) Solomonis subgroup of Spheno-
morphus (see Greer, 1967a)

5.) Macroscincus (see HofFstetter,
1949)

Area of Evolution
Australian Region

Australian Region

Southeast Asia, Philippines
Australian Region

Cape Verde Islands

1970 SKINKS REFERRED TO THE GENUS DASIA 19

LampioU'pis Fitzinger

Lamprolepix Fitzinger, 1843, Systcma Reptilium, p. 22 (Type species: Sciii-
ctis snuircigclina Lesson, 1830. hy monotypy).

Keneiixia Gray, 1845, Catalogue of the Specimens of Lizards in the Collec-
tion of the British Museum, p. 79 (Type species: Sciiuiis sniaragdina
Lesson, 1830. by monotypy).

Diagnosis. On the basis of its palatal characters ( Fig. 1 ) and
the single large pair of nuchal scales bordering the parietals, Lam-
prolepis is an alpha skink (Greer and Parker, 1968). Most alphas,
however, have either 11 premaxillary teeth (e.g., Anotis, Emoia,
Eiigongyliis, most Australian and Pacific Leiolopisma and ''Spheno-
morphus" bignelli, "S." minutus and "S." pseudornatus) , or a
spectacle in a movable lower eyelid, or permanent brille (e.g., all
alpha Leiolopisma and Ablepharus). They can thus be distin-
guished from Lamprolepis, which has but 9 premaxillary teeth
and a scaly lower eyelid. In these two characters Lamprolepis is
similar to the primitive alpha genus Mabiiya; in fact, the only way
Lamprolepis can be readily distinguished from all Mahiiya, except
the most geographically distant species in tropical America (ex-
clusive of M. atlantica), is by the absence of keels on the body
scales.

The only other alpha taxon not mentioned above is the African
Eumecia. This genus encompasses one or two attenuate species
with a reduced number of digits and is easily distinguished from
the more robust, pentadactyl Lamprolepis.

Referred species, nieiiwenhuisi Lidth de Jeude, 1905; smarag-
dina Lesson, 1830; vyneri Shelford, 1905.

Distribution. Two of the three species, nieiiwenhuisi and vyneri,
are confined to Borneo, whereas smaragdina is very widespread.
This species occurs from Formosa, the Philippines, Palawan and
the Sulu Archipelago, the Celebes and Lombok (but not Bali or
west) south and east through Micronesia (Palau Islands and the
Carolines east to the Marshalls), the Indo-Australian Archipelago,
New Guinea and Cape York (Australia) to the Solomon Islands

20 BREVIORA No. 348

and the Santa Cruz Islands.' In the Indo-AustraHan Archipelago
south of the Philippines, the western limit of suiarai^dhui s distribu-
tion follows Wallace's Line exactly (see fig. 7 on page 167 in
Mertens, 1930).

Reproduction. L. smaragdina is the only species in the genus
for which information about reproduction is available, and from
all accounts (Hediger. 1934; Alcala, 1966; and Fred Parker, per-
sonal observation ) as well as my own observations, it is clear that
the species consistently lays but two eggs in a clutch.

The eggs are laid inside the rotten wood and rubbish in hollow
branches and trunks (Fred Parker, personal observation), in the
"moist humus collected in rotting parts of tree trunks, exposed
roots, and points of intersection of main tree trunks ... in nest-
ing sites . . . very close to or several meters above the ground"
(Alcala, 1966) or, probably less frequently, on the ground under
logs or stones (Hediger, 1934).

Brown and Reyes (1956) found that the maximum incubation
period for snuiragdliui in the Philippines was 54 days, while
Parker reports that "an egg laid on 1 October . . . hatched on
27 November," indicating an incubation time of approximately
58 days.

Field notes. L. .snuiragdina is one of the best known lizards in
the Pacific. This notoriety is undoubtedly due to the species' great
abundance, its medium-large size and beautiful coloration, and its
propensity for the more or less bare parts of tree trunks. In an
effort to incorporate the best known aspects of the lizard's ecology
and behavior, I have chosen for presentation below the accounts
of four people, each of whom has seen smaragdina in the field in a
different part of its range.

' Reports of L. .snuinii^cliiui from Java and Borneo (de Rooij, 1915) are
probably in error. Mertens ( 1929) says that the species is certainly absent
from Java and attributes its inclusion in the Javanese fauna to a kind of
thoughtless recording of locality that was apparently fairly usual in regard
to species actually native to more eastern islands in the indo-Australian
Archipelago.

The single specimen on which the Bornean locality is based is not L.
<iiiuir(i<i(lin(i but /.. nieiiweuhiiisi (Mertens. 1929; Brongersma. 1933), a
species which is endemic to Borneo. Furthermore, Dr. Robert Inger, who
has been involved in recent studies on the herpetology of Borneo, writes
(letter of 17 July 1968): "So far as I know, l'>(i.\:a siiuiiai>clina does not
occur on Borneo, i have never seen any specimens in any museum here or
in Europe with a Bornean locality."

1970 SKINKS RF.FERRED TO IMF GENUS DASIA 21

According to Hcdiger (1934), who worked primarily in the
Bismarck Archipelago and in northern New Guinea, siuarui^dina
lives primarily in isolated trees, small groups of trees, and on the
forest edge, it is most usually found on tree trunks and resorts to
the tree branches only for protection from the rain and on cloudy
days. The lizard is extremely arboreal in its activities and ap-
parently descends to the ground only for an occasional insect and
sometimes to lay its eggs. When approached by an intruder, these
skinks characteristically scurry around to the other side of the
tree trunk just as squirrels often do. Although primarily insec-
tivorous, some individuals examined by Hediger had masses of
leaves and flowers in their stomachs.

In the Philippines snuinii^dina "has been found in dipterocarp,
freshwater swamps, and mangrove forests, wooded grasslands, and
coconut groves. It prefers areas with big trees that are exposed to
the sun, irrespective of the amount of ground cover. Its altitudinal
distribution is from sea level to about 600 meters" (Alcala, 1966).

Fred Parker, speaking of sinaniiidina on Bougainville and its
neighboring islands in the Solomons, says that it is "mainly an
arboreal species but during the day some can be found feeding on
the ground in undergrowth. When they are disturbed they run to
the nearest tree trunk. The preferred trees are those without thick
creeper growth, such as coconut palms and breadfruit trees. The
species is more common in open cleared areas, on trees still stand-
ing in gardens, and in small coconut plantings. In thick primary
forest they are rarely seen. Found all over the lowlands and into
the mountains but not very common over 2000 feet, except per-
haps in the Guava area."

Marshall (1951) had the opportunity of observing smaragdina
on the extreme northeastern periphery of its range on the islets of
Arno Atoll in the Marshall Islands. He noted that here "this
strictly arboreal skink is the diurnal counterpart of the Big Tree
Gecko I Gehyra oceanica | , and one or two can be seen on almost
every coconut trunk during their hours of activity from about
7:30 a.m. to 5 p.m. (if sunny). They may remain motionless for
long periods, pressed against the clear trunk, with the head held
out at an angle. On uninhabited islands where there is denser
vegetation, they will come lower in the leafy understory of vines
and shrubs but are rarely found on the ground. They occur on
just about all the islets of the atoll; though . . . none |were
found I on Autore, one was seen on tiny Rakijer, much smaller
than Autore, and containing only 3 or 4 coconut palms. A little

22 BREVIORA No. 348

islet next to Rakijer, which had only a growth of Scaevola, lacked
any lizards of any kind. . . . [Smaragdina] sleeps at night on the
trunk, wherever they happen to be at the end of the day, as evi-
denced by the fact that individuals will seldom be roosting at the
same spot on successive nights."

In contrast to this vast amount of information for smaragdina,
we know virtually nothing about nieuwenhuisi or vyneri. Ac-
cording to M. A. Smith (1931), two nieuwenhuisi were collected
at 3000 feet on Mt. Kinabalu, North Borneo, and a vyneri in the
Field Museum was collected 8 feet up a tree in forest.

