Full text of "Breviora"

See other formats

HARVARD UNIVERSITY

Library of the

Museum of

Comparative Zoology

MUS. COMP. ZOOU.

LIBRARY

MAY 2 4 1977
BREVIORA

HARVARD
UNIVERSITY

MUSEUM OF COMPARATIVE ZOOLOGY

Harvard University

NUMBERS 410-436
1973-1976

CAMBRIDGE, MASSACHUSETTS, U.S.A.

1977

CONTENTS

BREVIORA

Museum of Comparative Zoology

Numbers 410-436

1973

No. 410. The Color Pattern of Scmora michoacanensis (Duges)
(Serpentes, Colubridae) and Its Bearing on the Origin
of the Species. By Arthur C. Echternacht. 18 pp.
September 20.

No. 411. The Mandibular Dentition of Plagiomene (Dermop-
tera, Plagiomenidae). By Kenneth D. Rose. 17 pp.
December 28.

No. 412. Mylostoma variabile Newberry, An Upper Devonian
Durophagous Brachythoracid Arthrodire, with notes
on related taxa. By William J. Hlavin and John R.
Boreske, Jr. 12 pp. December 28.

No. 413. The Chanares (Argentina) Triassic Reptile Fauna. XX.
Summary. By Alfred Sherwood Romer. 20 pp.
December 28.

No. 414. Ecology, Selection and Systematics. By Nelson G.
Hairston. 21 pp. December 28.

No. 415. The Evolution of Behavior and the Role of Behavior
in Evolution. By M. Moynihan. 29 pp. Decem-
ber 28.

No. 416. Museums and Biological Laboratories. By Ernst Mayr.
7 pp. December 28.

No. 417. A New Species of Cyrtodactylus (Geckonidae) From
New Guinea With a Key to Species from the Island.
By Walter C. Brown and Fred Parker. 7 pp. Decem-
ber 28.

No. 418. Morphogenesis, Vascularization and Phylogeny in
Angiosperms. By G. Ledyard Stebbins. 19 pp.
December 28.

No. 419. Protopiychus, A Hystricomorphous Rodent from the
Late Eocene of North America. By John H. Wahlert.
14 pp. December 28.

1974

No. 420. Environmental Factors Controlling the Distribution of
Recent Benthonic Foraminifera. By Gary O. G.
Greiner. 35 pp. March 29.

No. 421. A Case History in Retrograde Evolution: The Onca
Lineage in Anoline Lizards. L A no/is amiectens
new species. Intermediate Between the Genera Anolis
and Tropidodactylus. By Ernest E. Williams. 21 pp.
March 29.

No. 422. South American Anolis: Three New Species Related
to Anolis ni^rolineatus and A. dissimilis. By Ernest
E. Williams. 15 pp. March 29.

No. 423. A New Species of Primitive Anolis (Sauria Iguanidae)
from the Sierra de Baoruco, Hispaniola. By Albert
Schwartz. 19 pp. March 29.

No. 424. The Larva of Sphindocis denticollis Fall and a New
Subfamily of Ciidae (Coleoptera: Heteromera). By
John F. Lawrence. 14 pp. June 28.

No. 425. Systematics and Distribution of Ceratioid Anglerfishes
of the Genus Lophodolos (Family Oneirodidae). By
Theodore W. Pietsch. 19 pp. June 28.

No. 426. Association of Ursus arctos and Arcfodus simus (Mam-
malia: Ursidae) in the Late Pleistocene of Wyoming.
By Bjorn Kurten and Elaine Anderson. 6 pp. Novem-
ber 27.

No. 427. The Stratigraphy of the Permian Wichita Redbeds of
Texas. By Alfred Sherwood Romer. 31 pp. Novem-
ber 27.

No. 428. A Description of the Vertebral Column of Ervops Based
on the Notes and Drawings of A. S. Romer. By
James M. Moulton. 44 pp. November 27.

No. 429. Anolis rupinae new species A Syntopic Sibling of A.
inonticola Shreve. By Ernest E. Williams and T.
Preston Webster. 22 pp. November 27.

1975

No. 430. Anolis marcanoi new species: Sibling to Anolis cyhotes:
Description and Field Evidence. By Ernest E. Wil-
liams. 9 pp. March 28.

No. 431. An Electrophoretic Comparison of the Hispaniolan
Lizards Anolis cyhotes and A. marcanoi. By T.
Preston Webster. 8 pp. March 28.

No. 432. Evolution and Classification of Placoderm Fishes. By
Robert H. Denison. 24 pp. March 28.

No. 433. South American Anolis: Anolis ihague. New Species
of the Pentaprion Group from Columbia. By Ernest
E. Williams. 10 pp. September 19.

No. 434. South American Anolis: Anolis parilis. New Species,
Near A. tuirus Williams. By Ernest E. Williams.
8 pp. September 19.

1976

No. 435. Two New Species of Chelus (Testudines: Pleurodira)
from the Late Tertiary of Northern South America.
By Roger Conant Wood. 26 pp. April 8.

No. 436. Stupendetuys geographicus. The World's Largest
Turtle. By Roger Conant Wood. 31 pp. April 8.

INDEX OF AUTHORS

BREVIORA

Museum of Comparative Zoology

Numbers 410-436

1973-1976

No.

Anderson, Elaine 426

Boreske, JohnR., JR 412

Brown, Walter C 417

Denison, Robert H 432

echternacht, arthur c 410

Greiner, Gary O. G 420

Hairston, Nelson G 414

Hlavin, William J 412

Kurten. Bjorn 426

Lawrence, John F 424

Mayr, Ernst 416

MouLTON, James M 428

Moynihan, M 415

Parker, Fred 417

PiETSCH, Theodore F 425

ROMER, Alfred Sherwood 413, 427

Rose, Kenneth D 411

Schwartz, Alber r 423

Stebbins, G. Ledyard 418

Wahlert, John H 419

Webster, T. Preston . ,. 429, 431

Williams, ErnestE 421, 422, 429, 430, 433, 434

Wood, Roger Conant 435, 436

B R E V I O R A

Miiseiiiii of Coiiipa]iliU&a^jv^^<2Blpgy

LlBftARV

us ISSN 0006-9698 ■'»^'^rfT

Cambridge, Mass. September 20,w©|^2 4^^f^tPER 410

THE COLOR PATT^^j^JW^'^O

Sonora michoacanensis i^Dugis/S^TV
(SERPENTES, GOLUBRIDAE) AND ITS BEARING
ON THE ORIGIN OF THE SPEGIES

Arthur C. Eghternaght

Abstract. The extensive variation in color pattern of the 31 known
specimens of Sonora michoacanensis is described and a model illustrating
the relationships of the major components presented. Sonora aequalis
Smith and Taylor is placed in the synonymy of Sonora michoacanensis
muiabilis Stickel from which it differs only slightly in color pattern. It is
suggested that S. michoacanensis evolved from a bicolor, banded ancestor
within the 5. semiannulata group or from a common ancestor at the southern
edge of the Mexican .Plateau following habitat shifts associated with
climatic changes during the Pleistocene. Sonora michoacanensis is inter-
preted as an imperfect Batesian mimic of elapid coral snakes (Micrurus
sp.) , intermediate irl an evolutionary sequence beginning with the bicolor,
banded ancestor and leading toward a more perfect, tricolor mimic. Known
locality records of S. michoacanensis are mapped and selected meristic
data presented in tabular form.

Introdugtion

The genus Sonora (Serpentes, Colubridae) is represented in
Mexico, at the southern Hmit of its range, by Sonora micho-
acanensis (Fig. 1). Sonora m. michoacanensis (Duges) is found
in arid to semiarid habitats from the upper Balsas Basin in
Puebla to the lower slopes of the Sierra de Coalcoman and
southeastern Colima, whereas S. m. mutabilis Stickel occupies
foothills of the Sierra Madre Occidental from southern Jahsco
to Nayarit and Zacatecas (Duellman, 1961; Zweifel, 1956).
The principal diagnostic difference between the subspecies is
that S. m. michoacanensis has an unmarked tail, whereas the
tail of 6*. m. mutabilis is banded. The two subspecies will be
considered together in the discussion of color pattern to follow.

The last review of this assemblage was by Stickel (1943).

BREVIORA

No. 410

106

Figure 1. Localities of documented specimens of Sonora iiiichoacanensis
in Mexico. Hollow circles: 5. ?/?. michoacanensis; solid circles: S. m. muta-
bilis. D. F. is the Distrito Federal.

His clear and concise discussion included a detailed description
of a single unusual specimen which Smith and Taylor ( 1 945 )
subsequently named, with no further description, Sonora
aequalis. Stickel had been unwilling to base a new species on
the single specimen because it was of unknown provenance and
because it difTered from S. m. mutabilis only in color pattern,
a character known to be highly variable in S. michoacanensis.
Stickel presented data on all 18 specimens of S. michoacanensis
(including ^S*. aequalis) then known but was able to examine
only 1 1 of these. The holotype of S. m. ?nichoacanensis was lost,
and he designated a neotype (Fig. 2), and described S. m. muta-
bilis. The recent discovery of a specimen intermediate in color
pattern to "typical" S. m. michoacanensis and S. aequalis and
the availabihty of 14 specimens of S. ynichoacanensis collected
over the 30 years since Stickel's paper ha\e made possible a
re-examination of the variation in color pattern of the species
and a reassessment of the taxonomic status of S. aequalis. Al-

1973

COLOR PATTERN OF SONORA

Figure 2. Neotype of Sonora michoacanensis micfioacanensis, BMNH 1946.
1.14.65.

though this paper emphasizes color pattern, I have summarized
meristic data for all known specimens (Tables 1 and 2) so that
these data will be available to others. Counts of ventral scales
were made according to the method of Dowling ( 1 95 1 ) and
do not include the anal scale. Counts of subcaudal scales exclude
the tip. For these reasons, data given here may differ slightly
from those presented by Stickel (1943: 114-115). Where
means are given for scale counts they are based only upon
specimens that I was able to examine myself. The color de-
scriptions are based on preserved specimens unless stated
otherwise.

Acknowledgements. William E. Duellman, Richard D. Estes,
Ernest E. Williams and Richard G. Zweifel have all read the
manuscript in its formative stages and I am grateful for their
thoughtful criticism. The research was funded in part by a
grant from the Boston University Graduate School (GRS BI-
.15-BIO). I am indebted to the following individuals and
institutions for the loan of specimens: William E. Duellman
(University of Kansas Museum of Natural History, KU), Her-
bert S. Harris (Personal Collection, RS-HSH), Hymen Marx
(Field Museum of Natural History, FMNH), Hobart M. Smith
and' Dorothy Smith (University of Illinois Museum of Natural
History, UIMNH), David B. Wake (Museum of Vertebrate
Zoology, MVZ), Charles F. Walker and Scott M. Moody (Uni-
versity of Michigan Museum of Zoology, UMMZ), Ernest E.
Williams (Museum of Comparative Zoology, MCZ) and Richard
G. Zweifel (American Museum of Natural History, AMNH).
Herbert S. Harris kindly provided a color slide of a living
snake, and A. F. Stimson was instrumental in obtaining data

4 BREVIORA No. 410

on, and photographs of, the three specimens in the British
Museum of Natural History (BMNH). Photographs of other
specimens were prepared by Frederick W. Maynard.

Variation of Color Pattern

It is almost impossible to exaggerate the extent of variation
in color pattern exhibited by the series of Sonora michoacanensis
presently a\'ailable for study. Only the pattern of the head
and neck seem relatively invariant. There is always a dark
"mask" on an otherwise pale head. The mask may include the
rostral and internasal scales, but typically begins between the
rostral and a line connecting the anterior margins of the orbits.
This dark area surrounds the eye and may extend forward on
the side of the head to include all or parts of the nasal, loreal,
preocular, anterior supralabials and those in contact with the
orbit, the postorbitals and the temporals. Dorsally it covers
the frontal, supraoculars and (often) parts of the prefrontals,
terminating with a crescentic posterior margin on the parietals.
There is a black or dark brown nuchal band (coUar) separated
from the mask by a light-colored band. The nuchal band may
completely encircle the body or may be interrupted midventrally.
The anterior margin of the nuchal band is variable in shape
but the posterior margin is usually straight across. The nuchal
band is followed posteriorly by a light-colored band, usually
three to fixt scales wide, which is, in turn, followed by another
dark band. The last is a "half-saddle," its anterior margin
straight across and its posterior margin crescentic. The half-
saddle may completely encircle the body or be interrupted at
the midline below.

One specimen (FMNH 37141, Fig. 3A) has no pattern what-
soever except that just described. All others have some dorsal
banding pattern. This overall dorsal pattern ranges from one
of only saddle-shaped triads consisting of a median gray band
abutted fore and aft by black {e.g., AMNH 74951, Fig.'4B) to
one of only broad black bands separated by a narrower gray
band corresponding to the median gray band of the triads
[e.g., KU 106286, Fig. 4C-4D). Individual snakes may have
combinations of triads and broad black bands (Fig. SB, 3E-3F).
Occasionally, the broad black bands are partially split by light
pigment extending up from the venter {e.g., MVZ 76714,
Fig. 3B). The light pigment (= ground color) may be ofT-
white, gray, salmon or flesh-colored but to comply with Stickel's

1973

COLOR PATTERN OF SONORA

Figure 3. Sonora michoacanensis michoacanensis: A. FMNH 37141,
dorsal; B. MVZ 76714, dorsal; C. UMMZ 109904, dorsal; D. UMMZ 109904,
ventral; E. FMNH 39129, dorsal; F. FMNH 39129, ventral.

6 BRE\aORA No. 410

(1943) terminolog)- it is referred to as red herein. The black
bands mav not reach the ventral scutes but if thev do, thev mav
or ma}- not extend across them to form rings. The same is true
for the black elements of the triads which may not reach the
\'entral scutes, may completely ring the body in such a way that
the median gray band is also a ring, or may be joined along
the midventral line so that the median gray band is incomplete.
All three possibilities are seen on UMMZ 109904 (Fig, 3D). If
a snake has both triads and broad black bands, it is usual for
the triads to be found anteriorly and the black bands posteriorly
[e.g., FMNH 39129, Fig. 3Ey.

Taylor ( 1937) provides a description of color-in-life of Sonora
michoacanensis michoacanensis from Guerrero and Jalisco. The
ground color is red or pinkish, the dark elements of the triads
black and the middle element of the triads yellow or gray-
cream. A single specimen from Colima is similarly colored
(Harris and Simmons, 1970), but Duellman (1961) described
the middle element of the triads as white in a series of specimens
from Michoacan.

A specimen of Sonora michoacanensis michoacanensis collected
in Jalisco by Percy CUfton (KU 106286, Fig. 4C-4D) is un-
usual in that none of the black bands is split by red and there
are no triads. None of the black bands except the nuchal and
that immediately posterior to it reaches the ventral scutes. The
broad black bands are expanded laterally just above the ventral
scutes and some contact adjacent, similarly expanded bands.
The black and gray bands (black and pale salmon in this
specimen) are subequal in width. This pattern is approached
in MVZ 76714 (Fig. 3B) but, prior to the discover\^ of KU
106286, no S. michoacanensis were known with a pattern en-
tirely of unsplit black bands alternating with gray bands of
approximately equal width. In this respect, KU 106286 re-
sembles Sonora aequalis (MCZ 6444, Fig. 4E-4F).

In addition to presence or absence of caudal banding, Sonora
michoacanensis michoacanensis and S. m. mutahilis differ in
the number of gray bands of females, the number of complete
triads of males, and the number of black bands unsplit by red
of males and females. Sexual differences are e\'ident for all
three of these characters in S. m. mutabilis, but not in S. m.
michoacanensis (Tables 1 and 2). In addition, there is a sta-
tistically significant (t = 3.91, P < .01 with 23 degrees of
freedom) difTerence between the subspecies in total (left plus
right) number of infralabials: The mean and standard devia-

1973

COLOR PATTERN OF SONORA

Figure 4.

Sonora michoacanensis mutabilis: A. UIMNH 18754, dorsal;
B. 'AMNH 74951, dorsal; C. KU 106286, dorsal; D. KU 106286, ventral;
E. MCZ 6444, dorsal; F. MCZ 6444. ventral. MCZ 6444 is the holotype of
Sonora aequalis.

8 BREVIORA No. 410

tions for S. m. michoacanensis are 13.5 ± 1.09, for S. m. muta-
bilis 12.1 ± 0.30. The number of infralabials is not sexuallv
dimorphic for either subspecies. It is notable that of the seven
S. m. michoacanensis with 13 fewer infralabials, three are from
near Coalcoman, Michoacan (UMMZ 106604-6), where a
single specimen (UMMZ 109904, Fig. 3C-3D) has one irregu-
larly shaped caudal band, possibly indicati\-e of intergradation.
Three other specimens with fewer than 14 infralabials (KU
23791, MCZ 33650) or indications of low numbers of infra-
labials (MVZ 45123) are from near Chilpancingo, Guerrero.
The seventh such specimen is the missing holotype from
"Michoacan" [Cope, 1884(1885)].

The Taxonomic Status of Sonora aequalis

The only known specimen of Sonora aequalis (MCZ 6444)^
is recorded as being from Matagalpa, Nicaragua, but Stickel
( 1 943 : 117) concluded that Matagalpa was most likely only
the shipping point for material collected by W. B. Richardson.
Other specimens in the same bottle as the snake and the locality
label were two Eurneces lynxe lynxe (fide Joseph R. Bailey in
Stickel, 1 943 : 1 1 8 ) , a lizard ^vhose range overlaps that of Sonora
michoacanensis mutabilis. This and other evidence led Stickel
to conclude that MCZ 6444 was found within or near the
range of S. 7n. mutabilis. The pattern of MCZ 6444 consists
of 26 black bnnds and 25 gray bands, the bands being all of ap-
proximately the same width (the basis for the name aequalis).
None of the black bands is split by red but se\'eral are xentrally
concave (Fig. 4F). The nuchal band completely rings the body,
but details in this region are obscure because of damage to the
specimen. None of the black bands on the body reaches the
venter and none is expanded laterally as in KU 106286. The
cephah'c pattern is the same as that of S. michoacanensis and
the tail is banded in triads as is characteristic of S. m. mutabilis.
The specimen is badlv faded and no colors other than black
and gray are apparent.

In vie\'/ of the great \'ariation in dorsal body pattern evident
within the su}:)species of Sonora michoacanensis, it does not
seem to mc that the differences between S. aequalis and S. m.

\Stickcl (1943: 117), in error, recorded tlu- snake as ;iii uiicatalogued
specimen in the University of Michi,gan Museum of Zooloi^v. How and
whv it got to Michigan and thence back to the Museum of Comparative
Zoolog\' remains a mystery.

1973 COLOR PATTERN OF SONORA 9

mutabiUs are great enough to warrant taxonomic recognition
of S. aequaUs. These differences are certainly no more startling
than those of the almost patternless FMNH 37141 (Fig. 3A).
KU 106286 (Fig. 4C-4D) seems to be a logical intermediate
in pattern between S. m. mutabilis and S. aequalis. Extensive
collecting in Mexico and Nicaragua over the last 30 years has
brought to light no additional specimens of S. aequalis, but a
number of additional specimens of "typical" (if that word is
admissable) S. michoacanensis have been collected in Mexico.
Of course, no additional specimens similar to FMNH 37141
have been found either.

It may be questioned whether it is any more justifiable to
"sink" a species on the basis of one specimen (KU 106286)
than it was to name one in the first place [S. aequalis, MCZ
6444). But the discovery of KU 106286 has provided an im-
portant link in what appears to be a continuum in pattern
variation extending from the pattern (or, rather, lack of pat-
tern) exhibited by FMNH 37141 to that of MCZ 6444 with the
presence or absence of caudal banding superimposed. The
possibility that KU 106286 is a hybrid of S. aequalis and S. m.
mutabilis cannot be ruled out, but its likelihood is reduced by
the absence of additional specimens of S. aequalis in collections
made over the past 30 years.

Relationships of the Components of Color Pattern
AND THE Origin of Sonora michoacanensis

Figure 5 illustrates my concept of the relationships of the
various components of dorsal color pattern of Sonora michoa-
canensis. Certainly no ontogenetic sequence is impHed, but the
initial stages (Fig. 5A-5B) may be interpreted to suggest some-
thing of the origin of the species. The ancestor of S. michoa-
canensis may have been patterned verv^ much like MCZ 6444.
Progressive erosion of the broad black bands (Figs. 5B-5D)
would yield triads (Fig. 5E). A complex genetic mechanism
would allow indi\ddual snakes to have various combinations of
triads and unsplit black bands or triads in varying numbers and
of varying distances apart. With the exception of the virtually
patternless FMNH 37141, the most consistent element of color
pattern is the gray band between adjacent unsplit black bands
or as the median element in a triad (Stickel, 1943: 116).

The banding pattern of MCZ 6444 is very similar to that of
the banded forms belonging to the Sonora semiannulata group

BREVIORA

No. 410

Figure 5. Diagiammatic representation of color pattern variation of
Sonora michoacanensis. The arrow spans one complete triad. Black r^
black, white :z= white or yellow, stippled rzz red. Upper figure of each
pair, lateral view; lower figure, dorsal view.

of southwestern United States and northern Mexico (Stickel,
1938: 184-186; Stebbins, 1966). MCZ 6444 and all Sonora
michoacanensis have 15 dorsal scale rows with no reduction as
do some members of the S. semiannulata group. Sonora
michoacanensis is distinguishable from members of the S. semi-
annulata group in morphology of the hemipenis (Stickel, 1943:
112), but the two groups are very similar in scutellation, teeth,
dentigerous bone structure, microscopic scale striation and,
generally, color pattern (Stickel, 1943: 110). It seems reason-
able to assume that, as Stickel ( 1 943 : 118) seems to have
suggested, S. michoacanensis had its origin within the S. semi-
annulata group or that the two groups had a common ancestor.
Members of the Sonora semiannulata group are presently
found (Stebbins, 1966) in the southern Warm Temperate and
Subtropical Climatic Zones as broadly mapped by Dorf (1959:
198). These major climatic belts shifted southward \vith glacial

1973 COLOR PATTERN OF SONORA 11

advance during the Pleistocene (Dorf, 1959: 195) and the
range of the S. semiannulata group or its ancestor may have
been depressed southward into the area presently occupied by
iS*. michoacanensis. Sonora michoacanensis may have differ-
entiated as a relict at the southwestern fringe of the Mexican
Plateau when climatic zones retreated northward with retraction
of ^Visconsin glaciation.

The Selective Significance of the Color Pattern

OF Sonora michoacanensis

A number of New World colubrid snakes have tricolor band-
ing patterns which are reminiscent of the red, black and yellow
or white patterns well known among the highly venomous coral
snakes (Elapidae). Considerable circumstantial evidence has
accumulated that the colubrids are mimics of those coral snakes
with which they are sympatric and are thus avoided by those
predators which have learned to avoid coral snakes (Dunn,
1954; Hecht and Marien, 1956; but see Brattsrom, 1955).
Three kinds of mimicry in snakes have been recognized ( Wickler,
1968: 118). Batesian mimicry where the model is highly
venomous and the mimic nonvenomous, Miillerian mimicry
where both models and mimics are highly venomous and rein-
force one another, and Mertensian mimicry where the model
is highly \'enomous and the mimic mildly venomous. Sonora
michoacanensis I?, a Batesian mimic of coral snakes of the genus
Micrurus (Hecht and Marien, 1956: 345).

The ranges of several species of Micrurus overlap or are con-
tained within the range of Sonora michoacanensis (Roze, 1967).
The basic color pattern of these elapids is one of black rings
bordered on either side by narrower yellow or white rings, these
triads being separated along the body by red. The order of the
colors in the triads is, therefore, different from that of S. michoa-
canensis. This difference is probably of little significance insofar
as mimicry is concerned, as the distinction is difficult to make,
even for a trained obser\'er, when the snakes are come upon
suddenly or when they are moving. Potential predators pre-
sumably have the same difficulty and Hecht and Marien (1956:
339) present evidence that the order of the colors is less im-
portant that the presence of the bright, contrasting colors
themselves. In other words, the mimic need not be an exact
renlica of the model to gain a selective advantage.

The concept of Batesian mimicry requires that the mimic be

12 BREVIORA No. 410

less abundant than the model. If relative abundance in museum
collections is an accurate reflection of relative abundance in
nature, this requirement is met in that Micrurus is much
better represented. It should, however, be noted that Sonora
michoacanensis is a secretive species and may not be as rare
as collections indicate. In a few areas where collecting has been
repeated or intensive, small series have been obtained (see list
of specimens ) .

There are two alternative hypotheses concerning the origin
of mimicry- : 1 ) The mimic evoh es in a single step by mutation
(Goldschmidt, 1945), and 2) the mimic evolves gradually
through selection of modifier genes improving upon an original
mutant that had itself a shght selective advantage (Fisher, 1930;
E. B. Ford, 1953). Sheppard (1959) strongly supports the
second hypothesis and suggests that mimetic patterns are con-
trolled by supergenes that have evolved stepwise. Recent experi-
mental work by H. A. Ford (1971) supports the alternative of
gradual evolution and pro\ides evidence that bird predators
avoid a new partial mimic, strongly preferring a familiar non-
mimetic form of prey.

If my interpretation is correct, Sonora michoacanensis evolved
from a bicolor, banded ancestor belonging to the S. semiannulata
group. Although bicolor members of this group are sympatric
with a coral snake {Micruroides euryxanthus) over much of
their range, relative numbers of specimens in museums suggests
the colubrid to be much the commoner snake. Thus, Batesian
mimicry could not develop. To the south, however, the Pleisto-
cene rehct population ancestral to S. michoacanensis may have
been small relative to the populations of Micrurus with which
thev evolved. If this was indeed the case, S. michoacanensis
may as yet have not been perfected as a mimic and should be
considered as intermediate in an evolutionary sequence leading
from a nonmimetic, bicolor, banded ancestor toward a snake
with a pattern of only triads. As there seems to be no geo-
graphic trend in color pattern except the presence or absence
of caudal bands and the generally better mimetic pattern of
male S. m. mutabilis (see below), the gradual perfection of
mimicry seems to be proceeding over the entire range of S.
yyiichoacanensis. The extreme variability in color pattern evi-
dent in the present population would result from lack of fixation
at each of the major and minor gene loci responsible for pattern.
This di\ersity of pattern would be tolerated because all of the
intermediate types are to some degree mimetic except those that

1973 COLOR PATTERN OF SONORA 13

have bicolor banding patterns {e.g., MCZ 6444 and KU
106286) or are nearly patternless {e.g., FMNH 37141). Such
extremes are expected at low frequencies where inheritance is
polygenic and where fixation has not occurred ( Strickberger,
1968). The pattern of S. michoacanensis may be regarded as
both protective in a mimetic sense and as concealing or dis-
ruptive (Brattstrom, 1955). Hecht and Marien (1956: 346)
have suggested that, "Banding may be an intermediate step
through which a disruptive pattern is converted to a ringed,
warning pattern, but functioning in both ways." It seems
equally likely that the disruptive stage is intermediate to banded
and tricolor, warning patterns.

An interesting and unexplained observation is that male
Sonora michoacanensis mutabilis are, by virtue of having more
complete triads (Table 2), better mimics than females and
than both sexes of S. m. michoacanensis. Among butterflies,
mimetic patterns are often sex-limited to females, as are other,
nonmimetic, polymorphisms (Sheppard, 1959: 137). E. B.
Ford (1953) has attributed this phenomenon to the importance
of visual stimuli in the courtship of butterflies. Females make
a choice of mates largely on the basis of visual cues and Ford
(1953: 68) reasons that a new color pattern in males might
not stimulate a female to copulate. In moths, where olfactory
courtship stimuli largely replace visual cues, both sexes may be
polymorphic (Sheppard, 1959: 137). Noble (1937) reviewed
the role of sense organs in the courtship of snakes and concluded
that chemical and tactile senses play the primary^ roles in sex
discrimination and courtship, respectively. Vision was found
to be important only in that movement attracts snakes during
the breeding season. Nothing at all is known of the behavior
of S. michoacanensis, but it seems unlikely that the sexual
dichromatism of S. m. mutabilis serves as an aid to sex dis-
crimination or courtship. There are no clues as to why sexual
dichromatism should be pronounced only in S. m. mutabilis
and not in S. m. michoacanensis.

The color pattern variation exhibited by Sonora michoa-
canensis is at least equaled by that of Sonora aemula Cope of
southern Sonora and Chihuahua, Mexico (Bogert and Oliver,
1945: 374; Zweifel and Norris, 1955: 244; Nickerson and
Heringhi, 1966: 136). Sonora aemula is rare in collections
(Nickerson and Heringhi knew of only ten specimens), but it,
like S. michoacanensis, is probably locally more abundant than
collections indicate. Five of the known specimens were found

14 BREVIORA No. 410

in or near Alamos, Sonora. The species is sympatric with both
Micruroides and Micrurus and one specimen {e.g., Arizona State
University No. 6611; Nickerson and Heringhi, 1966, fig. 1)
may ha\e typical MicruroidesAik^ triads (white-black-white),
S. 7nichoacanensis-\ike triads (black-white-black), or expanded
triads (black-white-black-white-black) like some Micrurus from
southern Mexico and Guatemala. The area between the triads
is red. Mimicry in S. aemula may be at the same stage of de-
velopment as that which I have suggested for S. michoacenensis,
as may mimicry in some species of the venustissimus and annu-
latus groups of the genus S cap hiodonto phis in Central America
(Taylor and Smith, 1943). Scaphiodontophis is a Batesian
mimic of both Micrurus and the mildlv colubrid Erythrolarnprus
(Hecht and Marien, 1956: 342).

Known Specimens of Sonora michoacanensis

The holotype of Contia michoacanensis Duges (Cope), 1884
(1885) (= Sonora michoacanensis) has been lost, and Stickel
(1943: 113) designated BMNH 1946.1.14.65 as neotype.
BMNH specimens have been recatalogued since Stickel's (1943)
paper and both old and new catalogue numbers appear in the
listing to follow. Stickel ( 1943 : 115) examined an uncatalogued
specimen of S. m. mutabilis in the American Museum of Natural
History which was "tied with" (Stickel, 1943) AMNH 19714-
19716, but the present whereabouts of this specimen is unknown
(W. H. Stickel and R. G. Zweifel, personal communications).
Zweifel (1956: 6) has questioned the locality data of all four
specimens. They are said to haxe been collected in Distrito
Federal, Mexico, but this is far remo\'ed from the range of the
subspecies as presently understood from well-documented speci-
mens (Fig. 1) and they are given as "Locality Unknown"
below. Stickel ( 1 943 ) cited specimens in the collections of
E. H. Taylor and H. M. Smith by field number. These speci-
mens have all been deposited in museums, and both field
numbers (preceded by "HMS") and museum catalogue num-
bers are gi\en below.

Sonora michoacanensis michoacanensis (18). COLIMA:
Between Tecoman and Boca de Apiza, RS 596 HSH.
GUERRERO: Chilpancingo Region, KET 23790-1, MCZ
33650, MVZ 45123; 16 km^ S Taxco, UTMNH 25063 (HMS
5440, holotype of Sonora erythrura Taylor, 1937); locality
unknown, unnumbered specimen in the Museo Alfredo Duges,

1973 COLOR PATTERN OF SONORA 15

Colcgio del Estado Guanajuato. MICHOACAN: Apatzingan,
FMNH 39128-9; Apatzingan, Hacienda California, FMNH
37141; 3.2 km E Coalcoman, 1364 m, UMMZ 109904-6;
12.2 km S Tzitzio, 1121 m, UMMZ 119457; 16 km S Uruapan,
MVZ 76714; locality unknown, BMNH 1946.1.14.65 (formerly
BMNH 1903.3.21, neotvpe), the holotype (presumed lost).
PUEBLA: 10 km SE Matamoros, UIMNH 41688.

Sonora michoacariensis rnutabilis (13). JALISCO: near
Magdalena, FMNH 105296 (HMS 4659, paratype), FMNH
105257 (HMS 4661, holotvpe), UIMNH 18754 (HMS 4660,
paratvpe); 6.5 km S Tecalidan, MVZ 71356. NAYARIT:
Jesus Maria, AMNH 74951. ZACATECAS: 8.8 km S Maya-
hua, 1212 m, KU 106286; Mezquital de Oro, BMNH 1946.1.
14.63 (formerly BMNH 92.10.31.42, paratype), BMNH
1946.1.14.64 (formerly BMNH 91.10.31.43, paratype). LO-
CALITY UNKNOWN: AMNH 19714-6 (paratypes), speci-
men "tied with" AMNH 19714-6 (presumed lost), MCZ 6444
(holotype of Sonora aequalis Smith and Taylor).

Literature Cited

BOGERT, C. M., AND J. A. OLIVER. 1945. A preliminary analysis of the

herpetofaima of Sonora. Bull. American Mus. nat. Hist., 83: 297-426.
Br.\ttstrom, B. H. 1955. The coral snake "mimic" problem and protec-
tive coloration. Evolution, 9: 217-219.
Cope, E. D. 1884(1885). Twelfth contribution to the herpetology of

tropical America. Proc. American phil. Soc, 22: 167-194.
DoRF, E. 1959. Climatic changes of the past and present. Contrib. Mus.

Paleont. Univ. Michigan, 13: 181-210.
Bowling, H. G. 1951. A proposed standard system of counting ventrals

in snakes. British J. Herp., 1: 97-99.
DuELLMAN, W. E. 1961. The amphibians and reptiles of ISIichoacan,

Mexico. Univ. Kansas Publ., Mus. nat. Hist., 15: 1-148.
DUNN^ E. R. 1954. The coral snake "mimic" problem in Panama.

Evolution, 8: 97-102.
Fisher, R. A. 1930. The Genetical Theory of Natural Selection. Oxford:

Clarendon Press, xiv + 291 pp.
Ford, E. B. 1953. The genetics of polymorphism in the Lepidoptera.

Advance. Genet., 5: 43-87.
Ford, H. A. 1971. The degiee of mimetic protection gained by new

partial mimics. Heredity, 27: 227-236.
GoLDSCHMiDT, R. B. 1945. Mimetic polymorphism, a controversial chapter

of Darwinism. Quart. Rev. Biol., 20: 147-164, 205-250.
Harris, H. S., and R. S. Simmons. 1970. A Sonora michoacanensis michoa-

canensis (Duges) from Colima, Mexico. Bull. Maryland herp. Soc.

6: 6-7.

16 BREVIORA No. 410

Hecht, M. K., and D. Marien. 1956. The coral snake mimic problem:

A reinterpretation. J. Morph., 98: 335-365.
NiCKERSON, M. A., AND H. L. Heringhi. 1966. Three noteworthy colubrids

from southern Sonora, Mexico. Great Basin Nat., 26: 136-140.
Noble, G. K. 1937. The sense organs involved in the courtship of Storeria,
Tliamnophis and other snakes. Bull. American Mus, nat. Hist., 73:

673-725.
RozE. J. A. 1967. A checklist of the New World venomous coral snakes
(Elapidac) , with description of new forms. American Mus. Novitatcs,

No. 2287, 60 pp.
Shei'pard, p. M. 1959. The evolution of mimicry; a problem in ecology

and genetics. Cold Spring Harbor Symp. Quant. Biol., 24: 131-140.
Smith. H. M., and E, H. Taylor. 1945. An annotated checklist and key

to the snakes of Mexico. U. S. natl. Mus. Bull., No. 187, iv + 239 pp.
Stebbins. R. C. 1966. A Field Guide to Western Reptiles and Amphibians.

Boston: Houghton Mifflin Co., xiv + 279 pp.
Stickel, W. H. 1938. The snakes of the genus Sonora in the United States

and Lower California. Copeia, 1938: 182-190.
. 1943. The Mexican snakes of the genera Sonora and

Cliiouactis with notes on the status of other colubrid genera. Proc.

biol. Soc. \Vashington, 56: 109-128.
Strickberger. M. W. 1968. Genetics. New York: The Macmillan Co.,

X + 868 pp.
Taylor. E. H. 1937. A new snake of the genus Sonora from Mexico, with

comments on S. miclioacanensis. Herpetologica, 1: 69-73.
, and H. M. Smith. 1943. A review of American sibynophine

snakes, with a proposal of a new genus. Univ. Kansas Sci. Bull.,

29: 301-337.
WiCKLER. "\V. 1968. Mimicry in Plants and Animals. New York: "World

Univ. Lib., McGraw-Hill Book Co., 255 pp.
ZwEiFEL, R. G. 1956. Additions to the herpctofauna of Nayarit, Mexico.

American Mus. Novitates, No. 1953, 13 pp.
, AND K. S. NoRRis. 1955. Contribution to the herpetology

of Sonora, Mexico. American Midi. Nat., 54: 230-249,

ADDED IN PROOF: Mr. Scott AL Moody has kindly called my attention
to an additional specimen of Sonora michoacanensis miitahilis obtained too
late for inchision in this study. The snake (UMMZ 131666) is typical of
the subspecies and was found at Prcsa de El Molino. El Molino in Jalisco,
Mexico.

1973

COLOR PATTERN OF SONORA

II.

§1

3£

«J

-^H

«o

iJL

O}^

s§

§'

UICQ

'id a-

93p

-5?

O— (S

«t iij X
•- C
m ts —

K LU U-
CO LJ>--'

u. < £
o — ts

<r 3 cc

ggtz

=) a. La

o: 2 cc:

g[uy

o z: :

t/7

t/0

Ili

a y

CO t/^ CO

N-> N-\ hO

-J -I _l

y a y

<_) o u

K K P

c/; c/; c/v

-3-

LU
U

I I § I I

CO CO CO CO c/:.

fe fe fe

LU
Q_

to N^ OO

9 I

CC CO
t I-

»-

1- t-

c/> OO

I i

oo oo Ln

cz> o r-^

IXJ LX5

e p^

LH C-J •— I OC

CO oo t~>-

gQijqLncncDr~^Lrit--CLni-i

.:3- I~^ ^T

tqj^tq^^ L/^^J^ccc^^^^^^^oc^r^^^

oo CNI en >— I

g:^5iSi5l^"c^"c5AL^^<^^5P^gr^

<-o LTi -a-

m oo

oo o ctj

r~^ oo r— I ' — I ' — I

5::^ t^ i:^ i:^ -i,

KjD r^ r*^ r-v, (X) r^

(^ r"^ r-^ r-^ r*^ r^

.a- -=r -^r r^ ^r

nK nK nS CNI rA

cr ^3- -3- cT ^r

rA rA nK rA nK

I — ^ CO i-^ 1^ r--
r^ oo r^ ?5 i~^

oooototor^i^r^^i-DuDuD

rArArAfA^rArAhArArA
rArArArAcrtAvrArArArA

LJ3 r^ r^ i^^
?5 ^ ?^ r^

r^^ LD i^^ r^ r —
^^ ■^^ ^^ ^^ >~

i^^ i_o i-o r-^ r^

I— ICSICNICSIOJCNICSICNICS;

?5 ^

ro U3 Ln
CNI -3- LA

^ ^

CNI K-\

.-H oo

^ :^

CNJ CNI CNI CNI

q S i:^

CO ir,

S 3 S R P^

CNJ CNI ■—1 CSl CNI CNI CNI

I — I LD Ln ■ — I
■— t o Ln CNI

' — iCsiN^ooc^r^CNi^rcDLn
CDr^LnrsiCNLncTiSncNj'— I

CNICNJCNj6JeNlCNli— ICNIt— ICNl

y y

■s.

y y y y

y y

t— I hn

- ■ — I CO

s ^ s

9 P=; ^

C=5

P^ Pi s

^ nr gr CNJ CNI r-^

:^ &S t

en cri LO ' — ' r-^

N~> fn fA Ln LO CO

m -3- I — ^ en
Ln

S g ^ ^
*^ ^ ^^ ctl

tJ tJ

m nz t^

g S

a a
a: cc

t- fao

O 3

C/3

•*■*

4Mi

c«

J:
o

Vl-I

>~.

i-,

u-l

•^H

•

CL,

c«

J—

c/)

c
M

BREVIORA

No. 410

t^CO

sy§

SSrS

=3 O 1—

(_)

1- LU

OCC

o tg cQ

LXJ

^ f—

03?

LLi ta

3 -J i-

U. < I

O " "^

*:2t

^ Li- LlJ
Z Z _l

_J > t-

9 z 5

S LU [Z

li.<X
O — i3

a:'

LU _ -

U. (/) _L,

o J 3

<
o

q: -3- -3-

ci! 3 a

^ ^

^ a

J -J

ix: uj

o (_)

OO CO

OO CO CO

i I i I i I I

r=H V— I— J— V- H- t—

OO do OO CO OO OO

_J h- I— ►- I— •— •"

W 5 S z: 5 2 2

— . —. LU

^ ;^ jq !q :q fii 3

Q-

O-

OO t>a OO

J- en o^ c~J CD

■3-

1—

>■

J-

►-

1—

V-

-J

111

□_

cvi cn

UD CnI OJ

CD C3

I —
CD

Csl

cr -3-

cn CD

tq t£) 3 ^ ^ I^ S

r-l I-^ N^ Ln LO CO K^
K~> OJ l-O csl Cs] CNl

i^i^gj^sai^^s*

o i"^ cn

3- ^n-

Ln tn i^

^-j-

.—1

■-H

.—1

i3,

N~\ N^ CO ^O-

P:^ i:^ ^ :^

cn N^

It^l

. in r:^

lis

ffi UJ

LlJ

c?5

UD CD CO CD UD to

l_D r^^ CD CD CO CD

.=r cr cr cr ■3- -3"

nS nK nS nK nK hA

.3- N~> 3- -3- ■3- -3-

lA CNi t^ nK nK rA

t^ t-^ r^ t-^ t^ r^

^ -^ -^v "~^ ^^ ^

r^ CD r^ t^ r^ i^^

CN) CNI CNI

CNJ CSJ CNI

CD
CO

CO LD
CD CD

nS rA

csi CM csl

Csl CSl Csl

CO LA N^

.3- cn ■3-

cn •3- 3-

3: g! tq ^ j::^ ^

^ 3- c^-

CSI CNI

SR §

00 00 N^

3- -. ;^

CD
00

CSI

CSJ

y y y y y y y
^ ^ ^ ^ ^ ^ ^

CD CD

y UJ. y y
^ > ¥■ ^

u3 LU

LJ_ LJ_ LL-

3"
LD LD • — I

ap 3- r^

a a a

. CD r^

LD ' — I cn cn

•— 1 cn fsj Jsi

r^ op cn cn

a a

I 8

a on
LU

§ g tq

2 -^ -

fe8

CSJ

S3 a sc s:

^ ^ i>i e:

LU
O

_)
0

I—

CSI

c/5

-a

So
-0

S
0

C/2

<J_1

(>i

tfl

•i-H

l«.S.£uV I 0 R A

UWARY

Museum of Comparative Zoology
JAM? 1974 ^

HAftVARO

us ISSN 0006-9698

CambSB^?^ Number 411

THE MANDIBULAR DENTITION OF

PL A GIOMENE
(DERMOPTERA, PLAGIOMENIDAE)

Kenneth D. Rose^

Abstract. The peculiar bilobate lower incisors and the anterior lower
premolars of the Early Eocene genus Plagiomene are described for the first
time. Several groups of mammals have independently acquired incisors with
divided crowns, but available evidence suggests that any resemblances to
Plagiotnene, except in the case of Recent dermopterans, can be attributed to
convergence. Nevertheless, the close resemblance between the incisors of
Plagiomene and those of certain Recent elephant shrews (Macroscelididae)
may be indicative of similar incisor function. The hypothesis that Recent
dermopterans (Galeopithecidae) are descended from Plagiomene or a closely
allied form (a view previously based primarily on molar morphology) is
strengthened by the specimens described here. A brief review of fossil forms
that have been referred to the Dermoptera is presented, and it is concluded
that, at present, only two fossil genera, Plagiomene and Planetetherium, can
with reasonable probability be assigned to the Dermoptera.

Introduction

The Early Eocene genus Plagiomene has been widely re-
garded as an early member of the Dermoptera, a view based on
the molar morphology, which is similar to that in living der-
mopterans. Fossil evidence of dermopteran e\'olution is ex-
tremely scarce. Although Plagiomene is better known than any
other fossil forms that may be considered Dermoptera, it is
represented only by dental and gnathic remains. Previous litera-
ture on fossil dermopterans (known forms of which are all
assigned to the family Plagiomenidae) is minimal, and has been

^Department of Vertebrate Paleontology, Museum of Comparative Zoology,
Harvard University.

2 BREVIORA No. 411

restricted to descriptions of parts of the dentition. None of the
anterior dentition has been described or adequately figured be-
fore, although the unusual incisors ha\'e been noted pre\'iously
(Jepsen, 1962, 1970; Van Houten, 1945). The nearly complete
lower dentition of Plagiomene described here (PU 14551, right
mandible, and PU 14552, associated left mandible) is significant
in pro\iding new e\idence that Plagiomene is related to and
possibly ancestral to extant dermopterans. In addition, an in-
complete right mandible, PU 13268, provides the first knowl-
edge of the deciduous premolars in Plagiomene.

Comparative material of Plagio?nene and other forms has
been examined during this study. Abbreviations used in the
text are as follows:

AMNH American Museum of Natural Historv, New York
MCZ Museum of Comparative Zoology (Mammalogy Col-
lection), Harvard Uni\ersity, Cambridge, Massachusetts
PU Princeton University Museum, Princeton, New Jersey
YPM Peabodv Museum of Natural Historv, Yale Uni\er-
sity. New Haven, Connecticut

Description

The lower dental formula of Plagiomene, 3.1.4.3, deduced by
Matthew (1918) from f ragmentar\^ specimens, is confirmed bv
PU nos. 14551 and 14552 (see Fig. 1 ).

The three lower incisors (Figs. 1, 2, 4) of Plagiomene are
semiprocumbent, with broad, bilobate crowns, of which the
mesial lobe is the larger. Faint longitudinal depressions on the
lingual sides of these larger lobes in Ii and U (see Fig. 1 lower)
are potential sites for further digitation of the incisor crowns. The
crowns are slighth- convex on the buccal surface and somewhat
concave lingually. The incisors diminish in size from U to U,
Ii being considerably larger than L,. They have an oval, mesio-
distally compressed cross section at the root. In the absence of
the crowns, Matthew (1918) inferred from the roots that the
incisors were small and unspecialized. The specimens discussed
here show this inference to have been incorrect. Expansion of
the incisors (mostly mesiodistally) occurs at the base of the
crowns and increases towards the tip. There are no cingula. A
small wear facet on the labiodistal surface of the mesial lobe of
left Ii suggests that upper incisors may have occluded with the
lower incisors. This is of interest because in the Recent forms, in

1973

DENTITION OF Plagiomeue

-Jo

•^H

•

HH*

4-1

■d

""^

"«*<

-d

-a

4-1

-U

c«

^-J

C«

■4-1

■^-)

-bJ

•.H

-d

.to

'*•

'r*

•^

,(o

•<«»

•»«>

<4-l

*«4

•^H

tJD -O

ft.

:§

H-1

BREVIORA

No. 411

■r

Figure 2. Occlusal view of left mandibular dentition, PU 14552.

1973

DENTITION OF Plagiomeue

' Figure 3. Occlusal view of right mandibular dentition, PU 14551.

BREVIORA

No. 411

5/^tA

Figure 4. Comparison of lower left incisors (I, at top) of Plagiomene
(above) and Cynocephalus (below) .

1973

DENTITION OF Plagiomene

TABLE I

MEASUREMENTS (in mm) OF MANDIBULAR TEETH

OF PLAGIOMENE

PU 14552

PU 13268
(Deciduous teeth)

maximum mcsiodistal length

2.3

maximum height of crown

3.4

(measured Ungually)

maximum mesiodistal length

2.0

maximum height of crown

2.5

(measured lingually)

maximum mesiodistal length

1.4

maximum height of crown

1.8

(measured lingually)

maximum length

2.0

maximum breadth

1.4

maximum length

1.7

maximum breadth

1.3

maximum length

2.8

2.3 (dP,)

maximum breadth

2.0

1.2 (dP,)

maximum length

3.7

3.5 (dP3)

maximum breadth

2.3

2.0 (dPa)

maximum length

4.3

4.3 (dP,)

maximum breadth, trigonid

2.8

1.9 (dP,)

maximum breadth, talonid

3.2

2.3 (dP,)

maximum length

4.3

maximum breadth, trigonid

3.0

maximum breadth, talonid

3.5

maximum length

4.0

maximum breadth, trigonid

3.1a

maximum breadth, talonid

3.3a

maximum length

3.9

maximum breadth, trigonid

2.4

maximum breadth, talonid

2.4

a,— approximate (tooth damaged)

which P is lost and P is reduced, the anteriormost upper teeth
have migrated distally, so that the lower, comblike incisors meet
an edentulous area during centric occlusion. The most com-
plete upper dentition known for Plagiomene, AMNH 15208
(Szalay, 1969: 241), shows diminishing tooth size anteriorly,
however, and does not preserve any incisors (except possibly

8 BREVIORA No. 411

P). This may indicate reduction or loss of the anterior upper
teeth as in extant dermopterans.

The single-rooted lower canine (Fig. 1) of Plagiomene is
premolariform, consisting of a large anterior cusp, which rises
above the crowns of the incisors and of Pi, and a prominent
but low heel. A low, incipient cusp is observed on the anterior
border. The canine is laterally compressed and its root is ellip-
tical in cross section.

The first premolar is a small, single-rooted tooth bearing one
major cusp that may be followed by a much lower, small cusp-
ule. Behind this is a still lower, incipient talonid cusp.

P2 (Figs. 1-3), a much larger tooth than Pi, is double-rooted
and "premolariform-semimolariform" (as defined by Szalay,
1969: 199). The prominent protoconid is preceded by a dis-
tinct though much smaller and lower paraconid, which is situ-
ated directly anterior to the protoconid (not anterolingual to it,
as in the teeth behind P2). The talonid is much broader and
longer than in Pi, but still consists of only a single distinct cusp,
homologous to the hypoconid.

The third premolar is semimolariform. The protoconid is the
largest cusp, and there is a conspicuous, lower paraconid an-
terolingual to it. A less prominent metaconid de\'elops from
the posterolingual border of the protoconid. Some individuals
{e.g., YPM nos. 24966 and 24971) have a small, lower cuspule
anterior to the paraconid. The trigonid is somewhat extended
anteroposteriorly and there is no trigonid basin. The talonid
is well de\'eloped, with both hypoconid and entoconid prom-
inent, and with a rudimentary^ hypoconulid. The talonid basin
is closed posteriorly but is open anteriorly in a deep buccolingual
valley separating the trigonid and the talonid. This feature is
more strongly expressed in the molariform teeth.

P4 is fully molariform, differing from Mi chiefly in its slightly
smaller size, but these two teeth are frequently almost indis-
tinguishable. The three trigonid cusps and two main talonid
cusps of P-i are large and sharp; the hypoconulid is lower and
smaller. Some specimens {e.g., YPM 23578) have a small
entoconulid anterior to the entoconid.

The lower molars have been pre\'iously figured and described
(Matthew, 1918), but a few features may be noted. M1-3 are
very similar to each other. The trigonid cusps are high and
sharp; the metaconid is usually as high as the protoconid or
higher, and the paraconid is somewhat lower. In the talonid a

1973

DENTITION OF Pldgiomene

5MM

Figure 5. Right mandible with functional deciduous premolars (dPj.^)
and unerupted jjemianent P2_4; Mj is in process of eruption. Lateral view
of PU 13268.

pronounced entoconulid is anterior to the entoconid on M2 and
Ms, and it is present on Mi in some individuals. Posterior to the
hypoconulid, the postcingulid rises in a broad cusphke projec-
tion. This is well developed in Mi and M2 and, to a lesser ex-
tent, in P4. In Ms the hypoconulid forms a small third lobe.
Ms is usually narrower buccolingually than the other molars.
The enamel of the molariform teeth is moderately crenulated,
particularly in the talonid. A prominent ectocingulid is present
on P3-M3 and posteriorly on P2. The posterior premolars and
the molars clearly demonstrate a tendency toward polycuspida-
tion, a characteristic of the Plagiomenidae.

The deciduous premolars preserved in PU 13268, dP2-4 (see
Figs. 5 and 6), are in general similar to their adult replace-
ments. They possess the same cusps in approximately the same
positions but are relatively longer anteroposteriorly and more
cqmpressed buccolingually. The talonid of dP2 is more molari-
form than in P2, exhibiting both a hypoconid and a small ento-
conid. The talonid of dPs is similarly more expanded than that
of the replacing tooth. In the trigonid of dPs the paraconid and
metaconid are somewhat more distinct and better separated
from the protoconid than in the permanent Ps. In dP4 as well,
the talonid is elongated and expanded relative to its condition
in P4, and the hypoconulid is much more pronounced, almost
forming a small third lobe as in Ms.

BREVIORA

No. 411

Figure 6. Occlusal view of PU 13268.

Discussion

Incisor specializations comparable to those occurring in Pla*
giomene are found in several other mammals. Incisors with
digitate crowns have evolved independently in several unrelated
groups, including Carnivora, Notoungulata, Macroscelidea, Der-
moptera, and Insectivora. Among these, carnivores such as
Canis and Ursus show tendencies toward digitation of the in-
cisor crowns, but to a less marked degree than in Plagiomene,
and there is surely no relationship involved. Patterson (1940)
described the deciduous incisors of the notoungulate "Progaleo-
pithecus" (^ Archaeophylus), so-named by Ameghino in refer-
ence to the dermopteran-like, pectinate incisor crowns, but there
is no reason to believe that Plagiomene is in anv wav related to
the Notoungulata.

Among the Insectivora, Nesophontes, a recently extinct Antil-
lean form (McDowell, 1958: fig. 3), possesses bilobate incisors
very similar to those in Plagiomene. Tenrec also shows a slight
tendency toward digitation of the incisor crowns. There is little
resemblance of the lower cheek teeth or the upper dentition of
these forms to Plagiomene, however. The superficial similarities
again may be attributed to convergence.

1973 DENTITION OF Plagiomeue 11

Certain Recent elephant shrews (Macroscelididae) bear a
remarkable likeness to Plaoiomene in the conformation of the
incisors; the most striking examples are Petrodromus and par-
ticularly Rhynchocyon. In the former, the crowns of the per-
manent incisors are bilobate, while the milk incisors {e.g., MCZ
26113) may have three or four lobes. The lower incisors of
Rhynchocyon are the closest to Plagiomene of any forms exam-
ined. They are, however, all approximately of equal size in
Rhynchocyon, in contrast to the decrease in size from Ii to Is in
Plaoiomene. The remainder of the macroscelidid dentition is

quite unlike that of Plagiomene. The most obvious contrasts
are the loss of M3 (in the majority of known macrosceHdids,
including both genera mentioned here) and the peculiar struc-
ture of the molariform teeth (PI, MJ, M?.). Macroscelidids are
not common in the fossil record, and of those known (Patterson,
1965; Butler and Hopwood, 1957), none show any particular
resemblance to Plagiomene. The family is unknown outside
Africa. Therefore, the similar form of the incisors in some Recent
macroscelidids is surely not indicative of any close relationship,
although it may reflect functional similarities.

Matthew (1918: 599) noted that the molars of the talpid
Myogale [^= Desmana) were of somewhat similar structure to
those of Plagiomene. Although he viewed this as "perhaps sig-
nificant of a real though remote affinity" {ibid.: 600), the
resemblances do not extend to the other teeth. It is unlikely
that Plagiomene is related to talpids.

Plagiomene has most frequently been compared with the
living dermopterans, Galeopithecidae {e.g., Matthew, 1918;
Romer, 1966; Szalay, 1969; Jepsen, 1970; among others), and
alliance with this group still appears to be the most likely possi-
bility. Matthew (1918) first suggested a relationship between
the two groups after studying the molars of Plagiomene, which
be described as "unlike any placental molars known to me
except those of Galeopithecus" {ibid.: 601). Indeed, the mo-
lariform teeth (P4-M3, as in Plagiomene) of extant dermopter-
ans show many features in common with Plagiomene: prominent
conules; absence of hypocone; paracone and metacone situated
well lingual to the buccal margin; low paraconid; presence of
an entoconuHd; talonid and trigonid separated by a deep bucco-
lingual valley; and crenulated enamel. Furthermore, PJ and,
to a lesser extent, P3 are molarized as in Plagiomene. Although

12 BREVIORA No. 411

the lower incisors of galeopithecids exhibit less resemblance to
those of Plagiomene than do most of the forms discussed above,
the long time inter\'al separating these two forms must be taken
into account. It seems highly probable that the comblike in-
cisors of galeopithecids must ultimately have been deri\'ed from
incisors with divided crowns such as those present in Plagiomene
(see Fig. 4). In fact, the form of I3 in extant dermopterans is
an approximate morphologic intermediate between the form of
the incisors in Plagiomene and the pectinate condition of Ii and
I2 in the living forms. The dental formula of the Galeopitheci-
dae differs from that of Plagiomene, in the loss of two ante-
molar teeth ( probably Pi and P2 ) ; this is easily explained,
however, for the reduction or loss of teeth is common in species
that evolve enlarged, specialized teeth, such as the pectinate in-
cisors of galeopithecids. In summary, the new evidence pro-
vided by the anterior dentition of Plagiomene strengthens the
view that it is in or near the ancestry of the Recent Dermoptera.

This view, however, has been questioned recently. Van Valen
(1967) regarded the Dermoptera as a suborder of the Insecti-
vora. He suggested {ibid.: 271) that the Galeopithecidae may
have been derived from Adapisoriculus (or an unknown related
form) rather than from the Plagiomenidae, which he considered
to be "unrelated to the Galeopithecidae" (although including
both Plagiomenidae and Galeopithecidae in the same super-
family of the Dermoptera, and placing Adapisoriculus in a
suborder separate from the Dermoptera).

From the preceding discussion, it is clear that incisors with
di\ided crowns have arisen independently in many unrelated
mammals and that such incisors function in various ways. Al-
though incisors of different general morpholog)^ are included in
this discussion, some of those mentioned above exhibit close
resemblances to those of Plagiomene. Based on these similari-
ties, incisor function in Plagiomeiie may have been close to that
in Nesophontes, Petrodromus, and Rhynchocyon, and probably
not so much like that in extant dermopterans. Unfortunately,
little is known of incisor use in anv of these forms. Flvinsr lemurs
are reported to use their comblike incisors "in scraping the
green coloring out of leaves" (Gregory, 1951: 387, quoting
H. C. Raven), in ingesting leaves (Winge, 1941), or in groom-
ing (Wharton, 1950). They are strictly herbivorous, feeding
mainly on leaves, but including shoots, buds, soft fruit, and
coconut blossoms in their diet (Wharton, 1950; Walker et al.,

1973 DENTITION OF Plapiomeue 13

1964; Medway, 1969). In contrast, macroscelidids are pri-
marily insectivorous, feeding largely on ants (Brown, 1964),
but almost nothing is known of how macroscelidids use their
incisors.

Hiiemae and Kay (1973) stress that incisors frequently func-
tion in processes other than food ingestion and, in fact, that
minimal use of incisors during ingestion in primitive mammals
provided the opportunity to develop incisor specializations un-
related to feeding. Therefore, it may not be correct to speculate
that the diet of Plagiomene was similar to that of macroscelidids
(indeed, differences in premolar and molar morphology would
seem to be against siich a supposition) ; but it does seem likely
that in both there are similarities of incisor function.

Fossil forms that have been assigned to the Dermoptera are
rare and are represented solely by jaws and teeth. Only two
monotypic genera, Plagiomene (from the Early Eocene of Wyo-
ming) and Planet ether ium} (from the latest Paleocene of Mon-
tana), can with reasonable assurance be referred to the family
Plagiomenidae, the only known family (in addition to the Re-
cent Galeopithecidae) referred to the order. Planetetherium
(Simpson, 1928, 1929; Szalay, 1969) is almost certainly the
direct ancestor of Plagiomene. It is known from only one lo-
cality, the Eagle Coal Mine at Bear Creek, Montana, where it
occurs in carbonaceous shale just above the coal layer (Van
Valen and Sloan, 1966). The site evidently represents an an-
cient swamp, and many of the mammals present (including
Planetetherium) were probably arboreal (Simpson, 1928; Van

^Giasse (1955: 1727, fig. 1698) reproduced drawings of isolated incisors,
from Simpson (1928: figs. 12 and 13) , and attributed the incisors to Planete-
therium. This is apparently an unintentional error, which may have oc-
curred because the description of the incisors (which Simpson, p. 14, stated
"cannot be definitely classified or correlated with cheek teeth as yet") im-
mediately followed the discussion of Planetetherium in Simpson's paper.
Simpson believed that the incisors in question belonged to insectivores or
primates, but he suggested no association with Planetetherium. The mor-
phologies observed differ substantially, indicating that more than one taxon
is involved. Inasmuch as Planetetherium is the most abundant form at Bear
Creek, it seems not improbable that it is among the forms represented by
the incisors. Szalay (1972: 25, figs. 1-9) has recently referred one of these
incisors, AMNH 22153, to the primate Carpolestes, a common occurrence at
Bear Creek. There is little evidence to confirm this allocation and, in fact,
the morphology of AMNH 22153 may be closer to what might be expected in
Planetetherium than in Carpolestes.

14 BREVIORA No. 411

Valen and Sloan, 1966; Jepsen, 1970). Planetetherium is by
far the most commonly found member of the Bear Creek fauna.
Se\'eral isolated teeth from the Early Eocene of France are
the basis for a new genus and species being described by D. E.
Russell, P. Louis, and D. E. Savage (in press) and regarded by
them as a plagiomenid dermopteran. Casts of the teeth show
features that suggest to me, however, that the new form may be
neither a plagiomenid nor even a dermopteran. More complete
evidence may in the future substantiate allocation of this form
to the Plagiomenidae, but I do not believe that presently avail-
able evidence is sufficiently convincing for such an assignment.
L. S. Russell (1954) proposed Thylacaelurus montanus based
on a maxillary fragment from the Kishenehn Formation (Late
Eocene ?), British Columbia, which he believed to have mar-
supial affinities. Although the specimen probably represents a
placental (McKenna, in Van Valen, 1965: 394), Van Valen's
(1967) allocation of the genus to the Plagiomenidae is unjusti-
fied (see also Szalay, 1969: 242). Its relationships will remain
obscure until further material is available.

Van Valen (1967) referred the Mixodectidae to the Der-
moptera. This move also seems unwarranted, but the resem-
blance of Elpidophorus to the plagiomenids may be significant.
This comparison is not new. Simpson (1936) first discussed this
similarity and suggested that Elpidophorus pro\ided a suitable
structural intermediate between the two families, but he rejected
Elpidophorus as an ancestor of Planetetherium on the grounds
that they were approximate contemporaries. This objection is
no longer valid, howe\^er, for the range of Elpidophorus has
since been extended back at least into Torrejonian time. Szalay
(1969) reviewed the status of relationships between the Plagio-
menidae and the Mixodectidae and concluded that a\'ailable
evidence does not support such ties. Nevertheless, the cheek
teeth (both upper and lower) of Elpidophorus are quite similar
to those of Plaoiomene, sufficientlv close to susrsrest that more
than con\'ergence may be involved. It is possible that Elpido-
phorus lies in or near the ancestry of the Plagiomenidae (cf.
Sloan, 1969: fig. 6).

The Picrodontidae were placed in the Dermoptera by Romer
(1966), but I concur with Szalay (1968: 32) that there is no
evidence to support this.

If the Plagiomenidae are truly related to the living fiying
lemurs, as seems probable on the basis of dental e\idence pre-

1973 DENTITION OF Plagiomeue 15

sented above and by Matthew ^1918) and Szalay (1969), the
Dermoptera have been distinct from other mammalian groups
since at least Late Paleocene time. Recent dermopterans have
acquired a peculiar suite of specializations (including in par-
ticular the dental specializations and the patagium) which is not
found in other mammals. In view of these considerations, recog-
nition of ordinal status for the Dermoptera (as accepted by
Simpson, 1945; Grasse, 1955; Butler, 1956; Walker, 1964; An-
derson and Jones, 1967; among others) seems fully warranted.

Acknowledgments

I am indebted to G. L. Jepsen and V. J. Magho, Princeton
University, for granting me the privilege of studying and de-
scribing the Princeton specimens of Plagiomene. G. L. Jepsen
also furnished me with drawings of the specimens prepared
several years ago by R. Bruce Horsf all.

Donald E. Savage kindly sent me a copy of a manuscript
(Russell, Louis, and Savage, in press) describing a new form
from the Eocene of France. Casts of the new specimens were
generously provided by D. E. Russell. I am grateful to Russell,
Louis, and Savage for graciously permitting me to include
herein a dissenting view on the allocation of this new species.

My appreciation is also extended to the following, who have
given me access to specimens under their care: Mary Dawson,
Carnegie Museum; Parish A. Jenkins, Jr., Department of Verte-
brate Paleontology, Museum of Comparative Zoology; Malcolm
McKenna, American Museum of Natural History; C. W. Mack,
Department of Mammalog)^, Museum of Comparative Zoology;
and Elwyn Simons, Peabody Museum of Natural History.

Finally I would like to thank Thomas M. Bown, John G.
Fleagle, F. A. Jenkins, Jr., G. L. Jepsen, and especially Br\an
Patterson for critically reading the manuscript and offering help-
ful suggestions and stimulating discussion. Laszlo Meszoly pre-
pared the drawings; photographs are by A. H. Coleman. The
illustrations were made possible through National Science Foun-
dation Grant GB-30786 to A. W. Crompton.

Literature Cited

Anderson, S., and J. K. Jones, Jr. 1967. Recent Mammals of the World.
New York: Ronald Press. 453 pp.

16 BREVIORA No. 411

Brown, J. C. 1964. Observations on the elephant shrews (Macroscelididae)

of Equatorial Africa. Proc. zool. Soc. London, 143(1): 103-120.
Butler. P. M. 1956. The skull of Ictops and the classification of the

Insectivora. Proc. zool. Soc. London, 126(3): 453-481.
Butler, P. M., and A. T. Hopwood. 1957. Insectivora and Chiroptera

from the Miocene rocks of Kenya Colony. Fossil Mammals of Africa,

no. 13. London: British Museum (Nat. Hist.) , 35 pp.
Grasse, p. 1955. Ordre des Dermopteres. In Grasse, P. (ed.) , Traite de

Zoologie. Paris: Masson, pp. 1713-1728.
Gregory, W. K. 1951. Evolution Emerging. Vol. I. New York: Mac-

millan. 736 pp.

Hiiemae, K. M., and R. R, Kay. 1973. Evolutionary trends in the dynamics
of primate mastication. Fourth Internat. Cong. Primatology. Cranial-
Facial Biology Symposium, M. R. Zingeser, ed. Basel: Carger. 30 pp.

Jepsen, G. L. 1962. Futures in retrospect. Yale Peabody Museum Report
for 1962, 3: 8-21.

. 1970. Bat origins and evolution. In "Wimsatt, W. A. (ed.) ,

Biology of Bats. Vol. L New York: Academic Press. 64 pp.

Matthew, W. D. 1918. Part V — Insectivora (continued), Glires, Eden-
tata. In Matthew, W. D., and Walter Granger, A revision of the Lower
Eocene Wasatch and Wind River faunas. Bull. Amer. Mus. Nat. Hist.,
38: 565-657.

McDowell, S. B., Jr. 1958. The Greater Antillean Insectivores. Bull.
Amer. Mus. Nat. Hist., 115: 117-214.

Medwav, L. 1969. The "Wild Mammals of Malaya. London: Oxford

University Press. 127 pp.
Patterson, B. 1940. The status of Progaleopithecus Ameghino. Field

Museum Nat. Hist., Geol. Ser., 8(3) : 21-25.
. 1965. The fossil elephant shreAvs (Familv Macroscelidi-
dae) . Bull. Mus. Comp. Zool., 133 (6) : 297-336.
RoMER, A. S. 1966. Vertebrate Paleontology. Chicago: Univ. of Chicago

Press. 468 pp.
RussrLL, D. E., P. Louis, and D. E. Savage, (in press) . Chiroptera and

Dermoptera of the French Early Eocene. Univ. Calif. Publ. in Geol. Sci.
Russell, L. S. 1954. Mammalian fauna of the Kishenchn Formation.

southeastern British Columbia. Ann. Rep. nat. Mus. Canada for 1952-

1953, 132: 92-111.
Simpson, G. G. 1928. A new mammalian fauna from the Fort Union of

southern Mcmtana. Amer. Mus. Novitates, No. 297: 1-15.
■ . 1929. A collection of Paleocene mammals from Bear

Creek. Montana. Ann. Carnegie Mus., 19: 115-122.
. 1936. A new fauna from the Fort Union of Montana.

Amer. Mus. Novitates, No. 873: 1-27.
. 1945. The principles of classification and a classification

of mammals. Bull. .\nRi. Mus. nat. Hist. 85: 1-350.

1973 DENTITION OF Plagiometie 17

Sloan, R. E. 1969. Cretaceous and Palcocene terrestrial communities of

western North America. Proc. North Amer. Paleont. Conv., part E:

427-453.
SzALAY, F. S. 1968. The Picrodontidae, a family of early Primates. y\mer.

Mus. Novitates, No. 2329: 1-55.
. 1969. Mixodectidae, Microsyopidae, and the insectivore-

primate transition. Bull. Amer. Mus. nat. Hist., 140: 195-330.
. 1972. Paleobiology' of the earliest Primates, hi Tuttle, R.

(ed.) , The Functional and Evolutionary Biology of Primates. Chicago:

Aldine, pp. 3-35.
Van Houten, F. B. 1945. Review of latest Paleocene and early Eocene

mammalian faunas. Journ. Paleo., 19(5): 421-461.
Van Valen, L. 1965. Paroxyclaenidae, an extinct family of Eurasian

mammals. Journ. Mammal., 46(3) : 388-397.
. 1967. New Paleocene insectivores and insectivore classi-
fication. Bull. Amer. Mus. nat. Hist., 135: 219-284.
Van Valen, L., and R. E. Sloan. 1966. The extinction of the multi-

tuberculates. Syst. Zool., 15 (4) : 261-278.
Walker, E. P., et al. 1964. Mammals of the World. Vol I. Baltimore:

Johns Hopkins Press.
Wharton, C. H. 1950. Notes on the life history of the flying lemur.

Journ. Mammal., 31: 269-273.
Winge, H. 1941. The Interrelationships of the Mammalian Genera. Vol. I,

Copenhagen: C. A. Reitzels Forlag. (English translation.)

B R E V I 0 R A

iMu s^fftff^Tf Comparative Zoology

J^H^ t974

US ISSN 0006-9698

Cambridge, Mass. 28 December 1973 Number 412

HARVARD

um^m^c

OMA VARIABILE NEWBERRY,
AN UPPER DEVONIAN DUROPHAGOUS
BRAGHYTHORAGID ARTHRODIRE,
WITH NOTES ON RELATED TAXA

William J. Hlavin^

and
John R. Boreske, Jr.^

Abstract. All known gnathal elements of the durophagous aithrodire
Mylostoma from the Late Devonian (Famennian) Cleveland Shale of Ohio
show that the inferognathal and posterior palatopterygoid elements increase
in size and maintain a constant shape during growth, Avhile the anterior
palatopterygoids are paired elements in the juvenile condition which fuse
into a single median gnathal in the adult. Dinognatlius is a synonym of
Mylostoma. Mylosioma variahile, Mylostoma eurhinus, and Mylostoma new-
berryi are here considered the only valid taxa. Mylostoma eastmani from
the Grassy Creek Shale of Missouri (Famennian) is now considered a syno-
nym of M. variahile; it was based on undiagnostic gnathal characters. The
fusion of anterior gnathal elements is suggested as a possible origin of the
median gnathal in the enigmatic arthrodire Biingartius and possibly also
in the selenosteid Paramylostoma.

Introduction

, Newberry (1883: 146) described a left inferognathal from
the Cleveland Shale member of the Ohio Shale Formation ( Late
Devonian, Famennian) as Mylostoma variahile, referring to it as
a "dipterine ganoid" on the basis of the similarity of its gnathal
element to those of Dipterus and Ceratodus. In 1893, a concre-
tion containing the virtually complete cranial, thoracic, and

^Cleveland Museum of Natural History, Cleveland, Ohio, and Boston Uni-
versity, Boston, Massachusetts.

^Museum of Comparative Zoology, Harvard University, Cambridge, Mas-
sachusetts.

2 BREVIORA No. 412

ventral shield of a single indi\'idual ^vas collected from the Cleve-
land Shale exposures at Brooklyn, Ohio, and was obtained by
the American Museum of Natural History, with the counterpart
being acquired by the Museum of Comparati\'e Zoology. Dean
( 1901 ) described both specimens as Mylostoma variabile, placing
the taxon within the Arthrodira. Eastman (1906) reviewed the
jaw mechanics of Mylostoma as well as the morphology of its
gnathal elements and concluded that Mylostoma was an arthro-
dire with a gnathal apparatus specialized for crushing.

Hussakof (1909: 268) described Dinognathus ferox as "an
imperfectly definable genus and species of arthrodire" on the
basis of an isolated median gnathal. Eastman (1909) made a
hypothetical reconstruction by placing the Dinognathus ferox
t\pe of dentition over the inferognathals of Mylostoma terrelli
and placing the posterior palatopter)'goids of M. terrelli on the
labial side of the Dinognathus ferox median gnathal. Dunkle
and Bungart (1945) described Dinognathus eurhinus, a second
species of Dinognathus, on the basis of a median gnathal with
general morphology differing from that of D. ferox, but with
features giving evidence for a similar function.

A recently discovered specimen (CMNH 8120) represents a
complete set of jaw elements of an adult Mylostoma variabile.
This specimen, along with other specimens in the Museum of
Comparative Zoologv (MCZ), American Museum of Natural
History ( AMNH),'bberlin College (OC), and the Cleveland
Museum of Natural History (CMNH) has enabled this study
of the morphology of the functional region of the inferognathals
and palatopterygoids through various size-growth stages. Evi-
dence of the fusion of the anterior palatopterygoids has been
observed in the adult, aiding in the synonymy of mylostomatid
taxa that were based oh undiagnostic character-states of the
anterior palatopterygoids.

Order Arthrodira

Family Mylostomatidae

Mylostoma variabile Newberry, 1883

Mylostoma variabile Newberry, 1883: 146
Mylostoma terrelli Newberry, 1883: 147
Dinognathus ferox Hussakof, 1909: 268
Mylostoma eastmani Branson, 1914: 62

Holotype. OC 1 300, left inferognathal.

Paratypes. MCZ 1435, left anterior palatopterygoid ; MCZ

1973 MYLOSTOMA VARIABILE 3

1436, right posterior palatoptcrygoid ; AMNH 42G, left anterior
paIatopter)goid ; and AMNH 43G, right anterior palatopter)'-
goid.

Type locality and horizon. Sheffield Lake, Ohio. South Shore
of Lake Erie, T 7 N, R 17 W, Lorain County, Ohio; Cleveland
Shale member of the Ohio Shale Formation.

Age. Famennian ( Late Devonian ) .

Hypodigm. Cleveland Shale member of the Ohio Shale For-
mation, Ohio: AMNH 7526, nearly complete disarticulated cra-
nial and thoracic shields (counterpart = MCZ 1490) ; CMNH
8129, left and right inferognathals, left and right posterior pala-
topterygoids, median gnathal; AMNH 7915, 10701, CMNH
6094, median gnathals; MCZ 1429-1431, CMNH 5080, 5150,
5177, 6095, 6224, 7256, 7643, 7705, OC 1483, inferognathals;
AMNH 44G, 3290, 3588, 3591, MCZ 1437-1438, 13271-
13274, OC 1301, 1429, CMNH 5022, 5795, 7694, palatoptery-
goids. Huron Shale member of the Ohio Shale Formation,
Ohio: MCZ 13275, right inferognathal. Grassy Creek Shale
Formation, Missouri: University of Missouri collections, median
gnathal, posterior palatoptcrygoid.

Revised diagnosis. Cranial shield having a wide lateral width
and short anteroposterior length similar to that of the titanich-
thyids. Postorbital element bordered posteriorly by paranuchal;
centrals not in contact with marginals and are anteriorly sep-
arated by pineal. Anterior palatopterygoids of juvenile fuse to
form median gnathal in adult. Suborbitals narrow and long,
orbits large. Median dorsal short without well-developed keel.
Median gnathal of Mylostoma variabile possessing a greater
width than length and less deeply excavated on either side of
the longitudinal ridge than that of Mylostoma eurhinus.

Systematic Discussion

The holotype of Mylostoma variabile Newberry (1883: 146)
is a left inferognathal, the size of which indicates that it belongs
to a young adult of the species. The paratypes, comprising the
anterior and posterior palatopterygoids, are characteristic of the
known palatopterygoids of Mylosto?na. Dean (1901) described
the most completely known specimen of M. variabile (MCZ
1490, AMNH 7526). This specimen represents a young in-
dividual of the species ( Plate 1 ) . All of the elements comprising
the upper and lower jaw apparatus are well preserved and are

4 BREVIORA No. 412

the basis for Eastman's (1907) reconstruction of the mylosto-
matid dentition.

A second species, M. terrelli Newberry (1883: 147), repre-
sents the left inferognathal (MCZ 1430) of an individual larger
than the holotype of M. variabile. Hussakof (1909: 268) be-
lieved the specific variations in this specimen could be attributed
only to an age difference in M. variabile, and recommended that
M. terrelli become a synonym of M. variabile.

A third species of Mylostoma, M. newberryi Eastman (1907:
224) is based on a pair of dental elements identified as the
anterior portions of left and right inferognathals (OC 1302) and
the posterior portion of a smaller left inferognathal (MCZ 1439) .
These dental elements were originally described by Newberry
(1889: 165) as belonging to M. variabile because of their dis-
tinctive narrowness and triangularity, which he believed demon-
strated diversity in the species. Earlier, Eastman (1906: 22;
fig. E) figured these plates as pre-anterior palatopterygoids as
part of his reconstruction of the upper dentition of M. variabile.
This reconstruction is misleading since these pre-anterior pala-
topterygoids are not present in the MCZ 1490 and AMNH 7526
specimens. We believe that Eastman realized this a year later
and established M. newberryi to include these "extra" plates.
Morphologically, the dental plates represent the functional region
of the inferognathal in a juvenile mylostomatid, having a very
thin and narrow attachment with the blade of the inferognathal.
This functionally weak attachment between the two areas in this
bone may be a result of either an extremely early growth stage
or a pathologic condition, the latter being here suggested as an
explanation for the abnormal osteological conditions in the jaw
elements of the dinichthyid Hussakofia ( Cossmann ) .

Branson (1914) described Mylostoma eastmani on the basis of
an isolated posterior palatopterygoid from the Famennian Grassy
Creek Shale of Louisiana, Missouri. This specimen, along with
an element referred to by him as an "occipital" (= nuchal) of
Dinichthys rowleyi (correctly identified as a Dinognathus-\ike
median gnathal by Dunkle and Bungart, 1945), comprises the
only known occurrence of Mylostoma outside the Ohio Shale
Formation. The character-states established by Branson (1914)
for Mylostoma eastmani are undiagnostic since they do not differ
from those of M. variabile, and we therefore include Mylostoma
eastmani as a synonym of Mylostoma variabile. This occurrence,
however, extends the distribution of this genus outside of the
Appalachian Basin onto the mid-continent.

1973

MYLO STOMA VARIABILE

/0^mm:\

••.'••-•;;5-i>-!'>.\ v';-,! '• -J'-;''-".-;;;.' • /

Figure 1. Median gnathal elements (after Dunkle and Bungart, 1945) :
A, Mylostoma (= Dinognathus) eurhinus CMNH 5063; B, Mylostoma varia-
bile {= Dinognathus ferox) CMNH 6094; d = dorsal, v = ventral.

6 BREVIORA No. 412

Hussakof (1909: 268) described Dinognathus ferox (Fig. IB)
from a single median gnathal (AMNH 7915) resembling the
mvlostomatid dentition but ha\ins: uncertain affinities. Eastman
(1909) felt that D. ferox represented the fused part of the an-
terior palatopterygoids of an adult Mylostoma, but he lacked the
appropriate specimens needed to prove this hypothesis. Dunkle
and Bungart (1945), in describing Dinognathus eurhinus from
a median gnathal (CMNH 5063; Fig. lA), did not advocate
Eastman's ideas on fusion of the anterior palatopterygoids and
opposed his hypothesis on anatomical grounds, which they felt
were contradictory to the generalized pattern of jaw elements in
all arthrodiran fish. They considered his reconstruction of the
Dinognathus median gnathal as a dorsal gnathal element of
Mylostoma to be invalid, arguing that the median gnathal could
not have been derived from the fusion of the anterior pair of
mylostomatid palatopterygoid elements.

A recently discovered specimen (CMNH 8129; Plate 2) rep-
resents a complete set of gnathal elements belonging to an adult
M. variabile. This specimen consists of typical right and left
inferognathals, right and left posterior palatopterygoids, and a
Dinognathus ferox median gnathal. The discoxery of this speci-
men, which lacks the anterior palatopterygoids but has posterior
palatopterygoids and inferognathals associated with the D. ferox
median gnathal element, confirms Eastman's hypothesis that the
median gnathal of D. ferox represents the fusion of the anterior
palatopterygoids in the adult mylostomatid (Fig. 2). A survey
of all known existing mylostomatid palatal dental plates shows
them to fall into three size categories : ( 1 ) the posterior pala-
topterygoids, having a size-growth range from ju\^enile to adult,
(2) the anterior palatopterygoids, all representing juvenile speci-
mens of varying degrees but none approaching the adult size of
their corresponding posterior palatopterygoids, and ( 3 ) the me-
dian gnathals or fused anterior palatopterygoids, which all cor-
respond to the adult size of the inferognathals and posterior pab.-
topterygoids of the genus Mylostoma.

In \iew of this evidence, it is su2:s:ested here that the taxonomv
of the Mvlostomatidae ma\ be revised as follows: the 2:enus
Dinognathus becomes a synonym of Mylostoina; Mylostoma
variabile, the type species, includes also Dinognathus ferox,
Mylostoma terrelli, and Mylostoma eastmani as synonyms;
"Dinognathus'' eurhinus becomes a valid species of Mylostoma;
Mylostoma newberryi, a species known only from the anterior
portions of its inferognathals, is included within the M}losto-

1973

MYLOSTOMA VARIABILK

Figure 2. A, Eastman's (1907) reconstruction of the upper jaw apparatus
of Mylostoma variabUe, displaying the paired anterior palatopterygoids (AP)
of the juvenile condition (reconstruction based on AMNH 42G-43G, 3591,
and MCZ 1437) ; B, Reconstruction of the tipper jaw apparatus of Mylo-
stoma variabile, displaying the median gnathal (MG) of the adult condition
(fused anterior palatopterygoids; reconstruction based on CMNH 8129) ;
PP = posterior palatopterygoids.