Relationships within the genus. The two Bomean species of
Lamprolepis, nieuwenhuisi and vyneri, are extremely closely re-
lated, perhaps even conspecific, and both are seemingly more primi-
tive than smaragdina. Both nieuwenhuisi and vyneri, for example,
have retained supranasal scales, whereas smaragdina has lost them;
nieuwenhuisi and vyneri also lack the single enlarged heel scales
characteristic of male smaragdina (Gandolfi, 1907, and Fig. 5,
this paper), and in this lack the two species may also be primitive.^

Within smaragdina itself Mertens (1929) has suggested that
the Moluccas "subspecies," moluccarum, is the most primitive liv-
ing population. This argument was based on two lines of reason-
ing: first, this population is more or less centrally located in regard
to the other named populations of smaragdina and can therefore
be looked upon as occupying a central area from which the species
could have dispersed evenly in all directions, and second, the light
grey to greyish brown ground color of moluccarum — in contrast
to the unusual (for skinks) green color in most other named forms

1 Although the single enlarged heel scale of male Lamprolepis smaragdina
is very similar to the paired enlarged heel scales of Dasia and Apterygodon.
even to the point of having a glandlike appearance, the scales have probably
differentiated independently in each group. The best evidence for this lies
in the fact that the nearest living relatives of the two groups' common
ancestor, i.e., the primitive southeast Asian Mabiiya (see below), lack
such highly differentiated heel scales. An histological examination of the
scales" morphology and perhaps a chemical analysis of the cell products (if
the scales are glandular) would make an interesting extension of the com-
parison of these heel scales.

970

SKINKS REFERRED TO THE GENUS DASIA

23

3
O

c

ca

.«>

fT>-vV\ xi

•a

ri^(\ ^P\/

u

c

vtXXvVl

rt

) vCZC-/ 1/

/y

^^

A^~c4jr / /

I

E

^■^sPCv'/ V\A

^

E

00 ^

0)

1 1 c

.

(Q

1 1 ca

E

■5 5

.22

heel" scales of a male (snout-vent leng
laragdina from Bougainville, Solomon

vW5

1 ^

\

**

V

^

)"-( ^vJ^^sA^

^K /

u

VfL / ^Vv

1

■= .''

rTV I *

vV \

■" 5^

rS(1 1

fry

o^

/ /

A 1 f

(jk/ ^^^^ /

^A-\

C =-

>vV-C7/6

c2^—

_a>

E -

CD

E

ro

00

'^' II

1 1

D :/;

o c

— OJ

U. —

.^^

c

o

>

24 BREVIORA No. 348

of siuaragdina^ — is probably most similar to the drab ground color
of most other ground inhabiting skinks that are presumably most
similar to the ancestor of smarai^dina.

As they stand, Mertens' arguments are overdrawn, but they do
point the way; thus, while it is naive to place too much emphasis
in pinpointing the geographically central area of a species as the
center of its dispersal, I do think that the origin and early distri-
bution of .sinarai>dina is more likely to have occurred in the western
part of its range, in an area close to its nearest relatives and well
known as a source area, than in the remote Pacific island chains
where .smanii^dina reaches its easternmost limits. And while it is
probably true that a drab color pattern is more likely ancestral
to a bright green one than vice versa, it is unnecessary to imply
that the amount of green in a color pattern is strongly correlated
with evolutionary age. In this more diffuse argument the two lines
of evidence mesh as they were supposed to in Mertens' argument,
for the subdued, more achromatic (primitive) color patterns are
characteristic of the populations in the western part of the range,
where, as I said, smamgdina probably first arose and from whence
it dispersed, whereas the predominately green (advanced) color
patterns are characteristic of the populations in the eastern part of

1 All the attempts to name subspecies of snuiragdina (Barbour. 1912 and
1921; de Rooij, 1915; Mertens, 1929) have relied heavily on differences in
color pattern. This method, however, has several shortcomings. First, the
taxonomically important greens and yellows are lost in preservation so that
the remaining structural and melanistic color patterns are very unlike, and
generally not correlated with, the patterns in life, and second, there is often
a good deal more continuous variation, or possibly in some cases, discrete
polymorphism, in single breeding populations than those who ascribe to the
current subspecies concept of Miniraf^cliiui would like to admit (some of this
pattern variation in single breeding populations is described by Taylor.
1922; Hediger, 1934; Marshall. 1951; and Brown and Marshall. 1953).

This is not to say. however, that differences between populations of
sniciniiidina do not exist. There are clear differences in color pattern be-
tween populations, but they are of a statistical nature and not absolute:
also, judging from preliminary observation, there are discernible propor-
tional differences between populations. Whether these differences can be
worked out to the point of establishing a new subspecies concept for
smumgdiiia is problematical, for liic basic data-gathering task wi|l be Her-
culean. In the moHntime, however, and by way of taking a step toward
unraveling this larger problem, there is a real need to understand the
genetic and ecologic aspects o^ ct)lor pattern variation in populations of
smamgdina.

1^70 SKINKS REFERRED TO THE GENUS DASIA 25

the range, i.e., in the Pacific island chains, which arc more likely
to receive colonists than to send them.

1 can add one additional piece of color pattern data to the argu-
ment that the western populations of sinanii>clina arc generally
more primitive than the eastern populations. This is the fact that
in western sinanii^ilina there is often a whitish, "paint brush"
streak extending from the tlank onto the base of the tail which
passes directly over the back leg. This streak is also present in the
Bornean vyneri and iiieuwenhuisi, i.e., the closest living relatives
of primitive western smaiagdina, but it is absent in the smarcii^dina
from the eastern part of the species range, i.e., in those populations
derived from the more primitive western sinarciiidina. Presumably,
therefore, the white tlank stripe of vyneri, nieunenhni.si, and primi-
tive western smaiagdina has been lost in the evolution of the ad-
vanced eastern smaiagdina.

Relationships with other genera. Lamprolepis is most similar to
Mabuya and ditters from all the Old World species of this genus
only in having smooth instead of keeled body scales. The New
World species of Mabuya (except atlantica) also have smooth
body scales but the great geographical distance between the
smooth-scaled Lamprolepis and the smooth-scaled Mabuya argues
for the independent evolution of this condition in both groups.

If one assumes that the most geographically proximate species
of Mabuya are probably most like the ancestors of Lamprolepis,
then the loss of pterygoid teeth has been another feature in the
evolution of Lamprolepis from Mabuya. All southeast Asian
Mabuya have pterygoid teeth, whereas the palatal teeth have been
lost in all the New World species and in about half the African
species, but in only one of the south Asian species (bibroni) exam-
ined to date, it looks, therefore, as if the pterygoid teeth, like
keeled body scales, may have been lost independently in both
Lamprolepis and the advanced species of Mabuya.

it thus seems that both Dasia (sensu stricto) and Lamprolepis
are independently derived from a Mabuya-X\k& ancestor, and they
mi^ht, in fact, have shared the same ancestor. Since this time,
however, the two taxa have diverged to a great extent. Dasia
evolved a more extensive secondary palate (by incorporating the
pterygoid bones), a slower gait, and an unusual juvenile color
pattern but retained, as a primitive feature from its Mabuya-\\kQ
ancestry, pterygoid teeth and keeled body scales. Lamprolepis, on
the other hand, retained a basically Mabuya-\ikQ palate, gait, and
color pattern (except in the greenish eastern populations of

26 BREVIORA No. 348

smaragdina) but lost the pterygoid teeth and keeled body scales
(and later the supranasal scales in smaragdina) of their Mahuya
ancestry (Fig. 6).

Zoogeography. The significance of Wallace's Line to the zoo-
geography of Lainprolepis probably lies in its coincidence with the
eastern edge of the Sunda Shelf, i.e.. the eastern limit of the Asian
land mass that would result with normal Pleistocene lowering of
sea levels. With such lowering of sea levels, Borneo, the island
harboring the two primitive species of Lamprolepis, would be con-
nected to the Asian mainland along with Sumatra and Java, while
the area to the east of this land mass, i.e., the area inhabited by
the advanced species of Lamprolepis, would still be an archipelago,
albeit with an increased average island size.

The dynamics of this distribution might have been somewhat as
follows: The ancestor of vyneri and nieuwenhuisi probably arose
from a primitive Mabiiya-WkQ skink on the land mass of the Sunda
Shelf when this area was a dry extension of the southeast Asian
mainland. This vyneri-nieuwenhuisi ancestor reached the island
archipelago to the east of the Sunda Shelf and here gave rise to
smaragdina, which then spread east with great success but which
could not spread west back onto the dry land regions of the Sunda
Shelf. The inability of smaragdina to spread westward onto the
land of the Sunda Shelf was certainly not the result of an inability
to cross water gaps, as its success in the east proves, but was
probably due instead to an inability to penetrate the complex,
basically mainland continental fauna of the Sunda Shelf land mass.

ACKNOWLEDGMENTS

1 would like to thank first and foremost Mr. James P. Bacon, Jr.,
Dr. Walter C. Brown, Dr. Robert F. Inger, and Mr. Fred Parker
for relating to me their field experiences with some of the animals
discussed in this paper. Dr. Inger also very generously permitted
me to examine two specimens of Lamprolepis vyneri in the Field
Museum of Natural History. In addition, I have benefitted greatly
from discussions with Mr. Ross Kiester about Lamprolepis smar-
agdina in Micronesia.

Dr. Ernest E. Williams of the Museum of Comparative Zoology
has read the manuscript in several drafts and has offered many
useful criticisms.

Mr. Ian T. Riddel took the photograph for Figure 2 and Mr.
Laszlo Meszoly did the drawing for Figures 4 and 5.

1970

SKINKS REFERRED TO THE GENUS DASIA

27

Hi

S

■&|<= ™

rtl c ^

E

o. c
o

X

S/J

c

n

a>

2

o

tr

>>

-**

lU

c

s:

<

h-

c<

c

^

o

w

X

CS

"S

tr- «

-J

— Q.

c
o

'W

c

i/i

■o

a>

(•1

OJ

4/1

OJ

E

J3

o

i/>

1*0

TO

o

OJ

<D

E

"55

5

on

TO

o

O)

o

^

>-N

wo

en

u^

TO

■o

•D

s

TO

o

o

OJ

O

C

5;

5

a.

c

TO

1_

o

.tr

OJ

Q.

tj

TO

QJ

^

■q.

O.

<u

CD

J:m:

^*o

TO

c

QJ

QJ

t_>

■o

LO

O

•

E

OJ

k—

Q.

"S

(/I

(/I

"qj

E

O

-o

03

a>

TO

o

TO

CO

"cD

J3

4-0

QJ

L

fc/1

>N

un

Q>

O

TD

TO

T3

"O

O.

5

s

TO

O

"o

v_
QJ

O

QJ
O

cz
o

UO
TO

1

QJ
QJ

c

TO

^-

3

"q.

"q.

c

>-,

CTi

J^

UO

as

E

"c

«/1

■o

re

s

QJ
O)

D-

X3
TO
<yi
O
O

"o

c:

un

un

x:

TO

TO
O.

a;
on

uo

JZ
-D

o

1

ut

QJ
TO

>->

_QJ
TO

!•!

TO
TO

/^v

o

d

o
en

O

O
(J

en

O

"O

TO

-i

^

U^

>-<

><

CU

c

_aj

ir

c

u

0^

QJ

o

13

."ti

Q)

Q.

o

C

O

TO

QJ

Z)

k.

O-

"q.

c

>^

cn

J^

cn

cL

o

c
<u

o

(4—

c
o

X

>

C
o

x:

c
c

c

_ y^ OJ

i £

o
n.

o

C

c
c

< i= -c

c

c -^

•- c
x: o

o JO

Wi

28 BREVIORA No. 348

Work for this paper was carried out while 1 was first a National
Science Foundation predoctoral and then later a postdoctoral fel-
low. An NSF grant to Dr. Williams (GB 6944) also provided
funds for the completion of the paper.

LITERATURE CITED

Alcala, a. C.

1966. Populations of three tropical lizards on Negros Island, Philip-
pines. Unpublished Ph.D. thesis, Stanford University; 269 pp.

Annandai h, N.

1906. New and interesting lizards in Colombo Museum. Spolia Zey-
lanica 3(11): 189-192.

Barbour, T.

1912. A contribution to the zoogeography of the East Indian Islands.

Mem. Mus. Comp. Zool. 44( 1 ) : 1-203.
1921. Reptiles and amphibians from the British Solomon Islands.

Proc. New England Zool. Club 11: 91-112.

BOUI ENGliR, G. A.

1887. Catalogue of the Lizards in the British Museum (Natural His-
tory). Vol. 3. 2nd ed. London, British Museum: vii -|- 575 pp.

Brongersma, L. D.

1933. Herpetological notes. I. Lygosonui nieiiwenhuisii van Lidth de
Jeude. Zool. Meded. Leiden 16: 1-2.

Brown, W. C, and J. T. Marshall, Jr.

1953. New scincoid lizards from the Marshall Islands, with notes on
their distribution. Copeia 1953(4): 201-207.

Brown, W. C. and A. Y. Reyes

1956. Philippine lizards. Observations on the incubation period and
on hatchlings of several oviparous species. Silliman J. 3(2):
139-143.

Cantor, T.

1847. Catalogue of Reptiles Inhabiting the Malayan Peninsula and
Islands. Calcutta: Baptist Mission Press. 157 pp.

Deraniyagala, p. E. P.

1931. Some Ceylon lizards. .Spolia Zeylanica 16(2): 139-180.
1953. A Colored Atlas of Some Vertebrates from Ceylon. Volume

Two. Tetrapod Reptilia. Ceylon Nat'l. Mus. Publ.. Ceylon

Gov't. Press; vii + 101 pp.

1970 SKINKS REFF.RRFD TO THE GENUS DASIA 29

Cjandoih. H.

1907. Ein sekundiirer Geschlectsiinterschied bei Lyf>osoma smarag-
iliniim (Less). Zool. Anzeig. 32(7): 186-188.

Grkkr, a. H.

1967a. A new generic arrangement for some Australian scincid lizards..
Breviora 267: 1-19.

1967b. The generic relationships of the African scincid genus Eumecia.
Breviora 276: 1-9.

1970. A subfamilial classification of scincid lizards. Bull. Mus. Comp.
Zool. 139(3): 151-183.

Greer, A. E., and F. Parker

1968. Geomyersia glabra, a new genus and species of scincid lizard
from Bougainville, Solomon Islands, with comments on the
relationships of some lygosomine genera. Breviora 302: 1-17.

Hediger, H.

1934. Beitrag zur Herpetologie und Zoogeographie Neu Britanniens
und einiger umliegender Gebiete. Zool. Jahrb. (Syst.) 65(5-6):
441-582.

Hendrickson, J. R.

1966. Observations on the fauna of Palau Tioman and Palau Tulai.
5. The reptiles. Bull. Nat'l. Mus. Rep. Singapore 34: 53-71.

HOFFSTETTER, R.

1949. Les reptiles subfossiles de Tile Maurice. T.-Tes Scincidae.
Ann. Paleontologie 35: 45-72.

Marshall, J. T., Jr.

1951. Vertebrate ecology of Arno Atoll, Marshall Islands. Atoll Res.
Bull. No. 3: ii + 38 pp.

Mertens, R.

1929. Die Rassen des Smaragdskinkes, Dasia siuaragciinum Lesson.
Zool. Anzeig. 84(9-10): 209-220.

1930. Die Amphibien und Reptilien der Inseln Bali. Lombok. Sum-
bawa und Flores. Abh. senckenberg. naturf. Ges. 42(3): 115-
344.

Mittleman, M. B.

1952. A generic synopsis of the lizards of the subfamily Lygo-
sominae. Smithsonian Misc. Coll.. 117(17): 1-35.

de Room, N.

1915. The Reptiles of the Indo-Australian Archipelago. 1. Lacertilia,
Chelonia. Emydosauria. Leiden: E. J. Brill. Ltd.. xiv -p 384 pp.

30 BREVIORA No. 348

Smith, M. A.

1931. The herpetology of Mt. Kinabalu, North Borneo. 13.455 ft.
Bull. Raffles Miis. 5: 3-32.

1935. The Fauna of British India. Including Ceylon and Burma. Rep-
tilia and Amphibia. Vol. II — .Sauria. London: Taylor and
Francis, xiii + 440 pp.

1937. A review of the genus Lygosonni (Scincidae: Reptilia) and its
allies. Rec. Indian Mus.. 39(3): 213-234.

1949. Notes on a second specimen of the skink Dasia xuhcaeniica
from southern India. J. Bombay Nat. Hist. Soc. 48(3): 596-
597.

Tayi OR, E. H.

1915. New species of Philippine lizards. Philippine J. Sci., Sec. D.
10(2): 89-109.

1922. The Lizards of the Philippine Islands. Gov't. Philippine Islands,
Dept. Agricult. Nat. Res. Bureau of Sci. Publ. No. 17: 1-269.

1963. The lizards of Thailand. Univ. Kansas .Sci. Bull. 44(14): 687-

1077.

Taylor. E. H.. and R. E. Elbel

1958. Contribution to the herpetology of Thailand. Univ. Kansas
Sci. Bull. 38(13): 1033-1189.

(Received I 1 December 1969.)