8 BREVIORA No. 412

matidae but its affinities with the other species of Alylostorna
cannot be determined until additional material becomes avail--
able.

Comparison With Other Arthrodires Having A
Similar Jaw Apparatus

As presently constituted, the family Mylostomatidae embraces
the following genera: Mylostoma (= Dinognathus), Dinomylos-
toma, and possibly Tafilalichthys. Eastman (1906) described
Dinomylostoma, which is restricted to the medial Frasnian Shales
of New York and Kentucky, as being phylogenetically the most
primitive of the mylostomatids. Although incompletely known,
it is morphologically and chronologically transitional between
Dinichthys and Alylostoma. The inferognathal elements possess
a flat, narrow oral surface, not yet expanded as in Mylostoma.
The blade-length comprises approximately 45 percent of the
inferognathal, displaying the generalized condition of the ad-
ductor mandibulae muscles in the Frasnian mylostomatids, as
compared to the 60 percent blade-length attained by the arched
forward inferognathal elements of the Famennian Mylostoma.
According to Dunkle and Bungart ( 1 943 ) , this specialized con-
dition increases the length of the adductor mandibulae muscles
to produce a more powerful bite. The anterior dorsal gnathal
elements of Dinomylostoma display features transitional between
the dinichthyid anterior supragnathals and the mylostomatid
anterior palatopterygoids. The posterior gnathal elements, how-
ever, have become completely specialized into well-defined my-
lostomatid posterior palatopterygoids. This gnathal condition is
paralleled to a less specialized degree by the Frasnian pholidosteid
Malerosteus, described by Kulczycki (1957) from the Holy
Cross Mountains of Poland.

It is interesting to note that the enigmatic arthrodire Bungar-
tius perissus Dunkle, which is known from a single complete
adult specimen, lacks the anterior supragnathal element. The
jaw elements preserved represent the corresponding right and
left inferognathals, the posterior supragnathals, and a wtII-
developed median gnathal. In this case, Dunkle (1947: 104)
considered the "anterior supragnathal element either \estigial or
completely absent." The absence of the anterior supragnathal
elements in the adult Bungartius parallels the absence of these
elements in the adult Mylostoma. The median gnathal is uniquely
restricted to these two genera and we believe it has developed

1973 MYLOSTOMA VARIABILE 9

through the fusion of the anterior supragnathal elements during
srrowth. This condition mav occur also in the selenosteid Para-
mylostoma Dunkle and Bungart, in which the jaw mechanism is
represented by an inferognathal specialized for crushing, and an
associated posterior supragnathal. The anterior supragnathal
and/or median gnathal is unknown in this genus.

The gnathal condition, suggesting a durophagous habit, while
not exclusively restricted to the Mylostomatidae as demonstrated
by Bungartius, Paraniylostoma, and Malerosteus, has achieved
its highest degree of specialization in the genus Mylostoma. This
gnathal condition as manifested within other families of arthro-
dires is believed to represent diverse attempts of broader adapta-
tion and efficiency of the feeding mechanisms at the pachyosteo-
morph le\el of organization as suggested by Miles (1969).

On the basis of an isolated cranium, Lehman (1956) de-
scribed Tafilalichthys lavocati as a new brachythoracid arthro-
dire from the Famennian of Southern Morocco. Obruchev
(1964), in his review of this genus, suggested that Tafilalichthys
lavocati might be a mylostomatid, since the cranium is morpho-
logically similar to that of Mylostoma as described by Dean
(1901). No gnathal elements are yet known from T. lavocati,
and therefore no positive assignment to the Mylostomatidae can
be made at this time. However, the close relationship of the
North American Famennian arthrodiran taxa to the Moroccan
arthrodiran remains, as well as a review of the Cleveland Shale
Arthrodira, will be of considerable interest in documenting the
phylogenetic and paleozoogeographic relationships within the
Mylostomatidae.

The stratigraphic range of Mylostoma is relatively short, re-
stricted to the Famennian (Late Devonian) time in North
America. At this time the brachythoracid arthrodires achieved
their highest level of adaptive radiation before extinction.

ACKNOW^I.EDGMENTS

Thanks are due to J. -P. Lehman and Daniel Goujet ( Museum
National d'Histoire Naturelle, Paris), Farish A. Jenkins, Jr.
and Robert H. Denison (Museum of Comparative Zoology),
Richard Estes (Boston University), and William E. Scheele
(Cleveland Museum of Natural History) for their helpful sug-
gestions. This research was supported in part by grants from
the Albion Foundation and Sigma Xi to Hlavin.

10 BREVIORA No. 412

Literature Cited

Branson, E. 1914. The Devonian fishes of Missouri. Univ. Missouri Bull.,

15(31): 59-74.
Dean, B. 1901. On the characters of Mylostoma Newberry. Mem. New

York Acad. Sci., 2 (3) : 101-109.
DuNKLE, D. 1947. A new genus and species of arthrodiran fish from the

Upper Devonian Cleveland Shale. Cleveland Mus. Nat. Hist. Sci. Publ.,

8(10) : 103-117.
, AND P. BuNGART. 1943. Comments on Diplognathus mirabilis

Newberry. Cleveland Mus. Nat. Hist. Sci. Publ., 8 (7) : 73-84.

AND . 1945. Preliminary notice of a remarkable

arthrodiran gnathal plate. Cleveland Mus. Nat. Hist. Sci. Publ., 8 (9) :

97-102.
Eastman, C. 1906. Structure and relations of Mylostoma. Bull. Mus. Comp.

Zool., 50(1).: 1-29.
. 1907. Mylostomid dentition. Bull. Mus. Comp. Zool., 50 (7) :

211-228.

1909. Mylostomid palatal dental plates. Bull. Mus. Comp.

Zool., 52 (14) : 261-269.

HussAKOF, L. 1909. The systematic relationships of certain American
arthrodires. Bull. Amer. Mus. Nat. Hist., 26: 263-272.

KuLCZYCKi, J. 1957. Upper Devonian fishes from the Holy Cross Moun-
tains (Poland) . Acta Pal. Polonica, 2 (4) : 285-380.

Lehman, J.-P. 1956. Les arthrodires du Devonien Superieur du Tafilalet
(Sud marocain) . Notes Mem. Serv. Geol. Maroc, 129: 1-70.

Miles, R. 1969. Features of placoderm diversification and the evolution of
the arthrodire feeding mechanism. Trans. Roy. Soc. Edinburgh, 68 (6) :
123-170.

Newberry, J. 1883. Some interesting remains of fossil fishes, recently dis-
covered. Trans. New York Acad. Sci., 2: 144-147.

. 1889. The Paleozoic fishes of North America. Monog. U.S.

Geol. Surv., 16: 1-340.

Obruchev, D. 1964. Class Placodermi. hi Osnovy Paleontologii 11. Mos-
cow: Nauka, pp.1 68-260.

1973

MYLOSTOMA VARIABILE

'T" *^yf^»^

^_ — I

ocrri

Plate 1. Mylostoma variabile (displaying cranial, thoracic, and ventral
shields) , juvenile: A, MCZ 1490; B, counterpart AMNH 7526; SO = sub-
orbital.

BREVIORA

No. 412

\ '»>;-'v >^'

r"'^^*^

"|gi-v

5cm

Plate 2. Mylostoma variabile CMNH 8129; jaw elements of an adult
showing left and right infeiognathals (IG) , left and right posterior pala-
topterygoids (PP) , and a median gnathal = fused anterior palatopterygoids
(MG) .

S-/VA ^(a.>Mo

B^HoJEoouV I O R A

LIBRARY

11 seiim^of ^Comparative Zoology
JAM? 1974 ^ ^^

JAN

US ISSN 0006-9698

CAMBRiDGKajl^j^f^gl*^, December 1973 Number 413

THE GHANARES (ARGENTINA)

TRIASSIG REPTILE FAUNA

XX. SUMMARY

Alfred Sherwood Romer

Abstract. A brief account is given of the geologic setting of the Tiiassic
tetrapod faunas found in South America; the nature of the Chanares reptile
fauna is summarized, and this fauna is compared with other Triassic as-
semblages in South America and other continents.

In nineteen pre\'ious papers in the Museum of Comparative
Zoology Breviora^; an account has been published of the reptile
fauna ifrom the Triassic Chafiares Formation of Argentina col-
lected by the La Plata-Harvard expedition of 1964-65; this
series includes, in addition to papers written by myself, contribu-
tions by C. Barry Cox, Parish A. Jenkins, Jr., James A. Jensen,
and Arnold D. Lewis. Except for a future detailed study of the
skull of the cynodont Pro baino gnat hus by Edgar F. Allin and
myself I have no further plans for publication on the Chafiares
fauna. The present paper is intended to furnish a short summary
of the results of the 1964-65 expedition. Except for a few forms
recently described from the Chanares Formation, a recent paper
by Bonaparte (1972) gives a succinct summary of all known
reptiles from the South American Triassic, so that detailed ref-
erences are unnecessary below.

As noted in previous papers in this series, I am deeply in-
debted to the National Science Foundation for grants for collec-
tion, preparation, and publication of the Chanares fauna.

Geologic Setting
Until the last few decades, almost nothing was known of the

^Breviora Nos. 247, 252, 264, 295, 333, 344, 352, 373, 377, 378, 379, 385, 389,
390, 394, 395, 396, 401, and 407.

2 BREVIORA No. 413

Triassic tetrapod faunas of South America. Now, however, tetra-
pods are known from fi\e discrete areas of Argentina and south-
ern Brazil:

( 1 ) The El Tranquilo Formation of Santa Cruz Province of
Patagonia. From the upper part of this formation, ob\'iously of
Late Triassic age, have been collected prosauropod dinosaur
remains. These ha\'e been studied by Casamiquela, but the
results have not been published; they appear to pertain to the
European genus Plateosaiirus.

(2) The Puesto Viejo Formation, in southern Mendoza Prov-
ince. Undescribed fragmentary remains are present in the lower
part of the formation; from the upper part, Bonaparte has
described a primitive but somewhat specialized traversodontid
gomphodont Pascualgnathus and, most interestingly, forms in-
distinguishable from Cynognathiis and Kannemeyeria, the most
characteristic genera of the Cynognathus zone of the Upper
Beaufort beds of South Africa. The Scythian age of this forma-
tion is obvious.

(3) The Cacheuta Basin. In the precordillera west of Men-
doza is a series of beds of Triassic age, the Cacheuta Series. I
have elsewhere (Romer, 1960) given a brief resume of the
geology. Four formations have long been recognized; in ascend-
ing order they are the Las Cabras, Potrerillos, Cacheuta and Rio
Blanco; recently a basal Rio Mendoza Formation has been dis-
tinguished. Rusconi, in various publications (as Rusconi, 1951)
has described \'ertebrates from these beds, including various
fishes, many of uncertain systematic position, and from the
Cacheuta Formation, flat-skulled amphibians of the genus Pel-
orocephalus [Chigutisauriis], which, although comparable in
many regards to the brachyopids of other Gondwana continents,
appears not to pertain to that group. Reptilian remains are rare;
in the older collections there was, apart from a few scraps, only
the postcranial skeleton of a primiti\c thecodont, Cuyosuchus.
More recently an indeterminate jaw from the Potrerillos Forma-
tion has been described as Colbertosaurus, and Bonaparte has
described the gomphodonts Andescynodon and Rusconiodon and
a kannemeyeriid dicynodont, Vinceria from the Rio Mendoza
Formation. Because the flora of the Cacheuta Series is of the
Dicroidium type present in the Late Triassic, Stipanicic (1969)
believes the Cacheuta beds to be relati\elv Late Triassic in as^e.
Howe\er since the Dicroidhim flora extends well down toward
the level of the Upper Beaufort beds of South Africa, Bona-
parte's belief (1966, etc.) that part of the Cacheuta Series is

1973 CHANARES SUMMARY 3

relatively Early Triassic in age is reasonable. Unfortunately the
reptile fauna is as yet too fragmentary in nature for adequate
comparisons to be made.

(4) Santa Maria Formation. From this Triassic formation in
southern Brazil a few bones were early sent to the British Mu-
seum; major collections were later made by and for Huene,
whose full results were published in 1944; further collections
have been made by Price and White for Harvard University, by
Colbert for the American Museum, and by Price for the Bra-
zilian Geological Survey. The Santa Maria Formation has been
described by Beltrao ( 1 965 ) and by Bortoluzzi and Barbarena
(1967). The vertebrate remains are confined to the upper part
of the formation, and there is no known difference in the age
of the beds between the three major collecting areas — near the
city of Santa Maria, in the region of Chiniqua, west of that city,
and in the Candelaria region, well to the east.

The fauna is varied, but the nature of preservation is such that
structural details are frequently obscure and many forms are
imperfectly known. Included are the procolophonid cotylosaur
Candelaria; the rhynchosaur Scaphonyx [Cephalonia] ; a number
of thecodonts including Cerritosaurus, Rauisuchus, Prestosuchus,
Hoplitosuchus, Procerosuchus; a fragmentary postcranial skeleton
that appears to be a primitive saurischian, Staurikosaurus and a
questionable second dinosaur, represented by a few vertebrae
and limb bones; two carnivorous cynodonts, Chiniquodon and
Belesodon; the gomphodont cynodonts Traversodon and Gom-
phodontosuchus; the dicynodonts Barysoma, Dinodontosaurus
and Stahleckeria.

As discussed later, the Santa Maria Formation seems surely to
be equivalent to the Los Rastros Formation of the Talampaya
basin.

(5) The Talampaya basin or Villa Union-Ischigiialasto
cuenca. This is the largest and most richly fossiliferous of the
bope-bearing South American Triassic areas. It lies on the
boundary between La Rioja and San Juan provinces, between
the Sierra de Safiogasta on the east and the Rios Bermejo and
Guandacol on the west, and extends from the region of Villa
Union on the north to the Sierra de Valle Fertil on the south.
Faults are numerous, but in general the Triassic beds can be
grouped in two areas, east and west of the flat alluvium-covered
Talampaya plain, the two areas being essentially the two limbs
of a major syncline, with various formations present in reverse
order on the two sides of the plain. The area to the west of the

4 BREVIORA No. 413

plain is the better known and here the formations identified are
much thicker than on the east. This region was explored by
earlier geologists, but first adequately studied by Frenguelli
( 1 946 ) ; his account has been modified and corrected by later
workers, such as Groeber and Stipanicic (1953) and Ortiz
(1968). To the northwest, in the region of Cerro Bolo there is
an exceedingly thick series of beds that appear to extend con-
tinuously upward from the Carboniferous "Paganzo I" to the
Late Triassic; this region was studied by de la Mota, whose
work, unfortunately, remains unpublished. To the southwest the
series, as far as published results are concerned, terminates below
in the presumed Triassic "Paganzo III." For much of the west-
ern border this last is absent; if included, the major formations,
in descending order, are :

Los Colorados Formation,

Lschigualasto Formation,

Los Rastros Formation,

Tarjados Formation (= Paganzo III).

As described by Frenguelli, the Los Colorados beds were
termed the Gualo Formation, a mistake corrected bv Groeber
and Stipanicic. The lower part of the Los Rastros Formation
was synonymized by Frenguelli with the Ischichuca Formation;
as pointed out by Ortiz this is incorrect, for the type Ischichuca,
in the Cerro Bolo region, is synonymous with the main carbon-
bearing beds of the Los Rastros. The lowest redbeds were
thought by Frenguelli to represent the Permian "Paganzo II,''
whereas, as Ortiz states, they are the redbeds of "Paganzo III,"
or Tarjados.

Fragments of vertebrate skulls were reco\'ered by Frenguelli
from the Ischisfualasto Formation and described bv Cabrera in
1943. The richness of fossils in this formation was disclosed by
the Har\'ard-Buenos Aires Museum expedition of 1958 (Romer,
1966). For many years, from 1958 on, the lschigualasto beds
were worked by expeditions from the Instituto Lillos of Tucu-
man, at first under O. A. Reig, later with great success by J. F.
Bonaparte. The rich reptile fauna includes the rhynchosaur
S ca phony x: the thecodonts Proterochampsa, Saurosuchus, Ven-
aticosuchus, Triassolestes, Aetosauroides and Argentinosuchus ;
the rare saurischian dinosaurs Herrerasaurus and {?)Ischisaurus;
the ornithischian Pisanosaurus ; fragmentan- remains perhaps
representing the carnivorous cynodont Chiniquodon; the gom-
phodonts Exaeretodon, Proexaeretodon and Ischignathus; the

1973 CHANARES SUMMARY 5

dicynodont Ischigualastia. Except for representatives of Ischi-
gualasto forms in transitional beds, no reptiles are known from
the Los Rastros beds or the underlying Tarjados Formation.
Abo\e the Ischigualasto Valley rise the high clifTs of the Los
Colorados. Except for a single dicynodont, Jachaleria, the faunal
content of most of the thick series of Los Colorados redbeds is
unknown; from the few meters available at the summit of the
cliffs Bonaparte has described (1972b) a fauna of very late
Triassic age, including the thecodonts Riojasuchus, Pseudhes-
perosuchus and N eoaetosauroides ; the primitive crocodilian
H emiprotosuchus ; the prosauropod Riojasaurus; and fragmen-
tary materials comparable to Tritylodon.

We are here concerned mainly with beds lying to the eastern
side of the basin, which was little studied by earlier workers;
Jensen and I (1966) have discussed the geology here. Most of
the formations present can be matched with those on the west
side of the \alley, although they appear to be much thinner here.
The formations present (all adequately represented along the
course of the Arroyo de Agua Escondida) are, in descending
crder :

-Los Colorados Formation,

Ischigualasto Formation,

Los Rastros Formation,

Chafiares Formation,

Tarjados Formation,

Talampaya Formation.

These formations are presumably underlain by the Carboni-
ferous and Permian beds of "Paganzo F' and "Paganzo II,"
which are exposed on the slopes of the Safiogasta Range, east of
a major north-south fault at the western margin of the moun-
tains; in the area studied, however, we have not seen a contact
between "Panganzo II" and the base of the Talampaya beds.
The latter formation is best exposed in the clifTs forming the
walls of the "Puerta de Talampaya," where 180-200 meters of
these beds are present. They mainly consist of soft sandstones,
but with occasional "cobbles." No fossils of anv sort have been
found. They appear to be purely continental in nature and are
not improbably Early Triassic in age, or possibly Late Permian.

Unconformably above the Talampaya beds are the hard sand-
stones of the Tarjados Formation, some 385 meters in thickness
at the i\rroyo de Agua Escondida. These beds correspond, ap-
parently, to part or all of the sandstones elsewhere termed

6 BREVIORA No. 413

"Panganzo III." For the most part they are red, but in the
southern part of the area studied the upper beds are white in
color. Fossils are rare, but a few fragmentary dicynodont re-
mains have been found in the upper layers. They are presumably
Early Triassic in age.

On the irregular upper surface of the Tar j ados sandstones
lie unconformably the 75 meters of the volcanic ash deposits
constituting the Chaiiares Formation. The uppermost layer of
the Tar j ados, about half a meter thick, forms an uneven, undu-
lating surface of hard resistant materials suggesting hydrothermal
action. Obviously there was major volcanic activity in the region
at that time. The Chaiiares sediments show none of the laver-
ing that would be expected if the ash had been laid down in
water; presumably there was merely a covering of the then
existing surface with tremendous quantities of volcanic ash in
Pompeii-like fashion. Bearing out such a conclusion is the fact
that no trace of water-dwelling amphibians or fishes have been
discovered in the Chaiiares and — more significant — almost
all the numerous reptile remains found are in the lowest few
meters of the ash deposits. Apparently the ash falls resulted in
the local extermination of the vertebrate fauna.

As Jensen and I noted in 1966, it is not customary' in Argen-
tina to give a formation name to a set of beds of such limited
thickness. I believe, however, that it is warranted in this case
because of the distinctive nature of the sediments, and most
especially, because of the vertebrate fauna contained in them.
Bonaparte (1967) suggested that the Chaiiares beds are
equivalent to those of the Ischichuca Formation, the type section
of which lies in the Cerro Bola region. However, both Ortiz
(1968) and I (1971) have shown that this is incorrect. Bona-
parte informs me that light-colored beds, which may be compa-
rable to those of the Chaiiares, are present below the typical Los
Rastros in the southwestern part of the basin, and that he has
collected reptiles of Chafiares type there. I have not \'isited this
area. Ortiz includes these beds in the Los Rastros Formation,
and if one does not wish to distinguish a separate Chafiares
Formation, one might include it in the Los Rastros — despite
the marked contrast in the nature of the sediments — but could
not, of course, consider these beds as part of the so-called
"Ischichuca."

Conformably above the Chafiares ash beds are the Los Rastros
sediments of shales, clays, and sandstones, with intercalated
carbonaceous layers, similar in nature to the beds of this forma-

1973 CHANARES SUMMARY 7

tion in the western part of the basin. Because of numerous
faults it is impossible to determine the thickness of the Los
Rastros in this region, but it is obviously much less than the
estimated 600 meters found west of the Ischigualasto Valley.

Only a limited exposure of Ischigualasto Formation sediments
is present in this region; the thickness observed is but 175
meters, as compared with 400-500 meters in the type area.
Abo\'e the Ischigualasto Formation are present Los Colorados
beds, only 95 meters thick; whether this is the total amount
originally deposited or whether they were originally thicker and
later reduced by erosion before deposition of overlying Tertiary
sediments is uncertain.

The Chanares Fauna

Below are listed the reptiles discovered in the 1964-65 expedi-
tion and described in earlier papers in this series. A few forms
are represented by fairly complete specimens; others are known
only from fragmentary materials. Much further collecting is
possible; one may hope that if and when such collecting can be
done, much better material of many of the forms already de-
scribed may be obtained and additions be made to the faunal
list:

Dicynodonts :

Chanaria platyceps
Dinodontosaurus brevirostris
Dinodontosaurus platygnathus
Kannemeyeriid indet.

Gomphodont cynodonts:
Massetognathus pascuali
Massetognathus teruggii
Alassetognathus major
Megagomphodon oligodens

, Carnivorous cvnodonts:
Probelesodon lervisi
Probelesodon minor
Probainognathus jenseni

Thecodonts :

Luperosuchus fractus
Lagerpeton chanarensis
Lagosuchus talampayensis
Lagosuchus lilloensis
Chanaresuchus bonapartei

8 BREVIORA No. 413

Gualosuchus reioi
Gracilisuchus siipanicicorum
Lewimchus admixtus

Dicynodonts. In contrast to the wealth of dicynodonts in the
later Permian, the group in the typical Triassic deposits is re-
stricted to a few forms of relatixely large size (their place as
herbivores appears to ha\e been taken oxer mainly by rhyncho-
saurs and gomphodonts). In the Chanares beds such forms are
present, but only in modest numbers, dicynodont specimens
constituting but perhaps 5 percent or so of the total of reptiles
collected. A few postcranial remains suggest the presence of a
kannemeyeriid ; apart from this, three types of dicynodonts are
present, all of which are assigned by Cox to the characteristically
Middle Triassic family Stahleckeriidae — Chanaria platyceps,
Dinodontosaurus platygnathus, and D. brevirostris. Chanaria is
a form not present elsewhere; howe\er, the Dinodontosaurus
species are quite similar to the genotypic form from the Santa
Maria Formation ( presumably of somewhat later age ) .

As also mentioned below, ecologic factors tend to separate
stratigraphically and topographically the three common herbi-
vore groups — dicynodonts, gomphodonts and rh) nchosaurs —
of the South American Middle Triassic fossiliferous areas. In
the Santa Maria beds, dicynodonts and rhynchosaurs are, so to
speak, "allergic" to one another; rhynchosaurs abound in the
deposits near Santa Maria city but are unknown in the two
other major fossil beds in this formation where dicynodonts are
abundant. At Ischigualasto all known dicynodonts ha\e been
found in a stratigraphically narrow band, about half-way up the
formation, and quite distinct from higher levels where gompho-
donts abound, and frotn lower levels where rhynchosaurs are
plentiful. Ill the Chanares beds, as noted abo\e, almost all
fossils are from the lowest part of the formation, but 1 ha\e the
impression that all dicynodonts collected were from the \ery
base, within a meter or two of the unconformity with the
Tarjados sandstones, whereas other types tended to occur up to
a dozen or so meters higher.

Gomphodonts. Gomphodont cynodonts arc the dominant
herbivores in the Chanares beds; more than half of all specimens
collected in the 1964-65 expedition were members of this group.
Nearly all clearly pertain to a single genus, Massetognathus. In
the first box of fossils received in Cambridge, Massachusetts,
there was present a considerable series of specimens that seemed
to sort out clearly into two size groups, and hence I descril^ed

1973 CHANARES SUMMARY 9

them as belonging to two species, M. pascuali and M. teruggii.
As I noted later, the full collection, when received, broke down
such a clear distinction. Dr. James Hopson tells me that in
primitive African cynodonts which he has been studying, a very
considerable size range is to be found; this suggests that M.
pascuali and M. teruggii merely represent populations of two
sizes of the same species. However, as my tables show, the size
distribution is heavily weighted above the peak that one may
reasonably believe to represent mature adults, and the presence
of two common species of Massetognathus is still a not unreason-
able assumption. Still further, the size range of specimens that
seem to belong to this genus is such that I find it impossible to
believe that the amount of growth necessary to reach the size of
the largest specimen can have been possible if a single species
(or even two species) had been present, and hence have with
some confidence given the name Massetognathus major to this
relatively enormous skull.

Nearly all the gomphodonts in the collection appear to be
reasonably assignable to a single genus. However, two rather
large individuals are clearly distinctive, and I have given the
name Alegagomphodon oligodens to this rare form.

The Chafiares gomphodonts are clearly members of the family
Traversodontidae, a group to which all known South American
gomphodonts belong ( and also forms present in the Manda beds
of East Africa). In the Santa Maria beds of Brazil gomphodonts
are less common, and are represented mainly by the genus
Traversodon. This genus may well have descended from Mas-
setognathus, but its remains are too poor to allow a detailed com-
parison. The Ischigualasto traversodontids are obviously much
more ad\'anced types.

Rhynchosaurs. Quite as significant as the presence of certain
forms in a given formation is the absence of expected types.
Most Triassic reptile faunas, except those of the very earliest and
very latest parts of the period, are notable for the presence of
rhynchosaurs, often in great abundance. In our Chafiares col-
lections there is not the slightest trace of a rhynchosaur ( despite
the fact that identifiable elements of this type of animal, most
especially upper tooth plates, are readily preserved and readily
recognized ) .

Why are no rhynchosaurs present? It is not because they had
not yet evohed, for although the Chafiares beds date from a
fairly early time in the Triassic, primitive rhynchosaurs were
already present in the Cynognathus Zone, definitely earlier, and

10 BREVIORA No. 413

were abundant in the Manda beds of East Africa, which (as
discussed later) are probably somewhat earlier than the Chaiiares
Formation. Quite certainly rhvnchosaurs had evolved bv the
time of formation of the Chaiiares beds and (although there is
no proof) may have been present in Argentina at that time.

Their absence here is quite surely, as I have suggested else-
where (Romer, 1973), attributable to some ecologic factor.
Rhynchosaurs and gomphodonts, in South American deposits at
least, seem to be basically incompatible.^ In the Ischigualasto
beds, rhvnchosaurs are exceedingly abundant in the lov/er part
of the formation, but in our 1964-65 expedition we found no
specimens in the upper half of the beds. On the other hand, on
our expedition we found gomphodonts to be very rare in the
lower part of the Ischigualasto Formation but very abundant in
the upper half of these deposits. Rather surely the contrast is
related to the type of plants present; the rhynchosaurs fed on
some type of plants having a hard-shelled "seed" for which the
"cracking" dentition of these forms was a necessity; the gompho-
donts, as the grinding character of their teeth and the absence
of a cracking device indicate, fed upon some different types of
plant materials. In the Santa Maria Formation, gomphodonts
are not as conspicuous as in the Ischigualasto and Chanares
beds, but such gomphodonts as are present there are absent in
the beds near Santa Maria city where rhynchosaurs alone are
present. If, as is probable, rhynchosaurs were present in South
AmxCrica in Chanares times, they would presumably have been
of a relatively primitive type, comparable to Stenaulorhynchus of
the Manda beds rather than the more advanced genus present
at Santa Maria and Ischigualasto.

Carnivorous cynodonts. In the Permian and earliest Triassic
the typical carnivores are therapsids; during the Triassic car-
ni\orous therapsids are reduced and disappear, to be replaced by
archosaurs (but giving rise to the earliest mammals before dis-
appearing completely). In the Chaiiares beds, thecodont archo-
saurs were becoming abundant, but carni\orous cynodonts were
still present and modestly abundant. They are interesting in
being more ad\'anced than Thrinaxodon and Galesaurus of the
earliest Triassic and without the somewhat specialized features
seen in Cynognathus, the common form in the Late Beaufort of
South Africa. Probelesodon lewisi is quite clearly ancestral to

'Charig tells mc, however, that there is no evidence for this in the Manda
beds of East Africa.

1973 CHANARES SUMMARY 11

Belesodon of the somewhat later Santa Maria beds; apparently
two species are present, P. lewisi, fairly common, and a smaller
form, Probelesodon yninor. More interesting is Probainognathus,
in which a starthng advance is the presence of a socket — a
glenoid cavity - — in the squamosal for attachment of the jaw.
This, however, is only a half-way stage in the development of the
mammalian system of jaw suspension, for this glenoid is for the
reception of an articular body of the lower jaw formed by a
fusion of the posterior elements of the reptilian jaw type; the
dentary bone, which in mammals articulates with the squamosal,
is as yet not quite in touch with the squamosal. The teeth of
Probainognathus are usually worn and show only the main
fore-and-aft row of cusps present in the teeth of primitive mam-
mals and seem to lack the row of basal "cusplets" found in early
mammals. For this reason it was thought for a time that Pro-
bainognathus could not be on the direct line of ascent to mam-
mals. However, Hopson has studied a little-worn dentition in
which these cusps are present and hence it may be reasonably
considered to be a true pre-mammal, or at least very close to the
actual ancestral line.

Thecodonts. Although carnivorous cynodonts still survived,
thecodonts were well on their way toward succeeding them as
dominant carnivores. In earlier years we knew little of this group
except for a few primitive forms in the Early Triassic and ( apart
from the specialized phytosaurs) only a few survivors in the
Late Triassic, where the thecodonts were already being succeeded
by the dinosaurs descended from them. One could have rea-
sonably assumed that were Middle Triassic beds well known,
the thecodonts would be discovered to be a varied group, with a
variety of forms leading in different directions — toward ptero-
saurs, bird ancestors, crocodilians and dinosaurs. Our increased
knowledge of Middle Triassic fossil deposits in recent decades
has gone far toward verifying this assumption, for although
many phyletic lines are far from clear, it is obvious that during
the middle part of the Triassic the thecodonts were undergoing
a rapid radiation into a wide diversity of types. The only large
Chaiiares form is Luperosuchus, represented only by an incom-
plete skull, which appears to be a member of the prestosuchid
(or rauisuchid) assemblage, of uncertain relationship. No close
affinities are known for Lewisuchus or the two small long-legged
types, Lagosuchus and Lagerpeton, represented mainly by hind
legs. Chanaresuchus and Gualosuchus are long-snouted, prob-
ably amphibious forms related to Cerritosaurus of the Santa

12 BREVIORA No. 413

Maria and Proterochampsa of Ischigualasto; once suggested as
crocodilian ancestors, the proterochampsids do not seem to be
related to that group, but are not impossibly related to the
phytosaur pedigree. A progressive form is Gracilisuchus, related,
it would appear, to Ornithosuchus of the later Triassic, which
has suggestive resemblances to primitive theropods, although it
is far from certain that the ornithosuchids are ancestral to these
dinosaurs. The Chanares thecodonts, as was stated, increase con-
siderably our knowledge of thecodont diversity, but as vet do
little toward establishment of any major archosaur evolutionary
lines.

Comparison With Other Faunas

As knowledge of Middle Triassic faunas has increased, ideas
as to the stratigraphic position and interrelations of these faunas
ha\e been expressed by a variety of workers, such as Bonaparte,
Colbert, Cox, Reig, and myself. I shall here merely consider the
interrelationships of these faunas from the point of view of the
Chanares assemblage. I have recently re\iewed the Triassic
faunas in a plenary paper (1972) for the Second Gondwana
Symposium, and hence full documentation here seems unneces-
sary.

As I pointed out some years ago (1966) Triassic faunas may
be roughly divided into three successive groups, (A) early,
(B) intermediate, and (C) late, although it is obvious that
such distinctions cannot be completely clear-cut, and transitional
assemblages may be expected. A-type faunas have long been
known from the Upper Beaufort beds of South Africa, contain-
ing mainly therapsids, although with early members of other
groups, notably thecodonts. C-type faunas are almost ubiquitous,
being known from redbeds Late Triassic deposits in Eiuope,
North America, South Africa, China, and (now) South Amer-
ica. In such faunas dinosaurs are already prominent, and their
thecodont predecessors are still present, whereas therapsids are
practically extinct (although the earliest mammals descended
from them have now appeared ) .

As to B-type faunas, these were until recently almost entirely
unknown, since deposits of Middle Triassic age in the northern
continents are mainly marine, and in South Africa the Molteno
beds, of Middle Triassic age, appear to be nearly barren of
fossils (although footprints are abundant). What should one
have expected in B-type faunas? Ob\'iously, a transition between

1973 CHANARES SUMMARY 13

A and C, with a gradual reduction of therapsids and an increase
in archosaurs, including a variety of thecodonts and the begin-
nings of the dinosaurs. The B-type faunas now known from the
southern continents do show these expected transitional features.
But, in addition, they show positive characteristics of their own,
in the great flourishing of gomphodont cynodonts and rhyncho-
saurs - — groups that had their beginnings in the A-type faunas of
the Early Triassic but seemed of little importance.

Let us first consider the South American situation. A-type
faunas are certainly present in the Puesto Viejo Formation and
not improbably in the Cacheuta series, as Bonaparte believes
(although the evidence is still scanty). The C-type is present
both in the upper part of the Los Colorados Formation, as now
being developed by Bonaparte, and in the El Tranquilo Forma-
tion. Between, we have in Argentina the succession Chanares-
Los Rastros-Ischigualasto, three formations that lie conformably
one above the other in the Talampaya basin. The Los Rastros
beds are almost barren of fossils, but it is, I think, generally
agreed that the Santa Maria Formation of Brazil is equivalent,
and thus, for vertebrates, our sequence may read Chafiares-Santa
Maria-Ischigualasto. All three clearly include B-type reptile
faunas. ^

The Chaiiares beds, earliest of the three, clearly are an early
part of the B complex. The gomphodonts are members of the
traversodontid family, and the diademodontids and trirachodon-
tid types present in the Scythian Cynognathus beds of South
Africa appear to be extinct. The carnivorous cynodonts are of
relatively ad\anced types — rather more advanced than Cyno-
gnathus. Rhynchosaurs are absent, but this, as noted above, ap-
pears to be due to some ecological factor, since primitive rhynch-
osaurs were already present in the A-type Cynognathus zone.
And, while few thecodonts were present in the Cynognathus
zone, thev are here alreadv varied in nature and in some cases at
least, of a progressive type.

The Santa Maria beds are quite surely later in age than the
Chafiares beds but, just as the presumably equivalent Los Ras-
tros beds lie in the break above the Chaiiares, the fauna of the
Santa Maria beds follows that of the Chafiares with some ad-
vances but without any major change. Among the dicynodonts,
Dinodontosaurus continues Httle changed into the Santa Maria.
Of gomphodonts, the Santa Maria Traversodon, although poorly
known, may well be descended with litde change from Mas-
setognathus. The Santa Maria carnivorous cynodont Belesodon

14 BREVIORA No. 413

appears to be but an enlarged edition of Probelesodon of the
Chanares. In both Chanares and Santa Maria beds, most of the
thecodonts are imperfectly known, but it is very probable that,
given more adequate material, several close comparisons may
come to be made, and Cerritosaurus of Santa Maria is very
similar structurally to Chanaresuchus of the earlier formation.
As Cox (1968) states, "the Chafiares fauna is only slightly
earlier than that of the Santa Maria." The only advance of any
note is that here (as might be expected) we have the first sign
of the evolution of dinosaurs from thecodonts in Staurikosaurus
Colbert and possibly the fragmentary materials described by
Huene as Spondylosoma.

Next above the Los Rastros Formation, without disconformity,
lies the Ischigualasto Formation, from which a very considerable
fauna is now known. The only dicynodont, Ischigualastia, is a
large form of no particular stratigraphic significance. Gompho-
donts of several genera — Exaeretodon, Proexaeretodon, Ischig-
nathus — are exceedingly abundant, especially in the upper part
of the formation. All are traversodonts that are more advanced
than those of the Chanares and Santa Maria beds. Carnivorous
cynodonts are rare and represented only by fragmentary remains
that ha\e been referred to the Santa Maria genus Chiniquodon.
Thecodonts are, again, fairly common and \'aried. Saurosuchus
is a relative of Luperosuchus of the Chanares but of larger size;
Proterochampsa is similarly a large member of the Chanare-
suchus-Cerritosaurus group. Triassolestes, originally thought to
be a dinosaur, is probably a thecodont, but perhaps a crocodi-
loid relative. Interesting is the presence of Aetosauroides, first
representative of a thecodont type that was to continue, ap-
parently little changed, to Late Triassic times. Of dinosaurs we
now have (although as rarities) the probable saurischians Her-
rerasaurus and Ischisaurus and, most interestingly, the oldest
known ornithischian, Pisanosaurus. Despite advances, we have
a close tie with the Santa Maria in that the common Ischigua-
lasto rhynchosaur Scaphonyx (thoroughly studied in an unpub-
lished thesis by Sill) is almost indistinguishable from the species
present in the Santa Maria. Chatterjee (1969) has suggested
that the Santa Maria localities containing Scaphonyx are later
than those containing the remainder of the fauna. But there is
no geological e\idence to support this suggestion; all the verte-
brate fossils, rhynchosaurs, dicynodonts and others, appear to
come from the relatively thin upper portion of the Santa Maria
Formation. In sum, the fauna of the Ischigualasto Formation

1973 CHANARES SUMMARY 15

is ad\'anced over that of the Santa Maria, but the difference is
not great, as Bonaparte has noted.

We lack any means of correlation of these South American
beds with the standard marine series, but since these faunas are
ob\'iously post-Scythian and pre-Norian, it is natural to suggest
a one-to-one correlation of Chanares-Santa Maria-Ischia^ualsto
with Anisian-Ladinian-Carnian. I have in the past expressed
doubts as to whether the horizon of the Ischigualasto Formation
was as high as the Carnian. In the European Keuper reptile
remains are known only from the upper, Norian, part of the
sequence and we have no knowledge of the reptile fauna of
Carnian times. Further, in the Ischigualasto Valley the Los
Colorados redbeds tower for some 400-500 meters above the
top of the Ischigualasto beds and, except for a single dicynodont,
our knowledge of the Los Colorados fauna is derived from the
vcv)^ topmost beds of this formation, so that it is possible that the
lower part of these beds are of Carnian age. However, consid-
eration of the faunas found in India and the northern continents
(discussed below) suggests that our B-type faunas continued into
Carnian days. It is thus very likely that the age of our B-type
Middle Triassic faunas conflicts with the classic division of the
Triassic into lower, middle and upper. Stratigraphically the
Middle Triassic includes Anisian and Ladinian, while the Upper
Triassic includes Carnian, Norian and Rhaetic ; as regards verte-
brates it is probable that the Middle Triassic includes Carnian
and Anisian and Ladinian as well, with the "upper" C-type
faunas restricted to the Norian and Rhaetic.

If one wishes to compare the Chaiiares and other South Amer-
ican B-type faunas with those of other continents, one naturally
turns first to South Africa, since current theories of continental
drift suggest that in the Triassic South America and Africa were
closely apposed to one another. If this was the case one would
expect similarities between the faunas of the two continents. But
even if the South Atlantic were then nonexistant, there would
remain a considerable distance between the Talampaya basin,
and even the Santa Maria region, and the fossiliferous beds of
east and south Africa. One should expect that there might be a
considerable difference between the reptile faunas of these regions
just as there is today a very considerable difference between the
reptile faunas of, for example, California and the Atlantic coast
areas of North America.

The African beds concerned are ( 1 ) the Molteno beds of the

16 BREVIORA No. 413

Stormberg Series cf South Africa, (2) the Ntawere beds of
Zambia, and (3) the east African Manda beds.

The Molteno beds are quite surely Middle Triassic in age and
should contain a fauna of the B-type. But while footprints are
tantalizingly abundant, actual fossils are rare, and such few as
ha\'e been described are of uncertain stratigraphic position and
may either come from the top of the Cynognathus zone (as in
the case of a cynognathid) or from the base of the redbeds (as
in the case of a traxersodont gomphodont ) .

The Ntawere beds are as yet not fully explored and as yet little
material has been described [cj. Cox, 1969). Two zones appear
to be present. The lower, in which Diademodon is present, may
well be equivalent to the upper part of the Cynognathus zone,
with an A-type fauna. The upper zone fauna includes two
dicynodonts — the stahleckeriid ^ambiasaurus and the kanne-
meyeriid Sangusaurus, two traversodont cynodonts, Luangwa
and a second form as yet undescribed, and fragments of theco-
dont«;. In default of fuller data, the age of this fauna is difficult
to determine. The presence of traversodonts suggests the B-type;
but tra\ersodonts occur at an Early Triassic age in Argentina
and may well ha\ e been as early in appearance in Africa.

Of especial interest is the Manda Formation of east Africa,
from which a very considerable fauna is known, owing to col-
lections made for Huene, by Parrington, and by an English ex-
pedition a decade ago. Unfortunately much of the known ma-
terial is undescribed or described in only preliminary fashion.
I am indebted to A. J. Charig for the faunal list given here.
There are three dicynodonts, Kannemeyeria, Tetragonias, and a
third undescribed form. No carnivorous cynodonts are as yet
described, but gomphodonts are numerous and \aried, including
the diademodontids Theropsodon and {?)Aleodon, the triracho-
dontid Cricodon and a \arietv of traversodontids of which the
only remains as yet described are assigned to four species of the
Q:enus Scalenodon. Some seven thecodonts have received names,
including the prestosuchids Mandasuchus and (?)Stag?iosuchus,
and fi\e further genera not assigned to families — Parringtonia,
Teleocrater, Hypselorhachis, Nyasasaunis and Pallisteria. The
abundant rhynchosaur remains pertain to the primitive genus
Stenaulorhynchus.

The abundance of gomphodonts and rhynchosaurs indicates
that we are dealing with a typical B-type fauna, and the presence
of Kannemeyeria and of diademodontid and trirachodontid
gomphodonts suggests a relati\'ely early age. The fauna is ob-

1973 CHANARES SUMMARY 17

\iouslv earlier than that found at Ischia^ualasto, and the Santa
Maria and Chanares faunas are the two South x\merican as-
semblages with which comparisons might reasonably be made.
On the whole, it is the Chanares fauna that seems to be the
closest. The absence of rhynchosaurs in the Chafiares beds re-
mo\'es one basis of comparison which might have been hoped
for. Not improbably some of the Manda thecodonts will show
aflfinities to Chafiares genera when fully described. Crompton
tells me that some of the Manda gomphodont specimens are
closely comparable to Massetognathus, but here again we must
await further publication. It is not unreasonable to expect that
when the Manda fauna is fully described it will prove to be
rather similar to that of the Chanares, but of a somewhat earlier
date.

In more northern regions — India, Scotland and Nova Scotia
— are assemblages that contain characteristic elements of the
B-type fauna but are usually considered as of Late Triassic age.
In the Maleri beds of India only three named tetrapods are
present. These are : ( 1 ) a stereospondylous labyrinthodont ge-
nerically identical with Metoposaurus, common in the Upper
Triassic of both Europe and North i\merica but otherwise un-
known in presumed "Gondwana" areas; (2) a phytosaur, diffi-
cult to assign to a given genus (the systematics of phytosaurs are
in a confused state) but representing a group unknown else-
where in "Gondwana" areas except in Morocco; (3) a rhyncho-
saur Parasuchus [Paradapedon] of an advanced type which
Chatterjee believes related to Scaphonyx of South America and
Hyperodapedon of Elgin. The presence of a metoposaur and
phytosaur in a supposed Gondwana region presents an interesting
geologic problem, but the question of the age of the Maleri is
almost equally interesting.

The Maleri is considered to be "Upper" Triassic; but while
"upper" in a stratigraphic sense, it may well represent a Carnian
fauna of our B-type. As regards phytosaurs, they are unknown
in Europe before the Norian, but this group obviously had a long
antecedent history (disregarding the question of the age of
Mesorhinus) . Metoposaurs, again, are "Upper" Triassic, but it
is not improbable that there may have been older antecedent
stages in the development of these peculiar stereospondylous
labvrinthodonts.

Rhynchosaurs, in the form of the advanced genus Hyper-
odapedon, are present in the Elgin beds of Scotland, which
Walker (1961) believes to be of Norian age. His conclusions

18 BREVIORA No. 413

mav be correct, and this mav mean a late survival of rhvncho-
saurs in Europe. But it must be pointed out that there is no
trace of a rhynchosaur in the Norian Keuper of continental
Europe, and hence it may be suggested that the Elgin beds are
pre-Norian, perhaps Carnian in age. The Elgin fauna is a sparse
one; there is nothing to represent the typical dinosaur fauna of
the continental Norian (the systematic position of Ornithosuchus
is questionable ) . Walker's correlation with the Norian is based
mainly on the presence of Stagonolepis, a close relative of
Aetosaurus of the continent. But we now know that the aeto-
saurid pattern was already present in the Ischigualasto beds in
the form of Aetosauroides [Argentinosuchus], which is still in-
completely known but appears to be a fully developed member
of this group.

Most interesting is the report by Baird (1962 and in litteris)
of the presence in beds in Nova Scotia which have been corre-
lated with the Newark series of the Atlantic seaboard of the
United States, of both of the most characteristic elements of the
B-type fauna — - rhynchosaurs and a gomphodont jaw ! The
Newark is a characteristically C-type series, as witnessed not so
much by the rare dinosaurian fossil remains as by the vast num-
bers of dinosaur footprints. Are we dealing in these Nova Scotia
finds with a very late sur\d\'al of gomphodonts and rhyncho-
saurs? Or — more probably, I think ^ — these supposed Newark
equi\'alents in Nova Scotia may, in their lower beds, extend
downward from Norian to Carnian age, into the time of exist-
ence of the B-faunas. Parenthetically, while the familiar red
Triassic deposits of the western United States — Chinle, Dockum,
Popo Agie — are usually considered as of quite Late Triassic
age, we find in them mainly metoposaurid amphibians and phy-
tosaurs, and little representation of the abundant dinosaurs found
in the European Norian, the redbeds of South Africa, the Late
Triassic of China and, apparently, in the Newark series proper.
Ls the nature of the faunas of these western beds associated
with ecological factors or are they of pre-Norian age?

References Cited

Baird, D. 1962. Rhynchosaurs in the late Triassic of Nova Scotia. Gcol.
Soc. Amer. Spec. Paper, 73: 107.

Beltrao, R. 1965. Paleontologia de Santa Maria e Sao Pedro do Sul, Rio
Grande do Sul, Brasil. Bot. Inst. Cien Nat. Univ. Fed. Santa Maria, 2:
1-114.

1973 CHAN ARES SUMMARY 19

Bonaparte, J. F. 1966. Chronological survey of the tetrapod-bearing Tri-
assic of Argentina. Brcviora, Mus. Comp. Zool., No. 251: 1-13.

. 1967. Comentario sobre la "Formacion Chanares" de la

cucnca Triasica de Ischigualasto-Villa Union (San Juan-La Rioja) .
Acta Geol. Lilloana, 9: 115-119.

— ■ 1972a. Annotated list of the South American Triassic

tetiapods. Proc. and Papers Second Gondwana Symposium (South
Africa, 1970) , Pretoria: 665-682.
. 19721^. Los tetrapodos del sector superior de la forma-

cion Los Colorados, La Rioja, Argentina. (Triasico Superior) . I Parte.
Opera Lilloana, XXIIL 1-183.

BoRTOLUZZi, C. A., AND M. C. Barbarena. 1967. The Santa Maria beds in
Rio Grande do Sul (Brazil) . Proc. Intern at. Symp. on Gondwana Strat.
and Paleont.: 169-195.

Cabrera, A. 1943. El primer hallazgo de terapsidos en Argentina. Notas,
Museo La Plata, 8, Paleont., No. 55: 317-331.

Chatterjee, S. 1969. Rhynchosaurs in time and space. Proc. Geol. Soc.
London, No. 1658: 203-208.

Cox, C. B. 1968. The Chanares (Argentina) Triassic reptile fauna. IV.
The dicynodont fauna. Breviora, Mus. Comp. Zool., No. 295: 1-27.

. 1969. Two new dicynodonts from the Triassic Ntawere forma-
tion, Zambia. Bull. Brit. Mus. (Nat. Hist.) , Geol., 17: 255-294.

Frenguelli, J. 1946. Consideraciones acerca de la "Serie de Paganzo" en
las provincias de San Juan y La Rioja. Rev. Mus. La Plata (N.S.) ,
Geol., 2: 313-376.

Grober, p. F., and P. N. Stipanicic. 1953. Geografia de la Rej)ublica
Argentina. Buenos Aires, H (Primera Parte) : Triasico: 13-141.

Ortiz, A. 1968. Los denominados estratos de Ischichuca como seccion
media de Formacion Los Rastros. Actas IH Jorn. Geol. Argentina, 1:
333-339.

RoMER, A. S. 1960. Vertebrate-bearing continental Triassic strata in Men-
doza region, Argentina. Bull. Geol. Soc. Amer., 71: 1279^1294.

. 1966. The Chanares (Argentina) Triassic reptile fauna. L

Introduction. Breviora, Mus. Comp. Zool,. No. 247: 1-14.

— ' . 1971. The Chanares (Argentina) Triassic reptile fauna. IX.

The Chanares Formation. Breviora, Mus. Comp. Zool., No. 377: 1-8.

1972. Plenary paper. Tetrapod vertebrates and Gondwana-

land. Proc. and Papers, Second Gondwana Symposium (South Africa,
1970). Pretoria: 111-124.

. 1973. Middle Triassic tetrapod faunas of South America. Act.

IV Congr. Latin. Zool. (Caracas, 1968), II: 1101-1117.
, and J. A. Jensen. 1966. The Chanares (Argentina) Triassic

reptile fauna. II. Sketch of the geology of the Rio Chanares-Rio Gualo
region. Breviora, Mus. Comp. Zool., No. 252: 1-20.

20 BREVIORA No. 413

RuscoNi, C. 1951. Laberintodontes Triasicos y Permicos de Mendma.. Rev.

Mus. Hist. Nat. Mendoza, 5: 33-158. "

Stipanicic, p. N. 1969. Las sucesiones Triasicas Argentinas. Gondwana

Stratigraphy, I. U. G. S. Symposium (Buenos Aires, 1967): 1121-1150.
Walker, A. D. 1961. Triassic reptiles from the Elgin area: SMgdnolepis,

Dasygnathus and their allies. Phil. Trans. Roy. Soc, London (B) , 244:

103-204. \ • ■■

B R E V I 0 R A

]ffV^^liii^Y^^^^™^P^^^*^'^^ Zoology

IAN 7 1974

US ISSN 0006-9698

Cambridge, Mas^. 28 December 1973 Number 414

UNIVfiRSlTt:

ECOLOGY, SELECTION AND SYSTEMATICS

Nelson G. Hairston^

Abstract. Three different kinds of ecological relationships between newly
separated species are examined, with the aim of establishing their expected
effects on the systematic differences between the species involved. In cases of
slight difference between the habitats of two products of recent speciation,
selection can be expected to favor specific competitive mechanisms, but
taxonomic differences would be expected to be slight, and examples of
hybrid superiority would be common. Where the habitats of the two species
are markedly different, as along a steep ecological gradient, adaptation to
the different places will result in species that become broadly overlapping
in habitat, and taxonomically different in many clearly adaptive characters.
Although this latter process leads to species with somewhat different food
habits, it would not lead to food specialization, even if the two species were
originally limited in abundance by food and in competition for it. True
food specialization, in the form of monophagy, is most likely to evolve in
the presence of a superabundance of several kinds of food, owing to in-
creased efficiency of handling, digestion and metabolism, and is improbable
among species in competition for food. Closely related monophagous species
should differ maikedly in a few characters, and hybrids should be inferior.
Examples of the three situations are described, plethodontid salamanders
being used for the first two and leaf-mining insects for the third.

Introduction

Classically, the relationship between systematics and ecology
has been approached by first taking systematics as the exploration
of genomic diversity, and then turning to ecology for explana-
tions that were secondary to the origin of differences. This
approach is epitomized by the recent comment to me that the
reproductively isolated entities within Paramecium aurelia could

^Museum of Zoology, the University of Michigan

2 BREVIORA No. 414

now be considered species because their isoenzyme patterns are
visibly different. Such a viewpoint surely gets the classification
much too far away from the biology. As an antidote, I propose
to examine the relationship from the standpoint that ecology
provides the set of opportunities that can be exploited by diversi-
fication of the genome. The approach is not original, as it is
the basis for the idea of adaptive radiation, but the impact of
ecology on systematics deserves reexamination. In this, we should
separate the passive background from the active; that is, those
factors that set the conditions, and those that are able and likely
to respond by evolving themselves. These two classes, unfortu-
nately, will not remain constant for us. For example, it would
be agreed that the distinction between nonliving and living parts
of the environment might provide such a preliminary classifica-
tion, but as far as I can discover, this is not the case. The dis-
tinction between the vegetation on one hand and the climate and
substrate on the other is clear enough. The physical gradients
provide the passive background, making physiological demands
on a potential additional plant species, and the various com-
peting species of plants provide the acti\'e counteradapting back-
ground, making ecological demands.

However, when we consider the active and passive background
of animals, particularly carnivorous ones, the distinction between
plants and the physical environment becomes less important than
the distinction between both of those on the one hand and other
animals on the other. Indeed, there are few cases of terrestrial
predators which are distributed concordantly with even the
dominant plants, and when this coincidence does occur, the
plants are used in a nonliving context, as when they are required
for nest sites.

This example provides the opportunity to emphasize the dis-
tinction between selection for physiological adaptation and selec-
tion in response to the ecological pressures of competition and
predation. It is to the latter to which I wish to address myself
principally, but I first give an example of the simultaneous opera-
tion of both. This will be followed by a description of what
seems to me to be an unusual opportunity to investigate the
ecological interaction between one species and several geograph-
ically varying populations of another, closely related one. From
that, I hope to be able to generalize some about a fruitful in-
vestigation of other kinds of systematic consequences of ecological
phenomena.

1973 ecology and systematics 3

^,: An Analysis of the Exploitation of a

: ' Undimensional Gradient

As has been emphasized by Dunn (1926), (Hairston, 1949),
Organ (1961) and others, the evolution of the Dusky Sala-
manders of the genera Desmognathus and Leurognathus is de-
scribable in terms of adaptation to a linear series of habitats from
aquatic to terrestrial.

This unidimensional array of pertinent physical environments
facilitates the analysis of each species' most immediate biological
environment: namely, its closest relatives.

My own early analysis showed that the coexistence of five
species was possible, when they used the entire physical gradient
from completely aquatic to terrestrial. The species involved are
Leurognathus marniorata, Desmognathus quadrainaculatus, D.
monticola, D. ochrophaeus and D. wrighti. The distribution of
the four species of Desmognathus is shown in Figure 1. With
no further information, however, it was not possible to determine
whether more species could be accommodated in this presumably
competitive series.

Some years later. Organ was able to provide a tentatively
negative answer when he investigated the ecological distribution
of the same four species of Desmognathus in an area where a
fifth species, D. fuscus, was found. He found that at nearly every
location, the maximum number of species present was four.
D. fuscus could coexist either with D. quadramaculatus at high
elevations or with D. monticola away from large streams at lower
elevations but not with both.

Thus, the limit imposed by the presumably competitive rela-
tionships seems to have been reasonably well established, but a
more detailed look at the data suggests that steepening of the
moisture gradient may reduce the number of species that can be
accommodated from the competitive standpoint. At high eleva-
tions, atmospheric moisture, however expressed, is as great far
from water as it is over a stream at low elevations (Hairston,
1 949 ) . This correlates very well with the combined vertical and
horizontal distributions of the two most terrestrial salamanders,
Desmognathus ochrophaeus and D. wrighti. D. ochrophaeus is
confined to a zone near streams at low elevations, none having
been found more than 15 feet from a stream at elevations below
3000', but its distribution is unrelated to surface water above
4500 feet. D. wrighti, with its distribution unrelated to water
in summer, apparently cannot compete with its congeners close

BREVIORA

No. 414

DISTRIBUTION OV DESMOGNATHUS

BLACK MOUNTAINS
(3000-6500')

NANTAHALA MOUNTAINS
(2300')

r~i

D. quadramaculatus

^ ■

t n r~i r ]

D. monficola

V7^

mn^

'/a

'//■

^■y,.

D. ochrophae

I
I
I

5-
9

10-
14

D. wrighti

15- 20- 25 +
19 24

□ q

a.llll

D. aeneus

0 I- 5- 10- 15- 20- 25+
4 9 14 19 24

NUMBER OF FEET FROM STREAM

Figure 1. The ecological distribution of the species of the salamander
genus Desmognathiis in two different mountain ranges in North Carolina.

1973 ECOLOGY AND SYSTEMATICS 5

to Streams at low elevations, and cannot persist away from
streams there because of the lower moisture.

It is therefore with some interest that one notes the coexistence
of four species of Desmognathus at low elevations (down to
2200 feet) in the Nantahala Mountains. D. wrighti does not
occur at low ele\'ations, but a study of the ecological distribution
of the genus shows the presence of a terrestrial species, D. aeneus.
This species, which is the size of D. wrighti, but more slender,
was found closer to streams than wrighti usually is in summer,
but clearly occupies the same general position at the terrestrial
end of the environmental gradient ( Fig. 1 ) . It seems anomalous
that it should be present, although D. wrighti is unable to occupy
the corresponding habitat at low elevations near its range. It
was postulated above that this inability is related to reduced
moisture at low elevations. This suggests that there may be a
climatic \'ariation that permits the existence of a low-altitude
terrestrial Desmognathus in the Nantahala Mountains. An
examination of rainfall records reveals that such is the case. In
the Coweeta Experimental Forest, the location of the distribu-
tional study, the average annual rainfall ranges from 75 inches
at 2240 feet to 93. inches at 3870 feet. This is appreciably higher
than the rainfall at comparable elevations elsewhere in the
Southern Appalachians. For example, at the foot of the Great
Smoky Mountains, Bryson City, N.C. has an average annual
rainfall of 52.12 inches. At the foot of the Black Mountains,
Montreat and North Fork have 53.61 and 51.78 inches respec-
tively, and between the Smokies and the Blacks, the French
Broad Valley receives from 38.45 inches at Enka to 47.61 at
the Asheville-Hendersonville x\irport.

Among other locations at comparable elevations in the South-
ern Appalachians, only the region from Brevard to Highlands,
N.C. receives as much rain as the general area south and west
of the Little Tennessee River. Comparable rainfall is found
elsewhere only at high elevations (71.20 inches at Mt. Mitchell,
6684' in the Black Mountains, and 81.71 inches at Clingman's
Dome, 6643' in the Great Smoky Mountains).

The end of the series of species seems to be determined by
climate, with high rainfall permitting the addition of a small
terrestrial species. On larger and higher mountains, when the
tops are (or once were) covered with conifer forests and rainfall
is high, the terrestrial species is Desmognathus wrighti, which is
confined to elevations above 3500 feet; in that part of the moun-
tains where the rainfall is high, even at low elevations, Des-

6 BREVIORA No. 414

mognathus aeneus occupies the terrestrial end of the series. In
other areas, the series stops with the third species, D. ochro-
phaeus. It does not appear possible for another species to enter
the series in the midle, as shown by the situation with D. fuscus
at \Vhite Top Mountain in Virginia. Competition thus seems
to determine how similar any pair of species can be and still
coexist. When the climate would require the next most terrestrial
species to o\erlap the habitat of D. ochrophaeus to too great an
extent, only three species are found.

This situation seems to present an unusually clear example of
the e\'olutionary exploitation of a simple environmental gradient
and of the limits of this diversifying exploitation that are set by
competitive interactions. The limits to "species packing" are
demonstrated as clearly as post-facto analysis could permit.

Moreover, it provides a miniature model for the early stages
in the e\'olution and diversification of the family Plethodontidae.

Post-Speciational Events :
Increased Competition or Coexistence?

The kind of analysis made in the preceding section differs
from large numbers of published descriptions only in being a
little more tidy than most. If the field is to progress, such state-
ments will become the beginning of studies at the interface of
ecologv' and systematics, rather than representing final conclu-
sions. The choice among investigations of ecological distribution
should depend upon the respective opportunities that they pre-
sent for experimental tests of hypotheses of systematic status or
ecological processes. One of the points which I wish to make
most strongly is that experimentation related to ecological inter-
actions can yield important information about evolutionary
events, provided that care is taken to select appropriately favor-
able situations for study. One such situation that seems to be
especially suitable for field manipulations is represented by two
species of Plethodon, an exclusively terrestrial genus of sala-
manders. The location is also the Southern Appalachians.

Plethodon jordani is endemic to the southern Appalachians.
Through much of its range, it is confined to higher elevations,
resulting in a fragmented distribution consisting of a number of
isolated populations, many of which are morphologically dis-
tinct from each other. These populations have been studied
repeatedly, and have been classified as belonging to as many as
four distinct species (Grobman, 1944). Whenever specimens

1973 ECOLOGY AND SYSTEMATICS 7

have been taken from intermediate locations, they are inter-
mediate in color between the adjacent different populations.
This discovery led to the eventual inclusion of all of these popu-
lations within Plethodon jordani and the recognition of seven
subspecies (Hairston and Pope, 1948; Hairston, 1950). The
subspecies are no longer recognized, largely because at least some
of the color characters are distributed independently of one
another. The situation as it is presently known is described by
Highton (1970, 1971) and by Highton and Henry (1970), who
add the electrophoretic patterns of plasmaproteins to the char-
acters for which distributional data are available.

Plethodon glutinosus is widespread throughout the eastern
United States. In the Southern Appalachians, it tends to occur
at lower elevations than those at which P. jordani does, and I
ha\'e suggested that the sharp altitudinal replacement of the two
species is the result of competitive exclusion (Hairston, 1949,
1 95 1 ) . Although easily recognizable color differences are known
for at least four geographically distinct parts of the P. glutinosus
population (Highton, 1962, 1970, 1971), the population in the
area discussed herein consists of only one of these. P. glutinosus
is thus morphologically more uniform than is P. jordani. The
above-mentioned altitudinal separation of the two species is not
the case everywhere, however. Over the southeastern part of the
range of P. jordani, the two species occur together over nearly
the entire range of altitudes available, indicating that competition
does not play a significant role in their distributions. This ob-
servation, reported by me for a few vertical transects (Hairston,
1951) has been confirmed and extended by Highton. The fact
that in this area P. jordani occurs at lower elevations and P-
glutinosus at higher elevations than elsewhere strengthens the
conclusion that in the areas of altitudinal replacement, there is
intense competition in the narrow vertical zones of overlap. It is
this geographical difference in ecological relationship between the
two species that provides an unusual opportunity to investigate
the phenomenon of competition in the field, and to obtain evi-
dence on the sequence of evolutionary events accompanying
competitive interactions between two similar species.

The above account is oversimplified from the taxonomic stand-
point. Over most of the area west of the French Broad River,
the two species are distinct, but Highton has found hybrids at
appropriate elevations on some of the mountains, and intergra-
dation is so extensive in the Nantahala Mountains that the local
form of P. jordani was once described as a subspecies of P.

8 BREVIORA No. 414

glutinosus (Bishop, 1941). Highton has called specimens from
intermediate elexations a hybrid swarm. Two detailed vertical
transects in the Southeastern Nantahalas at Coweeta Experi-
mental Forest show that simple explanations of the relationship
are unlikely to be satisfactory. The forest has two more or less
parallel roads that ascend to the top of the mountain. The roads
di\'erge slowly from the foot of the mountain at 2200 feet, being
a little more than one mile apart at 3200 feet and around two
miles apart at the points where they reach the top of the ridge at
4100 and 4500 feet, respectively. In October, 1971, a transect
was carried out along the more northern road, to be referred to
as the Shope Creek Road. The con\'entional expectation would
be of continuously increasing similarity to P. jordani and de-
creasing similarity to P. glutinosus with increasing altitude. The
comparison was made on the basis of color alone, no other
known character being of value in that part of the range. Four
different color characters are possible. P. jordani is character-
ized by red legs and a pale belly; P. glutinosus has extensive
white spotting, especially on the sides, and a black belly. A
population of P. jordani 10-15 miles to the east has extensive
brassy spotting on the back, as well as some white spotting on the
sides, but at present seems to be distributed discontinuously from
the Nantahala population. A few specimens from the transect
had brassy spots, but were too few to yield meaningful informa-
tion. Arbitrary scales were established to compare the relative
amount of red on the legs, white spotting, and darkness of belly
color. Six to 20 specimens were collected at each of 11 eleva-
tions from 2200 to 4300 feet. For each collection, an average
intensity of each character was established by five different ob-
servers, and the results pooled. The three characters changed in
exactly the same way along the transect. The results for two of
them are shown in Figure 2. The reversal of the expected trend
led to a transect of the southern road (Ball Creek) in 1972.
The results, shown in Figure 3, conform to the original expecta-
tion, but do not agree with the Shope Road transect, which was
repeated in 1972 with \irtuallv identical results to those obtained
in 1971 (Fig. 2).

Although the 3800-foot site is located on an east-west ridge,
the same is true of all higher sites, and no obvious vegetational
differences could be seen to account for the difference between
the transects — impressions confirmed in the records from 69
widely dispersed rain gauges (Dils, 1957).

\Vhate\er the eventual explanation for these anomalous data.

1973

ECOLOGY AND SYSTEMATICS

ALTITUDINAL VARIATION IN COLOR CHARACTERS IN PLETHODON
ALONG SHORE CREEK WATERSHED

Q
UJ

q:
o

3
O

1971
1972

2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200

ALTITUDE (feet)

— 1 1

1— T

•■' '■• \

p 1

1 :

\ \

1—

/ j

•

^ 2

/ 1

\ >

/ 1

'' 1

;

o 3

O— .

;

^,-^-'

. y

^,---r^

_..d'

V'"'

•

1 1

1 '

2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200

ALTITUDE (feet)

Figure 2. The vertical distribution of two color characters in the sala-
mander genus Plethodon along the Shope Creek transect in the Nantahala
Mountains in North Carolina. The scale for white spotting has been in-
verted because white spots are characteristic of the low-altitude species.

BREVIORA

No. 414

ALTITUDINAL VARIATION IN COLOR CHARACTERS IN PLETHODON
ALONG BALL CREEK WATERSHED

2300 2500 2700 2900 3100 3300 3500 3700 3900 4100 4300

ALTITUDE (feet)

2300 2500 2700 2900 3100 3300 3500 3700 3900 4100 4300

ALTITUDE (feet)

Figure 3. The vertical distribution of two color characters in Plethodon
along the Ball Creek transect, for comparison with Figure 2.

1973 ECOLOGY AND SYSTEMATICS 11

they reflect complications in the relationship between the two
species, and further in\'estigations may reveal or at least suggest
\'ery local selective forces.

The situation in the Nantahalas gives a strong indication of
close taxonomic relationship between P. glutinosus and P. jor-
dani, and is thus useful information in suggesting ecological and
especially evolutionary questions about the two species elsewhere
in the Southern Appalachians where hybridization is absent or
very rare.

Current ev^olutionary theory would explain the observed eco-
logical distributions in these other areas in the following manner :
assuming, as seems likely, that Plethodon glutinosus and P. jor-
dani share a common ancestor in the not very remote past, the
speciational event separating them left two species with adjacent
geographical ranges and very similar ecological requirements.
Plethodon jordani presumably occupied the southern part of the
Blue Ridge physiographic province, and the relevant part of
P. glutinosus occupied the adjacent part of the Piedmont prov-
ince. With a warming climate, glutinosus has invaded the val-
leys of the Blue Ridge province, but competition from jordani
has prevented glutinosus from extending its range to the tops
of at least some of the mountains, notably the Great Smoky
Mountains, the Black Mountains, and the Unicoi Mountains.
Throughout most of the rest of the area of common distribution,
one or both species have evolved into ecologically divergent
directions, with the result that competitive exclusion no longer
operates, and the two species coexist over a wide range of eleva-
tions. This situation would represent character displacement in
the use of some ecological requirement as yet unidentified. In
the areas of competitive exclusion, the vertical overlap of 200
feet represents the uncertainty of outcome of competition owing
to climatic variability, P. jordani being favored by cool, wet
years and P. glutinosus by the reverse conditions.

' Thus, in conventional theory and as far as numerous observa-
tions have revealed, we have the same two species coexisting in
some areas and in intense competition in others. Geographic
variation in color of P. jordani provides independent identifica-
tion of representatives from the two ecologically different popu-
lations, and this and other features make it feasible to undertake
experimental manipulations to test the accuracy of the interpre-
tations that I and others have made of the present distributions
of the local populations of the two species. This should be done
by reciprocal removal experiments and by exchanging numbers

12 BREVIORA No. 414

of Plethodon jordani between the two areas of presumably dif-
ferent ecological relationships. Inasmuch as they difTer in color
pattern, the introduced individuals and their descendents would
be readily identifiable for an indefinite number of years after the
start of the experiments.

The most obvious first test of the interpretations would be to
remo\'e each species separately from different plots in the differ-
ent areas where competition is and is not expected. If the in-
terpretation is correct, the remaining species should show a much
greater response in the area of narrow vertical overlap than in
the area of wide vertical o\'erlap.

Whatever the outcome of these simple removal experiments,
they would help resoh^e an implicit contradiction in ecological
theory. This is the conflict between the often used theory that
distributional overlap between closely related species implies an
appreciable amount of competition (Levins, 1968; MacArthur,
1968) and the converse that the same overlap implies that com-
petition is reduced or absent (Crombie, 1947; Hairston, 1951;
Brown and Wilson, 1956; MacArthur, 1972: 29 ff). This con-
flict is rarely stated overtly, but its resolution could have a pro-
found effect on ecological theory, including much that has been
written about niche breadths and community matrices.

The implications of the simple removal experiments are more
directly ecological than they are evolutionary. The combination
of ecological and systematic situations provides the opportunity
for more sophisticated experiments whose results could yield im-
portant insights into the recent influence of natural selection on
the direction of evolution in the several populations of Plethodon
jordani. These experiments would consist of reciprocal trans-
plants of populations of P. jordani between an area of narrow
o\'erlap and one of wide overlap. The subsequent changes in the
transplanted jordani populations and in the P. glutinosus popu-
lations newly exposed to the foreign jordani would re\eal the
direction of recent evolution with respect to interspecific com-
petition.

If P. jordani from the area of wide overlap survived in the
area of narrow overlap, and the P. glutinosus population in-
creased, the interpretation would be that in the area of wide
overlap, P. jordani has evolved so as to decrease its competitive
interaction with glutinosus. If P. glutinosus has evohed in the
same way, the reciprocal experiment should result in no change
in the glutinosus population, and it might result in an increase in
the jordani population introduced from the area of narrow over-

1973 ECOLOGY AND SYSTEMATICS 13

lap, because the jordani would not be meeting as much compe-
tition as it had been experiencing before the experiment.

Con\ersely, if the P. jordani transplanted from the area of
narrow o\erlap increases in the area of wide overlap at the ex-
pense of the local P. glutinosus, it would be necessary to conclude
that recent e\'olutionary history had produced a specialization
in jordani for some specific competitive mechanism.

A decrease in and eventual disappearance of jordani moved
from the area of wide overlap, combined with an increase in the
local glutinosus, would be interpreted to mean the evolution of a
specific competitive mechanism in that population of glutinosus.

The complete set of possible experimental outcomes and their
interpretations is given in Tables 1 and 2. Specifically omitted
from the tables are the highly necessary controls. For the re-
moval experiments, the only controls required are undisturbed
plots containing both species. The reciprocal transplantation of
populations of P. jordani will require elaborate controls. First,
one must be satisfied that the salamanders can be moved at all
and continue to thrive. This will require transplanting animals
within an area where their ecological relationships appear to be
constant. Assuming the success of such an experiment, it will
also be necessary to provide assurance that they are physiologi-
cally capable of existing in the remote area where the competi-
tive relations are presumably different. For this control, it will
be necessary to first remove both species from a plot and then
introduce the foreign jordani. Its survival would assure an
interesting result on those plots where it was introduced into
contact with glutinosus. The failure of any of these controls
would of course mean that the main experiment in reciprocal
transplantation of populations was a failure. This is a gamble
taken by anyone planning a controlled experiment.

If the controls succeed, the experiment should permit one to
choose with confidence between the following hypotheses: First,
that after speciation natural selection has favored ecological
diversification with resultingly greatly lowered competition and
a greatly increased area of coexistence; and second, that after
speciation and reinvasion, natural selection has favored the de-
velopment in at least one species of mechanisms to increase its
competitixe ability and thus exclude the congener from all or
nearly all of its range. The ability to choose between the two
hypotheses would greatly advance our ability to interpret sys-
tematic-distributional data from a large array of situations where
post facto conclusions are all that can be expected.

BREVIORA

No. 414

TABLE 1. The plan and possible outcomes with their interpretations of
experimentation in the area where Plethodon jordani and P. glutinosus over-
lap broadly in vertical distribution. All controls are described in the text.

MANIPULATIONS

OUTCOME

INTERPRETATION

Local glutinosus has a competi-

Disappearance

tive adaptation to foreign

of moved

jordani and local jordani has

jordani.

evolved ecological character
displacement.

(I)

Combined with a decrease in

Replace

abundance of glutinosus, means

with

that introduced jordani had

jordani

evolved a specific competitive

from area

Persistence

mechanism against glutinosus.

of narrow
overlap.

of moved
jordani.

(II)
Combined with constant gluti-
nosus population, means that

Remove
jordani.

local glutinosus has evolved eco-

logical character displacement.

Means that there was no

No change in

competition with jordani.

abundance of

Leave

glutinosus.

local

glut 171 OS us

Means that there was some

alone.

Increase in
abundance of
glutinosus.

competition at a low level.

Means that there was no

No change

competition with glutinosus.

in abundance

(Reciprocal of A 2 a)

Leave

oi jordani.

Remove

local

glutinosus.

jordani
alone.

h.
Increase in

Means that there was some
competition with glutinosus at a

abundance

low level. (Reciprocal of A 2 b)

of jordani.

1973

ECOLOGY AND SYSTEMATICS

TABLE 2. The plan and possible outcomes with their interpretations of
experimentation in the area where Plethodon jordani and P. glutinosus have
a narrow zone of vertical overlap. All controls are described in the text

MANIPULATIONS OUTCOME

Remove
jordani.

Remove

glutinosus.

Replace
with
jordani
from area
of wide
overlap.

2.
Leave
local

glutinosus
alone.

Leave
local
jordani
alone.

Disappearance
of moved
jordani.

Persistence
of moved
jordani.

No change in
abundance of
glutinosus.

h.
Increase in
abundance of
glutinosus.

No change
in abundance
of jordani.

Increase in
abundance
of jordani.

INTERPRETATION

Local glutinosus has a specific
competitive adaptation to all
jordani; glutinosus should
increase in abundance.

(I)
If glutinosus increases in

abundance or remains stable,

indicates that introduced

jordani has evolved ecological

character displacement with

respect to all glutinosus.

(II)
If glutinosus decreases, indicates
specific adaptation by area I
glutinosus to coexist with all
jordani; especially strong if
combined with A 1 b (II) of
Table 1.

Means that original hypothesis
of competition was false. Total
distribution pattern hard to
interpret. Expect other bad
results. Habitat disturbed?

Confirms original hypothesis of
competition. Should increase
more than in A 2 b of Table 1.

Means that original hypothesis
of competition was false,
especially with A 2 a. (Same
interpretation)

Confirms original hypothesis of
competition; jordani should
increase more than in B 1 b of
Table 1.

16 BREVIORA No. 414

Specialization and the Results of
Ecological Interactions

The e\'olutionary result of competitive interactions has been
the subject of a great deal of speculation, most of it stressing
specialization for different resources. This interpretation requires
scrutiny, since it implies that differential specialization is a prob-
able result of competition for resources, and the observation of
different food habits among coexisting related species has been
interpreted as a\'oidance of competition.

Such an interpretation, to be accepted even provisionally,
should require an examination of alternate hypotheses to explain
the observation. One such hypothesis that has not been explored
adequately, is that specialization carries advantages in efficiency
of handling, digesting or metabolizing the food, and that com-
petition need not be invoked at all. Thus, competition is easily
shown not to be a necessary condition for the evolution of food
specialization. The subject will be pursued to examine the ques-
tion of the sufficiency of competition as an explanation. If spe-
cialization for one kind of food is regarded as a derived state, as
either of the aboxe hypotheses assumes, then polyphagy must be
regarded as the starting point for any reconstruction. Assuming
that such is the case, and that the members of a species are ex-
periencing intraspecific competition for food, an individual of this
species which tended to specialize would be at a disadvantage
whene\er its specialty became scarce, since, in becoming a spe-
cialist, it would be expected to lose some ability to handle or
digest the remaining kinds of food. The only ways for such a
specialist to remain at an advantage would be to begin by being
so efficient at obtaining the special food as to overcome the
expected periodic scarcity, or else in some way to avoid the ex-
pected trade-off in efficiency with regard to other kinds of food.
The probability appears to be very low in either case. Thus, for
food-limited species polyphagy should be the rule.

With an initially polyphagous species that has a superabun-
dant supply of food, the situation is quite different. Any geno-
type increasing specialization is likely to be favored because of
the benefits of increased efficiency. No penalty is attached to
this tendency, because under the terms stated, none of the various
kinds of food is ever in short supply. Therefore, contrary to
routinely accepted theory, specialization for different foods
should be characteristic of species that are not in competition,
and the claim is hereby advanced that prior competition is

1973 ECOLOGY AND SYSTEMATICS 17

neither a necessary condition nor a suflficient one to explain the
coexistence of closely related species each specializing on a dif-
ferent food.

How is such a claim to be tested? One way would be the
laborious one of field experimentation testing for the means of
limitation of population size in a large series of related species,
some of which were monophagous and some polyphagous. If
the former are consistently limited through means other than the
supply of their food resources, and the latter show a consistent
tendency to be food-limited, the claim would be strongly sup-
ported. Rigorous proof of a series of events in evolutionary his-
tory is, of course, not possible, and in the present instance, even
if the experiments had the expected outcomes, the counterclaim
could always be made that the specialists had been released from
competition by becoming specialists and therefore would have
to be limited in abundance by some other factor.

A post facto test of the claim that food specialization implies
the absence of prior competition for food can be suggested in the
following manner. Among a number of species whose food is
well documented, there should be no particular relationship be-
tween the degree of specialization and the number of specialized
species per species of food. If, on the other hand, specialization
represents an evolutionary "escape" from competition for food,
the advantage gained should be reflected in a tendency to be the
only such species feeding on the food species in question. Thanks
to an extensive table by Needham, Frost and Tothill (1928),
this test can be made in the case of leaf-mining insect species.
There are 435 species of plants that serve as hosts. Of these 289
are fed on by only one species of leaf miner; 82 are fed on by
two species, and 64 are fed on by three or more species of leaf
miners. On the hypothesis that the distribution of the insect
species is by chance among the three groups of plant species, the
expected distribution can be calculated by tabulating for each
insect species its host plant species with respect to the number of
insect species that the host plant supports. Thus, for each spe-
cialist, only one plant species will appear in the table; for those
feeding on two plant species, both plant species will appear in
the table, and the same system continues for insects feeding on
three or more species of plants; each plant species will appear
separately in the appropriate part of the table. After the removal
of those records involving plants determined only to genus, and
prorating those appearing more than once in the table, there
remain 426 records of the plant species, classified according to

18 BREVIORA No. 414

TABLE 3. The number of species of plants attacked by varying numbers of
species of leaf-mining insects. The insect species have been separated ac-
cording to the specificity of their food habits. The figures in the table have
been calculated on the assumption of no relationship between the degree of
specialization of the insect and the number of species of insects supported by
its food plant (s) .

Number of species of insect per species
2 of host plant

U o - -

O a. U

^ w y:

«4-l — —

^ ^ -y.

^ (-1 ^
^ O y.

TABLE 4. The observed distributions of plant species for comparison with
the expected distributions in Table 3.

Number of species of insect per species
60 of host plant

O '> w

a- ~ '

or more

99.47

28.31

21.87

47.21

13.44

10.38

3 or more

136.41

38.83

30.00

*^ -y.

"ti a-

or more

94.00

37.50

18.31

48.00

11.00

12.10

3 or more

134.00

38.50

32.95

Z o

the number of insect species feeding on them. In the absence
of a relationship between specificity of feeding by the insect and
the number of insect species supported by the host, these 426
records should be distributed in the ratio 289 : 82 : 64 for each
group of insects : those found on one species of plant, those found
on two species and those' found on three or more species. The
expected distributions are given in Table 3.

If specialized species of insects tend to specialize on plant spe-
cies for which there is little competition, there should be an
excess of species in the first column for species with one host, and
a corresponding deficiency in the third column for the same row.
That such is not the case is shown in the observed distribution
(Table 4). Three of the specialists are confined to a plant spe-
cies that supports them and ten other species of leaf miners;
four are confined to a plant species that supports them and
eight other species of leaf miners. At the other end of the scale,
one species of leaf miner which lives on 37 different plant species
is the only species feeding on 19 of these plants. Thus, these data

1973 ECOLOGY AND SYSTEMATICS 19

provide no support for the hypothesis that specialization for spe-
cific food items arises as a direct result of interspecific competi-
tion, and the data do support the hypothesis that such specializa-
tion arises in the presence of ample food of various kinds. The
data, incidentally, are also consistent with other kinds of evi-
dence indicating that the terrestrial herbivore trophic level is
predator-limited as a whole (Hairston, Smith, and Slobodkin,
I960).