) ' Ni A- cL Tao. bv: i<;i^^bo>iot ^>^u late

MUS. COMP. ZOOL.
LIBRARY

B R E V I O R- A

HARVARD

Mmseiuiim of Comparative iJ^oolo^y

Cambridge. Mass. 15 May 1970 Number 349

The Structural Niche of Anolis scriptus on Inagua
Anthony L. Laska ^

Abstract. The only member of its genus on Inagua Island, Southern
Bahamas, Anolis scriptus leucophaeus, is believed to be derived from A.
cristatellus of Puerto Rico, which shares the latter island with 10 congeners.
As has been demonstrated in the case of other "solitary" anoles on other
islands, A. scriptus has a broader niche than its ancestor on a multi-species
island.

INTRODUCTION

Differences in structural niche among sympatric species of
Anolis in the West Indies have been described recently by several
authors (Collette, 1961; Ruibal, 1961; Rand, 1962, 1964; Scho-
ener, 1968). The significance of intraspecific sexual dimorphism
and its ecological relationships has been discussed by Rand ( 1967),
Schoener (1967), and Schoener and Gorman (1968). Rand and
Rand (1967) suggest that the lone species of anole on Curasao,
Anolis lineatus, occupies a somewhat broader structural niche
than most species of the Greater Antilles. Similar field observa-
tions were made of Anolis scriptus, which occurs with no congeners
on Inagua. Results of a six-day stay in February of 1967 suggest
a fairly broad niche for this species, as for lineatus.

Anolis scriptus leucophaeus is a medium-sized anole. The mean
of the largest third of all male specimens examined was 67.6 mm
(n = 36) in snout-vent length, while corresponding females aver-
aged 46.3 mm (n = 70). The closest relative of this species is
A. cristatellus cristatellus of Puerto Rico (Gorman et al., 1968).

Great Inagua is located about 110 km north of Haiti, approxi-
mately 100 km from easternmost Cuba. About 70 by 30 km, it
is predominantly flat, the highest point being East Hill at 40 meters.
Lower hills dot the eastern half of Inagua's coasdine. A large,

' Department of Biology, Tulane University, New Orleans, La. 70118

2 BREVIORA No. 349

shallow, saline lake fills a substantial area in the western interior
of the island, and brackish mud flats are common. Inagua receives
little rain, a yearly mean of 612 mm; and the annual temperature
range is 22 to 30°C.^ Cactus, thorn scrub, and low bushes are
common in the dry flatlands, while stands of buttonwood and larger
trees are found on low-lying ground. Sizeable palmetto groves
occur on low aeolian hills along the shores in the regions studied
to the northwest and south.

The low profile of Inagua indicates that this island is relatively
young. This young age, combined with the movement of currents
toward the northwest and the presence of submerged island banks
between Puerto Rico and the southern Bahamas, accounts for the
close relationship between the Anolis faunas of Puerto Rico and
Inagua (Gorman et al., 1968).

FIELD METHODS

For this study anoles were observed and collected in western
and southern parts of Great Inagua and on Sheep Cay, off North-
west Point. They were found most frequently in palmetto groves
and stands of large trees but were rarely seen in thorn scrub, al-
though several sub-adults were observed in dense patches of cactus
and adder's tongue in Matthew Town. No apparent intraspecific
variation was found among scriptiis collected on Inagua proper
or when these were compared with specimens taken from Sheep
Cay.

To determine perch preference of this species, perch height
and diameter were recorded for anoles observed at five principal
localities: Northwest Point, Sheep Cay, Matthew Town, and neigh-
boring areas of Salt Pond Hill and Conch Shell Hill. A total of
168 individual observations was made with an effort to avoid
repetition. All but 23 entries were seen on trunks of palmetto
trees and trunks and branches of larger trees. Those exceptions
were in clumps of cactus, on stumps, stone walls, fences, and
major trunks in brush piles. The exceptions are not included in
calculations.

RESULTS AND DISCUSSION

Most of the lizards recorded were perched above ground level
and below 12 feet. The data in Table 1 suggest this species dis-
tributes itself vertically as a function of sex and size of individuals.

1 These are means of the extreme temperatures between 1954 and 1968.

1970 ANOLIS SCRIPTUS ON INAGUA 3

With four zones available (following Rand, 1964), 59 percent
of adult males were observed between 6 and 10 feet, 68 percent
of adult females perched between 3 and 5 feet, while 74 percent
of juveniles were seen within 3 feet of the ground. Several juveniles
were found on the ground or in leaf litter. Results of X'^ tests
show that adult males were found significantly higher (34>5';
22<5') than aduh females (4>5'; 33<5'; P<.00l). Adult fe-
males occurred significantly higher (29 > 3'; 8<3') than juveniles
(5>3'; 14<3'; P<.001).

Table 2 compares A nolis script us with A . lineatus of Cura9ao
(Rand and Rand, 1967), both of which occur with no congeners.
Rand's (1964) figures for the structural niche of A. cristatellus
cristatellus are included for comparison between A. scriptiis and
its nearest relative. The vertical distribution of scriptus is nearly
as broad as that of lineatus but its preferred perch diameter ap-
pears more restricted. Perch diameter limitation for scriptus is
probably a reflection of the species' association with palmetto
trees, which are prevalent on much of the island. Since the
typical palmetto trunk measured 3 inches, a high number of entries
for scriptus is found in the 1/2-3 inch column in Table 2. Uni-
formity in available perch diameters is reflected by a lack of sta-
tistical significance for comparisons of preferred perch diameter
for scriptus males, females, and juveniles in Table 1 .

Total height of the perch was incidentally recorded for 41 of
the observations and the ratio of perch position to total tree height
was calculated. Means for these ratios are given in Table 3 for
males, females, and juveniles. Since most of the trees in the
localities did not exceed 12 feet, those over 12 feet were eliminated
in calculating the second series of figures; the ratios b^ing based
on 22 remaining individuals.

When tree heights of 12 feet and less are considered, the posi-
tion of lizards relative to the tree crown is higher than when taller
trees are included, as indicated by a substantially greater ratio
for the former. Although an optimum perch height relative to the
ground is suggested by these data, the inherent bias of the ob-
server's position may have some influence.

Of the juveniles observed, 74 percent were less than three feet
above ground level, away from most of the larger adults that might
try to catch and eat them. Since these juveniles are not competing
for food with many larger anoles, they have a better chance of
developing faster in this niche (Rand, 1967).

4 BREVIORA No. 349

Scriptus, as a solitary anole, occupies a broader structural niche
than typical species occurring with congeners. But its niche is not
nearly so broad as that filled by the totality of a mixed Anolis
fauna (Rand, 1964).

ACKNOWLEDGMENTS

I thank E. E. WilHams for several critical readings of the manu-
script, numerous useful comments and continuing interest. I am
also grateful to A. S. Rand, and G. C. Gorman and T. W. Schoener
for critically reading the manuscript and making valuable sugges-
tions. Thomas Schoener supplied measurement data for A.
scriptus. Charles Bremer, of Morton Bahamas Limited, and J. R.
Clackson, Chief Meteorologist of the Bahamas Department of
Civil Aviation, provided climatic information for Inagua. 1 thank
M. Hecht for helpful comments and Shirley Laska for assistance
in the field and numerous discussions. Field work was supported
by Grant GB-6944 to E. E. Williams.

LITERATURE CITED

CoLLETTE, B. B. 1961. Correlations between ecology and morphology
in anoline lizards from Havana, Cuba, and southern Florida. Bull.
Mus. Comp Zool.. 125: 137-162.

Gorman, G. C, R. Thomas, and L. Atkins. 1968. Intra- and interspe-
cific chromosome variation in the lizard Anolis cristatellns and its
closest relatives. Breviora, No. 293: 1-13.

Rand, A. S. 1962. Notes on Hispaniolan herpetology. 5. The natural
history of three sympatric species of Anolis. Breviora, No. 154: 1-15.

. 1964. Ecological distribution in anoline lizards of Puerto

Rico. 'Ecology, 45: 745-752.

. 1967. Ecology and social organization in the iguanid lizard.

Anolis lineatopus. Proc. U.S. Nat. Mus., 122 (3595): 1-79.
Rand. A. S.. and P. J. Rand. 1967. Field notes on Anolis lineatus in Cu-

ragao. Studies on the Faunas of Curasao and other Caribbean Islands,

24: 112-117.
RuiBAL, R. 1961. Thermal relations of five species of tropical lizards.

Evolution, 15: 98-111.
Schoener, T. W. 1967. The ecological significance of sexual dimorphism

in size in the lizard Anolis conspersiis. Science, 155: (3761): 474-477.
. 1968. The Anolis lizards of Bimini: resource partitioning in

a complex fauna. Ecology, 49: 704-726.
ScHOENiiR, T. W., AND G. C. GoRMAN. 1968. Some niche differences in

three Lesser Antillean lizards of the genus Anolis. Ecology, 49:

819-830.

(Received 15 March 1970.)

1970

ANOLIS SCRIPTUS ON INAGUA

0

•v
o

o

O rt

c

k> Vi-I

t>

oi

X)

E

z

e2

■*

V£)

o

>i-i

>o

■<1-

VO

m

*

^

^

__

r)

Tf

*

5

V

o

m

rJ

rJ

M

VO

(N

i-

^

<-n

'^

rl

O

m

<N

m

r-

(^

ON

A

<^

■«-•

m

r<^

^H

<N

(r\

*

H

'"^

-t

rn

00

13

^

1— t

«S

m

"rt

3

V

E

[>

pn

>— 1

o

o

-*

«i-i

C/5

^

r»-)

(N

"O

m

<s

<N

<— 1

VO

*— (

A

<N

O

^-H

•^

Tf

OS

^H

H

^

»-H

•-H

'c

V

>

r*^

^

rn

o

Tj-

3

1

1-H

^H

1— >

A

^— 1

rr,

Tf

-i-i

1-H

m

«/->

00

t^

f2

<S

cn

^

T-H

^H

3

V

■o

<

A

(N

00

t-~

as

■s

1-^

ro

o\

rri

vo

,o

f»^

«r>

H

*-t

C/)

:^

t-H

rt

<

6

V

1

-^

<s

(N

:^

m

^

00

A

'"'

^

x

(U

o

S

,.^

b

b

>o

r<^

5

Oh

■•3

to
c

A

1

ro

V

o
H

•^^

'^

JZ

fc!

.£P

^

'53

CL,

.c

I,

a

^

■»->

a

4>

^

4l<

3

O

c

t«

c

• 1-^

CO

U

J3

U

c

o

s

♦J

Ui

V

"O

c

3

«i

«J

3

"O

>

■a

.S

•o

u

CB

■4-*

B

a

rr

T1

a

c

3

3

T3

C/5

>

n

T3

a.

C

CAI

1)

u

t)

c

(U

«

>

3

E

t/i

(/I

3

3

a

a

c«

r/l

u

<1>

C3

«

E

E

4J

<*j

4-^

"-i

3

-o

<

<

*

*

«

BREVIORA

No. 349

Table 2. A comparison among Anolis scriptiis of Inagua, A. Uneutiis
of Curasao, and A. c. cristatelliis of Puerto Rico.

Percent of Number of Each Species Recorded

Inagua

scriptus

Curasao
lineatiis

Puerto Rico
cristatelliis

Perch height

Perch diameter

<3 ft.

3-5 ft.

6-10 ft.

>10 ft.

>3 in.

5^-3 in.

<K in.

24%

41%

32%

3%

27%

70%

3%

32%

33%

27%

7%

67%

31%

2%

47%

40%

11%

1%

65%

27%

7%

Table 3. Ratio of perch height to tree height for Anolis scriptus.
Tree height Adult males Adult females Juveniles

6 ft. - 26 ft.

.42

.37

.14

6 ft. - 12 ft.

.67

.41

.20

MUS. COMP. ZOOL.
OO MOT CIRCULATE ^"^'^'^''^

B R E V I 0"W%

I I r-^ I \ V r-\ ( \ L_

UNIVERSITY

Mmseimim of Comparative Zoology

Cambridge. Mass. 15 May 1970 Number 350

FIMBRIA AND ITS LUCINOID AFFINITIES (MOLLUSCA;

BIVALVIA)

Kenneth Jay Boss

Abstract. Fimbria is the only surviving genus of the Fimbriidae, a
family of bivalve mollusks that flourished during the Mesozoic. The con-
chological and anatomical features of Fimbria show that the family be-
longs to the superfamily Lucinacea and that it is closely allied to the
Lucinidae. The systematic position of the Fimbriidae in relation to other
lucinoid families is discussed, and a possible phylogeny for the superfamily
is re-constructed.

INTRODUCTION

The relationship of the Fimbriidae with other lucinoid bivalves
has been questioned by McAlester (1966), who suggested that
additional anatomical and morphological data would be necessary
before certain families, the Fimbriidae included, could be unequiv-
ocally assigned to the Lucinacea. It has been usual to include
Fimbria and its fossil relatives in the Lucinacea (Dall, 1895;
Thiele, 1935; Franc in Grasse, 1960; Eberzin in Orlov, 1960;
Newell, 1965; Moore, 1969). Dall in Zittel (1900) considered
the Fimbriidae an offshoot of the Lucinidae. The reason for this
assignation was conchological, based on the structure of the dental
and ligamental elements of the hinge of Fimbria. Anatomical
characteristics supplementing the features of the shell allow the
definite inclusion of the Fimbriidae in the Lucinacea.

The species of Recent Fimbria were reviewed by Lamy (1921)
and Nicol (1950), and according to the latter there are only two
living species, both distributed in the Indo-Pacific area. During
the Mesozoic, particularly the Jurassic and Cretaceous, the genus
Fimbria was rich in number of species and was widely distributed
throughout warm and temperate seas (Pictet, 1855; Stoliczka,
1871). Both species-number and distribution dwindled during
the Tertiary.

BREVIORA

No. 350

THE ANATOMY OF FIMBRIA FIMBRIA! A

Mantle. The mantle is thin, translucent over much of its sur-
face and open ventrally in an extensive pedal gape. Its edge is
fringed with two rows of tentacles. It is fused anteriorly above
the anterodorsal surface of the anterior adductor muscle (Fig. 1 )
and posteriorly, ventral to the incurrent aperture (Fig. 2). The
fusion of the mantle lobes anteriorly is much like that in Myrtea
spinifera (Montagu) as figured by Allen (1958, fig. 7c) and pro-
vides a broad anterior incurrent pathway by which water and sus-
pended matter are drawn into the mantle cavity.

Muscles. The adductor muscles are large and are differentiated
into 'quick' and 'catch' fibers. The posterior adductor muscle is
irregularly oblong and rounded, whereas the anterior adductor is
larger and subsemilunate, with the ventral portion wider or more
extensive.

Siphons. Incurrent and excurrent apertures are developed pos-
teriorly (Fig. 2). The incurrent opening consists of a simple
rounded aperture, laterally flanked by two ridges of tentacles; ven-
trally, iJie muscular lobes of the mantle are fused to form a narrow

Figure 1. Diagrammatic view of anterior portion of Fimbria. I. Point
of anterior mantle fusion. 2. Anterior pedal retractor muscle. 3. Ante-
rior adductor muscle. 4. Inner mantle fold. 5. Outer mantle fold.

970

THE LUCINOII) FIMBRIA

isthmus, and dorsally they are joined in a broad union, which sepa-
rates the excurrent siphon. The excurrent siphon consists of a
hirge muscuhir tube which is capable of b^ing invaginated and
Hes, when turned inside out, in the suprabranchial chamber. The
siphonal retractor muscles are not developed, and the orifice of

Figure 2. Diagrammatic view of posterior portion of Fimbria. 1. Point
of fusion of outer mantle fold. 2. Secondary papillate fold. 3. Extent
of tissue grade mantle fusion. 4. Inner mantle fold. 5. Posterior adduc-
tor muscle. 6. Excurrent invaginable siphon. 7. Incurrent aperture.
8. Outer mantle fold.

4 BREVIORA No. 350

the aperture is flanked by two rows of papillae. The posterior
aspect of Fimbria (Fig. 2) is similar to that of Codakia orbicu-
laris (Linnaeus) as is illustrated by Allen (1958, fig. 9b).

Foot. The foot in Lucina has been studied by Barrois (1885).
In Fimbria, it is large, subtrigonal, somewhat laterally compressed,
longitudinally grooved or creased, heeled posteriorly, and pointed
anteriorly (Fig. 3). In addition to its strong intrinsic longitudinal
and circular muscles which surround a central sinusoidal lumen,
the foot has a pair of anterior pedal retractors that attach to the
shell just posterodorsal to the anterior adductor muscle and a
pair of posterior pedal retractors that insert just anterodorsal to
the posterior adductor muscle.