It is now worthwhile to examine the kinds of divergence that
would be likely under the selective force of interspecific compe-
tition. It is assumed, and will probably be conceded, that com-
petition is likely to be most intense between close relatives, here
interpreted as those most recently separated by speciation. It is
further assumed that newly separated competing species will be
in contiguous but largely nonoverlapping ranges. If the differ-
ences between the adjacent places were great enough, the pro-
cess of adaptation to the separate local conditions would be
likely to result in species that were different in many ways, in-
cluding the acquisition of different kinds of food, even if both
species were limited in abundance by their food supplies. Selec-
tion might now favor either of two quite different courses: the
production of competitive mechanisms specifically against the
neighboring species, or further divergence by each species in ob-
taining food in those parts of the others' range most like its own.
The first would sharpen the boundary between the two species,
as is the case with Plethodon jordani and P. glutinosus over parts
of their distribution; the second course would be expected to
lead to broadly overlapping but different ecological distributions,
such as are exemplified by the species of Desmognathus. These
two courses, as well as the third and noncompetitive course pro-
posed earlier, would have quite different consequences from the
standpoint of systematics. The continued highly competitive situ-
ation should result in few differences, and it is easy to imagine
situations in which hybrids would be at an advantage. The two
spdcies of Plethodon in the Nanthala Mountains may provide an
example. Where the species become differentially adapted to
place, it would be expected that many differences would be
favored, and that eventually these would become the large dif-
ferences that characterize higher categories. It would be easy to
place Desmognathus aeneus and D. quadramaculatus in different
genera, were it not for the existence of two species intermediate
between them in morphology. Finally, in the noncompetitive
situation, it might be expected that selection would produce few

20 BREVIORA No. 414

differences, but those would be ver\- distinct, and would be such
as to put hybrids at a severe disadvantage.

What is being suggested here is that an analysis of the sys-
tematic and distributional relationships provides clues to the eco-
logical forces that have been operating on the species in question.
In the case of one such situation, there has been proposed a series
of experimental tests designed to permit a choice among the eco-
logical and selectional events that led to the present systematic
relationships. Without such planned experiments, we are com-
mitted at best to accepting "natural experiments," the conditions
of which may be unknown to us, and which nearly always lack
the elements of controls and of experimental design that promote
definitive answers to specific questions. Manipulations will not
be possible for all situations, but if the different ecological causes
and their systematic effects that I have suggested can be con-
firmed for a few specific cases, predictive power would be added
to the simple analyses to which we are now confined.

References Cited

Bishop, S. C. 1941. Notes on salamanders with descriptions of several new
forms. Occ. Papers Mus. Zool., Univ. of Mich., No. 451: 1-21.

Brown. W. L., and E. O. Wilson. 1956. Character displacement. Syst.
Zool., 5: 49-64.

Crombie, a. C. 1947. Interspecific competition. J. Anim. Ecol., 16: 44-73.

DiLS, R. E. 1957. The Coweeta Hydrologic Laboratory. U.S. Dept. Agri-
culture Forest Service Southeastern Forest Experiment Station, Asheville,
N.C. ii + 40 pp.

Dunn, E. R. 1926. The salamanders of the family Plethodontidae. Smith
College Anniversary' Pubis, xii + 441 pp.

Grobman, a. B. 1944. The 'distribution of the salamanders of the genus
PletJiodon in the eastern United States and Canada. Ann. New York
Acad. Sci., 45: 261-316.

Hairston, N. G. 1949. The local distribution and ecology of the pletlio-
dontid salamanders of the Southern Appalachians. Ecol. Monogr., 19:
47-73.

. 1950. Iiucrgradation in Appalachian salamanders of the

genus Plethodon. Copeia, 1950(4) : 262273.

1951. Interspecies competition and its probable influence

upon the vertical distribution of Appalachian salamanders of the genus
Plethodon. Ecology, 32: 266-274.
, AND C. H. Pope. 1948. Geographic variation and spccia-

tion in Appalachian salamanders {Pletfwdon jordatii Group) . Evolu-
tion, 2: 266-278.

1973 ECOLOGY AND SYSTEMATICS 21
, F. E. Smith, and L. B. Slobodkin. 1960. Community

structure, population control, and competition. Amer. Natur. 94: 421-
425.

HiCHTON, R. 1962. Revision of North American salamanders of the genus

Plethodon. Bull. Fla. State Museum, 6: 235-367.
. 1970. Genetic and ecological relationships of Plethodon jor-

dani and P. glutinosus in the Southern Appalachian Mountains.

Pp. 211-241 in Th. Dobzhansky, M. K. Hecht, and W. C. Steere (eds.) ,

Evolutionary Biology, Vol. 4. New York: Appleton-Century-Crofts.
. 1971. Distributional interactions among eastern North Amer-

ican salamanders of the genus Plethodon. Pp. 139-188 in P. C. Holt
(ed.) , The Distributional History of the Biota of the Southern Appa-
lachians. Research Div. Monograph 4. Blacksburg, Va.: Virginia Poly-
technic Inst.

-, AND S. Henry. 1970. Variation in the electrophoretic migra-

tion of plasma proteins of Plethodon jordani, P. glutinosus, and their
natural hybrids. Pp. 241-256 in Th. Dobzhansky, M. K. Hecht, and
W. C. Steere (eds.) , Evolutionary Biology. Vol. 4. New York: Appleton-
Century-Crofts.

Levins, R. 1968. Evolution in Changing Environments. Princeton, N.J.:
Princeton Univ. Press, x + 120 pp.

MacArthur, R. 1968. The theory of the niche. Pp. 159-176 in R. C.

Lewontin (ed.) , Population Biology and Evolution. Syracuse, New York:

Syracuse University Press.
. 1972. Geographical Ecology. New York: Harper & Row.

xviii -f 269 pp.

Needham, J. G., S. W. Frost, and B. H. Tothill. 1928. Leaf-mining in-
sects. Baltimore, Md.: Williams and Wilkins Co. viii + 351 pp.

Organ, J. A. 1961. Studies on the local distribution, life history, and pop-
ulation dynamics of the salamander genus Desmognathus in Virginia.
Ecol. Monogr., 31: 189-220.

• I'i.: ■■■' .'■>' ;n^

r.ry: . ''*^^'

B R E V I O R A

Miiseiin:j^^j^f^£jim^a^ Zoology

■■^^BAfifeYr 0006-9698

Cambridge, MAS^^fiJ^f^I^ECj^i^^ER 1973 Number 415

THE EJW^fiJ^ON OF BEHAVIOR
AND THE RoKe'dF'bEHAVIOR IN EVOLUTION

M. MOYNIHAN^

Abstract. Modern behavior studies are, or should be, primarily concerned
with problems of causation. The immediate causes of particular behavior
patterns are being analyzed at the physiological and biochemical levels. The
ultimate causes, selection pressures, are being studied by ecologists and
ethologists. Unfortunately, there is little contact between the two lines of
investigation at the moment. Doubtless a new synthesis will be achieved in
the future. It does not, however, appear to be imminent. In the meantime,
the results of behavior studies in th^ field or in the laboratory in semi-
natural conditions can still be of use to the evolutionary biologist. They
may be most helpful in revealing the details, mechanics, of certain ecological
processes, which are themselves the regulators or determinants of evolutionary
events. Some examples from recent work on cephalopods, monkeys, and birds
may illustrate the sorts of data that are both available and relevant.

Introduction

I have been asked to talk about my own work on animal
behavior and related subjects, and also to say something about
possible further developments of behavioral studies in general.
The prospect of thus anticipating the future is not entirely grati-
fying. It seems to me that current research on animal behavior
has reached a difficult, awkward, almost embarrassing stage. As
is the case with any subject, there are numerous false starts and
unrewarding pursuits. Some questions being asked by workers
in the field are hardly worth posing. The answers are self-
evident or easily predictable. Some other questions are devoted
to more significant problems, but apparently cannot be answered
with the techniques currently available, at least not the tech-
niques actually being used. More important, the various kinds

^Smithsonian Tropical Research Institute

2 BREVIORA No. 415

of studies that are proving to be useful and successful are becom-
ing increasingly disparate in both methods and objectives.

This anomalous situation is, of course, the result of historical
factors. It might be instructive, therefore, to give a brief resume
of some aspects of the past, in order to explain the present unease
and to pro\'ide or re\ eal a reasonable rationale for some of the
continuing work — my own included.

Many biologists, the majority of evolutionary biologists and
"natural historians," probably would agree that the most stimu-
lating school of behaviorists in this century was that of the "ethol-
ogists." Ethology as such may be difficult to define. In theory,
the term could be applied (without paying too much attention
to its classical deri\ation) to the whole of the science of behavior.
In fact, it is usually restricted to a particular approach to the
subject, based upon Darwin (1872) and other pioneers such as
Heinroth (1911), Whitman (1899 and 1919), Huxley (1914),
and Craig (1918), and perhaps influenced by some early ideas
of Freud or his predecessors, but largely developed in continental
or Teutonic Europe in the 1930's and 1940's and subsequently
widely diffused, first in the English-speaking world and then
elsewhere in the next decade.

This school was distinguished by a concentration upon large
segments or sequences of behavior in natural or semi-natural
conditions, especially social (inter-indi\idual behavior and the
reactions that were called at the time "innate," i.e., species-
typical or (often by implication) species-specific. Among the
better known products of the school which may ser\^e to illustrate
its original range of interests were papers by Lorenz {e.g., 1931,
1935, 1941), Lorenz and N. Tinbergen (1938), N. finbergen
(1932, 1935, 1936, 1939, 1940), Makkink (1936), Kortlandt
(1940), Seitz (1940 and 1941), and Baerends and Baerends
(1950).

Another characteristic of the first ethological studies was a pre-
occupation with causes, not only long-term components such as
selection pressures affecting beha\'ior in the course of evolution
but also short-term or even immediate causes, external and in-
ternal states and stimuli and internal mechanisms producing
particular acts at particular instants in time. The latter interest
entailed a considerable amount of rather ambitious and detailed
model-building, the dc\elopment of concepts and terms such as
"Innate Releasing Mechanism," "reaction specific energy," "dis-
placement" activities, and "hierarchies" of instincts. The state of
the art at this stage is beautifully summarized in N. Tinbergen

1973 EVOLUTION AND BEHAVIOR 3

( 1951 ) . Unfortunately, most of the models proved to be descrip-
tive of the overt manifestations of behavior but not explanatory
or usefully predictive. They did not correspond very closely to
the actual e\'ents within a behaving animal. (This sort of dis-
crepancy between the perceived and the real is an occupational
hazard of model-building. There may be comparable gaps in
ecological models — a topic that will be mentioned later. )

The responses of ethologists to their logical and methodological
difficulties were exceedingly diverse :

1. The original mainstream of effort was impeded and re-
duced but did not dry up completely. There were hopeful and
ingenuous attempts to redefine and refine the classic concepts
(see, for instance, Bastock et al., 1953; Hinde, 1954a and 1954b;
Morris, 1957; Blest, 1961). Some of these attempts may have
been helpful in minor ways, but I think that it would be fair to
say that they did not do very much to resolve the basic dilemma.
There was a push to render descriptions more precise, by adop-
tion of mathematical and pseudo-mathematical means of nota-
tion, often with an infusion of information theory and cybernetic
terminology, and by increased* use of improved photographic
and other kinds of recording equipment. Examples are too nu-
merous to cite, but many can be found in recent issues of the
journals "Behaviour" and "Animal Behaviour" and the bibli-
ographies of the general surveys of Hinde (1970), Eibl-Eibesfeldt
(1970), and Marler and Hamilton (1967). All too often, they
have merely told us what we already knew or assumed, at dis-
tressingly greater length and elaboration than we were prepared
to cope with.

2. Perhaps a more practical response was switching of atten-
tion to groups of animals and special problems that had been
neglected in earlier years. Several bends in the river or new
channels which are in some danger of becoming oxbows but are
at least picturesque. There has been a great deal of strictly
etholoarical work on a variety of "lower" mammals such as mar-
supials, rodents, and carnivores [e.g., Kaufmann, in press; Klei-
man, 1972; Leyhausen, 1956; Kruuk, 1972; Schaller, 1972;
Ewer, 1963, 1968, and 1973), and an enormous proliferation of
studies and surveys of primates {e.g., Altmann, 1967; Chance
and Jolly, 1970; Crook, 1970; DeVore, 1965; Dolhinow, 1972;
Imanishi and Altmann, 1965; Jay, 1968; Jolly, 1966 and 1972;
Kummer, 1968 and 1971; van Lawick-Goodall, 1971; Morris,
1967a; Movnihan, in press a; Fetter, 1962; Poirier, 1972; Rey-
nolds, 1968; Rosenblum and Cooper, 1968; Rowell, 1972;

4 BREVIORA No. 415

Schaller, 1963; Struhsaker, 1969). Many of these papers were
indirect reflections of a strong interest in human beha\ior, both
as it is and as it may be supposed to have been at some earlier
time in the Pliocene or Pleistocene; and there have also been
attempts to apply conventional ethological insights to some of the
urgent problems of modern man {e.g., Lorenz, 1963; Russell
and Russell, 1968; Morris, 1967b; Martin, 1972) with amusing
results (critics have tended to dismiss both the good and bad
suggestions and interpretations as impertinent sensu stricto, but
it may be hoped that some of them will eventually be incorp-
orated into the intellectual background of the well-informed
citizen ) .

The most fashionable of the special subjects has been what
might be broadly called "communication." Different aspects of
the subject ha\'e been tackled at many different levels and in
many different areas. There have been analyses of the various
ways in which information, true or false, can be transmitted
among individuals of the same or different species, and also of
the means by which transmission can be prevented or inter-
rupted. One of the aspects of interspecific communication that
has attracted investigation and speculation is mimicr\ , not onh'
the long known Batesian and Mullerian types but also aggressive
and social and e\'en more recondite forms. Relevant publications
include Brower et al. ( 1 960, and many other papers from the
same school); Rand (1967); Robinson (1969); Moynihan (in
press b), and an extensive discussion and summary in Wickler
(1968). The methods by which predators discover and recog-
nize prey, with or without the baffles of mimicry and crypsis,
have been studied by many workers. The papers of Robinson
and his collaborators {e.g., 1969, 1971a, 1971b) reveal some of
the factors that may corne into play. Research on intra-specific
communication has been primarily concerned with the e\en more
variegated "languages" used in more complex social situations
("social" in the ever\^ day sense of the term). It has involved
description, decipherment, and efforts to detect and formulate
the general rules, the "grammar and syntax," of a multiplicity
of signal systems. There have been sur\'eys and comparisons of
the signals of different groups of animals {e.g., Tembrock, 1959;
Lanyon and Tavolga, 1960; Busnel, 1963; Sebeok, 1968), some-
what abstract discussion of theorv {e.g., W. J. Smith, 1965 and
1969; Moynihan, 1970; Cullen, 1972; Mackay, 1972), and
detailed accounts of particular systems, ranging from the phero-
mones of insects {e.g., the work of E. O. Wilson and his col-

1973 EVOLUTION AND BEHAVIOR 5

leagues) through bird "song" {e.g-, Thorpe, 1961 ; Hinde, 1969)
to the non-\erbal movements and expressions of children and
adults in contemporary western and other human societies {e.g.,
Gofifman, 1971; Blurton Jones, 1967 and 1972; Argyle, 1972;
Eibl-Eibesfeldt, 1972). These studies may have implications for
related fields. They have, for instance, at least made available
to "real" linguists such as Chomsky, Lenneberg, etc., some useful
background material and evolutionary perspective.

3. However valuable such works may be, they would appear
to be di\Trsions from the classical behavioral point of view.
Most active students are proceeding, and probably will continue
for the foreseeable future, in one or the other of two different
directions, two new mainstreams. Those who are preoccupied
with immediate causes are going into physiology in earnest, lab-
oratory research on hormones, nerve cells, receptor organs, at the
deepest or lowest, even molecular, level. I cannot say anything
about this. Results are obviously flowing in, but the subject is
complex and not my major interest and I am not competent to
discuss it.

4. Ethologists who are more concerned with ultimate causes
are exploring connections or interfaces among behavior, ecology,
and evolution.

This has been my own preference. I may, therefore, be able
to illustrate sonle of the positive virtues and negative drawbacks
of the approach by citing particular cases from my own experi-
ence. In recent years, I have been engaged in observation and
analysis of three groups of animals, cephalopods. New World
primates, and passerine birds (and some "near passerines" such
as hummingbirds ) , in the field in natural or semi-natural condi-
tions.

Examples

1. I was attracted to cephalopods for several reasons. They
provide remarkable examples of evolutionary and ecological
convergence. Beginning with a molluscan body plan, they have
acquired large size, good eyes, large brains, and (in many spe-
cies) active and predatory habits. They have become similar to
many fishes and other aquatic vertebrates in these respects. (The
convergence is discussed at length in Packard, 1972.) They
have also evolved unique or peculiar characters such as distinc-
tive methods of buoyancy control, color changes, and jet pro-
pulsion. Combinations of some of these features have finally

6 BREVIORA No. 415

allowed them to invade the laboratory, to serv^e the neurophysi-
ologist. I would say, without being an expert, that some of the
operations of their central nervous systems and their handling of
visual information must be better known than the corresponding
processes of any other animals with the possible exception of
man. See, for instance. Young (1964 and 1972), Wells (1962),
and the many papers of Sutherland and his co-workers.

In these circumstances, it is noteworthy that the social be-
havior of cephalopods has not been studied in anything like the
detail that might, off-hand, have been expected. (There are
technical reasons for this comparative neglect. Most cephalopods
do not li\'e long in captivity and/or are difficult to follow in the
field.) Such work as has been done on the subject has been
une\'enly distributed. The great majority of living species of the
class can be assigned to one or the other of three diversified and
flourishing orders. Using the terminology of Jeletzky ( 1 966 ) ,
these may be called Teuthida (including the squids), Sepiida
(cuttlefishes and their relatives), and Octopida (octopi and
argonauts ) . There are more or less lengthy published accounts
of the social behavior in the laboratory of the common European
cuttlefish. Sepia officinalis (L. Tinbergen, 1939; Holmes, 1940),
and the common octopus. Octopus vulgaris (e.g., Packard and
Sanders, 1971; Wells and Wells, 1972), but relatively little on
other species, only bits and pieces on some reactions of a few
other sepiids and octopi and several kinds of squids, mostly
Loligo spp., in the laboratory or in the field (see references in
Lane, 1957, and Moynihan, in press b).

I was delighted, therefore, to encounter a species of squid,
Sepioteuthis sepioidea, in the San Bias Island region of the At-
lantic coast of Panama which is quite unusually easy to observe
in the wild under natural conditions. Mr. Arcadio Rodaniche
and I seized the opportunity to look at its social behavior. We
have now been observing it at monthly intervals for over two
years.

The species occurs inshore in moderately or very shallow
waters over turtle grass and coral. It is often extremely abun-
dant. It is a true squid, but rather cuttlefish-like in shape,
adapted for "hovering," and much less rapidly or continuously
mobile than most other squids (see also Boycott, 1965). It is
both predator, eating small fishes and crustaceans, and prey,
being eaten by large fishes such as barracuda and snappers (and
perhaps many other animals, including birds, Brown Pelicans,
etc.). Individuals of the species tend to scatter singly or in pairs

1973 EVOLUTION AND BEHAVIOR 7

or trios to hunt more or less actively at night, but they congregate
in large groups in the daytime to wait for prey to come to them.
The daytime groups may be almost completely stationary for
long (several hour) periods. Even when they are less sluggish,
they tend to keep within rather small territories or home ranges.
Groups are easily habituated to the presence of human observers.
( In fact, one of the few technical problems of working with the
species is to keep from getting too close to retain perspective and
an overall view.) Individuals in groups are not shy about per-
forming a variety of elaborate social reactions, including the full
range of "courtship" and copulatory patterns, before human ob-
serv^ers. Thus, they have provided us with a superfluity of data.

What have been the results?

In one sense, they have been disappointingly conventional.
The social behavior of Sepioteuthis is essentially vertebrate-like
in basic articulation and organization. There do not seem to be
any general principles of molluscan behavior apart from those
shared by most other complex animals of other phyla. But this
squid does exhibit or illustrate a whole series of interesting special
adaptations which may be correlated with, causally related to,
one significant aspect of its ecology — and many of which may
also be characteristic of other cephalopods and for the same
reasons.

S. sepioidea populations are highly structured. Not only do
individuals repeatedly leave and rejoin groups, but even the
groups are formed of sub-groups which may be separate at some
times, with obvious hostility and territorial defense among them-
selves, yet completely integrated at other times. There also are
size and (presumably) age classes that assort themselves in par-
ticular spatial arrangements according to particular temporal
and physical circumstances. The system is both intricate and
flexible, apparently at least as much so as those of such mam-
malian carnivores as lions, African hunting dogs, and Spotted
Hyenas.

The system is mediated by signals, both ritualized (mostly
displays) and unritualized. As far as we can tell, all the signals
are visual. (Cephalopods seem to be deaf, and we did not
detect, see, any indications of the use of pheromones or other
means of olfactory communication.) The visual signals include
postures and movements and many color changes. The number
of ritualized patterns is quite high. The basic components of
the ritualized repertory may not be more numerous than the
corresponding elements in the repertories of certain birds and

8 BREVIORA No. 415

fishes (see Moynihan, 1970), but they can be combined and
recombined almost endlessly. It is not uncommon to see an
animal adopt two or three, even four or five, color patterns
simultaneously, each color on a particular part of the body,
while performing a series of movements, especially of the fins or
arms, in very rapid succession. The effect is Protean. A squid
is quite able to transmit a variety of different signals in difTerent
directions to difTerent recei\'ers, different kinds of onlookers, all
at nearly or completely the same times. As visual signal systems
go, the cephalopod versions must be unique in their combinations
of speed and diversity or multiplicity and perhaps efficiency.

Comparison of the known patterns of Sepioteuthis, Sepia,
Octopus, and some other cephalopods has revealed some sugges-
ti\'e similarities and contrasts. Some displays are very distinct,
obviously not homologous, in the different species. Others are
very similar. Some of these are relatively simple. They may well
have become ritualized independently in each of the phyletic
lines. But at least four major displays are both extremely com-
plex, exaggerated, and "unexpected," and yet strikingly similar
in many details (of causation and function as well as form) in
the \-arious species. These displays would appear to have be-
come ritualized before the lines diverged from one another. As
the divergence must have occurred well before the end of the
Mesozoic, perhaps most probably in the Late Triassic, the pat-
terns are not only old but also have been remarkably conservative
during evolution. To my knowledge, they ha\e been more con-
servative than any patterns of other groups so far recorded in the
literature. One of the reasons whv some or all of them have been
stable is apparent when they are compared with the other dis-
plays of the same species that have changed more considerably
or de\eloped more recently. The latter tend to be shown to only
a few individuals or types of individuals. The conservative sig-
nals, on the other hand, are designed to influence a great number
and di\ersity of receivers, different age, size, and sex classes of
the same species and/or individuals of other species, especially
potential predators. This may be a general rule, applicable to
most animals. All other things being equal, the more widely
reflected or broadcast a signal, the more conservative it will be,
the more narrowly reflected or broadcast, the more Ukely it is to
be changeable in evolutionary time.

The role of predation should be emphasized in connection
with cephalopods. There is good evidence (see Moynihan, in
press c) that several or many of the living members of the class

1973 EVOLUTION AND BEHAVIOR 9

are favorite prey of marine birds and mammals almost through-
out the seas and oceans of the world. They must, therefore, be
themselves enormously abundant in many areas. (Common as
it is, Sepioteuthis has a fairly restricted distribution in the tropical
Atlantic. Other squids must have larger populations. The total
numbers of cephalopods in any given area are difficult to esti-
mate precisely, as many species are nocturnal and most are diffi-
cult to catch with the traditional gear of marine biologists, but
the birds and mammals probably are more efficient collectors.)
There also is evidence that the enormous biomass of cephalopods
is di\ided among fewer "packets," i.e., species, than is that of
their nearest competitors, the marine fishes. This could be both
cause and consequence of their relatively greater attraction for
predators.

It may be assumed that many of the extinct cephalopods ex-
hibited some or all of the demographic and ecological charac-
teristics of their living relatives. If so, it seems likely that preda-
tion pressure could have been the major impulse for a series of
evolutionary events. Some of the probable steps can be listed
briefly and crudely. The ancestors of the majority of living
cephalopods presumably reduced, internalized, and in some cases
lost, their originally external shells to gain greater maneuverabil-
ity and powers of escape. This "freed" their skin for other uses,
including the elaboration of color change mechanisms. The de-
velopment of gregarious habits may well have been another
(even earlier?) anti-predator adaptation (Brock and Riff en-
burgh, 1960). The habit of living in groups puts a premium
upon the development of complex signal systems. For vulnerable
marine animals, a visual communication system has definite
adxantages. (Visual signals can be turned off instantaneously
whene\er necessary or desirable, unlike olfactory cues, and they
are perhaps less apt to be noticed at a distance by dangerous
receixers than are acoustic signals, especially in murky waters or
around reefs or vegetation. And, of course, short range signals
are perfectly adequate as long as the animals are close together. )
Once the skin has become speciaHzed for color changes, it prob-
ably is not easily transformed for other purposes such as the
development of new kinds of armor or spines. This restricts the
choice of further anti-predator adaptations. It has already been
mentioned that whatever displays may have to be shown to
potential predators are conservative. As many or most of these
patterns are also used in intraspecific encounters, they may tend
to impede fundamental changes in the type, although certainly

10 BREVIORA No. 415

not the details, of the signal system as a whole. Other char-
acters of cephalopods such as their rapid growth, relatively short
life spans, special arrangements and care of eggs (see, for in-
stance, Packard, op. cit., and Wells, op. cit.), and even their
preference for reproducing only once in a lifetime, in "big bangs"
(Gadgil and Bossert, 1970), could also be explained as responses
to intense predation. (And the need to synchronize reproductive
moods in a hurry, without much time for trial and error, must
add another premium for both gregariousness and the elabora-
tion of signals.)

The series is an illustration of some of the ways in which
ecology and beha\dor can interact to determine the course of
evolution, each step opening up some possibilities and foreclosing
others.

2. The New World primates are a variegated family of mon-
keys of some 11 to 13 genera and many species. I have obser\ed
representatives of all the genera at irregular intervals over 15
years. Some species have been observed only in captivity, at the
field station on Barro Colorado Island and in zoos in \Vashing-
ton, London, Paris, and Amsterdam; but many others have been
studied at considerable length in the wild, in the central part of
the isthmus of Panama, to the west in the province of Chiriqui,
and to the south in the upper part of the Amazon basin, in the
Caqueta and Putumayo regions of Colombia.

For most biologists, the primary significance of the American
monkeys is that they represent a wide and independent adaptix^
radiation. They have occupied most of the habitats available to
primates. In this respect, they are more or less strictly equivalent
to the two other radiations of modern primates, the (Recent and
Pleistocene) lemuroids of Madagascar, and the so-called Old
World monkeys and apes, the "Catarrhini," of tropical Asia and
Africa and some adjacent areas, of which man is a specialized
offshoot. The New World forms may thus provide a useful
check to hypothesis and speculation about the evolution of pri-
mates in general and man in particular. I should also like to
claim that they are interesting in themselves.

They range from very small (the Pigmy Marmosets of the
genus or sub-genus Cebuella) to moderately large (the howlers,
Alouatta, and the spider monkeys, Ateles). They show a great
diversity of types of locomotion, from squirrel-like scrambling
and/or vertical clinging and leaping among the marmosets and
tamarins {Saguinus, Leontideus, Callirnico, and Callithrix in
addition to Cebuella), through quadrupedal "springing," walk-

1973 EVOLUTION AND BEHAVIOR 11

ing and pacing in such forms as Saimiri and Cebus, to brachia-
tion or semi-brachiation with the supplementary use of a pre-
hensile tail in Ateles. (The classification and details of locomo-
tion are discussed in Erikson, 1963, and Napier and Walker,
1967.) At least two species of Cebus, capucinus and apella,
come down to the ground with appreciable frequency. All or
most of the species of other genera are thoroughly arboreal. One
genus, Aotus, is nocturnal; the rest are diurnal. They all tend to
be nearly omnivorous on occasion ; but most of the smaller forms,
many of the tamarins and probably the marmosets of the genus
Callithrix, seem to prefer insects whenever they can get them,
while some of the larger forms are essentially herbivorous, taking
various assortments of fruits of particular kinds and ages, as well
as buds and leaves and even twigs and bark. At least one form,
Cebuella, has specialized in sap-sucking. (The sap-sucking is
described in Moynihan, in press d. The best general accounts of
more conventional feeding habits and regimes, unfortunately lim-
ited to the Panamanian species, are in Hladik and Hladik, 1969,
and Hladik ^^ fl/., 1971.)

In the course of my own studies, I have attempted to discover
and analyze the social behavior and structures of different spe-
cies and combinations of species, to determine how such com-
plexes are held together (or apart as the case may be), and to
identify some of the selective forces involved, to tie the observed
behavior to particular aspects of ecology. The results sum-
marized below are taken from Moynihan (in press a) ; this book
also lists references to papers and unpublished notes of other
workers.

Two extreme types of intraspecific social organization can be
recognized without much difficulty: the restricted "nuclear"
family group and the large band. The former seems to be the
basic social unit of Aotus, Callimico; two species of Callicebus,
moloch and torquatus; and, in some circumstances, Pithecia
monacha. Bands are characteristic of Pithecia melanocephala,
Alouatta villosa, Alouatta caraya, Lagothrix, Saimiri, and some
or all forms of Cebus and Ateles. As might be expected, there
are intermediate conditions, complications, and exceptions. One
type of intermediate is the "extended" family of some species of
Saguinus, e.g., juscicollis, graellsi, midas, and Cebuella and prob-
ably many other marmosets. Intermediates can also be flexible,
intermittent or recurring. Small families of some species may
join one another in some circumstances. It also is normal or
usual for neighboring small families of most species to perform

12 BREVIORA No. 415

certain responses, e.g., anti-predator reactions, in common. (This
is evidence that they do form a real social community.) Con-
versely, large bands may split up into smaller sub-groups tempo-
rarily, or reveal traces of sub-group organization within the
bands without actual splitting. This appears to be most common
in Saimiri and some form of Ateles. (The sub-groups are not
usually families but rather cephalopod-like age and sex classes.)

The adaptixe value of such variance is surprisingly obscure.
It seems to be characteristic of American monkevs that there is
little general correspondence between basic types of intraspecific
organization and either habitat or food preferences. There are
species that li\'e in bands and species that live in small family
groups among the primarily or exclusively vegetarian forms.
There also are both kinds of species, or at least forms that usually
live in bands and forms that live in extended family groups,
among the animals that prefer insect food when available. The
proportions of highly to poorly gregarious species and individuals
are much the same in many of the stages of succession from
young second-growth scrub to mature forest in many areas. Per-
haps even more remarkable, density of populations also appears
to be largely irrelevant in this connection (if not for other as-
pects of social behavior — ^ see below). Both Callicebus moloch
and Saimira usually are abundant and concentrated where\'er
thev occur. Thev are concentrated in different v/avs, but the
average number of individuals per unit of time and area may be
high in both cases. Both Aotus and Cebus albijrons can be
described as dispersed. The albijrons occur in rather large bands,
but the bands themselves are scattered.

These facts would suggest that almost any type of social or-
ganization can permit or facilitate almost any kind of exploita-
tion of the environment within the range of niches occupied by
American monkeys at the present time. Presumably, because
most of them are more "generalists" than "specialists," they have
been able to choose among alternative strategies to achieve
similar ends.

Much more restricted are the modalities or techniques by
which particular social systems are maintained. The ritualized
signal systems of these animals are not only adaptive but are
quite obviously so, down to the finest details. They include vis-
ual, acoustic, olfactory, and tactile patterns (Moynihan, 1967).
Of these, the visual and acoustic seem to be usually most im-
portant. The basic elements, the deep structures, of the repertory
of sounds may be nearly identical in all species, with the possible

1973 EVOLUTION AND BEHAVIOR 13

or probable exception of Alouatta. It is not difficult to trace
homologies among most of the vocalizations of most of the spe-
cies, and much of the information encoded is almost uniform or
strictly equi\'alent throughout. The forms and frequencies of
particular patterns are, however, very different in different spe-
cies. The differences seem to depend upon the distances over
which sound usually need to be transmitted, the carrying proper-
ties of the medium (the numbers and kinds of obstructions likely
to be encountered ) , and the presence or absence of other possible
sources of relevant information, features of the external and/or
social circumstances and other types of signals. In fact, this
means that both the physical forms of the patterns and the
methods of encoding information are closely correlated with
social structure, density of population, activity rhythms, and
density of vegetation, as well as vulnerability to predation and
diversity of appropriate receivers. The ritualized visual signals
are more heterogeneous but equally easy to explain in terms of
the same factors.

Some New World primates are involved in, or are the foci of,
specialized and stereotyped interspecific social reactions. Such
reactions may take either positive or negative forms, "friendly"
joining and following or hostile fighting or avoidance. They may
occur among two or more species of monkeys and /or between
monkeys and other animals such as squirrels {Sciurus granat en-
sis, S. variegatoides, Microsciurus sp.), birds of prey such as
Harpagus bidentatus and Leucopternis albicollis (these small
hawks do not attack the monkeys themselves, but rather take the
arthropods, lizards, etc., flushed by them), and even flycatchers
{e.g., Myiozetetes, Tyrannulus, Lagatus, Elaenia, Megarynchus) .
The combinations of positive and negative responses can be com-
plex, and the interspecific relations of a single species may be
different in different areas. It is possible, nevertheless, to detect
certain general rules or trends.

There are apparent correlations among interspecific bonds,
feeding habits, and territorial behavior. The monkeys that are
most likely to mingle with other species are forms such as Calli-
cebus moloch and Alouatta villosa. They are vegetarian, taking
items such as leaves, buds, and berries that are abundant and
evenly distributed, and have small territories or large territories
through which they move slowly. Individuals and groups of
these species seldom find themselves in situations with which they
are not thoroughly familiar or have not had time to inspect
carefully beforehand. Conversely, the establishment of friendly

14 BREVIORA No. 415

interspecific bonds is characteristic of such forms as Saimiri,
Cebus apella, and Ateles paniscus s.l. They are omnivorous or
preferentially insectivorous or feed on plant materials that are
dispersed or distributed in irregular clumps. They tend to have
large territories through which they move rapidly. They must be
precipitated into unfamiliar situations rather frequently. They
must also, therefore, have more need of extra companions of the
same or other species, to act as scouts or sentinels, than do spe-
cies of more sedentary or cautious habits.

On logical grounds, one would suppose that the various kinds
of interspecific social behavior should be adjusted to intensities of
competition^ as well as particular ecological facies. It would be
expected that species that do not compete at all, or compete as
little as may be feasible for animals that occur in the same areas,
would usually tend to ignore one another. There are many ap-
parent examples of such behavior among the New World pri-
mates. It would also be expected that species that compete verv'
strongly would tend to exclude one another from wide areas and
entire regions. Again there are apparent examples among the
American monkeys.

Presumably either of these extreme types of interspecific be-
havior can be transformed into the other in the course of time.
It would be interesting to know the intermediate stages. Data
from observations of the New World primates and their asso-
ciates would suggest that the following progression (quoted from
Moynihan, in press a) may be common as intensity of competi-
tion increases: "When competition becomes slightly more than
minimal, the species will tend to ignore one another in most
circumstances but will exhibit overt and active hostility toward
one another occasionally. (If it is only desirable or necessary to
drive off rivals infrequently, it may be worth taking the risk of
fighting. ) When competition is stronger, it may be advantageous
for the competitors to join up with one another. (If you can't
lick 'em . . .) When competition becomes stronger yet, it may
become imperative to avoid one another. First by a\'oiding per-
sonal encounters while still ranging over the same areas at much

'I am employing such terms as "complete" and "competition" in the
broadest possible sense. Two animals are considered to be competing with
one another whenever one preoccupies, permanently or temporarily, any
resource that would otherwise be likely to be used by the other. Among
primates and birds, competition for preferred observation posts, singing
perches, safe sleeping quarters, etc., may be quite as important as compe-
tition for food.

1973 EVOLUTION AND BEHAVIOR 15

the same times. Then by claiming exclusive territories or by
elaborating some form of temporal segregation. (Segregation by
differential timing may have peculiar advantages, but it can only
work when the species involved are not too numerous.) From
the claiming of exclusive territories, there may be no more than
a small step to complete allopatry. It seems very probable that
the process can also go in the opposite direction, through the
same stages but in reverse order, and that the direction of
change can be reversed repeatedly, with or without reaching the
extreme conditions at either end."

3. Most of my recent work on birds has been conducted in the
Andes.

The higher reaches of these mountains provide a wealth of
material for students of biogeography. They include a large
series of habitats and biotas that differ from those of the sur-
rounding lowlands in several respects {e.g., temperature, endemic
species). The northern part of the Andes is extremely complex
in structure, with separate cordilleras, chains of mountains, and
a scattering of single peaks and massifs. The central and south-
ern parts are simpler, more unified in general or overall form,
but still varied in details of terrain and cHmate. As a result,
many of the higher altitude habitats and biotas are distributed
in patches, partly or wholly isolated from one another. They
are essentially ijisular. They differ from oceanic islands, how-
ever, in not being impoverished. The higher Andes have "com-
plete" or "balanced" floras and faunas. They are inhabited by
many kinds of organisms which have occupied most of the
obvious niches or ecological roles, exploited most of the available
opportunities. They are, therefore, ideal for analyses of some
aspects of insular evolution. The effects of isolation and adapta-
tions to facilitate or impede invasions can be studied per se,
quite apart from the possible distortions of "accidental" barriers
or "sweepstake" phenomena.

I have concentrated upon interspecific behavior among two
groups of species of a particular "life zone." Observations were
begun in 1959 and have continued off and on until the present.
The results are being analyzed and written up. Many details
remain to be settled, but the general sense of the bulk of the
data is clear.

The life zone is the one that Chapman (1917 and 1926)
called "humid temperate." The term is perhaps misleading —
"cold humid tropical" might be more suitable (see comments in
Moynihan, 1971). The zone is best developed around 2800-

16 BREVIORA No. 415

3300 m in most areas. Its natural \egetation would be more or
less dense forest and "alpine" scrub (Weber, 1969). Some of
this survi\es apparently intact. The rest has been replaced by
secondary bush, gardens, hedges, crop fields, pastures, etc. For-
tunately, substantial numbers of the native birds have been able
to occupy and even flourish in some (the lusher) of these man-
made habitats. They are still easily observable. The distribution
of the zone is eccentric within the Andes. It must cover almost
the whole of the northern Andes at appropriate elevations, i.e.,
it is scattered among islands, most of which are small, a few of
which are large but long and narrow. It is much broader and
more nearly continuous in the central Andes, in all or most of
central Ecuador and northern Peru. It becomes progressi\'ely
narrower toward the south, even though the Andes themselves
remain broad. The apparent discrepancy is due both to the relief
of the mountains and the nature of the prevailing wind systems
(briefly summarized in Murphy, 1936). Rain falls off at an
unequal rate. The principal southern extension of the zone is
along the eastern slope of the chain, down into central Bolivia.
It is dissected by the deep valleys of rivers flowing to the Ama-
zon. In effect, the southern extremities are a series of narrowly
linked narrow peninsulas.

My own observations have ranged from the Sierra de Merida
in Venezuela and the Sierra Nevada de Santa Marta in northern
Colombia down through central and southern Colombia, Ecua-
dor, and Peru to northern Bolivia, the Yungas of La Paz, at
altitudes between 2400 and 3700 m. This is nearly the full
length of the cold humid tropical zone, with the addition of
some fringe areas of adjacent zones.

One of the groups of species studied could be called the
''Diglossa cluster." It includes six species or superspecies of the
genus, flower-piercers, which may be called carbonaria, lafres-
nayei, albilatera, baritula, cyanea, and coerulescens (this is the
nomenclautre and classification of Zimmer, 1929; HeUmayr,
1935; and de Schauensee, 1970 — Vuilleumier, 1969, suggests
a slightly different arrangement, and other refinements are con-
ceivable), as well as the conebill Conirostrum cinereum and
some hummingbirds such as Colibri coruscans, Aglaeactes cupri-
pennis, and Ramphoynicron microrhynchufn. All these birds are
nectarivorous to a greater or lesser extent.

The other group includes many more species of different sub-
families, families, and at least one more order. For want of a
better name, I shall call it the "tanager cluster." It includes a

1973 EVOLUTION AND BEHAVIOR 17

variety of closely related montane tanagers, mostly black and
blue with touches of yellow, buff, or red, of such genera as
Anisognathus, Buthraupis, and Iridosornis (and also the "Plush-
capped Finch," Catamhlyrhynchus, hardly distinguishable from
Iridosornis in appearance or habitus in the field^) ; other tanagers
of rather different stocks [e.g., Chlorospingus, Cnemoscopus,
Hemispingus, Chlorornis) ; finches of the genus Atlapetes; some
other conebills (especially Conirostrum sitticolor — see Moyni-
han, 1968) ; warblers of the very different genera Myioborus and
Basileuterus'- ; a few flycatchers {e.g., Uromyias and Megacer-
cuius spp.); the occasional hummingbird {e.g., Ensifera and
Coeligena) ; a few woodpeckers [e.g., Piculus rivolii in Vene-
zuela) ; and many furnariids and dendrocolaptids [Margarornis,
Synallaxis, Cranioleuca, etc.). And at least one squirrel in the
western cordillera of Colombia [Sciurus granatensis again!).
The association includes frugivores (different species taking dif-
ferent fruits ) , insectivores ( catching different insects in different
ways), a new nectarivores, and many types with very mixed
diets. Different species also prefer different levels of vegetation,
from the highest tree-tops down to the ground.

The chief peculiarity of both clusters, the one that drew my
attention, is that their members show pronounced intraspecific
geographic variation in their interspecific behavior. More pre-
cisely, individuals of a single species or superspecies react very
differently to individuals of other species in different regions
(often the same other species in each of the regions). The vari-
ation affects different types of interspecific behavior in the two
clusters, hostility in the Diglossa association and "friendliness"
in the tanager association, but the trends are roughly parallel in
both, although inverse and complicated by certain exceptions.
The exceptions themselves are sometimes revealing.

The situation is roughly as follows :

^The classification of the "New World nine-primaried songbirds" is in
need of further revision. Some of the supposed families and subfamilies of
the group appear to be polyphyletic in origin. Some of the genera cmrently
assigned to one family may be more closely related, phylogenetically, to some
of the genera assigned to other families than to other genera assigned to the
same family. Terms such as "warbler," "tanager," and "finch" are little more
than short-hand descriptive labels for certain ecological categories.

^In the case of these Andean birds, it seems probable that a revised scheme
would place the Plush-capped Finch in the same tribe as the tanagers it so
much resembles, and also link Basileuterus to Hemispingus rather than to
Myioborus.

18 BREVIORA No. 415

Many members of the tanager cluster extend throughout all or
most of the cold humid tropical zone. All show tendencies to
form or join mixed species flocks in some areas and regions (this
is the prescriptive reason why they have been assigned to the
cluster). In general, individuals of the same species behave in
similar ways in the northern and southern extremities of the zone,
but very differently in the central part. They show a high degree
of interspecific gregariousness in the western and central Cordil-
leras of Colombia (the western cordillera is alwavs extremely
northern, "far out," in the behavior of its inhabitants — see also
below). In these regions, the birds occur in mixed flocks most
of the time, and most of the flocks are large, cohesive, complex
in structure, and stable (maintained for hours on end and often
re-formed on successive days). In the eastern cordillera of Co-
lombia and the Sierra de Merida, the birds still show a con-
siderable amount of interspecific gregariousness, but mixed flocks
are formed somewhat less frequently and tend to be smaller,
looser, and simpler in structure on the average (the decline may
be more evident in the eastern cordillera than in Venezuela ) .
In central Ecuador and central and northern Peru, interspecific
gregariousness is slight. In fact, quite absent in some localities.
Even when and where mixed flocks are formed, they are always
small and simple, and usually loose and sustained for only a few
minutes. The trend is reversed in southern Peru and northern
Bolivia. Mixed flocks become larger, more stable, cohesive, and
complex again ( rather more so in Bolivia than in Peru, but never
as much so as in the western cordillera of Colombia ) .

It is obvious that the development of flocking depends upon
several factors. There are positive correlations among densities
of populations, thickness of vegetation, and frequency and elab-
oration of interspecific gregariousness within regions. But these
cannot account for the whole of the major geographic trends.
They do not explain the exceptions. There must be something
else involved. This would appear to be an "invasion" or "fron-
tier" effect. Interspecific gregariousness seems to go up with
exposure to, or anticipated number of, invasions from or into
other regions of the same life zone or an adjacent zone, the warm
or hot humid zone of lower elevations.

The western cordillera of Colombia is the least continuous of
the major chains of the Andes. Its patches of cold humid habi-
tats are comparatively small. The populations of these small
islands must include a relatively very high proportion of indi-
viduals near the frontiers of their patches and a low proportion

1973 EVOLUTION AND BEHAVIOR 19

of individuals at the centers of patches, away from the frontiers.
The same must be true of the populations of the narrow penin-
sulas of the zone in the far south. Birds on the frontiers must
encounter strays from other zones and stray into other zones more
often than do birds from the centers. It would seem that this is
one of the causes of interspecific gregariousness. The evidence
is somewhat restricted, but I think convincing. In central Ecua-
dor, I worked along one transect from the top edge of cold
humid forest and scrub down into the upper reaches of warm
humid forest. Interspecific gregariousness is essentially nil in the
higher part of the cold humid zone, but increases abruptly at
the exact point where occasional strays from the warmer zone
begin to appear with some appreciable, if still low, frequency.
(The increase is "intrinsic." It is always apparent, whether or
not strays are present at the moment.) The remarkably high
degree of gregariousness of the birds of the central cordillera of
Colombia, higher than would be expected of its not particularly
northern or isolated position, may also be correlated with the fact
that it is exposed to invasions from the nearby chains on either
side as well as from the immediately adjacent lowlands.

What is the functional significance of this apparent connection
of interspecific gregariousness with frontiers, strays, and inva-
sions? The advantages of mixed flocking from the point of view
of a straying bird in an unfamiliar area are obvious, and much
the same as in the monkeys cited above. By associating with
experienced local individuals, a stray may be able to discover
and identify food and/or danger relatively rapidly. The ad-
vantages for the "hosts" of a stranger are more problematical.
Of course, they are acquiring a companion who may be of use
in various ways. They are also encouraging or tolerating a com-
petitor. Perhaps one of the reasons that they do so is that they
may become strangers in their turns. Some of them must also
stray into adjacent life zones, where they will also need the help
of local inhabitants. It may be diflficult for an animal to join and
follow strangers without also developing some tendency to allow
itself to be joined and followed by strangers. (The roles of joiner
and joined are easily distinguishable in some areas such as parts
of Panama — see, for instance, Moynihan, 1962a — but they
are less clearly distinct in these Andean flocks. In any case, both
roles often reflect similar states of mind.) It seems to be char-
acteristic of most animals that they cannot, at least do not, sup-
port very great qualitative differences in kind of social responses.
A species that is comparatively aggressive in one class of social

20 BREVIORA No. 415

encounters also tends to be aggressive in other encounters. Sim-
ilarly, a species that is gregarious in some circumstances usually
tends to be gregarious in other circumstances.

This "extrapolation" may have been favored in Andean birds
because the boundaries of their life zones have been fluctuating,
repeatedly shifting back and forth in recent geological histor\'
(see Simpson-Vuilleumier, 1971 ). Many of the birds of the cold
humid zone must have had to invade new areas, and cope with
invaders from other areas, again and again in response to secular
climatic changes, quite apart from or in addition to the normal
stravinsf that would have occurred even if the frontiers had been
fixed and permanent.

The species of the Diglossa cluster show another contrast be-
tween individuals of the central part of the cold humid zone and
those of the northern and southern extremities of the zone. Some
aspects of their interactions in central Ecuador have been de-
scribed in Moynihan (1963). Each of the local species has its
own, partly unique, series of ecological preferences, but the
ranges of most species are broadly overlapping. The territories
of indi\'iduals of different species are often completely overlap-
ping. Indi\iduals of different species may use the same perches,
move along the same pathways, feed in the same places on the
same types of foods. But they almost never do so simultaneously.
They are almost always kept a few meters apart, at any gi\'en
instant of time, by some avoidance mechanisms. There is also
mutual inhibition of "Song" among individuals of different
species of Diglossa and Conirostrum cinereum, although not
among individuals of the same species. The whole thing can be
summed up as rigid and continuous social segregation. In the
western cordillera of Colombia and in northern Bolivia, on the
other hand, many of the species are separated microgeographi-
cally, each largely or completely confined to a particular facies
of habitat slightly different from the facies of all or most of the
others. This may be due to fighting. On the rare occasions when
individuals of different species that are usually separated do
happen to come together, they usually fight, actually attack, one
another. There is no visible avoidance mechanism. Thus, the
microgreographical segregation may be encouraged or imposed
by reactions among individuals but it is not continuously social
in the same wav as in central Ecuador. Conditions are more or
less intermediate in the Sierra de Merida, the eastern and central
Cordilleras of Colombia, and many areas of Peru, with all com-

1973 EVOLUTION AND BEHAVIOR 21

binations of partial overlaps, incomplete avoidance and inhibi-
tion, and more frequent and prolonged overt disputing.

The variations of the birds of the Diglossa cluster are also
correlated with factors such as density of vegetation and inter-
specific competition. They do not, however, include frontier
effects. They would seem to be more concerned with size of local
populations and competition within regions rather than invasions
by strays from without {Diglossa individuals are very sedentary).
Indi\iduals of the small northern and southern populations may
hope to fight off all or most of their not very numerous com-
petitors with relative ease. Individuals of the larger central pop-
ulation probably could not fight off their more numerous com-
petitors without exhausting themselves in the process or taking
unacceptable risks of physical injury.

It will be noticed that different adaptations for coping with
interspecific competition may tend to produce different diversity
gradients in the two assciations. In the Diglossa cluster, species
dixersity at any given point is least at the extremities and prob-
ablv CTeatest at the center of the cold humid zone. In the
tanager cluster, species diversity must often be greatest at par-
ticular points in the extremities and least at the center.

Comments

The sorts of work cited above are perhaps typical of a con-
temporary approach to ethology. I should hope that they would
suggest certain conclusions about studies of behavior and the
relationships of such studies to analyses of evolutionary processes.

Beginning with the purely ethological aspects, it seems evident
that causation is the crucial problem. Studies of ultimate causes,
natural selection, seem to be proceeding fairly well. At least,
there are no theoretical or basic methodological difficulties in-
volved. Studies of proximate causes, physiology, may also be
making progress, perhaps more rapid and exciting progress. But
there is very little contact between the two lines of investigation,
least of all when vertebrates provide the working material.
Doubtless, there will be a new and sophisticated synthesis of the
two approaches at some date in the future. I do not expect to
see it in my own (research) life time. I should also imagine
that, when it comes, it will be largely due to an expansion of
concern and efforts by physiologists. They would seem to be in
a better practical position to develop the necessary techniques
than are the field-oriented "natural historians."

22 BREVIORA No. 415

Meanwhile, there is still a lot that the ethologist can do for
for the evolutionist.

Beha\'ioral information can help to illuminate the evolution of
particular groups of animals. They have, for instance, increased
our knowledge of the phylogenies of many vertebrates such as
ducks and geese (Lorenz, 1941; Delacour and Mayr, 1945;
Johnsgard, 1965), gulls and terns and their relatives {e.g.,
Moynihan, 1962b), and cichlid fishes [e.g., Baerends and Baer-
ends, op. cit.). As taxonomic characters, however, behavior
patterns are no more and no less valuable than any other char-
acters. They may be more useful in some cases than in others,
more useful than other features in some groups, less useful in
other groups. They should continue to be considered, to be taken
fully into account, in systematic studies. But I would suggest that
they can make a more significant contribution to the analysis of
evolution by providing concrete, immediate, information to help
explain certain ecological phenomena, developments, and inter-
actions which are themselves among the causes of evolutionary
changes.

A substantial proportion of current and recent ecological re-
search has been devoted to such matters as competition, co-
existence, partitioning of resources, invasions of new areas and
habitats, replacement, and extinction (see the works of Hutchin-
son, Mac Arthur, Wilson, and others). There has been a stimu-
lating sequence of papers with models and diagrams, mathe-
matical formulae and other elaborations of symbolic logic, to
describe and summarize the results of interactions among indi-
viduals and species at present, as they probably were in the past,
and as they may be expected to be in the future and always.
What seems to me to have been lacking in many or most of these
discussions is attention to some of the details of the ongoing
processes as well as their end products, how and why they actu-
ally work in fact and in nature, the mechanics by which the final
results are achieved. A great many questions have been left
hanging in air. What do individuals of the same or different
species really do when they come face-to-face with one another?
Or when they occur in the same areas without necessarily en-
countering one another directly? What are the forms of compe-
tition? Who moves where, and why and when? How are spe-
cific resources found, used, preoccupied, defended? What are
the relevant clues? How does replacement occur on a day-to-
day or year-to-year time scale? What are the adaptations which
permit or facilitate supplants and invasions? How are these

1973 EVOLUTION AND BEHAVIOR 23

adaptations used in life and why are they effective? Why are
some adaptations more effective than others that could have
been used instead? Is there any consistent relation between size
of area inhabited and probabiUty of success? Are there some
species that are really specialists in competition? If so, why?
And how do they manage it?

These are the kinds of questions which behaviorists should be
able to answer, in whole or in part. I think that many behavior-
ists are trying to find the answers now. I hope and expect that
they will continue to do so.

Literature Cited

Altmann, S. a. 1967. Social Communication Among Primates. Chicago

and London: Univ. Chicago Press. 392 pp.
Argyle, M. 1972. Non-verbal communication in human social interaction.
In R. A. Hinde (ed.) , Non-verbal Communication: 243-268. Cambridge:
Cambridge Univ. Press. 441 pp.
Baerends, G. p., and J. M. Baerends. 1950. An introduction to the ethol-
ogy of cichlid fishes. Behaviour, Suppl. 1: 1-242.
Bastock, M., D. Morris, and M. Moynihan. 1953. Some comments on

conflict and thwarting in animals. Behaviour, 6: 66-84.
Blest, A. D. 1961. The concept of ritualisation. In W. H. Thorpe and
O. L. Zangwill (eds.) , Current Problems in Animal Behaviour: 102-124.
Cambridge: Cambridge Univ. Press.
Blurton Jones, N.G. 1967. An ethological study of some aspects of social
behaviour of children in nursery school. In D. Morris (ed.) , Primate
Ethology: 347-368. London: Weidenfeld and Nicolson.

. 1972. Non-verbal communication in children. In

R. A. Hinde (ed.) , Non-verbal Communication: 271-296. Cambridge:
Cambridge Univ. Press. 441 pp.
Boycott, B. B. 1965. A comparison of living Sepioteuthis sepioidea and
Doryteuthis plei with other squids, and with Sepia officinalis. J. Zool.,
147: 344-351.
Brock, V. E., and R. H. Riffenburgh. 1960. Fish schooling: a possible

factor in reducing predation. J. du Cons., 25: 307-317.
Brower, L. p., J. V. Z. Brower, and P. W. Westcott. 1960. Experimental
studies of mimicry. 5. The reactions of toads (Bufo terrestris) tO'
bumblebees (Bombus americanorum) and their robberfly mimics {MaU
lophora bomboides) , with a discussion of aggressive mimicry. Amer.
Natur., 94: 343-355.
BusNEL, R. -G. 1963. Acoustic Behaviour of Animals. Amsterdam, London^

New York: Elsevier Publ. Co. 933 pp.
Chance, M. A., and C. J. Jolly. 1970. Social Groups of Monkeys, Apes,,
and Men. London: Thames and Hudson. 224 pp.

24 BREVIORA No. 415

Chapman, F. M. 1917. The distribution of bird-life in Colombia: a con-
tribution to a biological survey of South America. Bull. Amer. Mus.
Nat. Hist., 36: 1-720.

1926. The distribution of bird-life in Ecuador. A contri-
bution to the study of Andean bird-life. Bull. Amer. Mus. Nat. Hist.,
55: 1-784.

Craig, W. 1918. Appetites and aversions as constituents of instincts. Biol.
Bull., 34: 91-107.

Crook, J. H. 1970. Social Behaviour in Birds and Mammals: Essays on the
Social Ethology of Animals and Man. London and New York: Aca-
demic Press. 492 pp.

CuLLEN, J. M. 1972. Some principles of animal communication. In R. A.
Hinde (ed.) , Non-verbal Communication: 101-122. Cambridge: Cam-
bridge Univ. Press. 441 pp.

Darwin, C. 1872. The Expression of Emotions in Man and Animals.
London: Murray. 374 pp.

Delacour, J., AND E. Mayr. 1945. The family Anatidae. Wilson Bull., 57:
3-55.

De Schauensee, R. M. 1966. The species of birds of South America. Nar-
berth, Pa.: Livingston. 577 pp.

Devore, I. 1965. Primate Behavior. New York: Holt. 654 pp.

DoLHiNow, P. 1972. Primate Patterns. New York, etc.: Holt, Rinehart and
Winston, 425 pp.

Eibl-Eibesfeldt, L 1970. Ethology. The Biolog)- of Behavior. New York,
etc.: Holt, Rinehart and Winston. 530 pp.

. 1972. Similarities and differences between cultures in

expressive movements. In R. A. Hinde (ed.) , Non-verbal Communica-
tion: 297-312. Cambridge: Cambridge Univ. Press. 441 pp.

Erikson, G. E. 1963. Brachiation in the New World monkeys. Symp. Zool.
Soc. London, 10: 135-164.

Ewer, R. F. 1963. The behaviour of the meerkat, Suricata suricatia
(Schreber) . Z. Tierpsychol., 20: 570-607.

. 1968. Etholog)^ of Mammals. London: Logos Press. 418 pp.

1973. The C-arnivores. Ithaca: Cornell Univ. Press. 494 pp.

Gadgil, M., and W. H. Bossert. 1970. Life historical consequences of
natural selection. Amer. Natur., 104: 1-24.

GoFFMAN, E. 1971. Relations in Public. New York, etc.: Harper and Row.
396 pp.

Heinroth, O. 1911. Beitrage zur Biologic, namentlich Ethologie und
Psychologic der Anatiden. Proc. 5th Int. Orn. Congr.: 589-702.

Hellmayr, C. E. 1935. Catalogue of birds of the Americas. Part VIII,
Alaudidae-Compsothlypidae. Publ. Field Mus. Nat. Hist., zool. ser., 13:

1-542.
HiNDE, R. A. 1954a. Factors governing the changes in strength of a par-
tially inborn response, as shown by the mobbing behaviour of the
chaffinch (Fringilla coelehs) : I. The nature of the response, and an
examination of its course. Proc. Roy. Soc, B, 142: 306-331.

1973 EVOLUTION AND BEHAVIOR 25
. 1954b. Factors governing the changes in strength of a par-

tially inborn response, as shown by the mobbing behaviour of the chaf-
finch (Fringilla coelebs) : II. The warning of the response. Proc. Roy.
Soc, B, 142: 331-358.
. 1969. Bird Vocalizations. Cambridge: Cambridge Univ. Press.

394 pp.

1970. Animal Behaviour: a Synthesis of Ethology and Com-

parative Psychology. 2nd edition. New York: McGraw-Hill. 876 pp.
. 1972. Non-verbal Communication. Cambridge: Cambridge

Univ. Press. 441 pp.
Hladik, a., and C. M. Hladik. 1969. Rapports trophiques entre vegetation

et primates dans la foret de Barro Colorado (Panama) . Terre et Vie,
1969: 25-117.

Hladik, C. M., A. Hladik, J. Bousset, P. Valdebouze, G. Viroben, and
J. Delort-Laval. 1971. Le regime alimentaire des primates de I'ile de
Barro-Colorado (Panama) . Folia primat., 16: 85-122.

Holmes, W. 1940. The colour changes and colour patterns of Sepia offi-
cinalis L. Proc. Zool. Soc. London, A, 110: 17-36.

Huxley, J, S. 1914. The courtship of the Great Crested Grebe. Proc. Zool.
Soc. London: 491-564.

Imanishi, K., and S. a. Altmann. 1965. Japanese monkeys. Published by
S. A. Altmann (Edmonton?) . 151 pp.

Jay, p. 1968. Primates. Studies in Adaptation and Variability. New York,
etc.: Holt, Rinehart and Winston. 529 pp.

Jeletzky, J. A. 1966. Comparative morphology, phylogeny, and classifica-
tion of fossil Coleoidea. Univ. Kansas Paleontol. Contr., MoUusca, 7:
1-162.

Johxsgard, p. a. 1965 Handbook of Waterfowl Behaviour. London: Con-
stable. 378 pp.

Jolly, A. 1966. Lemur Behavior. Chicago: Univ. Chicago Press. 187 pp.

. 1972. The Evolution of Primate Behavior. New York: Mac-

Millan. 397 pp.

Kaufmann, J. H. In press. Social ethology of the Whip-tailed Wallaby,
Macropus tarryi, in northeastern New South Wales, Australia.

Kleiman, D. G. 1972. Social behavior of the Maned Wolf (Chrysocyon
brachyurus) and Bush Dog (Speothos venaticus) : a study in contrast.
J. Mammal., 53: 791-806.

KoRTL'ANDT, A. 1940. Wcchselwirkuug zwischen Instinkten. Arch, neerl.
Zool., 4: 442-520.

Kruuk, H. 1972. The Spotted Hyena. Chicago: Univ. Chicago Press. 335 pp.

Rummer, H. 1968. Social Organization of Hamadryas Baboons. Chicago:
Univ. Chicago Press. 189 pp.

. 1971. Primate Societies. Chicago: Aldine. 160 pp.

Lane, F. W. 1957. The Kingdom of the Octopus. London: Jarrolds.

Lanyon, W. E., and W. N. Tavolga. 1960. Animal sounds and communi-
cation. Washington: Am. Inst. Biol. Sci. 443 pp.

26 BREVIORA No. 415

Lawick-Goodall, J. VAN. 1971. In the Shadow of Man. Boston: Houghton

Mifflin. 297 pp.
Leyhausen, p. 1956. Verhaltensstudien an Katzen. Z. Tierpsychol. Beiheft^

2: 1-120.
LoRENZ, K. 1931. Beitrage zur Ethologie soziales Corviden. J. Ornithol.. 79:

67-120.
. 1935. Der Kumpan in der Umwelt des Vogels. J. Ornithol. 83:

137-213 and 289-413.
. 1941. Vergleichende Bewegungsstudien an Anatinen. J. Orni-

thol., 89 (Suppl.) : 194-294.
. 1963. On Aggression. New York: Harcourt, Brace and World.

306 pp.
, AND N. TiNBERGEN. 1938. Taxis und Instinkt-handlung in der

Eirollbewegung der Graugans. I. Z. Tierpsychol., 2: 1-29.
MacKay, D. M. 1972. Formal analysis of communicative processes. In

R. A. Hinde (ed.) , Non-verbal Communication: 3-25. Cambridge:

Cambridge Univ. Press. 441 pp.
Makkink, G. F. 1936. An attempt at an ethogram of the European avocet

(Recurvirostra avosetta L.) with ethological and psychological remarks.

Ardea, 25: 1-60.
Marler, p., and W. J. Hamilton. 1967. Mechanisms of Animal Behavior.

New York: John Wiley and Sons. 771 pp.
Martin, R. D. 1972. Concepts of human territoriality. In P. J. Ucko,

Ruth Tringham, and G. W. Dimbleby (eds.) , Man, Settlement and

Urbanism: 427-445. London: Gerald Duckworth.
Morris, D. 1957. "Typical intensity" and its relation to the problem of

ritualisation. Behaviour, 11: 1-12.
. 1967a. Primate Ethology. London: Weidenfeld and Nicolson.

374 pp.
. 1967b. The Naked Ape. London: Cape. 252 pp.

Moynihan, M. 1962a. The organization and probable evolution of some
mixed species flocks of neotropical birds. Smithson. Misc. Coll., 143 (7) :
1-140.

. 1962b. Hostile and sexual behavior patterns of South

American and Pacific Laridae. Behaviour, Suppl., 8: 1-365.

. 1963. Inter- specific relations between some Andean birds.

Ibis. 105: 327-339.

1967. Comparative aspects of communication in New

World primates. In D. Morris (ed.) , Primate Ethology: 236-266. Lon-
don: Weidenfeld and Nicolson. 374 pp.
1968. The "Coerebini": a group of marginal areas, habi-

tats, and habits. Amer. Natur., 102: 573-581.
. 1970. Control, suppression, decay, disappearance, and re-

placement of displays. J. theor. Biol., 29: 85-112.
. 1971. Successes and failures of tropical mammals and

birds. Amer. Natur.. 105: 371-383.

1973 EVOLUTION AND BEHAVIOR 27
In press a. The New World Primates. Studies of the Evo-

lution of Intelligence, Gregariousness, and Language. Princeton: Prince-
ton Univ. Press.

In press b. Conservatism of displays and comparable stereo-

typed patterns among cephalopods.
In press c. Some distinctions of cephalopods; their causes

and evolutionary consequences.

In press d. Notes on the ecology and behavior of the

Pygmy Marmoset, Cehuella pygmaea, in amazonian Colombia.
Murphy, R. C. 1936. Oceanic Birds of South America. Vol. I. New York:

MacMillan. 640 pp.
N.AJiER, J. R., AND A. C. Walker. 1967. Vertical clinging and leaping —

a newly recognized category of locomotor behaviour of primates. Folia

primat., 6: 204-219.
Packard, A. 1972. Cephalopods and fish: the limits of convergence. Biol.

Rev., 47: 241-307.
, AND G. D. Sanders. 1971. Body patterns of Octopus vulgaris

and maturation of the response to disturbance. Anim. Behav., 19:

780-790.
Fetter, J. -J. 1962. Recherches sur I'ecologie et I'ethologie des lemuriens

malgaches. Mem. Mus. Nat. Hisc. Natu., n. s., 27: 1-146.
PoiRiER, F. E. 1972. Primate Socialization. New York: Random House.

260 pp.
Rand, A. S. 1967. Predator-prey interactions and the evolution of aspect

diversity. Atas do Simposio sobre a Biota Amazonica, 5 (Zoologia) :

73-83.
Reynolds, V. 1968. The Apes. London: Cassell. 296 pp.
Robinson, M. H. 1969. Defences against visually hunting predators. Evol.

Biol., 3: 225-259.
, and J. Olazarrl 1971. Units of behavior and complex

sequences in the predatory behavior of Argiope argentata (Fabricius) :

(Araneae: Araneidae) . Smithson. Contr. Zool., 65: 1-36.

-, and B. Robinson. 1971. The predatory behavior of the

ogre-faced spider Dinopis longipes (Araneae: Dinopidae) . Amer. Mid-
land Natur., 85: 85-96.
Rosenblum, L. a., and R. W. Cooper. 1968. The Squirrel Monkey. New

York and London: Academic Press. 451 pp.
RowELL, T. 1972. Social Behaviour of Monkeys. Middlesex: Penguin.

195 pp.
RussELL, C, AND W. M, S. RussELL. 1968. Violence, Monkeys and Man.

London: MacMillan. 339 pp.
ScHALLER, G. B. 1963. The Mountain Gorilla. Ecology and Behavior.

Chicago: Univ. Chicago Press. 431 pp.
. 1972. The Serengeti Lion. Chicago: Univ. Chicago Press.

480 pp.
Sebeok, T. a. 1968. Animal Communication. Bloomington: Indiana Univ.

Press. 686 pp.

28 BREVIORA No. 415

Seitz, a. 1940. Die Paarbildung bei einigen Cichliden. I. Z. Tierpsychol.,

4: 40-84.
. 1941. Die Paarbildung bei einigen Cichliden. II. Z. Tierpsychol.,

5: 74-101.
SiMPSON-VuiLLEUMiER, B. 1971. Pleistocenc changes in the fauna and flora

of South America. Science, 173: 771-780.
Smith, ^\^ J. 1965. Message, meaning, and context in ethology. Amer.

Natur., 99: 405-409.
. 1969. Messages of vertebrate communication. Science, 165:

145-150.
Struhsaker, T. 1969. Correlates of ecology and social organization among

African cercopithecines. Folia primat., 11: 80-118.
Tembrock, G. 1959. Tierstimmen. Wittenberg: A. Ziemsen Verlag. 286 pp.
Thorpe, \V. H. 1961. Bird Song: the Biology of Vocal Communication and

Expression in Birds. London: Cambridge Univ. Press. 143 pp.
TiNBERGEN, L. 1939. Zur Fortpflanzungsethologie von Sepia officinalis.

Arch, neerl. Zool., 3: 323-364.

TiNBERGEN, N. 1932. Uber die Orientierung des Bienenwolfes {Philanthus
triangulum Fabr.) . Z. vergl. Physiol., 16: 305-335.

. 1935. Uber die Orientierung des Bienenwolfes II. Die

Bienenjagd. Z. vergl, Physiol., 21: 699-716.
. 1936. Zur Soziologie der Silbermowe (Larus a. argentatus

Pontopp.) . Beitr. Fortpflanzungsbiol. Vogel, 12: 89-96.
. 1939. The behavior of the Snow Bunting in Spring. Trans.

Linn. Soc. New York, 5. 94 pp.

. 1940. Die Ubersprungbewegung. Z. Tierpsychol., 4: 1-40.

. 1951. The Study of Instinct. Oxford: Oxford Univ. Press.

228 pp.
VuiLLEMiER, F. 1969. Systematics and evolution in Diglossa (Aves, Coere-

bidae) . Amer. Mus. Novitates, No. 2381: 1-44.
Weber, H. 1969. Zur natiirlichen Vegetations — gliederung von Siidamer-

ika. In E. J. Fittkau, J. lilies, H. Klinge, G. H. Schwabe, and H. Sioli

(eds.) , Biogeography and Ecology in South America, vol. 2: 475-518.

The Hague: W. Junk.
Wells, M. J. 1962. Brain and Behaviour in Cephalopods. London: Heine-

mann. 171 pp.
, AND J. Wells. 1972. Sexual displays and mating of Octopus

x'lilgaris Cuvier and O. cyanea Gray and attempts to alter performance

by manipulating the glandular condition of the animals. Anim. Beliav.,

20: 293-308.
Whitman, C. O. 1899. Animal Behavior. Biological Lectures, Delivered at

Marine Biology Laboratory, Wood's Hole, Massachusetts. No. 17: 285-

338. Boston: Ginn and Co.
. 1919. The behavior of pigeons. Carnegie Inst. W^ash-

ingion Publ., 257: 1-161.
Wickler, W. 1968. Mimicry in Plants and Animals. New York: McGraw-
Hill. 253 pp.

1973 EVOLUTION AND BEHAVIOR 29

Young, J. Z. 1964. A Model of the Brain. Oxford: Oxford Univ. Press.
348 pp.

. 1972. The Anatomy of the Nervous System of Octopus vul-
garis. Oxford: Oxford Univ. Press. 690 pp.

ZiMMER, J. T. 1929. Variation and distribution in two species of Diglossa.
Auk, 46: 21-37.

. 'OO:: -m

Kt-!

• • •- .

•.■•'•..■•.

B R E V I O R A

MU0. COMP. ZOOL

Museum of ^fWffi^arative Zoology

Cambridge, Mass. ^^[^^j^qber 1973 Number 416

UNiVfiRSlTYi
MUSEUMS AND BIOLOGICAL LABORATORIES

Ernst Mayr

When Professor Crompton invited me to give a short after-
dinner address on the occasion of the opening of the wing, he
added that he wanted to publish it. This posed a challenge to
me to come up with something that is worth being printed.
However, I consider this invitation less of a challenge than a
welcome opportunity to present some thoughts on museums and
their role in science.

The speakers this afternoon have rightly emphasized that the
opening of the Museum's laboratory wing is a milestone in the
history of the MCZ. It is an occasion to look back to the days
of its founding and an occasion to look forward to its future.
It is also an occasion to ask some searching questions. For
instance, someone unacquainted with biology and intolerant of
anything but his own hobbyhorse, might ask, "Why do we still
need natural history museums?" Such a question is quite legiti-
mate, for I am a strong believer in the principle that the
legitimacy and continuing value of traditional rituals and insti-
tutions should be challenged from time to time. How, then,
would we answer this question?

The role of museums in science, and their image in our so-
ciety, is changing from decade to decade. When natural history
was revived during the Renaissance and during the 17th and
18th centuries, it expressed at first man's wonder and bewilder-
ment at the enormous variety of life. This "diversity of nature"
has been a key concept in man's world picture from the days
when the Lord told Adam to give names to all the creatures in
the field to the present day when species diversity is one of the
central themes in the work of the ecologists.

The rich treasures brought back from exotic countries in the
18th and 19th centuries by voyages and expeditions, combined

.--■*■

2 BREVIORA No. 416

with the steady rise of a more and more scientific attitude in
Western man, resulted in a changed concept of organic diversity.
No longer was it merely a source of wonder but naturalists
began to raise questions concerning the reasons for the existence
of so many and such strange organisms and about the meaning
of their peculiar distribution in Asia, Africa, the Americas, and
Australia.

I am not claiming that naturalists were always interested only
in the most lofty generalizations because there was hardly a
naturalist who was not also infected by that strange virus called
the collector's fever. Perhaps no one was more affected by this
disease than the founder of the MCZ, Louis Agassiz, who cheer-
fully pawned ever\thing he owned in order to acquire more
specimens. Indeed, it is said that only a few decades ago this
Museum still had unopened boxes of collections from Louis
Agassiz's days.

These collections, however, were not merely the useless gather-
ings of pack rats. It was their study which helped bring about
a conceptual revolution — the establishment by Darwin of the
theory of evolution, to a considerable extent based on Darwin's
own researches during the voyage of the "Beagle" and the sub-
sequent working out of his collections. And the proposal of the
theory of evolution was only one of several such conceptual
revolutions in the history of natural history.

The diversity of nature has been considered, ever since Dar-
win, a documentation of the course of evolution. Research in
the pathway of evolution indeed turned out to be an incredibly
rich gold mine. And it was the museums that established and
maintained leadership in this type of research. The historians of
biology have clearly determined that the crucial advances in the
modern interpretation of species, of the process of speciation, and
of the problems of adaptation were made by systematists.

One of the greatest conceptual revolutions in biology, the
replacement of essentialism by population thinking, was intro-
duced into biology by museum systematists. From systematics
it was brought into genetics by workers like Chetverikov, Timo-
feeff-Ressovsky, Dobzhansky, Sumner, and Edgar Anderson, all
of whom had either been trained as systematists or had worked
closely with systematists.

Again and again the students in special branches of biology
such as biogeography have gone back to systematics for material
and for novel ideas.

The speakers this afternoon have documented sufficiently how

1973 MUSEUMS AND LABORATORIES 3

important museums and systematics are. But this raises another
question, which is: "Why is systematics so important?" And this
leads right on to the further question of the position of syste-
matics in biology as a whole. I pointed out a dozen years ago
that, in spite of all of its unitary characteristics, biology really
has two major divisions; indeed, one can speak of two biologies.
In the first one, functional biology, "How?" questions are the
important ones. This is the biology that deals with physiological
mechanisms, developmental mechanisms, metabolic pathways,
and with the chemical and physical basis of all aspects of life.
To use modern technical language, this part of biology ultimately
deals both with the translation (decoding) of genetic programs
into components of the phenotype and with their subsequent
functioning. This type of biology played a decisive role in dis-
proving conclusively all vitalistic notions and in establishing firmly
that nothing happens in organisms that is in conflict with the
laws of chemistry and physics. This is the biology which inter-
prets all cellular and developmental processes, both the normal
ones and such abnormal ones as the origin of cancer.

The other biology is interested in the genetic programs them-
selves, dealing with their origin and evolutionary change. It
continuously asks "Why?" questions, for instance:

Why is there such a diversity of animal and plant life?

Why are there two sexes in most species of organisms?

Why is the old faunal element of South America seemingly
related to that of Africa while the new one is related to that of
North America?

Why are the faunas of some areas rich in species and those of
others poor?

Why are certain organisms very similar to each other, while
others are utterly different?

In the last analysis, all questions in this part of biology are
evolutionary questions, and museum-based collections are ulti-
mately needed to find the facts for posing and answering all of
these questions.

At this point some of the more perceptive members of this
audience will think that I have painted myself into a corner.
Why, they will say, do you need a laboratory wing when the
method of systematic and evolutionary biology is the comparative
method, based on observations? Why do you have to perform
experiments?

The explanation for the seeming contradiction is that I have
told only part of the story. Systematics, as it was defined by

4 BREVIORA No. 416

G. G. Simpson, "is the scientific study of the kinds and diversity
of organisms and of any and all relationships among them.''

This definition has two consequences: First, it means that
the systematist also must ask "How?" questions, like "How do
species multiply?" or "How does an evolutionary line acquire
new adaptations?", or "How did the phyletic line leading to
Man emerge from the anthropoid condition?".

All these evolutionary questions deal with the historv' of
changes, and, most importantly, with the causation of changes.
Translated into Darwinian language, each of the questions I
have just posed can also be stated in the following terms:

"What were the selection pressures responsible for causing the
stated evolutionary changes?"

Not only is it often necessary to make use of experiments to
answer this type of question, but, more importantly, many of
such questions cannot be answered — or at least not completely
— simply by the study of preserved material.

Since the investigation of diversity includes the study of rela-
tionships, organisms must be studied alive and in the field. In
the last 150 years there has hardly been an outstanding sys-
tematist Vv^ho was not, at the same time, an outstanding field
naturalist, and who could not have been called, with equal
justification, an ecologist or a student of behavior. This is, by
no means, a recent development. Re-reading recently Louis
Agassiz's "Essay on Classification," published in 1857, I was
astonished to find what stress he placed on the study of the
"habits of animals," as he put it.

"Without a thorough knowledge of the habits of animals,"
he said, "it will never be possible to determine what species are
and what not." Today we would call this a biological species
concept. He goes on to say that we want to find out ''how far
animals related by their structure are similar in their habits, and
how far these habits are the expression of their structure." He
continues, "How interesting would be a comparative study of
the mode of life of closely allied species." Indeed, Agassiz pro-
poses a program of study which is virtually identical with that
of the founders of ethology more than 50 years later: "The more
I learn about the resemblances between species of the same
genus and of the same family . . . the more am I struck with the
similarity in the very movements, the general habits, and even in
the intonation of the voices of animals belonging to the same
family ... a minute study of these habits, of these mo\ements.

1973 MUSEUMS AND LABORATORIES 5

of the \oice of animals cannot fail, therefore, to throw additional
light upon their affinities."

An interest in the behavior of animals is still a tradition in
the MCZ, more than 100 years later. Half of my Ph.D. students
in the last 20 years, for example, did their theses on problems of
beha\ior. One of the outstanding characteristics of the so-called
new systematics is the concern with the attributes of the living
animal. Variation, adaptation, speciation, and evolutionary
change cannot be fully understood unless the field work is sup-
plemented by experimental research in population genetics, the
analysis of protein and chromosomal variation in populations,
the study of the relations between adaptation and functional
morphology, to give merely a few examples. Laboratories for
such studies are a major component of the new wing. Environ-
mental physiology, another aspect of animal adaptation of great
interest to the evolutionist, is being studied at the Countway
Laboratories of the Concord Field Station.

The outside world has been largely oblivious to these develop-
ments and, I am sorry to say, unfortunately so have also many
svstematists. For the modern svstematist, however, all this seems
to be a perfectly natural development. Anyone who has read
books Hke Huxley's New Systematics (1940) or my own Sys-
tematics and the Origin of Species (1942) knows to what an
extent all these mentioned activities have been part of systematics
for at least 30 years. The new wing gives us an opportunity to
help correct the false image about museums which is still widely
held, and replace it by the new concept, the beginnings of which
were already outlined by Louis Agassiz 1 1 6 years ago.

The new wing signals to the outside world that the MCZ
is not merely a repository of collections but a biological research
institute that differs from the other laboratories in the Biological
Laboratories only in the nature of the subject matter. While
the emphasis in much of the Biological Laboratories is on cells
and the molecular constituents of cells, the major emphasis in
the MCZ is on the whole organism, on the diversity of organisms
and on their evolution. Since closest contact between the two
groups of investigators is of the utmost mutual benefit to both of
them, the organization of the Department of Biology was modi-
fied in recent years in order to integrate the staffs of the two
groups. Research and teaching are the objectives of both of
them.

In this day and age science is no longer conducted merely for
its own sake. Science is no longer the tenant of an ivory-tower.

6 BREvioRA No. 416

Without wanting to minimize in any way the indispensabihty of
basic science, we now realize that the scientist also has social
obligations. When optimistically inclined he will say that he is
helping to build a better world; when pessimistically inclined
he will say he is trying to prevent a further deterioration of this
world.

But he cannot do this unless he has a sound understanding of
Man and of the world in which he lives. And it is precisely the
study of diversity and of evolutionary history which has made a
major contribution toward the development of a new image of
Man.

In the pre-Darwinian literature, and also, in much of certain
types of contemporar)^ literature, man is conceived as a static
being, created within an equally static nature that is subservient
to him. Ever since Darwin this concept has increasingly been
replaced by a new image, an image of an evolved and still
evolving man, part of the evolutionar)' stream of the whole living
world. And this new image, the direct product of evolutionary
and natural history studies, is of critical importance, not only
for our personal concept of the world in which we live, but also
for such quite practical issues as man's relation to the environ-
ment, to the natural resources, and indeed even to the inter-
action among men.

It is about time we realize that the future of mankind is not
something "written in the stars," something controlled by ex-
ternal forces, but that it is we humans ourselves who hold the
fate of our species in our hands. We now have a fairly good idea
what the major ills of mankind are and it has become quite
clear that only a few of them are susceptible to purely techno-
logical solutions. Instead, most of them are of a beha\'ioral-
sociological nature and require a change in our value systems,
a change one is not likely to accept unless one has a far better
understanding of nature, of the dynamics of populations, of the
biological basis of behavior, and of other components of the
biology of organisms, than most of those have who are responsi-
ble for policy decisions.

It will require a deeper understanding of the mentioned prob-
lems and it will require massive education based on the findings
that emerge from the type of researches that we are planning.
During the planning of the wing we sometimes referred to it as
a new "center for environmental and behavioral biology." Al-
though this title was not officially adopted, it is indeed an apt

1973 MUSEUMS AND LABORATORIES 7

description of the focus of attention of the investigators in our
new facihty.

There may be some who have not kept up with recent devel-
opments in biology and who might consider it far-fetched to
claim that the mentioned problems fall within the area of
interest of systematics. And yet with systematics defined as the
science of biological diversity and with the organism defined as
something living and not merely a preserved specimen, a solid
chain of links is formed from the systematics of Linnaeus through
that of a Louis Agassiz to that of the modern evolutionary sys-
tematist and population biologist.

I add my vote of thanks to those who have made the creation
of this new center of environmental and behavioral biology pos-
sible. I predict that it will have an impact on our knowledge
and our thinking that will reach to the far corners of the earth.

^ -(V ^\(J rr) 6

B KMJU 0 R A

LIBRARY

Vliiseiini of Comparative Zoology

JAN 7 m

JAN

US ISSN 0006-9698

HARVARD ~ ~Z

Cambridge, MASgj|^|^8_DjjMyBER 1973 Number 417

A NEW SPECIES OF CYRTODACTYLUS

(GEGKONIDAE) FROM NEW GUINEA

WITH A KEY TO SPECIES FROM THE ISLAND

Walter C. Brown^

AND

Fred Parker-
Abstract. A new species of Cyrtodactylus from New Guinea is described.
The type locality is Derongo at an altitude of 1300 feet on the Alice River
tributary system to the upper Fly River, in western Papua, New Guinea.
A key to the species of Cyrtodactylus which have been recorded from New
Guinea is also provided (see de Rooij, 1915, for descriptions of most of the
species) .

Introduction

Of the nine species of Cyrtodactylus previously recorded from
New Guinea, known ranges of at least two (C. sermowaiensis
and vankampeni) are restricted to one or two localities. The
species described in the present paper may also exhibit a limited
range, for although the junior author has collected extensively
in papuan New Guinea for several years, no specimens have
been collected thus far outside of the type locality in the head-
waters of the Fly River.

Inger (1958) calls attention to the usefulness of the pattern of
the enlarged scales in the preanal region and on the under sur-
face of the thighs as characteristics for distinguishing species of
Cyrtodactylus, and uses it in the key to the species from the
Philippines and Borneo. We have found these characters sim-

^Califomia Academy of Sciences and Menlo College, Menlo Park, California
94025

^P.O. Box 52, Daru, Papua, New Guinea

2 BREVIORA No. 417

ilarly useful in separating most of the New Guinea species. We
have not had the opportunity to examine specimens of C. novae-
guineae.

Cyrtodactylus derongo new species

Holotype. Museum of Comparative Zoology Rl 26205, an
adult female, collected by Fred Parker in the Derongo area at
an ele\'ation of 1300 feet, Alice River system, tributary to the
upper Fly River, Papua, New Guinea, 8 April 1969.

Paratypes. Museum of Comparative Zoologv Rl 26203,
126204,' and 126206, Papua New Guinea Museum R995, and
American Museum of Natural Historv 103910, same data as the
holotype.

Diagnosis. A Cyrtodactylus with small scales on postero-
ventral surface of thighs meeting the enlarged scales of antero-
ventral surface at a sharp boundary; the rows of enlarged
femoral scales forming a continuous series with preanal rows;
enlarged preanal scales posterior to the pore series absent; dorsal
ground color dark brown with very faint darker blotches en-
closing irregular rows of large, white tubercles ( Fig. 1 ) .

Description. A moderately large Cyrtodactylus; four adult
females measure 105-112 mm snout-vent length, one specimen
81 mm in snout-vent length is immature; head about one and
one-half times its breadth; eye, large, its diameter about one-
third of the length of the head and about equal to its distance
from, the nostril; diameter of ear opening less than half its dis-
tance from the eye; head covered with granules, very small
posteriorly and somewhat larger anteriorly; scattered, moderate-
sized, pointed tubercles as far anterior as the interorbital region ;
rostral large, rectangular, its breadth about 60 percent of its
length, nostril bordered by the rostral, supranasal, first labial
and 3 small shields; upper labials 11 or 12; lower labials 1 1 to
13; supranasals large, separated by 1 or 2 scales; one large pair
of postmcntals in contact posteriorly for about half their length;
distinct lateral fold lacking, but its normal position marked by a
row of flattish tubercles separated from one another by several
smaller scales; in the mid-body region, 20 irregular lines of dorsal
tubercles between the aforementioned rows of flattish scales; 15
to 1 8 rows in the axillary region ; some of the tubercles are white
and tend to form widely separated irregular transverse lines, 8
to 10 between the nape and the hind limbs; undersurface of

1973

CYRTODACTYLUS DERONGO

Figure 1. Dorsal view of Cyrtodactylus derongo, MCZ 126205, type specimen.

4 BREVIORA No. 417

head with small granules; venter with about 46 to 48 rows of
scales at the mid-body between the ventrolateral rows of tuber-
cles, small and granular laterally, but merging gradually with
the large cycloid scales of the mid-venter; the large preanal-pore
scales in a \'ery shallow "/\" continuous with a row of femoral-
pore scales that are gradually reduced in size along the femur;
those anterior to the pore row somewhat enlarged, flattish scales
on both the thighs and the preanal region, the latter merging with
those of the \enter; posteriorly the pore series is met abruptly by
small granular scales in both the preanal and femoral regions;
24 to 26 rows of lamellae and scales beneath the fourth toe;
tail only slightly depressed, with square or rectangular plates
toe; tail only slightly depressed, with square or rectangular plates
on the \entral surface and with everv fourth or fifth scale dis-
tinctlv enlar2:ed.

Snout-\'ent length of holotype 105 mm.

Color ( in preservative ) . The dorsum is dark reddish brown
with 9 or 1 0 very faint series of darker blotches each enclosing
two to several large white tubercles; the latter tend to form very
irregular, widely separated, transverse rows; in the inter\'ening
areas the tubercles are dark or have a faint whitish tip; scat-
tered white tubercles also occur on the posterior part of the head,
the dorsal surfaces of the limbs and the base of the tail; venter
lighter brown, most dilute on the head and throat, each scale
marked by a xarying number of small brown spots and flecks.
In life, the dorsal ground color is dark purplish brown; the
venter is paler and more translucent. The iris is deep brown.

Habitat note. The specimens of Cyrtodactylus derongo were
collected from crannies and hollows in trees in dense rain forest.
Natives state the species is completely arboreal. Two other spe-
cies of Cyrtodactylus, papuensis and mimikanus, are sympatric
with derongo, and were observed both on the forest floor and on
trees a few feet above the ground. A possible fourth species, also
arboreal, was observed in the same area but specimens are not
availal)le for identification.

Comparisons. Differs from other Indo-Australian species of
Cyrtodactylus in the rather uniformly dark ground color of dor-
sum marked by large white tubercles. The color pattern is remi-
niscent of that of Underwoodisaurus milU, but in the latter the
white patches in\'olve small surrounding scales, and the patches
may be fused into partial or complete transverse bands. Com-
pared to other New Guinean species, C. derongo is somewhat

1973 CYRTODACTYLUS DERONGO 5

intermediate in size along with mimikanus, marmoratus, papuen-
sis, and pelagicus, and in contrast to the diminutive vankampeni
and the larger loriae, louisiadensis and novaeguineae. It also
differs from other species, with the possible exception of novae-
guineae (not examined), in the pattern of enlarged preanal
and femoral scales, and in lacking enlarged scales posterior to
the pore series in the preanal area. C. pelagicus and vankampeni
exhibit no or only ver\^ slightly enlarged scales in the pore series;
loriae, louisiadensis, mimikanus, marmoratus, and papuensis ex-
hibit 3 to 8 or 9 short rows of large scales posterior to the pore
series in the preanal area.

Key to Cyrtodactylus From New Guinea

1. a. Preanal region, or both preanal and femoral regions, with one or more

rows of distinctly enlarged scales 3

b, Preanal and femoral regions covered by relatively uniform small scales,
even the pore series not distinctly enlarged 2

2. a. Dorsal rows of tubercles at mid-body 22-24, usually 10 at region of fore

limbs; 8-12 preanal pores, femoral pores absent pelagicus

b. Dorsal rows of tubercles at mid-body 10-12, usually 6 at region of
fore limbs; 45-50 preanal and femoral pores in a continuous series.

vankampeni

3. a. Dorsum usually marked by a pattern of light and dark bands or

distinct dark , blotches of varying size; or if melanistic, lacking promi-
nent, white tubercles 4

b. Dorsum dark brown with very faint darker blotches enclosing promi-
nent, white tubercles, which tend to form narrow, irregular, partial or
complete transverse series; a continuous series of preanal and femoral
pore scales (females) preceded anteriorly by several rows of enlarged
scales, those in the preanal region merging with those of the venter;
no enlarged scales posterior to the pore series in the preanal region
derongo

4. a. One or more rows of enlarged femoral scales; upper labials usually not

greater than 12 ^

b. No enlarged femoral scales; 12-14 upper labials; 10 11 broad lamellae

' under basal portion of fourth toe; dorsum with a double or united

series of 5 or 6 rather large dark blotches between ear region and base

of tail, separated by light bands variably marked by 3 or 4 smaller

dark blotches; males without pores sermoivaiensis

5. a. Enlarged preanal pore scales in a shallow "/\" chevron 7

b. Enlarged preanal pore scales compressed into a narrow "/\" sunk in a

groove in males with 8-14 pores 6

6. a. Seven to 9 moderately narrow, dark, irregularly margined bands or

series of blotches between the ear region and the groin; 8-10 preanal

6 BREVIORA No. 417

pore scales bearing pores in males, preceded anteriorly by 1 or 2 rows
of much enlarged scales and followed posteriorly by a narrow cluster
of 8-12 enlarged preanal scales; preanal series widely separated from
a single row of much enlarged femoral scales; no femoral pores.

papuensis

b. Seven to 9 irregularly margined, dark bands or blotches between the
ear region and the groin; 12-14 preanal pore scales bearing pores in
males, preceded anteriorly by several rows of enlarged scales merging
with those of the venter and followed posteriorly by several rows of
enlarged scales which diminish gradually; several rows of enlarged
femoral scales continuous with the enlarged preanal series; a short

series of 4-6 femoral pores separated from the preanal series

marmoratus

7. a. Dorsum with five broad, dark, rather even-margined, transverse bands

or double series of blotches between the ear region and the groin;
26-28 irregular rows of rather small, unikeeled tubercles between lateral
folds at mid-body; a continuous series of enlarged preanal and femoral
pore scales bearing 38-80 pores^ for several males examined but in each
instance reaching the distal end of the femur, both preceded anteriorly
by several rows of enlarged scales merging with those of the venter in
the preanal region, followed posteriorly by several rows of enlarged

preanal scales that diminish gradually loiiisiadensis

b. Dorsum with 5 to 8 broad to narrow dark bands or series of blotches,
usually with irregidar margins, between the ear region and the groin;
20-22 irregular rows of tubercles between lateral folds at mid-body; a
continuous or interrupted series of preanal and femoral pore scales,
some bearing pores in males 8

8. a. Dorsmn with 5 dark transverse bands or series of blotches between the

ear region and the groin; males with a continuous series of preanal and

femoral pores 9

b. Dorsum with 7 or 8 dark transverse bands or series of blotches between
the ear region and the groin; a series of enlarged preanal scales bear-
ing 12-14 pores in males; often separated by 3 or 4 somewhat smaller
scales from the pore-bearing femoral series; in males the latter bearing
a median group of 0-5 pores and a distal group about 5-11 pores on
either side; both preanal and femoral series preceded anteriorly by
several rows of enlarged scales which in the body region merge ^vith
those of the venter; and in the preanal region also follo^\•ed posteriorly
by several rows of enlarged scales which gradually diminish in size.
mimikanus

9. a. A continuous series of preanal and femoral pores extending the length

of the femur, bearing in males an uninterrupted series of 60-70 pre-

^This wide range may reflect population differences, since in our small
sample those with the lowest number of pores were from Australia and those
with the largest number from the Solomon Islands.

1973 CYRTODACTYLUS DERONGO 7

anal and femoral pores; preanal pore series preceded by several rows
of enlarged scales merging with those of the venter, and followed
posteriorly by 3 or 4 rows of enlarged scales; femoral scales anterior to
the pore series exhibiting a gradual reduction in number of scales and
a resultant strongly tapered appearance; 20-24 lamellae and enlarged
scales beneath the fourth toe; small, roundish tubercles absent from

throat loriae

b. A series of enlarged preanal and femoral pore scales, bearing a con-
tinuous series of 38-42 preanal and femoral pores in males; 28-33
lamellae and enlarged scales under the fourth toe; throat with some
scattered small rounded tubercles (from description by Brongersma,
1934) novaeguineae

Literature Cited

Brongersma, L. D. 1934. Contributions to Indo-Australian Herpetology.

Zool. Meded., 17: 161-251, 2 pis.
DE Rooij, N. 1915. The Reptiles of the Indo-Australian Archipelago. I.

Lacertilia, Chelonia, Emydosauria. Leiden, xiv + 384 pp.
Inger, R. F. 1958. A new gecko of the genus Cyrtodactylus with a key to

the species from Borneo and the Philippine Islands. Sarawak, Mus.

Journ., 8: 261-264.

~t\J(\ r\( \ >

B R-feaT^ I 0 R A

Museum JAS^toOTparative Zoology

HARVfcTissN o()or>-9r>98

Cambridge, Mass. 28 December 1973 Number 418

MORPHOGENESIS, VASCULARIZATION AND
PHYLOGENY IN ANGIOSPERMS^- -"

G. Ledyard Stebbins^

Abstract. Evidence is reviewed to support the hypothesis that vascular
strands in the angiosperm flower which some botanists have regarded as
"vestigial" can be understood better if they are regarded as the result of
irregularities in development, which provides no indication with respect to
the alternatives of phylogenetic reduction vs. amplification. Nevertheless,
the concept of the conservatism of vascidar anatomy is supported by the
proljability that genes acting late in development can more easily give rise
to mutations that can become incorporated into a harmonious genotype
than can genes that act early in development. Examples from the develop-
ment of achenes in various genera of the family Compositae show that size
of mature achene is not necessarily correlated with complexity of vascular
anatomy, and that this anatomy may reflect the particular course of develop-
ment, particularly the time when procambial initials are differentiated. In
this family, genera that are generally regarded as more closely related to
each other tend to have more similar developmental patterns than those
that are more distantly related.

Ever since the 19th-century research of Celakovsky (1896),
botanists have asked the question: "Is the arrangement of vascu-
lar bundles in the organs of higher plants a more reliable guide
thar) outward form to homology and the direction of evolution?"
Until verv recently, the usual answer has been affirmati\'e
(Eames, 1931, 1961; Puri, 1951, 1952; Melville, 1962), al-

^Much of the material in this paper is reproduced from the author's book:
Flowering Plant Evolution Above the Species Level, Harvard University
Press (in preparation) , through kind permission of the Press.

-This paper is respectfully dedicated to my former teacher and mentor,
Ralph H. Wetmore, who was largely responsible for developing my interest
in comparative plant anatomy.

•Department of Genetics, University of California, Davis

2 BREvioRA No. 418

though botanists have differed widely with respect to interpre-
tations of anatomical structure. In particular, single vascular
bundles that appear to have no function have been designated
as "vestigial." They have been interpreted as vestiges of organs
that are no longer formed, and therefore as indicating wide-
spread, predominant trends of reduction. Furthermore, the
concept of "fusion" has been adopted to interpret situations in
which two related species or genera differ with respect to the
number of parallel bundles found in an organ. If a form has
two parallel bundles in a particular position, it is regarded as
more generalized or primitive than a related form that has only
one bundle in that position.

During the last decade, botanists have become increasingly
skeptical of such notions. An extreme form of this skepticism
has been expressed by Carlquist (1969). x\fter an extensive
review of the entire problem, he reaches the following conclusion
(p. 334) : "Anatomy of flowers can be studied meaningfully
only in relation to adaptations for particular modes of pollina-
tion, dispersal and allied functions."

In my opinion, neither the rigid interpretations of Eames,
Puri, Melville and their followers nor the complete skepticism of
Carlquist are justified. Later in this article, examples are given
to show that when comparing even such similar and certainly
homologous structures as the achenes of different Compositae,
one finds many exceptions to a supposed correlation between
organ size and complexity of \'ascularization. On the other hand,
se\'eral examples exist in the literature to show that supposed
"vestigial ]:)undles" can be associated with either increase or
decrease in numbers of parts. One of the clearest of these was
presented long ago by Murbeck (1914). In two species belong-
ing to the family Rosaceae, Comarum palustre and Alchemilla
vulgaris (sens. lat. ), he found rare de\iations from the normal
or modal number of calyx lobes, in both an upward and a down-
ward direction. In Alchemilla, for instance (Fig. 1), the normal
number of lobes is four, but occasional flowers have three lobes
and others have five. Most important, however, is the fact that
among 3-lobed as well as among 4-lobed calyces are examples
in which one of the lobes is larger, and may have a double-
pointed apex, as well as extra vascular bundles. According to
the classical interpretation, such 3-lobed calyces result from
a trend of reduction via "fusion," and the extra bundles found
in the larger lobe are "vestigial." If, however, this interpretation

1973

ANGIOSPERMS

Figure 1. Calyces of individuals of Alchemilla vulgaris, showing devia-
tions from the normal 4-merous condition in the direction of both decrease
and increase in lobe number, as well as intermediate situations with ab-
normal lobe number and structure. From Murbeck, 1914.

4 BREVIORA No. 418

is to be consistent, the larger lobes of the aberrant 4-lobed calyces
would have to be interpreted in the same way, and the conclu-
sion would have to be reached that the basic number of calyx
lobes in Alchemilla vulgaris is five rather than four. Such an
interpretation is contradicted by the fact that 4-merous calyces
are found throughout the genus Alchemilla, except for rare aber-
rant indi\iduals like those described by Murbeck. In Comarum
palustre, similar aberrant calyces have five lobes, one of which
is larger than the others and contains extra vascular bundles. If
one held strictly to the concept of reduction and vestigial bun-
dles, one would have to interpret these calyces as indicating that
the calvx of Comarum was oris^inallv hexamerous. Since hex-
amerous calyces are almost completely lacking, not only in the
family Rosaceae but also in the entire order Rosales, such an
interpretation is absurd.

A MORPHOGENETIC InTREPRETATION OF

"Vestigial Bundles"

These examples are best interpreted by discarding entirely the
concept of reduction and vestigial bundles, as well as any other
phylogenetic concept, and regarding them entirely in the light of
developmental genetics. The aberrant calyces found by Murbeck
are comparable to the aberrant corollas described by Huether
(1968) in Linanthus androsaceus, and shown by him to repre-
sent unusual gene combinations that render the plant more sus-
ceptible than normal individuals to producing aberrant pheno-
types, or phenodeviants, as a result of normal environmental
fluctuations during development. Deviations from the normal
or modal condition can occur in either direction. Using a de-
velopmental approach, they can be explained on the basis of a
formula that I suggested a few years ago (Stebbins, 1967). The
number of similar organs or parts that are produced in a

particular whorl can be represented by the quotient A" + ^. ,

where A" is the final number of parts, a'" is the total number of
meristematic cells that are capable of producing an A-type part,
and a' is the number of meristematic cell initials needed to pro-
duce a single A-type part.

Applying this formula to Murbeck's examples, one could sug-
gest that in the normal development of the calyx of Alchemilla,
the relation of a"^ to a' is on the order of 20 to 5, so that A" = 4.

1973 ANGIOSPERMS 5

In the extreme aberrants, a' remains the same, but a" has become
respectively 15 and 25. On the other hand, 3-lobed calyces of
which one lobe is lar2:er and has extra bundles would result from
values such as a"^ = 17 and a^ = 5, so that A" = 3.4. Similarly,
abnormal 4-lobed calyces would represent the quotient A" = 4.4,
resultins: from values of A"^ = 22 and a' = 5.

Morphogenetic evidence with respect to "vestigial" bundles
in the androecium of various species belonging to the order
Malvales has been obtained by van Heel (1966). He showed
that in several instances vascular bundles, which in the mature
flower were not associated with any recognizable structure,
ne\'ertheless appeared in a position where small stamen primordia
could be recognized in early stages of development. These
primordia later became enveloped by the growth of the sur-
rounding tissue, presumably produced by persistent intercalary
meristems. These examples could be regarded either as terminal
stages of a reduction series, or intermediate stages of a trend
toward amplification.

The most convincing evidence regarding the morphogenetic
significance of vascularization comes, however, from experiments
in which the conditions under which vascular tissue appears
have been determined, or have been altered in specific ways.
Only two such experiments are known to me. One of them, by
\\'etmore and Rier (1963), showed that vascular tissue arises
in callus tissue at positions that are at regular distances from
each other, and that their distributional pattern can be altered
as a result of relatively slight alterations in the nutritive medium.
Consequently, the appearance of a bundle in an unexpected
position requires only a slight shift in the distribution of nutri-
tional factors or in the balance of hormonal interactions within
the developing system.

In the other experiment, Torrey (1955, 1957) altered ex-
perimentally the number of protoxylem points in a pea root.
He^ found that when 0.5 mm of the distal portion of the root,
containino- onlv cells that are not visiblv differentiated, was iso-
lated and cultured in vitro, the great majority of cultures pro-
duced roots having the normal triarch condition. About 2 percent
of the cultures, however, which were tips of relatively small size,
produced at first diarch roots, which later reverted to the triarch
condition.

If to the culture he added indole acetic acid at a concentration
of lO""" molar, he obtained a greater proliferation of the cells

6 BREVIORA No. 418

from which vascular tissues are differentiated. As a result, he
converted the triarch to the hexarch condition, and found that
the latter condition persisted indefinitely. The number of pro-
toxylem points could, therefore, be increased or decreased, de-
pending upon the amount of meristem present when procambial
differentiation took place.

These two experiments suggest that much can be learned
about the processes that affect the pattern of vascularization
by various kinds of experimental approach. This is a field of
morphogenesis that has not yet been well developed but that
promises eventually to provide a bridge over which visible
changes in vascular anatomy can be related to specific alterations
of the genot\pe, as they affect developmental processes.

Vestigial Characters in Plants and Animals

The results just reviewed suggest that with respect to any
group of similar structures, such as parts of a perianth, stamens
in an androecium, or "carpels" in a gynoecium, evolutionary
change can involve either increase or decrease in number, and
that the anatomical features associated with either trend are
similar to each other. Vascular anatomy cannot tell us whether
or not the ancestors of a particular form had more or fewer
sepals, petals, stamens, or carpels.

The belief of plant anatomists that this is possible rests, in my
opinion, on a mistaken analogy with the genuine vestigial struc-
tures found in animals. These latter, such as the gill slits of the
x'ertebrate embryo and the vermiform appendix, have a complex
and distinctixe developmental pattern. The so-called "vestigial
bundles," on the other hand, are identical in structure with the
bundles that are unquestionably functional. Furthermore, the
procambial cells that form the xylem and phloem of these bun-
dles are probably differentiated from meristematic cells during a
single mitotic cycle (Olson et ai, 1969). More important, the
epigenetic sequence responsible for the formation of these bun-
dles is an exact repetition of a course of events that occurs in
many other parts of the plant; only the position w^here it occurs
is distincti\e.

A Developmental Hypothesis That Favors
Conservatism of Vascular Anatomy

The concept of vestigial bundles is part of a broader concept

1973 ANGIOSPERMS 7

that \'iews \'asciilar anatomy as more conservative than external
morphology. This concept has been rejected by Carlquist ( 1969)
as an "insufficient and fallacious framework on which most
phylogenetic interpretations of floral anatomy still rest." He
ne\'ertheless concedes that degree of union between vascular
bundles can be "conservative." Is there any logic to this ac-
ceptance of a part of the doctrine of conservatism, after most
of it has been rejected?

I belie\e that botanists must examine the problem from the
viewpoint of developmental genetics and morphogenesis, since
this brings us closer to the basic nature of evolutionary changes.
When we do this, we can recognize and emphasize the fact that
the procambial initials from which vascular bundles arise become
differentiated from the ground meristem at a very early stage
of the de\'elopment of primordia. Consequently, alterations of
vascular pattern require changes in the time of action of genes
that normally act very early in development. Alterations in the
action of genes that normally act at later developmental stages
can produce changes in size or form without altering the pattern
of vascularization.

Is there any logical reason for assuming that genes which
produce their effects at early stages of development are less likely
to play a role in evolutionary change than genes which affect
later stages? A positive answer to this question is the genetic
basis for recognizing Von Baer's principle of embryonic similar-
ity, which was used by Darwin (1872) as embryological evi-
dence for evolution, and has been applied more recently to
animal development by De Beer ( 1 95 1 ) , and to plants by the
present author (Stebbins, 1950). The reasoning is as follows.
Adult characteristics are assumed to be the products of epi-
genetic sequences of gene action in development, so that later
processes depend in part upon the nature of gene products pro-
duced at earlier developmental stages. Moreover, the action of
most genes is pleiotropic in the sense that their primary products
may have many secondary effects. The earlier is this primary
action, the greater is the amount of pleiotropy that is possible,
and the more widespread are the secondary effects of genes.
Hence mutations of genes affecting early stages are more likely
to produce profound alterations of development, and hence to
upset the entire developmental system, than are mutations of
late-acting genes. The milder alterations produced by these
latter mutations are more likely to adjust the individual in a

8 BREVioRA No. 418

harmonious fashion to new selecti\'e pressures than are the more
drastic effects produced by mutations of genes that act early in
de\'elopment. Hence, adaptive alterations of morphology are
brought about more often by ]ate-acting genes than by those
acting early in de\'elopment. In other words, genes acting early
in de\elopment tend to be conser\'ative with respect to the estab-
lishment of their mutations in populations. Among such genes
are those that affect the differentiation of procambial strands.

Relationships Between Organ Size and
Amount of Vascularization

In the remainder of this contribution, I would like to apply
the theoretical concept just developed to two situations. The first
is the relationship between organ size and amount of vasculariza-
tion. If vascularization is related only to adaptation and physio-
logical function, as Carlquist has assumed, then large organs
should always have a proportionately greater amount of vascu-
larization than homologous, smaller ones. On the other hand, if
preferential establishment of late-acting gene changes is a sig-
nificant factor, then the relationship between size and vascular-
ization w^ould ha\'e a historical or evolutionary component.

Among homologous organs having approximately the same
size, but different patterns of vascularization, one might postulate
that the one having the more complex pattern resembles most
closely the most primitive organ of the group in question, while
the simpler pattern has been derived by a process of reduction
that affected early stages of development, followed by a reversal
of evolutionary direction, in which increase in size was accom-
plished by establishment of genes acting late in development.
Similarly, in comparisons between homologous organs of very
different sizes, but having similar, relatixely simple patterns of
vascularization, one might postulate that the smaller organ more
nearly resembles a reduced, ancestral form, and the larger one
has been deri\'ed via secondary enlargement.

Ovary and Achene Development in the
Family Com po sitae

A good object for testing these hypotheses is the ovary and
achene in the family Compositae. In different genera of this

1973

ANGIOSPERMS

f f

12-28 VASCULAR STRANDS

I I

\ »

10 VASCULAR STRAKDS

// n

5 V^CULAR STRAKDS

Figure 2. Mature achenes of various species of Compositae of which the
development is recorded in Tables 1 and 2. A, Helianthus annuus, wild
form from east of Davis, Calif. B, Helianthus annuus, cultivated variety
from Department of Agronomy, University of California, Davis. C, Wyethia
glabra, from Cache Creek Canyon, Yolo County, Calif. D, Senecio cruentus,
cult. var. "stellata" (smaller heads) . E, Senecio vulgaris, from campus. Uni-
versity of California, Davis. F, Microseris nutans, from Wright's Lake,
Eldorado County, Calif. G, Tragopogon porrifolius, from Locke, Sacramento
County, Calif. H, Stephanomeria exigua ssp. coronaria, from Antioch,
Calif. I, Microseris douglasii, from south of Dixon, Solano County, Calif.

10 BREVIORA No. 418

family, an enormous range of size exists between mature achenes
having a length of 1.4 mm to achenes 20 times as long, and
many-fold greater in bulk (Fig. 2). With respect to anatomy,
the most complex patterns consist of 26 to 28 parallel bundles
traversing the ovar\^ and achene (Stebbins, 1940), while in the
simplest ones, only two bundles are present (Stebbins, 1937).
The poor correlation between size and complexity of vascu-
larization is shown in Figure 2, which illustrates the mature
achenes of ten forms belonging to this family. In three of these
(A-C), the ovary and achene are traversed by 12 to 28 parallel
vascular strands, while in the remaining three (G-J) only five
are present. In the first group, achene length ranges from
2.92 mm to 13.65 mm; in the second, from 1.4 mm to 5 mm;
and in the third, from 3.8 mm to 28.5 mm. I admit that the
largest example of the latter group, Tragopogon porrifolius, was
chosen to represent an extreme example of large size associated
with a relatively simple vascular pattern, so that one cannot
conclude from this tiny sample that an inverse correlation exists
between achene size and amount of vascularization. Neverthe-
less, the lack of a significant positive correlation in the family as
a whole seems to me highly probable on the basis of my acquaint-
ance with a large number of genera.

In order to discover more about the relationships between
vascularization and developmental patterns, I have compared
the ovaries of these species at four stages of development:
( 1 ) the smallest size at which procambial strands can be recog-
nized ; ( 2 ) the first appearance of xylem tracheids ; ( 3 ) anthesis ;
and (4) mature achenes. Since the Composite achene increases
far more in length than in width, mean length of the ovary at
each of these stages is a reliable indicator of overall size. The
stages were determined both from sectioned material and from
whole mounts cleared according to the schedule of Herr (1971)
and observed under Nomarski interference-contrast optics.

Preliminary results of this study are shown in Tables 1 and 2.
Table 1 gi\'es the mean lengths of the ovary and achene at four
different stages: differentiation of procambium; first differen-
tiation of xvlem strands, anthesis, and seed maturity. The
final column of this table gives the mean number of vascular
strands in the ovary at anthesis. Table 2 presents the mean
percentage growth increment for each interval between the
stages listed in Table 1. To obtain these values, the difference
between the length at a later stage and at the next earlier stage,

1973

ANGIOSPERMS

Table 1. Lengths of ovaries and achenes of some species and varieties of
Conipositae at selected stages.

Procambial Xylem
differen- differen-
tiation tiation

(P)

(X)

Xylem
strands
An thesis Maturity at

(A) (M) an thesis

0.253mm 0.631mm 11.25mm 13.65mm 12-17

Species or variety
Wyethia glabra
Helianthiis bolanderi
ssp. exilis

Helianthus annuus
wild (neai" Davis, Cal.)
Helianthus annuus
cultivated
Senecio cruentus
cult, small heads
Senecio cruentus
cult, large heads
Senecio vulgaris
Microseris nutans
Microseris douglasii
Stephanomeria exigua
Tragopogon porrifolius

Table 2. Proportional growth increments at successive stages of ovaries of
Conipositae. Symbols explained in Table 1, and in text.

0.198

0.291

2.01

2.92

19-21

0.251

0.38

1.596

5.52

18-24

0.208

0.442

9.90

13.65

26-28

0.234

0.732

0.868

1.43

0.228

0.61

1.41

1.66

0.186

0.772

1.135

2.35

0.294

0.997

1.366

5.04

0.194

0.999

1.67

4.96

0.205

0.524

1.449

3.86

0.242

0.934

1.912

28.5

X-P

Species or variety

Wyethia glabra

1.49

16.8

0.23

Helianthus bolanderi

ssp. exilis

0.47

5.91

0.45

Helianthus annuus

wild

0.51

3.20

2.46

Helianthus annuus

cultivated

1.12

21.40

0.38

Senecio cruentus

cult, small heads

2.12

0.17

0.65

Senecio cruentus

cult, large heads

1.70

1.31

0.18

Senecio vulgaris

3.10

0.47

1.07

Microseris nutans

2.39

0.37

2.69

Microseris douglassi

4.15

0.67

1.97

Stephanomeria exigua

1.56

1.77

1.66

Tragopogon porrifolius

2.85

1.05

13.91

BREVIORA

No. 418

WyetKia
glabra

Helian-thus
bolanderi

Helianthui
annuus wild

Small l^eacis Large heads

Senecto cruen+us

Senecio
vulgaris

HeliantKus
annuus cult.

Microseris
nutans

fiicroseris
douglasM

StepKanomeria
exigua

Traqopogon
porrifolius

Figure 3. Chart showing diagramatically the growth increments of ovaries
of Compositae, as recorded in Table 2.

1973 ANGIOSPERMS 13

i.e., the amount of growth during the interval, is divided by the
length at the earlier stage. In this way, growth during each
inter\'al between stages is expressed in proportion to the amount
of tissue or "meristematic capital" present at the beginning of
the interval under study. In Figure 3, the same results are pre-
sented graphically.

These figures show that the amount of growth which takes
place before the vascular pattern is laid down by procambial
differentiation is only a small percentage of the total growth of
the organ. Moreover, this percentage varies greatly from one
species to another. The size of the primordium at the time of
procambial differentiation is similar in all of the species studied,
ranging from 186 micra in Senecio vulgaris to 294 micra in
Microseris nutans. This range is far less than the extreme differ-
ences in size between mature achenes, so that the percentage of
growth in length that takes place before procambial differentia-
tion ranges from high figures to 14 to 16 percent in Senecio
cruentus to the extremely low figure of 0.9 percent in Tragopo-
gon porrifolius.

Two obvious conclusions can be made from these results.
First, developmental patterns differ widely from one species to
another of this family, and may even differ between varieties of
the same species, as in Helianthus annuus and Senecio cruentus.
Second, each of the tribes represented possesses a characteristic
series of patterns that are different from those found in other
tribes. In the Heliantheae, for instance, the greatest percentage
increase in size occurs between procambial differentiation and
xylem differentiation. The Cichorieae are more variable in this
respect, but show a greater tendency than other tribes toward
growth between anthesis and achene maturity.

A further conclusion can be drawn by comparisons between
members of the same tribe. In both of the comparisons between
cultivated varieties of the same species: wild vs. cultivated
Helianthus annuus and the two cultivated varieties of Senecio
cruentus, the greatest difference exists with respect to size in-
crease between xylem differentiation and anthesis, a stage during
which few or no mitotic divisions are taking place. In Senecio,
this is also the stage at which the greatest difference exists be-
tween the two species studied: S. vulgaris and S. cruentus. In
the Heliantheae, the two wild species of Helianthus differ most
from Wyethia glabra with respect to the increase at this stage,

14 BREVIORA No. 418

but the greatest difference between H. annuus and H. Bolanderi
is with respect to the stage between anthesis and seed maturity.
In the Cichorieae, the most divergent species, Tragopogon por-
rifolius, differs most from the others with respect to this last stage.

These results support, in general, the hypothesis that later
developmental stages are more easily modified at the level of
varieties and species than are early stages. In all of the varietal
and species comparisons, except for the species of Microseris,
stages after xylem differentiation differ more than do earlier
stages. Furthermore, the size of the primordium at the time of
procambial differentiation is strikingly similar among all of the
forms studied, at least in comparison to the much greater differ-
ences between their mature achenes. Finally, with respect to the
two examples of artificial selection for increased size, genetic
changes affecting later stages were established in preference to
those affecting earlier stages.

The comparison between the two species of Microseris pro-
vides a significant exception to the above generalization. The
annual species, M. Douglasii, differs from the perennial M.
nutans with respect to the smaller size of the o\'ary primordium
at the stage of procambial differentiation, and the proportion-
ally greater amount of growth that takes place between this stage
and that of xylem differentiation. This suggests that Af. Doug-
lasii arose from its perennial ancestor, which certainly was not
M. nutans, but may have been a species having a similar devel-
opmental pattern, via reduction in the size of the ovary primor-
dium, accompanied or followed by compensatory growth at later
stages. This reduction, which affected an early developmental
stage, may have been responsible for the reduction from ten
ovarian bundles, which is characteristic of M. nutans and other
perennial species of Microseris, to five bundles, as found in most
or all of the annual species, including M. Douglasii.

This small and admittedly inadequate sample supports, as far
as it goes, the hypothesis that large achenes having simple
vascular patterns are deri\ed by secondary enlargement from
smaller ones having similar vascularization. VV^ith respect to the
hypothesis that simplification of vascular pattern takes place via
a "bottleneck" of reduction that affects early developmental
stages, followed by secondary enlargement, the present evidence
is inconclusive. I hope, however, to obtain an answer to this
question when the study is complete.

1973 ANGIOSPERMS 15

A Basis For Differentiating Between Primary
AND Secondary Union of Parts

The second kind of situation that I would like to discuss con-
cerns the validity of vascular patterns as evidence for the phylo-
genetic origin of "fusions" and "adnations" between parts. This
topic has been much discussed in connection with the origin of
the inferior ovaiy, or epigyny (Douglas, 1957; Kaplan, 1967).
The extreme skepticism of Carlquist ( 1969) with respect to such
evidence has been challenged by Kaplan (1971), who in my
opinion has successfully answered many of Carlquist's criticisms.
At any rate, since diverse vascular patterns are found in various
genera having epigynous gynoecia, is association with other very
dififerent morphological characteristics as well as affinities to
various groups having perigynous or hypogynous gynoecia, this
evidence indicates strongly that the epigynous condition has been
evohed many times independently in different orders of plants,
by various evolutionary pathways.

In my discussion, however, I should like to focus attention on
the androecium. The "fusion" of stamens into bundles or a
tubular staminal column that includes the entire androecium is a
familiar feature in several plant families, particularly the Mal-
vaceae, Sterculiaceae, Hypericaceae (Guttiferae), Myrtaceae,
and some genera of Dilleniaceae. This "fusion" is generally
regarded as secondary (Eames, 1961), and in most instances
this conclusion is well justified. Developmentally, it is most often
brought about by a suppression of differentiation with respect to
stamen filaments. Instead of separate intercalary meristems that
produce the growth of each individual filament, a common
meristem elevates some or all of the anther primordia on a single
column, tube or sheath (van Heel, 1966).

Recent developmental studies, however, suggest that not all
"fusions" between stamens are of this secondary kind. In Pae-
oni'a (Hiepko, 1965) and Hypericum (Leins, 1964; Robson,
1972) careful analyses of the development of floral primordia
have shown that stamen bundles, not individual anther pri-
mordia, fit into the phyllotactic sequence that is followed by the
other floral parts. Furthermore, anther primordia arise not from
the undifferentiated meristem of the reproductive axis, but from
distinct primordia of stamen bundles. Their differentiation pre-
cedes the activity of the intercalary filament meristem, which in

16 BREVIORA No. 418

Paeonia and Hypericum ele\'ates each stamen upon a separate
filament.

The anatomical condition that follows this developmental
pattern is that of a common "trunk" vascular strand for each
cluster of stamens that are differentiated from the same bundle
primordium. The vascular strands that supply indi\idual sta-
mens di\erge from the "trunk" strand, not directly from the
floral axis.

Examination of the \'ascular anatomy of the mature androe-
cium in a number of relatively primiti\'e angiosperms, such as
Degeneria (Swamy, 1949), Hibbertia (Wilson, 1965), and
certain Annonaceae {Cananga, Goniothahnus, unpublished ob-
servations of the present author), has revealed the same kind of
bundle pattern in them. In most instances, this pattern is not
accompanied by an ob\ious clustering of the stamens in the
flower as view^ed externally. This condition leads me to believe
that, although in some instances such stamen bundles may have
been deri\'ed from single stamens by a process of multiplication
of another primordia, or "dedoublement," as Leins (1964, 1971 )
maintains, this has not always been so. Conclusions based upon
comparisons between o\ules and megasporophylls, which will be
presented elsewhere, have led me to believe that among known
fossil forms, those most nearly related to ancestors of the angio-
sperms are the cupule-bearing Pteridosperms such as Caytoniales
(Thomas, 1925) and Corystospermaceae (Thomas, 1933). If
this hypothesis is correct, then the structure of the microspro-
phylls in these forms should be considered. In no case do they
consist of flat structures bearing sporangia upon their surfaces,
as would be expected on the basis of the "classical" concept of
the origin of stamens (Eames, 1961 ). They are always branched,
and bear numerous microsporanma at the ends of the branches.
The stamen bundles in genera like Paeonia could be derived
from such microsporangiophylls by suppression of their branches.
This discussion can be summarized by stating the hypothesis
that "fusions" of stamens are of two kinds. The existence of
stamen bundles that are evident chiefly from examination of the
vascular pattern, and are seen with difficulty or not at all when
one examines the external structure of the flower, represents a
primary fusion, which takes place at the very earliest stage of
androecial de\elopment, and reflects an ancestral condition. On
the other hand, the staminal tube of the Malvaceae, and the
elevated clusters of stamens that are found in many genera of

1973 ANGIOSPERMS 17

Hypericaceae and Myrtaceae, as well as similar structures in
\'arious other families, are secondary in origin, and are produced
by intercalary meristems that appear relatively late in develop-
ment, after the anther primordia are fully differentiated. This
hypothesis is entirely in accord with that of conservatism of gene
complexes affecting early de\'elopmental stages.

A Plea For Further Research in the Field

OF MORPHOGENETIC TaXONOMY

The account which I have just given of the comparative de-
\'elopment of achenes in the Compositae reports only the begin-
ning of a small piece of research. Nevertheless, it shows that
careful comparisons between developmental patterns of selected
organs in a series of closely related forms can reveal similarities
and differences that are not evident from examinations of mature
organs. Moreover, some of these differences in pattern can serve
as a guide to evolutionary direction.

In their efforts to broaden their field, botanists have, in recent
years, been relying to an increasing extent on characteristics
other than external morphology. Cytotaxonomy, based upon
chromosomal differences, has been with us for a long time. More
recently, chemotaxonomy has increased in popularity, and is
yielding highly significant results. In my opinion, the essentially
undeveloped field of morphogenetic taxonomy also needs to be
developed. Its potential importance lies in the prospect that it
may contribute more to our understanding of morphological
taxonomy than any other field. The cytotaxonomist studies
chromosomes as they appear during mitosis, when the DNA is
condensed into neat packages, and the genes are inactive. In-
numerable studies in this field have shown us that the number
and shape of these "packages" is much less important for adap-
tation, survival, and ecological distribution than is the nature of
the genes contained in them. Chemotaxonomists, because of the
cornplexity of their field, have been forced to concentrate upon
certain compounds and properties largely because of technical
considerations that determine the ease of study rather than cri-
teria of evolutionary significance. We have, therefore, many
systematic comparisons of secondary and accessory compounds
such as phenolics and terpenes, as well as of a single property,
electrophoretic mobility, possessed by those proteins that are
easily isolated and recognized. Important as these investigations

18 BREvioRA No. 418

are, they explore only the fringes of the biochemical systems of
the organisms concerned.

The potential value of morphogenetic taxonomy arises from
the fact that adult structures appear as a result of patterned
sequences of gene action in development. Groups of genes are
acti\'ated and deactivated according to a specific program that
is controlled by a complex system of regulator genes (Britten
and Davidson, 1969). Morphological evolution must be based
ultimately upon mutations and recombinations of these par-
ticular genes. By developing the discipline of morphogenetic
taxonomy, botanists may be able to approach closer to an under-
standing of how these genes work, and how they change during
evolution.

Literature Cited

Britten, R. J., and E. H. Davidson. 1969. Gene regulation for higher cells:

a theory. Science, 165: 349-357.
Carlquist, S. 1969. Toward acceptable evolutionary interpretations of

floral anatomy. Phytomorpholog)', 19: 332-362.
Celakowsky, L. 1896. tjber den phylogenetischen Entwicklungsgang der

Blute. Sitzber. K. Bohm. Ges. Wiss. Math. nat. Kl., 1896: 1-91.
Darwin, C. 1812. The Origin of Species. 6th London Edition.
DeBeer, G. R. 1951. Embryos and Ancestors, Revised Edition. Oxford

University Press.
Douglas, G. E. 1957. The inferior ovary. II. Bot. Rev., 23: 1-46.
Eames, a. J. 1931. The vascular anatomy of the flower, with refutation of

the theory of carpel polymorphism. Amer. J. Bot., 18: 147-188,
. 1961. Morphology of the Angiosperms. New York: McGraw

Hill.
Heel, W. A., van. 1966. Morphology of the androecium in the Malvales.

Blumea, 13: 177-394. .
Herr, J. M., Jr. 1971. A new clearing-squash technique for the study of

ovule development in angiosperms. Amer. J. Bot., 58: 785-790.
HiEPKo, P. 1965. Das zcntrifugale Androecium von Paeonia. Ber. deu. bot.

Ges., 77: 427-435.
HuETHER, C. A., Jr. 1968. Exposure of natural genetic variability under-
lying the pentamerous corolla constancy in Linanthus androsaceiis ssp.

androsaceus. Genetics, 60: 123-146.
Kaplan, D. R. 1967. Floral morphology, organogenesis and interpretation

of the inferior ovary in Downingia bacigalupii. Amer. J. Bot., 54:

1274-1290.
. 1971. On the value of comparative development in phylo-

genetic studies — a rejoinder. Phytomorphology, 21: 134-140.

1973 ANGIOSPERMS 19

Leins, p. 1964. Die fiiihe Bliitcnentwicklung von Hypericum hookerianum

Wight ct Arn. iind H. aegypticum L. Ber. deu. bot. Ges., 77: 112-123.
. 1971. Das Androccium der Dicotylen. Ber. deu. bot. Ges., 84:

191-193.
Melville, R. 1962. A new theory of the angiosperm flower: 1. The

gynoeciutn. Kcw Bull.. 16: 1-50.
MuRBFCK. S. 1914. ubci die Baumcchaiiik bei Andeiungen ini Zahlen-

veihiiltnis der Bliite. Lunds Univ. Arsskr., N.F., Afd. 2, 11(3): 1-36.
Olson. K. C, V. W. Tibbits, and B. E. Struckmeyer. 1969. Leaf histo-
genesis in Lactuca sativa with emphasis upon laticifer ontogeny. Amcr.

J. Bot., 56: 1212-1216.
Purl V. 1951. The role of floral anatomy in the solution of morphological

problems. Bot. Rev., 17: 471-553.

. 1952. Placentation in angiosperms. Bot. Rev., 18: 603-651.

RoBSON, N. K. B. 1972. Evolutionary recall in Hypericum (Guttiferae) ?

Trans, bot. Soc. Edinburgh, 41: 365-383.
Stebbins, G. L. 1937. Critical notes on Lactuca and related genera. J. Bot.,

75: 12-18.
. 1940. Studies in the Cichoricae: Dubyaea and Snroseris,

endemics of the Sino-Himalayan Region. Mem. Torrey bot. Club, 19:

1-76.
. . 1950. \'ariation and Evolution in Plants. New York:

Columbia l^niversit-v Press. 643 pp.

1967. Adaptive radiation and trends of evolution in higher

plants. In Evolutionary Biology. Ih. Dob74ian.sky, M. K. Hecht, and

Wm. C. Steere, eds. \\A. 1: 101-142.
SwAMv, B. G. L. 1949. Further contributions to the anatomy of the

Degeneriaceae. J. Arnold Arb., 30: 10-38.
Thomas, H. H. 1925. The Caytoniales, a new group of angiospermous

plants from the Jurassic rocks of Yorkshire. Phil. Trans, roy. Soc.

London, B, 213: 299-313.
. . 1933. On some pteridospermous plants from the Meso-

zoic rocks of South Africa. Phil. Trans, roy. Soc. London, B, 222: 193-265.
Torrey, J. G. 1955. On the determination of vascular patterns during

tissue differentiation in excised pea roots. Amer. J. Bot., 42: 183-198.
. 1957. On the determination of vascular pattern formation

in regenerating pea root meristems grown in vitro. Amer. J. Bot., 44:

859-870.
Wetmore, R. H., and J. P. Rier. 1963. Experimental induction of vascular

tissues in callus of angiosperms. Amer. J. Bot., 50: 418-429.
Wilson, C. L. 1965. Ihe floral anatomy of the Dilleniaceae. I. Hibbertia

Andr. Phytomorphology, 15: 248-274.

B R E XJ.n R A

LIBRARY

Miiseiiiii of Comparative Zoology

JAI^Y 1974 ^^

us ISSN 0006-9698

Cambridge, Mass. 28 December .l9|-% Number 419

PROTOPTYCHUS, A HYSTRIGOMORPHOUS

RODENT FROM THE LATE EOCENE

OF NORTH AMERICA

John H. Wahlert^

Abstract. The North American late Eocene Protoptychus Scott possesses
an enlarged infraorbital foramen, a depression on the side of the snout
anterior to this foramen for the origin of the anterior part of the middle
masseter, tetralophate P*-M^ an enlarged incisive foramen, a deep pterygoid
fossa, and apparently no stapedial foramen or carotid canal. These char-
acters also occin- in the Caviomorpha. With regard to the zygomasseteric
structure and acquisition of an essentially molariform P^, Protoptychus is
more advanced than both its possible North American ancestor, which may
be either a paranlyid or Mysops, and Platypittamys, the most primitive
Deseadan (Oligocene) caviomorph. The Protoptychidae, on present evi-
dence, cannot be related closely to any rodents other than these. Pending
further knowledge, the family is retained in the Protrogomorpha, but the
possibility exists that it may be a specialized offshoot from the North
American caviomorph ancestry.

Introduction

In the course of studying the cranial foramina of North
American protrogomorphous and sciuromorphous rodents, I ex-
amined the type skull of Protoptychus (Princeton University
11235) and a second, much damaged facial region (PU 11230).
I was immediately struck by features that set this form com-
pletely apart from all others I had at hand. These were the
unusual shape and great posterior extent of the incisive foramen,
the large size of the infraorbital foramen, the flatness of the sides

^American Museum of Natural History, Vertebrate Paleontology Depart-
ment, Central Park West at 79th Street, New York, N.Y. 10024

2 BREVIORA No. 419

of the snout, and the depression of an area on the snout anterior
and extending somewhat dorsal to the infraorbital foramen. I
was led, finally, to conclude that Protoptychus is a primitive
hystricomorphous rodent possibly allied to the ancestry of the
South American Caviomorpha. The lower jaw is present in
specimens that I have not seen which belong to the Field Mu-
seum of Natural History; TurnbuU (personal communication)
is in the process of preparing these for description.

Taxonomic History of Protoptychus

The monotypic genus Protoptychus has had a checkered his-
tory in the literature of rodent taxonomy. Scott, in describing
the skull of Protoptychus hatcheri from the Uinta deposits of
Utah, stated: "That Protoptychus is an ancestral form of the
Dipodidae seems abundantly clear." 'Tt is not improbable that
the Heteromyidae were derived from some form related to Pro-
toptychus, though not from that genus itself" ( 1895 : 280, 286) .
Matthew (1910: 68) followed Scott in associating the genus
with the Dipodidae. Schlosser (1911: 427) created the sub-
family Protoptychinae as one of two di\isions of the family he
termed Geomyoidea. Miller and Gidlev (1918: 443) placed
the subfamily back in the Dipodidae. Wood (1935: 239-240)
stated that the tooth structure did not indicate close relationship
to the Geomyoidea, and he noted that Schaub's studies on the
jumping mice and dipodids eliminated them also as relatives of
Protoptychus. He suggested that, instead, ". . . Protoptychus
may represent an aberrant and sterile offshoot of the Ischyro-
mvidae." Wood (1937: 261) formally raised the taxon to
familial rank, Protoptychidae, as a division of the Ischvro-
myoidea. Simpson (1945: 78) and Wilson (1949: 99-100)
followed Wood's familial designation and placement of the
genus. A diagnosis of the family was published bv Wood in
1955 (p. 171).

Dentition
Figure 1, a and b

In most respects Scott's description of Protoptychus hatcheri
(1895) is accurate, but there are a few points that require re-
consideration. He failed to notice the presence of a minute,
peglike third premolar, and the revised dental formula (as noted

1973

PROTOPTYCHUS

7 mm

Figure 1. Dentition of Protoptychus hatcheri (PU 11235) : a. left cheek
teeth, view perpendicular to wear surface; b. left incisor, cross section.
Dentition of Mysops parvus (USNM 18043) : c. left cheek teeth, view per-
pendicular to wear surface; d. left incisor, cross section.

by Wilson, 1937: 450) is thus P C ?' M\ P'-M' are bra-
chyodont and notably higher crowned lingually than labially;
although quite worn, they are clearly four-crested (Fig. la).
The most conspicuous feature of the crown is a mesoflexus,
which is broadest at the labial side and ends, at this stage of
wear, near the middle of the tooth. The crowns of M^"^ are
grooved in the middle of the lingual side, the groove fading
away well before reaching the base of the enamel; P* possesses
only a vague suggestion of this groove.

Although the four molariform cheek teeth are lophate, the
cusps are still readily compared with those in paramyid teeth as
figured by Wood (1962: 8, fig. lA). On the labial side the
paracone and metacone flank the mesoflexus. The protocone is

4 BREVIORA No. 419

anterior to the lingual groove, and the hypocone, posterior; the
crown is quadrate in outline. The paracone and protocone form
the protoloph; the metacone and hypocone, the metaloph. The
hypocone and protocone are already joined in the slightly worn
M^, and the metaloph is more broadly connected with the hypo-
cone than with the protocone. A small, low mesostyle is present
on the molars and is closely associated with the metacone in the
first molar and with the paracone in the second and third
molars; it increases in size posteriorly. No trace of it is to be
seen in P^. The four molariform cheek teeth possess both an
anteroloph and a posteroloph. These are subordinate in im-
portance to the two main crests on M^"", and are nearly equal
to them in prominence in M^.

Scott remarked ( 1895 : 270) that "the transverse crests visible
on M^ of Protoptychus (and doubtless in the unworn state of
the other teeth, also) have a certain resemblance to the teeth of
squirrels and spermophiles . . . ." In this he is correct because all
retain in the upper dentition a relatively primitive arrangement
of cusps. He continued, "... but the fundamental character of
the tooth pattern is given by the enamel invaginations, which
tend to di\'ide it into two prisms. This arrangement is most like
that found in Pedetes, the Heteromyidae and Geornyidae." The
mesoflexus, however, is not an invagination of the enamel from
the lingual side of the tooth, it is simply a valley in the enamel
between two worn crests; the crown is not divided into two
prisms.

The incisor enamel as seen in a peel from the transverse break
appears to be pauciserial. Pauciserial and multiserial enamels
are similar, and a transverse section is not ideal for distinguish-
ing them; the enamel is certainly not uniserial. Scott did not
figure the incisor in cross section; the distribution of enamel

Figure 2. Skull of Protoptychus hatcheri (PU 11235); dorsal, lateral, and
ventral views; sutures diagrammatic.

Key: stippled areas: bone missing, crushed, or matrix covered; dark area
on snout: site of origin of masseter medialis; hatched areas: cross section
of bone; dashed lines: structine reconstructed.

Bones: ah — auditory bulla, as — alisphcnoid. / — frontal, ip — interpari-
etal, / — jugal. ^ — lachrymal, /// — maxilla, nist — mastoid, // — nasal,
occ — occipital, as — orbitosi^henoid, p — parietal, pi — palatine. /;/// —
premaxilla, sq — squamosal. Foramina: bf — buccinator, // — interorbital,
iof — infraorbital, isj — incisive, ;/ — jugular, tnj — masticatory, o/ — optic,
paj — post-alar fissure, plj — palatine, sj — stylomastoid.

1973

PROTOPTYCHUS

<^:..I^.^

1 cm

6 BREVIORA No. 419

on its front surface (Fig. lb) is similar to that in many small
Eocene rodents, e.g., some species of Paramys, and of Franimys,
Sciuravus, and Adysops. In transverse section the front of the
incisor is less bowed than in these forms and has a marked
posterolateral slant relative to the sagittal plane; it resembles the
incisor of Platypittamys in this respect.

Skull
Figure 2

Scott's description of the skull is adequate and accurate for the
most part, but a few additional points can be made. The pos-
terior extension of the nasal bones almost as far back as the
middle of the orbits is, to my knowledge, unique to Protoptychus
among rodents.

The auditory region is greatly inflated, and both the temporal
and mastoid portions of the skull participate in this inflation.
Scott stated that the "... mastoid bulla ... is divided by partial
septa into chambers, two of which are plainly shown, e\'en ex-
ternally, being bounded by deep grooves" (1895: 275). The
two \dsible septae are seen o-nly at the surface, and their extent
is unknown. The region closely resembles that in Chinchilla
except that there is no trace of a supraoccipital process that
reaches the squamosal. In Chinchilla partial septae are present
in the epitympanic sinus.

The parietal overlaps the dorsal epitympanic sinus laterally,
and a narrow process of the parietal extends posteriorly beside
the interparietal, apparently reaching the mastoid. Scott's dorsal
view of the specimen (p. 270, fig. 2) shows the process arising
from the parietal, although he incorrectly states in the text that
the squamosal "... appears to send out a process between the
parietal and the mastoid, which articulates with the interparietal"
(1895: 276). The compression of the posterior part of the
parietal and the unusual rectangularity of the interparietal seem
to be in response to the great dorsal inflation of the epitympanic
sinus. The back of the skull roof retains the primiti\e flatness
and sharp angle with the occipital surface; it does not curve
downward onto the occipital surface as it does in dipodids,
heteromyids, and those caviomorphs in which the auditory region
is also greatly inflated.

Many of the cranial foramina are preserved in the type speci-
men. The incisive foramina, unlike those of any protrogomor-

1973 PROTOPTYCHUS 7

phous rodent, are unusually long, extending back to the middle
of the fourth premolar, and their lateral margins are intersected
anterior to the middle by the premaxillary-maxillary suture.

The infraorbital foramen is conspicuously larger dorsoventrally
than that of any protrogomorphous rodent. The sides of the
snout are flattened, and the course of the incisor root stands out
as a swelling. Just anterior to the infraorbital foramen and ex-
tending somewhat dorsal to it is a depression on the side of the
snout; this area appears to have been the site of origin of the
anterior part of the medial masseter, which must have passed
through the infraorbital foramen. Protoptychus was hystri-
comorphous.

In the orbital region, three foramina are visible. The optic
foramen, of which only the ventral margin remains, is clearly a
large aperture in comparison with those of paramyids, and is
probably the structure which Scott (1895: 278) called "a large
sphenoid fissure." Antero ventral to the optic foramen in the
orbitosphenoid is a small aperture, possibly an interorbtial fora-
men. A foramen occurs in this position in various unrelated
rodents, e.g., Ischyromys, Geomys, and questionably in Castor,
and I attach no special taxonomic significance to its presence
here. In the floor of the orbit is a dorsal palatine foramen, which
transmitted the descending palatine artery. In Paramys this fora-
men shares a common opening with the sphenopalatine, whereas
in Protoptychus, as in Sciuravus, the foramen is in the orbital
floor posterolateral to the sphenopalatine foramen. The posterior
palatine foramen, the exit for the artery, is wholly within the
palatine, the primitixe condition for rodents.

The margin of the sphenoidal fissure and most of the region
where the aHsphenoid, parietal, frontal, and orbitosphenoid come
close together is crushed. The masticatory and buccinator fora-
mina open upward and forward, respectively, near the back of
the alisphenoid bone. Retention of separate foramina for the
masseteric and buccinator nerves is a primitive rodent character.
Posterior to the buccinator foramen there is an emargination of
the alisphenoid, which, with the anterior side of the bulla, makes
a foramen. A multiple aperture in the position is present in
Reithroparamys; there is no comparable foramen in other para-
myid skulls or in Sciuravus.

The postglenoid and the temporal foramen are absent, prob-
ably because of the greatly inflated bullae. The stapedial fora-
men, carotid canal, and mastoid foramen appear to be absent.

8 BREVIORA No. 419

but they (especially the last two) may have been obliterated
by the slight lateral crushing which the specimen has suffered.
The pterygoid fossa is very deep, and inadequately preserved
for full description.

Discussion

By the process of elimination it is possible to rule out relation-
ship to any rodent group except the Paramyidae, the genus
Alysops, and the Ca\'iomorpha. Of the protrogomorphous ro-
dents, all but the Paramyidae and Mysops are significantly dif-
ferent from Protoptychus.

In 1959 Wood (p. 359) thought that the Protoptychidae
might have been deri\'ed from the Sciuravidae; sciuravids are
primiti\e in most skull characters and in this respect could be
ancestral. However, the cheek teeth and their incipient crests
are not nearly so primitive. Unlike the condition in Protoptychus
and paramyids, the medial valley of the crown is open lingually
and blocked labially by the mesostyle. Wilson (1949: 91) noted
this and other characteristics of the cheek teeth as being markedly
different from those of most paramyids.

The cheek teeth of Protoptychus are advanced over those of
paramyids in that the third premolar is greatly reduced, the
fourth premolar and third molar are tetralophate, and the
metaloph is more closely connected with the hypocone than with
the protocone. The major cusps, howe\ er, are still readily identi-
fiable, and the anteroloph and posteroloph are not quite equal in
prominence to the crests formed by these cusps. The basic pat-
tern is most nearly comparable to that of Paramys and Reithro-
par amy s. Some reduction of the third premolar has already
occurred in Reithro paramys. Wood (1962: 248) tentatively
suggested derivation of Protoptychus from Reithro paramys but
stated, "On the other hand there are some undescribed specimens
(including skeletons) that seem to suggest other relationships
for Protoptychus-" These remain undescribed.

The cheek teeth of the Ischyromyidae (including only 7^-
chyromys and Titanotheriomys) are very similar. However, the
infraorbital foramen is much smaller, and the zygomatic plate is
tilted, indicating a trend toward a sciuromorphous type of masti-
catory musculature\ The dorsal palatine foramen is well inside

^Having examined the evidence, I agree with Wood (1937: 195) rather
than Black (1968: 275) on this point.

1973 PROTOPTYCHUS 9

the sphenopalatine foramen; the pterygoid fossa, though well
developed, is not nearly so deep; and there is a well-defined caro-
tid canal in ischyromyids.

The cylindrodontids\ specifically Ardynomys, which has four-
crested cheek teeth, differ in detail. The dorsal palatine foramen
is not separated from the sphenopalatine; the pterygoid fossa is
shallow, and the carotid canal is present although small.

The Eocene rodent that most closely resembles Protoptychus
is Mysops. There are three differences between the molariform
teeth of the two genera (cf. Fig. Ic and d). In Mysops the
anteroloph of P^ is not fully developed as a continuous crest;
the metaloph is incomplete and does not meet the hypocone,
though its trend is toward the anterior part of that cusp; and
whereas in Protoptychus the cusp is prominent, in Mysops it is a
very minor one. As seen in transverse section, the incisors of
Mysops are very similar to those of Protoptychus, but the an-
terior surface is more bowed. The alveolus for P^ indicates that
in Mysops the tooth was not reduced. A striking bit of evidence
for relationship between the two genera is that in Mysops the
length ratio of the incisive foramina to diastemal length exceeds
.60, a ratio greater than that known for any protrogomorphous
rodent (Wahlert, 1972). x\lthough the foramina do not extend
as far back as the first premolar, as in Protoptychus, their size
suggests a stage intermediate between a paramyid or sciuravid
and Protoptychus.

The Aplodontoidea, even the earliest ones, are so different in
cusp pattern that close relationship to them can be ruled out.
Prosciurids, which are most likely ancestral to aplodontoids,
differ in the same regard. In them the pterygoid fossa is not
deep, and there is a conspicuous stapedial foramen.

There is nothing about the dentition of Protoptychus that sug-
gests relationship to the Hystricidae, which, to judge from their
geologic record, mav have been of Oriental origin (Wood and
Patterson, 1970: 636).

The phiomyids, most notably Metaphiomys, bear some sim-
ilarity to Protoptychus in that they are hystricomorphous and
also have enlarged incisive foramina (Wood, 1968). The cheek

l^Vilson {e.g., 1949: 93) and Wood (personal communication) , on the
basis of dental similarity, place Mysops in the Cylindrodontidae. I hesitate
to accept this assignment because, in the one partial skull of the genus
(USNM 18043) , the incisive foramina are considerably longer relative to the
diastemal length than in Cylindrodon, Pseudocylindrodon, and Ardynomys.

10 BREVIORA No. 419

teeth, however, are quite different; the crown pattern of Pro-
toptychus is four-crested, whereas those of Phiomys and Meta-
phiomys are five-crested, the fifth crest being the mesoloph. Like-
wise the cheek teeth of the theridomyids differ in having five
crests.

Myomorphous rodents can be excluded from possible relation-
ship because the cheek tooth cusp pattern is essentially different.
All sciuromorphous forms can be eliminated because of their
zygomasseteric structure. Furthermore, the stapedial artery,
which may well have been lacking in Protoptychus, is retained
and its foramen is conspicuous in heteromyids and eomyids; in
sciurids the foramen is present although less easily seen.

The remaining group for consideration is the Caviomorpha.
The Caviomorpha are hystricomorphous ; many of the early
South American members of the group, e.g., the Deseadan
Cephalomys (Wood and Patterson, 1959: 343, fig. 21), Sal-
lamys and Incamys (Patterson and Wood, in preparation), and
se\eral Santacruzian genera illustrated in Scott ( 1 905 ) have
elongate incisive foramina. The living caviomorphs lack the
tympanic portions of both the stapedial and internal carotid
arteries (Guthrie, 1963: 478; Bugge, 1971: 532), as is quite
possibly the case in Protoptychus. The pterygoid fossa is \'ery
deep in caviomorphs.

The cheek teeth of Protoptychus are lophate and are based
on a series of four crests that are fully homologous with those of
primitive caviomorphs. Protoptychus retains a small but distinct
mesostyle on the molars which is lacking in caviomorphs, except
Branisatnys luribayensis, which has the cuspule on the second
molar (Hoffstetter and Lavocat, 1970: 172 and fig.); it lacks
the lingual valley, the hypoflexus, which is prominent in cavio-
morphs, but does have an indentation in that position. The
fourth premolar of Protoptychus is molariform, unlike those of
the more primitive Deseadan caviomorphs, Deseadomys, and
Platypittamys, but shows some resemblance to one specimen of
Sallarnys (Patterson and Wood, in preparation).

The incisors, as noted above, appear to have pauciserial
enamel. This is a plausible condition for a caviomorph relative,
since multiserial enamel was surely derived from pauciserial
( Korvenkontio, 1934; Wahlert, 1968: 13), and the two are not
very different, bands of the inner enamel layer in each being
several prisms wide.

The simplest taxonomic interpretation of Protoptychus is to

1973 PROTOPTYCHUS 11

call it a hystricomorphous member of the Protrogomorpha.
Structural details which are like those found in caviomorphs
would be attributed either to convergence or to parallelism stem-
ming from common ancestry within the Protrogomorpha. The
consequence of this interpretation would be that the hystrico-
morphous condition of the masseter and infraorbital foramen
arose more than once from the protrogomorphous condition, a
conclusion in keeping with the similar multiple origin of sciuro-
morphous musculature, e.g., independently in Titanotheriomys,
and with its presence as a component of the myomorphous con-
dition. Mysops may be a close relative of Protoptychus, but until
a good skull of the genus is known this can be taken as no more
than a possibility. The specialized characteristics of Protopty-
chus, especially those associated with the masseter and with the
auditory region, confirm the need for a separate family to receive
the genus.

Protoptychus could be a caviomorph, but, on the basis of the
earliest forms known, a rather complicated explanation would
be required. There are three anatomical barriers to placing
Protoptychus in the Caviomorpha: its precociously molariform
[i.e., four-crested ). fourth premolar, the lack of a distinct hypo-
flexus in the molars, and its hystricomorphous condition. Ac-
cording to Wood (1949) the most primitive Deseadan cavio-
morph^, Platypittamys, has only a slightly enlarged infraorbital
foramen, which did not transmit any part of the masseter, and
a simpler fourth premolar than any paramyid known at the time
of its description; whether the condition of the premolar was
primitive or reduced could not be determined. On the basis of
an undescribed Gray Bull paramyid. Wood and Patterson
(1959: 296-297) were able to ascertain that the absence of a
separate metaloph in the fourth premolar of Platypittamys and
some other Deseadan caviomorphs is primitive. The Gray Bull
paramyid, Franimys, was described by Wood in 1962 (pp. 139-
147). The fourth premolar is comparable and also simple.

Although the cheek tooth patterns of Protoptychus are closer
to those of Paramys, Reithroparamys, and Mysops, it is possible
to derive them from that of Franimys. The direct ancestor of
the South American Caviomorpha would then have been primi-

^The caviomorphs described by Hoffstetter and Lavocat (1970) from the
Deseadan of Bolivia are more advanced in that they already have enlarged
infraorbital foramina and the posteroloph in some is divided into two parts
(I do not agree that a mesoloph is present) .

12 BREVIORA No. 419

tive in comparison with its closely related North American con-
temporaries. Wood and Patterson (1959: 406) stated, "The
South American rodents were not descended from immigrants
from Wyoming, but rather from rodents that lived in some part
of middle America or southeastern United States, regions from
which the Eocene mammalian faunas are essentially unknown."
The rarity of Protoptychus in fossil collections supports the pos-
sibility that it, too, is based in a stock e\^olving elsewhere than
in the western United States.

Until the lower jaw of Protoptychus is described, however,
retention of the hystricomorphous Protoptychidae in the Pro-
trogomorpha seems advisable for the present, since a hystri-
comorphous skull can accompany a sciurognathus jaw [e.g.,
Pedetes). The similarities to caviomorphs are very suggestive
nevertheless. The future may reveal that Protoptychus was a
precociously specialized offshoot of the northern group from
which ca\'iomorphs arose.

ACKNOW^LEDGMENTS

I am indebted to Albert E. Wood and Brvan Patterson for
their guidance; to the vertebrate paleontology staff at Princeton
University for permitting me to study the specimens; and to
Barbara Lawrence and Charles Mack of the Mammal Depart-
ment, Museum of Comparati\'e Zoology, for making modern
comparati\'e material available to me. I would also like to
thank both Carol C. Jones for unbiased corroboration of my
views of structural details, and Katherine H. Wahlert for aid
with the manuscript.

References

Black, C. C. 1968. The Oligocene rodent Iscliyromys and discussion of the
family Ischyromyidae. Ann. Carnegie Mus., 39: 273-305.

BuGGE, J. 1971. The cephalic arterial system in New and Old World
hystricomorphs, and in bathyergoids, with special reference to the sys-
tematic classification of rodents. Acta Anat., 80: 516-536.

Guthrie, D. A. 1963. The carotid circulation in the Rodentia. Bull. Mus.
Comp. Zool., 128: 455-481.

HoFFSTETTER, R., AND R. Lavocat. 1970. Decouvcrte dans le Deseadien de
Bolivie de genres pentalophodontes appuyant les affmites africaines des
Rongeurs Caviomorphes. Compt. Rend. Acad. Sci. Paris, Ser. D, 271:
172-175.

1973 PROTOPTYCHUS 13

KoRVENKONTio, V. A. 1934. Mikroskopische Uiitcisucluingcn an Nagerin-

cisiven, unter Hinweis auf die Schmelzstiuktur dcr Backenzahne. Ann.

Zool. Soc. Zool.-Bot. Fcnnicae Vanamo, 2: i-xiv, 1-274.
Matthew, W. D. 1910. On the osteology and relationships of Paramys,

and the affinities of the Ischyiomyidae. Bull. Amer. Mus. Natur. Hist.,

28: 43-72.
Miller, G. S., and J. W. Gidley. 1918. Synopsis of the supcigeneric

groups of rodents. Jour. Washington Acad. Sci., 8: 431-448.

ScHLOSSER, M. 1911. Mammalia Saugetiere, p. 325-585. In K. A. von

Zittel, Grundziige der Paliiontologie, II Abt. — Vertebrata; neubearbeitet

von F. Broili, E. Koken, M. Schlosser. Munich and Berlin: R. Olden-

bourg.
Scott, W. B. 1895. Protoptychus hatcheri, a new rodent from the Uinta

Eocene. Proc. Acad. Natur. Sci. Philadelphia, 1895: 269-286.
1905. Paleontology. Part III. Glires. Repts. Princeton

Univ. Exped. Patagonia, 5: 384-487, plates LXIV-LXX.
Simpson, G. G. 1945. The principles of classification and a classification of

mammals. Bull. Amer. Mus. Natur. Hist., 85: 1-350.
Wahlert, J. H. 1968. Variability of rodent incisor enamel as viewed in

thin section, and the microstructure of the enamel in fossil and Recent

rodent groups. Breviora, No. 309: 1-18.
1972. The cranial foramina of protrogomorphous and

sciuromorphous rodents; an anatomical and phylogenetic study. Ph.D.

Thesis. Harvard Univ. 230 pp.
Wilson, R. W^. 1937. Two new Eocene rodents from the Green River

Basin, Wyoming. Amer. Jour. Sci., 34: 447-456.
. 1949. Early Tertiary rodents of North America. Carnegie

Inst. Washington Pub., 584: 67-164.
Wood, A. E. 1935. Evolution and relationships of the heteromyid rodents.

Ann. Carnegie Mus., 24: 73-262.
. 1937. Rodentia, pp. 155-269. In W. B. Scott, G. L. Jepsen,

and A. E. Wood, The mammalian fauna of the White River Oligocene.

Trans. Amer. Phil. Soc. (n.s.) , 28.
. 1949. A new Oligocene rodent genus from Patagonia. Amer.

Mus. Novitates, No. 1435: 1-54.
1955. A revised classification of the rodents. Jour. Mammal.,

36: 165-187.
. 1959. Eocene radiation and phylogeny of the rodents. Evo-

lution, 13: 354-361.
. 1962. The early Tertiary rodents of the family Paramyidae.

Trans. Amer. Phil. Soc. (n.s.) , 52: 1-261.

1968. Early Cenozoic mammalian faunas, Fayum Province,

Eg)pt. Part II. The African Oligocene Rodentia. Bull. Peabody Mus.
Natur. Hist., 28: 23-105.

14 BREVIORA No. 419
, AND B. Patterson. 1959. The rodents of the Deseadan Oli-

gocene of Patagonia and the beginnings of South American rodent evolu-
tion. Bull. Mus. Comp. ZooL, 120: 281-428.
, AND . 1970. Relationships among hystritognath-

oiis and hystricomorphous rodents. Mammalia, 34: 628-639.

Addendum

Since this manuscript was submitted, W. D. Tumbull (per-
sonal communication) has pro\ided me with a description of
the lower jaw in a Field Museum specimen of Protoptychus;
only the outside of the jaw has been prepared so far. Turnbull
states, "The masseteric fossa of the lower jaw is distinct but
shallow, and the angle is laterally offset and rather attenuated.
From the offset angle and the appearance of the junction of the
angle with the ramus, Fd say it had a well developed pars
reflexa to the masseter, but I'\'e not seen the medial side so
know nothing about its area of insertion." He concludes that
the jaw was probably quite hystricognathus. This evidence adds
support to the hypothesis that Protoptychus is related to the
caviomorph rodents through common ancestry either within the
paramyids or within a Middle American caviomorph population
that is as vet unknown.

B R E V I a^fipA

Museum of Comparative

us ISSN 0006-9698

Cambridge, Mass. 29 March 1974 iiNiV^'^'^^fi^^^

ENVIRONMENTAL FACTORS CONTROLLING

THE DISTRIBUTION OF RECENT

BENTHONIC FORAMINIFERA

Gary O. G. Greiner*

Editorial Introduction

Gary Greiner lost an eight-year battle with cancer and died
in January 1973 at the age of 31. His unconventional approach
to paleontology belied the painfully shy and unassuming charac-
ter that many might have taken, so wrongly, as marks of merely
ordinary ability. He was an original and radical thinker, limited,
frustrated, even exasperated, by the reception that must attend
unconventional ideas (be they right or wrong). And it was his
special tragedy that illness, with its ultimate and ineluctable re-
sult, struck even before he began his research and robbed him of
energy and time to test the ideas that flowed so readily.

Gary was captivated by D'Arcy Thompson's approach to
form — ■ to the reduction of organic complexity to a few, simple
generating factors related to physical forces in the environment.
D'Arcy Thompson overstated his case for the complex Metazoa,
but it represents an insight scarcely explored (though surely
more appropriate) for simpler Foraminifera. Gary asserted this
theme within a traditional area of natural history fundamentally
hostile to it ( f or amini feral systematics) ■ — an area that cata-
logues the specific, the unusual and the peculiar in preference to
extracting the simpler regularities that have both general sig-
nificance and frequent exceptions.

This paper represents Gary's views on the control of relative
abundances by a simple environmental factor. Specialists will
recognize some exceptions among forams in other parts of the

♦Request reprints from Stephen Jay Gould, Museum of Comparative
Zoology, Harvard University, Cambridge, Mass. 02138.

2 BREVIORA No. 420

world. They may disagree with his unsupported speculations on
the significance and mode of formation for different types of
calcareous walls. Yet the data on distribution are firm and
must be explained. We hope that readers will focus on the
power of Gary's unconventional approach, on his search for
reduction and cause in preference to elaboration and minute,
thoughtless description.

As an appendix, we attach the short text of a talk delivered
to the annual meeting of the Geological Society of America in
1970. It supplements, in a broader evolutionary context, the
central notion of physical control so central to the functional
theme of causal correlation between environment and form.
We report with the greatest regret that we were unable to re-
construct Gary's major work from his fragmentary notes and
copious data — a bold attempt to synonymize virtually all the
agglutinating Foraminifera of the Gulf of Mexico by showing
that the entire range of form (now attributed to several genera)
can be generated automatically by the interaction of a varying
environment and the few parameters (sensu Raup and Vermeij)
needed to specify construction of the seemingly complex fora-
miniferal test.

Gary wrote the following paper during a post-doctoral year
at the Museum of Comparative ^oology. It was our privilege
to have known, better than most others, such a courageous and
talented person.

Stephen Jay Gould
Alan D. Hecht

Abstract. The relative abundance distributions of the three major groups
of benthonic Foraminifera (agghitinated, porcelaneous, and hyaline calcare-
ous) from the northern Gulf of Mexico paralic environments have been
studied to determine the environmental factor, or factors, actually controlling
the distribution. The relative contribution of each type to the total fora-
miniferal fauna is related to temperature and/or salinity within each bay
studied, and to regional gradients in temperature and salinity (expressions
of climatic and physiographic interactions) throughout the northern Gulf
estuaries.

I conclude that these correlations can be explained on the basis of fora-
miniferal interaction with a single environmental factor — availability of
calcium carbonate for use in construction of tests. This factor depends, to
a large extent, on salinity and temperature in shallow, marine or brackish
waters.

Agglutinated Foraminifera do not require calcite to build their test; they
dominate the faunas in areas of low CaCOg availability. Porcelaneous Fora-

1974 FORAMINIFERAL DISTRIBUTION 3

niinifera employ no nucleating surface for cakite crystal growth; crystals
develop in a random array within a cytoplasmic layer. They dominate in
areas of high CaCOj availability, but diminish in abundance toward lower
values owing to difficulties in secretion of calcite. Hyaline calcareous
Foraminifera produce oriented calcite crystals grown on an organic nucleating
surface. This surface permits secretion of calcite for test construction in
areas of lower CaCOj availability than is possible for the porcelaneous types,
but the need for an ordered structure prevents their thriving in areas of
hyper-supersaturation. Calcareous Foraminifera can dominate agglutinated
types when CaCO;. is readily available, through occupation of niches un-
available to the latter (e.g., on marine plants) . Thus, hyaline calcareous
Foraminifera dominate in areas of intermediate CaCOg availability.

If we accept this simplistic approach to the study of Foraminifera. then
its ramifications might have far-reaching effects in the study of foraminiferal
paleoecology, since the applications would be independent of specific or
generic classification.

Introduction

Most ecologic studies of Recent Foraminifera have dealt with
distributions of the various species or genera present in a par-
ticular area, and with the correlation of these distributions with
various environmental parameters. The reasons for these corre-
lations are difficult to ascertain; hence, the applicability to the
fossil record of conclusions based on such correlations is often
doubtful. To extend ecological inferences of a particular faunal
group to paleontologic situations, an understanding of environ-
mental interactions with morphologic characteristics transcend-
ing specific or generic classifications should be sought.

I chose foraminiferal wall type as the character to investigate
(Greiner, 1969). In standard classifications (Loeblich and Tap-
pan, 1964), wall type is used to separate the three major groups
of Foraminifera into suborders — the Textulariina ( agglutinated
walls), the Miliolina (porcelaneous, calcitic walls), and the
Rotahina (perforate, hyahne calcareous walls). If the influence
exerted by the environment on the distribution of these separate
suborders could be recognized, the information gained could
reasonably be extrapolated to paleoecologic interpretations of
faunas as early as the beginning of the Mesozoic Era when cal-
careous Foraminifera were becoming abundant.

In the Recent, the relative contributions of each of these
groups to the total fauna vary systematically across the con-
tinental shelf, from one bav to another, and from boreal waters
to the tropics. That these changes are systematic and simple,
rather than sporadic and complex, suggests that the abundances

BREVIORA

No. 420

1974 FORAMINIFERAL DISTRIBUTION 5

of the foraniiniferal suborders are being controlled by some
general property of the environment, and that this property also
varies simply and systematically. I assumed that a careful analy-
sis of these distributions in relation to general environmental
parameters would result in correlations leading to an under-
standing of the actual controlling factor or factors. Depth, the
one factor suggested by Phleger (1960a) as most significant
in controlling distributions of foraminiferal species in offshore
traverses, can be essentially eliminated from consideration by
investigation of faunas in very shallow water bodies — bays,
lagoons, and sounds. Variation in the faunas can then be
ascribed to some other environmental factor, such as tempera-
ture, salinity, character of the substrate, or some critical com-
bination of several of these.

Foraminiferal faunas and general environmental parameters
have been described for many of the larger bays, lagoons, and
sounds adjacent to the northern Gulf of Mexico (Fig. 1). Since
we have adequate literature on these shallow water bodies and
since they form a geographic, as well as an environmental, con-
tinuum, they have been chosen for more complete analysis.

The purposes of this study are, then, to describe the relative
abundance distributions of the three major groups of benthonic
Foraminifera in the estuarine environments of the northern Gulf
of Mexico; to relate these distributions to physical and chemical
parameters of the environment; to review the more recent litera-
ture pertinent to the understanding of physiologic mechanisms
employed by the foraminifers in constructing each wall type;
and, finally, to summarize the environmental factors and relate
them to the physiologic processes of wall construction by these
protists, with a view to determining the actual causes of distribu-
tion at this morphologic level.

The results, it is hoped, will have a general significance for
the interpretation of the paleoenvironments and paleoclimates
of geologic epochs prior to those populated by species that still
exist today.

Previous Studies of Foraminiferal Ecology

The early works on Recent foraminiferal ecology {e.g., Parker,
1948; Phleger and Parker, 1951; Parker, Phleger, and Peirson,
1953; and Bandy, 1956) were largely taxonomic, with descrip-
tions of species distribution in relation to depth and geographic
position, based on relative abundances at each sample locality.

6 BREVIORA No. 420

Various environmental parameters were invoked to explain the
apparent natural breaks in faunal patterns. Since depth and
proximity to the shore and continental shelf break had been
measured, and since little else was known about the environment
of the open ocean, discontinuities in the distributions were cor-
related with these factors.

Later studies show similar approaches to the problem of causes
for the observed distribution patterns. A notable example is that
of Lidz (1965), who observed intercorrelations of various en-
vironmental factors and species distributions measured in Nan-
tucket Bay, Massachusetts. The most that could be said, based
on the correlations, is that all of the factors are interrelated and
correlated with one another, i.e., the environmental factors are,
to varying degrees, dependent variables. But nothing can be
said about actual causes of the foraminiferal distributions.

Phleger (1960a), in discussing the ecology and distribution of
Recent Foraminifera, states that the causes of depth zonation
and other distribution patterns are not clearly known. The fac-
tors involved (he states) are temperature, salinity, food, water
chemistry, pressure, currents, turbidity, turbulence, substrate,
biologic competition, disease, etc. And in summarizing this long
list, he states that at the present state of our knowledge it is not
possible to evaluate any one of these factors. In a later report
of the state of the field (Phleger, 1964), he indicates that
". . . there is little or no specific information on the interactions
between the patterns of benthonic foraminiferal faunas and the
natural environments which control these patterns."