Gills and palps. The true gills or ctenidia consist of single,
bilateral, inner demibranchs that are fused to the visceral mass
dorsally and to each other posteriorly. They also attach to the
muscular portion of the edge of the mantle between the ventral
incurrent and dorsal excurrent apertures, thus effectively forming
a septum that creates the dorsal suprabranchial chamber, into
which empty the reproductive, excretory and faecal products.

Although longitudinally wrinkled, the gills are non-plicate,
smooth or flat and apparently homorhabdic. They are thus similar
to other luncinoids (Ridewood, 1903). In the specimens
dissected, the demibranchs were retracted and thin, not thick and
turgescent as are the gills in Lucina (Mencgaux, 1889). The
filaments are interconnected, and the inner ascending lamella is
shorter than the outer descending lamella, so that the latter ap-
pears to be a supra-axial extension of the demibranch. Unlike
certain other lucinoids (Read, 1962; Hartman and Boss. 1966),
the gills are not distinctively colored by hemoglobin.

As in other lucinoids (Thiele, 1886), the labial palps of Fimbria
are very much reduced and consist of narrow flaplike lips border-
ing the mouth that accept, at their distal ends, the anteriormost
portion of the demibranchs (Fig. 4). The reduction of labial
palps is probably related to the increased efliciency of the ciliary
mechanisms of the anterior incurrent canal and permits the in-
gestion of relatively large food particles (Allen, 1958).

The relation between the labial palps and ctenidia has been
studied by Stasek (1963). Lucinoids, Fimbria included, exhibit
a pattern, the category III of Stasek, in which the ventral tips of
the anterior filaments of the inner demibranch are not inserted
into a distal oral groove. Further, the inner demibranch may be
fused to the inner palp lamella. Although the demibranch inserts

1970 THE LUCINOID FIMBRIA 5

between the palp lamellae in Fimbria, there is no fusit')n (Fig. 4).

Mantle i^'ills. The reduction of the true gills to only a single
demibranch and the existence of the anterior incurrent canal
probably were preadaptive anatomical features that aided in the
development of another unusual character in lucinoids, namely,
the mantle gill, noted by Duvernoy ( 1854: 1 15. pi. 5. f. 3) and
Semper (1880).

Pelseneer (1911) discerned three different types of mantle
gills and these have been discussed by Allen (1958). Fimbria is
somewhat unusual in the development of two large flaps on the
posteroventral surface of the anterior adductor muscle (Fig. 3).
A large pallial vessel courses through the mantle diagonally from
the mantle gill to the auricles. The tiaplike structures of the
mantle gills lie on either side of the channel of the anterior in-
current canal on the ventral surface of the anterior adductor
muscle.

Intestine. The morphology of the alimentary canal was not
studied in detail. A short esophagus enters the rather large
stomach anteroventrally and the combined style-sac and midgut
leaves it posteroventrally; the stomach is of the Type IV category
of Purchon (1958). The digestive glands surround the stomach
and are not developed into external visceral pouches. The in-
testine forms a sunple loop in the visceral mass and enters the
pericardium slightly to the right of the midline; the rectum passes
over the dorsal surface of the posterior adductor muscle and emp-
ties into the suprabranchial chamber near the nephroprocts.

Heart and kidney. The complex of the heart and kidney is
located posterodorsally and does not vary significantly from the
typical eulamellibranch pattern (Menegaux, 1890;Odhner, 1912).
The heart is simple and lies in an expansive, somewhat triangular
pericardium; two extremely large, thin-walled auricles receive
blood from the large palUal vessel and the ctenidia. The ventricle
is thick-walled and does not completely encircle the rectum; a
dorsal aorta extends anteriorly. Pericardial glands (White, 1942)
were not detected.

The kidney lies between the heart and the posterior adductor
muscle and consists of two bilateral portions. The anteroventral
reno-pericardial funnels collect fluid from the pericardial cavity;
the posterior saclike parts of the kidney are broadly connected
medially. The posterior portion of the kidney encircles the pos-
terior retractor muscles. The kidneys open via bilateral nephro-
procts to the suprabranchial chamber posterior to the gonadial
aperture near the excurrent invaginable siphon.

BREVIORA

No. 350

<U 4^ ■- O

•= o 2 ~

c o u c

E . b S

1^ "U

j: :r ^

•r- O) Q. ^

."2 =

n

11

J=

lm

<u

11

yi

r

O

-«.->

CL

O

dJ —

•a
<u

M

■- «

J^ •-

a. o

d c
!? <

>> 3

E S
o

■Si '-

c 2

C« U

03
W J:;

■" k.
■*- _-

o «
o o.

1.9

■3; 1-

3 w
== c

•- 5

.y E

C
03

>

03 _
•- 03

03

a

c o

> .-

.n 0-

c

I- <u

(U —

> tj

03 S2

03

r4

Q S
U

3 o

■- -3

:« 1-

o

c

C 5

CXI

CO

LTD

u

U. >

o

E

03

03
O

SO C u

970

THE LUCINOID FIMBRIA

Nervous system. The general pattern of the nervous system
of Fimbria is not significantly different from that of other lucinoids
or eulamellibranchs (Duvernoy, 1854). There are three main
ganglionic centers: cerebropleural, pedal, and visceral. The cere-
bropleural gangUa are located on the inner surface of the anterior
adductor muscle, just over the labial palps on either side of the
mouth; they are united via a medial cerebropleural commissure
(Fig. 4). Anteriorly, the ganglia give rise to nerves which supply

Figure 4. Diagrammatic view of the buccal area. 1. Outer labial palp.
2. Inner labial palp. 3. Ctenidium. 4. Mouth. 5. Cerebropleural gan-
glion with cerebropleural commissure. 6. Esophagus. 7. Cerebropleural-
visceral connective.

the anterior adductor muscle and the anterior pallial musculature.
Anterolaterally are branches to the labial palps, while extending
ventrally from the ganglia are the cerebropleural-pedal connectives
and posteriorly the cerebropleural-visceral connectives. Both
pedal and visceral ganglia are closely juxtaposed and fused; the
pedal ganglion gives rise to branches that innervate the intrinsic
pedal musculature as well as portions of the viscera. The visceral
ganglion is located on the inner surface of the posterior adductor
muscle. Rather large branchial nerves arise laterally from the
ganglion. The posterodorsal surface of the ganglion gives rise to
small, short renal nerves, whereas the large posterior pallial nerve
splays out over the ventral surface of the adductor muscle.

Specimens examined. (All Fimbria fimbriata (Linnaeus
1758)): Ifaluk Atoll, CaroHne ids. (USNM 616497); Bikini

8 BREVIORA No. 350

Lagoon, Bikini Atoll, Marshall Ids. (USNM 58471 1 ); NE side of
Ngaloa Ids.. Fiji Ids. (USNM 674578); E of Mathuvata, N coast
of Vanua Levu, Fiji Ids. (USNM 674577); Yanutha Id., N Coast
of Viti Levu. Fiji Ids. (USNM 674576).

DISCUSSION

The unusual anatomical features in the Lucinacea were noted
long ago by the famous naturalist Poli (1791), who figured the
Mediterranean Tellina [Lucina] lactea with its elongate, cylin-
drical foot and single demibranch. Some of the anatomical
characters of Fimbria were described by Valenciennes (1845a and
b), who noted the reduction of the ctenidia to single demibranchs
and the smallness of the labial palps. He showed that the foot
of Fimbria differed in shape from that of other lucinoids. Thiele
( 1935 ), following Dall, remarked that in Fimbria, the mantle was
fringed by two rows of papillae and the excurrent siphon was
retractile. Allen ( 1958) studied a number of species of lucinoids
and distinguished the Ungulinidae [Diplodontidael, Thyasiridac.
and Lucinidae anatomically.

Table 1 contrasts the living families of the Lucinacea. The
Fimbriidae are most closely allied to the Lucinidae. Both families
have in common: 1 ) an excurrent siphon consisting of a unique
invaginable tube; 2) an anterior point of fusion of the mantle
edges that is dorsal to the anterior adductor muscle; 3) the ctenidia,
consisting of a single inner demibranch on each side of the animal;
4) greatly reduced palps; 5) a tissue-grade fusion of the muscular
lobes of the mantle ventral to the posterior incurrent aperture.
Some species of the Lucinidae also develop accessory respiratory
organs or mantle gills and have a foot which is differentiated
posteriorly into a heel for digging. Fimbriids differ from lucinids
in the development of a bilaterally compressed subtrigonal and
pointed foot, thin demibranchs without large amounts of subtil-
amentar tissue, a double row of papillae along the mantle margin,
and the confluence of the ventral surface of the anterior adductor
muscle with the pallial musculature.

In addition, the shell of Fimbria is strong and heavy with thick-
ened dentition, elongately-ovate in shape and cancellately sculp-
tured (Fig. 5). The scar of the anterior adductor muscle is con-
tiguous with the pallial line, a configuration unlike that of lucinids,
in which the anterior adductor scar protrudes inwardly and is
separated from the pallial line.

1970

THE LUCINOID FIMBRIA

Figure 5. External view of left valve of Finihrici ftnihriatn. length =z
30.5 mm. from Yaniitha id.. N Coast of Viti Levii (USNM 674576).

Although the ecology of living Fimbria was not studied in the
field, it is known that the species are reef-dwellers, preferring rela-
tively shallow water in coralline sand (Nicol, 1950). The heavy,
strongly ribbed shell is an adaptation to the coarse substrate which
the species inhabits. Further, Fimbria probably lives near the
sand-sea interface, for it possesses the anterior-posterior water-
flow mechanisms of other lucinoids but lacks the vermiform foot
to construct a long anterior incurrent mucus tube.

In addition to the family-level characteristics, Fitubria may be
included in the Lucinacea because of the following anatomical
features: 1) an extensive pedal gape; 2) the enlargement of the
ventral surface of the anterior adductor muscle to facilitate an
anterior inhalant current; and, 3) the relatively poorly differen-
tiated and simple posterior incurrent aperture.

10 BREVIORA No. 350

i^^-

(U

O

o

^

:3

E

t^-

l-m

u.

c/:

o

<u

^

•o

03

T3

■a
"5

i;

CO

5

'^

c

—1

CI;

_o

c

'5

^

D

u—

c

CO

n

<u

u

aj

w

PS

u

CJ

1-

_c

3

CO

O

<*-

■4— •

"o

u

u.

o

3

lU

3

7

hJ

CO

t«

T3

<u

J .

CO

x:

O

c

<u

CO

"C

1)

.a

'S5
CO

C

_o

3

c

CO

B

f-

O

c3

4~<

vt^

til

x:

_C

«

'C

lU

CJ

>

c

'~

OJ

c

k>

XJ

CO

_o

3

c
'35

>
O

E

«*-(

3

O

u^

(U

c

(/5

_u

2

O

u

^^

p-

—

o

u«

c

.—

u—

2

3

E

O

c

•a

M-l

o

^^

u

c

_o

'e

OJ

'C

o

X

CC

OJ

c

C

II

CO

—

«

Un

u
D

t/5

u

UJ

>

E

B

OJ

o

o

."2

c

o

<4-l

Xi

e

"a:

o

o

«^H

3

c
o

ii^

O
C

CO

^ d)

u u u

-J CO 3

60 •- .3

oj i;: T3 z:

3 CO C

u.

<u

D.

.>.<

c

.^

o

C

ii>

D.

to

a

'E.

CO

c

a.

CO

1/)

tu

•a 3 CO

UJ

c

E -

2i

_C0

•T3

3

"a,

OJ

3

aj

<u

cO

cr

c/5

j=

a

Q.

C

.^

CO

1-

O

O

c

.^^

^

O
C

c

o

c

o

zz

^

s:

(U

CO

JZ

r-

^D.

L.

C

C

u

ni

'

U.

U.

'i7)

3
CJ

1)

y:

:s

o

V*—

O

c

c

^

O

c

O

wo
C

UJ

T3
CO

E

^-^

^-v

• "-

'"^

OJ

^^^^

L.

O

^~

ri

x

r<-.

1/5

^

C/J

>

^.^

(U

(U

3

o

u

•O

3

<U

CO

Uh

X

u

CO

C/J

<u

QJ

_cu

Q.

CO

n.

3

c/l

CO

en

C

CO

E

'EJj

CO
>

o
n.

<D

Urn

u
3

CJ

•a

■a
c

3

O

E

c

C

t/3

3

UJ

<U

>

C

3

X)

^

OJ

^■^

o

^"^

""^

3

—

u.

ri

i-

r<-i

^

c«

o

•a
c

CO

« x:

■^ ^ CJ
o u -ii

XJ "3 CO

XJ

o

c

CO -a

^ ' (U

X3 O

"c -^ 3

CO T3

CJ '3 OJ

T3 -5 1-

bl

^

"3

CO

c

.5

'u,

X)

3

y:

<u

CO

CO

OJ

C

o

E

c

1

CO

E

C/5

"3

E

•o

o

JZ

•«-•

yi

o

X

CJ

._-

<J-

CJ

J=

^

c

o

o

3

, f

CJ

o

D.

£

>

OJ

c

o

■a
8

CJ

o

E

3
T3

"cj
a.

•a

^

O

CO

OJ

E

a.

u
e/j

CO
C

CJ ?J -:;• ::: o

•a
CJ u- • • JZ

CJ >> CJ

CJ r: c
x: .E o

Ld CJ

3 t ^

g 3 TD

X r 2

OJ g_ t^J -^

"pr 3 j:: CO TJ

-'^ ♦^ ^ <- CJ

CO^CWy)-Xl U

c

-o

O

_CJ

JZ

CJ
CJ

(J

c

J=

CO

k.

,

Xi

CO

c

E

o

CJ

t/J

T3

L.

1-

■^-

CJ

x>

c

3

c

C ^ jn -♦— ' •-- ■-" •— r5

•EoS^co- 2 c ^

^-^-^S u. § l-t I ..

> •& - ^ g- S .H^ ^ ■.£ ^ g

^C3 _X>c^3ii

^-■i-rjum-^ yi .tSC

y)

'C

CO

a
S

*»^ *- y) 3

C CJ CJ y

._ 3 1- -r" t. •— CJ

-^ 2 .3 CO g CO o j= .^ Wj u

H < a. u. O £ i5

1970 THE LUCINOID FIMBRIA 11

The Finibriidae are more distantly related to other lucinoid fam-
ilies. The Thyasiridae are unique in having a highly difTerentiated
vermiform foot, external pouches for the digestive diverticula, both
outer and inner demibranchs, a posterior incurrent slit without
differentiated papillae, and a nontissue grade connection of the
muscular folds of the mantle beneath the incurrent slit (Allen,
1958; Nakazima, 1958; Kauffman, 1967). The Ungulinidae
I =Diplodontidae | possess posterior sensory papillae, both inner
and outer demibranchs (Duvernoy, 1842), moderately sized labial
palps, and a fusion of the mantle musculature over one-half the
surface of the anterior adductor muscle.

Working with the Thyasiridae, Ungulinidae | =Diplodontidae],
and Lucinidae, Allen (1958) suggested that these families vv-ere
related in an evolutionary series, starting with the Ungulinidae
I =:Diplodontidae), which were considered the most 'primitive"
family, and progressing through the Thyasiridae to the more 'spe-
cialized' Lucinidae. McAlester (1966) postulated that the rela-
tionship was actually reversed and that the Ungulinidae | =Dip-
lodontidae I and Thyasiridae were convergent with 'typical heter-
odont' bivalves. The single important heterodont feature referred
to by McAlester is the presence of both inner and outer demi-
branches in thyasirids and ungulinids; however, it is now known
that this feature may not be as important as previously thought,
since different species in the same genus (e.g., Astarte) may either
have a full complement of demibranchs or have only a single demi-
branch. Nevertheless, if the Paleozoic stock of lucinoids had only
a single demibranch, then the Ungulinidae [=rDiplodontidae] and
Thyasiridae, which branched from them in the late Paleozoic or
early Mesozoic, had to re-acquire this morphological feature. Thus,
McAlestefs interpretation of the phylogenetic relationship among
these lucinoid families is essentially correct. However, the geo-
logical record itself provides a better clue to phylogeny than do
inferences based on supposed anatomical features of fossil lineages.

The Lucinidae and Fimbriidae are geologically older than the
Thyasiridae and Ungulinidae | =:DipIodontidae| (Moore, 1969)
(Fig. 6). The lucinids may date from the Silurian; Ilionia Billings,
a Middle to Upper Silurian genus from Sweden and eastern Canada
is the first known member of the Lucinidae. However, if Bahinka
Barrande is considered a lucinoid (McAlester, 1965), the super-
family is traceable to the Middle Ordovician. Fimbria von Miihl-
feld appeared in the Bathonian of the Middle Jurassic and radiated
in the Cretaceous while other fmibriids were present earlier in the

12

BREVIORA

No. 350

1970 THE LUCINOID FIMFiRIA 13

Figure 6. Proposed evolutionary relationships of four living families
of the Lucinacea. Question marks indicate possible derivations of families.
Solid lines denote known lineages: the Lucinidae, from the Silurian (Ilionia
Billings); the Fimbriidae, from the Upper Triassic (Sclwfhaeiitlia Coss-
mann); the Thyasiridae, from the Middle Triassic (Storthodon Giebel),
and the Ungulinidae from the Upper Cretaceous i Brachy/neris Conrad).