A few, more current papers reflect this state of aflfairs and
illustrate attempts to define characteristics of foraminiferal popu-
lations (diversity, planktonic/benthonic ratios, general morphol-
ogy, etc.) which transcend specific or generic characteristics and
which are explicable in terms of the environment (Bandv and
Arnal, 1960; Bandy, 1964; Phleger, 1964; Stehh, 1966; Want-
land, 1967).

Funnell (1967) summarizes our knowledge of foraminiferal
ecology in a discussion of Foraminifera as depth indicators in the
marine en\'ironment. He suggests that since Foraminifera are
studied with relation to depth, and depth has so many factors
correlated to it, we can construct good interpretations for the
Tertiary of, say, the Gulf Coast as compared to the Recent Gulf
of Mexico, but that these same conclusions will not be neces-
sarily valid for the Tertiary of, for example, northwestern
Europe, or for the pre-Tertiary of the Gulf Coast.

1974

FORAMINIFERAL DISTRIBUTION

January

April

July

October

MEAN SURFACE TEMPERATURES (°F)

for the

GULF OF MEXICO

Figure 2. Mean surface water temperatures during four months of the
year for the Gulf of Mexico. (Redrawn from charts supplied by the National
Oceanographic Data Center, 1966.)

Clearly then, the causes of various trends in foraminiferal
faunas must be established, if situations in the fossil record fun-
damentally dissimilar to the time or area of Recent investigations
are to be treated profitably.

Physigo-Chemical Setting of the Gulf of Mexico

The coastal United States bordering the northern Gulf of
Mexico is generally a broad, low-lying plain. The near-shore,
shallow-water environments are made more complex by the
presence of many barrier islands closely paralleling the coastline
and often restricting the free interchange of river and open Gulf
waters. The presence of the barrier islands produces many bays,
lagoons, and sounds ( Fig. 1 ) , which harbor faunas distinct from
those of the open Gulf. The temperatures and salinities of the
water in these estuarine environments are a result of the inter-
action of various climatic and physiographic parameters of the
region.

There is a definite increase in mean annual temperature (re-

BREVIORA

No. 420

Jan.-Mar.

Apr.-Jun.

Jul .-Sept.

Oct.- Dec.

MEAN SURFACE SALINITIES
for the
GULF OF MEXICO

Figure 3. Mean surface water salinity during four seasons of the year
for the Gulf of Mexico. (Redrawn from charts supplied by the National
Oceanographic Data Center, 1966.)

fleeted in the Gulf surface water temperatures, Fig. 2; and in
the January normal isotherms of Fig. 9) from north to south
across the region. Since the bays are generally quite shallow, and
hence the water well-mixed by wind, temperatures in them tend
to correspond to air temperatures. Thus, mean annual water
temperatures in the estuarine environments around the northern
Gulf are lowest in Mobile Bay-Mississippi Sound and Sabine
Lake, and increase in the more southern bays, being highest in
Florida Bay and I^aguna Madre.

Salinity values in the bays similarly show an increase from
north to south. This is the result of several interrelated factors
— precipitation, runoff, evaporation, and salinity of adjacent
Gulf water ( Fig. 3 ) . The first three factors have been studied
by Thornthwaite (1948) and the net effect plotted on a map
as moisture budget isopleths ( reproduced here as part of Fig. 9 ) ,
which are an indication of moisture surplus (positive values) and
moisture deficit (negative values). In general, low salinity values
in the bays are associated with high moisture surpluses, as fresh-
water influx into an enclosed shallow water body prevents, to

1974 FORAMINIFERAL DISTRIBUTION 9

varying degrees, the encroachment of higher salinity Gulf water
{e.g., Mobile Bay-Mississippi Sound and Sabine Lake). (See
discussion by Phleger, 1954: 604). On the other hand, high
salinity \'alues are associated with moisture deficiencies. In this
case the evaporation of lagoonal water permits entrance of higher
salinity Gulf water and subsequent concentration of dissolved
salts {e.g., Laguna Madre). The general increase in salinity of
open Gulf water from north to south (Fig. 3) enhances this
estuarine environmental continuum of increasing salinity, ob-
served from Mobile Bay-Mississippi Sound to Laguna Madre
on the west and to Florida Bay on the east.

Through the interaction of these climatic and physiographic
factors, then, an environmental continuum of increasing temper-
atures and increasing salinities is produced in the shallow water
bodies under consideration here, from Mobile Bay-Mississippi
Sound and Sabine Lake with lowest values, through Matagorda
Bay and San Antonio Bay and environs on the west and Tampa
Bay and Charlotte Harbour on the east with intermediate values,
to Laguna Madre and Florida Bay with highest values.

Discussion of Foraminiferal Distributions

From published tables of species abundances in various estu-
arine environments around the northern Gulf of Mexico, I cal-
culated the relative abundance of individuals possessing each of
the three major wall types at given sample locations. This is
based on percentage of individuals in the total (living plus dead)
foraminiferal fauna. I then plotted these percentages on maps
of the sample distributions and contoured the values.

The relative abundance distribution of the three foraminiferal
groups will be discussed in detail for three of the estuarine en-
vironments — Mobile Bay-Mississippi Sound, Tampa Bay, and
Laguna Madre — and more broadly for the others, to demon-
strate correlations with temperature and salinity on the local
scale. Following this, I will consider the faunal dominance by
each of the groups through all the bays, lagoons, and sounds
adjacent to the northern Gulf to document similar correlations
with these environmental factors on a regional scale.

mobile BAY-MISSISSIPPI SOUND

The distribution of Foraminifera and possible ecologic factors
affecting the distribution in Mobile Bay, Alabama, have been
briefly mentioned by Walton ( 1 964 ) . Phleger ( 1 954 ) has made

BREVIORA

No. 420

re
o

o
o

r^ Q

3 ^

fac
r3

«3

••H

(I)

-4—)

<-U

• ^^

<-H

■4-J

C/3

-i

c/3

C/}

C/3

1/3

c/:

• -H

•4—1

I— ^

«J

•^H

til

1974 FORAMINIFERAL DISTRIBUTION 11

a similar but more detailed study of the Mississippi Sound.
Upshaw et al. (1966) have studied and described the environ-
ment, sediments, and microfauna from both areas plus a portion
of the adjacent continental shelf (Fig. 4).

There is a considerable moisture excess for this region
(Thornthwaite, 1948; and Fig. 9) . This results from many large
ri\'ei's discharging fresh water into Mobile Bay (Mobile and
Tensaw rivers) and Mississippi Sound (particularly the Pas-
cagoula River). The offshore, discontinuous island chain is an
effective barrier to ready mixing of this runoff with the open
Gulf water ( Phleger, 1 954 ) . However, some denser, more saline
water from the Gulf does enter Mississippi Sound by way of the
surge channels and mixes with the fresh water from the rivers
within this shallow water body. Thus, there is a steep salinity
gradient in bottom waters from the open Gulf (with usually
35°/oo), through the adjacent inlet (near 30° /oo), and into
Mobile Bay-Mississippi Sound (to < 5°/oo within 10 miles of
the Gulf).

From the foraminiferal distribution data of Upshaw et al.
( 1966, plate 4, reproduced here as Fig. 4), it is evident that the
agglutinated Foraminifera are relatively most abundant in water
with the lowest salinity values, and that they decrease in relative
abundance with increasing salinity. On the other hand, the
hyaline calcareous Foraminifera are associated with the more
saline Gulf water, diminishing in relative abundance as it is
diluted by fresh water within the bay and sound. Representatives
of the third group, the porcelaneous Foraminfera, are not found
within this restricted area, though they are present (up to 30%
or more) in the more saline Gulf water somewhat seaward of
the freshwater influence. Hence, the relative abundance dis-
tributions of two of the foraminiferal groups are correlated here
with water salinity values - — ■ hyaline calcareous directly, ag-
glutinated inversely.

TAMPA BAY

Bandy (1956) and Walton (1964) have made ecologic studies
of the Foraminifera of Tampa Bay and environs, including Old
Tampa Bay (Walton, 1964) and Hillsboro Bay (Bandy, 1956).

Bathymetrically, the bay can be divided into low sand and
grass flats of shallow depth (< 15 ft. of water) with superim-
posed relatively deep channels (Goodell and GorsHne, 1960).
Maximum depth in the bay is slightly more than 30 feet, which

BREVIORA

No. 420

• 2*30'

28*00

-27'50

after Bandy, 1956; and Walton, 1964

Figure 5. Bathymetry and sample locations for Tampa Bay, Florida.
(From Bandy, 1956; and Walton, 1964.)

1974 FORAMINIFERAL DISTRIBUTION 13

is that of most of the channels ( Fig. 5 ) . The sediments of
Tampa Bay are predominantly fine to very fine quartz sands
(Walton, 1964).

The salinity distribution pattern for Tampa Bay and environs
can be qualitati\'ely described as follows: In the channels dis-
secting the bottom topography, the water salinity is at a maxi-
mum near the mouth of the bay complex (somewhat above
'normal' marine), with a very slight gradient to lower salinities
in Hillsboro Bay. The adjacent shoal waters have a similar
gradient, from near normal marine salinity at the mouth of
Tampa Bay to lowest salinities (just slightly above that of river
water) in upper Hillsboro Bay. Since the salinity in the channels
is everywhere higher than that of the adjacent sand and grass
flats, there is also a positive gradient from shallow to deep water.

The relati\'e abundance distributions of the agglutinated and
the porcelaneous Foraminifera are shown in Figures 6 and 7,
respectively. The changing contributions of these two groups
and that of the hyaline calcareous group reflect the salinity
gradients just discussed.

These foraminiferal distributions clearly demonstrate a strong
correlation between salinity and the relative abundances of each
of the three groups. Highest salinity waters characteristically
have high percentages of the porcelaneous type associated with
them. In successively lower salinities, the hyaline calcareous type
and then the agglutinated type reach their maximum relative
abundances.

LAGUNA MADRE

Laguna Madre is located within the semi-arid climatic zone
of Thornthwaite ( 1 948 ) , and, hence, has a more or less persis-
tent, marked moisture deficiency (Fig. 9). There are no major
rivers flowing into the area, and there is only very slight fresh-
water inflow from ephemeral streams during local rainfall (Rus-
nak, 1960). The excess of evaporation over precipitation allows
the normal marine Gulf waters to enter the shallow basins
(average depth, about 25/2 ft-) of Laguna Madre and causes
the water there to be generally hypersaline. Chlorinities in the
northern basin range from 22 to 45°/ 00 CI and in Baffin Bay
from 1 to 45°/oo CI; the southern basin, with lower salinities,
has up to 35°/oo CI (Phleger, 1960b).

The temperature of the lagoonal water reflects that of the air
(Phleger, 1960b) ; and because of the positive thermal gradient

BREVIORA

No. 420

-28*00

»3'45

27°50

57°40

27*30

Figure 6. Relative abundance distribution of agglutinated Foraminifera
from Tampa Bay, Florida. (Data from Bandy, 1956; and Walton, 1964.)

1974

FORAMINIFERAL DISTRIBUTION

82*30'

-27*50'

• 2*45

28*00'

-27*30

Figure 7. Relative abundance distribution of porcelaneous Foraminifera
from Tampa Bay, Florida. (Data from Bandy, 1956; and Walton, 1964.)

16 BREVIORA No. 420

in this area from north to south (Espenshade, 1960), the relative
abundance distribution of the foraminiferal groups can be cor-
related with this parameter.

The foraminiferal populations are dominated by the porce-
laneous types in nearly all samples studied (Phleger, 1960b)
(Fig. 8).'

In Laguna Madre, the foraminiferal distributions are related
to both salinity and temperature. Low salinity areas (Baffin
Bay) are dominated by hyaline calcareous species; high salinity
areas by porcelaneous species. But within the hypersaline en-
\'ironments, the relati\'e proportions of the two types are corre-
lated with temperature — porcelaneous (most abundant in the
southern basin) directly, hyaline calcareous inversely.

General Discussion of the Distributions

I have shown that the relative abundance distributions of the
three groups of benthonic Foraminifera are closely related to
salinity distributions, and occasionally to temperature gradients,
within several shallow-water environments adjacent to the Gulf
of Mexico. The relationship on a local scale shows a gradient
of maximum relative abundances for the three groups, from
agglutinated forms in low salinity waters, to hyaline calcareous
forms in waters of intermediate salinities, to porcelaneous forms
in waters of highest salinity. Each of the various t) pes does not
necessarily dominate the fauna at its maximum, but only reaches
its peak relative abundance there for the bay or estuary under
consideration.

Some modifications to this sequence occur. Most can be ex-
plained as the simple displacement of either or both of the end-
member groups — the agglutinated and the porcelaneous types
— from the sequence. Thus, for example, in the Mobile Bay-
Mississippi Sound environment, the porcelaneous forms are not
present, and the sequence ends with the hyaline calcareous maxi-
mum. However, at the opposite end of the spectrum, the agglu-
tinated types not only reach their maximum, but completeh
dominate the upper bay fauna to the exclusion of any calcareous
forms. This situation is correlated with a much hie^her runoff
and consequent lower salinity for this estuary than for most of
the others.

On the other hand, the samples from Laguna Madre yielded
almost no agglutinated Foraminifera while the hyaline calcareous
forms reach their maximum abundance in waters of the lowest

1974

FORAMINIFERAL DISTRIBUTION

«l

• »H

• w^

)->

C/5

,,,«^

(£>

:-!

1— <

S-i

«4-l

3
X5

• ^H

4-1

M-l

•4-J

^ — ■

,'"

(1)

i-i

4-1

18 BREVIORA No. 420

salinity and temperature. Thus, the sequence is still preserved,
but with one end-member excluded. This area is characterized
by higher than "normal" marine salinities and by high tempera-
tures.

Intermediate faunal and environmental situations are present
and, as might be expected, are located geographically between
these end-member dominances. Matagorda Bay and environs
is a good example of these conditions. The foraminiferal fauna
is everywhere dominated by the hyaline calcareous types, to the
near exclusion of agglutinated and porcelaneous types. This can
be correlated with intermediate regional temperatures and with
the close balance of run-off plus precipitation against evapora-
tion, the latter of which produces a salinity near the normal
value for the open Gulf in that region (which is slightly below
"normal" marine; cf. Fig. 3).

This sequence in maxima of the relative abundances of the
three groups of Foraminifera is present not only on a local scale
within a bay or lagoon, but also on a regional scale, across the
entire northern Gulf of Mexico. Just as on a local scale, the
trend is correlated with salinity and with temperature.

There are two regional trends in the environment (or climate)
which we must recognize. These are : 1 ) the gradual increase
in temperature from north to south; and 2) the gradual shift in
moisture budget from a marked surplus in the Mississippi Delta
region, westward through a moisture balance near Matagorda
Bay, to a marked moisture deficit in the region about Laguna
Madre, and eastward to a near balance, but definite surplus,
alonsf most of the Florida coast.

The relati\'e abundance distributions of the three groups of
benthonic Foraminifera together with isotherms of the January
normal temperature and with the moisture budget zones (after
Thornthwaite, 1948) of the coastal region of the northern Gulf
of Mexico have been summarized in Figure 9. The sequence in
maxima of relative abundances of the Foraminifera, correlative
with a sequence of environmental factors, can be seen as a re-
gional continuum from Mobile Bay, along the Louisiana and
Texas coasts, to Laguna Madre. The Mobile Bay foraminiferal
fauna is dominated, for the most part, b) agglutinated species.
This is correlated with excessive moisture and consequent low
salinity water within the bay. In addition to this, mean annual
temperatures are here near the lowest for the Gulf area.

Sabine Lake, the next area of study to the west, again has a
fauna dominated by agglutinated species (Kane, 1967) and is

1974

lORAMINII FRAL DISTRIBUTION

>50X PORCELANEOUS FORAMINIFEHA
JANUARY NORMAL ISOTHERMS
MOISTURE BUDGET ISOPLETHS

FLORIDA BAY

Figure 9. Dominance of agglutinated, hyaline calcareous, and porcelaneous
Foraminifera in the northern Gulf of Mexico paralic environments; including
January normal isotherms after Espenshade (1960) , and moisture budget
isopleths after Thorn thwaite (1948) .

included, essentially, within the same environmental zones as
Mobile Bay. The moisture surplus is actually less than in the
previous area, but this is compensated by the greater restriction
from mixing with the open Gulf.

To the southwest, Matagorda Bay and environs is within a
warmer climatic zone and is also within a zone of only very
slight moisture surplus. Emphasizing this lower moisture surplus
is the lack of large rivers discharging fresh water into the bay.
The result is relatively warm water with salinities near, but some-
what less than, those of the adjacent Gulf. Commensurate with
this rise in water temperature and salinity over that of Sabine
Lake and Mobile Bay is a shift in the foraminiferal fauna. Here
the hyaline calcareous forms dominate (Lehmann, 1957; Shen-
ton, 1957).

San Antonio Bay and environs displays an anomalous, but
explicable, reverse in the environmental and faunal sequence
(Parker, Phleger, and Peirson, 1953; Phleger, 1956; Shepard
and Moore, 1955 and 1960), despite its position within the
climatic trend to higher temperatures and greater moisture de-
ficiency. The influx of fresh water to the bays from the relatively
large San Antonio-Guadalupe River system is the cause of the
much lower water salinity values here than in Matagorda Bay,
which is in a general area of greater moisture surplus. There is
a correlative shift in the foraminiferal fauna to one dominated

20 BREVIORA No. 420

in the upper bay by agglutinated forms. The central and lower
bay fauna is dominated by hyaline calcareous forms as in Mata-
gorda Bay, but there is still a higher proportion of agglutinated
types in the former area, commensurate with the lower average
salinity there.

Laguna Madre, in a climatic zone of high annual tempera-
tures and marked moisture deficiency, completes the faunal se-
quence obsen^ed on a smaller scale in some of the bays around
the Gulf. Agglutinated Foraminifera are virtually absent from
all samples taken in the lagoon (Phleger, 1960b). In the north-
ern basin of the lagoon, the proportion of hyaline calcareous
specimens is slightly less than that of porcelaneous. And the
porcelaneous types overwhelmingly dominate the southern basin.
Hence, there is a direct correlation between temperature and the
proportion of porcelaneous forms in the bottom sediment.

An environmental continuum and faunal dominance sequence
similar to that just described can be documented for the Florida
Gulf coast and correlated with climatic trends from Mobile Bay
to Florida Bay. The change in moisture budget is not so dra-
matic as to the west, as a surplus is maintained along the entire
coast to the tip of Florida (Fig. 9). However, the temperature
gradient is even steeper, making Florida Bay approximately 6°C
warmer than Laguna Madre during January, though both are at
comparable latitudes.

The sequence in maxima of relative abundances of the three
benthonic groups is developed and can be correlated with the
general environmental trend to higher salinities and higher tem-
peratures to the south. After Mobile Bay, with its overwhelming
dominance of agglutinated Foraminifera, the next area to the
south is Tampa Bay.. The whole foraminiferal sequence is de-
veloped here, but the hyaline calcareous types dominate the
fauna over the greater part of the bay, except in the deep chan-
nels. Charlotte Harbour and vicinity has a similar fauna, largely
dominated by hyaline calcareous forms (data after Bandy,
1954), though the whole sequence is again present. Both of
these areas are similar environmentally and climatically. Both
are in the wet subhumid zone and both receive limited drainage
from the surrounding, low-lying, karst topography. There is
some difference in latitude and hence, in mean annual tempera-
ture, but this is minimal. Thus, the two areas have very similar
foraminiferal faunas.

The fauna of Florida Bay is dominated in the near-shore,
lower salinity areas by hyahne calcareous types, and by por-

1974 FORAMINIFERAL DISTRIBUTION 21

celaneous types seaward, toward the keys (Lynts, 1962). This
fauna! composition is similar to that of Laguna Madre, but with
a slightly greater proportion of hyaline calcareous types. Thus,
despite its location within a wet subhumid climatic zone
( Fig. 9 ) , comparable in this respect to Matagorda Bay, it has a
fauna similar to that of a lagoon within a semi-arid zone. Mata-
gorda Bay and Florida Bay both have only very small rivers
emptying into them. The differences between Florida Bay and
Mata2:orda Bav, and the similarities that the former has with
Laguna Madre can perhaps be explained on the basis of salinity
of adjacent Gulf water, and on the basis of temperature.

The salinity of the open Gulf water replacing that evaporated
from Florida Bay is somewhat higher than that entering Laguna
Madre, and considerably higher than that available to Mata-
gorda Bay ( Fig. 2 ) . Mean annual temperature at Florida Bay
is somewhat higher than at Laguna Madre and considerably
higher than at Matagorda Bay (Fig. 9). Thus, though the
water of Florida Bay is diluted by runoff and precipitation simi-
lar to that for Matagorda Bay, it can be more quickly reconsti-
tuted to a higher salinity owing to greater evaporation and easier
mixing with waters more saline than "normal" marine. It is also
possible that the high proportion of porcelaneous Foraminifera
should be correlated with the higher temperatures there, as I
postulated for Laguna. Madre.

To summarize the distributions and correlations discussed in
this section, the following conclusions can be drawn. On a local
scale, i.e., within a bay, lagoon, or other shallow-water environ-
ment, there is a succession of relative abundance maxima from
agglutinated, through hyaline calcareous, to porcelaneous types;
this is correlated with a trend in salinity or temperature values
from low to high for the area. Also, either or both of the end-
member types can be displaced from the sequence with commen-
surate shifts in the salinity and temperature gradients. These
gradients are the most obvious factors of the environment to
which the faunal sequence can be related. There are essentially
uniformly shallow depths over most of the areas, and no ap-
parent correlation of the faunal groups with bathymetry. Where
several different sediment types are present within a single bay
area, they are generally correlated with depth and, hence, not
correlated with the fauna. In some areas, such as Florida Bay
and Laguna Madre, a relative abundance sequence in the fora-
miniferal types is correlated with the temperature or salinity

22 BREVIORA No. 420

gradient in each bay despite the uniformity of bottom sediment

type.

Regionally, the same foraminiferal sequence is present — man-
ifested in the various types dominating the population from bay
to bay in succession. This sequence is again correlated with a
general trend in salinity and temperature. This trend in the
shallow-water environmental continuum is explicable in terms of
climate and physiography of the adjacent coastal plain. The
main climatic factors necessary for explanation are moisture
balance and temperature. The influence of physiography on
the local en\'ironment is evident in the amount of runoff carried
into the various areas of investigation.

Environmental Factors Controlling
Distribution of Foraminifera

I present the hypothesis that the actual environmental factor
controlling the distribution of Foraminifera is the availability of
calcium carbonate (dependent, to a great extent, on salinity,
temperature, and depth of water) ; or the ease with which these
one-celled organisms can extract and precipitate CaCOs for
their test from the surrounding water.

Chemistry. Revelle (1934), in discussing the physico-chemical
factors affecting the solubility of calcium carbonate in seawater,
stated that, from the mass law equation Ca++ X CO3 ==
^CaCOa, three parameters control the solubility of CaCOs:
concentrations of calcium and carbonate ions and the value of
the temperature-dependent constant ^CaCOs. "These factors
are in turn dependent on salinity, temperature, hydrostatic pres-
sure due to depth below the surface, carbon dioxide content,
and the concentration of hydrogen and hydroxyl ions, as indi-
cated by the /?H" (Revelle, 1934: 103-104). Revelle and
Fairbridge (1957: 256) conclude that the two most important
processes facilitating the precipitation of calcium carbonate
probably are : ( 1 ) an increase in temperature, which lowers the
solubility of CO2, thus increasing the carbonate ion concentra-
tion; and 2) evaporation, which increases the calcium ion con-
centration and carbonate alkalinity.

These two processes, governing the carbonate ion and calcium
ion concentrations, respecti\ely, can be equated with increasing
temperature and increasing salinity. Thus, in low salinity and
low temperature environments calcium carbonate will not be
easily precipitated, owing to low calcium and low carbonate ion

1974 FORAMINIFERAL DISTRIBUTION 23

concentration, the latter being largely a result of increased solu-
bility of CO2 in the water. On the other hand, waters with high
salinites and high temperatures, with their relatively high cal-
cium and carbonate ion concentrations, are saturated or super-
saturated with respect to calcium carbonate, as in tropical and
subtropical surface seawater (Chave and Schmalz, 1966). In
these areas calcium carbonate will be precipitated most readily.
Thus, all of the environmental parameters tend to increase
the availability of calcium carbonate from the Mississippi Delta
region toward the Rio Grande on the west, and toward Florida
Bay on the east. This trend is closely correlated with the ob-
sen^d trend in relative abundance distributions of the foramini-
feral groups studied (see Fig. 9 for a summary of climatic factors
and the f oraminif eral distributions ) .

From these observations, it is apparent that agglutinated Fora-
minifera are relatively most abundant in areas with the lowest
availability of calcium carbonate. Porcelaneous Foraminifera,
on the other hand, are associated with high availability of cal-
cium carbonate, and often dominate the foraminiferal faunas of
warm, saline tropical or subtropical waters. Finally, the areas
characterized by intermediate calcium carbonate availability are
dominated by the hyaline calcareous Foraminifera. This gen-
eralization is true on nearly all scales of observation: within a
bay or lagoon, among several adjacent bays of a region, on con-
tiguous portions of the continental shelf (Greiner, 1970), and
on a worldwide scale.

Mechanism. The agglutinated Foraminifera do not require
the precipitation of calcium carbonate in construction of their
tests. They utilize the available sediment grains, cementing them
together with a predominantly organic material (Hedley, 1963;
Towe, 1967). They are therefore free of restriction to any of
the marine or estuarine environments. The calcareous Fora-
minifera (both hyaline calcareous and porcelaneous), on the
other hand, require calcium carbonate for the construction of
their tests. The extent to which its availability is required de-
pends upon the ability of the organism to concentrate and secrete
( or allow precipitation of ) calcium carbonate against ( or within )
the chemical environment of the water. I suggest that a funda-
mental distinction between the hyaline calcareous and the por-
celaneous Foraminifera lies herein.

Electron microscope studies (Hay, Towe, and Wright, 1963;
Towe and Cifelli, 1967; Lynts and Pfister, 1967) have shown
that there is a radical difference between the shell structure of

24 BREVIORA No. 420

porcelaneous Foraminifera and that of the hyahne calcareous
types. In the porcelaneous wall there is a thick, inner layer with
a three-dimensionally "random" array of elongate crystals and
a pavement-like, surface veneer that in part exhibits preferred
orientation. The hyaline calcareous wall, on the other hand, is
made up of calcite crystals with a preferred orientation, the
whole wall being penetrated by numerous pores, which are visi-
ble under the light microscope as well (Towe and Cifelli, 1967).
These observ^ations are consistent with the general separation
(Loeblich and Tappan, 1964) of the hyaline calcareous and
the porcelaneous wall t}pes on the basis of perforations of one
type and porcelaneous appearance of the other.

Lynts and Pfister (1967) have pointed out the differences in
crystallization of the wall as obser\'ed for these two test types.
One species with a hyaline calcareous wall was obser\'ed in the
process of chamber formation (Angell, 1967a and b). The fora-
minifer, when beginning to add a new chamber, extended a
portion of its protoplasm through the aperture of the test, form-
ing a bulbous drop with the exact shape of the prospecti\T
chamber. An organic sheath formed on the surface of the drop.
Shortly thereafter, protoplasm was again exuded (through the
new aperture in the organic sheath) and covered, in a thin film,
the surface of the new, tectinous chamber wall. Calcite crystals
were then observed to nucleate on the organic surface and to
grow upward (perpendicular to the surface) within the exuded
cytoplasm, until the calcareous wall was complete. Observations
by Towe and Cifelli ( 1967) suggest that other hyaline calcareous
species also nucleate calcite crystals for test formations on an
organic base."^'

Arnold ( 1 964 ) , while obserxdng chamber formation of a por-
celaneous species (similar to that of hyaline calcareous types up
to the secretion of calcite ) , noted that the calcite crystals grew
in "random" fashion within the oroanic matrix formed bv the
exuded cytoplasm of the protist, not upon an organic nucleating
surface (see Fig. 10 for a diagrammatic comparison of crystal
growth in the two types). Lynts and Pfister (1967) have pointed
out this difference between these two test types, and Towe and
Cifelli (1967), likewise, conclude that porcelaneous wall struc-
ture is significantly different from hyaline calcareous.

I suggest that the absence of a nucleating surface for the se-

*Subseqiient work by Towe (1972) suggests that this may not be true
for all Foraminifera in this group.

1974

FORAMINIFERAL DISTRIBUTION

SCHEMATIC DEVELOPMENT OF HYALINE
CALCAREOUS WALL TYPE

iXUDiO
PROTOPLASM

CAICITI CRYSTALS

ORGANIC
NUCLEATING
SURFACE

EMBRYONIC CRYSTAL
STAGI

FINISHED C AlCITE
WALL

SCHEMATIC DEVELOPMENT OF PORCELANEOUS

WALL TYPE

EMRRrONIC CRYSTALS

EXUDED
PROTOPLASM

INNER
O

RGANIC ^

EMBRYONIC CRYSTAL
STAGE

FINISHED CAICITE
WALL

Figure 10. Diagrammatic sketch illustrating differences in test wall
calcification in porcelaneous and hyaline calcareous Foraminifera. See text
for discussion.

26 BREVIORA No. 420

cretion of calcite by the porcelaneous Foraminifera dictates that
they Hve within an environment of readily available calcium
carbonate ^ — ^ at the point of "saturation'' or even "supersatura-
tion."*^^ The nucleating surface employed by the hyaline cal-
careous Foraminifera, however, allows them a greater range of
habitable en\'ironments. Because of this, they can do well both
in normal marine and in slightly hypersaline conditions, and are
pre\'ented from thriving only within areas of low calcium car-
bonate availability (usually low salinity) and areas of "hyper-
supersaturation" (see below for further discussion).

In the very low salinity environments, where the availability
of calcium carbonate is below the threshold required by hyaline
calcareous forms, the agglutinated Foraminifera will predomi-
nate. This is so simply because the agglutinated species are not
restricted by such a boundary, while the calcareous types are.
As waters with more readily available calcium carbonate are
approached, more hyaline calcareous forms will be present,
thus diminishing the relative abundance of agglutinated types.
Though the agglutinated types are not excluded from en\iron-
ments of high calcium carbonate availability, they are subordi-
nate in abundance to the calcareous forms there. This can be
explained by the ability of calcareous forms to di\ersify and
occupy ecological niches not as readily a\'ailable to the aggluti-
nated types {e.g., marine plants), as the construction of an ag-
glutinated test ties the protist to its source of raw material —
the bottom sediments. (Again, this is a relative situation. I am
aware that some agglutinated types may live on marine plants
utilizing the fine sediment dust that clings to their surfaces for
test construction.)

The porcelaneous types reach their maximum relati\e abim-
dance under en\'ironmental conditions of maximum a\'ailability
of calcium carbonate — the tropics and subtropics with high
temperatures and hypersalinities. Their proportion of the total
fauna decreases in the direction of lower calcium carbonate
availability, toward lower temperatures as in Laguna Madre, or
toward lower salinities as in Florida Bay. This is so because they
have greater difficulty in secreting calcite in these environments,

**I use the terms "saturation," "supersaturation," and "hyper-super-
saturation" in a relative sense. Though these terms do have definite
meanings in chemistry, it is difficult to say at what point a sea-water
solution is "saturated" with respect to CaCO.j in a natural environment,
and even more difficult to state tlie relation of the foraminiferids to some
precise value of saturation. They can be related relatively, however.

1974 FORAMINIFERAL DISTRIBUTION 27

while the hyaline calcareous types are seemingly not hindered in
this process until very low salinities or temperatures are reached.
The porcelaneous types can completely dominate the fauna in
en\ironments of \^ery high calcium carbonate availability owing,
perhaps, to the unordered crystalline nature of their test walls.
Hyaline calcareous types would perhaps be unable to secrete
well-ordered crystals in an environment of calcium carbonate
"hyper-supersaturation."

Consequently, Foraminifera with hyahne calcareous walls
reach their maximum relative abundance in areas of intermedi-
ate calcium carbonate availability, where the porcelaneous types
are greatly diminished owing to problems of calcite secretion.

Summary. An hypothesis has been proposed to explain the
obser\'ed foraminiferal sequence correlated with salinity and
temperature gradients within estuarine environments. The en-
vironmental factor thought to control the distributions of major
groups is the availability of calcium carbonate utilized in test
construction by two of the types. This factor is dependent mainly
on temperature, salinity, and CO2 content of the water.

This hypothesis adequately explains the observed distributions
of these groups; it explains, through physiologic interaction with
the en\'ironment, the correlation between foraminiferal groups
and temperature and salinity gradients; and it ultimately ex-
plains the correlation of these groups with climatic factors. The
fact that this correlation exists between the foraminiferal se-
quence and the environmental factors reducible to calcium car-
bonate availability, and the fact that this relationship can be
explained by varying abiHties of the foraminifers to construct
tests suggest, that these organisms secrete calcite in near-equi-
librium with their environment. This implies, further, that these
protists are unable to concentrate and precipitate calcium car-
bonate from the seawater in \'ery great chemical opposition to
their surroundings and that they are, in this sense at least,
simple organisms, dependent on, and controlled to a great extent
by, their environment.

Geologic Significance of Results

The understanding of environmental effects on the distribution
of organisms and on the modification of phenotypes is essential
to the interpretation of paleoenvironments. The purpose of this
study has been to gain some understanding of factors governing
the distribution of Foraminifera in Recent environments. The

28 BREVIORA No. 420

difficulty in learning the causes of distribution of any particular
species is apparent, and geologic applicability of such knowledge
is severely limited by the geologic range of the species. In this
light, I have sought to determine the environmental control on a
characteristic of the fauna that transcends the specific level of
classification and w^hich is amenable to paleoecologic extrapola-
tion. I have shown that Foraminifera are distributed within the
Recent environment in a fashion covariant with certain factors
summarized as the availability of CaCOs. The proposition that
the availability of CaCOs is indeed the cause of their relative
abundance distribution is supported by a credible explanation,
on the physiologic level, of foraminiferal test construction.

The understanding of distributions at this level depends only
on a knowledge of the wall types, not on individual character-
istics of a taxonomic group. Much can be learned concerning
salinity and temperature distributions in ancient seas and estu-
aries through use of Foraminifera at this morphologic le\'el.
With a more thorough understanding of the causes of plank-
tonic distributions and changes in foraminiferal diversity on the
continental shelf, more can be learned of paleobathymetry and
location of shore-lines.

Since work with the Foraminifera at this level circum\'ents the
problems associated with extending interpretations of \'arious
Recent taxa back in time, application of the principles gained
can be extrapolated through the Mesozoic to the beginnings of
the calcareous Foraminifera. One major assumption must be
made for the interpretation of fossil faunas. This is that the
ability of Foraminifera to secrete calcite for particular wall types
within a given environment of CaCOs availability has not
changed appreciably since the corresponding development of
each test type. This assumption, it would seem, is a fair one; if
the crystalline structure within the wall of Jurassic porcellaneous
Foraminifera is similar to that found in Recent specimens of
that wall type, it is reasonable to assume that the physiologic
processes that produced it were similar.

Perhaps a more important inference can be drawn from the
results of this study. If the Foraminifera depend to such an
extent on the availability of CaCOs in specific degrees of satura-
tion or supersaturation within the environment for secretion of
their tests, then they cannot readily concentrate these ions physio-
logically and hence cannot easily act in chemical opposition to
their surroundings. This implies further that other aspects of
foraminiferal tests are subject to simple control by the en\iron-

1974 FORAMINIFERAL DISTRIBUTION 29

ment. I suggest that such factors as general test morphology,
apertural position and number, and chamber number may be
go\erned not strictly genetically (as is implied by the erection of
specific or generic groups based on these characters), but by
the macro- or microenvironment of the living individual. This,
then, is an open avenue for research. If environmental causes
for \arious morphological characteristics can be derived, im-
measurable, paleoecologic value can be attributed to Foramini-
fera.

References

Angell, R. D. 1967a. The test structure and composition of the fora-

minifer Rosalina Floridana. J. Protozool., 14: 299-307.
. 1967b. The process of chamber formation in the fora-

minifer Rosalina Floridana (Cushman) . J. Protozool., 14: 566-574.
Arnold, Z. M. 1964. Biological observations on the foraminifer Spirolo-

culina hyalina Schulze. Univ. Calif. Pub. Zool., 72: 1-78.
Bandy, O. L. 1954. Distribution of some shallow-water foraminifera in the

Gulf of Mexico. U.S. Geol. Surv. Prof. Pap. 254-F.
. 1956. Ecology of Foraminifera in the northeast Gulf of

Mexico. U. S. G. 5. Prof. Paper 274-G.
, 1964. General correlation of foraminiferal structure with

environment. In Imbrie, J., and N. D. Newell (eds.) , Approaches to
Paleoecology. John Wiley, pp. 75-90.
AND Arnal, R. E. 1960. Concepts of foraminiferal paleoecology.

Amer. Assoc. Petrol. Geol., Bull. 44: 1921-1932.
Chave. K. E. and Schmalz, R. F. 1966. Carbonate-seawater interactions.

Geochim. et Cosmochim. Acta, 30: 1037-1048.
ESPENSHADE, E. B., Jr., cd. 1960. Goode's World Atlas: Rand-McNally.
FuNNELL, B. M. 1967. Foraminifera and Radiolaria as depth indicators

in the marine environment. Marine Geol., Special Issue, 5/6: 33-47.
GooDELL, H. G. AND GoRSLiNE, D. S. 1960. A sedimentologic study of

Tampa Bay, Florida. Internat'l Geol. Cong., 21st Session, Norden, pt.

XXIII.
Gre'iner, G. O. G. 1969. Environmental factors causing distributions of

Recent Foraminifera. Ph.D. Thesis, Case Western Reserve University,

195 pp.
. 1970. The distribution of Recent benthonic foramini-
feral groups on the Gulf of Mexico continental shelf. Micropaleontology,

16(1): 83-101.
Hay, W. H., Towe, K. M., and Wright, R. C. 1963. Ultra-microstructure

of some selected foraminiferal tests. Micropaleontology, 9: 171-195.
Hedley, R. H. 1963. Cement and iron in the arenaceous Foraminifera.

Micropaleontology, 9: 433-441.

30 BREVIORA No. 420

Kane, H. E. 1966. Sediments of Sabine Lake, the Gulf of Mexico, and
adjacent water bodies, Texas-Louisiana. Jour. Sed. Petr., 36: 608-619.

. 1967. Recent microfaunal biofacies in Sabine Lake and en-
virons, Texas-Louisiana. Jour. Paleont. 41: 947-964.

Lehmann, E. p. 1957. Statistical study of Texas Gulf coast Recent fora-
miniferal facies. Micropaleontology, 3: 325-356.

LiDZ, L. 1965. Sedimentary environments and foraminiferal parameters:
Nantucket Bay, Massachusetts. Limn, and Oceanog., 10: 392-402.

LoEBLicH, A. R. AND Tappan, H. 1964. Sarcodina, chieflv "Thecamoebians"
and Foraminifera. In Treatise on Invertebrate Paleontology, Part C,
Protists 2, V. 1 and 2, R. C. Moore, ed. Lawrence, Kansas: LTniv.
Kansas and Geol. Soc. America, pp. C55-C164.

Lynts, G. W. 1962. Distribution of Recent Formaminifera in Upper
Florida Bay and associated sounds. Contrib. Gush. Found. Foram.
Res. XIII, pp. 127-144.

AND Pfister. R. M. 1967. Surface ultrastructure of some

tests of Recent Foraminifera from the Dry Tortugas, Florida. J. Pro-
tozool., 14: 387-399.

Parker, F. L. 1948. Foraminifera of the continental shelf from the Gulf
of Maine to Maryland. Bull. Mus. Comp. Zool., 100: 213-41.

, Phleger, F. B., and Peirson, J. F. 1953. Ecology of Fora-
minifera from San Antonio Bay and environs, southwest Texas. Gush.
Found. Foram. Res., Special Pub. 2.

Phleger, F. B. 1954. Ecology of Foraminifera and associated organisms
from Mississippi Sound and environs: Amer. Assoc. Petrol. Geol., Bull.
38: 584-647.

. 1956. Significance of living foraminiferal populations along

the central Texas Goast. Gontrib. Gush. Found. Foram. Res. \\\. pp.
106-151.

. 1960a. Ecology and Distribution of Recent Foraminifera.

Baltimore: Johns Hopkins Press.

1960b. Foraminiferal populations in Laguna Madrc. Texas.

Sci. Repts. Tohoku Univ., 2nd Ser. (Geol.), Special \'ol. 4: 83-91.
. 1964. Foraminiferal ecology and marine geology: Marine

Geol., 1:16-43.
and Parker. F. L. 1951. Ecology of Foraminifera. Xoitli-

west Gulf of Mexico. Geol. Soc. .Aimerica, Memoir 46.
Revelle. R. 1934. Physico-chemical factors affecting the soluhiliiv of

calcium carbonate in sea water. Jour. Sed. Petr.. 4: 1031 10.
AND FAiRBRmGE, R. 1957. Carbonates and carbon dioxide. In

Treatise on Marine Ecology and Palcoecology, Vol. 1. Ecology. J. ^V.

Hedgpeth, ed. New York Gity: Geol. Soc. America, Memoir 67.
RusNAK. G. A. 1960. Sediments of Laguna Madre, Texas. /;/ Recent

Sediments, Northwest Gulf of Mexico, F. P. Shepard et al., eds. Tulsa:

.\mer. Assoc. Petrol. Geol. Pub.
Shenton, E. H. 1957. A studv of the Foraminifera and sediments of

1974 FORAMINIFERAL DISTRIBUTION 31

Matagorda Bay, Texas. Trans. Gulf Coast Assoc. Geol. Soc, v. VII,
pp. 135-150.

Shepard, F. p. and Moore, D. G. 1955. Central Texas coast sedimenta-
tion: characteristics of sedimentary environment, Recent history, and
diagenesis. Amer. Assoc. Petrol. Geol., Bull., 39: 1463-1593.

. AND . 1960. Bays of central Texas coast. In

Recent Sediments, Northwest Gulf of Mexico, F. P. Shepard, et al., eds.
Tidsa: Amer. Assoc. Petrol. Geol. Pub.

Stehli, F. G. 1966. Some applications of foraminiferal ecology. Proc.
2nd W. African Micropaleo. Coll. (Ibadan, 1965) , pp. 223-240.

AND Creath, W. B. 1964. Foraminiferal ratios and regional

environments. Amer. Assoc. Petrol. Geol., Bull. 48: 1810-1827.

Thornburv, W. D. 1965. Regional Geomorphology of the United States.

New York: John Wiley and Sons.
Thornthwaite, C. W. 1948. An approach toward a rational classification

of climate. Geog. Rev., 38: 55-94.
Towe, K. M. 1967. Wall structure and cementation in Haplophragmoides

canariensis. Contrib. Gush. Found. Foram. Res. XVIII, pp. 147-152.
. 1972. Invertebrate shell structure and the organic matrix

concept: Biomineralization, 4: 1-13.

and Cifelli, R. 1967. Wall ultrastructure in the calcareous

Foraminifera: crystallographic aspects and a model for calcification.

Jour. Paleo., 41: 742-762.
Up-shaw, C. F. and Stehli, F. G. 1962. Quantitative biofacies mapping.

Amer. Assoc. Petrol. Geol., Bull. 46: 694-699.
, Creath, W. B., and Brooks, F. L. 1966. Sediments and

microfauna off the coasts of Mississippi and adjacent states. Miss. Geol.,

Econ.. and Topo. Surv., Bull. 106.
Walton, W. R. 1964. Ecology of benthonic Foraminifera in the Tampa-

Sarasota Bay area, Florida. In Papers in Marine Geology, Shepard

Commemorative Vol., R. L. Miller, ed. New York: MacMillan.
Wantland, K. E. 1967. Recent benthonic Foraminifera of the British

Honduras shelf. Ph.D. Thesis, Rice Univ.

Appendix

On the Construction of Calcite Walls
IN Foraminifera

The Ihing calcareous Foraminifera have been divided into
two suborders on the basis of general test wall construction : the
Miliolina have nonporous, porcelaneous walls ; the Rotaliina ha\'e
a glassy appearance, and are penetrated by numerous pores.
Studies with the electron microscope have upheld this basic
distinction and have revealed the crystal arrangements underly-
ing and producing this difTerence, as seen by the light micro-
scope.

32 . BREVIORA No. 420

The miliolid test wall is composed of two layers of calcite
rhombs or needles: an inner, "randomly" oriented layer (which
is the thicker) and an outer, pavement-like layer, one rhomb
thick, with the rhombs oriented parallel to the surface. The
crystallization process in miliolid foraminifers has been observed
and reported by Arnold and by Lynts. The process is as follows :
Cytoplasm is extruded through the aperture ; it then takes on the
form of the new chamber. A layer of fibrous organic matter is
deposited on the surface of the chamber and will become the
"inner organic lining" of the test. After the new aperture is
formed, cytoplasm is again extruded, but this time it covers the
new chamber in a thin organic sheath, which is to act as ihe
crystallizing matrix.

Mineralization then occurs in two waves of crystal growth,
with the rhombs being nucleated either spontaneously or by
properly patterned organic molecules, but at many "randomly"
placed sites throughout the sheath of matrix. This results in
growth of the crystals in a nonoriented fashion within an im-
miscible solvent. This will be contrasted with the result of
oriented crystal s^rowth in the rotaliids.

Thus the randomly oriented rhombs in the inner layer are
the result of randomly oriented crystal nuclei. What special
mechanism operates to orient the surface rhombs? I believe this
is simply the result of surface tension at the protoplasm-seawater
interface, acting on the elongated crystals to align them parallel
to that surface. No biological directives are required; it is a
simple, physical process. No special crystallizing mechanisms
should be sought, and no adapti\'e significance can be attached
to this pa\^ement-like surface layer.

The mineralization process in a hyaline calcareous foraminifer
has been watched and reported by Angell, and a mechanism
for this process has been proposed by Towe and Cifelli on the
basis of the electron microscopic study of test wall sections. The
process is as follows: Cytoplasm is extruded through the aper-
ture, and takes on the shape of the chamber to be formed. A
fibrous ors^anic laver is secreted to cover the chamber. The cvto-
plasm is again extruded through the new aperture and covers
the new chamber in a thin organic sheath. To this point the
process is similar to that of the miliolids, but the fibrous organic
layer, which was merely an inner lining for the miliolid, has
taken on a new function. It apparently acts as a template for
calcite nucleation. CrystalHzation then takes place beginning on
this template, with the calcite crystals growing upward within

1974 FORAMINIFERAL DISTRIBUTION 33

the organic sheath. When the process is completed, there are
crystals and pores oriented perpendicular to the test surface.

It is my opinion that these pores and crystals are simply the
result of oriented crystallization of two immiscible substances
from an originally miscible solution — the cytoplasmic sheath.
My analysis of the process is as follows: The entire fibrous sur-
face of the new chamber can act as a nucleation template. How-
ever, as crystallization commences, both calcite and organic mat-
ter are coming out of solution. Since these are immiscible as
solids, there will be separation of the two phases. Organic mat-
ter will be excluded from the calcite crystal lattice and will
migrate toward, and collect in, relatively equally spaced organic
plugs on the template surface (the "pore processes" of Angell).
As crystallization continues, the same process will result in the
upward growth of the two separated phases: calcite will con-
tinue to crystallize on calcite, and organic matter on the pore
processes. The final result is a wall with oriented calcite crystals,
penetrated by organic plugs, which upon death and decay will
leave the characteristic "pores" of the hyaline calcareous fora-
minifers.

The results of the same process can be observed on a macro-
scopic level, and in an even more convincing manner, in your
home refrigerator. Most ice cubes exhibit "pore" structures
amazingly similar to those of the hyaline calcareous Foramini-
fera. They are formed by the entrapment of gases formerly
dissolved in the water, which must come out of solution during
crystallization. If freezing proceeds from the top down, the gas
cannot escape into the atmosphere, and space within the cube
must be provided. As crystallization proceeds the water becomes
saturated with the gas, and as it comes out of solution, it tends
to gather into bubbles at more or less equally spaced sites at the
ice surface. This, I am suggesting, is analogous to the separation
and collection of organic matter into the "pores" of foraminiferal
walls during their mineralization.

The total volume of pore space in the ice cube is dependent
on the amount of dissolved gas at the onset of crystalHzation, but
the pore size and density is related to rates of crystallization, as
indicated by a few simple experiments which I conducted. The
faster the cooHng rate of the ice, the smaller, and hence more
closely spaced are the pores. This is reasonable, as greater mi-
gration of the excluded molecules is possible with slower cooling.

The extension of the original bubbles, and hence the elonga-
tion of the pores, is the result of simple physical processes. As

34 BREVIORA No. 420

crystallization continues ice will tend to extend alreadv-existinor
ice crystals, and the gas will collect at sites already occupied b\
gas. When the entire solution is used up, crystallization stops
and the analogy is complete.

Thus "pores" are de\'eloped in the crystallization of ice with-
out the ,need of biologically derixed genetic directives, and, I
suggest, the same mechanism operates in the calcification of
foraminiferal walls. Surely no "adapti\^e significance" can be
ascribed to ice cube pores. Likewise, I believe we err in search-
ing for a "purpose" in the construction of foraminiferal pores.
I think the pores are simply the result of the simultaneous crystal-
lization of two immiscible substances upon a nucleation template.
Pores do not develop in the porcelaneous walls because nucle-
ation of the calcite crystals is at many sites, scattered throughout
the matrix, and exclusion of organic matter from the lattice dur-
ing crystal growth is accomplished by merely pushing it aside;
whereas, in the rotaliid wall, calcite is being nucleated over an
entire surface, necessarily forcing the organic matrix to gather
at particular sites. Thus, it was the mode of calcification, the
organic nucleating surface, which was selected for, and which
has adaptive significance, not the "pores." However, this does
not exclude the possibility that foraminifers use these "pores" in
the quest for specialized adaptations. By increasing the ratio of
organic matter to CaCOs (quite possibly through genetic con-
trol), it is possible to reduce the calcite wall to a mere lattice
work composed almost entirely of pore space, as in the genus
Globigerinoides, thereby lightening the test in preparation for a
planktonic habit. Thus, the very enlarged pores of Globigeri-
noides are in a close-packed condition resulting in hexagonal
openings and consequent inter\'ening small triangular calcite
pedestals serving as bases for the growth of the spines character-
istic of this genus. The spine growth can be simply ascribed to
the continued crystallization of calcite in the direction it was
started — a common phenomenon in crystal growth.

The factors of pore density and total porosity in recent plank-
tonic Foraminifera have been studied bv Be and are found to be
related to environment in a gross way. I suggest that total poros-
ity will be related to some factor or factors that govern the
matrix to calcite ratio (perhaps this is entirely genetic) and that
pore size and density will be found to be related to factors
go\erning rates of crystallization. And this might more closely
correlate with en\ironmental parameters. Perhaps in areas of
CaCO?, supersaturation crystallization will be most rapid, result-

1974 FORAMINIFERAL DISTRIBUTION 35

ing in many, minute pores spread over the test, as opposed to
larger, more widely spaced pores that might be found in regions
of en\'ironnientally controlled slow rates of crystallization.

In summary, I would like to emphasize that this is purely a
hypothesis for pore formation based on other hypotheses for cal-
cification mechanisms in Foraminifera, and quite possibly the
whole matter is more complex than what I have presented here.
However, I believe it is important to refresh our thinking by
coming to problems from new angles, by making analogies in the
biological world with things or processes in the purely physical
or chemical world. I especially think that Foraminifera are much
less complicated biologically than most workers currently sup-
pose. Much of their activity, their feeding, their shell construc-
tion can be duplicated in completely nonbiological systems.
Much of their shell morphology is predictable from a purely
geometrical point of view; for example, consider the stacking of
different sized spheres. Thus, in my opinion, Foraminifera, per-
haps more than any other group of organisms, can be utilized in
paleoecological studies, because they are basically simple physico-
chemical systems; they do not exert much biological pressure
against the environment, and hence they are closely governed by
the environment; that is, they must work within the confines of
molecular forces such as surface tension and crystal growth
processes.

Foraminifera must be examined in this new light if we are to
advance in our understanding of them. Foraminifera are not
molluscs; they do not have their sophisticated biological systems;
we must stop looking at them as if they do.

^-''

B R E V l.a.K A

LIBRARY

Miiseiiiii of Coiiiparative Zoology

us ISSN 0006-96'

Cambridge, Mass. 29 March ^^^^ygi^j^j^^^^^^ "^^^

A CASE HISTORY IN RETROGRADE EVOLUTION:

THE ONCA LINEAGE IN ANOLINE LIZARDS.

I. ANOLIS ANNECTENS NEW SPECIES,

INTERMEDIATE BETWEEN THE GENERA

ANOLIS AND TROPIDODACTYLUS.

Ernest E. Williams

Abstract. A new anole species bridges the gap between the genus Anolis,
diagnosed by the presence of adhesive subdigital pads under phalanges ii
and iii, and Tropidodactylus, diagnosed by the absence of such pads: Anolis
annectens has typical anoline transverse lamellae with microscopic hairs
and free distal margins only under phalanx ii; the third phalanx has only
keeled infradigital scales as in the species onca currently referred to the
monotypic genus Tropidodactylus. The genus Tropidodactylus is formally
synonymized with Anolis. A morphological series in the reduction of the
anoline adhesive pad that culminates in the condition seen in the species
A. onca is described.

The genus Tropidodactylus was erected in 1885 by Boulenger
in the second volume of his classic Catalogue of the Lizards in
the British Museum (Natural History) to receive the single
species described as Norops onca by O'Shaughnessy in 1875.

Neither the genus nor the species has received much atten-
tion since their description. They have, up to the present, been
very poorly known. The validity of the genus has not been ques-
tioned, since the difference between Tropidodactylus and Anolis
in the defining character of digital structure has seemed a sharp
and important one: all Anolis (including all those species classi-
cally referred to Norops) have under phalanges ii and iii ex-
panded adhesive digital pads, the smooth, flattened, transverse
lamellae of which are provided with microscopic hairs (Ruibal
and Ernst, 1965; Killer, 1968; Maderson, 1970; Lillywhite and
Maderson, 1 968 ) . The adhesive pad may be narrower or wider,
may be sharply set off ("raised") from phalanx i or not so set

BREVIORA

No. 421

JBC

Figure 1. Left: toe of an anole showing the "Norops" condition; Right:
toe showing the typical AnoUs condition.

off ( the Norops condition ) ( Fig. 1 ) and may ha\'e more or
fewer lamellae, but there is always some expansion, always
smooth trans\erse flattened lamellae under phalanges ii and iii,
and always microscopic hairs. Tropidodactylus, as known from
the single species onca, has been belie\ed to differ in the com-
plete absence of the hairs and of smooth lamellae and in the
presence of multiple keels on the infradigital scales. Although
in general habitus, including the presence of a large and typically
anoline dewlap, the species onca has unmistakably the appear-
ance of Anolis and is often so identified in collections; the digital
difference has always been regarded as quite worthy of generic
distinction. It seemed to support this position that, according to
Ruthven (1922), Tropidodactylus seemed more terrestrial than
any Anolis: "All of the specimens taken (17) were on the
ground. It is very shy and at the slightest cause for alarm dashes
into a hole."

However, Etheridge (1960) was unable to find any osteologi-
cal character on which to separate onca generically and re-
garded this species as the terminal, most specialized member of
his beta section of the genus Anolis. He was willing to retain
the genus only on the basis of "the e\'olutionary significance of
the loss of typical anole subdigital lamellae and the accompany-
ing alteration in mode of life."

George Gorman (1969), describing the karyotype of onca,

1974 ANOLIS ANNECTENS 3

found it to resemble closely two of the more primitive (or "typi-
cal") members of the beta group within Anolis [A. lineatopus
and A. opalinus). The onca karyotype (2n=30) with seven
macrobi\alents and eight microbivalents is characteristic of this
group within Anolis, and onca even resembles A. opalinus in
clear heteromorphism in chromosome pair seven. The only ob-
vious difference found by Gorman was that pair seven appeared
relati\ely smaller in onca than in the two compared Jamaican
anoles, "i.e. it might be considered an intermediate between
macrochromosomes and microchromosomes."

New collections of onca have been made by James Collins on
Margarita Island (reported by him in 1971), by Carlos Rivero-
Blanco and Abdem R. Lancini on the mainland of Venezuela
in and near Coro, by Bryan Patterson and the members of his
paleontological expedition in the same region, by the author,
A. S. Rand and A. R. Kiester on the neighboring Paraguana
Peninsula, and by the author, Jane Peterson, K. Miyata and
R. Salvato on the Paraguana isthmus and on the east side of the
Goajira Peninsula.

However, very surprisingly, as a summary of our knowledge
of the species onca was being prepared, a unique specimen in
the collection of the Field Museum of Natural History demon-
strated the existence of a new species that is an ideal intermedi-
ate between the genera Tropidodactylus and Anolis as currently
conceived. Differing trivially from onca in color and in some-
what greater size of the dorsal scales, it differs sharply in having
smooth lamellae under phalanx ii of the fourth toe, but keeled
scales under phalanx iii. It thus becomes impossible to make a
separation of two genera in the fashion that has hitherto been
customary. It is necessary either to describe a new monotypic
genus for the new species or to submerge Tropidodactylus in the
svnonvmv of Anolis. I choose the latter course and describe the
new species as :

Anolis annectens new species

Holotype: FMNH 5679, adult male.

Type locality: Lago de Maracaibo, collected by W. H. Os-
good between late January and early March, 1911.

Head (Fig. 2) : Head stout, a little longer than tibia. Head
scales unicarinate, 10 scales across snout between second canthals.
A shallow frontal depression. Nasal scale separated from rostral
by two intervening scales.

Supraorbital semicircles separated by one row. Supraocular

BREVIORA

No. 421

Figure 2. Head of A. aJinectens Holotype, dorsal view.

disk ill-defined, consisting of about 12 keeled scales, the largest
anteromedial, the disk separated by two rows of granules from
the scales of the supraciliary rows. Two parallel elongate supra-
ciliaries continued posteriorly by a double series of moderately
enlarged scales. Canthus distinct, canthal scales 5, second canthal
scale largest. Loreal rows 6, the lowest row not much larger than
those above it. Interparietal almost equals ear, separated from
the supraorbital semicircles by 2 scales. Temporal and supra-
orbital scales keeled, smallest in center of temporal region, dor-
sally grading into larger scales surrounding interparietals. Scales
behind interparietal somewhat smaller than those lateral and
anterior to it.

Suboculars separated from supralabials by one row of scales,
anteriorly not continued to canthal ridge, posteriorly ending
abruptly with one enlarged scale. Ten supralabials to center
of eye.

1974 ANOLIS ANNECTENS 5

Mental wider than long, in contact posteriorly with 7 unicari-
nate scales between infralabials. No differentiated sublabials,
but scales in center of throat smallest.

Trunk: Middorsal scales enlarged, hexagonal, keeled, grading
laterally into much smaller but keeled flank scales. Ventrals
much larger than dorsals, unicarinate, the keels in line. Post-
anal scales not differentiated.

Dewlap: Large, scales smaller than ventrals, keeled, arranged
in widely spaced rows except at the edge.

Limbs and digits: Hand and foot scales mul tic annate. No
scales on limbs as large as ventrals, unicarinate. Eight rather
narrow lamellae under phalanx ii of fourth toe, scales under
phalanx Hi of fourth toe multicarinate.

Tail: Tail round, all scales keeled, only ventral scales larger
than dorsals, 4 scales above, 3 below.
Size: 67 mm, snout-vent length.

Color (in alcohol) : Grey-brown with vague blotching and
mottling on flanks, limbs and tail more distinctly barred. A
round dark spot above each shoulder and a smaller spot between
these on the neck middorsally. Narrow oblique transverse bands
connect the shoulder spots across the middle of the back. A
transverse band directly in front of shoulder, indistinct on the
right side. On posterior midline two black spots, one in front of
sacrum, and one between two lateral sacral spots. Dewlap scales
are white, with black pigment around the bases of some of them.
Comparison with onca. Scales: The variation in squamation
seen in onca O'Shaughnessy completely includes that of the
single specimen of annectens except in two regards: the greater
size of the dorsal scales in annectens (Fig. 3) and the presence
under phalanx ii of smooth lamellae (Fig. 4).

Color: The color of annectens may also differ from that of
onca but the variability of onca is so great that the rudimentary
pattern seen in the type of A. annectens seems easily derivable
from an onca pattern. Nevertheless there is no exact or nearly
exact match in any of the extensive series of onca. The shoulder
spots of onca are roughly triangular or elongate, not round, as in
annectens. The neck spot and the two posterior midline spots
of annectens are not exactly matched in onca. The peculiar
distribution of dark pigment on the dewlap skin in annectens is
again without real parallel in onca.

Color in life of annectens is unknown. However, it may be
presumed from its similarity to onca that at least in body pattern
the preserved animal gi\^es a very fair impression of the live ani-

BREVIORA

No. 421

Figure 3. Dorsal scales. Above: A. onca. Below: A. annectens Hole-
type.

mal. Dewlap color, however, cannot safely be inferred from
specimens long preserved and this might be important.

A good description of color in life by William E. Duellman of
onca from 3 km SW of Cumana in Sucre, Venezuela, follows:
"Dorsum light brown mottled with gray, gray brown and black.
Venter creamy white, lightly flecked with grayish brown. Tail
medium brown above, cream below. Dorsolateral motthng on
WED 28685 forms more or less distinct paravertebral blotches
which are gray centrally and outlined in black. Dewlap bright
greenish ochre centrally, becoming orange peripherally, the whole
dewlap reticulated with bright orange brown and bearing white
scales. Iris bronze. Lining of throat bluish gray." [WED field
notes.]

Distribution: The locality for A. annectens is, most unfortu-
nately, inexact. It is not known whether Osgood collected the

1974

ANOLIS ANNECTENS

) '

\l\l

Figure 4. Fourth toe lamellae. Left: A. onca. Right: A. annectens Holo-
type. In the center the 4th toe of Anolis ("Norops") auratus is shown in
ventral and lateral view for comparison.

8 BREVIORA No. 421

specimen himself or had it brought to him, but the very inex-
actitude of the data and the absence of any further field notes
for the specimen (H. Marx, personal communication) indicate
most probably that Osgood did not have personal knowledge of
the collecting site. One additional specimen — an Anolis auratus
(formerly Nor ops auratus) — in the Field Museum received
from Osgood has the same inexact data. A. auratus is an animal
occurring in open grassy lowlands and, less abundantly, in bar-
ren thorn scrub with much cactus. It is a natural first hypothe-
sis that A. annectens has a similar ecology.

Osgood (1912) reports the itinerary of his 1911 expedition to
western Venezuela and eastern Colombia rather fully. Only
two of his stations are plausible for A. annectens in terms of the
expectation of a lowland grassy or arid habitat: El Panorama
8 miles due east of Maracaibo and the Empalado Savannas 30
miles further east. It is more probable, howe\'er, that both A.
annectens and A. auratus were among "the few specimens from
other places . . . obtained from natives in Maracaibo." How-
ever, an effort to discover annectens by collecting in a variety
of habitats on both sides of the Lago de Maracaibo in November
1972 and iVugust 1973 was unsuccessful. For the present no
better localization of this extraordinary annectant species is pos-
sible.

The distribution of A. onca is much better known, though
some of the literature records are clearly errors of identifica-
tion or of locality. The British A^Iuseum types described by
O'Shaughnessy were cited as from "\^enezuela'' and "Domin-
ica." The latter locality is certainly erroneous (Barbour, 1914^).
Specimens reported by Marcuzzi (1954) and Aleman (1952,
1953) from interior Venezuela are misidentifications. I record
below only the localities that I ha\'e personally \erified by ex-
amination of specimens (see Fig. 5) :

COLOMBIA. Guajira. Cabo de la Vela: FMNH 165159.
Two hours E El Cabo, near Cabo de la Vela: MCZ 85441.
El Cardon, S Cabo de la Vela: RNHL 7707. Maicao: USNM
115067. Manaure and Pajaro areas: USNM 151517-23. Media
Luna, E Cabo de la Vela toward Bahia de la Protete: MCZ
85440. Puerto Lopez, E shore Bahia Tucacas: MCZ 81554.
Rancheria del Cabo de la Vela: ZMA 14916. Riohacha:
UMMZ 54799, 54801-07, 54810-13; MCZ 14637.

^On Barbour's inquiry Boulenger wrote "The Tropidodactylus onca was
purchased of a dealer (Mr. Cutler) . The locality Dominica is, therefore,
open to doubt."

1974 ANOLIS ANNECTENS 9

\TNEZUELA. Distrito Federal. Near Caracas: USNM
107321. Falcon. Bahia de las Piedras, Paraguana: RNHL 7708
(3). Bariinu, Buenavista, Paraguana: ZMA 14917. Cerro de
Machuruia, E Santa Ana, Paraguana: RNHL 7709. El Mainon
ca. 5 km N ITrumaco: MCZ 133453. Isthmus of the Paraguana
Peninsula: MCZ 133456-65. Istmo de Medanos: UCV 272,
300, 461, 488. Los Algodones, 28 km NW Coro: MCZ
112386-98. Los Chipes, 41 km W Coro: MCZ 112399-407.
Medanos de Coro: ILS 743. Paraguana Peninsula: MCZ
133264-65 (hatched in Panama from females taken in Vene-
zuela), UCV 485. Parque Los Orumos, Coro: MCZ 139349-
50. Punta Baroa, Paraguana: UCV 448, 561. Rio Condore,
vicinity of Urumaco: MCZ 133455. Rio Seco on Caribbean
between Coro and Urumaco: MCZ 132735. Urumaco: MCZ
132734. Monagas. San Antonio de Maturin: MCZ 14648.
Margarita Island. Boca del Rio: ILS 578. Between El Agua
and Puerto Fermin: MCZ 109014, 110068-70. Near El Agua
on road from Punta de Piedras to Porlamar: MCZ 110064.
Ensenada de la Guardia, Laguna Arestinga: MCZ 110067.
Guacuco: UCV 364, USNM 139072, MCZ 110057. Laguna
Arestinga: ILS 102. Las Morites: ILS 1208. Las Robles:
USNM 79226-27. "S Las Robles, Porlamar: RNHL 7710 (3).
Matasiete: ILS 584. Morro de Moreno: RNHL 7711. Porla-
mar: ZMA 14918 (2). Plantio oeste de la Asuncion: ZMA
14915 (2). Three kilometers west of Porlamar: MCZ 110397.
Salamanca: ILS 561, 1231. Sucre. Cumana: KU 117080.
2.5 km SW Cumana: KU 117079. 3 km SW Cumana: KU
104369-70. Zulia: south of Paraquaipoa, W side Lago de Mara-
caibo: MCZ 139352.

Many of these localities are coastal, but although Collins
belie\'es onca to be a beach animal on Margarita Island, some
verified continental localities are well inland {e.g., Urumaco,
Falcon, Venezuela). x\ll localities, however, are extremely arid
lowland, usually within the zone called in the Holdridge terms
adopted by Ewel and Madriz (1968) for Venezuela "monte
espinoso tropical." A few records appear to lie in an adjacent
zone, "monte muy seco tropical." A few lie outside e\'en this
zone, i.e., USNM ^7321 "near Caracas" and MCZ 14648 "San
Antonio de Martin." These, howe\'er, are very imprecise locali-
ties. Figure 5 shows the distribution of "monte espinoso tropical"
and "monte muy seco tropical" for Venezuela according to Ewel
and Madriz. The Colombian localities are similar.

Howe\'er, the observations of the field party in the summer

10 BREVIORA No. 421

of 1973 suggest that the requirements of onca are more specific
than just the zone "monte espinoso tropical," Search for addi-
tional specimens of A. annectens led us into zones clearly within
the mapped areas but in which onca was apparently absent.
Anolis auratus was taken in these areas. (See the ecological
remarks below.)

Ecology

The relictual digital pad of annectens would seem to imply a
somewhat greater arboreal adaptation than that of onca. But
how terrestrial is onca?

No more recent obser\^er supports the apparent implication of
Ruthven's (1922) statement that onca uses burrows. On Mar-
garita Island Collins (1971) looked particularly into this point.
He remarks: "At times, a specimen being pursued would run
into a large hole in the sand opening into a tunnel. It should
be noted, however, that these holes are resting places made by
ghost crabs {Ocypode) and are not dug by Tropidodactylus.
It should also be noted that this was a rather infrequent mode
of escape, used by the lizard only when almost completely ex-
hausted." Collins points out that onca does climb when the
vegetation permits this. Where the vegetation was only a mat
of vines and branches, onca would clamber over or into this.
However, "in the area just north of Punta Montadero where
Mallotonia, a woody-stem plant, is dominant, the animal's be-
havior was very different. Here, when sighted, the lizard was
always on the ground. When pursued, the majority of animals
observ^ed would merely run among the ground cover. A few
specimens, however, were observed to climb the Mallotonia,
some to a height of 30.0 cm. Their climbing was clumsy and
ineffective."

Collins also took one animal sleeping on a branch of a low
bush.

On the continental mainland the observations of Carlos Riv-
ero-Blanco in July and August, 1970, are very useful. He re-
ports nine specimens taken on trunks of planted trees in a park
(Parque Los Orumos in Coro) within one meter from the
ground and two more taken in the same park from low branches
between one meter and a meter and a half above ground. Else-
where, in more natural situations, he reports them from piles of
dry branches and inside hollow dried cardon and cactus branches
partly buried in sand. He reported, however, that the local

1974

ANOLIS ANNECTENS

CARIBBEAN SEA

00 200

Figure 5. Map of the distribution of A. onca. Shading shows two vege-
tational zones in Venezuela (after Ewel and Madriz) : black :=. "monte
espinoso tropical;" cross-hatched = "monte muy seco tropical." + marks
known localities for onca.

people said that onca could be seen on the branches of a local
spineless euphorbiacean.

One of the animals obtained by the Patterson party in July
to August, 1972, had been taken on the outside wall of the
doctor's house in Urumaco. Again, most specimens observed
were among the branches of piles of dead plants buried in sand
( at Rio Seco, one to a pile ) . In another area one specimen was
seen lying motionless on a cobble in the full sunlight. Another
in still another area was seen on open ground in full sunhght,
\'ery cryptically colored and detected only by its motion.

The November 1972 party found males widely spaced out on
top of the pipe line that runs much of the length of the Isthmus
of the Paraguana Peninsula. Some were displaying. Others had
climbed to the top of posts, including fence posts. The re-
mainder, taken by day, including all females and the one juven-
ile, were on the ground in bare open spaces. None were seen

12 BREVIORA No. 421

in vegetation. Only one individual -- a female — was taken at
night, sleeping on a low bush less than a foot above the ground.

The August 1973 field party found onca primarily inside the
low thorn bushes that are very characteristic of the Paraguana
Peninsula, apparently coming out of the depths of these early
in the day and clambering around within these bushes much
more often than outside of them. Individuals were indeed seen
on the ground and both returning to and emerging from the
thorn bushes, but less frequently. Males were seen on the pipe
line and on fence posts but were not seen perched on rocks in
open sun in August as they had been so often in November.
The small thorn bushes were shared to some extent with young
Cnemidophorus lemniscatus, which climbed skillfully within and
on top of the bushes. As during the earlier visit to the Para-
guana, no onca were seen under or on the occasional large, quite
extensive thorn bushes.

The small thorn bushes of the Paraguana Peninsula provide
a very dense matrix in which climbing without adhesive pads is
obviously easy. The compact bases masked by grass also provide
places of concealment for onca and very probably sleeping sites. ^

The August 1973 party searched for onca and annectens in
many areas between Coro and Maracaibo, but only located onca
again S of Paraguaipoa on the east side of the Goajira Peninsula
( = the west side of Lago de Maracaibo ) . This area closely
matched the Paraguana Peninsula in appearance and especially
in the character of the vegetation, inchiding the sparse cover of
thorn bushes of small to moderate size.

Anolis auralus was repeatedly observed in areas in which onca
was lac king and never where onca occurred. It is clear that
auralus is less stenotopic than onca. It has been seen in lush
grassland, abundantly on a fence row beside a cattle pasture, and
sparsely in bare and harsh thorn scrub, often in situations that
seem climatically more rigorous than those from which onca is
known.

Aridity is certainly not the factor determining the presence of
onca. A special vegetational structure does seem characteristic
but there is another feature that may be even more important.
The notes by Rivero-Blanco call attention to the constant high
wind in the areas in which he observed onca. The November
1972 and August 1973 field parties also found the winds an

further data on the ecology, including thermal ecology, of onca will be
presented by Kenneth Miyata.

1974 ANOLIS ANNECTENS 13

impressive feature of the Paraguana isthmus. The onca locality
on the east side of the Goajira Peninsula was similarly windy.
The Patterson group, working well inland at Urumaco, were
constandy buffeted by wind also. Such winds may be a real
hazard and difficulty for lizards, preventing any strongly arboreal
adaptation, and wind in combination with aridity and sparse
\egetation may delimit the habitat of onca.

Discussion

The majority of iguanid lizards have infradigital scales with
multiple longitudinal keels. Tropidodactylus onca in this regard
appears by "the rule of parsimony" to have retained a primitive
condition. Why then do Etheridge, Gorman and myself con-
sider onca the derived extreme in anoles rather than the most
primiti\e surviving species? The hypothesis that a reversal of
evolution has produced a rather perfect simulacrum of a primi-
tive character state seems prima facie less plausible and more
complicated than a view that accepts an apparently primitive
character as genuinely so.

The argument is in fact a simple one : in no other regard does
onca seem primitive. In every character that Etheridge's skeletal
analysis regards as important, onca stands closest to the most
deri\^ed members of the beta section of Anolis. Etheridge ( 1960:
60) comments: "Except for the absence of specialized lamellae,
it is in no way distinguished from other anoles. Other features
of the genus, e.g. the absence of both splenial and angular, ab-
sence of pterygoid teeth, reduction of the parasternum ( = in-
scriptional ribs, Etheridge, 1965) etc. indicate that ''Tropido-
dactylus" is a specialized rather than a primitive anole. Accord-
ing to Ruthven (1922), the genus is strictly terrestrial, yet all
other features which mark the anoles as arboreal lizards are
present. Evolutionary loss of the anoles' specialized lamellae,
rather than retention of the pre-anole condition, probably offers
the most reasonable explanation of the [loss of] lamellae in
Tro pidodactylus ."

In karyotype also onca departs very much from the 1 2 macro-
chromosome-24 microchromosome pattern that occurs repeatedly
in primitive anoles, other diverse groups of iguanids (and in
other lizard families) and is believed to be primitive for the
Sauria generally (Webster, Hall, and Williams, 1972). The
primitive karyotype is found in many members of Etheridge's
alpha section of Anolis but in no betas, and, as Gorman (1969)

14 BREVIORA No. 421

has noted, onca belongs karyotypically, as in skeletal characters,
to one of the more highly derived groups of beta anoles.

Two external features are very characteristic of most Anolis —
the throat fan or dewlap and the adhesive pad with microscopic
hairs. Both are sometimes reduced within the genus (Williams,
1963). Both onca and annectens, however, have the dewlap
very large and very mobile, extremely well developed. A. onca
is known to use the dewlap very actively (observations of the
field party in November 1972), flashing it repeatedly, a derived
and not a primitive feature (Rand and WiUiams, in prepara-
tion).^ Of the two most basic anole characters, it is only the
second — the adhesive pad — that is absent in onca and tran-
sitional in annectens.

Some of the species that show the first stages of the degrada-
tion of the digital pad have been separated taxonomically as the
genus Nor ops. Schmidt (1939), describing the Mexican species
A. barkeri, called attention to the difficulty, made obvious by
more than one generic assignment for several of the species, of
making consistent distinctions between the genera Anolis and
Norops. Schmidt himself, though he placed barkeri in Anolis,
recorded the terminal phalanges of barkeri as "less distinctly set
off from the widened portion than in the normal Anolis.''

Moreover, it is now clear that any definition of Norops based
on digital features includes species that cannot be closely related.
Anolis aequatorialis and A. mirus of the trans- Andean regions of
Ecuador and Colombia have Norops-typc digits but are mem-
bers of the alpha subdivision of the genus (Etheridge, 1960;
Williams, 1963) and hence are on the other side of a basic
dichotomy within anoles from Cuban A. ophiolepis, Mexican
A. barkeri, A. tropidonotus, Colombian A. notopholis. central
Brasilian and northern Bolivian A. meridionalis, and northern
South American and Panamanian A. auratus, all anoles with
Norops-type digits (or an approach to them but belonging to
Etheridge's beta subdivision ) .

Even within the beta subdivision the species showing the
Nor ops-type condition are not closely related to one another.
Figure 6 adapts Etheridge's 1960 figure of beta anole relation-
ships to show the independent origin of the species of ''Norops.''
The numerals refer to the number of attached and free inscrip-
tional ribs; both the total number and the number of attached
ribs tend to decrease from primitive to advanced forms.

^Dewlap "flashing" is very characteristic of the possibly related forest
species, Anolis chrysolepis.

1974

ANOLIS ANNECTENS

" N 0 ro p s" 0 phi o]e p[s

"Tropidodactylus" onca 2 : 2

MAINLAND
BETAS

WEST INDIES
BETAS

Figure 6. A dendrogiam of relationship within the beta anoles. Modified
from Etheridge (1960) .

BREVIORA

No. 421

Figure 7. (from Collette, 1961) . Feet of five Cuban and one mainland
species of Anolis showing lamellae on the third toe of the left hind foot:
(a) alutaceus, (b) angusticcps, (c) sagrei, (d) caroUneusis, (e) porcatus,
(f) eqiiestris. Not to scale.

1974 ANOLIS ANNECTENS 17

Phylooeny apart, Anolis species can be arranged in a sequence
showing clear morphological stages in retrograde evolution.

1 . Narrowing of the digital dilations.

^Vithin any local Anolis fauna of more than a few species,
there are several conditions of the adhesive pads which Collette
(1961) has related to "arboreality." The broadest digital pads
are found in those species — "crown," "trunk-crown" and
"twig" anoles of Rand and Williams (1969) — which live high
in the trees or use twigs and leaves as perches {e.g., A. porcatus
and A. equestris in Figure 7e, f [copied from Collette, 1961]).
There is also some correlation with size, but those species spe-
cializing on the lower trunks and the ground — "trunk-ground"
anoles of Rand and Williams (1969) — have strikingly nar-
rowed pads although they may be larger than some of the com-
pared species [e.g., A. sagrei in Fig. 7c) .

2. Reduction of the number of digital lamellae.

While there is an evident functional difference between a wide
and a narrow pad in terms of area of adhesive surface, it is not
functionally obvious what the number of transverse smooth
"lamellae" has to do with adhesion, especially since many of the
lamellae in those species with the highest numbers are far distal,
crowded, small and much narrowed {i.e., at the tapering distal
end of the pad ) . It is, however, an empirical generalization
(and not only for Anolis; cf. Hecht, 1952 for the gecko Aristel-
liger) that the number of lamellae has a positive correlation with
size and with climbing efficiency. Correspondingly, those anole
species which climb least and use the ground more show fewer
lamellae than species of the same size with more arboreal habits.
Again contrast A. sagrei in Figure 6 with A. porcatus.

3. Loss of distinctness of the anterior margin of the pad {that
part under phalanx ii) as against the scales under phalanx i.

This is the character — the loss of "raised" character of the
pad — that has classically defined Nor ops {e.g., Boulenger) (see
Fig. 3 center: ''Nor ops" auratus) and is the maximal degree of
dedifferentiation of the pad seen except in annectens and onca.

The functional meaning of this stage is again unclear. But it
should be pointed out again that the phenomenon is not anoline
only and that genera have classically been recognized in the
Gekkonidae on whether the claw arises at the end of the ad-
hesive pad or "within the pad," i.e., dorsal to it, in the latter case
providing the pad with a projecting lip just as in Schmidt's
"normal Anolis."

4. The fourth and next to final stage in this retrograde series

18 BREVIORA No. 421

is found in annectens. As an intermediate between "Nor ops''
and Tropidodactylus it is interesting and perhaps unexpected.
In annectens the scales under phalanx iii are no longer either
wide or smooth; they are instead narrow and keeled. Under
phalanx ii, however, there is a residual pad, very narrow, it is
true, and the lamellae few in number, but still recognizably a
remnant of the classic anoline pad. The area under phalanx ii
is in any anole the region of the pad's maximum width (and
presumed effectiveness). One must assume that there is still
some selective value to the presence of a minimal adhesive pad
in annectens. However, the partial reversion to keeling in an-
nectens and the total reversion in onca may, perhaps, be more
easily understood in terms of morphogenetic patterns than in
terms of direct function in the environment: supradigital scales
are usually keeled in Anolis; unkeeled scales there are excep-
tional. The modified scales underneath the digit — the adhesive
pad — are obviously a specialized and limited morphogenetic
field. The distinctness and perfection of this field must be main-
tained by a continuing functional need greater than the cost in
ontogenetic complexity of maintaining the speciaUzed field. A
reversion to the keeled condition of the infra-digital scales, first
under phalanx iii and then also under phalanx ii, may therefore
be no more than the spread of the morphogenetic field of the
supradigital scales around and under the digit once the utility —
i.e., the selective value — of and hence the need for local dif-
ferentiation of very specialized adhesive lamellae has diminished.

5, The culmination of the retrograde series in onca is in one
regard imperfect. Hatchling onca have what appear macro-
scopically to be lamellae under phalanges ii and iii, not keeled
scales. First discovered in the collection of the Leiden Museum,
the only preser\^ed collection to have any very small specimens,
it is now confirmed on hatchlings from eggs laid by captive
female onca in Panama.

The "lamellae" of onca hatchhngs are astonishing enough to
require histological study. How closely do these lamellae match
the lamellae of "normal" Anolis? Hatchhngs and near hatch-
lings 27-30 mm in snout-vent length show "lamellae"; juveniles
just a few millimeters larger (34 mm, 41 mm) already show
keeled infradigital scales. How is this sharp ontogenetic change
accomplished?

A proper study of this question would be a digression here.
The problem has been referred to P.F.A. Maderson and he will
be reporting on it. Some of his preliminary observations are.

1974 ANOLIS ANNECTENS 19

howe\'er, germane at this time. The "lamellae" of hatchling
one a are pseudo-lamellae without the "hairs" (spinules) of the
true lamellae of an Anolis adhesive pad. They also lack the
spikes characteristic of larger juveniles (almost equal 34 mm
snout-vent length) and of adults of onca. In contrast annectens
has under phalanx ii anoline hairs and the lamellae have the
free distal edge that is characteristically anoline.