The diagrams indicate anatomical features: in the fimbriid-lucinid line-
age, invaginable excurrent siphon, rounded incurrent aperture and inner
demibranch only; in the thyasirid-ungulinid lineage, no excurrent invagi-
nable siphon; slitlike incurrent aperture and both outer and inner demi-
branchs.

14 BREVIORA No. 350

Upper Triassic {Schajluieutlia Cossmann) and even in the Carbon-
iferous {Scaldia de Ryckholt) (Eberzin in Orlov, 1960; Moore,
1969).'

The Ungulinidae | =Diplodontidae| probably arose in the Up-
per Cretaceous {Brachymeris Conrad and Felaniella Dall) and
were certainly present in the Paleocene, whereas the Thyasiridae
are known positively from the Cretaceous and probably from the
Middle Triassic {Storthodon Giebel). Thus, the phylogenetic dia-
gram presented by McAlester (1966) can be extended and aug-
mented with the insertion of the Fimbriidae near to the Lucinidae,
from which stem they diverged in the early Mesozoic or late Paleo-
zoic (Fig. 6).

ACKNOWLEDGMENTS

Dr. J. Rosewater of the United States National Museum (USNM)
loaned preserved specimens for dissection. The manuscript was
critically read by Professor B. Kummel, Dr. R. D. Turner, and
Mr. R. I. Johnson.

LITERATURE CITED

Allen. J. A. 1958. On the basic form and adaptations to habitat in
the Liicinacea (Eulamellibranchia). Phil. Trans. Roy. Soc, ser. B,
241: 421-484.

Barrois, T. 1885. Les glandes du pied et les pores aquiferes chez ies
lamellibranches. Lille, 166 pp.

Dall. W. H. 1895. Tertiary Fauna of Florida. Part 3. A new classi-
fication of the Peiecypoda. Trans. Wagner Free Inst. Sci. Philad.,
3(3): 483-565.

. In K. A. von Zittei. 1900. Text-book of Palaeontology.

London: MacMillan and Co.. 1, part ii.

DuvERNOY, G. L. 1842. Memoire sur Tanimal de i'onguiine couieur de
laqiie {Ungiilina rubra Daud.) et sur les rapports de ce Mollusqiie
acephale. An. Sci. Nat.. 18(2): 110-122.

. 1854. Memoires sur le systeme nerveux des mollusques

acephales. Mem. Acad. Sci. Inst. Fr.. 24: 3-212.

1 Scaldia de Ryckholt from the Visean and Tournaisian Dinantian of
the Lower Carboniferous has only recently been placed in the Fimbriidae
by Chavan /// Moore (1969). Previously it was considered in the Edmon-
diidae ( Vokes. 1967). Ryckholt's original figures indicate that Scaldia
has a pallia! sinus, a characteristic which prevents the genus from being
considered Lucinacean and which removes it from the Fimbriidae.

1970 THE LUCINOID FIMBRIA 15

Ebf.rzin. a. G. //( Orlov. Y. A., ed. 1960. Osnovi Paleontologii. Mol-
liiiski, Dvustvorchatie, 300 pp., Moscow.

Franc, A. In Grassc. P. 1960. Traite de Zoologie. 5(2): 1053-2219.

Hartman. O.. and K. J. Boss. 1965 |1966|. Antoiihniiinia viridis. a

new inqiiiline annelid with dwarf males, inhabiting a new species of

pelecypod, Liiciiui fosicri. in the Mozambique Channel. An. Mag.

Nat. Hist., 8( 13): 177-186.
Kauffman, E. G. 1967. Cretaceous Thyasini from the western interior

of North America. Smithson. Misc. Collect., 132(1): 1-159.

Lamy, E. 1921. Revision des Lucinacea vivants du Museum d'Histoire
Naturelle de Paris. J. Conch., Paris, 65(3): 233-318.

McAlester, a. L. 1965. Systematics, affinities, and life habits of Bah-
inka. a transitional Ordovician lucinoid bivalve. Palaeontology,
8(2): 231-246.

. 1966. Evolutionary and systematic implications of a tran-
sitional Ordovician lucinoid bivalve. Malacologia, 3(3): 433-439.

Menegaux, a. 1889. De la turgescence et de la branchie dans les Lu-
cines. Bull. Soc. Philomath., Paris, 1(2): 130-132.

. 1890. Recherches sur la circulation des lamellibranches

marins. Thesis, Besangon, 296 pp.

Moore, R. C, ed. 1969. Treatise on Invertebrate Paleontology. Part N.

Vol. 2 (of 3), Mollusca 6. Bivalvia, pp. N 491-N 951.
Nakazima, M. 1958. Notes on the gross anatomy of Conchocele dis-

jiincta. Venus, Kyoto, 20(2): 186-197.
Newell, N. D. 1965. Classification of the Bivalvia. Am. Mus. Novi-

tates, no. 2206: 1-25.
Nicol, D. 1950. Recent species of the lucinoid pelecypod Fimbria. J.

Wash. Acad. Sci., 40(3): 82-87.
Odhnlr, N. 1912. Morphologische und phylogenetische Untersuchungen

iiber die Nephridien der Lamellibranchien. Z. wiss. Zool., 100:

287-391.

Pelseneer, p. 1888. Anatomy of the deep-sea Mollusca. CHAL-
LENGER Reports, Part LXXIV, 27: 1-46.

. 1911. Les lamellibranches de Texpedition du SIBOGA. Partie

anatomique. SIBOGA-Expeditie, Monographie 103a, 125 pp.

PiCTET, F. J. 1855. Traite de Paleontologie, Paris: Bailliere, 3.

PoLi, I. X. 1791. Testacea utriusque siciliae. Parma. 1.

PuRCHON, R. D. 1958. The stomach in the Eulamellibranchia; Stomach
Type IV. Proc. Zool. Soc. London. 131(4): 487-525.

Read, K. R. H. 1962. The hemoglobin of the bivalved mollusc, Flicuoidcs
pcctinatiis Gmelin. Biol. Bull. Mar. Biol. Lab., Woods Hole, 123:
605-617.

16 BREVIORA No. 350

RiDEWooD, W. G. 1903. On the structure of the gills of lamellibranchs.

Phil. Trans. Roy. Soc. ser. B, 194: 147-284.
Semper. K. 1880. Die natiirlichen Existenzbedingungen der Thiere.

Leipzig: Brockhaus, 1: 299 pp.
Stasek, C. R. 1963. Synopsis and discussion of the association of cteni-

dia and labial palps in the bivalved Mollusca. Veliger, 6(2): 91-97.
Stoliczka, F. 1870-1871. Memoirs Geol. Survey India. Paleontologia

Indica. Cretaceous Fauna of Southern India, 3: 1-537.
Thiele, J. 1886. Die Mundlappen der Lamellibranchiaten. Z. wiss.

Zool. 45: 239-272.
. 1935. Handbuch der systematischen Weichtierkunde. Jena:

Gustav Fischer, 2: 779-1022.
Valenciennes, M. A. 1845a. Sur Torganisation des lucines et des cor-

beilles. C.R. Acad. Sci.. Paris, 20: 1688-1692.
. 1845b. On the organization of the Lucinae and of Corhis.

An. Mag. Nat. Hist., 16: 41-45.
VoKES, H. E. 1967. Genera of the Bivalvia, a systematic and biblio-
graphic catalogue. Bull. Amer. Paleo., 51(232): 112-394.
White, K. M. 1942. The pericardial cavity and the pericardial gland

of the lamellibranchia. Proc. Malac. Soc. Lond., 25: 37-88.

(Received 16 March 1970.)

ADDENDUM: While this paper was in press, J. A. Allen and J. F. Turner
(1970. Pacific Science, 24(2): 147-154) published 'The morphology of
Fimbria fiinhricita (Linne) (Bivalvia; Lucinidae)' in which they proposed
placing the genus Fimbria in the family Lucinidae.

r\/^r<r ^g<^-^• /-«fr%

Harvard MCZ LIbrai

3 2044 066 302 779

Date Due

CISICULATE