Hatchling onca, thus, though they seem superficially very dif-
ferent, are on their way to the adult onca infradigital condition.
The lamellar field, to return to that interpretation of the em-
bryological basis of these several conditions, is already extremely
weakened at the time of hatching and soon thereafter is wholly
substituted for by the field that produces spikes and keeling.

We have here emphasized a morphological series. The onca
hatchling is in this regard an intermediate in the series but a
very difTerent intermediate from adult annectens. The onca
hatchling already shows a breakdown of the lamellae and ad-
hesive pad and in the adult the breakdown is total. Annectens
is on another pathway. The pad under phalanx iii — always in
Anolis the least significant portion of the total adaptation — has
in annectens gone completely; retrograde evolution is for this
area complete. But under phalanx ii the pad is only narrowed
and the lamellae reduced in number; the latter are still fully
pilose, presumably still fully adhesive. A habitat for annectens
more genuinely "arboreal" than that of onca does seem plausible.

ACKNOW^LEDGMENTS

Work was supported by National Science Foundation grants
B 1980 IX and GB-37731X. I am grateful for assistance in the
field to A. Ross Kiester, A. Stanley Rand, Jane Peterson, Ken-
neth Miyata, and Richard Salvato. The Curators at the Field
Museum of Natural History (fmnh), the United States National
Museum (usnm), Kansas University (ku), the Universidad
Central de Venezuela (ucv), the Zoologisches Museum Amster-
dam (zma), and the Rijksmuseum van Natuurlijke Historic
Leiden (rnhl) have generously loaned material.

References

Aleman, G. C. 1952. Puntes sobre reptiles y anfibios de la region Baruta-
El-Hatillo. Mem. Soc. Cienc. Nat. La Salle, 12: 11-30.

• . 1953. Contribucion al estudio de los reptiles y batracios

de la Sierra de Perija. Mem. Soc. Cienc. Nat. La Salle, 13: 205-225.

20 BREVIORA No. 421

Barbour, T. 1914. A contribution to the zoogeography of the West Indies,
with especial reference to amphibians and reptiles. Mem. Mus. Comp.
ZooL, 44: 205-359.
BouLENGER, G. A. 1885. Catalogue of the lizards in the British Museum

(Natural History) 2: xiii + 497 pp. London.
CoLLETTE, B. 1961. Correlations between ecolog\^ and morphology- in ano-

line lizards from Havana, Cuba, and southern Florida. Bull. Mus. Comp.

Zool., 125: 135-162.
Collins, J. 1971. Ecological observations on a little known South Ameri-
can anole: Tropidodactylus onca. Breviora, No. 370: 1-6.
Etheridge, R. 1960. The relationships of the anoles (Reptilia: Sauria:

Iguanidae) : an interpietation based on skeletal morphology, xiii +

235 pp. University Microfilms, Ann Arbor, Michigan.
. 1965. The abdominal skeleton of lizards of the family

Iguanidae. Herpetologica, 21: 161-168.
EwEL, J. J., AND A. Madriz. 1968. Zonas de Vida de Venezuela. Ministerio

de Agricultura y Cria, Caracas. 265 pp.
Gorman, G. C. 1969. Chromosomes of three species of anoline lizards in

the genera AnoUs and Tropidodactylus. Mammalian Chromosomes

Newsletter, 10: 222-225.
Hecht, M. K. 1952. Natural selection in the lizard genus AristelUger.

Evolution. 6: 112-124.
Hiller, U. 1968. Untersuchungen zum Feinbau und zur Funktion der

Haftborsten von Reptilien. Z. Morph. Tiere. 62: 307-362.
Hummelinck, p. \\L 1970. A survey of the mammals, lizards and mol-

lusks. Fauna of Curacao, Aruba, Bonaire and the ^'enezuela Islands.

Vol. 1: 59-108.
Lillvwhite, H. B., and P. F. A. Maderson. 1968. Histological changes in

the epidermis of subdigital lamellae of AnoJis carolinensis during the

shedding cycle. J. Morph., 125: 379-402.
Maderson, P. F. A. 1970. Lizard glands and lizard hands: models for

evolutionary study. Forma et Functio, 3: 179-204.
Marcuzzi, G. 1954. Nota.? sobre zoogeografio y ecologia del medio xero-

filo venezolano. Mem. Soc. Cienc. Nat. La Salle, 14: 225-260.
Osgood, W. H. 1912. Mammals from western Venezuela and eastern

Colombia. Field Mus. Nat. Hist., Zool. Scr., 10: 29-66.
O'Shaughnessv, a. W. E. 1875. List and revision of the species of Anolidae

in the British Museum collection, with descriptions of new species.

Ann. Mag. Nat. Hist. 15(4): 270-281.
Rand. A. S.. and E. E. Williams. 1969. The anoles of La Palma: aspects

of their ecological relationships. Breviora, No. 327: 1-19.
RtiBAL, R.. AND \\ Ernst. 1965. The structure of the digital setae of

lizards. J. Morph., 117: 271-294.
RiTHVFN. A. G. 1922. The amphibians and reptiles of the Sierra Nevada

de Santa Marta, Colombia. Misc. Publ. Mus. /ool. \ri'h.. 8: 1-69.
Srii\TiDT. K. P. 1939. A new lizard from Mexico, with a note on the

genus Voro/K. Field Mns. Xat. Hist.. Zool. Scr., 24: 71 0.

1974 ANOLIS ANNECTENS 21

Webster, T. P., W. P. Hall, and E. E. Williams. 1972. Fission in the
evolution of a lizard karyotype. Science, 177: 611-613.

Williams, E. E. 1963. Studies on South American anoles. Description
of Anolis mirus, new species from Rio San Juan, Colombia, with com-
ment on digital dilation and dewlap as generic and specific characters
in the anoles. Bull. Mus. Comp. Zool., 129: 463-480.

1969. The ecology of colonization as seen in the zoo-
geography of anoline lizards in small islands. Quart. Rev. Biol., 44:
345-389.

' ' U\ ^ r I <_) T I y~i

^CJ

B R E V I 0 R A

Mii^^^^^^ip^^^lQ^^iiiparatiYe Zoology

LIBRARY

us ISSN 0006-9698

CAMBRiDAftRNfi^^.19749 March 1974 Number 422

'^Ag^ffith AMERICAN A NOUS:
+t¥ftge^NJE\V SPECIES RELATED TO
ANOLIS NIGROLINEATUS AND A. DISSIMILIS

Ernest E. Williams

Abstract. Three new Anolis species are described from widely scattered
localities in Colombia and Venezuela. Together with Anolis nigrolineatus
and Anolis dissimilis they appear to represent a natural subgroup of the
punctatus group of South American alpha anoles.

The lizard fauna of South America is poorly understood but
more than that it is little known. It is, for example, very prob-
able that there are many lizard species to be discovered in the
continent's remoter and more obscure areas. The three new
anoles here described are cases in point: they are from areas
quite remote or obscure - — one from a small river valley in
Santander and the poorly known states of Tachira and Trujillo
in Venezuela, another from a camp in remote Caqueta in
Colombia, and still another from a mission in the delta at the
mouth of the Orinoco.

More interesting, however, than the existence of new species
in little explored areas is the close resemblance of these newly
discovered, perhaps isolated anoles to species occurring at very
great distances from them. The most extreme instance is the
similarity of the anole from the mouth of the Orinoco to a form
from Madre de Dios Province in Peru. However, the distances
between the other forms that must be compared are relatively
small only in the context of the immensity of South America.

Even in South America it is quite unusual to be compelled to
describe related species from such small samples as are available
for the three new forms (one, one and five), especially when
these are spread over so wide an area with no series available
for any locality. This may point to a special difficulty pecuUar
to small arboreal species. The fauna of open formations is usu-

2 BREVIORA No. 422

ally obtainable in some appreciable numbers wherever it occurs.
The species of forests are rarer or more difficult to obtain, but
most probably both. Those elements of the forest fauna that
occur well up in the trees or at least in thick \'egetation are likely
to be the last to be known. On morphology and affinity, al-
though only for one is anything known directly of the ecology,
the present three new species appear to belong to this most diffi-
cult group.

All three anoles are so close to Anolis nigrolineatus and Anolis
dissimilis (Williams, 1965) that they, like these, must be as-
signed to the punctatus group of the alpha section of South
American anoles.

A. nigrolineatus (Williams, 1965) was described from two
specimens, both with questionable localities in southeastern Ecua-
dor. Two additional specimens have since been discovered in the
collections of the University of Michigan. These not only provide
the first good locality for the species (Playas de Montalvo, Prov.
Los Rios, Ecuador) but provide a better comparison with the
new but very closely related species from eastern Colombia and
western Venezuela which I call :

Anolis nigro punctatus new species

Holotype: ILS 21, an adult male.

Type locality: El Diamante, Norte de Santander, Colombia.

Paratypes (all adult females). ILS 20: Toledo, Norte de
Santander, Colombia; MCNC 5395, Villa Paez, Edo Tachira,
Venezuela; MCZ 136175, Quebrada Honda on road from Tru-
jillo City to San Lazaro, Edo Trujillo, 4700 feet.

Diagnosis. Close to A. punctatus (cf. the slightly swollen snout
in the male) but differing in color and squamation. Closer still
to A. nigrolineatus but difTering in wider head, apparently larger
size (male 72 mm in snout-vent length rather than 46 mm), in
the absence of the narrow middorsal black hne and of the broad
black spot in the dewlap. Nostril without a differentiated an-
terior nasal scale ( Fig. 1 ) . An apparently greater number of
lamellae under phalanges ii and iii of the fourth toe (21-22
rather than 18-19).

Description. (Paratype variation in parentheses.) Head:
Head scales flat, obscurely wrinkled. Seven scales (7-10) across
snout between second canthals. Five scales (6-8) border rostral
posteriorly. Circumnasal scale separated from rostral by one
scale (or in contact). Four scales between supranasals. Snout

1974

THREE NEW ANOLIS SPECIES

Figure 1, Anolis nigropunctatus Holotype. Dorsal view of head.

,-«:2S?p?c)0'

Figure 2. Anolis nigropunctatus Holotype. Lateral view of head.

BREVIORA

No. 422

Figure 3. Anolis nigropunctatus Holotype. Underside of head.

somewhat swollen, protuberant, overhanging lower lip (snout
not swollen in ?).

Supraorbital semicircles separated medially by 2 scales (2 or 1
or in contact) and from the supraocular disks of each side by a
single row of subgranular scales. Supraocular disk of 9 (8-12)
indistinctly wrinkled scales. Supraciliaries 1-2, continued pos-
teriorly by granules. Canthus distinct, canthals 5 (5-6), second
and third canthals longest (third longest). Loreal rows 5 (4-5),
uppermost largest (uppermost largest or subequal) .

Temporals and supratemporals granular, grading into enlarged
scales surrounding interparietal (obscure in second female),
which is smaller than the small round ear (almost equals ear)
and separated from the supraorbital semicircles by three (1-4)
scales. Several of the scales surrounding interparietal larger than
that scale (or 2/3 that size). Scales posterior to interparietal
grading gradually into dorsal granules. No enlarged supratem-
poral rows (indistinct supratemporal rows).

Suboculars weakly keeled, in contact with supralabials, grad-
ing posteriorly into the supratemporal granules and anteriorly
separated from canthals by one scale. Seven supralabials to
center of eye.

Mental semidivided, each part almost as wide as deep (wider

1974 THREE NEW ANOMS SPECIES 5

than deep), the whole in contact with 3 (4) throat scales
between large, smooth sublabials which indent it. Sublabials
enlarged, two (3) in contact with infralabials. Gular scales
smallest medially, grading laterally toward sublabials.

Trunk: Middorsal scales not differentiated from flank scales
(two middorsal rows slightly enlarged), obtusely keeled. Ven-
trals larger, smooth, quadrate, imbricate, in transverse rows (not
imbricate). Lateral chest scales obtusely keeled (smooth).

Dewlap: Large (smaller in ?, extending only between fore-
limbs), extending nearly to middle of belly. Scales at the edge
much longer than ventrals (in ? smaller than or equal to ven-
trals). Lateral scales narrow, elongate, in well-spaced rows
(close packed in $) , separated by naked skin.

Limbs and digits: Scales on limbs smooth or unicarinate,
largest on both arm and hind limb (smaller than ventrals).
Supradigital scales multicarinate. Twenty-one (22) scales under
phalanges ii and iii of fourth toe.

Tail: Compressed, without verticils or dorsal crest. Two
distinctly keeled middorsal rows; the ventralmost two rows even
more distinctly keeled. Greatly enlarged postanals (absent in
9) present. Scales behind vent smooth.

Color (as preserved) : d above brown, irregularly punctate
with black; below light brown with a few small lateral black
spots. Dewlap, both scales and skin, light. $ same as above
except with a broad middorsal zone light brown, mottled and
lined with grey and dewlap with light scales and pigmented skin.
Size: Type (snout-vent length) 72 mm. Paratypes: 60, 56,
55 mm.

Comment. A. nigropunctatus (see Table 1) is extremely close
to A. nigrolineatus but quite adequately distinct. The two newly
discovered specimens of nigrolineatus (UMMZ 84114-15) fully
confirm the scale and color characters noted in the original
description and have the same small size. In the feature of a
simple single scale (nasal or circumnasal scale) surrounding the
nostril, I regard nigropunctatus as more primitive than nigro-
lineatus. The scale called "anterior nasal" in the latter I believe
to be a modification of a scale originally anterior to that sur-
rounding the nostril, now become enlarged and triangular, over-
lapping the anterior margin of the primitive circumnasal scale.
The higher number of toe lamellae in nigropunctatus accord
with its larger size.

Ecological notes are available only for MCZ 136175 for which
J. A. Rivero records: "On leaves three feet from the ground at
edge of road near a stream."

BREvioRA No. 422

flj

1— 1

C/1

-c

4-^

>-H

► -H

c«

^ A

o -S

CI ^ ^ ^ A

o-Sl^o.St;^ ^^A

^ S^ H I § s

-Q Q -5 cs c o i-

4-J

•"^

■^^

1— (

U-l

'rt

■4->

c/;

■^^

o
U

•l-H

l-H

•2

1— t

o
>

l-H

C3
*->

O
o

4-1
(J
<U

-5

c
o

t/3

'(5

3
u

• 1— 1

1/3

■4-J

3
O

' - ^

CM «0

' 4^ V

S ^ V3 ^ C ^

:S 3 .2 H -i o

Q w j- c 2 "C _

;^ = 3 p c ^

c«

4->

3
O

W5
V3

</i

—

s->

U
•l-H

1—

4-)

• •iH

S-i

i-H

4-1

</)

•-^

1974

THREE NEW ANOLIS SPECIES

i-H p-H T^ O *^

A ^ I Ji S 00

OS (J -

3 CM

.bo

•2 -s -H OS (JJ

« ^A o77+c 2

o
.to

K C/} 4-J M

4-J

-d

TJH

•4— >
• l-H

13
c

,12

"7^

c«

4-J

<y5

• i-^

"73

"rt

CI.

,_«

flj

•fH

ITl

c«

• 1-^

c«

C/3

«J

c«

-d

4-1

-a

4-1

O
r;
55

BREVIORA

No. 422

'i^tidtcQ^V^^^Zt^-Jt^----'^

Figure 4. Anolis caquetae Holotype. Dorsal view of head.

The two remaining undescribed species appear to be closest
to A. dissimilis, but the species geographically more remote is
more similar than that which is spatially intermediate. The latter
is clearly the primitive member of the series and, coming from
the upper Rio Apaporis, is within the Amazonian faunal prov-
ince but in one of the remoter peripheral parts of that region.
I name it after the Department of Colombia from which it
comes.

Anolis caquetae new species

Holotype: MCZ 131 176, an adult male.

Type locality: Camp Soratama, Upper Apaporis, Caqueta,
Colombia.

Diagnosis. Close to A. dissimilis but without the tail crest
characteristic of that species and with a different coloration.

Description. Head: Most head scales smooth, some on the
anterior snout unicarinate. Scales in frontal depression distinctly
smaller than surrounding scales. Ten flat scales across snout
between the second canthals. Eight swollen scales bordering
rostral posteriorly. Nasal scale anterior to canthal ridge with
one lower and one anterior scale separating it from rostral (see
Fig. 5 ) . Seven swollen scales between supranasals. Snout some-
what swollen, protuberant, overhanging lower lip.

Supraorbital semicircles separated from each other by a single
row of small scales, in contact with the supraocular disks, which
consist of 24-28 enlarged smooth scales grading into granules

1974

THREE NEW ANOLIS SPECIES

Figure 5. Anolis caquetae Holotype. Lateral view of head.

Figure 6. Anolis caquetae Holotype. Underside of head.

10 BREVIORA No. 422

anteriorly, posteriorly and laterally. A single enlarged supracili-
an^ continued posteriorly by granules. Canthus distinct, canthal
scales 7, the third canthal largest, then diminishing gradually
forward. Loreal rows 5, the lowest distinctly the largest. Tem-
poral and supratemporal scales granular, grading into enlarged
scales lateral to the interparietal. A weakly indicated double
supratemporal row of large granules extending posteriorly from
the orbit. Interparietal very large, much larger than the small
round ear opening, in contact with the supraorbital semicircles.
Scales lateral to the interparietal distinctly enlarged, but those
posterior to it hardly larger than the dorsal granules, about equal
to the supratemporal and temporal granules.

Suboculars smooth, broadly in contact with supralabials,
grading into large granules behind the eye; anteriorly grading
into loreals. Seven supralabials to the center of the eye.

Mentals deeper than wide, in contact with 4 throat scales
between the sublabials. Sublabials large, wide, three to four in
contact with infralabials. Central throat scales small, not grading
into sublabials, swollen, \'aguely keeled.

Trunk: Middorsal scales granular, swollen, smooth, not dif-
ferentiated from flank scales. Ventrals larger than dorsals,
weakly keeled, imbricate.

Dewlap: Dewlap small, scales larger than ventrals, close set.

Limbs and digits: Hand and foot scales obscurely multicari-
nate. Largest arm and leg scales unicarinate, those of the arm
somewhat larger than ventrals. Twenty-two lamellae under
phalanges ii and iii of fourth toe. Postanals?

Tail: Tail compressed with two middorsal rows obtusely
keeled and the two midventral rows larger, sharply keeled. Ver-
ticils not evident. Lateral caudal scales increasing in size toward
ventrals.

Color (as preserved) : Dorsum brown with a narrow black
vertebral line bifurcating on nape. Broad oblique transverse
banding of obscure dark blotches, limbs obscurely banded. Belly
and throat light brown, sparsely punctate with darker. Tail very
obscurelv banded.

Size (snout-vent length) : 57 mm.

Comment. Like a number of South American anoles that do
not seem closely related {e.g., A. jacare, A. nigropunctatus),
A. caquetae has a double row of scales surmounting the tail
rather than the more usual one. This is very different from the
tail crest of a single row of enlarged triangular scales character-
istic of A. dissimilis. This difference does not seem, however, a

1974 THREE NEW ANOLIS SPECIES 11

bar to the close relationship. A similar if less extreme difference
exists between A. nigropunctatus and A. nigrolineatus. In other
details of squamation A. caquetae and A. dissimilis are very
much alike (Table 1). They differ strikingly, however, in color
and pattern. The dark dorsal color of dissimilis with the light
line from supralabials to shoulder has no elements of similarity
to the middorsal dark line and broken crossbanding of A. caque-
tae. On the other hand, the vestigial dark line may indicate
relationship to A. nigrolineatus, which in squamation (Table 1)
differs most prominently in features associated with the huge size
of the interparietal in A. caquetae.

The last species requiring description comes from the delta of
the Orinoco. I have therefore named it :

Anolis deltae new species

Holotype: (MCN) 2031, adult male.

Type locality: Mission Araquaimujo, Delta Amacuro, Ter-
ritorio Federal, Venezuela.

Diagnosis. Very close to A. dissimilis including the presence
of a distinctive tail crest, but with a blunter, shorter head, a
differentiated anterior nasal scale, a larger interparietal with
larger scales surrounding the interparietal and more lamellae
under phalanges ii and iii of fourth toe.

Description. Head: Most head scales smooth, swollen, a few
obtusely keeled. Eight scales across snout between second can-
thals. Six scales border rostral posteriorly. Anterior nasal scale
in contact with rostral. Four scales between supranasals. Scales
in frontal depression smaller than surrounding scales.

Supraorbital semicircles in contact, separated from the supra-
ocular disks on each side by one row of scales. Supraocular disks
of 12-14 strongly enlarged scales. Supraciliaries one on each
side, continued by granules. Canthus distinct. Canthal scales 6,
the second and third largest. Loreal rows 4, the lowermost
largest.

Temporals and supratemporals subgranular, grading into en-
larged scales surrounding interparietal. Interparietal very large,
larger than ear, in contact with supraorbital semicircles. Scales
behind interparietal grading gradually into dorsal granules. Sub-
oculars in contact with supralabials, grading posteriorly into
supratemporal granules, anteriorly separated from the canthals
by one to two scales. Seven supralabials to the center of the eye.

Mental wider than deep, in contact with four throat scales,

BREVIORA

No. 422

->t> — '-w'-

Figure 7. Anolis deltae Holotype. Dorsal view of head.

Figure 8. Anolis deltae Holotype. Lateral view of head.

1974 THREE NEW ANOLIS SPECIES 13

Figure 9. Anolis deltae Holotype. Underside of head.

set in a gentle forward arc between sublabials. Sublabials en-
larged, two in contact with infralabials on each side. Gular
scales subequal centrally but grading laterally into sublabials.

Trunk: A few middorsal rows slightly enlarged, obtusely
keeled, grading into flank granules. Ventrals larger, smooth,
quadrate, imbricate, in transverse rows.

Dewlap: Large, extending nearly to midbelly. Scales at edge
as large as ventrals. Lateral scales narrow, elongate, in rows
separated by naked skin.

Limbs and digits: Largest limb scales unicarinate, almost
equal ventrals. Supradigital scales obscurely uni- or bicarinate.
Twenty-four lamellae under phalanges ii and iii of fourth toe.

Tail: Most of tail missing but a distinct crest on the portion
present. Enlarged postanals absent. Scales behind vent smooth.

Color (as preserved) : Straw. A series of broad but vague
darker blotches middorsally. Obscurer and quite irregular spots
and mottling on flanks. Belly with vague markings. Above and
below head and limbs very obscurely mottled. Dewlap skin and
scales light.

Size (snout-vent length) : 58 mm.

Comment. The tail crest of A. deltas and A. dissimilis in
particular is a highly distinctive common feature. It is entirely a
crest of slightly raised keeled scales that gives the appearance of
a serrate upper border to the tail, not at all like the huge tail

BREVIORA

No. 422

80"

70"

60°

50"

40°

A nigrolineatus

• nigropunctatus

A caquetae

O dissimilis

10'

—10"

-20°

Figure 10. Distribution of the Anolis of the A. nigrolineatus subgroup.

fins supported by vertebral spines of the considerable number of
VV^est Indian species that have compressed crested tails — not
therefore impressive except that it is very unusual in South
America. Even the South American giants (the latifrons group
sensu stricto) , though they have compressed tails, lack any sort
of crest. The closest resemblance in tail type is perhaps provided
by the anoles of the pentaprion group (Myers, 1971) in which
the serrate crest, however, is surely convergent, since these are
beta anoles belonging to quite a distinct section within the genus
Anolis.

A. deltae is quite different from dissimilis in color and pattern,
closer in this to A. caquetae which it resembles also in the
strongly enlarged interparietal. It differs, however, from both
species in the enlarged scales behind the interparietal, markedly
larger than the dorsals.

Discussion. The fi\'e species that have been discussed here are
perhaps a natural subgroup — the A. nigrolineatus subgroup —

1974 THREE NEW ANOLIS SPECIES 15

of the punclatus species group. They are all allopatric and the\
ring changes on just a few characters. If they are such a group,
there are two series on the basis of affinities and geography —
an inner series, peripheral to Amazonia proper, in the upper
reaches of Amazonian tributaries and the Orinoco, and an outer
series with one species west of the Andes in Ecuador (almost at
the southern Hmit of Anolis species west of the Andes) and
another in valleys in the northern and northeastern continuation
of the Andes in Colombia and Venezuela.

So far as current information extends, none of these overlap
with the two larger Amazonian species of the punctatus group
— A. punctatus itself and A. transversalis. These widespread
species, which show little geographic variation, lie internal to
even the inner series of the nigrolineatus subgroup, A. punctatus
with a very wide distribution in the Brazilian Atlantic forest,
Amazonia and in the Guianas, A. transversalis at least partly
sympatric with punctatus in western Amazonia. With South
American anoles so little known, this apparent geographic pat-
tern could well be factitious. However, A. punctatus and A.
transversalis are among the first collected of anole species wher-
ever they occur. Their absence from the collections that record
the dissimilis-caquetae-deltae series may therefore be real.

Acknowledgments

Research on South American anoles has been supported bv
NSF grants B- 1980 IX and GB-37731X. I am grateful to Dr.
Juan Rivero and Dr. Fred Medem for the gift of material, and
to Brother Niceforo Maria (Instituto La Salle (ILS), Bogota),
Frof. Ramon Lancini (Museo de Ciencias Naturales (MCN),
Caracas ) and Dr. Charles Walker ( University of Michigan
A^useum of Zoology (UMMZ) ) for the loan of specimens.

Literature Cited

My^rs, C. W. 1971. Central American lizards related to Anolis pentaprion:
two new species from the Ckjrdillera de Talamanca. Amer. Mus, Nat.
Hist., Novitates, No. 2471: 1-40.

Williams, E. E. 1965. South American Anolis (Sauria, Iguanidae) : Two
new species of the punctatus group. Breviora, No, 233: 1-15.

B R E V JUQlK a

LIBRARY

iiseiim of Comparative Zoology

us ISSN 0006-9(

— — HARVARD

Cambridge, Mass. 29 March l^^i^HVERSTT^'^^^^^ ^^"^

A NEW SPECIES OF PRIMITIVE ANOLIS

(SAURIA IGUANIDAE) FROM THE

SIERRA DE BAORUCO, HISPANIOLA

Albert Schwartz^

Abstract. A new species of primitive anole is described from the Sierra
de Baoruco in the Republica Dominicana. The species is compared with
its relatives occultus (Puerto Rico) and darlingtoni and insolitus (His-
paniola) . Data on the ecology of the new species, in relation to A. insolitus
and A. occultus, are presented.

On the Antillean islands of Puerto Rico and Hispaniola occurs
a small group of anoles which has been known from only three
species, two of which were only very recently discovered and
named. The earliest discovery of a member of this trio of lizards
was that of Anolis darlingtoni Cochran, of which the holotype
and still only known specimen was taken by P. J. Darlington in
1934 at Roche Croix on the northern slopes of the Haitian Mas-
sif de la Hotte on the Tiburon Peninsula at an elevation of about
5000 feet (1525 meters). Cochran (1935) named this species
Xiphocercus darlingtoni in recognition of its resemblances to
X. valencienni Dumeril and Bibron from Jamaica. The genus
Xiphocercus is now in the synonymy of Anolis; the two species
are' very similar in general habitus and habits but are not closely
related. Etheridge (1960: 92) stated that although these two
species were externally similar, they differed in critical osteo-
logical details (caudal vertebrae, number of attached and float-
ing chevrons, and presence of autonomic septa). X. valencienni
was like other Jamaican anoles in osteological characteristics and
X. darlingtoni Uke several Haitian species. It seemed obvious
that these two species were erroneously associated at the generic

^Miami-Dade Community College, Miami, Florida 33167.

2 BREVIORA No. 423

level, and that they represented a convergence between repre-
sentatives of two anoline stocks of Jamaica and Hispaniola.

The second member of this complex of anoles was discovered
on Puerto Rico in 1963 by Juan A. Rivero in the Cordillera
Central near Cerro de Punta at an elevation of 1338 meters.
Anolis occultus was described by Williams and Rivero (1965)
from a suite of specimens from various upland Puerto Rican
localities and at the same time Thomas ( 1 965 ) summarized the
ecological data and field observations that he had accumulated
while collecting the majority of the type-series. Later, Webster
(1969) presented further information on the ecology of this
forest-dwelling species.

The third member of the trio was first secured by Clayton E.
Ray and Robert R. Allen in 1963 at La Palma, La Vega Prov-
ince, Republica Dominicana, at an elevation of 3500 feet (1068
meters) in the Dominican Cordillera Central. Anolis insolitus
was described by Williams and Rand ( 1 969 ) from six specimens
taken at the type-locality. These authors also made extensive
comparisons between darlingtoni, occultus, and insolitus, which
form a small complex of primitive anoles. That they are distinct
species is unquestioned. But WiUiams and Rand (1969: 10)
noted that "Certainly the most plausible assumption based on
current evidence is that darlingtoni and insolitus are geographic
representatives ... of one stock. This assumption, however,
leaves the extreme size disparity of these allopatric species with-
out easy explanation." At the time this statement was written,
the largest known insolitus had a snout-vent length of 34 mm
and the holotype of darlingtoni has a snout-vent length of
72 mm. The allusion of WilHams and Rand to these two species
as "geographic representatives" is due to the fact that one {dar-
lingtoni) occurs on the Hispaniolan south island whereas the
other (insolitus) occurs on the Hispaniolan north island. These
two terms have come into common usage among herpetologists
who deal with Hispaniolan amphibians and reptiles, since they
apply to two island masses, formerly separated, but now joined
by the low-lying Cul de Sac-Valle de Neiba plain. These two
islands have, to a large extent, distinctive faunas; there has
naturally been some invasion and interchange of species, but this
has been primarily of lowland forms. The montane faunas of
these two paleoislands remain remarkably distinct today, and it
is only reasonable to assume that these montane faunas, despite
a common origin in many cases, have been completely discon-
tinuous for a very long period.

1974 Anolis sheplani 3

Williams and Rand (1969: 10) also pointed out that of the
21 Hispaniolan species of Anolis, seven had been described
within the last ten )ears; they also stated that they felt that the
list of species presented in their summary was incomplete and
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."

Under the sponsorship of two National Science Foundation
o-rants (G-7977 and B-023603) between 1968 and 1972, I col-
lected in the Republica Dominicana; comparable collections
were made by Richard Thomas in Haiti. In the former country,
we were successful in securing specimens of two new species of
Anolis. The description of one of these (Schwartz, 1973) has
already been completed. Although this species, from the Cordil-
lera Central, is a large and exceptionally handsome lizard, it
does not add materially to our knowledge of the Antillean history
of the genus Anolis. It is a species living in deciduous forest of
the Central uplands at elevations above 5400 feet (1647 meters),
and as far as present evidence indicates, it is an endemic Cordil-
lera Central species of the monticola complex.

The second species is far more interesting and intriguing.
This anole is an inhabitant of hardwood forests in the Sierra de
Baoruco, the easternmost massif of the chain of three montane
masses on the Hispaniolan south island. It is in the Massif de la
Hotte, the westernmost of this chain of three ranges, that A.
darlingtoni occurs. Thus, we now know of two species of this
group of anoles from the Hispaniolan south island. The doubts
expressed by Williams and Rand concerning the geographical
equivalence of darlingtoni and insolitus have been shown to have
a sound basis, since there is little question that this new species
is the south island analogue of the north island insolitus, and
that the larger darlingtoni stands alone among other members
of the group as a much larger lizard. Details of the relationships
between all four species will be presented by WiUiams and
Eth'eridge in a separate publication; it is my aim herein to
describe the new species, give details of its variation, and com-
pare it with the three remaining species, as well as to present
field observations made during 1971.

The first specimen of this new taxon was obser^^ed by myself
on the night of 29 August 1971, as it slept on a dry hanging
\ine under a low \'ine canopy shelter adjacent to the road in the
Sierra de Baoruco. Its sleeping posture and general configura-
tion, despite the fact that it was some ten feet (3.1 meters) above

4 BREVIORA No. 423

me, attested that it was a species related to A. insolitus and A.
occultus. Because of the peculiar situation where the lizard slept,
I was reluctant to make the attempt to secure it. This reluctance
was due to the fact that I and my companions have spent many
nights and days collecting in the Sierra de Baoruco since 1963
without seeing a lizard of this sort. Bruce R. Sheplan was in-
vited to make the attempt at retrieving the lizard, and he very
carefully ascended the muddy road cut, crawled gingerly beneath
the vine canopy without disturbing the vegetation, and handily
secured the lizard. We later learned that there was no need for
such care in dealing with this Anolis, since, like insolitus and
occultus, it is extremely tolerant of any sort of nocturnal disturb-
ance and determinedly clings to its perch despite disturbances.
A second specimen was secured later the same evening from a
similar sleeping situation only 15 feet (4.6 meters) from the first
individual. Two more visits to the same general area yielded a
total of 16 lizards: it is obvious that at least locallv this new
species is not rare, but on the other hand its ecological require-
ments (and these can be deduced only from its sleeping sites)
may be extremely rigid. The locality itself is not difficult of ac-
cess and to my eye is little different from many other regions irt
the Sierra de Baoruco uplands, areas such as the Las Auyamas-
Valle de Polo region which have been extensively collected. Still,
the new species is known only from one fairly circumscribed
area. In honor of Mr. Sheplan, whose care and interest not only
were responsible for the first two specimens but also for most of
the subsequent material, I propose that the new species be named

Anolis sheplani new species

Holotype. National Museum of Natural History (USNM)
194015,' an adult male, from 13.0 mi. (20.8 km)'SE Cabral,
3200 feet (976 meters), Barahona Province, Republica Domi-
nicana, taken by Bruce R. Sheplan on 29 August 1971. Original
number Albert Schwartz Field Series (ASFS) V30309.

Paratypes. ASFS V30310, same data as holotype; Carnegie
Museum (CM) 52300, same locality as holotype, 30 August
1971, D. C. Fowler; ASFS V30326,' USNM 194016-17, CM
54140-41, American Museum of Natural History (AMNH)
108822, Museum of Comparative Zoology (MCZ)' 125641-42,
12.3 mi. (19.7 km) SE Cabral, 3300 feet (1007 meters), Bara-
hona Province, Republica Dominicana, 30 August 1971, D. C.
Fowler, A. Schwartz, B. R. Sheplan; MCZ 125691, ASFS

1974 Anolis sheplani 5

V30824-26, 12.3 mi. (19.7 km) SE Cabral, 3300 feet (1007
meters), Barahona Province, Republica Dominicana, 9 Septem-
ber 1971, A. Schwartz, B. R. Sheplan.

Diagnosis. A species of the darlingtoni-occultus-insolitus group
of anoles, distinguished from all other species by the combina-
tion of: 1) small size (males to 41 mm, females to 40 mm
snout-vent length) and strong lateral compression; 2) modally
2 rows of loreal scales (modally 3 or 4 in other species) ;
3) supraorbital semicircles modally separated by 1 row of scales
(3 rows in occultus, 1 row in darlingtoni and insolitus) ; 4) su-
praocular semicircles separated from interparietal scale by 1
scale on each side (4 scales in occultus, 1 scale in darlingtoni
and insolitus) ; 5) modally 1 enlarged scale in supraorbital disk
(no enlarged scales in occultus, 2 in insolitus, 5 in darlingtoni) ;

6) rostral scale in contact posteriorly with 5 small scales (9
scales in occultus, 5 scales in insolitus, 6 scales in darlingtoni) ;

7) 4 distinct canthal scales (10 indistinct small canthal scales in
occultus, 4 distinct canthals in insolitus, 5 in darlingtoni) ; 8) su-
pralabials to center of eye 8 ( 10 in occultus, 7 in insolitus, 7 or 8
in darlingtoni) ; 9) 4—6 scales (mode 5) between second canthal
scales (9—14 in occultus, 2-6 in insolitus with a mode of 4, 5 in
darlingtoni) ; 10) a distinct supraciliary row of scales but no
scale enlarged (no differentiated supraciliaries in occultus) ;
11) no postorbital, supratemporal, or occipital spines (present
in insolitus); 12) no distinct supratemporal line of enlarged
scales (present and the series enlarged and terminating in a spine
in insolitus); 13) interparietal scale ovoid, much larger than
external auditory meatus (equal in occultus) ; 14) canthal ridge
strong (weak in occultus); 15) middorsal scales small, smooth,
subequal, with a longitudinal series of isolated spine-like scales
separated by about 6 to 8 small flat scales, no specialized spine-
like scales on neck (no modified middorsal scales in occultus;
nape scales slightly smaller than middorsals and no specialized
spine-Hke scales in darlingtoni; nape scales forming a low nuchal
crest as far posteriorly as about insertion of forelimbs, followed
by low rounded and isolated bosses, composed of about 8 small
rounded scales, the bosses separated by about 5 or 6 small dorsal
scales in insolitus) ; 16) ventral scales smooth and distinctly larger
than dorsal scales (about equal in darlingtoni), juxtaposed, in
often poorly defined transverse rows; 17) dewlap large, slotted
(= inset), in both sexes, pale peach in males, brown with a
cream border in females (pinkish gray in both sexes of occultus;
rich mustard, brown, orange or orange-ocher in both sexes of

6 BREVIORA No. 423

insoUtus; color unknown and dewlap not slotted in darlingtoni) ;
18) limb scales smooth, those on anterior face of thigh as large
as ventrals (smaller than ventrals in occultus, weakly carinate in
darlingtoni) \ 19) supradigital scales smooth (multicarinate in
darlingtoni) ; 20) tail round with a continuation of the e\enly
spaced middorsal spines, dorsal caudal scales larger than xentrals,
smooth to weakly unicarinate, ventral caudal scales much larger,
strongly unicarinate (no dorsal caudal scale modification in
occultus, dorsal scales very small, granular, ventral caudal scales
larger, smooth, and smaller than ventrals; dorsal caudal scales
modified into a series of irregularly spaced large triangular scales
in insolitus, dorsal and ventral caudal scales unicarinate and ven-
tral caudals larger than ventral scales) ; 21) a pair of enlarged
postanal scales in males (none in occultus) ; 22) general colora-
tion \'ery pale (almost white) but capable of pale tan to dark
brown phases, or lichenate blotching of these two colors with a
row of tiny dark brown dots down middorsal line, these dots
the enlarged median dorsal spinose scales; a small black to dark
brown nuchal dot and a broad dark sacral U in the pale phase:
two black radiating lines from the eye onto the temporal region
and a ventral radiating line from the eye which, \'entrally, forms
one of a maximum series of five incomplete transverse dark
brown to black lines crossing the throat, the most posterior at
the anterior end of the slotted dewlap; venter white.

Description of holotype. An adult male with the following
measurements and scale counts: snout-\Tnt length 40 mm, tail
length 43 mm; 4 canthal scales; 5 snout scales at level of second
canthal scales; 3 vertical rows of loreals; supraorbital semicircles
separated by 1 row of scales; 1 scale on each side between the
interparietal and the supraorbital semicircles; subocular scales
and supralabial scales in contact; 1 large scale in the supraocular
disk ; 2 postmental scales ; 6 small scales in contact with the ros-
tral scale posteriorly; 8 supralabials to center of e^e; 14 sub-
digital lamellae on phalanges II and III of fourth toe. Colora-
tion of holotype. When collected at night, very pale tan (almost
white), but capable of limited metachrosis to pale tan at one
extreme and dark l^rown at the other; often assuming a lichenate
blotched pattern of pale tan and dark brown, with a row of tiny
dark brown dots down the dorsal midline, these dots correspond-
ing to the indi\idual enlarged and spaced spinose middorsal
scales; in the pale phase, a black to dark brown nuchal dot and
a dark broad sacral U; tail banded red-brown and tan, the
red-brown bands narrow, fi\'e in number including the tail tip,

1974 Anolis sheplani 7

and separated by tan interband areas that are twice the width
of the dark bands; a pair of fine black lines radiating onto the
temples from the eye on each side, and a fine black line extend-
ing \'entralh from the eye across the supralabials onto the throat
where it forms the central of five incomplete dark crossbands
across the throat, the most posterior of which is at the angle of
the jaws; dewlap large, slotted, very pale peach, venter very pale
tan laterally, white centrally.

Variation. The series of A. sheplani consists of 16 specimens
of which one (MCZ 125691) has been skeletonized and upon
which no external counts or measurements were taken. Of the
remaining 15 lizards, nine are males and six are females. The
largest male has a snout-vent length of 41 mm (MCZ 125641)
and the largest female 40 mm (ASFS V30310). Both sexes thus
seem to reach about the same adult size; males are easily dis-
tinguished at any age by the presence of a pair of enlarged post-
anal scales. The series includes four young lizards with snout—
\'ent lengths between 20 mm and 25 mm. The canthal scales are
large and clearly delimited and always 4. There are between
4 and 6 scales across the snout at the level of the second canthals
( mode 5 ) . The loreal rows are either 2 or 3 ( mode 2 ) . The
supraocular semicircles are either in contact or separated by
1 or 2 rows of scales ( mode 1 ) . The scales between the inter-
parietal and the supraocular semicircles are almost always 1
bilaterally, although two specimens have 2 scales in this position
unilaterally. The subocular scales are always in contact with the
supralabial scales, of which there are between 7 and 10 (mode
8) to the center of the eye. There is modally only 1 enlarged
scale in the supraorbital disk, but three lizards have 2 scales
(the second enlarged but much smaller than its companion) in
the disk. The postmental scales vary between 2 and 5 (mode 4)
and there are 4 to 8 small scales (mode 5) in posterior contact
with the rostral scale. In further discussion of scutellar charac-
ter§, I follow the schema established by Williams and Rand
(1969) for this group of anoles.

Head: Narrow, elongate. Head scales large, smooth, smallest
anteriorly. Nostril circular, nasal scale separated from rostral by
3 small oval scales. Rostral scale wide, low, in contact with 4
to 8 small scales posteriorly.

Supraorbital semicircles large, weakly con\'ex, the scales
slightly boss-like, either in contact or separated by 1 or 2 rows of
smaller scales. A much less distinct row of many small oval
scales along the supraciliary margin on each side, no elongate

BREVIORA

No. 423

■F>4

o 2

2 W

«3

^ O

o -
o X

C CO

C '

^o .

1974

Anolis sheplani

!^:V:

■.«;■•■.■

'•^:;'s= ».•■■■■'•■■ .".■ • ■-i"?S'-^

u
&

to
.to

BREMORA

No. 423

o
X

•^

1974 Anolis sheplani 11

siipraciliary scale. Posterior and interior to the supraciliary row,
3 or 4 rows of small scales or granules of which the most interior
are largest, surrounding the single (occasionally two) enlarged
scale in the supraorbital disk. Canthal ridge of 4 scales well
defined, second canthal longest, diminishing in size anteriorly,
anteriormost posterior to nostril and separated from it by the
posterior portion of the nasal scale. Loreal rows 2 or 3, the scales
varying in shape between elongate rectangular and quadrangu-
lar. No distinct supratemporal Hne or row of scales. Temporal
scales small, flat, about 14 between the enlarged postocular
scales and the external auditory meatus. Supratemporal scales
flat and gradually larger than temporals, not forming a U-shaped
crest behind the interparietal region. Interparietal ovoid, very
much larger than tiny external auditory meatus, separated on
each side usually by 1 (occasionally 2) scale from the supra-
ocular semicircles. Scales surrounding interparietal flat, without
prominent tubercles or spines. External auditory meatus very
tiny, elliptical, placed far ventrally, just dorsal to the comissure
of the mouth.

Suboculars directly in contact with supralabials, anteriorh
grading into loreals, posteriorly continuous with the enlarged
postoculars. Seven to 10 supralabials to center of eye.

Mental large, semidivided, wider than deep, in contact with
2 to 5 small granular postmental scales; 1 infralabial and 1 sub-
labial in contact with mental on each side. Throat scales smooth,
elongate anteriorly, becoming more granular and ovoid posteri-
orly, gradually merging with the ventral scales.

Trunk: Dorsal scales small, smooth, slightly larger on flanks,
and merging with the \entral scales; a middorsal series of in-
di\idual spinose crest scales, separated by about 6 to 8 unmodi-
fied dorsal scales, this middorsal series of spinose scales continued
onto the dorsal caudal midline. Ventrals larger than dorsals,
smooth, rounded, and in transverse rows that may be slightly
irregular.

Dewlap: Large; present in both sexes, slotted (= inset), pale
peach in males, brown with a cream border in females, scales
lar2:e and arranged in rows, larger than throat scales and about
the same size as ventrals; marginal dewlap scales crowded and
about the same size as throat scales adjacent to dewlap.

Limbs and digits: Limbs short, tibial length about equal to
distance from tip of snout to center of eye. Thirteen to 17 lamel-
lae under phalanges II and III of fourth toe. Scales of limbs

12 BREViORA No. 423

smooth, those of anterior surface of thigh sUghtly smaller than
ventrals. Supradigital scales smooth.

Tail: Round non-verticillate, with a median series of widely
spaced spinose scales, their apices directed posteriorly, separated
from each other bv about 3 to 5 smaller, smooth to weaklv uni-
carinate dorsal caudal scales. A pair of enlarged postanal scales
in males. Scales behind vent and around base of tail smooth.
Four to 6 ventral rows of much enlarged unicarinate caudal
scales.

Color in life: The coloration and pattern of A. sheplani have
been given both in the diagnosis of the species and in the descrip-
tion of the holotype and need not be repeated in detail. The
lizards are capable of limited metachrosis (they have no green
phase) between very pale tan (almost white) while sleeping and
brown when disturbed or active. In the pale phase there is a
brown nuchal dot, a broad dark sacral U, and a median dorsal
series of dark brown to black dots. An intermediate pigmental
condition involves a lichenate tan-and-brown phase. The dew-
lap is pale peach in males, dark brown with a cream border in
females; although the dewlap is well developed in both sexes, it
is slightly larger in males than in females.

Comparisons. The diagnosis gives details of comparisons be-
tween sheplani and the three remaining species of the group
{darlingtoni, occultus, insolitus) , and these need not be repeated.
However, there are some salient differences that I wish to em-
phasize. Of the four species, sheplani most closely resembles
occultus in snout-vent length; females of both species reach a
snout-vent length of 40 mm, whereas the largest male occultus
(ASFS V5489) I have examined has a snout-length of only
35 mm; Williams and Rand (1969: 13) noted maximally sized
occultus at 34 mm snout-vent length (sex not stated), but
Williams and Rivero ( 1 965 : 7 ) gave 42 mm as the size of the
largest occultus (sex not stated) examined by them. A. sheplani
is smaller than A. insolitus (maximally sized male 47 mm —
ASFS V22502; female 44 mm — ASFS V31614), and much
smaller than A. darlingtoni (holotype male, 72 mm). Of the
four species, only occultus males lack enlarged postanal scales.
The spinose or tuberculate head scales, and the supratemporal
line of enlarged scales which terminates in a spine, are absent in
sheplani, as well as occultus and darlingtoni; these features are
distinctive of insolitus. Scales between the second canthals are
very numerous in occultus (9-14) and very many less in the
other species, with insolitus hciving 2-6 (mode 4) and sheplani

1974 Anolis sheplani 13

4-6 (mode 5). A. darlingtoni has 5 scales between the second
canthals. Loreal rows are modally 2 in sheplani, 3 in darlingtoni
and in insolitus, and 4 in occultus. The supraorbital semicircles
are modally separated by 3 scales (2-5) in occultus, by 1 row
of large scales in insolitus, by 1 row of small scales (0-2) in
sheplafii, and by 1 row of small scales in darlingtoni. Scales
between the interparietal and the supraorbital semicircles are
modally bilaterally 4 in occultus (range 2-6), and 1 scale in the
other species (range 0-2 in insolitus, 1-2 in sheplani, 1 in dar-
lingtoni). The supraocular disks in occultus have no enlarged
scales, whereas in sheplani there is 1 (occasionally 2) enlarged
scale in this area, in insolitus 1 to 6 ( mode 2 ) , and 5 in darling-
toni. Scales posteriorly in contact with the rostral are 6-10 in
occultus (mode 9), 4-7 in insolitus (mode 5), 4-8 in sheplani
(mode 5), and 6 in darlingtoni. The canthal scales are poorly
defined and very numerous (7-12; mode 10) in occultus,
whereas all sheplani have 4 distinct canthals, insolitus modally
has 4 distinct canthals (range 3-6), and darlingtoni has 5.
There are 9-11 supralabials to the eye center in occultus (mode
10), 6-8 (mode 7) in insolitus, 7-10 (mode 8) in sheplani, and
7 or 8 in darlingtoni.

The dewlap color in occultus is pinkish gray, whereas that of
insolitus varies between rich mustard, brown, orange or orange-
ocher; in neither of these species is the dewlap color sexually
dichromatic, whereas the dewlap is strongly sexually dichro-
matic in sheplani.

Thomas (1965: 15-16) gave a resume of the color repertory
of occultus; the pattern of this species consists of a dark cephalic
figure or interocular trangle; dark radiating eye lines; four zones
of transverse body banding (scapular, dorsal, lumbar, sacral) ;
a single or paired lumbar spot; and a fine reticulum of dark lines
which frequently appears as small ocelli. The ground color of
occultus varies through shades of gray through olive-brown,
olive, yellow-green to dirty orange, to a lichenate off-white or
very light gray and black or very dark gray. In insolitus, the
dorsum is grayish green or grayish brown, irregularly marbled,
with a distinctive pale green supra-axillary crescent, a white
subocular spot, and a black postorbital spot. In life, the supra-
axillary crescent is extremely clear, and it, plus the black post-
orbital spot, are ready recognition features of the species. At
night while asleep, insolitus may often be a very pale tan or
white, very much in the fashion of sheplani. The coloration of
darlingtoni in life is unknown, but Williams and Rand (1969:

14 BREVIORA No. 423

1 1 ) have an excellent figure showing the basic design of the
holotype. Conspicuous details of the pattern are a large dark
postocular blotch and a generally transversely banded (about
five fragmented bands) dorsal pattern.

One structural feature is interesting. A. occultus has the
median dorsal scales unmodified into any sort of spines or crest
scales. In sheplani, there are isolated spinose scales along the
dorsal midline, the scales separated widely by small dorsal scales.
In insolitus, there are low raised bosses that are coxered by
"rosettes" of scales, slightly larger than their surrounding scales,
the bosses separated by unspecialized dorsal scales. These raised
"bosses" with the rosettes of scales become slightly less conspicu-
ous posteriorly, and on the tail are replaced by laterally com-
pressed and spaced individual triangular scales as part of the
same dorsal series. A. darlingtoni lacks specialized middorsal
scales.

Field observations. All specimens of A. sheplani were taken
in a very circumscribed area between 3200 and 3300 feet (976
and 1007 meters) in the Sierra de Baoruco. The immediate area
where the lizards were secured is high mesic deciduous forest,
somewhat modified by the cultivation of coffee and cacao. The
high original forest trees have been retained as shade cover for
the cultivated plants. The general aspect is rich, wet, and very
well wooded. A newly constructed highway ascends the north-
ern slope of the Sierra de Baoruco between Cabral in the Valle
de Neiba and the settlements of Las Auyamas and Polo in the
Baoruco uplands. At a distance of 10.4 miles (16.6 km) south
of Cabral, an unpaved but quite good road takes ofT to the
southeast of the main highway and terminates abruptly at the
settlement of La Lanza. The road apparently formerly went
from La Lanza to the coastal town of Paraiso, but this section
is no longer passable. At a distance of between 1.9 and 2.6 miles
(3.0 and 4.2 km) from the intersection, the road has been cut
into a gradually sloping mountain side. Below the road there
are high-canopied cajetales and cacaotales; abo\'e the road, and
separated from it by a road-cut bank that varies from 2 to 10
feet (0.6 to 3.1 meters) in height, is an area of second-growth
trees, saplings, shrubs, and weed and grass patches, the arbores-
cent vegetation heavily interlaced with li\'ing and dead \ines,
primarily those of a purple-flowered member of the Con\ol\ula-
ceae. In many places along this limited stretch of road, there
are dense mats and curtains of vines; it was within and under
these mats that A. sheplani was encountered. The species is far

1974 Anolis sheplani 15

outnumbered by Anolis hendersoni Cochran, which sleeps in pre-
cisely the same situations, and one Anolis singularis Williams
was also found sleeping syntopically with A. sheplani.

Sleeping sites of A. sheplani are bare twigs and vines within
and beneath the curtains and mats of vines. The lizards sleep
exposed and are easily seen since they are very pale. They are
not easily disturbed by movement of the collector, jostling of the
\ines, or flashlight. On those rare occasions when an individual
was disturbed, it opened its eyes, clutched the twig or vine more
tightly, and, if pressed, moved unhurriedly away from the source
of disturbance. We never saw A. sheplani either scurry away or
drop to the ground in the fashion of other anoles when disturbed
at night. Rather, their reaction to complete disturbance (for
instance, touching the lizard or breaking the twig or vine to
collect it) only caused the lizard to cling more tightly to its sub-
strate. The lowest Hzard was taken at a height of 3 feet, the
highest 14 feet, above the ground; this gives a sleeping range of
3 to 14 feet (0.9 to 4.3 meters). It is probable that A. sheplani
sleeps even higher on vines in the canopy, but at this location
the trees in general are fairly low (perhaps 20 feet — 6.1 meters
— average height) and thus the vines are low. It is significant
that we never encountered A. sheplani below the road in this
same area, despite suitable vine mats and curtains; on the lower
side of the road the forest is much less disturbed and the canopy
is much higher. In neighboring situations, even within a few
meters, A. cy botes, A. coelestinus and A. distichus were also
found sleeping.

It is instructive to compare the sleeping sites and general be-
havior of A. sheplani with that of A. insolitus and A. occultus.
I have the impression that insolitus is an inhabitant of much less
disturbed situations than sheplani. The known localities for in-
solitus, which now number seven, are invariably gallery forest
along rixers or streams. At some localities for insolitus, the forest
has been slightly disturbed by planting of coffee and cacao, but
in 'general the canopy is high and dense, and vines and lianas
are abundant and conspicuous (but often quite high). Conse-
quently, sleeping sites of insolitus are not restricted to sheltered
spots beneath vine mats or curtains. Regularly, specimens of
insolitus have been taken completely exposed on the tips of twigs,
vines, and branchlets, at heights above the ground between Z
and 25 feet (0.6 and 7.6 meters). On occasion, A. insolitus
ha\e been taken sleeping on green leafy shrubs rather than on
bare twigs and vines. At the type locality, however, during a

16 BREvioRA No. 423

verv heavy and continuous rain, most insolitus were secured in
sheltered situations under vine mats or curtains, and two indi-
viduals were found sleeping on top of each other on a pendant
vine. In summary, the sleeping sites of A. insolitus are regularly
much more exposed than are those of A. sheplani.

Thomas (1965) and Webster (1969) have both commented
upon the habits of A. occultus in Puerto Rico. Northeast of
Guayama, Thomas reported occultus "sleeping at night in tangles
of dead (or leafless) vines and twigs along both sides of the
path, four to ten feet abo\'e the ground" on a forested hillside,
and north of Sabana Grande Thomas recorded this species sleep-
ing at heights of 4 to 15 feet (1.2 to 4.6 meters) on dead vines.
Finally, south-southeast of Villa Perez, A. insolitus was en-
countered asleep in the same sorts of situations 5 to 1 2 feet ( 1 .5
to 3.7 meters) above the ground. Webster reported sleeping
sites of seven A. occultus at a locality south of Palmer as "long,
exposed twigs, . . . twigs near leaves, . . . and the upper surface
of a broad, stiff leaf." Webster also located six additional A.
occultus sleeping on H\ing twigs near leaves, one on a long dead
twig, and at the tip of a very long descending branch, and a
juvenile on a dead fern. Both Thomas and Webster commented
on the habit of occultus of clinging tightly to twigs when dis-
turbed; this habit is shared with A. sheplani as noted above.
The same is true of A. insolitus; on one occasion, we cut from
the tree the small branch upon which an insolitus slept, and the
lizard remained clinging to the branchlet during the entire
operation. On another occasion, a pendant vine upon which an
insolitus slept was dehberately broken above and below the lizard
and then accidentally dropped onto the ground in leaf litter and
herbaceous growth. When the vine was located, the now wide-
awake insolitus was seen to be still clinging tightly to the vine!

Remarks. I have little doubt that A. sheplani is more closely
related to A. insolitus than to A. darlingtoni, despite the fact
that the latter species occurs on the south island along with
sheplani (although the sole darlingtoni locality is removed some
310 kilometers to the west of those for sheplani). It is truly
puzzling, considering the intensive (albeit local) collecting ac-
tivity on the Hispaniolan south island in Haiti, most especially
in the mountains above Port-au-Prince ( Montague Noire, Mome
I'Hopital) and in the Massif de la Hotte (Les Platons, Castillon)
that no further specimens of A. darlingtoni have been encoun-
tered. I suspect that the habits of this species will be found to
be very like those of the remaining members of the complex; if

1974 Anolis sheplani 17

so, then nocturnal collecting with emphasis on dead vines,
branches, twigs, etc., in sheltered locales may well be the secret
of securing more A. darlingtoni. Considering the apparently
very narrow ecological situations that A. sheplani favors, and
the fact that the uplands of the Sierra de Baoruco in the Las
Auyamas-Polo region have presumably been well collected since
the 1920's, there is always the possibility that A. darlingtoni has
equally stringent ecological requirements that have been over-
looked or that may be very restricted in the Massif de la Hotte.
Likewise, I have little doubt that A. sheplani will be encountered
elsewhere in the Sierra de Baoruco and (or a related form) in
the Massif de la Selle and its associated ranges.

The knowledge that the darlingtoni group of anoles occurs on
both the north and south Hispaniolan islands should spur interest
in ascertaining the presence of similar species of this small group
in other Hispaniolan ranges. Most pertinent is the Sierra de
Neiba, that range which borders the Valle de Neiba on its north-
ern side, just as the Sierra de Baoruco borders the low-lying
\'alley on its southern side. If insolitus and sheplani are more
closely related to each other than either is to darlingtoni, it would
seem likely that some member of this group of anoles occurs in
the uplands of the intervening Sierra de Neiba. On this premise,
we visited that range both during the day and at night during
1971, but to no avail. The forests are mesic and viney, alto-
gether suitable situations for members of this group of lizards.
The canopy is generally high, however, and this may make it
more difficult to secure related anoles if they occur in this range.
However, in similar high-canopied forests south of El Rio in the
Cordillera Central, A. insolitus was easily observed. It may well
be that there is no member of the darlingtoni group in the Sierra
de Neiba, but this range is so poorly known herpetologically that
one cannot with certainty dismiss the absence of a related species
there.

The elevational distributions of the four members of the dar-
lingtoni complex are interesting. A. occultus in Puerto Rico is
known to occur between elevations of 2300 and about 4389 feet
(702 and 1338 meters), whereas the known altitudinal ranges
of the other species are: darlingtoni, 5000 feet (1525 meters);
sheplani, 3200-3300 feet (976-1007 meters) ; and insolitus,
3500-5800 feet (1068-1769 meters). Although the data on
darlingtoni and sheplani are limited, insolitus seems to reach
higher elevations in the Cordillera Central than any species does
elsewhere. This may at least in part be due to the fact that no

18 BREVIORA No. 423

mountains in Puerto Rico or the Sierra de Baoruco reach such
high ele\ations as do the mountains within the area known to
be inhabited by insolitus.

WilHams and Rand (1969: 9) noted that "It would be a pos-
sible 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 between closely
related species, particularly if they are sympatric, are not un-
usual, but as far as known, these two species are widely allo-
patric, and the size difference is extreme." More recently col-
lected and larger numbers of A. insolitus show that the supposed
extreme difference in size (= snout-\'ent length) between dar-
lingtoni and insolitus is not so striking as Williams and Rand
supposed. In fact, insolitus, which reaches a maximum known
snout-vent length of 47 mm (not 33 mm) but which is none-
theless still smaller than darlingtoni, rather bridges the size gap
between smaller occultus and sheplani and larger darlingtoni.
The size discrepancy for members of the complex, which Wil-
Uams and Rand felt might argue against relationships among
these lizards, is not so striking as they supposed.

Specimens examined. Anolis occultus: PUERTO RICO,
20.9 km NNE Guavama, 2300 feet (702 meters) (ASFS
V4891-92, V4901, V5017-18); 13.7 km N Sabana Grande,
2800 feet (854 meters) (ASFS V5489-91, V5494) ; 13.7 km
S Palmer (ASFS V6662-65); 10.6 km SSE Villa Perez, 3400
feet (1037 meters) (ASFS V6196-97) .

Anolis insolitus: REPUBLICA DOMINICANA, La Vega
Province, La Palma, 14 km E El Rio, 3500 feet (1068 meters)
(ASFS V18739, V18947-19, V22546-53, V31705-10) ; 1.9 mi.
(3.0 km) SW El Rio, 3900 feet (1190 meters) (ASFS V31656-
63); 16 km SE Constanza, 5250 feet (1601 meters) (ASFS
V22502-05); 16.4 km SE Constanza, 5500 feet (1678 meters)
(ASFS V31614); 18 km SE Constanza, 5800 feet (1769
meters) (ASFS V19096); 18.5 km SE Constanza, 5800 feet
(1769 meters) (ASFS V31581-82). Peravia Province, 6.b mi.
(10.4 km) NW La Horma, 5400 feet (1647 meters) (ASFS
V31933-37, V31973-74); 8.1 mi. (13.0 km) NW La Horma,
5800 feet (1769 meters) (ASFS V31927-28).

Anolis darlingtoni: HAITI, Dcpt. du Sud, Roche Croix,
Massif de la Hotte, ca. 5000 feet (1525 meters) (MCZ 38251).

1974 Anolis sheplani 19

Literature Cited

Cochran, D. M. 1935. New reptiles and amphibians collected in Haiti by
P. J. Darlington. Proc. Boston Soc. Nat. Hist., 40 (6) : 367-376.

Etheridge, R. E. 1960. The relationships of the anoles (Reptilia: Sauria:
Iguanidae) ; an interpretation based on skeletal morphology. Univ.
Microfilms, Inc., Ann Arbor, xiii + 236 pp. 11 figs., 10 maps.

Schwartz, A. 1973. A new species of montane Anolis (Sauria, Iguani-
dae) from Hispaniola. Ann. Carnegie Mus., 44 (12) : 183-195, 3 figs.

Thomas, R. 1965. A new anole (Sauria, Iguanidae) from Puerto Rico.
Part II. Field observations on Anolis occuUus Williams and Rivero.
Breviora, Mus. Comp. Zool., No. 231: 10-16. 2 figs.

Webster, T. P. 1969. Ecological observations on Anolis occultus Williams^
and Rivero (Sauria, Iguanidae), Breviora, Mus. Comp. Zool., No. 312:
1-5.

Williams, E. E., and J. A. Rivero. 1965. A new anole (Sauria, Iguanidae)
from Puerto Rico. Part I. Description. Breviora, Mus. Comp. Zool.,
No. 231: 1-9, 18. 5 figs.

, AND A, S. Rand. 1969. Anolis insolitus, a new dwarf anole

of zoogeographic importance from the mountains of the Dominican
Republic. Breviora, Mus. Comp, Zool., No. 326: 1-21. 6 figs.

JUL 8 W74

B R E V fcvftR A

Miiseiiin of Comparative Zoology

us ISSN 0006-9698

Cambridge, Mass. 28 June 1974 Number 424

THE LARVA OF SPHINDOCIS DENTICOLLIS FALL

AND A NEW SUBFAMILY OF GIIDAE

(GOLEOPTERA: HETEROMERA)

John F. Lawrence^

Abstract. The larva of Sphindocis denticollis Fall is described, and its
biology is briefly discussed. A new subfamily of Ciidae — the Sphindociinae
— is proposed for Sphindocis and is formally characterized, while the Ciidae
and Ciinae are redefined. Speculations are made concerning the phylogenetic
relationships of the family Ciidae.

The monotypic genus Sphindocis Fall is based on a very
interesting fungus-feeding beetle {S. denticollis) that is known
only from the Transition Zone forests of the northern California
coast. The genus was originally placed in the family Ciidae
(Fall, 1917), but it was recently removed from that family and
tentatively placed in the Tetratomidae (Lawrence, 1971). At
the suggestion of R. A. Crowson, I made a more detailed study
of the Sphindocis larva, comparing it and the adult with various
Ciidae, Tetratomidae, Pterogeniidae, and related Heteromera.
As as result, I have come to the conclusion that Sphindocis
represents the closest living relative or sister group of the Ciidae
and should either be returned to that family or form the basis
for a new group of equal rank. The former alternative appears
more reasonable, since the number of families in the Heteromera
is already excessive. The following treatment includes a descrip-
tion of the Sphindocis larva, the proposal of a new subfamily
for the inclusion of this genus, and a recharacterization of the
family Ciidae and subfamily Ciinae.

The larval description is based on more than 50 specimens
collected with adults in the fruiting bodies of Trametes sepium

^Museum of Comparative Zoology, Cambridge, Mass. 02138.

2 BREVIORA No. 424

Berkeley growing on dead branches of madrone {Arbutus Men-
ziesii) at the following localities in California: Alpine Lake,
Marin County; 1 mi. N Piercy, 2 mi. N Piercy, 3 mi. S Leggett,
and 4 mi. W Leggett, Mendocino County. Another eight speci-
mens were collected without adults in a fruiting body of Poria
cinerascens (Bresadola) Saccardo and Sydow growing on a
Douglas fir {Pseudotsuga Menziesii) log at Alpine Lake. A
single pupa was dug out of madrone wood beneath a fruiting
body, which may indicate that the beetles require the woody
substrate for pupation.

Most of the terms used in the larval description are those
found in standard works, such as Boving and Craighead (1931)
and van Emden (1942). For the three labial sclerites, I have
used the terms prementum, mentum, and submentum, although
Anderson (1936) has indicated that these are not homologous
in all groups. Terminology for the ventral thoracic sclerotizations
follows Watt (1970), while various other terms have been taken
from Crowson (1955), Glen (1950), Rozen (1958, 1960), St.
George (1924), and Snodgrass (1935).

I wish to thank H. B. Leech and the California Academy of
Sciences, San Francisco, for the loan of specimens; J. T. Doyen
for collecting adults and larvae of Sphindocis; and R. A. Crow-
son and E. Mayr for their encouragement.

DESCRIPTION OF THE MATURE LARVA OF
Sphindocis denticollis Fall

Body elongate and subcylindrical, lightly sclerotized except
for head, anterior part of prothoracic tergum, and pygidium
(upper part of ninth abdominal tergum). Length about 5 mm;
width about 0.7 mm.

Head (Figs. 1-3) exserted, obliquely prognathous, subglob-
ular, strongly convex dorsally, except for a broad, shallow
concavity (c) extending from the middle of the epicranial stem
to the upper part of the frontoclypeal triangle (fc) ; heavily
sclerotized and yellowish brown in color, with fairly coarse
and irregular punctation; vestiture consisting of numerous short
setae and several longer ones, the origins of which are shown
in Figures 1-3. Epicranial stem (es) about half as long as head
width; frontal arms (fa) somewhat V-shaped and extending to
antennal ridges (ar), which conceal antennal insertions; endo-
carina absent. Frontoclypeal area (fc) bearing two parallel,
transverse sulci (ts) near epistomal margin (em). Epicranial

1974

LARVA OF SPHINDOCIS

4\^

Plate 1

Figures 1-10. Sphindocis denticollis Fall, larva (1 line = 0.125 mm for
1-3, 9; 0.063 mm for 4, 7, 8, 10; 0.025 mm for 5, 6) . Fig. 1. Head capsule,
dorsal view, mandibles and ventral mouthparts removed (dots = setal ori-
gins) . Fig. 2. Head capsule, ventral view, right mandible and ventral
mouthparts removed. Fig. 3. Head capsule, lateral view. Fig. 4. Labrum-
epipharynx, dorsal view. Fig. 5. Epipharynx, median portion. Fig. 6. Left
antenna, lateral view. Fig. 7. Right mandible, dorsal view. Fig. 8. Left
mandible, ventral view. Fig, 9. Ventral mouthparts and gular region, ven-
tral view. Fig. 10. Apex of left maxilla, dorsolateral view.

4 BREVIORA No. 424

halves (eh) each bearing a ventral ridge (vr) which extends
posterad from mandibular articulation, parallel to the hypostomal
ridge (hr), and forms with the latter a support for the ventral
mouthparts, which are large and protracted. Ocelli (oc) 5 in
number, arranged as in Figure 3. Anteclypeus (ac) a short,
lightly sclerotized band. Labrum (lb) transversely oval, with
setae and spines as in Figure 4; epipharynx with 4 median
groups of very short setae or sensillae ( Fig. 5 ) ; tormae ( to )
symmetrical, joined posteriorly by a narrow bridge. Antenna
(Fig. 6) fairly short, less than 1/10 as long as head width,
3 -segmented, segments about equal in length, II slightly nar-
rower than I and bearing a sensory appendix (sa) that is longer
than III and ventral to it, III about half as wide as II and
bearing a terminal seta almost five times its length; antennal
insertion separated from the mouth cavity by a narrow bar.
Gula (gu) not well defined; gular sutures absent and no suture
between gula and submentum (sm). Posterior tentorial pits (pt)
and tentorium (tn) as in Figure 2.

Mandibles (Figs. 7 and 8) symmetrical, large and wedge-
shaped, with two apical teeth of unequal lengths, an obtuse tooth
on the cutting edge, and a lightly sclerotized retinaculum (rt) ;
mola absent; a seta located on the dorsal surface near the middle
of the lateral edge. Maxillae (Fig. 9) free almost to base of
mentum; mala (ma) obHquely rounded, its apex armed with 5
stout spines and several finer setae; inner edge of mala (Fig. 10)
bearing a dorsal laciniar lobe (la), located at the level of the
palpifer (pf) and bearing 2 stout apical spines and several long
setae at base; stipes (st) elongate; cardo (ca) subtriangular ; a
large, articulating sclerite (as) between stipes and submentum;
palp 3 -segmented. Labium with a short prementum, a sub-
quadrate mentum (rne), and a submentum (sm), which is
raised above the gula but is not separated from it; ligula (li)
short and rounded, bearing 4 setae at apex; palp 2-segmented.
Hypopharynx (hy) subquadrate, without a sclerome; hypo-
pharyngeal bracon (hb) lightly sclerotized except at base of
hypopharynx.

Prothorax ( Fig. 1 1 ) slightly longer than meso- or metathorax,
its tergum (prt) well developed and extending onto lateral sur-
faces, heavily pigmented anteriorly, becoming very lightly pig-
mented posteriorly, with a median ecdysial suture; vestiture
consisting of numerous short setae and 3 transverse rows of
setae consisting of 12 (anterior edge), 8 (anterior third), and
10 (posterior third) setae; sternum consisting of a large, tri-

1974

LARVA OF SPHINDOCIS

pap i ^

"^'-^

Plate 2

Figures 11-14. Sphindocis denticollis Fall, larva (1 line = 0.063 mm for
12; 0.250 mm for 11, 13, 14). Fig. 11. Prothorax and mesothorax, ventral
view, legs removed. Fig. 12. Prothoracic leg, coxa and part of trochanter
not sTiown. Fig. 13. Apex of abdomen, lateral view. Fig. 14. Apex of
abdomen, ventral view. Figures 15-21. Sphindocis denticollis Fall, adult
male (1 line = 0.063 mm for 15, 19; 0.250 mm for 16, 17, 20; 0.125 mm
for 18, 21) . Fig. 15. Antennal club. Fig. 16. Prothorax, ventral view, right
coxa removed. Fig. 17. Meso- and metathorax, ventral view, left mesocoxa
and metacoxae removed. Fig. 18. Metendosternite, dorsal view. Fig. 19.
Apex of protibia. Fig. 20. Abdomen, ventral view. Fig. 21. Aedeagus,
ventral view.

6 BREVIORA No. 424

angular cervicosternum (cv), a triangular basisternum (bs), and
vaguely defined sternellum, epistemum, and epimeron; coxal
cavities (cc) large and obliquely oval, separated by a little less
than y^ their greatest diameter. Mesothoracic tergum less ex-
tensive than that of prothorax and lightly pigmented except for
a transverse carina (tc) at the anterior fifth; several long setae
scattered on shield; each side w^ith two laterotergites (It), the
anterior of which bears a biforous spiracle (sp) with the air
tubes facing dorsad; sternal areas not well defined, coxae slightly
smaller and broader than those of prothorax. Metathorax simi-
lar in structure, but with no spiracle on the anterior laterotergite.
Legs about equal in size, with a large conical coxa, triangular
trochanter (Fig. 12, tr), the femur (fe) and tibiotarsus (ti)
about equal in length, and the claw (cl) bearing two setae.

Abdominal segments 1 to 8 slightly convex dorsally and
strongly so ventrally; tergal shields lightly pigmented, each with
an anterior carina and several long setae; each side with a single
laterotergite (Fig. 13, It), just above which is the spiracle with
the air tubes facing posterad. Ninth abdominal segment (Figs.
13 and 14) longer than those preceding it, with a large tergum
bearing a heavily pigmented, circular, concave, declivous py-
gidium (py), lined along 34 of its circumference with saw-like
teeth; ninth sternum reduced in size, bearing at its apex a row
of anteriorly projecting asperites (asp) ; tenth tergum lunate,
partly separating ninth tergum and sternum, bearing 3 papillae
(pap) at its apex; tenth sternum reduced and pygopod-like,
bearing 5 papillae in front of anal opening (Fig. 14).

This larva differs from that of any other known ciid in lacking
an endocarina and having 3-segmented antennae, a maxillary
articulating sclerite, biforous spiracles, and subanal asperites on
the ninth sternum. The presence of an endocarina has never
been noted for the Ciidae, probably because it is directly beneath
the epicranial stem and does not extend anterad of the frontal
arms, as it does in various other Heteromera. The epicranial
stem in Sphindocis is an ecdysial line, whereas in other Ciidae
it coincides with an internal ridge. The reduced antennal seg-
mentation in most ciid larvae represents a fusion of the last two
segments. Symmetrical mandibles also occur in other Ciidae, but
asymmetry appears to be the more common condition. Biforous
spiracles appear to be unique to Sphindocis, but a peculiar type
of accessory air tube has been obserxed in at least one other ciid
funpubhshed). The concave pygidium of Sphindocis, which
occurs in other Ciidae, such as Cis melliei (Coquerel, 1849),

1974 LARVA OF SPHINDOGIS 7

in the tenebrionid genus Meracantha (Hyslop, 1915), and in
\arious other substrate-dwelling beetle larvae, represents a type
of defensive adaptation, which Wheeler (1928) termed phrag-
mosis. The fruiting body of Trametes sepium is often resupinate
with a fairly thin context, and the concave and heavily sclero-
tized pygidium in Sphindocis serves to block the shallow larval
tunnel against predators or parasites.

CHARACTERIZATION OF THE FAMILY CIIDAE

AND ITS SUBFAMILIES

CiiDAE Leach

AV^ith the general characters of the Polyphaga: Cucujoidea.

Adult. Form variable, usually oval to elongate, convex. Size
0.5-6.0 mm. Head globular, without neck, often strongly de-
clined, partly concealed by pronotum, without stridulatory files.
Eye oval, entire, fairly coarsely facetted. Frontoclypeal area
with a distinct suture, often raised in males to form a ridge,
horns, or tubercles. Antennal insertion concealed from above
by frons. Antenna 8- to 1 1 -segmented, with a 2- or 3-segmented
club, club segments often with multi-pronged sensillae (absent
in Sphindocis). Mandible bidentate, with a simple cutting edge
and a quadrangular mola without ridges or tubercles. Maxilla
with an articulated lacinia and 2-segmented galea {Sphindocis)
or a fixed lacinia and 1 -segmented galea (Ciinae), palp 4-seg-
mented, the terminal segment not securiform. Labium with
ligula reduced, palp 3-segmented. Pronotum margined laterally
and posteriorly, anterior edge usually produced forward, some-
times bearing horns in male. Prosternum variable, long or short,
concave to carinate, coxae globose or transverse, sometimes pro-
jecting, contiguous to broadly separated, without internalized
lateral extensions, trochantin usually concealed; procoxal cavities
open internally, narrowly open or closed externally (posteriorly).
Elytra not striate, humeri tuberculate, epipleura very narrow,
extending almost to apex. Scutellum small and subtriangular,
sometimes absent. Wing venation often reduced, subcubital fleck
present, anal region with four veins {Sphindocis) or only one
(Ciinae). Mesosternum transverse, sometimes extremely re-
duced, coxae globose and narrowly separated, coxal cavities not
closed outwardly by sterna, trochantins exposed or not. Meta-
sternum subquadrate, with or without median suture, without
coxal lines, coxae narrow, transverse, subcontiguous. Metendo-
sternite with a long median stalk {Sphindocis) or none (Ciinae),

8 BREVIORA No. 424

anterior tendons arising near the apices of the lateral arms.
Tarsal formula in both sexes 4-4-4 (occasionally 3—3—3), tarsi
simple, the first three segments small and subequal, terminal
segment elongate, claws simple. Trochanters oblique, completely
(Ciinae) or only partly {Sphindocis) separating coxa from
femur. Tibial spurs usually absent; 2 reduced spurs in Sphin-
docis. Outer edge of protibia often expanded and modified at
apex. Abdomen with 5 visible stemites, the first 2 (III and IV)
connate {Sphindocis) or not (Ciinae). First visible sternite
(III) without coxal lines, often with a median pubescent fovea
in male. Aedeagus of inverted heteromeroid type, with ventral
tegmen and dorsal median lobe.

Larva. Body elongate and subcylindrical, lightly sclerotized,
except at anterior and posterior ends. Head subglobular, ob-
liquely prognathous, with well-developed epicranial stem and
Y-shaped frontal arms, endocarina present (Ciinae) or not
[Sphindocis) ; ventral epicranial ridge present behind mandib-
ular articulation. Ocelli usually 5, occasionally fewer or none.
Antennal insertion concealed from above and separated from
mouth cavity by a narrow bar. Antenna short, 2- or 3 -seg-
mented, with a long sensory appendix on segment II and a very
long terminal seta. Gular area short, sutures present or absent.
Mandibles large and wedge-shaped, usually somewhat asym-
metrical, with 2 apical teeth, a simple cutting edge, often with
a lightly sclerotized retinaculum, mola usually absent. Maxilla
free at least to middle of mentum, with a narrow articulating
membrane (Ciinae) or a large articulating sclerite {Sphindocis)
between stipes and submentum; mala obliquely rounded, inner
edge with a dorsal laciniar lobe; palp 3-segmented. Labium with
short prementum, subquadrate mentum, and elongate submen-
tum, the last separated from gula by suture or not; ligula short
and rounded, with 2 or 4 setae; palp 2-segmented. Hypo-
pharynx without sclerome. Thoracic terga well developed and
extending onto sides; prothorax slightly larger than meso- or
metathorax; prosternum with a large triangular cervicoster-
num; procoxae large and fairly close together; spiracle annular
(Ciinae) or biforous {Sphindocis), located on anterior latero-
tergite of mesothorax. Legs fairly short and broad, subequal;
claw with 2 setae. Abdominal spiracles located above latero-
tergites. Ninth tergum variously modified, usually heavily sclero-
tized and with urogomphi; tenth sternum reduced and pygopod-
like; anal opening surrounded by several papillae.

1974 LARVA OF SPHINDOCIS 9

Sphindociinae, New Subfamily

Adult. Antenna 11 -segmented, with 3 -segmented club (Fig.
15); club segments without multi-pronged sensillae. Maxilla
with an articulated lacinia and a 2-segmented galea. Pronotum
(Fig. 16) with lateral margins broadly crenulate, so that several
round teeth are formed; procoxal cavities with a slight lateral
extension, which may expose part of trochantin. Mesocoxal
cavities ( Fig. 1 7 ) with exposed trochantins ( t ) . Metendosternite
(Fig. 18) with a long stalk (s), a narrow lamina (1), and the
anterior tendons (at) near the apices of lateral arms. Hindwing
with well-developed anal region, bearing 4 veins and a wedge
cell. Trochanter (Fig. 17, tr) of heteromeroid type, obliquely
joined to femur so that the latter is in direct contact with coxa
at one point. Tibial apices (Fig. 19) simple, with 2 reduced
spurs. Abdominal sternites III and IV connate (Fig. 20), III
with a median pubescent fovea in male. Aedeagus (Fig. 21)
with a large basal piece (bp), with two apical condyles (cd),
a well-sclerotized ventral paramere (pm) with 2 pairs of setae
near its base, and a membranous median lobe with 2 lateral
struts (Is).

Larva. Head without endocarina, with 5 ocelli. Antenna 3-
segmented. Mandibles symmetrical, without mola and with
lightly sclerotized retinaculum. Maxilla free almost to base of
mentum, with a large articulating sclerite between stipes and
submentum. Spiracle biforous. Ninth tergum bearing a concave
pygidium surrounded by saw-like teeth; ninth sternum bearing
a row of asperites.

CiiNAE Leach

Adult. Antenna 8- to 10-segmented, with a 2- or 3 -segmented
club; club segments with at least 4 multi-pronged sensillae.
Maxilla with a reduced and fixed lacinia and a 1 -segmented
galea. Pronotum with lateral margins never broadly crenulate
or. toothed; procoxal cavities without lateral extension, trochan-
tin always concealed. Mesocoxal cavities with trochantins con-
cealed. Metendosternite with median stalk very short and
broad, so that arms may appear to arise independently. Hind-
wing with reduced anal region bearing a single vein. Trochanter
of normal type, oblique but completely separating coxa and
femur. Tibial spurs absent on all legs, apices of tibiae, especially
protibiae, variously expanded and modified. Abdominal sternites
free, III often with a median pubescent fovea in male. Aedeagus

10 BREVIORA No. 424

with a small basal piece, without condyles, paramere variously
modified at apex but without basal setae, and median lobe
sclerotized and without lateral struts.

Larva. Head with endocarina, ocelli 5 or less. Antenna 2-
segmented. Mandibles often asymmetrical, with or without mola
and retinaculum. Maxillae free to about middle of mentum,
without an articulating sclerite at its base. Spiracles annular.
Ninth tergum variously modified, usually with two urogomphi;
ninth sternum without asperites.

This subfamily includes all members of the family except
Sphindocis.

DISCUSSION

The major justification for uniting Sphindocis and the Ciidae
is the joint possession by the two groups of at least one feature —
the distinctive laciniar lobe of the larval maxilla — which is
certainly unique and derived. This particular type of structure is
found in no other cucujoid beetle, and it is sufficiently complex
and similar in the two groups to make convergence unlikely.
There is no reason to believe that the cleft malar apex of the
Zopheridae, Cephaloidae, and Synchroidae, or the various teeth,
spines, or hooks (to which the word uncus is often appHed) of
Anaspis, the Oedemeridae, and various other Heteromera are
homologous to this laciniar lobe. The loss of the mandibular
mola and of the hypopharyngeal sclerome in the larva are also
derived features, but it would be difficult to demonstrate their
uniqueness. The lightly sclerotized and tooth-like "retinaculum"
of the larval mandible appears to be unique in the Heteromera,
but similar structures occur in a number of Clavicornia, sug-
gesting that the character may be primitive. In the adult stage,
the reduction of the ligula and the presence of an abdominal
fovea in the male may both represent synapomophic conditions,
but most other adult characters are shared by one or more
Heteromera. The abdominal fovea is rare in this section of the
Cucujoidea, although some Mycteridae and at least one myceto-
phagid have abdominal tufts or patches of hairs in the male.
Foveae similar to those of ciids, however, do occur in certain
Erotylidae among the Clavicornia (Delkeskamp, 1959).

The erection of a new subfamily for Sphindocis is based on
numerous differences between this genus and all of the remaining
ciids. In larval Ciinae, the antennae are reduced to two seg-
ments, an endocarina is present, the maxillary articulating area

1974 LARVA OF SPHINDOCIS 11

is reduced to a narrow membrane, the spiracles are annular
without a pair of contiguous air tubes, the ninth sternite lacks a
row of asperites, and the gula and submentum are not fused,
while in the adults of this subfamily, the antennae always have
less than 1 1 segments, the club segments bear multi-pronged
sensillae, the galea has only a single segment, the lacinia is not
articulated, the anal region of the hindwing has only a single
vein, the pro- and mesotrochantins are concealed, the trochanters
are not heteromeroid, the tibial spurs are lacking, the abdominal
stemites are free, and the median lobe of the aedeagus is sclero-
tized. Most of these characters are derived and several are ob-
viously associated with reduction in size (hindwing, antennal
segments of adult and larva, adult maxilla ) . The development
of large and complex hygroreceptor sensillae on the antennal
club probably represents an improvement in the ability to locate
fungus sporophores, while the formation of a larval endocarina,
reduction of the maxillary articulating area, the further enclosure
of the pro- and mesocoxae, and the loss of tibial spurs may have
been associated with the utilization of a tougher fungus substrate.

The relationships of the Ciidae to other heteromerous families
are still somewhat obscure, and a detailed discussion must await
a study now in progress on adult and larval Heteromera. Crow-
son (1966) suggested that the Ciidae, along with the Pteroge-
niidae, Tetratomidae, and Mycetophagidae, might be direct off-
shoots from a biphyllid-byturid type of heteromeran ancestor,
and that the Pterogeniidae might represent the sister group of
the Ciidae. I have agreed basically with Crowson's views ( Law-
rence, 1971), while allowing for the possibility that the ciids
have evolved directly from a clavicorn ancestor, perhaps related
to Cryptophilus or Setariola in the Languriidae.

The Pterogeniidae resemble ciids both as adults and larvae,
but the similarities may be due to the fact that both groups
inhabit the woodier fungi. Adult pterogeniids differ from the
Ciidae in having filiform antennae, securiform maxillary palps,
a '5-5-4 tarsal formula, internally closed procoxal cavities, and
distinct lateral lobes on the aedeagus. The larvae of Pterogenius
and Histanocerus, which are being described elsewhere, differ
from those of ciids in having a characteristically curved epi-
cranial stem, an extensive mandibular mola with transverse
ridges, a well-developed and molar-like hypopharyngeal sclerome,
and no laciniar lobe on the maxilla.

The row of asperites at the apex of the ninth sternite in the
Sphindocis larva is found outside the group only in the genus

12 BREVIORA No. 424

Prostomis, which has been placed in a separate family of un-
certain affinities. The row of asperites in the larvae of Pythidae,
Pyrochroidae, and Othniidae is always at the base of the ninth
sternite and is apparently not homologous to that of Sphindocis.
The Prostomidae differ from ciids in having closed front and
middle coxal cavities in the adult and a simple mala, well-
developed mola, and hypopharyngeal sclerome in the larva.

The Tetratomidae have also been considered as a possible
sister group of the Ciidae, and certain characters of both adult
and lars^a tend to support this hypothesis. Adults of the Tetra-
tomidae (excluding Mycetoma, removed by Crowson, 1966, and
Viedma, 1966) and the related Mycetophagidae are similar to
ciids in having internally and externally open procoxal and
laterally open mesocoxal cavities, while the pisenine tetratomid
Eupisenus elongatus (LeConte) bears a striking superficial re-
semblance to Sphindocis. The procoxal cavity in all tetratomids
has a distinct lateral extension that exposes the trochantin; in
Sphindocis there is a slight extension of the cavity, while in the
Ciinae it is absent. The hindwing of Sphindocis is similar to
that of tetratomids in having a wedge cell and subcubital fleck
and differs in having four rather than five anal veins, while the
metendosternite is essentially of the tetratomid type with a re-
duced lamina. In the Ciinae, both the hindwing and the meten-
dosternite have undergone extreme reduction and modification.

The male genitalia of the Tetratomidae are variable, and
Miyatake (1960) has described and illustrated two major types:
that of Pisenus, with the basal piece ventral and bearing two
ventral accessory lobes in addition to the parameres, which are
free; and that of the Tetratomini, with the basal piece dorsal
and bearing only parameres, which are at least partly fused
together. In the genus Pent he (Penthini) the genitalia are of
the tetratomine type, but in Eupisenus, a distinctive type occurs
with the basal piece ventral and the parameres fused into a
single piece notched at the apex; moreover, this single paramere
bears near the base two clusters of six or seven setae, which are
in the same positions as the two pairs of setae in Sphindocis.
The median lobe is also like that of Sphindocis in being mem-
branous with lateral struts that meet at the apex.

The larv^ae of Tetratomidae are also quite variable, but they
differ consistently from those of Ciidae in having lyre -shaped
frontal arms and no laciniar lobe on the maxilla. The mandible
of Pisenus resembles that of the Mycetophagidae in having a
mola with transverse ridges grading into tubercles or asperites on

1974 LARVA OF SPHINDOCIS 13

the ventral surface (Hayashi, 1971; 1972). In Eupisenus, the
niola is simple and concave and is bordered by two rows of teeth
that grade into tubercles both dorsally and ventrally. In the
Tetratomini (Crowson, 1964) the mola is further reduced with
only three or four teeth, while in Pent he there is no mola. The
hypopharyngeal sclerome, which can often be correlated with
molar development, is well developed and tooth-like in Pisenus,
consists of a transverse band in Eupisenus and the tetratomines,
and is barely sclerotized in Penthe. It would not be difficult to
derive the simple mandible and unsclerotized hypopharynx of
the Ciidae from a form like Eupisenus, and it is also possible
that the "retinaculum" of the Ciidae represents a remnant of
the molar teeth in tetratomids, rather than a carry-over of the
clavicorn retinaculum.

LITERATURE CITED

Anderson, W. H. 1936. A comparative study of the labium of coleopter-
ous larvae. Smiths. Misc. Coll., 95 (13) : 1-29.

BoviNG, A. G., AND F. C. Craighead. 1931. An illustrated synopsis of the
principal larval forms of the Coleoptera. Ent. Amer. (N.S.) , 11: 1-351,
125 pis.

CoQUEREL, C. 1849. Observations entomologiques sur divers Coleopteres
recueillis aux Antilles. Ann. Soc. Ent. France, ser. 2, 7: 441-454, pi. 14.

Crowson, R. a. 1955. The Natural Classification of the Families of Cole-
optera. London: Lloyd. 187 pp.

. 1964. Observations on British Tetratomidae (Col.) , with

a key to the larvae. Ent. Mon, Mag., 94: 82-86.

. 1966. Observations on the constitution and subfamilies of

the family Melandryidae. Eos, 41: 507-513.
Delkeskamp, K. 1959. Sekundare Geschlechtsmerkmale bei Erotyliden.

Wiss. Zeit. Martin-Luther-Universitat Halle-Wittenberg, Math.-Nat.,

8 (6) : 1089-1098.
Emden, F. van. 1942. Larvae of British beetles — III. Keys to families.

Ent. Mon. Mag., 78: 206-226, 253-272.
Fall, H. C. 1917. New Coleoptera — VI. Canadian Ent., 49: 163-171.
Glen, R. 1950. Larvae of the elaterid beetles of the tribe Lepturoidini

(Coleoptera: Elateridae) . Smith. Misc. Coll., 111(11): 1-246.
Hayashi, N. 1971. On the larvae of Mycetophagidae occurring in Japan

(Coleoptera: Cucujoidea) . Kontyu, 39: 361-367.
1972. On the larvae of some species of Colydiidae, Tetratomi-
dae and Aderidae occurring in Japan (Coleoptera: Cucujoidea) . Kontyu,

40: 100-111.
Hyslop, J. A. 1915. Observations on the life history of Meracantha con-

tracta (Beauv.) . Psyche, 22: 44-48, pi. 4.

14 BREVIORA No. 424

Lawrence, J. F. 1971. Revision of the North American Ciidae (Coleop-

tera) . Bull. Mus. Comp. Zool., 142: 419-522.
MiYATAKE, M. 1960. The genus Pisenus Casey and some notes on the

family Tetratomidae (Coleoptera) . Trans. Shikoku Ent. Soc, 6: 121-135.
RozEN, J. G. 1958. The external anatomy of the larva of Nacerdes mela-

nura (Linnaeus) (Coleoptera: Oedemeridae) . Ann. Ent. Soc. America,

51: 222-229.

1960. Phylogenetic-systematic study of larval Oedemeridae

(Coleoptera) . Misc. Publ. Ent, Soc. America, 1 (2) : 35-68.
St. George, R. A. 1924. Studies on the larvae of North American beetles

of the subfamily Tenebrioninae with a description of the larva and

pupa of Merinus laevis (Olivier) . Proc. U. S. Nat. Mus., 65 (1) : 1-22,

pis. 1-4.
Snodgrass, R. E. 1935. Principles of Insect Morphology. New York: Mc-
Graw-Hill. X + 667 pp.
ViEDMA, M. G. de. 1966. Contribucion al conocimiento de las larvas de

Melandryidae de Europa (Coleoptera) . Eos, 41: 483-506.
Watt, J. C, 1970. Coleoptera: Perimylopidae of South Georgia. Pacific

Ins. Mon., 23: 243-253.
Wheeler, W. M. 1928. The Social Insects. Their Origin and Evolution.

New York: Harcourt-Brace. xviii + 378 pp.

JUL 8 1P74

B R E V^T-O^ R A

Miisenni of Comparative Zoology

us ISSN 0006-9698

Cambridge, Mass. 28 June 1974 Number 425

SYSTEMATIGS AND DISTRIBUTION OF

GERATIOID ANGLERFISHES OF THE GENUS

LOPHODOLOS (FAMILY ONEIRODIDAE)

Theodore W. Pietsch^

Abstract. The genus Lophodolos of the family Oneirodidae is reviewed
on the basis of all known material. Two species are recognized, L. acantho-
gnatlius Regan and L. indicus Lloyd. Lophodolos dinema Regan and Tre-
wavas is considered a junior synonym of L. indicus Lloyd. The tentative
distribution of each species is plotted and a key to the species of the genus
is provided.

INTRODUCTION

The genus Lophodolos was erected by Lloyd (1909a) to in-
clude a single species, L. indicus, on the basis of a specimen
collected from the Indian Ocean by the Royal Indian Museum
Survey Ship Investigator. Since that time three additional
species have been described: L. acantho gnathus Regan (1925),
to which have been referred more than 60 specimens from the
Atlantic and western Pacific oceans; L. lyra Beebe (1932),
synonymized with L. acantho gnathus by Regan and Trewavas
(1932); and L. dinema Regan and Trewavas (1932), repre-
sented by a single specimen from the South China Sea.

The number of female specimens of Lophodolos has doubled
since the appearance of Bertelsen's ( 1 95 1 ) monograph on the
Ceratioidei. In spite of extensive information gained from this
increase in material, taxonomic study of the genus is by no means
complete. Metamorphosed males are unknown; thus, species
are based only on females. The separation of species is based
on only a few characters, the most important being the morphol-
ogy of the esca and the length of the illicium. Differences in

^Museum of Comparative Zoology, Cambridge, Massachusetts 02138

2 BREVIORA No. 425

these two characters merge in specimens less than 25 mm stand-
ard length, making differentiation particularly difficult. Never-
theless, the material presently known appears to represent only
two forms: L. acanthognathus Regan (1925) and L. indicus
Lloyd (1909a).

METHODS AND MATERIALS

Standard lengths (SL) were used throughout. Measurements
were taken on the left side of the fish whenever possible and
rounded to the nearest 0.5 mm in specimens greater than 20 mm,
and to the nearest 0.1 mm in specimens less than 20 mm. To
insure accurate fin-ray counts, skin was removed from the pec-
toral fins and incisions were made to reveal the rays of the dorsal
and anal fins. lUicium length is the distance from the articula-
tion of the pterygiophore of the illicium and the illicial bone to
the dorsal surface of the escal bulb, excluding escal appendages.
Terminology used in describing the various parts of the angling
apparatus follows that of Bradbury (1967). Definitions of terms
used for the different stages of development follow those of
Bertelsen (1951: 10-11). "^

Locality data is given for primary type material only. Com-
plete locality data for all specimens examined may be obtained
by writing to the author.

The generic diagnosis (much of which is taken from osteologi-
cal evidence presented elsewhere: Pietsch, 1974) and descrip-
tion are based on 98 metamorphosed females ranging from 6.0
to 77.0 mm (metamorphosed males are unknown). Larvae were
described by Bertelsen ( 1 95 1 : 1 06 ) . Study material is deposited
in the following institutions :

BMNH British Museum (Natural History), London.
BOC Bingham Oceanographic Collections, Peabody Mu-

seum of Natural History, Yale University.
BZM University of Bergen Zoological Museum.

CAS California Academy of Sciences, San Francisco.

FMNH Field Museum of Natural History, Chicago.
GNM Natural History^ Museum of Goteborg.
IMC Indian Museum, Calcutta.

ISH Institut fiir Seefischerei.

LACM Los Angeles County Museum of Natural History.
MCZ Museum of Comparative Zoology, Harvard University.

NIO National Institute of Oceanography, Surrey, England.

NYZS New York Zoological Society.

1974 ANGLERFISHES OF THE GENUS LOPHODOLOS 3

ROM Roval Ontario Museum.

SIO Scripps Institution of Oceanography, La Jolla.

SU Stanford University (collections now housed at the

California Academy of Sciences, San Francisco).

UMML University of Miami Marine Laboratory.

USNM United States National Museum, Washington.

ZMUC Zoological Museum, University of Copenhagen.

ACKNOWLEDGEMENTS

I thank Erik Bertelsen and Karel F. Liem for critically read-
ing the manuscript and offering valuable suggestions. Thanks
are also due the following persons and their institutions for mak-
ing material available : Robert J. Lavenberg and Jerry Neumann
(LACM); Erik Bertelsen and j0rgen Nielsen (ZMUC); Ger-
hard Krefft (ISH); Richard H. Rosenblatt (SIO); W. B.
Scott (ROM) ; William N. Eschmeyer and Tomio Iwamoto
(CAS); Alwyne Wheeler (BMNH) ; C. Richard Robins
(UMML) ; Nigel Merrett and JuHan Badcock (NIO) ; Robert
H. Gibbs, Jr. (USNM); Robert K. Johnson (FMNH) ;
Thomas A. Clarke, Hawaii Institute of Marine Science, Uni-
versity of Hawaii; and MicheL Legand, Office de la Recherche
Scientifique et Technique Outre-Mer, Noumea, New Caledonia.
A. G. K. Menon of the Zoological Survey of India, Calcutta,
kindly provided information and a sketch of the esca of the
holotype of Lophodolos indicus. Finally, I thank Patricia Chaud-
huri for the fine illustrations.

SYSTEMATICS
Genus Lophodolos Lloyd, 1909a

Lophodolos Lloyd, 1909a: 167 (type species Lophodolos indicus Lloyd, 1909a,
by original designation and monotypy) . Fowler, 1936: 1337, 1339-1340,
1365, fig. 560 (brief description after Regan, 1926; in key) . Pietsch,
1974: in press (osteology; relationships) .

Lophodolus (emended or erroneous spelling of Lophodolos by various
authors) .

Oneirodes Murray and Hjort, 1912: 104, fig. 90 (in part; erroneous designa-
tion; type species Oneirodes eschrichtii Liitken, 1871, by original desig-
nation and monotypy) .

Lophodulus (erroneous spelling of Lophodolos by various authors) .

Diagnosis. The genus Lophodolos is distinguished from all
other genera of the family Oneirodidae by the following charac-
ters : dorsal profile of frontal bones concave ; ventromedial exten-

% BREVIORA No. 425

sions of frontals absent; posterior end of frontal in contact with
respective prootic; pterosphenoid absent; pterygiophore of illi-
cium emerging between or behind sphenotic spines; symphysial
and sphenotic spines extremely well developed; medial ends of
hypobranchials II (as well as hypobranchials III) approaching
each other on the midline (see Pietsch, 1974: in press).

In addition, Lophodolos is unique in having the following
combination of characters: snout short, mouth large, cleft ex-
tending past eye; vomerine teeth absent; anterior end of pterygio-
phore of illicium exposed, its posterior end concealed under skin ;
articular spines present, quadrate spine larger than mandibular
spine; angular spine present; pharyngobranchials I and II ab-
sent; pectoral lobe short and broad, shorter than longest rays of
pectoral fin ; operculum bifurcate ; suboperculum slender through-
out length, its upper end tapering to a point, its lower end
rounded, with a small anterior projection in some adolescent
specimens; skin naked, covering caudal fin to some distance from
fin base.

Description. Body relatively long, slender, not globular; jaws
equal anteriorly; lower jaw with an unusually long symphysial
spine; oral valves well developed, lining inside of both upper
and lower jaws; two nostrils on each side at end of a single short
tube; labial cartilage well developed (Pietsch, 1972a: 31);
angular bone terminating as a well-developed spine; eye sub-
cutaneous, appearing through a circular, translucent area of
integument; gill opening oval in shape, situated just postero-
ventrad to pectoral lobe; skin naked (embedded dermal spines
cannot be detected microscopically in cleared and stained speci-
mens) ; lateral line papillae as described for other oneirodids
(Pietsch, 1969, 1972b) ; ovaries paired; pyloric caeca absent.

Illicium length 11. 1 to 138.0 percent of SL, becoming longer
proportionately with growth ( Fig. 1 ) ; anterior end of pterygio-
phore of illicium exposed, emerging on head between or behind
sphenotic spines, its posterior end concealed under skin; esca
with a pair of unpigmented, bilaterally placed appendages arising
from distal surface.

Teeth slender, straight, all depressible, and weakly set (easily
damaged or lost), in overlapping sets as described for other
oneirodids (Pietsch, 1972b: 5, fig. 2) ; teeth in lower jaw larger
and more numerous than those in upper jaw; number of teeth
in lower jaw 200 to 280 (based on five specimens, 57.0-
77.0 mm) ; pharyngobranchial II absent; pharyngobranchial III
well developed and bearing numerous teeth.

1974 ANGLERFISHES OF THE GENUS LOPHODOLOS

100

— I- T ■ 1 1 1 — r

o L. acantho^nathus

— 1

•

•
•

• L. indicus

•

UJIUUI u

•

• •

1 40
3

•
•

•

o
o

.d..OOOO o oo^oo _ o _

0 10 20 30 40 50 60 70 88

Standard Length in mm

Figure 1. Relationship of illicial length and standard length for species
of Lophodolos.

3.0

2.5

2.0

u 1.5

— 1 1 1 1—

o L. acantho^nathus

1 - I-

Q O

• L. indicus

o
o o

•

o o o

o • •

• •
•

•

o o o o o

O O • •
o

•

• •

•

1.0 -

0 10 20 30 40 50 60 70 80

Standard Length in mm

Figure 2. Relationship of escal bulb width and standard length for
species of Lophodolos.

BREVIORA

No. 425

'•J

-a
c

-Ji

■■J

.T5
I

Tt^

--^ oo in a:

-H xo ;0 (M

C^ 00 CC ^

^ CO — . jO

-^ p-H c>r

CO o 00 — •

(M OO

(M C^

—I oc

■^ l^ -H

CM
SO

lO r—tlCi

• •

-C

•«o

.«

r-<

«o

S-i

•**»

<*•**

t4H

■TS

c«

■?i.

•?iH

,;^

1974 ANGLERFISHES OF THE GENUS LOPHODOLOS 7

Color in presentation dark brown to black over entire external
surface of body except for bulb and appendages of esca (escal
appendages and unpigmented distal portion of escal bulb silvery-
white in unpreserved specimens of L. acanthognathus; E. Bertel-
sen, personal communication) ; oral cavity and guts except for
outer surface of stomach wall unpigmented.

D. 5-8, first ray of dorsal fin reduced to a small stub; A.
4-7; P. 17-21 (Table 1); pelvics absent; C. 9 (2 unbranched
- 4 branched - 3 unbranched ) ; branchiostegal rays 6 ( 2 + 4 ) .

Relationships. Lophodolos appears to be the most derived
genus of the thirteen oneirodid genera. It is extremely special-
ized in many ways and, although probably derived from a
Microlophichthys-likt ancestor (a relatively primitive member
of the family ) , it shows little resemblance to any other oneirodid
(seePietsch, 1974).

Distribution. Both species of Lophodolos have a wide hori-
zontal distribution, and occur in all three major oceans of the
world. Lophodolos indicus has not been taken in the western
Atlantic where 82 percent of the material of L. acanthognathus
has been collected. On the other hand, L. acanthognathus is rep-
resented in the eastern Pacific by only three specimens ( Fig. 9 ) .

Since virtually all collections of Lophodolos were made with
nonclosing nets, vertical distributions are based on the maximum
depths reached by fishing gear for each capture. Metamor-
phosed specimens were taken between approximately 650 m
and an unknown lower limit. All specimens larger than 30 mm
(37 individuals) were captured by nets fished below 1000 m;
62 percent of these were captured by nets fished below 1500 m.
Material of both species from any one geographical area was
insufficient to show whether there is any vertical separation be-
tween the two forms.

Comments. The original spelling of the generic name Lopho-
dolos (Lloyd, 1909a), is reestablished as the "correct original
spelling," as provided by Article 32(a) of the International Code
of Zoological Nomenclature.

Key to the Females of the Species of the Genus Lophodolos

lA. Length of illicium less than 25 percent of SL in specimens 30 mm and
larger (Fig. 1) ; width of escal bulb 4.4-6.7 percent of SL in specimens
25 mm and larger (Fig. 2) ; length of escal appendages 10.2-20.9 percent
of SL in specimens 25 mm and larger, 8.7-22.2 (usually greater than
10.0) percent of SL in specimens less than 25 mm (Figs. 3-4) ; length of
sphenotic spine 4.1-9.2 (usually greater than 6.0) percent of SL in
specimens 30 mm and larger (Fig. 5) ; length of quadrate spine 2.9-6.5

8 BREVIORA No. 425

(usually greater than 3.5) percent of SL in specimens 30 mm and larger
(Fig. 6) ; D. 5-7 (Table 1) L. acanthognathus Regan, 1925.

IB. Length of illicium greater than 25 percent of SL in specimens 30 mm
and larger (Fig. 1) ; width of escal bulb 2.1-4.0 percent of SL in speci-
mens 25 mm and larger (Fig. 2) ; length of escal appendages 1.2-5.0
percent of SL in specimens 25 mm and larger, 4.2-10.5 (usually less
than 9.0) percent of SL in specimens less than 25 mm (Figs. 3-4) ;
length of sphenotic spine 1.9-6.0 (usually less than 5.0) percent of SL
in specimens 30 mm and larger (Fig. 5) ; length of quadrate spine
1.6-5.0 (usually less than 3.0) percent of SL in specimens 30 mm and
larger (Fig. 6) ; D. 6-8 (Table 1) L. indicus Lloyd, 1909a.

Lophodolos acanthognathus Regan, 1925

Figure 3

Oneirodes n. sp. Murray and Hjort, 1912: 104, fig. 90 (erroneous designation;
specimen referred to L. acanthognathus by Nybelin, 1948) .

Lophodolus acanthognathus Regan, 1925: 563 (original description; two
specimens; lectotype designated by Bertelsen, 1951, ZMUG P92104,
12.0 mm; DANA Station 1358 (5), western North Atlantic, 28°15'N,
56°00'W; 3000 m wire; 1530 hr; 2 June 1922) . Regan, 1926: 30, pi. 6,
fig. 1 (brief description; one additional specimen) . Regan and Tre-
wavas, 1932: 83 (description after Regan, 1926; five additional speci-
mens; L. lyra Beebe, 1932, a synonym of L. acanthognathus) . Gregory,
1933: 402, 404, figs. 274, 276A, 277 (osteological comments; specific name
misspelled acanthagnathus in fig. 277) . Beebe, 1937: 207 (45 specimens
listed from Bermuda) . NybeUn, 1948: 86-89, Text-fig. 9, table 20
{Oneirodes n. sp. of Murray and Hjort, 1912, referred to L. acanthog-
nathus; description of an additional specimen; comparison with previous
descriptions; geographic, bathymetric distribution) . Bertelsen, 1951:
107, figs. 64-65, tables 21-22 (synonymy; description; comparison with
all known material; DANA material listed; comments; in key) . Grey,
1955: 299 (one additional specimen) . Grey, 1956: 255 (synonymy;
vertical distribution) .

Lophodolus lyra Beebe, 1932: 96-98, fig. 28 (original description; about 40
specimens; holotype, USNM 170949, 47.0 mm; GLADISFEN Net 111,
32°12'N, 64°36'W; 1463 m; 27 July 1931). Koefoed, 1944: 7, pi. 3,
fig. 3 (misidentifications; description; three specimens including Onei-
rodes n. sp. of Murray and Hjort, 1912) .

Lophodolos acanthognathus, Fowler, 1936: 1340, 1365, fig. 560 (corrected
spelling; brief description after Regan, 1926) . Pietsch, 1972a: 35, 45
(osteological comments) . Pietsch, 1974: in press (osteology; relation-
ships) .

Material. Seventy-six female specimens, 6.0-70.0 mm:
BMNH 4(18.0-26.0 mm); BOC 3; BZM 3(8.5-51.0 mm);
FMNH 1(9.5 mm); GNM 1(56.0 mm); ISH 6(46.0-70.0

1974

ANGLERFISHES OF THE GENUS LOPHODOLOS

Figure 3. Esca of Lophodolos acanthognathus, LACM 10011-9, 38.0 mm,
left lateral view. Drawn by Patricia Chaudhuri.

10 BREVIORA No. 425

10-

5 8

2 •

1 ' . 1 ' i " "J '1 "■ 1

oL acanthognathus

•

• Lindicus

O o

o o o

•

o o

Ooo„OoO° , . ,
o Oo oO ° • • • -

c
1 1 — 1 1 1 '

•
•

•

0 10 20 30 40 50 60 70 80

Standard Length m mm

Figure 4. Relationship of escal appendage length and standard length
for species of Lophodolos.

mm); LACM 4(22.5-38.0 mm); MCZ 2(18.0-65.0 mm):
ROM 13(17.0-57.0 mm); SU 32(6.0-32.0 mm); USNM
2(10.0-47.0 mm); ZMUC 5(8.5-40.0 mm).
Diagnosis. See key to species.

Description. Illicium short, 11.1-23.1 (Fig. 1); width of
escal bulb large, 4.2-9.0 (Fig. 2); escal appendages long, 8.7-
22.2 (Figs. 3-4); sphenotic spines long, 4.1-9.2 (Fig. 5);
quadrate spines long, 2.9-6.5 (Fig. 6) ; D. 5-7 (only one speci-
men had D. 7, ISH 500/73); A. 4-6; P. 17-21 (Table 1)
(measurements in percent of SL; spine lengths based on speci-
mens greater than 30 mm, fin ray counts on specimens greater
than 20 mm).

Rest of characters as for genus.

Distribution. Lophodolos acanthognathus is known from
both sides of the Atlantic. The vast majority of specimens (82
percent, including all type material) have been collected from
the western half of this ocean as far east as 26°W, between 58°N
and 25 °N. In the eastern Atlantic the range extends from ap-
proximately 48°N, 18°W, southward, off the southern tip of

1974

ANGLERFISHES OF THE GENUS LOPHODOLOS

4.5

oL. acanthognathus

•

4.0

• L. indicus

■

E 3.5

•

• ^B

•

g 3.0

o o •

•
• •

• •

■q.

■^ 2.5

o
c:

■

•

2.0

•
• o

•

1.5

o •

1 1 . 1 —

40 50

Standard Length in mm

Figure 5. Relationship of sphenotic spine length and standard length for
species of Lophodolos.

Portugal and the continental slope of Africa to2°S, 26°W. A
single record is known from the central South Atlantic at ap-
proximately 40°S, 26°W.

In the Indo-Pacific region L. acanthognathus is represented
by three specimens: one from the Bay of Bengal, Indian Ocean
(at approximately 7°N, 60°E), and two from the South China
and Celebes seas. Three records are known from eastern Pacific
Equatorial waters: on the equator at 139°W and from off the
coast of Peru. The lectotype was collected from the western
north Adantic at 28°15'N, 56°00'W (Fig. 9).

On the basis of maximum depths reached by fishing gear,
metamorphosed L. acanthognathus are vertically distributed be-
tween approximately 650 m and an unknown lower limit. All

12 BREVIORA No. 425

■ i.V

- r r—

— I

■ )

2.5

•

o L. acanthognathus

•

• L. indicus

•

' ^"

5 2.0

o *

•

'o.

*^ 1.5

• .

•

"rS

•

1.0

•

■

20 30 40 50 60 70 SO

Standard Length in mm

Figure 6. Relationship of quadrate spine length and standard length for
species of Lophodolos.

specimens larger than 30 mm (19 individuals) were captured
by nets fished below 1000 m; 58 percent of these were captured
by nets fished below 1500 m.

Comments. Specimens of L. acanthognathus larger than ap-
proximately 30 mm can easily be separated from L. indicus on
the basis of illicial and escal appendage lengths alone (see key to
species). Smaller specimens, especially those less than 20 mm,
are difficult to identify, and require a combination of meristics
and counts, all of which overlap between the two species:
illicial and escal appendage lengths, width of escal bulb, and
dorsal fin ray counts (See Figs. 1-2, 4, Table 1). In some
cases, geographic distribution may provide additional data for
identification; L. indicus apparently does not occur in the western
North Atlantic where approximately 82 percent (62 indi\iduals)
of the known material of L. acanthognathus has been collected
(Fig. 9).

The holotype L. lyra Beebe (1932) compares well with the
known material of L. acanthognathus ; the name is retained as a
synonym of L. acanthognathus following Regan and Trewavas

(1932).

1974 ANGLERFISHES OF THE GENUS LOPHODOLOS 13

Lophodolos indicus Lloyd, 1909a
Figures 7-8

Lophodolos indicus Lloyd, 1909a: 167 (original description; single specimen;
holotype, IMC 1024/1, 53.0 mm; INVESTIGATOR Station 307, off
Kerala (formerly Travancore) , southwest coast of India: 0-1624 m) .

Lophodolus indicus, Lloyd, 1909a: pi. 45, fig. 7 (holotype figured) . Regan,
1925: 563 (comparison with L. acanthognathus) . Regan, 1926: 30 (brief
description after Lloyd, 1909a; comparison with L. acanthognathus) .
Regan and Trewavas, 1932: 83 (after Lloyd, 1909a, Regan, 1926) . Ber-
telsen, 1951: 108 (description after Lloyd, 1909a, Regan and Trewavas,
1932; comparison with all known material of Lophodolos) . Grey, 1956:
255-256 (synonymy; vertical distribution) .

Lophodolus dinema Regan and Trewavas, 1932: 83, pi. 4, fig. 3 (original
description; single specimen; holotype, ZMUC P92105, 43.0 mm; DANA
Station 3716(2), South China Sea, 19°18'N, 120°13'E; 3000 m wire; hot-
tom depth 3225 m; 1400 hr; 22 May 1929). Bertelsen, 1951: 108 (de-
scription; comparison with all known material of Lophodolos) . Grey,
1956: 255 (synonymy; vertical distribution) .

Material. Twenty-two female specimens, 9.5—77.0 mm: IMG
1(53.0 mm); ISH 5(36.0-75.0 mm); LACM 4(32.5-71.0
mm) ; MCZ 4(30.0-64.5 mm) ; NIO 1 (57.0 mm) ; SIO 5(9.5-
77.0 mm); UMML 1(23.0 mm); ZMUC 1(43.0 mm).

Diagnosis. See key to species.

Description. Illicium long, 15.2-138.0 (Fig. 1); width of
escal bulb small, 2.1-5.2 (Fig. 2) ; escal appendages short, 1.2-
10.5 (Figs. 4, 8); sphenotic spines short, 1.9-6.0 (Fig. 5);
quadrate spines short, 1.6-5.0 (Fig. 6) ; D. 6-8; A. 5—7; P. 17—
21 (Table 1) (measurements in percent of SL; spine lengths
based on specimens greater than 30 mm, fin ray counts on speci-
mens greater than 20 mm) .

Rest of characters as for genus.

Distribution. In the Atlantic Ocean, L. indicus appears to be
restricted to the eastern side ; seven specimens are known from off
the continental slope of Africa from 20°N, 21°W, east to the
Gulf of Guinea and south to approximately 18°S, 10°W. The
remaining material (15 specimens) is rather evenly distributed
across the Indian and Pacific oceans between approximately
4°S and 30°N. The holotype was collected off the southwest
coast of India ( Fig. 9 ) .

On the basis of maximum depths reached by fishing gear,
metamorphosed L. indicus are vertically distributed between
approximately 750 m and an unknown lower limit. All speci-
mens larger than 30 mm (18 individuals) were captured by

t_^

BREVIORA

No. 425

»n

IT!

o
c

1974

ANGLERFISHES OF THE GENUS LOPHODOLOS

figure 8. Esca of Lophodolos indicus, MCZ 47559, 58.0 mm, left lateral
view. Drawn by Patricia Chaudhuri.

nets fished below 1000 m; 67 percent of these were captured by
nets fished below 1500 m.

Comments. Large specimens of L. indicus (greater than ap-
proximately 30 mm) are easily distinguished from L. acantho-
gnathus on the basis of ilHcial and escal appendage lengths alone
(see key to species). Smaller specimens are more difficult to
identify (see comments under L. acanthognathus).

Lophodolus dinema Regan and Trewavas (1932) was de-
scribed as new on the basis of an escal morphology differing
from that of L. indicus. These differences, however, are un-
doubtedly the result of damage. The esca of the holotype of L.
indicus, originally described by Lloyd (1909a: 167) as being
''hard but . . . covered with short, shreddy filaments," has lost

BREVIORA

No. 425

"-^

■e

-»»

•?i.

•-I

(/I

• »^

tf3

U-4

S-i

4..)

i~,

•4-^

1/5

■O

fTl

!-l

bf)

a
a,

Si-

cr>

•»-H

1974 ANGLERFISHES OF THE GENUS LOPHODOLOS 17

the two bilaterally placed appendages found in the holotype of
L. dinema and in all known specimens of Lophodolos. Although
I did not see it, the poor condition of the esca was confirmed by
a sketch made from the holotype of L. indicus provided by
A. G. K. Menon of the Zoological Survey of India. Discrepan-
cies in illicial length (Bertelsen, 1951: 107) are also more ap-
parent than real. A plot of illicial length against standard length
( Fig. 1 ) shows the holotype of L. dinema to compare well with
the material here considered to be L. indicus. In the absence
of any significant differences, L. dinema is here synonymized
with L. indicus.

Species Ingertae Sedis
Lophodolos biflagellatus Koefoed, 1944, nomen nudum.
Lophodolus biflagellatus Koefoed, 1944: 7.

Comments. This name was used by Koefoed in a manuscript
dated 1918 (not seen by me), and later mentioned in published
form (Koefoed, 1944: 7) without application to a description
or type.

SPECIES RELATIONSHIPS

Lophodolos acanthognathus and L. indicus are distinguished
on the basis of five characters: ilHcial length, escal bulb width,
escal appendage length, sphenotic spine length, and quadrate
spine length. For most of these characters it is difficult, if not
impossible, to know whether a character state represents a primi-
tive or a derived condition. The longer ilHcium of L. indicus
(Fig. 1), however, is surely a derived state; an increase in illicial
length is a trend found within other oneirodid genera {Dol-
opichthys, Oneirodes, and Chaenophryne; Pietsch, 1972b, 1974).
The width of the escal bulb of L. acanthognathus is like that of
nearly all other oneirodids; the considerably narrower bulb of
L. indicus (Fig. 2) is most likely a derived condition. Lopho-
dolos acanthognathus has significantly longer escal appendages
than L. indicus (Fig. 4), perhaps representing a derived state;
longer escal appendages and filaments are found in the more
derived species of Oneirodes (Pietsch, 1974). The sphenotic
and quadrate spines of L. acanthognathus are long relative to
those of L. indicus; either character state, long versus short, may
represent the derived condition. From this character analysis, it
is reasonable to speculate that L. indicus is the more derived
member of the genus.

18 BREVIORA No. 425

LITERATURE CITED

Beebe, William. 1932. Nineteen new species and four post-lan^al deep-sea

fish. Zoologica (N.Y.) , 13: 47-107.
1937. Preliminary list of Bermuda deep-sea fish. Zoo-
logica (N.Y.) , 22(14) : 197-208.
Bertelsen, E. 1951. The ceratioid fishes. Ontogeny, taxonomy, distribu-
tion and biology. Dana Rep., 39, 276 pp.
Bradbury, Margaret G. 1967. The genera of Batfishes (family Ogcoce-

phalidae) . Copeia, 1967: 399-422.
Fowler, H. W. 1936. The marine fishes of West Africa. Based on the

collections of the American Museum Congo Expedition, 1909-1915.

Bull. Amer. Mus. Nat. Hist., 70 (2) : 607-1493.
Gregory, W. K. 1933. Fish skulls. A study of the evolution of natural

mechanisms. Trans. Amer. Phil. Soc., 23: 75-481.
Grey, Marion. 1955. Notes on a collection of Bermuda deep-sea fishes.

Fieldiana: Zool., 37: 265-302.
1956. The distribution of fishes found below a depth of

2000 meters. Fieldiana: Zool., 36 (2) : 75-337.
KoEFOED, E. 1944. Pediculati from the "Michael Sars" North Atlantic

Deep-sea Expedition 1910. Rep. Sci. Res. "Michael Sars" Exped. IV.,

2(1): 11-18.
Lloyd, R. E. 1909a. A description of the deep-sea fish caught by the

R.I. M.S. Ship "Investigator" since the year 1900, with supposed evidence

of mutation in Malthopsis. Mem. Indian Mus., Calcutta, 2(3): 139-180.
. 1909b. Illustrations of the Zoolog)^ of the Royal Indian

Marine Survey Ship Investigator under the command of Commander

W. G. Beauchamp, R.I.M. Fishes, Part 10, Plates 44-50.
LiiTKEN, Chr. Fr. 1871. Oneirodes eschrichtii Ltk. en ny gronlandsk Tud-

sefisk. Oversigt over det Kongl. Danske Vidensk. Selsk. Forhandl., 1871:
56-74.
Murray, J., and J. Hjort. 1912. The Depths of the Ocean. London:

Macmillian and Co., Limited, xx + 821 pp.
NvBELiN, O. 1948. Fishes collected by the "Skagerak" Expedition in the

eastern Atlantic 1946. K. Vet. O. Vitterh. Samh. Handl., Ser. B, 5(16):

1-93.
PiETSCH, T. W. 1969. A remarkable new genus and species of deep-sea

anglerfish (family Oneirodidae) from off Guadalupe Island, Mexico.

Copeia, 1969: 365-369.
. 1972a. A review of the monotypic deep-sea anglerfish

family Centrophrynidae: taxonomy, distribution and osteology. Copeia,

1972: 17-47.

1972b. Ergebnisse der Forschungsreisen des FFS "Walther

Herwig" nach Siidamerika. XIX. Systematics and distribution of cera-
tioid fishes of the genus Dolopichthys (family Oneirodidae) with the
description of a new species. Arch. Fischereiwiss., 23 (1) : 1-28.

1974 ANGI.ERFISHES OF THE GENUS LOPHODOLOS 19

. h>74. Osicology and relationships ol deep-sea angleilishes

ot' the family Oneiiodidae witli a review ol the genus Oneirodes Liitken.
Bull. Nat. Hist. Mus. Los Angeles Co., Sti., 18: 1 11.'}.
Reg.\n, C. T. 1925. New ceratioid lishes troin the N. Atlautie, the Caril)-
beau Sea. and the Gulf of Panama, colleeted by the 'Dana." Ann. Mag.
Nat. Hist., Ser. 8, 8 {&Z) : 561-567.

. 1926. The pediculate lishes of the suborder ( .eratioidca.

Dana Oceanogr. Rep., 2,, 45 pp.
Regan, C. 1'., and E. Trewavas. 1932. Deep-sea anglerfishes (^Ceratioidea) .

Dana Rep., 2, 113 pp.

APR 2 1 1977

ARVARO

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

Miiseiiin of Comparative Zoology

us ISSN 0006-9698

Cambridge, Mass. 27 November 1974 Number 426

ASSOCIATION OF URSUS ARCTOS AND
ARCTODUS SIMUS (MAMMALIA: URSIDAE)
IN THE LATE PLEISTOCENE OF WYOMING

BjORN KURTEN^ AND ElAINE AnDERSON^

Abstract. The first substantiated association of Ursus arctos and Arctodus
simus (Mammalia: Ursidae) from a local fauna south of Alaska is reported
from Little Box Elder Cave, a late Pleistocene site in Converse County,
Wyoming. Ursus arctos, the grizzly or brown bear, entered the area from
Alaska at the end of the Wisconsin glaciation, and may have been a factor
in the extinction of Arctodus simus, the great short-faced bear.

INTRODUCTION

The late Pleistocene Carnivora of Little Box Elder Cave, Con-
verse County, Wyoming, were described by Anderson (1968),
who noted the presence of the grizzly or brown bear, Ursus arc-
tos L. The material consists of a number of loose teeth and foot
bones, most of which belong to a large form of this species. Ex-
tended comparison has shown, however, that at least one and
probably two specimens must be referred to a distinct species
and genus, the extinct short-faced bear, Arctodus simus (Cope).
This is the first substantiated record of association between these
two species of bears south of Alaska. The material is in the
University of Colorado Museum (UCM), Boulder.

MATERIAL

Ursus arctos L., Brown bear

UCM 22289, right Mi. This tooth belonged to a young in-
dividual and shows hardly any trace of wear. The posterointernal

^Museum of Zoology, N. Jarnvagsgatan 13, Helsinki, Finland
'730 Magnolia St., Denver, Colorado 80220

2 BREvioRA No. 426

part of the talonid has been lost. The remainder of the tooth is
well preserved and is similar to the Mi in Recent U . arctos from
Alaska except for its somewhat larger size ( Table 1 ) . The lower
carnassial of Arctodus simus, in possessing a powerful trenchant
trigonid, differs markedly from the specimen at hand, and is also
considerably larger, as shown in Table 1 .

UCM 49-iO and 22290, left and right M2. As in No. 22289,
these two teeth are quite unworn and probably belonged to the
same individual. The left tooth is much damaged, while the
right one is intact except for a missing piece of the protoconid,
and the loss of the anterior root. This root, however, is preserv^ed
in the left M2. As far as can be seen from the preservation, the
two teeth are mirror images of each other.

The occlusal surface is strongly marked, with large, well-
developed cusps delimited by furrows. Protoconid, metaconid
and entoconid are all duplicated; the two cusps in a pair are
subequal in size except for the protoconid, where the anterior
cusp is noticeably bigger. The posterior rim of the tooth forms
a small hypoconulid. A well-developed external cingulum curves
around the hind edge of the tooth; there are no cingula in front
or internally.

Despite its large size, approximating to the average in A.
simus, M2 has typical U. arctos characters. The second molar of
A. simus differs in being narrow posteriorly, in lacking an ex-
ternal cingulum, and in having a markedly inward slope to its
outer wall, as well as in various morphological details of the
occlusal surface.

Although these M2 are larger than those of present-day Alas-
kan Ursus arctos, late Pleistocene specimens of comparable size
are known from Alaska (Table 2). An analogous decrease in
size within U. arctos since late Pleistocene times has been docu-
mented in Europe (Kurten, 1959, 1968).

USM No. 52-73, first phalanx, probably from the manus.
The relatively small size of this bone, which has a length of
37 mm and measures 11.6 mm transversely in the middle, leads
us to regard it as most probably being U. arctos. It agrees in
size with those of present-day grizzly bears.

Arctodus simus (Cope), Short-faced bear

UCM 22288, left M^ This tooth belonged to an old indivi-
dual, and the four principal cusps have worn down to the same
level as the cuspules. The anteroexternal corner of the tooth, the

1974 ASSOCIATION OF URSUS ARCTOS & ARGTODUS SIMUS 3

inner and outer roots, and some other portions of the crown have
been lost; the anterior and posterior roots are preserved.

In spite of the damage, there can be no doubt about the char-
acteristic outline of the Arctodus M". It is approximately tri-
angular in shape, being broad in front and tapering rapidly
posteriad; the inner wall is straight but the outer wall has a
rounded bulge at the base of the metacone. In Ursus arctos, M"
is a relatively longer and narrower tooth, and is not similarly
tapered towards the hind end of the talon. The specimen
matches closely the M" in specimens of Arctodus simus from
Alaska with which it has been compared. Its size is close to the
average for A. simus from Potter Creek Cave, California (Kur-
ten, 1 967 ; see also Table 3 ) .

UCM No. 7-56, left pisiform. The morphological characters
and relatively slender build of this bone lead us to refer it to
A. simus. It compares closely with the specimen from Rancho
La Brea figured by Merriam and Stock (1925), and deviates in
various respects from a specimen of U. arctos of comparable
size; the last mentioned is conspicuously heavier in build, as
shown by the measurements (Table 4). In No. 7-56, the shaft
is slenderer and the distal boss much more flattened than in
U. arctos. The size of the specimen is about the same as in ^.
simus from Rancho La Brea and Frankstown Cave (see Kurten»
1967: 35, Table 14).

DISCUSSION

The bear fossils were found at various levels and locations in
the cave. As Anderson (1968) pointed out, there has been some
reworking of the unconsolidated deposit by rodents, especially
Neotoma cinerea. This probably accounts for the fact that the
two brovvn bear M2 lay at different levels. The age of the fauna
as a whole is late Wisconsin.

As far as we know there is no other locality south of the Fair-
banks District, Alaska, that shows an association between these
two species. Ursus arctos has been reported from the Rancho
La Brea tar pits, which have also yielded A. simus, but as shown
by Kurten ( 1 960 ) the only specimen definitely referable to the
former comes from the postglacial Pit 10, where the short-faced
bear is not present. The ursine bear of the main Rancholabrean
fauna is the black bear, Ursus americanus Pallas, of which a
very large form was present in North America during the Wis-
consin. Its large size has led to confusion with the grizzly bear.

'4 BREVioRA No. 426

We suggest that U. arctos entered western United States at the
end of the Wisconsin glaciation, presumably through the corridor
between the Cordilleran and Laurentide ice fields, from Alaska,
which it had colonized some time earlier. Once it had penetrated
south of the ice sheet it extended its range far beyond its limits
in historical times, as shown by finds in eastern Canada and the
United States { Guilday, 1 968 ) . The great short-faced bear may
have become extinct either as a result of competition with U .
arctos or because its prey became extinct, or for some other
reason; but the exact date of its extinction cannot yet be stated.

ACKNOWLEDGEMENTS

We would like to thank Peter Robinson, University of Colo-
rado Museum ; Richard Tedford, American Museum of Natural
History; and John L. Paradiso, Bird and Mammal Laboratories,
National Museum of Natural History, for letting us examine
specimens in their care. We are indebted to National Science
Foundation Grant No. GB 31287 to Professor Bryan Patterson
for aid in carrying out this and other work on North American
Pleistocene mammals.

ABBREVL\TIONS

The following abbreviations are used in the tables:
F : AM — Frick Collection, American Museum of Natural His-
tory, New York.

LBEC-UCM ^ Little Box Elder Ca%'e, University of Colorado

Museum, Boulder.
LJSNM — National Museum of Natural Histon, AVashington,

D. C.
M — Mean

N — Number in sample
O.R. — Observed Range
S.D. — Standard Deviation

REFERENCES

Anderson, E. 1968. Fauna of the Little Box Elder Cave, Converse County,
Wyoming. The Carnivora. Univ. Colorado Stud. Earth Sci., No. 6: 1-59.

GuiLDAv, J. E. 1968. Grizzly bears from eastern North America. Amer.
Midi. Nat., 79(1) : 247-250.

KuRTEN, B. 1959. Rates of evolution in fossil mammals. Cold Springs Har-
bor Symp. Quant. Biol., 24: 205-215.

1974 ASSOCIATION OF URSUS ARCTOS & ARCTODUS SIMUS 5

1960. A skull of the grizzly bear (Ursus arctos L.) from Pit

10, Rancho La Brea. Contrib. Sci. Los Angeles Co. Mus., 39: 1-7.

1967. Pleistocene bears of North America. 2. Genus Arctodus,

short-faced bears. Acta Zoologica Fennica, 117: 1-60.

1968. Pleistocene Mammals of Europe. London: Weidenfeld

and Nicholson, and Chicago: Aldine Press. 317 pp.
Merriam, J. C, AND C. Stock. 1925. Relationships and structure of the
short-faced bear, Arctotherium, from the Pleistocene of California. Publ.
Carnegie Inst. Washington, 347 (1) : 1-35.

Table 1. Measurement of Mi in Ursus arctos and Arctodus simus.

N O.R. M S.D.

Trigonid length
U. arctos

Recent, Alaska — USNM 40 14.3-17.0 15.90±0.09 0.60

LBEC UCM 22289 I - 18.0

A. simus

Pleistocene 18 21.8-26.1 23.01 ±0.33 1.42

Trigonid width
U. arctos
Recent, Alaska — USNM 40 8.2-11.1 9.94 ±0.11 0.68

LBEC UCM 22289 1 - U.O

A. simus

Pleistocene 20 14.1-16.8 15.50 ±0.16 0.73

Table 2. Measurements of M2 in Ursus arctos.

LBEC Pleistocene, Alaska

UCM 22290 F:AM A-200-6671
Length 30.6 32.3

Anterior width 19.5 19.5

Posterior width 20.0 20.8

Table 3. Measurements of M2 in Arctodus simus and Ursus arctos.

N OR. M S.D.

Length
A. simus

Pleistocene 27 33.3-41.4 37.60 ±0.4 2.1

' LBEC UCM 22288 1 - 35.5

U. arctos
Pleistocene, Alaska 1 - 45.0

Anterior width
A. simus
Pleistocene 27 21.3-25.8 23.66 ±0.23 1.20

LBEC UCM 22288 1 - ca. 22.5

U. arctos
Pleistocene. Alaska 1 - 24.0

6 BREVIORA No. 426

Table 4. Measurements of Pisiform Bone in Arctodus simus and Ursus arctos.

A. simus U. arctos

LBEC UCM 7-56 USNM 199252

Greatest length 56.0 54.0

Greatest proximal diameter 31.0 32.0

Distal boss, long diameter 34.3 30.7

Distal boss, short diameter 18.5 21.8

Least width of shaft 15.4 16.8

'M/i-' APR 2 11977

B R E V I O R-A

Museum of Comparative Zoology

us ISSN 000&-9698

Cambridge, Mass. 27 November 1974 Number 427

THE STRATIGRAPHY OF THE
PERMIAN WICHITA REDBEDS OF TEXAS^

Alfred Sherwood Romer'^

Abstract. A description is given of the topography of the Hmestones and
sandstones that form the formation boundaries between the six units com-
prising the continental redbeds of north central Texas; the results are
presented in two maps.

The Early Permian redbeds of Texas, those of the Clear Fork,
and even more notably those of the stiU earlier Wichita Group,
are of major importance in the history of vertebrates. These are
the oldest beds in which there is present an abundant fauna of
continental type. In earlier. Carboniferous deposits of various
areas have been found a very considerable number of amphibian
types, and even, in the late Carboniferous, early reptiles. But
almost without exception Carboniferous deposits containing tetra-
pod vertebrates represent coal-swamp conditions, and it is not
until we reach the Texas Wichita redbeds at the beginning of
Permian times that we find a truly continental fauna. Speci-
mens, to the number of several thousands, representing scores of
amphibian and early reptile types, have been collected in these
beds for nearly a century. It is clear that these beds, with more
than a thousand feet of deposits, represent a very considerable
period of time during which a fair amount of evolutionary
progress and faunal change took place. Farther to the south
arid southwest the Wichita beds are mainly marine in nature,
with identifiable limestones, and there competent stratigraphic
work has been done. But with the transition to continental beds
to the north and east the limestones fade out, and almost nothing

^This paper was in essentially completed form at the time of Professor
Romer's death in November, 1973. Miss Nelda Wright kindly finished the
task of preparing the manuscript and maps for publication. (Ed.)

^Museum of Comparative Zoology, Harvard University.

2 BREVIORA No. 427

has been done in the past to sort out the sequence of formations
in the redbeds portion of the Wichita.

In default of work here by the geologists, I decided a number
of years ago (1960) that although not a stratigrapher or proper
geologist, I myself must attempt to work out the sequence of
formations in the Wichita beds.

The task, at first, seemed almost hopeless. Except on the
fringes of the area, limestones, to serve as formation boundaries,
were almost nonexistent. Sandstones could be observed here and
there, but it seemed probable that these were channel sandstones
of limited extent. The one saving grace was that almost all of
the area concerned was oil-bearing, and that in consequence
thousands of well logs were available. In these logs, identifiable
limestone markers of late Carboniferous age could be located.
Assuming (hopefully) that deposition of sediments was fairly
uniform over the area concerned, it would be possible to lay out
a sequence of formations by calculating the distance to the sur-
face from such limestones and thus plot out a rough stratigraphic
sequence.

A further discouragement lay in the fact that for almost aU
of the area no topographic maps were available. Apart from
highway maps and blueprint land-ownership maps of the counties
concerned ( drawn up for the benefit of oil lease men ) , the only
sources available were Department of Agriculture air photos,
which show streams, hills and roads, but do not, of course, give
any indications of elevation.

All in all, the prospect was most discouraging. But as I began
work, I found that both nature and man rendered valuable aid.
( 1 ) As I said above, surface markers to distinguish formation
boundaries appeared to be lacking. This proved, however, not
to be the case. Upon study of the sandstones encountered, many
of them proved to be wide ranging, and could be followed for
a considerable distance cross countr)'. Further, in most cases
limestones that, to the southwest, were used as formation
boundary markers, were found to change s^radually to the north-
east into sandy limestones, then into "limey" sandstones and
straight sandstones, which could be traced across the entire area
concerned.^

^Had I read more carefully Cummins' last paper (1897) on the Wichita-
Albany problem, I would have seen my discovery of this condition to have
been anticipated by him. He states: "We found that a limestone in the
Albany Division . . . gradually changed in composition to a calcareous sandy
clay. . . . other limestone beds in the Albany division when traced to the
northeastward would gradually pass into sandstone."

1974 PERMIAN WICHITA REDBEDS 3

(2) Major aid came from another source. As noted above,
almost no topographic maps of the area were available when I
began to work. At about this time, however, an arrangement
was made between the Texas Water Development Board and
the Topographic Branch of the U. S. Geological Survey, to map
a larger area, including almost every bit of the Wichita redbeds
region, on a scale of 1 : 24,000. The work proceeded rapidly
and presently proofs and finally finished sheets of the whole
area became available. These were of inestimable value to me
— most notably in giving accurate elevations ( doing elevations by
aneroid in the highly variable weather conditions of the Texas
prairies is most unsatisfactory).

( 3 ) A final aid in this work came as a result of the decision
of the Texas Bureau of Economic Geology to prepare a geological
map of the State, at a scale of 1 : 250,000, under the direction
of Virgil Barnes. One of the first sheeets attempted was the
Sherman Sheet, along the north border of the State. The Cre-
taceous covers most of the territory, but much of the western
margin, in Montague County, lay in the Permian. Almost no
definite formation markers were available in this area, but it
was found (as I had found) that certain sandstone beds could
be traced for a considerable distance. These were followed out
by J. H. McGowen westward across Montague County and into
Clay County to the west. These sandstones were merely given
numbers on the published Sherman Sheet; I found, however,
that certain of them were identical with formation boundaries
that I had been following eastward. In almost every instance,
McGo wen's findings and mine coincided. It was a pleasure to
have my work independently confirmed and, in fact, in a few
areas in Clay and Montague Counties, I saved my weary feet
and accepted McGowen's findings in completing my course over
to the Cretaceous boundarv.

I owe thanks to a very considerable number of people and
institutions for aid during the course of this work. Notably, I
am deeply indebted to my wife, Ruth Hibbard Romer, who
accompanied me on almost all of my trips to the area, furnished
my transportation and day after day picked me up, footsore and
weary, after a long trek across the cow pastures. John Kay, con-
sulting geologist of Wichita Falls, who is an authority on the
geology of the Wichita region, aided throughout with encourage-
ment, advice, and specific data. The Gulf and Humble Oil
companies gave me access to their well log collections and to
unpublished maps, surface and subsurface, and the first-named

4 BREVIORA No. 427

company presented me with a large collection of duplicate well
logs. Robert Roth of Wichita Falls gave useful advice in Wichita
stratigraphy. Robert Craig, oil geologist of Olney, gave me the
use of a very valuable series of well logs of Young County. I
am indebted to Frank Gouin, oil geologist of Duncan, Okla-
homa, for interesting discussions of the Montague County beds.
Virgil Barnes aided greatly by making available to me Mc-
Gowen's tracings. Adolph H. Witte of Clay County, who has
done much work in archaeology and paleontology, gave much
helpful advice. The maps here published were drawn by Carol
Jones.

I cannot refrain from mentioning the late Fred B. Plum_mer,
of the University of Texas and the Bureau of Economic Geology,
who first interested me in the stratigraphy of these beds and who,
had he not died at an unseemly early age, would have been
deeply interested in the present work.

It is impossible in a short space to give thanks to the many
landowners who have allowed me to wander over their pastures.
My wife and I are most especially indebted to Mr. and Mrs.
G. F. Boone and L. D. Boone of Godwin Creek, whom we have
long cherished as valued friends, to James R. Parkey who has
given us ready access to various areas that he owns in the Little
Wichita country, and John Robinson of Archer City, ever hos-
pitable to "bone hunters."

I am much indebted to the National Science Foundation for
support of part of my earlier Texas work, for support of a final
trip to the Texas beds in 1973, and for publication of this paper.

WICHITA STRATIGRAPHY

The first student of the Wichita beds was W. F. Cummins.
Originally a frontier preacher, he was engaged by Cope to col-
lect fossil vertebrates in the Texas redbeds, then turned geologist,
served on the Texas Geological Survey during the few years of
its existence, and later became geologist for the Southern Pacific
Railroad. In his early work for the Texas Sur\^ey, Cummins
(1891) believed conditions to differ north and south of the
Brazos River. He established a Cisco Division as forming the
uppermost section of the Carboniferous in the northern area.
Included in the Cisco were the coal beds (which he lumped at
that time as "Coal number 7" and, as seen on his plate VII,
considered the top of the Cisco to lie not far above this coal).
The typical coals of this area are contained in the Harpers\ille

1974 PERMIAN WICHITA REDBEDS 5

Formation of most writers. In the northern region he believed
the Cisco to be directly overlain by the Wichita beds, which thus,
as later identified, begin with the Pueblo Formation (in which
are found the lowest redbeds in southeastern Archer County).
In this northern area he believed the top of the Wichita beds to
lie at a double limestone seen along the Big Wichita River a
few miles west of the Archer-Baylor county line (1891: 402).
This limestone is clearly the Bead Mountain Limestone, forming
the boundary between Belle Plains and Clyde formations. Cum-
mins' original Wichita thus included, in ascending order, the
Pueblo-Moran-Putnam-Admiral-Belle Plains formations of later
terminology; the Clyde Formation, later considered an integral
part of the Wichita, was in this discussion thought to be a lower
element of the Clear Fork.

Farther south, beyond the Brazos in Young and Stephens,
Throckmorton and Shackleford counties, Cummins found a dif-
ferent situation. Above the Cisco are formations that are mainly
marine in nature, which he did not realize were identical with
his Wichita beds to the north. He believed these beds, which he
termed the Albany Division, to be a terminal part of the Car-
boniferous intercalated between the Cisco and the Permian red-
beds. The upper boundary of the Albany beds (1891 : 404) he
believed to He between California Creek and the Clear Fork,
about on the Shackelford-Haskell County boundary. He thus
considered the Lueders as the top of his Albany beds, above
which lay the Clear Fork redbeds.

Two years later (1893, especially p. 223) Cummins began to
suspect that his Albany beds were merely a different facies of
the Wichita beds. And in 1897 he confirms this suspicion, and
definitely traces certain "Albany" beds northward into the
"Wichita" region with a transformation of their character from
marine to continental in nature. As a result, the term "Albany"
was abandoned and the pre-Clear Fork Permian beds were
termed Wichita — although some confusion remained as to
boundaries between Cisco and Wichita and between Wichita
and Clear Fork.

For many years little was added to our knowledge of these
beds. Adams (1903) and Gordon (1911 (with others), 1913)
confirmed Cummins' identification of the Wichita and Albany,
and Gordon reasonably concluded that in the northern area the
beds from the Bead Mountain Limestone to Lueders should be
included in the Wichita.

6 BREVIORA No. 427

A landmark in the history of the group was the publication
in 1922 of "Stratigraphy of the Pennsylvanian Formations of
North-Central Texas" by F. B. Plummer and R. C. Moore.
While their attention was centered on the late Carboniferous,
the Wichita formations were discussed as well. The beds which
Cummins considered to constitute his Cisco division were divided,
in ascending order, into the Graham, Thrifty, and Harpersville
formations (the last including the coal beds). Cummins con-
sidered all higher beds as part of his Wichita. But since at the
time of publication of Plummer and Moore's paper the Carbonif-
erous-Permian boundary was believed to be at a considerably
higher level, three further formations — Pueblo, Moran, and
Putnam — were included by them in the Cisco, and only the
formations lying above the Coleman Junction Limestone at the
top of the Putnam Formation — Admiral, Belle Plains, and
Clyde formations and, finally the Lueders Limestone — were
considered to constitute the Wichita Group.

Subsequent to the publication of Plummer and Moore's basic
work, the stratigraphy of the Cisco and Wichita has been dis-
cussed by a number of workers. For example, Sellards, in the
comprehensive "Geology of Texas" (1933), follows in general
Plummer and Moore, but since by that time it was generally
agreed that the Carboniferous-Permian boundary had been
placed too high in the section, the Moran and Putnam forma-
tions were included in the Wichita Group. In 1940, M. G.
Cheney, oil geologist and an able student of Texas geology, pro-
posed a radical change in treatment. Former "groups" became
"series"; the former formations became "groups" and were
subdivided into rather thin formations. Durino^ the years pre-
ceding this publication the invertebrate paleontologists had estab-
lished a sequence of marine Permian beds in West Texas, termed
the Wolfcamp and Leonard Series, the base of the Wolfcamp
being considered the base of the Permian. Chenev nrooosed
abandoning the established terms "Wichita" and "Clear Fork"
and substituting the West Texas local terminolos^v. The evidence
of foraminifera indicates that the base of the Wolfcamn can be
equated with a point in the Waldrip shales, somev/hat below
the top of Harpersville. Cheney solves this problem by abolishing
the Harpersville "series," the top levels being included in the
Pueblo, and the rest of the Harpers\ille being lumped with the
Thrifty. The foraminiferal evidence indicates equivalence of
the top of the Wolfcamp with about the middle Admiral. Cheney

1974 PERMIAN WICHITA REDBEDS 7

hence reduced the Admiral by half, adding the upper part of
the formation to the Belle Plains.

Moore returned to the Texas redbeds region in 1949 with the
study of the geology of the Permian in the Colorado River region.
He followed Cheney in part, by including the upper part of the
Harpersville in the Pueblo, and including the upper part of the
Admiral in the Belle Plains. However, he refused to raise the
"formations" to "series" level. Furthermore, he retained the
term "Wichita Group" for formations from the Pueblo Forma-
tion (expanded) to and including the Lueders, but parallels
Cheney in also noting "beds of Wolf camp age" and "beds of
Leonard (?) age" at the levels given by Cheney.

In this present attempt at interpreting the stratigraphy of the
Wichita beds, I have essentially follo^ved Plummer and Moore.
The finer subdivisions proposed by Cheney may be followed in
the marine section, but are impossible to sleuth out in the con-
tinental beds. Nor can the subdivision proposed by him within
the Harpersville and Admiral formations be readily followed
in the continental areas with which we are concerned. I have
adopted the base of the Pueblo as the base of the Wichita. This
is in accord with Cummins' original definition of the Wichita,
since the actual base of the redbeds type of deposit is at the
base of the Pueblo Formation. Although I am far from certain
that the base of the Wolfcamp of West Texas has any neces-
sary relation to the true Carboniferous-Permian boundary,
this equivalent is but slightly below the base of the Pueblo.
It is generally overlooked by invertebrate paleontologists that,
considering that the extent of the Permian was for a long time
(and still is) a rather vague and ill-defined matter, the real
point in question is not the base of the Permian but the top of
the Carboniferous, a matter for settlement by paleobotanists. But
both invertebrate and botanical evidence agree that the Permian
base is a short distance below the base of the Pueblo, and since
this exact point cannot be accurately determined in the conti-
nental beds, the slightly higher Saddle Creek Limestone, which
can be readily followed, seems a satisfactory point for Cisco-
Wichita division.

Methods. The results of my field work are shown on the three
accompanying maps, on which I have attempted to exhibit the
subdivision of the beds into six successive formations, from the
underlying Cisco beds of the Carboniferous up to the Clyde
Formation and the Lueders Limestones, which cap the Wichita

8 BREVIORA No. 427

and form the boundary with the overlying Clear Fork. The
formation boundaries, as traced, were at first entered on the
air photographs, later on the 1 : 24,000 topographic sheets. It is,
of course, impractical to publish them on this scale. Maps 2
and 3 are executed on a two-miles-to-the-inch scale, which will,
I think, be sufficient for future workers to locate the horizon of
their finds with reasonable accuracy.

The method followed was to pick up each successive limestone
used as a formation boundary where already known and map-
ped, in the southwestern part of the region, and then follow it
northward and eastward cross-country as it changed toward and
to the condition of a sandstone. In some areas a continuous
tracing was possible. Over much of the region, however, the
rolling prairie surface makes this impossible, and I have had
to seek out occasional small outcrops or detached slabs in the
pasture grass, much in the fashion of a "paper chase.^' Under
such conditions, of course, it was possible to stray from one
sandstone to another, above or below. But over most of the ter-
ritory there exists such a profusion of well logs that a check on
elevations above the underlying hmestones of the Cisco Group
was present as a corrective.

All the stratigraphic studies mentioned earlier have been made
in the region to the south of the true redbeds area; almost no
previous attempts at stratigraphic subdivisions of the continental
beds have been made. The sole exception was that in the 1920's,
a time at which it was believed that the Coleman Junction Lime-
stone represented the Carboniferous-Permian bondar\^, a recon-
naissance was made of the probable course of this horizon from
the point at which the limestone disappears in southwest Archer
County north and east to the Red River (Timms, 1928). Some
years ago (1958) in a general essay on the redbeds and their
fauna I included a rough sketch of the probable formation
boundaries in the redbeds area.

The general area to be considered is bounded on the north by
the Red River; to the west by the Clear Fork beds above the
Lueders Limestone, running north to south through Wilbarger,
Baylor and Throckmorton counties. To the east, in Montague
County, the Wichita beds disappear beneath the Cretaceous
deposits. To the south we reach the base of the Wichita beds
along a line somewhat south of the Jack County boundan'. To
the southwest the formations of the Wichita Group continue, but
gradually change from continental to marine beds - — that is,

(

1974 PERMIAN WICHITA REDBEDS 9

from "Wichita" type beds to sediments of "Albany" nature. South
of the Brazos River vertebrate fossils become scarce, and very
few have been found in the Wichita beds beyond the southern
boundary of Throckmorton County.

The geologic structure of the area is a simple one. The area
is in general a northern continuation of the Bend arch. In
eastern Young County and northward the beds dip to the north ;
west of this line, the dip is to the northwest (Hubbard and
Thompson, 1926). In the southern part of the region the dip
is on the order of 40-50 feet to the mile. Farther north the
dip decreases, and in the upper beds, found on the surface
toward the Red River, the beds are nearly horizontal. Near the
river, the deeper beds in certain areas have been strongly affected
by the east-west Electra arch and, farther east, by the Muenster
arch. Arch activity, however, appears to have ceased before
deposition of the surface beds here, and in general, these struc-
tures have had no effect on the surface stratigraphy. To the east,
in southern Montague County we encounter the margin of the
Fort Worth basin, with strong dips to the east and northeast in
the lower beds.

One tends to think of the change in the nature of the Wichita
beds as being a north-to-south shift from continental to marine.
Actually it seems that it is an east-to-west transition. The general
redbeds area appears to have been a lowland, with ( presumably )
high land to the east and a sea to the west. As is known from
well logs, the Wichita redbeds formations became mainly marine
west of a line extending from central Wilbarger County south
through central Baylor and Throckmorton counties. In the
eastern parts of these counties there are occasional persistent
limestones, but redbeds tend to dominate and almost no lime-
stones persist east of the east line of these counties.

As an aid to future workers who wish to check — or correct
— my findings, I herewith add some detail as to the nature of
my work on the various formation boundaries.

The Saddle Creek Limestone

As noted above, I consider the Pueblo Formation to be the
basal member of the Wichita group; and I consider the Saddle
Creek Limestone, at the top of the Harpersville, as furnishing a
close approximation to the Carboniferous-Permian boundary.

The Saddle Creek Limestone is well developed in the more
marine sections of the Wichita to the south, and can be followed

10 BREVIORA No. 427

north as far as the Clear Fork of the Brazos, not far south of
the Young County line. It can be traced into southwestern
Young County only with difficulty and with doubt. Plummer
and Moore identify it for a distance west of the Salt Fork south-
west of Newcastle, but it is probable that this is the somewhat
lower Belknap Limestone, as is also presumably the case of the
supposed Saddle Creek in this area marked on the 1937 Co-
operative geological map (Plummer and Fuqua, 1937) . Lee and
colleagues (1938; cf. Cheney, 1940: 91 and fig. 10) figure the
Saddle Creek, although with some doubt, at the head of Ratliff
Branch in southwestern Young County, Here the limestone,
feebly developed, is part of a thick sandstone layer that can be
readily followed to the north and east across Young County,
where it lies in proper relation to the underlying limestones in
the Harpersville.^ From the point mentioned above, the sand-
stone beds here accepted as the Saddle Creek equivalent turn
westward along the south margin of the valley of Gibbens Creek,
cross that creek and run northeastward along the north side of
this valley to reach a prominent bluff close to the Brazos and
directly west of Fort Belknap. The Saddle Creek Limestone
then turns west, and becoming less well marked, descends down
the west side of the valley of Postoak Creek and reaches a bluff
south of the Salt Fork at the mouth of Elm Creek. It continues
west south of Elm Creek, to disappear into the Salt Fork allu-
vium about a mile east of Proffitt. The Saddle Creek reappears
on the north bank, only obscurely west of the mouth of Paint
Creek (California Creek), but east of that creek capping Deer
Head Bluff north of the Salt Fork bottoms. East of this bluff
it turns northward west of Big Skid Creek and can be traced
with some difficulty eastward across the flat country' at the head
of this creek and then southward along a low ridge west of
Peveler Creek. Returning northward to cross this last creek, the
outcrop continues eastward along the hills north of Newcastle
to a prominent point about four miles northeast of Newcastle
and a mile west of Salt Creek. From this point a series of out-
liers extends northeastward toward Jean, but the main outcrop

^Galloway, in an interesting study of the Harpersville in subsurface (Gal-
loway and Brown, 1972) , gives a surface map on which the assumed Saddle
Creek Limestone is shown for several areas in Young and Jack counties.
Different areas indicated on this map, however, show the supposed Saddle
Creek at several different levels, ranging from that of mv assumed Saddle
Creek up to that of the Camp Colorado, nearly 200 feet higher.

1974 PERMIAN WICHITA REDBEDS 11

turns northward along the west margin of the Salt Creek valley,
descending to cross this creek a mile northeast of True cemetery.
The outcrop turns southeastward for two miles, swings north-
ward to cross Little Salt Creek, then southward and again north-
ward to obscurely circumnavigate a flat area east of Jean.
The outcrop turns south for about three miles, then north for
six miles to Farmer, at a level of about 11 50 feet, mainly follow-
ing the base of the hills west of the road leading from State
Highway 199 north to Farmer.

Southeast of this area, the country rises to the Loving region.
My well records for this area are sparse, but it seems probable
that there were several outliers of the Saddle Creek in this area,
the principal ones being at the Loving settlement and along a
ridge running eastward toward the county line. Galloway (Gal-
loway and Brown, 1972) considers these beds to lie within the
underlying Harpersville Formation, presumably because he gen-
erally places the Saddle Creek member at a higher level strati-
graphically than I do.

From Farmer the main outcrop runs eastward two miles along
a ridge between two tributaries of Brushy Creek, then westward
south of these tributaries to a point north of Farmer. North of
this tributary it runs eastward along a ridge, which becomes
prominently developed, for about three miles, with outliers on
Rattlesnake Mountain and Bare Mountain, and then turns
northward, only to turn westward up a further northern branch
of Brushy Creek. After crossing this branch near its head, the
Saddle Creek comes east again several miles to Spy Knob.
Thence the outcrop runs for some miles northwest, then north-
east, then southeast, in so doing outlining a semicircle around
the margins of the Prideaux structure (highly important in the
days of shallow oil production ) . After crossing the Windthorst-
Loving highway, the Saddle Creek outcrop (now in southeastern
Archer County) runs eastward along the southern margin of a
ridge for several miles, almost reaching the West Fork of the
Trinity River. It then returns westward north of this ridge and
then turns north and northwest, to subside to the level of the
West Fork near its crossing by the Windthorst-Loving road, at
about 1,000 feet.

We are now entering a wild region, where the West Fork and
its tributaries have cut deep valleys, capped by sandstones and
covered by scrub oak and tangles of vines, making a very com-
plicated pattern. As noted below, the main outcrop of the Saddle

12 BREvioRA No. 427

Creek extends eastward north of the West Fork along a general
line south of the north border of Jack County, with a general
elevation of about 1,000 feet close to the county boundary, but
somewhat higher farther south. To the south, beyond the West
Fork, are large areas of hills and plateaus, sandstone capped,
which lie at higher levels, and which, by such well-log evidence
as is available to me, indicate them to be extensive outliers of
the Saddle Creek. ^ The most westerly of major outliers of this
sort is one whose southwestern extremity is at Markley and ex-
tends northeast about five miles to a point south of the mouth
of Brushy Creek and runs eastward a similar distance alonsr the
north side of Plum Creek. Much larger is a tableland that oc-
cupies the area betv/een the valleys of Plum Creek and Cameron
Creek and extends from three to five miles south of the West
Fork and includes an area of 20 square miles or so. Farther
east a smaller outlier lies between Cameron Creek and Roberts
Prairie Branch and a final, still smaller, outlier is found east
of this branch. Farther southeast, it is probable that the top of
the Indian Hills attains the Saddle Creek level.

After crossing the W^est Fork, the main outcrop of the Saddle
Creek, as noted above, runs eastward, roughly parallel to the
Jack-Clay county boundary and some miles to the south. For
the first mile or so below the crossing- there is little evidence of
the presence of the Saddle Creek in the alluvial river bottom,
but east of the Jack County line it is visible as a sandstone low
down toward the river level. Its eastward course is a zig-zag
one, the outcrop running to the north up successive creek vallevs
and rising southward to bluffs north of the "West Fork. A mile
east of the x\ntelope-Jacksboro highway it extends a mile to the
north up the valley of Flat Creek, where its elevation drops
somewhat below 1,000 feet, and then returns southward to a
river bluff at 1 ,040 feet — an elevation that matches that of the
outlier south of the ri\'er. Four miles east of the highway cross-
ing, it runs north a short distance in a \'alley in the Mount Lebo
region, then returns south to cap a high river bluff at about
1 ,050 feet. A mile further east lies Lodge Creek, a major north-
ern tributary of the West Fork; the outcrop extends north up

^Galloway (Galloway and Brown, 1972) believes these sandstones to lie
within the Harpersville; but this belief is due to the fact that the outcrop
to the north, which he indicates as the Saddle Creek, is quite surely the
Camp Colorado, nearly 200 feet higher in the section.

1974 PERMIAN WICHITA REDBEDS 13

this valley to well toward the county line southwest of Shannon,
dropping below the 1,000-foot level in elevation. East of this
creek the West Fork tends to swing to the southeast, and the
main outcrop, continuing eastward, tends to leave the river,
althouo^h east of Lods^e Creek outliers form bluffs at about
1,050 to 1,080 feet. The Saddle Creek again extends well to
the north up Turkey Creek, next to the east, but beyond this
creek the outcrop turns eastward around the margins of the
creek valley, sending, however, a high ridge southward and then
westward to reach an elevation close to 1,100 feet. Next to the
east is Jones Creek, which the Saddle Creek ascends to Postoak
settlement. East of Postoak the Saddle Creek extends southward
several miles along a high but narrow ridge, bifurcate distally,
with an elevation now over 1,100 feet. East of this ridge the
Saddle Creek runs northward up the north fork of Crooked
Creek, to end in a "flat" about two miles in circumference, where
there are few exposures except in road margins. Descending
this creek branch, it runs about two miles east to ascend the east
branch of Crooked Creek to a deep valley north of Galliher
Mountain. East of this branch it runs southeast and east for
about four miles along the summit of gentle slopes, past Truce
church, rising as it goes, to reach a ridge at the southwest corner
of Montague County at an elevation of about 1,150 feet. It then
turns northward alon? a bluff for somewhat over two miles,
losing altitude, to en^er the southeast corner of Clay County at
about 1,090 feet. There are certainly outHers to the northeast
of this bluff, and I have mapped sandstone ledges here that are
probably Saddle Creek equivalents. From the southeast corner
of Clay County the Saddle Creek turns westward along the foot
of the hills south of Newport.

From this point eastward my subsurface data are not sufficient
for me to be certain of the position of the Saddle Creek. There
is certainly a sharp dip to the northeast, where we are entering
the Fort Worth basin. It appears to be represented in hiUs north
and northeast of Newport along the course of Big Sandy Creek
toward and to the Montague County line and on northeast
to Prairie Branch. Crossing this branch it appears to be con-
tinued by sandstones following the north shores of Lake Amon
G. Carter, and then following for some distance up the valleys
of Jones Creek and East Jones Creek, disappears under the
Cretaceous about four miles south of Bowie.

14 BREvioRA No. 427

Camp Colorado Limestone

The uppermost member of the Pueblo Formation is the Camp
Colorado Limestone, which separates the Pueblo from the
Moran Formation. It has long been known farther south, and
is rather incompletely shown on the southwestern part of the
geological map of Young County (Plummer and Fuqua, 1937),
running north close to the Throckmorton County line northward
toward Elm Creek. From west of Murray in southwestern Young
County, it runs northward about three miles along the west edge
of the Fish Creek drainage area, then turns back southwest for
two miles east of Dr\' Branch of Elm Creek, then traces north-
ward west of Dry Branch on one side or the other of the county
line. It follows the west side of Drv Branch almost to Elm
Creek, ending this course in a prominent bluff. It then turns
back south along gentle slopes east of Meyers Branch, which it
crosses about two miles south of Elm Creek. The Camp Colo-
rado is not exposed along its course down the west side of Meyers
Branch except at the foot of the bluff west of the branch close
to Elm Creek. A mile west of this point the Camp Colorado
can be seen at the bottom of the channels of Elm Creek and its
tributary Bush Knob Creek.

North of Elm Creek slopes are gentle, but occasional traces
of the Camp Colorado can be made out as it runs northeast-
ward, gaining slowly in elevation and for some distance lying
close to the state highway from Newcastle to Throckmorton. By
two miles east of the county line it can be traced along the slopes
of low hills north of this highway. It then turns northward along
a low bluff to disappear in the Brazos alluvium near the mouth
of Boggy Creek. During this segment of its course the Camp
Colorado is gradually losing its calcareous nature and is in
process of changing into a sandstone.

The Camp Colorado reappears on the east bank of the Brazos
a mile to the north, in a low bluff west of the mouth of Rabbit
Creek. It is obscure in crossing this creek, but east of this it
ascends up a small tributary of the creek to the divide between
Rabbit and Paint creeks, with a large outlier to the south. It
then runs about three miles to the northeast along the west slopes
of the Paint Creek valley, crosses this creek and swings east and
south to a prominent south-facing bluff on the Jeffries ranch.
Here it sharply reverses direction, and runs north and somewhat
east, descending to cross Salt Creek somewhat over a mile south
of Olney. East of Salt Creek it swings for a mile up the valley

1974 PERMIAN WICHITA REDBEDS 15

of Willow Pond Creek, then turns back southwest to run east-
ward along gentle slopes for two miles to Pleasant Valley church.
It then turns northward and somewhat eastward (poorly ex-
posed) for two miles to gain the east- west ridge separating the
Brazos drainage from that of the West Fork of the Trinity River.
It crosses to the north through a low spot in this ridge, but
outliers extend eastward along this ridge for about four and
one-half miles. The main outcrop turns west, not far from the
Young-Archer County line, to swing around the headwaters of
the South Fork of the West Fork of the Trinity River. It con-
tinues northeastward for about eight miles down the west side
of this fork, with conspicuous outliers on the east side of this
creek. Crossing the West Fork proper, it continues eastward on
the side of this small river, keeping at a level of about 1,050
feet not far from the creek for about 10 miles. Beyond this point
the West Branch is gradually descending and swinging to the
southeast and the Camp Colorado, keeping at roughly 1,050
feet, gradually diverges from the river, running some distance
up Waters Branch and Darnell Branch as it approaches the
Archer-Clay County line. It runs eastward north of Antelope
and here meets the westward end of a line Pss, traced by
McGowen for the Sherman Sheet of the Texas geological map
mentioned earlier. From this point eastward my tracing of the
Camp Colorado outcrop and McGowen's Pss coincide almost per-
fectly (except that I am doubtful of certain southern outliers of
his where, I think, the south-to-north dip of the beds is not fully
taken into account). The outcrop continues eastward close to
the Jack County-Clay County boundary, at an elevation close
to 1,050 feet. It dips northward up the valley of Flat Creek,
just east of Antelope, farther to the north up the valley of
Willow Creek, west of Shannon and again up a small valley
near that settlement. The outcrop continues east, at the top of
low south-facing hills, turning north up the valley of Turkey
Creek west of Prospect and, to a lesser degree, up a small branch
east of that settlement. It then runs south two and one-half
miles to a hiU two miles west of Postoak and then runs north-
east along the west slopes of Jones Creek for a half a dozen
miles. Thence it continues eastward in an irregular course,
again capping south-facing hills, for another half dozen miles,
entering the drainage of Big Sandy Creek north of Newport.
Near the Clay-Montague County line it turns west up the valley
of Prairie Branch; it then follows eastward down the north side

16 BREvioRA No. 427

of Prairie Branch to about the county line, then retreats north-
west up a branch of this creek toward Vashti before returning
eastward, and, after some miles, turning for some distance up
East Prairie Branch for about one and one-half miles. East of
this creek it runs eastward along bluffs well north of Lake Amon
G. Carter (with a deep "incision" for Trail Creek). West of
Briar Creek it swings northward for about four and one-half
miles to a point west of Bowie, and then, after returning' some
distance down the east bank of this creek, turns eastward to end
beneath the Cretaceous cover.

Sedwick Limestone

Sedwick limestone, being the upper element of the Moran
Formation is, again, well developed in the counties to the south-
west of the region with which we are here concerned. It is
shown, in somewhat incomplete fashion, on the 1937 Throck-
morton County map (Hornberger, 1937), running north and
somewhat east toward Elm Creek. I began tracing this lime-
stone at a point about two and one-half miles west of the Young-
Throckmorton County line, and about three miles south of Elm
Creek. The Sedwick here is following north a ridge between
Mevers Branch and an unnamed small creek to the west. With
a slight interruption the Sedwick follows this ridge to within
about half a mile of Elm Creek and then turns back southwest
to a crossing of this unnamed creek. I could not trace the
Sedwick down the even slopes west of this creek until, within
about a mile of Elm Creek, the limestone is seen on a low ridge.
The Sedwick then turns back southwest, east of Bush Knob
Creek, to cross that creek at about three and one-half miles
south of its mouth. Subsurface logs indicate that it again turns
northward, but I found no surface indication of it until it is
exposed in the bed of Elm Creek at a ranch road crossing some
miles to the northwest.

North of Elm Creek, in a fashion comparable to the Camp
Colorado a few miles to the east, indications of the limestone
gradually become apparent, and it gradually ascends the north
slopes of the Elm Creek Valley in a zig-zag fashion, until, about
a mile west of the county line, it crosses north out of the Elm
Creek drainage into that of small western tributaries of the
Brazos, along which it runs northward to Bogg)' Creek, east of
Elbert. In this stretch the Sedwick maintains its character as

1974 PERMIAN WICHITA REDBEDS 17

a somewhat sandy limestone, and is accompanied by a shale
layer containing Myalina. At Boggy Creek it turns westward,
and is traceable to a point south of Elbert. It is not exposed
north of the creek, although the Myalina bed is definitely present.
Two miles east of Elbert the Sedwick again becomes visible and
can be followed to the west for three miles to a point south of
Leopard Creek. For the next four miles north and northeast to
a bluff on the west bank of the Brazos, little is seen of the Sedwick
(now a calcareous sandstone), for a curious reason. A local
rancher, now deceased, had apparently become deranged from
his services in the First World War, and seems to have spent
most of the remainder of his life building beautiful stone walls
(which have no obvious function) and appears to have incorpo-
rated in them nearly all sandstones visible in the area.

The Sedwick appears at the base of the bluff mentioned above,
and then disappears into the Brazos bottoms. A mile to the
north, somewhat over a mile below the Spring Creek settlement,
the Sedwick is seen emerging along a low bluff. From this point
it runs eastward and northward, crossing Spring Creek and then
following the north side of Bitter Creek. This is farming country,
but the general course of the Sedwick can be followed from
slabs of calcareous sandstones seen here and there in the fields
and field margins. South of Bitter Creek are low hills, capped
by sandstones that are obviously Sedwick outliers. More im-
portant, well logs strongly indicate that the sandstones capping
the hills west of Padgett, several miles to the south, are also
Sedwick outliers.

The Sedwick crosses Bitter Creek about four miles east-north-
east of Spring Creek settlement and then turns south to become
clearly visible in slopes lying along the Olney-Spring Creek high-
way. Farther east the country is quite flat, exposures are rare,
and were it not for the aid of well logs it would have been
extremely difficult to follow this bed. The course is slighdy north
of east, into the northwestern end of the Salt Creek drainage, to a
point at the west end of the settled Olney area, then north past
the Lutheran church into Archer County. The course now runs
north along the west side of a narrow valley which is running
northward toward the Little Wichita River. East of this valley
there develops a large outlier bounded (except to the south)
by well-developed bluffs. The main outcrop follows the valley
northward to about four miles north of the Young-Archer County
line, then turns southwest, circling most of the headwaters of

18 BREVIORA No. 427

Mesquite Creek and the two Olney reservoirs. Following down
the west side of these reservoirs, the outcrop continues north close
to the paved north-south road (farm road 2178) for two and
one-half miles, then turns east along the low divide between
Mesquite Creek and the South Fork of the Little Wichita River
to the region of their junction. Here the outcrop is nearly lost
in the alluvium, but having crossed Cottonwood Creek, it runs
southeastward east of that creek ( with outliers to the south ) .
South of Bobcat Bluff the outcrop swings east and north to the
region of the former settlement of Anarene. We find here the
watershed between the West Fork of the Trinitv to the south and
creeks tending north to the Little Wichita. The divide is marked
by a west-east line of hills, and a long series of Sedwick outliers
runs eastward along them to ( and a bit beyond ) the Archer-Clay
County line. From Anarene the main outcrop (poorly indicated
for some distance) runs northeastward down the west side of
Onion Creek. The northern dip of the Sedwick and the gentle
gradient of the creeks running north to the Little Wichita are
almost equivalent, and the course of the Sedwick to the east,
all the way to Montague County, is a complicated one, the
outcrop dipping to the north in each creek valley, and returning
south between creeks. The outcrop follows Onion Creek north
to a point four miles southeast of Archer City, then retreats
southeast for three and one-half miles, onlv to turn north aeain,
to follow Little Onion Creek to within a mile of the Archer City-
Windthorst highway. After a short retreat to the south, it again
advances northward down the valley of West Little Postoak
Creek to a Doint north of the highway. It then turns south,
circling the Windthorst hill, and then (with faint outcrops for
the most part) follows a tortuous course — for a short distance
north down a tributary' of East Little Postoak Creek, and, further
to the east, a mile or more down the valley of that creek. East
of Windthorst I find the west termination of McGowen's trace
of his sandstone PI, and his line is thus that of the Sedwick east
of here.

The Sedwick sandstone now travels southeastward for half a
dozen miles, with a major outlier to the south, paralleling the
course of East Little Postoak Creek upward to its headwaters.
Turning east, it dips slis^htlv into the headwaters of Deer Creek,
and then runs eastward to the East Fork of the Little Wichita.
Here it performs a complicated course. The Sedwick Sandstone
runs north some miles down the west bank of the fork, then

1974 PERMIAN WICHITA REDBEDS 19

turns back west up Joy Creek past the settlement of that name;
then back down the valley of the Fork five more miles, and up
a western tributary to Midway School. Finally, after continuing
obscurely some distance farther down the west side of the Fork,
it turns southeast and ascends the east side of the East Fork
Valley for some eight miles, leaving to the west a substantial
outlier in the region of Friendship cemetery. From a point
about two and one-half miles northwest of Vashti, it turns
northward a short distance down Smith Creek, and then east
across the Clay-Montague County line. The main line of out-
crop now extends eastward across the headwaters of Belknap
Creek, a southern tributary of the Red River, dipping down to
the north along this creek and several of its tributaries before
reaching the cover of the Cretaceous about five miles north of
Bowie.

Coleman Junction Limestone

Capping the Putnam Formation and underlying the Admiral,
Coleman Junction Limestone is shown with a considerable degree
of accuracy on the geological map of Throckmorton County
(Hornberger, 1937), running north-northeast from a point a
short distance east of Throckmorton City to cross the Brazos
west of Spring Creek settlement a few miles south of the Baylor
County boundary. North of the river the Coleman Junction
runs eastward, gradually rising in elevation, barely enters Young
County at its northwest corner, and then continues northeast
into Archer County rising gently as it goes, crossing Spring Creek
and the headwaters of Bitter Creek to attain the level of the
plateau east of Megargel, and, turning north, is present on east-
ward-facing bluffs about five miles east of Megargel (in an oil
field that was highly important in the shallow oil days). The
Coleman Junction has long been known to extend this far north
and, as noted above, Timms in 1928 attempted to sleuth out
the general continuation of this unit north, east and north to
the Red River (cf. Sellards, 1933: fig. 11). Although this was
hastily done, detailed tracing shows that the line he plotted was
essentially correct. A sandy lime, turning gradually into sand-
stone, continues northeastward from this point, high up on the
west slopes of the valley of the South Fork of the Little Wichita
River, but gradually descending toward the left bank of the
South Fork, to reach after 14 miles the west side of the fork
about two and one-half miles west of Archer City, at the June-

20 BREvioRA No. 427

tion of state highway 25 and farm road 210. The outcrop turns
west and then disappears into the alluvium of the Middle Fork.
From this point east and northeast the line of the Coleman
Junction equivalent, as proved by well logs, follows the valley
of the Little Wichita east and northeast for more than 20 miles,
to the one-time settlement of Halsell, in Clay County. It is
possible that in part some of the lowest sandstones north of the
river are at the Coleman Junction level; on the other hand,
well logs prove the existence of a number of outliers of this
sandstone to the south of the main "line of march," extending
to the neighborhood of Archer City and to high buttes south-
west of that town; further outliers are present south of the
river west of Scotland.

At Halsell, exposures now concealed under the waters of Lake
Arrowhead show the Coleman Junction equivalent to reappear
on the east bank of the Little Wichita and run southwest, rising
gently, for several miles. Emerging above the lake level, it
swings east, along slopes following the north side of the Deer
Creek valley which develop into good bluffs north of Deer Creek
settlement. There I find the western end of McGowen's tracing
of his sandstone P4, which is thus the Coleman Junction equiva-
lent. Three miles west of Midway School the outcrop reaches
a high point at the Myers triangulation marker and enters the
drainage of the East Fork of the Little Wichita. The Coleman
Junction now follows down the west side of this valley in an
irregular northeasterly direction for about nine miles to a point
opposite Kola siding on the Fort Worth and Denver railroad,
and about six miles northeast of Blue Grove. From this point
the main line of Coleman Junction obviously turns east^vard
past Kola switch and on to the bluffs three miles north of
Bellevue and two to three miles west of the Clav-Montaffue
County line. However, the northward dip of the Coleman Junc-
tion and the gradient of the East Fork are almost identical. In
consequence the Coleman Junction equivalent extends north-
ward in a complicated fashion down the vallev of the East Fork
and an eastern branch of this fork extends as far north as Dick-
worsham switch. This was obviously mapped competently by
McGowen and I have not retraced this area.

At the bluffs north of Bellevue the Coleman Junction leaves
the East Fork drainas^e for that of Belknap Creek and turns
northward, gradually descending the western slopes of that valley
into western Montague County (with a number of outliers to

1974 PERMIAN WICHITA REDBEDS 21

the east) and finally, about three miles east of Ringgold, dis-
appears into the Belknap Creek alluvium and perhaps reaches
the Red River, only about two miles to the north.

I have done little work east of Belknap Creek. North of a
west-east line running past Belcherville and Nocona, bounded
on the east by the Cretaceous and north by the Red River, is
a triangular area which McGowen, I am told, found difficult
to interpret and which I, studying it in more superficial fashion,
found equally puzzling. A sandstone running eastward along
the line mentioned is essentially equivalent to the Coleman
Junction, and hence all of the area under consideration is pre-
sumably as high as the Admiral Formation, lying above the
Coleman Junction, and McGowen found here several sandstone
beds suggesting to him, I am told, that we are here dealing with
a deltaic condition. On the other hand, Frank Gouin has pointed
out to me that in the region of Lake Nocona there is a well-
developed anticline, presumably connected with the Muenster
arch, which brings relativelv low strata to the surface. On the
map I have merely indicated the lowest sandstones, which may
be roughly Coleman Junction equivalents.

Elm Creek Limestone

The top member of the Admiral Formation, Elm Creek Lime-
stone, appears on the 1937 map of Throckmorton County,
running north-northeast from the neighborhood of Throckmorton
to the Baylor County line not far west of the Brazos. This
limestone has not previously been mapped further north. Enter-
ing Baylor County, this limestone is present in a river bluff across
the river from Round Timber settlement, and is visible in a
similar bluff east of the river near Round Timber. In between,
however, the limestone follows a verv circuitons course. It turns
westward, gradually descending in elevation along the branches
of WafTon Creek, and finnllv reaches t^e brpid aHuvial valley
of the Brazos River at the foot of a bhiff phont two miles north-
west of Round Timber. Across the river, at t^e month of a small
creek two miles north of Round Timber, t^e l^mes+^one is seen
on the north bank of the "Rrazos. The co-'T'^f'^' fmr^-i this point
north and east toward Westover is flat anrt-Vnltnr-^l land, but
occasional exposures, mainlv in his^hwav di+r^ps ^how t^e Elm
Creek to follow a circular course, abon^ t^'^o miVs no^th from
the river, then abont three mi^es east an-^ i-^^-n V»orV southwest
toward Round Timber — the limestone p-pininqr some elevation

22 BREvioRA No. 427

and becoming more readily traceable in this last part of its
circuit.

For several miles east of Round Timber the ground is covered
by river sands and the Elm Creek Limestone is not visible.
Beyond this sandy area, however, the limestone can be followed
(although with some difficulty) northward a bit west of the
Bavlor-Archer Countv line to reach the west side of Briar Creek,
about four miles northeast of Westover and just west of the
county line. From here the limestone runs (rather obscurely)
northeast, west of Briar Creek and then, a mile or so north of
the Seymour-Archer City highway, turns west and southwest
into the valley of Godwin Creek. Here the situation is a con-
fusing one. The Elm Creek is here a double limestone, and the
dip of the beds is almost exactly equivalent of the slope down-
ward to Godwin Creek, so that the two beds, prominently ex-
posed, form a confusing pattern. The two beds gradually reach
the creek level about four miles southwest of their first appear-
ance in the eastern slopes, and then run north, poorly exposed,
to cross the Little Wichita River above its junction with Godwin
Creek. North of the river the limestone is better exposed, and
gradually ascends the slopes, and crosses Slippery- Creek about
five miles south of Dundee.

A mile or so east of this creek the limestone disappears and
(contrary to the usual condition in the Wichita beds) has no
immediate sandstone continuation. However, well logs clearly
show that the bed continues east at the foot of the bluffs south
of Black Flat. East of that settlement the stratum, as shown by
the subsurface, is continued along the north side of the valley
of Plum Creek (locally termed Rattlesnake Canyon). However,
beyond this point, five miles south of Mankins, the bed dis-
appears into the flat prairies of the Holliday Creek valley and
for the next six miles can only be traced by well logs, until a
sandstone at an appropriate elevation appears in the Hull-Silk
oilfield three miles south of Holliday. This runs eastward for
five miles, forms a conspicuous bluff, and then turns north to
disappear into the Holliday Creek alluvium.

Beyond this point the main outcrop is to be found only north
of Holliday Creek and, farther on, north of the Big Wichita
River. However, to the northeast there is a verv extensive series
of outliers, covering much of northern Clay County. Along the
divide between the Big ^Vichita and Little Wichita rivers is a
scattered series of outliers, with elevations somewhat o\'er 1,000

1974 PERMIAN WICHITA REDBEDS 23

feet, from the northeast corner of Archer County and the south-
east corner of Wichita County into the western margin of Clay
County, just east of the Wichita Falls-Henrietta highway and
railroad, where the sandstone is present on a low hill at about
1,030 feet.

This marks the beginning of a large series of outliers covering
much of northern Clay County. The beds here are much affected
by the Electra arch structure, but with one conspicuous excep-
tion (mentioned later) this structure had become inactive by
the time of deposition of the Elm Creek equivalent, and the
beds are almost horizontal, lacking the northern dip seen farther
south; for the most part the sandstones, which I believe equiva-
lent to the Elm Creek, average about 950 feet above sea level.
Except along the Big Wichita River there are few bluffs, and
exposures are far from continuous along the gently rounded
hills of the region. The major outlier is one covering the higher
ground extending northeastward past Dean, Petrolia, and Byers.
From the southwest corner, at the county line, its borders can be
followed northward and then eastward around the vallev of Duck
Creek, eastward and then northwestward to the region of Dean,
following the upper slopes of the valley of Turkey Creek. After
running eastward for nearly ten miles, the outcrop turns north-
west, to circle about the Petrolia oilfield just southeast of that
town. The outcrop runs eastward again for four miles before
turning northward again, to run along the upper slopes of small
creeks running eastward into the Red River. There are further
small outliers along the high ground east of Petrolia, the last of
this series only a short distance west of the Stanfield community.
The east side of the main outlier can be traced as far north as
Byers. The bed, however, appears to continue about two miles
north of this town, and then swings sharply southwestward, east
of the Big Wichita River. Exposures generally close to the 950-
foot level can be followed along this course for about 14 miles,
to a point two and one-half miles NNW of Dean. Here the
supposed Elm Creek Sandstone equivalent, as well as beds above
and below, are turned up almost vertically, turn sharply to the
northwest and disappear into the Big Wichita alluvium. Subsur-
face maps show the presence here of a marked syncline, pre-
sumably related to the Electra arch structure but representing
an "adjustment" that took place at a much later date than
formation of the arch structure. Two miles southwest of this
area, the presumed Elm Creek Sandstone appears again east of

24 BREvioRA No. 427

the Big Wichita and, running south close to the county line,
reaches the hill mentioned above where the circuit of this major
outlier was begun.

The main outcrop of the Elm Creek member, as determined
by well logs, runs northeastward to Wichita Falls north of Holli-
day Creek, but is visible only in a few places north of Lake
Wichita and south of Allendale. Returning westward south of the
Big Wichita, it is well exposed for most of the way west for ten
miles, when it disappears into the river alluvium. East of Iowa
Park it appears north of the river, but there are only occasional
exposures to plot its course eastward, south of Sheppard Air
Force Base and the municipal airport, then on eastward north
of the Big Wichita, past Friberg School and onward past Thorn-
berry in Clay County to a point south of Charlie. East of this
point the sequence is interrupted by the course of a former
channel of the Red River but farther to the east, between the
Red River and the Big Wichita, Pumpkin Ridge forms a con-
spicuous outlier. Excellent subsurface logs are present for this
northernmost part of Clay County, and it is clear that the Elm
Creek Sandstone turns northward, west of the old river channel
and then west along the Red Ri\^er bluffs (where possible ex-
posures are largely covered by soil ) . Coming west into Wichita
County, this member dips a bit southward into the valley of
Gilbert Creek and a southern branch of this creek, and then
vanishes into the Red River bottoms.

Bead Mountain Limestone

Bead Mountain Limestone, forming the boundary, has long
been known to run northeast across Baylor County, and part
of its course is shown on the 1937 cooperative map of that
county (Garrett, 1937) and on the similar map of Wichita
County. Locally it has been termed the Rendham Limestone
in Baylor Countv and, farther north, the Beaverburk Limestone.
In contrast to all lower members, it can be traced as a limestone
all the way to the Red River. In southern Baylor County, it
crosses the Brazos River about eight miles south of Seymour
and, rising to the east, crosses Deep Creek and then forms the
summit of east-facing bluffs as it runs northward on the west
side of the Godwin Creek \'allev east of the former Endand
settlement and the England cemetery. It crosses Daggett Creek
near its head and then swings eastward for some miles (not

1974 PERMIAN WICHITA REDBEDS 25

clearly seen) and becomes exposed in bluffs south of the Little
Wichita River. Turning west, it descends to cross the Little
Wichita as a limestone ledge about two miles east of Fulda
station on the Wichita Valley Railroad. Turning eastward it
can be readily followed for some miles and then, more obscurely,
it can be seen to cross the Wichita Falls-Seymour railroad and
highway just east of the Baylor-Archer County line. It now
turns northward, presently forming a conspicuous bluff which, in
an outlier, forms the southern margin of the dam of the Diversion
Reservoir on the Big Wichita River. The limestone turns west
up the south side of the river, and, since the dip of the beds
and the slope of southern tributaries of the river are almost
identical, has an intricate pattern. The outcrop runs southward
up the valley to two small creeks west of the dam and then,
west of the county line, strikes the valley of Brushy Creek up
which it runs almost to the height of land and the Wichita Falls-
Seymour railroad and highway. It then descends again north to
the river bluffs, but three miles farther west encounters Boggy
Creek, up which the Bead Mountain extends for about two and
one-half miles. Beyond Boggy Creek the limestone reaches the
river level about a mile west of the bridge leading from Fulda
to "Sweetly Cruz" camp. North of the river the limestone
descends to the Diversion Lake dam, keeping (as would be
expected) close to the lake level. Below the dam the Bead
Mountain runs to the northeast (Fischer, 1937) along the bluffs
north of the Big Wichita, for some six miles, then turns west
to descend into the Beaver Creek vaUey, crossing that creek
about two miles east of the Wilbarger County line. Its course
from this point east up onto and along the ridge north of
Beaver Creek and the Big Wichita, and then back south of
Beaver Creek, to a point southeast of Fowlkes Station on the
Fort Worth and Denver railroad, is shown on the 1937 coopera-
tive map of Wichita County. Until this present study it was
unknown beyond a point north of Beaver Creek about six miles
west of Iowa Park. I have, however, been able to trace it north
to the Red River. In contrast to its strength farther west, the
Bead Mountain here is thin and sandy in nature. The country
between this point and the Red River is flat, with few exposures,
but through occasional small exposures, mainly in road cuts, I
have been able to plot its general course, northward and then
eastward around the headwaters of North Buffalo Creek, Lost
Creek and Stevens Creek, then over a low divide to follow the

26 BREvioRA No. 427

north side of the Gilbert Creek valley northeast nearly to Burk-
burnett. The deeper beds here are much disturbed in relation
to the Electra arch, but this structure appears to have become
inacti\'e by the time of deposition of the Bead Mountain, and
the surface beds here are nearly horizontal. For a short distance,
near Burkburnett, no exposures of the Bead Mountain Limestone
are seen, but turning west, it is occasionally visible in the slopes
south of Wildhorse Creek, which it crosses about two miles
northeast of Clara. It then attains the south bluff of the Red
River, where it is clearly visible in the cuts of two roads which
descend to the river bottoms northeast of Clara. It descends to
the west, and reaches the level of the Red River alluvium north
and a short distance west of Clara.

Leuders Limestone

The Leuders, now generally regarded as a formation, has
long been recognized as the top of the Wichita beds, separating
them from the Clear Fork. I have not studied the Leuders in
detail. Several members are shown in the 1937 cooperative map
of Bavlor County, crossing the Brazos in the "canyon" of that
ri\'er below Seymour and running north past Mavbelle and the
Kemp Lake dam. I do not 1 now of any detailed mapping of
the Leuders in Wilbarger County; this limestone senes crosses
Beaver Creek in the central part of the county and then, as
stated by Wrather (1917) trends northeast toward Harrold. It
appears to be represented by sandy lim.estones farther northeast,
along the lower course of China Creek, toward the Red River.

REFERENCES CITED

Adams, G. I. 1903. Strati.^aphic relations of the Red Beds to the Carbon-
iferous and Permian in northern Texas. Bull. Geol. Soc. Amer., 14: 191-
200.

Armstrong, J. M. 1937. Geologic map of Jack County, Texas, revised. Univ.
Texas, Bur. Econ. Geol.

Barnes, V. 1967. Geologic atlas of Texas, Sherman sheet, Univ. Texas,
Bur. Econ. Geol.

Cheney, M. G. 1940. Geology of north-central Texas. Bull. Amer. Assoc.
Petrol. Geol., 24: 65-118.

Cummins, W. F. 1891. Report on the geology of northwestern Texas. 2nd
Ann. Rept. Geol. Surv. Texas: 359-552.

. 1893. Notes on the geology of northwest Texas. 4th Ann.

Rept. Geol. Surv. Texas: 177-238.

1974 PERMIAN WICHITA REDBEDS 27
1897. Texas Permian. Trans. Texas Acad. Sci., 2: 93-98.

Fischer, R. W. 1937. Geologic map of Wichita County, Texas (revised) .

Univ. Texas, Bur. Econ. Geol.
Galloway, W. E., and L. F. Brown, Jr. 1972. Depositional systems and

shelf-slope relationships in Upper Pennsylvanian rocks, north-central

Texas. Rept. Invest. No. 75, Univ. Texas Bur. Econ. Geol., 62 pp.
Garrett, M. M., A. M. Lloyd and G. E. Laskey. 1937. Geologic map of

Baylor County, Texas, revised. Univ. Texas, Bur. Econ. Geol.
Gordon, C. H. 1913. Geology and underground waters of the Wichita

region, north-central Texas. U.S. Geol. Surv., Water-Supply Pap., 317:

1-88.
Gordon, C. H., G. H. Girty and D. White. 1911. The Wichita formation

of northern Texas. Jour. Geol., 19: 110-134.
Hornberger, J., Jr. 1937. Geologic map of Throckmorton County, Texas,

revised. Univ. Texas, Bur. Econ. Geol.
Hubbard, W. E., and W. C. Thompson. 1926. The geology and oil fields

of Archer County, Texas. Bull. Amer. Assoc. Petrol. Geol., 10: 457-481.
Lee, W., C. O. Nickell, J. S. Williams, and L. G. Henbest. 1938. Strati-
graphic and paleontologic studies of the Pennsylvanian and Permian

rocks in north-central Texas. Publ. Univ. Texas, No. 3801: 1-252.
Moore, R. C. 1949. Rocks of Permian (?) age in the Colorado River

valley, north-central Texas. U.S. Geol. Surv., Oil and Gas Invest.,

Prelim. Map 80, 2 sheets.
Plummer, F. B., and F. B. Fuqua. 1937. Geologic map of Young County,

Texas, revised. Ufiiv. Texas, Bur. Econ. Geol.
, AND R. C. Moore. 1922. Stratigraphy of the Pennsylvanian

formations of north-central Texas. Bull. Univ. Texas, No. 2132: 1-237.
Romer, A. S. 1958. The Texas Permian Redbeds and their vertebrate

fauna. In Studies on Fossil Vertebrates, Essays presented to D. M. S.

Watson (T. S. Westoll, ed.) . London: Athlone Press, pp. 157-179.
Sellards, E. H. 1933. The pre-Paleozoic and Paleozoic systems in Texas.

In The Geology of Texas. Vol. 1 (Stratigraphy) , pt. 1. Bull. Univ.

Texas, No. 3232: 15-238.
TiMMS, V. E. 1928. Cisco-Wichita contact in northern Texas and southern

Oklahoma. Map, Roxana Petroleum Corporation.
Wrather, W. E. 1917. Notes on the Texas Permian. Bull. Southwestern

Assoc. Petrol. Geol., 1: 93-106.

BREVIORA

No. 427

|99*W

WILBARGER

WISE

33* H

Map. 1. General area of north-central Texas showing where members of
the Wichita redbeds are exposed.

1974

PERMIAN WICHITA REDBEDS

tV =

- s

° i

lal

= = 2

u O o

s « t)

9 t? r'^

= 11

b - E

> c ^

> •:: ,j<

■^ < £

No. 427

PERMIAN WICHITA REDBEDS

o -^ t ^

•p O U -

" "g .. "

H o 5 =

^ o ^ S

a, u '^ ^

<= , c

»i 0.0 »

s ^ !2

'G CJ aj "^

S - o "

" s a

"3 " § s

~ 3 e

y- O O

2 5: s

c ea U -d §

° _ S- c

'^ « £ i.

! I s 1 1

o <: 2 '" "

"S . s o g

Si = £ B S

a >2 3 : «

1 - ' £
o I ^

/V/' Mhn APR 2 11977

B R E V I 0 B^-M'

'Museum of Comparative Zoology

us ISSN 0006 9698

Cambridge, Mass. 27 November 1974 Number 428

A DESCRIPTION OF THE VERTEBRAL COLUMN

OF ERYOPS BASED ON THE NOTES AND

DRAWINGS OF A. S. ROMER

James M. Moulton^

Abstract. This paper includes an illustrated description of the vertebral
column and ribs of Eryops megacephalus Cope, based principally on notes
and drawings prepared by A. S. Romer. The paper examines closely re-
gional variation in the column. The descriptions, originally written of the
Eryops now mounted in the Museum of Comparative Zoology (MCZ 1539) ,
are amplified by reference to other specimens. The paper includes data on
growth stages and regional variation in the vertebral column and ribs, which
will be useful in interpretation of Eryops postcranial remains.

INTRODUCTION

This publication was to have been based on collaborative
work with Alfred S. Romer, but his untimely death on Novem-
ber 5, 1973 prevented this. Fortunately, his notes and drawings
on the postcranial anatomy of Eryops have been available to me
and are here incorporated; his handwritten descriptive working
notes are only slightly modified. The paper presents a general-
ized description of the vertebral column of Eryops, and drawings
of a set of presacral and postsacral ribs. The principal concern
in preparing this material has been that Professor Romer's ob-
servations should be available to paleontologists. To Professor
Romer's observations, I have added others which appear to be
of interest.

Gregory (1951, I: 253) called Eryops "the best known" of
all rhachitomous labrinthodonts ; Williston (1914) called it "the
most famous' of the Temnospondyli. But despite the detailed
descriptions of various parts — skull ( Sawin, 1 94 1 ) , teeth ( Stick-

^Department of Biology, Bowdoin College, Brunswick, Maine 04011.

2 BREVIORA No. 428

ler, 1899), forelimb (Miner, 1925), ilio-sacral attachment f Ol-
son, 1936a) — no account of the vertebral column as a whole
is available.

In familiarizing myself with Eryops material, I gratefully
acknowledge the help of discussions with Ernest E. Williams,
Nelda Wright, Robert L. Carroll, Thomas S. Parsons, John R.
Bolt, Keith S. Thomson, Bryan Patterson and Bobb Schaeffer,
and to Carroll, Patterson, Wilhams and Wright I extend thanks
for critical reading of all or of large portions of my manuscript.
I appreciate the opportunity to study specimens in the following
institutions : the Redpath Museum of McGill University with Dr.
Carroll, the Cleveland Museum of Natural History (CMNH)
through David H. Dunkle, the Field Museum of Natural History
(FMNH) through Dr. Bolt, the Peabody Museum of Yale Uni-
versity through Dr. Thomson, the American Museum of Natural
History (AMNH) ) through Eugene S. Gaffney, and the Pratt
Museum of Amherst College through Walter P. Coombs; and
I was aided by valued correspondence with several of those
mentioned above and also with Robert E. DeMar, Everett C.
Olson, A. L. Panchen, F. R. Parrington and Peter P. \"aughn.
A loan of Eryops avinoffi material from the Clex^land Museum
is gratefully acknowledged.

The staff of the Museum of Comparative Zoology, and espe-
cially Professors A. W. Crompton and Parish Jenkins, Jr., Direc-
tor and Associate Curator of Vertebrate Paleontology, have been
very generous with their hospitality and have made the Museum
a most rewarding place to work during spring term of 1973-74.
I am indebted for travel and research funds to Bowdoin College.

Eryops material has been described from the Carboniferous
and Permian of an area extending from New Mexico to Prince
Edward Island fLangston, 1953, 1963; Olson and Vaughn,
1970), the bulk of it from the lower Permian of Texas where
it is the common large form (Romer, 1958). Both the geological
range occupied by Eryops and the length of time it survi\'ed are
grounds for suspecting that more than one Eryops species existed
(Williston, 1914; Romer, 1943, 1947, 1952). But in the ab-
sence of a sound anatomical basis for separating species (Romer,
1947, 1952; E. C. Olson, personal communication), the bulk
of Eryops material from the Permian is now generally assigned
to Eryops megacephalus Cope, 1877. Appreciation of the extent
of speciation in Eryops must await a distinction between specific
differences and those due to growth and accidents of preserva-

1974 VERTEBRAL COLUMN OF ERYOPS 3

tion. Recognized as a distinct species, however, is Eryops avinoffi
(Romer) from the Pennsylvanian of West Virginia and lower
Permian of Pennsylvanian (Romer, 1952; Vaughn, 1958).
Photographs of its dorsal vertebrae have been published (Mur-
phy, 1971).

It is to Cope then that we are indebted for the original de-
scription of Eryops from Texas Permian material collected by
Jacob Boll, his friend and collector (Cope, 1877; Osbom, 1931 :
486), and himself a recognized scientist (see e.g. Broili, 1899:
61) and practicing geologist. Bom in Canton Aargau, Switzer-
land on May 29, 1828, Boll died alone of appendicitis in a tent
on the Pease River near its confluence with the Red River in
Texas on September 29, 1880 (A. S. Romer, personal com-
munication), lamented by his friend Cope (1884). Eryops
material was given a prominent place in Cope's collection (Os-
born, 1931: frontispiece; 587) and figured frequently in his
publications. Cope's paleontological collections, purchased for
the American Museum of Natural History {idem, Chapter 6),
included materials Boll had collected. One specimen, AMNH
4183, from which I believe Cope's most frequently reproduced
figures of vertebrae were drawn (see, for example, Cope and
Matthew, 1915), is still accompanied by Boll's penciled, signed
field label dated 1-12-80 from the North Fork of the Little
Wichita River, which, together with the Big Wichita, Boll ex-
plored scientifically for over six months from December, 1879
(Boll, 1880). While studying this material in the American
Museum collections on March 28, 1974, I happened to turn
over the old field label, and there was a penciled poem, also
signed 'Boll', which read as follows:

"Nun wirst du mit noch manchen andern
Zum Sitze des Professors wandern.
Geistreich denkend wird er dich erwecken,
Aus deinen Triimmem dich zusammensetzen.
Der Nachwelt wird er dann erzaehlen,
Wie du gebaut in deinen Zahnen,
Wie du gelebt und wie verschwunden,
Benennen dich und was gefunden."

For help in transcription, I am indebted to B. Werscheck of the
American Museum of Natural History.

Cope's publications dealing significantly with the vertebral
column of Eryops appeared in the years 1877, 1880 (a,b), 1881,
1882, 1884, and 1890, a number of them repeating the same

4 ' BREvioRA No. 428

left lateral and ventral views of portions of the vertebral column
which first appeared in 1880 (Cope, 1880b); some of Cope's
discussions of rhachitomous vertebrae (1878a,b; 1897; 1898)
omitted them, but they finally appeared in Cope and Matthew
(1915). Later diagrams of Eryops vertebrae or of generalized
rhachitomous vertebrae, often drawn to emphasize particular
points, are seldom more convincing than those Cope drew 'from
life'.

Cope (1880a,b; 1881), Broili (1899), Branson (1905), Case
(1911, 1915), WiUiston (1918), Watson (1919), Olson
(1936b), Rockwell, Evans and Pheasant (1938), Romer (1947,
1966), Gregory (1951), Panchen (1967, Parrington (1967),
Thomson and Bossy (1970), and Williams (1959) collectively
provide a description of the Eryops vertebral column and its
evolution, often with special attention to typical dorsal vertebrae.
The papers of Cope (1880b) and Case (1911) provide the most
complete accounts. Further, a paper on another rhachitome,
Edops (Romer and Witter, 1942), makes several points about
the vertebrae of Eryops and provides a measure of differentiation
within the rhachitomes. A photograph of Eryops caudal verte-
brae from the MCZ mount (MCZ 1539) has been published
(Romer and Witter, 1941) with a description of dermal scales
(see also Williston, 1915); caudal vertebrae have also been
illustrated by Cope (1890). Diagrams of Eryops and other
rhachitomous vertebrae are generally shown in lateral view; it
is not easy to comprehend the three-dimensional form without
the actual specimen in hand. The deficiency of anterior and
posterior views is corrected by several of Romer's figures in the
present paper. Anterior views of dorsal vertebrae are provided
by Broili (1899) and Rockwell et al (1938). Branson (1905)
and Cope (Cope and Matthew, 1915) show the atlas in anterior
view, while Cope (idem) and Olson (1936b) show side views
of atlas and axis, articulated and disarticulated respectively;
Cope [idem) shows a somewhat distorted atlas (AMNH 4183)
articulated with the axis in anterior view. Photographs of
mounted Eryops skeletons have been published (Miner, 1926;
Romer, 1 943 ) , as well as drawings of the entire skeleton ( Case,
1911; Gregory, 1951).

An illustrated description of the whole vertebral column and
ribs had long been planned by Romer (1943, 1947, 1958). His
drawings with others showing particular points are here pre-
sented with a description prepared largely from his handwritten

1974

VERTEBRAL COLUMN OF ERYOPS

'^t.'^f-.'^" ■w^"","?'

^.....^'

ife

Plate L Mounted skeleton of Eryops, MCZ 1539 (from Roraer, 1943)
X 1/21.

notes. Observations on variations in size and form in Eryops
vertebrae are also included. A future study should focus on
vertebral growth in Eryops megacephalus, a matter of consider-
able interest only touched on in the present paper. This paper
examines variation in structure throughout the vertebral colurnn,
and reconstructs the probable structure in life of the vertebral
column of Eryops from the dissociated jumble of neural arches,
pleuro- and intercentra to which the vertebral column of Eryops
and other rhachitomes is often reduced in the fossil state.

The Eryops mount in the Museum of Comparative Zoology
(Plate I), the "most perfect (skeleton) yet discovered" (Romer,
1943), is a not quite full-grown animal (A. S. Romer, personal
communication). Vertebrae of larger size and more massive
construction than those in the mount are not uncommon in the
collections I have studied. The MCZ mount is however com-
parable in size to similar mounts in the Cleveland, Field, Pea-
body, American and Pratt Museums collected over a considerable
span of years, suggesting that full-grown (or larger species) of
Eryops for some reason lent themselves less well to preservation
or were rarer than smaller animals. The specimen in the Pratt
Museum, from Geraldine, Texas, is probably the youngest of the
mounts studied; it is somewhat smaller than the MCZ mount
which measures over the tops of the neural spines 187.5 cm
muzzle to tail tip, with a presacral vertebral column of 71.9 cm
and a postsacral length of 80.6 cm. The skull measures 35 cm.

From well-preserved Eryops material such as that which fur-
nished the mounts for the MCZ and Pratt Museum, Romer
(personal communication) was able to "make outlines of the
whole set of vertebrae, clear to the tip of the tail, and each rib";

6 BREVIORA No. 428

drawings from those outlines illustrate this article. Complete
tails and even complete presacral series of Eryops vertebrae have
not been common finds, and understandably controversy has
arisen over tail lengths and vertebral numbers. The MCZ mount
is taken to be correct until better information is available; it
displays 22 presacral vertebrae, two less than the primitive num-
ber (Romer, 1947; Vaughn, 1971), and 37 postsacral vertebrae,
a total, with the single sacral, of 60. The paired proatlas atop
the bisected atlas is well shown in its correct relationships in the
Field and Pratt mounts ( Fig. 1 ) . Presacral-postsacral counts of
five other Eryops mounts are: 22 — 44, 21 —51, 22 — 30, 22
— incomplete postsacral series, and 22 — 46.

With Case (1915) we are inclined to believe that the bifur-
cated caudal spines in Eryops provided dorsal accommodation
for tendons, which in primitive forms are the chief support of
the axial column (Olson, 1936b) ; the Eryops arrangement sug-
gests a tail of reasonable length which may have been held off
the ground. Tail length in Eryops should be resolved because
it is of significance in understanding locomotion. Former esti-
mates have varied from Cope's of a medium-length tail (1880b)
to a stump nearly coccygeal (1884), the latter seconded with
some reservation by Case (1915), to Williston's admission of
ignorance and his drawing of Eryops with its tail concealed by
vegetation (Williston, 1914). Romer's orthometric linear unit
(Panchen, 1966) has not been applied to Eryops in estimating
a length for the tail.

The following descriptions unless otherwise stated are based
on vertebrae in the MCZ and Pratt Museum mounts of Eryops.

PRESACRAL VERTEBRAE BEHIND THE AXIS
(DORSAL VERTEBRAE)

(Figs. 1-4; 9 I; 10; 11; measurements in Table 1)

Each vertebra consists of four ossifications : neural arch, paired
pleurocentra behind the neural arch and a single intercentrum
ahead and below. The neural arch terminates dorsally in a
neural spine that, for an amphibian, is of considerable height.
In a mid-dorsal, the height of the spine above a line through the
center of the zygapophyses is 56 mm, when the vertebral length
is 35 mm, a ratio of 1.6. Spine height increases to 73 mm in
the last presacral, and the height-length ratio approaches 2.
There is a gradual decrease in spine height anteriorly — it being

1974

VERTEBRAL COLUMN OF ERYOPS

o
Ah

4-1

> cj

2 X

i-i

BREVIORA

No. 428

!-(
>

-a,

ID
4-1

l-H

a.
u

1974

VERTEBRAL COLUMN OF ERYOPS

>
o

■i-i

Pi
u
u

BREVIORA

No. 428

1974

VERTEBRAL COLUMN OF ERYOPS

o
•o

■0

BREVIORA

No. 428

1>
X

4-1

(L)

o
c

4-1

1974

VERTEBRAL COLUMN OF ERYOPS

w
u

3
u

(L)

1—1

4-1

in
O

-a

BREVIORA

No. 428

S-i
(U

-4-1

a
S

O
u

■l-l

(4
u

Vi
CO

o
a,

S-I

3
bo

1974 VERTEBRAL COLUMN OF ERYOPS 15

57 mm in vertebra 5 — and a rather sharp decrease associated
with transition to the skull, it being 44 mm at the axis.

In reasonably mature specimens, the tops of the spines become
expanded, subcircular and rugose; they surely lay in the dermis.
The appearance in some cases is of dermal ossifications fused to
the spine tips (Fig. 14 B), but there is no evidence of separate
ossifications. Expanded spine tips may be lacking in young
specimens. The width of the spine shaft is about 2/3 of the
anteroposterior dimension, although sometimes the neural spines
are considerably more flattened than this. The spines often
assume a diamond form in cross-section with lateral ridges in
the upper part which expand into the sides of the dorsal rugosity.

Minor spines, spurs and flanges are not uncommon on neural
spines and elsewhere (Fig. 9 I; two spines on a neural spine,
AMNH 4183 ; spine on transverse process of vertebra 18, AMNH
4280; flange on spine of postsacral 10, MCZ 1539). Some of
these may be artifacts of preservation, as is undoubtedly the
flattening observed in some neural spines. A remarkable flexi-
bility of Eryops skeletal material either shortly after death be-
cause of drying cartilage (see p. 22) or changes during preserva-
tion is suggested by the twisted neural arches and spines one not
infrequently encounters in collections (Fig. 9 D; sacral vertebra
of MCZ 2669, for example) .

The upper part of the neural spine is keeled both anteriorly
and posteriorly. In the lower part of the spine, the keel bifurcates
into two divergent ridges which pass into the zygapophyses ven-
trally. Secondary ridges may be present within the groove en-
closed by the ridge pairs. Both grooves tend to become reduced
in depth in very large vertebrae. The anterior groove may extend
more than halfway up the spine, more so in the anterior part of
the vertebral column than posteriorly. In the last presacrals, the
anterior groove is limited to 1 /3 of the spine height and becomes
relatively shallow. The point of bifurcation of the ridges at the
top of the grooves is often recognizable in side view by a marked
angularity in the contour of the spine, and the spine shaft is
broadest between these points. The posterior groove deepens
ventraUy into a deep pit between the posterior zygaphophyses.

The zygaphophyses are of the normal primitive tetrapod type
and are readily comparable with, for example, those of many
pelycosaurs in size, contours, inclination and relative position.
As usual in labyrinthodonts and pelycosaurs, but in contrast to

BREVIORA

No. 428

Figure 9. Based mainly on Eryops MCZ 1539 and 1883, all X .5. (A)
Atlas and axis with their intercentra, in anterior view, proatlas removed.

(B) Eryops occipital region, atlas, axis and right proatlas, anterior at top.

(C) Eryops occipital region from below showing anterior intercentra. (D)
Eryops axis MCZ 1883, anterior view. (E) Eryops axis MCZ 1883, right
lateral view, anterior flange reconstructed. (F) Eryops axis MCZ 1883 in
dorsal view. (G) Eryops atlas and proatlas, left elements from medial side,

(H) Eryops vertebra 4, posterior (above) and anterior views. (I) Eryops
vertebra 6, posterior (1.) and anterior views.

1974

VERTEBRAL COLUMN OF ERYOPS

Figure 10. All X .5. (A) Eryops vertebra 13, posterior (1.) showing
position of notochord and anterior views. (B) Eryops vertebra 13, posterior
views, with (1.) and without reconstructed cartilages surrounding bony cen-
tra. (C) Eryops vertebra 21, posterior view. (D) MCZ 1828, left view, show-
ing matrix (dark stippling) occupying position postulated for cartilage about
centra of presacral vertebrae. (E) Reconstruction of two dorsal vertebrae
showing cartilage reconstructed about centra and rib head. (F) Eryops
vertebrae 23 and 24, right view, showing facets for rib articulation (large
stippling) .

BREVIORA

No. 428

•>-- *-t^T^-?r^Cw^r^

mmne^fi^vmi j.iy>w *i.wniyy w.T!W>piwiyi?f t'^ywr**^

■''-^^^

bJO

•-H

^-^

■J

I = 2

5 X

c5 si

£ I

f3 bb

'-I o

S 3

I <y -

1 tM "

1974 VERTEBRAL COLUMN OF ERYOPS 19

cotylosaurs, the zygaphophyses are situated close together with
but a short interval between the medial surfaces of the com-
ponents of each pair. The typical dorsal zygapophyses are tilted
so that the posterior zygaphophyses face about 45° out from the
median plane and 45° up from the horizontal plane, and diverge
30° laterally from the median plane; the anterior ones face 45°
in, 45° down and diverge 30° laterally. The angles of the zyga-
pophyses vary somewhat throughout the length of the column,
the posterior zygapophyses tending to be nearer the horizontal
and smaller anteriorly than posteriorly. The face of the posterior
zygapophyses is quite flat throughout, the anterior concave.

Continuing to consider a mid-dorsal vertebra, below the level
of the zygapophyses the neural arch divides into two pedicels;
between them these form a well-defined roof and lateral walls
for the neural canal, which is subcircular in outline, as the center
of the floor is unossified. The degree of closure of the pedicels
below the neural canal, however, is a function of age and size
or both (Fig. 11). One can demonstrate sacra (MCZ 2604,
4305) and a caudal vertebra (MCZ 3316, Fig. 11) with a
completely ossified neural canal, and a whole series of dorsal
vertebrae in which it is nearly closed ventrally (MCZ 3316,
Fig. 11). Where the floor is unossified, cartilage probably
formed a ventral apex to the neural arch between intercentrum
and pleurocentrum in life.

Laterally, the surface between the anterior and posterior zyga-
pophyses is smooth, but there is a depression, usually rather
shallow, behind and below anterior zygapophysis. At about
the midpoint of the length of this depression a ridge develops
that swings down and back into the dorsal edge of the transverse
process, presumably associated with the passage of a segmental
blood vessel.

The anterior and posterior margins of adjacent vertebrae,
below the zygapophyses, form the posterior and anterior margins
respectively for the intervertebral gaps that afforded exit for the
spinal nerves. These margins do not, however, form ventral
boundaries for the gaps.

The posterior surface of the neural arch on each side, from
the level of the neural canal floor down over the pedicel, in-
cludes a very large unfinished area which faces as much inward
and downward as backward. It is subquadrate in form, but
rounded in the dorsolateral margin. This surface corresponds

20 BREvioRA No. 428

to that on the anterior surface of the pleurocentrum and is
articulated with the anterior face of that element, although
obviously with an intervening thickness of cartilage. The rough-
ened anterior face of the pedicel, continuous with the posterior
face at the ventral edge, is much smaller and irregularly shaped.
The upper portion, adjacent to the spinal canal, is subcircular
with a pronounced convex mass of bone projecting backward
and inward. The more ventral portion of this surface slants
downward and outward, narrowing rapidly, becoming concave
rather than convex, and twisting so as to face a little inward.
This surface matches the posterior face of the next anterior
pleurocentrum to a moderate degree and undoubtedly apposed
it; there must have been a considerable thickness of cartilage
between the two.

The transverse process is rather variably developed. It is
typically wedge-shaped in section and at the distal articular sur-
face broad above, narrower below. Typically, the dorsal margin
arises in a ridge projecting laterally beyond the surface of the
arch pedicel. It faces backward and downward so that the
articular surface in a mid-dorsal vertebra faces back about 40°
and about 30° downward, in anterior vertebrae more directly
laterally.

In a mid-dorsal, the articular surface for the rib extends down-
ward to form the most ventral part of the arch ossified; typical
anterior vertebrae are similar. Posteriorly the articular area
becomes reduced to the dorsal part of the articulation. In more
anterior dorsals, there are two distinct portions : ( 1 ) a broader
oval dorsal area meeting the tubercle; (2) a thinner ventral
extension. Posteriorly, the ventral part disappears and the upper
part becomes thin; anteriorly the upper part remains thick and
the ventral part tends to thicken as well, until the articular sur-
face becomes a unit.

The measurements of Eryops dorsal vertebrae presented in
Table 1 are based on AMNH 4280, which includes a set of
dorsal vertebrae to which definite numbers can be assigned, and
MCZ 1539, the mounted specimen. From the information pro-
vided by these two specimens, it has been possible to estimate
the position of isolated Eryops presacral vertebrae through the
size ranges most abundant in collections I have studied. Meas-
urements of isolated Eryops vertebrae have been published by
Cope (1877, 1878a,b) and Case (1911).

1974

VERTEBRAL COLUMN OF ERYOPS

i-i

V
V

o
1*5

3
tn

C/3

•5

.a.

.1^

■*-» .^— V

♦J «*-' c«

c« CT5

i« O w
qj ~^ rt

•^H CO

u be a;

apop
(MCZ

h? C r"

F— ^

1—1

1— 1

lf5

C^T

I-H

Oiig

•<f

-"f

''f

■"t^

"*<

Tfl

■^

'"^

pine

39)

v'^ 2

^ O 5

Oj "^

■" 03 TS

2 rs3

CT>

Tt^

»-H

l-H

»o

Dis

from
rticula

of neu
(MC;

»r5

l>.

t>.

i:^

•

»-H

CTJ

— . W

SlitH

t«

3 £ 2

ophysi
i 1539)

htof n
above
h ante]

xf5

ITJ

t-~

!>•

bO D •"

K&-

cj XS 3

Great

widt

f artic

Cvj

i>-

''^^

^-H

ical

:ula

00
CM

nclinati
om vert
rib artic

«-*^

o
O

^z
:§

i-H

1-H

TfH

'^l

csr

^'^

a;th oi
ular fa
ransve

*^ HI

«J •" 3

c^r

csr

-1 J30

ertebra
umber

tTi

'tf

i«

l-H

»o

©^

l-H

r—l

l-H

P-H

l-H

c^r

> G

22 BREVIORA No. 428

THE INTERCENTRA OF PRESACRAL VERTEBRAE

(Figs. 1-4; 9 I; 10; 12; 13 A-C)

The dorsal intercentra are of the usual rhachitomous type,
being crescents as seen in anterior and posterior view, convex
side down. They are wedge-shaped in side view, apex upward.
Concavities on their external surfaces may mark the paths of
blood vessels. The inferior surfaces tend to descend as flanges
anteriorly and posteriorly, least so in the posterior dorsals. A flat
longitudinal ridge tends to develop mid-ventrally, best seen in the
dorsal region. The surface may be notched posteriorly at the
area of rib capitulum articulation. This is not well seen in young
individuals and may be lacking in fairly large animals. It is most
emphasized anteriorly in the presacral column, and at the sacrum
(Fig. 12).

The anterior, posterior and dorsal surfaces are rough and un-
finished, and presumably were continued in cartilage. The dorsal
notch is a rounded longitudinal depression, occupied in life by
the notochord and surrounding tissues. Four hummocks of bone,
two fore and two aft, are arranged on either side of the depres-
sion and may represent centers of ossification ( Fig. 1 3 A ) . These
hummocks show with varying clarity, sometimes are completely
obscured, and are illustrated as ridges by Broili (1899). Seen in
side view the anterior pair of hummocks is slightly more ventral
than the posterior in dorsal intercentra; the posterior hummocks
are closer to the top of the intercentra.

Cartilage, with which the intercentrum was continuous, may
have surrounded the notochord in life (Romer, 1947), but no
ring intercentra have been found. Coossification of the pleuro-
centra occurs below the neural canal (Fig. 13 E), above the
notochord (MCZ 2622 and 1652). Such a coossified piece may
in turn coossify with the intercentrum to form a type of ring
centrum in which all three elements are distinguishable (MCZ
2604 and 2562). A completely coossified vertebra has also been
studied (FMNH UR745). Such remains are perhaps the best
evidence of a vertebral column of ossified pieces embedded in a
matrix of cartilage in Eryops.

Intercentra that were broken during life are rarely found. Two
dorsal intercentra have been found (MCZ 2621, 4306; Fig. 13
B, G), which I think were so broken; a third (MCZ 4305) is
cracked diagonally on the dorsal surface. Each break is at an
angle clockwise to the anteroposterior axis (2621, 8°; 4306, 30°;

1974

VERTEBRAL COLUMN OF ERYOPS

Figure 12. The presacral and sacral (23) intercentra of Eryops in ventral
view, anterior uppermost, X .6.

BREVIORA

No. 428

Figure 13. All X .5. (A) Eryops presacral intercentrum showing paired
protuberances, anterior uppermost. (B) MCZ 2621, presacral intercentrum,
ventral view, anterior at top, showing inclination of healed break. (C) MCZ
4306, intercentrum, as in (B) . (D) MCZ 4307, a left pleurocentrum in
anterior, lateral, posterior and medial views. (E) MCZ 2591, anterior view
of coossified pleurocentra. (F) MCZ 4325, left and right sacral rib central
articulations in ventromedial views. (G) MCZ 2085, right sacral rib central
articulations in ventromedial view. (H) MCZ 2621, right sacral rib central
articulations in ventromedial views. (I) Sacral vertebra and right sacral rib
of Eryops, right view, pleurocentra not shown.

1974 VERTEBRAL COLUMN OF ERYOPS 25

4305, 30°). That these breaks occurred in younger animals is
evidenced by the small size of one intercentrum (MCZ 4306)
and the appearance of extensive growth after healing in the
other (MCZ 2621).

A fragment of the atlas intercentrum still clings to the left
element of the atlas in AMNH 4183 (omitted by Cope and
Matthew, 1915: pi. 12).

THE PLEUROCENTRA OF PRESACRAL VERTEBRAE

(Figs. 1^; 9 I; 10; 13 D, E)

The pleurocentra are paired ossifications, the centers for which
are situated dorsal to the notochord rather close to the midline.
Study of articulations of components of the vertebrae indicate,
however, that they must have been situated in pleurocentral
cartilages of much larger size. Such cartilages would have ap-
peared rhomboidal in side view, their longer sides articulating
anterodorsally with the arch of the same vertebra, anteroventrally
with their own intercentrum, posterodorsally with the next pos-
terior neural arch, and posteroventrally with the next posterior
intercentrum.

Their contours indicate that the ossified pleurocentral elements
came close to the ventral margin of the column but did not reach
it externally; restoration of the cartilage suggests that the car-
tilaginous pleiirocentra probably did not gain contact with each
other ventrally ( Fig. 1 0 B ) . Dorsally, however, they were ob-
viously in broad contact beneath the spinal cord; occasional
coossifications in old specimens would suggest that the cartilages
may have been continuous below the floor of the neural canal.
The conjoined elements would have given in end view the ap-
pearance of an inverted crescent with the two horns closely
approximated ventrally. The cartilaginous pleurocentra could
have closely approximated those seen in ossified form in Tri-
merorhachis.

' The paired centers of ossification of the pleurocentra appear
to have been situated far dorsally where there is a globular mass
of bone from which ossification proceeded slowly toward the
ventral part of the element. The pleurocentra appear to be
feebly ossified, and much of their surface area is unfinished in
aU but very old specimens. The more anterior pleurocentra are
in general less ossified, and far anterior ones are almost unknown
(see also Branson, 1905). A fifth pleurocentrum in the MCZ

26 BREvioRA No. 428

mount is finished on almost none of its surface, a fourth is a tiny
nubbin on one side only and coossified with the arch, and there
are no traces in material known to me of pleurocentra 1 and 2.

Exceptionally the two pleurocentra may abut medially, as they
do in sacral vertebrae in two mature specimens (MCZ 2669 and
4305 ) . There are cases in which the pleurocentrum has coossi-
fied with the neural arch, as on one side in two different sacra
(MCZ 4305, 2604), and cases of coossification with the inter-
centrum behind ( FMNH 60 ) , or at one level with intercentrum
and at another with neural arch (MCZ 1387), or with both in
the same vertebra (FMNH UR745). Such cases are suggestive
of a continuum of cartilage, the vertebral pieces embedded in it,
similar to what Parrington has proposed.

The pleurocentra are likely to abut in the caudal region ( MCZ
1787 and 2634), even to the point of occluding the notochordal
canal (Fig. 15 F). The anterodorsal face of a pleurocentrum,
that which articulates with the neural arch of its vertebra, is
nearly flat and forms essentially a quadrant of a circle with a
curved margin laterally and ventrally. In life this surface faced
somewhat up and out as well as anterior and was apposed to
the neural arch, although separated by at least a film of cartilage
from it. The posterior surface is irregular, convex above, and
apposed to but rather far from the anterior margin of the neural
arch. The medial and posterior surfaces present a continuous,
rough, curving form.

The external surface is in great measure finished. It is wedge-
shaped in external view, narrow above, broadening and then
tapering below. The margins curve up sharply anteriorly and
posteriorly so that the pleurocentrum is externally concave in
section; the curved margins are best defined above. The groove
between the margins conveyed a spinal nerve. It narrows dor-
sally and at the very top turns anteriorly above the anterior
articular surface to blend smoothly into the lower wall of the
neural canal. The constant mismatch between the large surfaces
on the neural arches for articulation with pleurocentra and cor-
responding anterior articular facets of the pleurocentra collected
at the same time and place is a measure of the extent of cartilage
beyond the borders of the ossified pleurocentra.

THE ATLAS-AXIS COMPLEX
(Figs. 1; 9 A-G; 14; PL I)

The neural arch of the atlas is highly specialized. The two

1974 VERTEBRAL COLUMN OF ERYOPS 27

sides may be separate (MCZ 1883) or coossified (AMNH 4183;
Case, 1911). In the former case, each side consists of a stout
pedicel and slender half arch and neural spine directed dorso-
posteriorly. The pedicel is wedge-shaped with two broad articular
surfaces, anteroventral and posteroventral. The anterior surface
is for articulation with an exoccipital; the posterior is finished
above (MCZ 1883), rough below where it articulated with the
intercentrum of the axis. Each articular surface is a quadrant
of a circle with a common straight ventral margin. The posterior
surface is somewhat concave, not flat as usual. Internally there
is a well-marked curved area for the side wall of the neural canal.
At the base of the spine on each side is a flat tubercle, a well-
defined anterior zygapophysis to seat the proatlas. Each half-
spine is a thin rod, posteriorly and dorsally directed close to the
axis spine. A tubercle or slight flange on the lower edge of the
half -spine rested in life on the corresponding anterior zygapophy-
sis of the axis.

The atlas intercentrum, seldom preserved, appears to ossify
late. That associated with Sawin's (1941) specimen is a very
flat crescent, with the outline of a slight notochordal space above,
and the anterior edge with a central depression. There is only
one pair of mounds, and the back surface is unexceptional. The
front is subdivided into two articular areas facing rather laterally
as well as anterodorsally, and obviously covered with much car-
tilage in life. The intercentra of atlas and axis have no capitular
facets.

Each proatlas is a small neural arch, the short neural spine
slanting back and upwards, its tip being irregularly rugose. At
its base is an articular facet for the atlas tubercle. The anterior
limb defines the upper edge of the foramen for the first spinal
nerve and appears to barely touch the exoccipital region of the
skull above and lateral to the foramen magnum; there is no
formed facet.

There was undoubtedly restricted motion of the head, in the
absence of a neck ; the atlas-skull joint probably acted as a dorso-
ventral hinge.

The axis neural arch is in many respects an ordinary one
(Fig. 10). The neural spine is however elongated anteroposteri-
orly. The spine slants backward and then angles up in its longer
dorsal portion, relative to a plane through the zygapophyses.
The spine is wedge-shaped in frontal section, and is generally
thicker posteriorly than anteriorly. There is a variable but gen-

BREVIORA

No. 428

r- a

be C

00
00

"H o

'"'

2 "rt

r-*

c\r

a; o

J Q

^' M

o '

^ ^

-^^

c^ X

-4—1

^^ y'^V

- S^

1/3

JZ O

be K

?►

i^'

B ^

rt ^^

-C ^

rv ^

4— >

"be

"J ^

;-i

P ffi

" z

W-I >^

u <i

!-(

•i-H

O 0^

G
as

k] ^

on CO

(-1

1/3

^^ ^

.-H -H

CU (>;

t« o

i>*

csr

'><

•

(A) A
83, 324

CM
O

ffi

5 "^

c
o

-1^

.^rvi

■4— >

• •»

i^ o

(>J

1974 VERTEBRAL COLUMN OF ERYOPS 29

erally prominent angle posteriorly, toward which a ridge is di-
rected on either side from the widest point of the dorsal surface.
Development at the front of the neural spine is very variable,
depending in part on preservation. It is likely that a well-devel-
oped thin flange occurred on the front of the axis spine for a
median ligament to the occiput. The anterior zygapophyses are
much reduced to small flattened areas for articulation with the
arch of the atlas. The pedicel and transverse processes are not
specialized. The intercentrum of the axis is flat-bottomed, broad,
and the posterior end has a rounded projecting keel. The axis
is the most anterior vertebra to bear a rib. Constancy of form
of the axis spine is illustrated in Figure 14 A.

VERTEBRA FOUR

Figs. 1; 9 H; PI. I)

This vertebra, with its specialized neural spine and anterior
zygapophyses, makes up for restrictions in movement at the
occiput. The posterior zygapophyses are normal, but the
anterior ones are greatly expanded and nearly horizontal, thus
permitting freedom of motion in the horizontal plane, together
with some rotation vertically. The spine is much reduced ( Case,
1911; Romer, 1943), a fact apparently not revealed by Cope's
material. The spines of vertebrae 3 and 5 are inclined toward
each other above that of 4; they are therefore distinctive. Their
neural spines like that of the axis are somewhat elongate antero-
posteriorly, and their facing edges are thinned, suggesting a spe-
cial connection taking up the movement between vertebrae 3
and 4. These features are illustrated in Figure 1. In the Field
Museum mount the spine of vertebra 4 curves slightly forward.

THE SACRUM

(Figs. 4; 10 F; 15 A; 12; 13 F-I)

The spine of the sacral vertebra is high and slants backward;
in the MCZ mount the highest spine is that of vertebra 26, three
behind the sacral (Table 2). The anterior zygapophyses are
quite large, being the most posterior of an increasing size series.
The posterior zygapophyses comprise approximately half the area
of the anterior. The transverse process is enormously developed
for articulation with the large tuberculum of the sacral rib, and
the intercentrum bears a large facet for the capitulum. The facet
may impinge upon the pleurocentrum (Fig. 15 A; FMNH

BREVIORA

No. 428

1974

VERTEBRAL COLUMN OF ERYOPS

Table 2. Some measurements of Eryops Anterior Caudal Vertebrae

(^fCZ 1539)

Vertebra

Hei

gilt of

Neural

Spine

ram

Greatest
Length of
Vertebra

1
2
3
4

15
16

35 mm
35 mm

35 mm
38 mm

36 mm

35 mm

35 mm
31 mm
31 mm

29 mm

26 mm

Figure 15. All X .5. (A) Eryops sacral vertebra (23) , spine omitted, in
posterior (1.) , anterior and left views, the latter of MCZ 4305. (B) Eryops
vertebra 27 (caudal 4) , posterior (1.) and anterior views. (C) Eryops
vertebra 33 (caudal 10) , posterior (1.) and anterior views. (D) Eryops
vertebra 43 (caudal 20) , posterior (above) and anterior views. (E) MCZ
2634, right view, showing matrix (dark stippling) occupying position postu-
lated for cartilage about centra of postsacral vetebrae. (F) MCZ 1787,
posterior view, spine missing, showing closure of notochordal canal by
pleurocentra. (G) AMNH 4183, left view, showing fusion of two successive
chevrons. (H) MCZ 4325, left lateroanterior view of one to three caudal
vertebrae showing perforation of dorsal expansion of neural spine on the
left side for segmental blood vessel.

32 BREVIORA No. 428

UC60 and UC117), this being a characteristic of old and of
very large specimens. The sacral rib may fuse to or coossify
with its central articulations (FMNH UC117). The coossifica-
tion of elements is not uncommon at the sacral vertebra, al-
though the degree of fusion may differ on the two sides (MCZ
4305,2669,2604).

THE CAUDAL VERTEBRAE

( Figs. 5-8 ; 1 5 B-H ; 11 ; the measurements in Table 2 )

The total number of caudal vertebrae in Eryops is about 40.
The number of vertebrae may vary, but the possession of ribs on
the first eight caudals with chevrons beginning on the eighth
vertebra is taken as typical. The proximal caudal neural arches
are closely comparable to the presacral ones in their general fea-
tures, with less anteroposterior extension at the top, and with the
zygapophyses placed more closely together. In the trunk region,
the shaft of the neural spine tends to curve back and then up,
whereas in the caudal the longer part reaches upward before the
backward bend, this curvature being more pronounced pos-
teriorly. The heights of the caudal spines gradually decrease and
the tops change from an oval outline and become bifurcated, at
about caudal 4, into two abbreviated horns with rounded sum-
mits, one on each side, directed first posterolaterally (4-10),
then laterally (11, 12), and then anterolaterally (13-20). Be-
hind caudal 20, bifurcation is not noticeable. The horns are not
always symmetrical; one may be anterior to the other. They
were covered by skin in life (Romer and Witter, 1941). Near
vertebrae 20 to 22, the neural spine tips are altered, becoming
single again. By this point, the spine is much shortened with a
strong back-and-up curve, is thin from side to side, and is rather
long anteroposteriorly.

The zygapophyses are closer together and more sharply tilted
than in the dorsal vertebrae, and there is a reduction in size.
In the first dozen caudals, the sides of the neural arch tend to
be somewhat concave between the zygapophyses, as in the dorsal
vertebrae. After that they are quite flat. In the MCZ mount
transverse processes with broad but thin ends that gradually
narrow occur on the first seven caudals and exceptionally on one
side of the eighth. Behind the eighth, the pedicels are smooth,
although convex and swollen along their posterior borders. Each
vertebra, and hence its pedicels, becomes relatively and increas-
ingly narrow in the tail, so that the sides of the pedicels are more

1974 VERTEBRAL COLUMN OF ERYOPS 33

vertical. The surfaces facing the pleurocentra and intercentrum
are similar to those in the trunk for most of the length. In old
specimens, the floor of the neural canal may be complete (Fig.
11), suggesting that cartilage extended through the area in
younger specimens. The pedicels narrow below the spinal nerve
foramina.

In the proximal part of the caudal column, each pleurocen-
trum tends to broaden at the top, flatten on the lateral surface,
and extend relatively far down. They tend to become relatively
large and more important, and distally may approach the em-
bolomerous ring type (MCZ 2634). In the sacral region espe-
cially, the two bony pleurocentra become closely approximated
dorsally, and the ventral ends tend to approach one another more
closely than elsewhere. It is possible that in mature specimens
they fused into a ring, but no such specimens have been seen,
although intercentrum and pleurocentra together may coossify
into a ring centrum. Pleurocentrum enlargement and coossifica-
tion of vertebral elements in the sacral region may be adaptations
for terrestrial life.

In the MCZ mount, the first seven intercentra of the tail lack
a haemal arch; the first chevron is on the right side of vertebra
8, the left side presenting a transverse process and rib. This
count may have varied depending on the extent of the coelom
in the cloacal region. The proximal intercentra are like those of
the trunk, but capitular facets are well marked and the inter-
centra are more convex ventrally than dorsal intercentra. A
medial ventral groove appears in intercentrum 7 for the caudal
blood vessel which posterior to vertebra 7 courses through the
foramina of the haemal arches. These arches tend to develop a
keel on the front and to be flat behind, and to develop small
terminal cartilages. The shafts gradually become shorter, the
foramina occupying a progressively greater extent of their length.
Distally, the ends become flattened and tend to become antero-
posteriorly oriented, shoe-shaped expansions.

'To a greater or lesser extent, the neural spines of Eryops ver-
tebrae show lateral grooves where segmental blood vessels have
coursed. On each of three caudal vertebrae of MCZ 4325, near
the front of the bifurcated spine series, a shallow groove appears
on the left side of the neural spine perforating or indenting the
dorsal tuberosity of the neural spine (Fig. 15 H). These three
are unique in the collections I have studied ; presumably all came
from the same animal.

BREVIORA

No. 428

3
O

u
u

i-H

t*3

«3

>
o

3
01

1974

VERTEBRAL COLUMN OF ERYOPS

I— (

>-(
bo

BREVIORA

No. 428

Sh
<U
>

•a

1974

VERTEBRAL COLUMN OF ERYOPS

CO
V

Xi
u

-4-1
i-l

n
u
u

■4-1
f/i

o
a.

I— I

u
Hi

38 BREVIORA No. 428

THE RIBS

(Figs. 13F-I; 16-19)

Of the 22 presacral vertebrae, the first (atlas) lacks a rib; in
some mounts the last presacral rib, that of vertebra 22 has been
omitted, while in others a misplaced iHum spans too many ribs.

The head of aU presacral ribs is expanded, with only a con-
striction separating capitulum and tuberculum. The tubercular
part is somewhat thicker than the capitular. The articular sur-
faces are somewhat concave and unfinished, suggesting a carti-
laginous surface coat. The proximal rib ends are inclined clock-
wise from the vertical, as are the corresponding articular surfaces
of the transverse processes (Table 1). The ribs are flat distally.

There are thickenings of the shaft continuous with capitulum
and tuberculum, and there may have been considerable variation
in the form of the uncinate processes. The distal rib ends are
unfinished, except those of the posterior presacrals, and presum-
ably ended in cartilage, but this is uncertain. In the MCZ
mount, uncinate processes are shown on the 2nd through the
13th ribs, reducing in size and disappearing as the ribs shorten
posteriorly. Caudal ribs lack these structures.

DISCUSSION

It is clear that — distortion apart — individuals of Eryops
were variable in the following details of their spinal columns:
extra processes and exostoses; closure of neural canal; degree of
definition of capitular facets ; relative sizes of neural arches, inter-
centra and pleurocentra ; degree and asymmetry of coossification ;
shapes of neural spines and of atlas, axis and the special fourth
vertebra; angles of inclination of the neural spines, and details
of configuration of their dorsal expansions. Nevertheless, a clear
picture emerges of a repeated series of ventral intercentra, dorsal
neural arches and paired dorsolateral pleurocentra probably
separated or held together in life by cartilage, which may have
been continuously woven among the centra or interrupted be-
tween vertebrae anterior to each intercentrum ; it is not clear
from the fossil record which was the case. Presumably cartilage
was more extensive in younger than in older animals. The un-
finished articular surfaces of vertebral elements clearly reflect
their continuation in cartilage.

Arrangement of vertebral elements varies considerably in exist-
ing Eryops reconstructions. Fossilized pieces, even when found

1974 VERTEBRAL COLUMN OF ERYOPS 39

adjacent, are, unless coossified, often difficult to fit exactly to each
other, presumably due to the missing cartilage. We have found
no reason to quarrel with Cope's (1890) description, which is
an excellent guide to vertebral arrangement in Eryops: "The
neural arch rests exclusively on the pleurocentrum, which in
turn adheres to the intercentrum behind it by its long side, and
to that in front by its short side or end", and of caudal vertebrae
"... the pleurocentra descend further than in the dorsal region,
rarely to the inferior face of the column, and separating the
intercentra from mutual contact." These points are illustrated
in Figures 10 and 15.

As regards the function of the rhachitomous vertebral column,
two views have been advanced. Cope (1884) proposed a coat
sleeve on a semiflexed arm as a model of the flexible cylinder to
which he earlier (1883) had likened the column of Eryops. He
saw the osseous elements of the rhachitomous vertebral column
distributed through a sheath of softer tissue around the noto-
chord, like segments of the skin of an orange — segments of a
sphere, as it were. "If you take a flexible cylinder, and cover it
with a more or less inflexible skin or sheath, and bend that
cylinder sidewise, you of course will find that the fractures of
that part of the surface will take place along the line of the
shortest curve, which is on the side; and, as a matter of fact,
you have breaks of very much the character of the segments of
the Permian batrachia" (1883: 276). In a coat sleeve cover-
ing the semiflexed arm, the folds represented to Cope the
fractures in the flexible cylinder, the intervals between elements,
and the spaces between folds the elements themselves. Cope left
it to future investigations to determine the applicability of his
model to the history of the vertebral column (1884: 32).

Parrington (1967) suggested a geodetic spiral, presenting the
rhachitomous vertebral column as a series of discrete ossicles in
a cartilage continuum, allowed to twist by virtue of the embed-
ding of the rather rhomboidal osseous elements interdigitated in
a cartilage matrix. Such twisting, Parrington suggested, would
have been essential for amphibious tetrapods like labyrinthodonts
on coming ashore in order to maintain a center of gravity upon
a triangle of three legs while bringing the fourth leg forward for
the next step. Coalescence of neural arches and neural spines in
certain armored rhachitomes has led Vaughn (1971) to question
whether or not Parrington's model can have applied to locomo-
tion in these particular labyrinthodonts. On the other hand, the

40 BREvioRA No. 428

flexibility in vertebral column which Parrington's model provides
would, it seems to me, lend itself ideally to the stereotyped loco-
motion probably imposed upon a large, tailed amphibian such
as Eryops by extension of the supracoracoideus muscle, between
coracoid and humerus, to the forearm through the coraco-radialis
proprius, as I have discussed it for living urodeles (Moulton,
1952). While the arrangement may have relieved Eryops from
decisions leading to more complicated locomotory patterns, the
simultaneous adduction of the forelimb and flexing of the fore-
arm, re-establishing at each step the triangle of three legs as
envisoned by Parrington, would have abetted the twisting of a
spirally organized vertebral column and vice-versa. It is noted
that Miner ( 1925) questions the occurrence of the coraco-radialis
proprius in Eryops. Thomson and Bossy have argued (1970)
that the temnospondyl and anthrocosaur amphibian lineages
represented different experiments in a terrestrial vertebral column,
both based on the principle of a geodetic spiral enunciated by
Parrington.

The spiral pattern suggested by Parrington seems reasonable
as a device for strengthening a vertebral column like that of
Eryops subject to the stresses of locomotion on land. Are there
evidences of the proposed torsion in fossil material? I believe so.
Two intercentra broken and healed during life (MCZ 2621,
4306 ) , and one that developed a shallow dorsal split also during
life (MCZ 4305) have been encountered (p. 22). Inasmuch
as each occurred at an angle clockwise from the primary axis
(MCZ 2621, 8°; 4306, 30°; 4305, 30°), I suggest that these
breaks may have occurred in young animals and that they may
reflect the twisting hypothesized by Parrington in his spiral model.
Such breaks are not common in fossil collections, the ones de-
scribed being unique among the intercentra I have studied.

At present, the detailed pattern of evolution of vertebral centra
is unsettled. Recent papers of special significance are those of
Williams (1959), Panchen (1967), Thomson and Vaughn
(1968) and Thomson and Bossy (1970). Despite gaps in our
knowledge of the details, there is a general concensus that some
form of the rhachitomous vertebra was the primitive amphibian
type; however, increasing evidences of variation in vertebral
pattern among primitive amphibians greatly complicate the pic-
ture (R. L. Carroll, personal communication). Eryops itself has
moved along the temnospondylous line from the most primitive
labyrinthodont condition (Romer, 1947). In suggesting that the

1974 VERTEBRAL COLUMN OF ERYOPS 41

amphibian centrum is homologous throughout, but differently
subdivided in different hneages, Panchen (1967) introduced an
idea open to examination by determining the attachments of
myosepta to the vertebrae, for in all tetrapods, it is clear since
the important review by Williams (1959), caudal and cranial
half sclerotomes of successive somites unite, resulting in alterna-
tion of vertebrae and primary muscle segments. Panchen saw
the vertebral margin of the myoseptum with its segmental blood
vessel providing the dividing line between intercentrum and pleu-
rocentrum. In temnospondyls he saw the myoseptum moving
posterodorsally, ultimately to the stereospondyl condition, leaving
an increasingly large intercentrum ahead of the myoseptum until
the pleurocentrum disappeared. Anteroventral movement of the
myoseptum on the anthrocosaur line would have resulted ulti-
mately in the loss of the intercentrum, and in an amniote centrum
formed from the pleurocentrum posterior to the myoseptum.

While I have no new evidence on the course of the interseg-
mental blood vessels in relation to the centra in labyrinthodonts,
the pathway for the blood vessels and myosepta postulated by
Panchen (1967: 28) as applicable to fossil material is supported
by the three neural arches of caudal vertebrae (p. 33) which
are grooved and perforated on the left side almost certainly for
the passage of segmental blood vessels. A similar pathway on
dorsal vertebrae of Eryops could easily have been continued
along the tops of the transverse processes (p. 20), behind the
well-defined ridge already described, then dropping behind the
rib blades almost exactly as Panchen describes and illustrates
(1967: fig. 5A). Since the courses of segmental blood vessels
have rarely been preserved in labyrinthodont vertebrae (Pan-
chen, 1967: 28), these three clearly marked caudal vertebrae
assume a special significance to our understanding of vertebrae
and muscle segments in Eryops.

The broadly flat form and orientation of most of the trunk
ribs in Eryops probably did not allow for much lateral undula-
tion, such as suggested by Thomson and Bossy (1970: 11) for
Ichthyostega. The tail, however, would have served as an excel-
lent swimming organ ; reconstructions that show it as flexible and
leaning toward one side on land may be close to the truth. That
it was strengthened by dorsal tendons seems likely from the
bifurcate nature of some of the spines.

42 BREvioRA No. 428

LITERATURE CITED

Boll, J. 1880. Geological examinations in Texas. Amer. Natur. 18: 26-39.
Branson, E. B. 1905. Structure and relationships of American Labyrintho-

dontidae. J. Geol., 13: 568-610.
Broili, F. 1899. Ein Beitrag zur Kenntniss von Eryops megacephalus

(Cope) . Palaeontographica, 46: 61-84,
Case, E. C. 1911. Revision of the Amphibia and Pisces of the Permian of

North America. Publ. Carnegie Inst. Wash., No. 146, pp. 1-179.
. 1915. The Permo-Carboniferous red beds of North America

and their vertebrate fauna. Publ. Carnegie Inst. Wash., No. 207, pp. 1-176.
Cope, E. D. 1877. Descriptions of extinct Vertebrata from the Permian and

Triassic formations of the United States. Proc. Amer. Phil. Soc, 17:

182-193.
. 1878a. Descriptions of extinct Batrachia and Reptilia from the

Permian formation of Texas. Proc. Amer. Phil. Soc, 17: 505-530.

1878b. The homology of the chevron bones. Amer. Natur.,

12: 319.

1880a. Second contribution to the history of the Vertebrata

of the Permian formation of Texas. Paleontological Bulletin No. 32
(June 5, 1880), pp. 1-22.

. 1880b. Same title. Proc. Amer. Phil. Soc, 19: 38-58.

1881. Same title, figures. Paleontological Bulletin No. 32

(May 2, 1881) , pp. 162-164.
. 1882. The rhachitomous Stegocephali. Amer. Natur., 16:

334-335.

1883. The evidence for evolution in the history of the extinct

Mammalia. Science, 2: 272-279.
. 1884. Batrachia of the Permian period of North America.

Amer. Natur., 18: 26-39.
. 1890. On the intercentnim of the terrestrial Vertebrata. Trans.

Amer. Phil. Soc, 16: 243-253.

-. 1897. Recent papers relating to vertebrate paleontology. Amer.

Natur., 31: 314-323.

1898. Syllabus of lectures on the Vertebrata, with an intro-

duction by H. F. Osborn. Philadelphia: University of Pennsylvania.
xxxv + 135 pp.

-, AND W. D. Matthew. 1915. Hitherto unpublished plates of

Tertiary Mammalia and Permian Vertebrata. Monograph Series No. 2,
Amer. Mus. Nat. Hist.

Gregory, W. K. 1951. Evolution emerging, vol. I and II. New York: The
Macmillan Co. xxvi + 736 pp., 1013 pp.

Lancston, W., Jr. 1953. Permian amphibians from New Mexico. Uni-
versity of California Publications in Geological Sciences, 29: 349-416.

. 1963. Fossil vertebrates and the late Palaeozoic red beds

of Prince Edward Island. Nat. Mus. Canada, Bull. No. 187.

1974 VERTEBRAL COLUMN OF ERYOPS 43

Miner. R. W. 1925. The pectoral limb of Eryops and other primitive
tetrapods. Bull. Amer. Mus. Nat. Hist., 51: 145-312.

MouLTON, J. M. 1952. Studies on the derivatives of inverted heteropleurally
transplanted forelimb buds of Ambystoma maculatum, with particular
attention to the heterotopic shoulder region. Ph.D. Thesis, Harvard
University. 379 + xii pp.

Murphy, J. L. 1971. Eryopsid remains from the Conemaugh Group. Brax-
ton County, West Virginia. Southeast Geol., 13: 265-273.

Olson, E. C. 1936a. The ilio-sacral attachment of Eryops. J. Paleontol.,
10: 648-651.

. 1936b. The dorsal axial musculature of certain primitive

Permian tetrapods. J. Morphol., 59: 265-311.

, AND P. P. Vaughn. 1970. The changes of terrestrial verte-

brates and climates during the Permian of North America, forma et
functio, 3: 113-138.

OsBORN, H. F. 1931. Cope: master naturalist. Princeton, N.J.: University
Press, xvi + 740 pp.

Panchen, a. L. 1966. The axial skeleton of the labyrinthodont Eogyrinus
attheyi. J. Zool., 150: 199-222.

. 1967. The homologies of the labyrinthodont centrum. Evo-
lution, 21: 24-33.

Parrington, F. R. 1967. The vertebrae of early tetrapods. In Probleraes
actuels de paleontologie, ed. by J.-P. Lehman. Paris: Centre Nat. Rech.
Sci., pp. 269-279.

Rockwell, H., F. G. Evans and H. C. Pheasant. 1938. The comparative
morphology of the vertebrate spinal column: its form as related to func-
tion. J. Morphol., 63: 87-117.

RoMER, A. S. 1943. Recent mounts of fossil reptiles and amphibians in the
Museum of Comparative Zoology. Bull. Mus. Comp. Zool., 92: 331-338.

. 1947. Review of the Labyrinthodontia. Bull. Mus. Comp.

Zool., 99: 1-368.

. 1952. Late Pennsylvanian and early Permian vertebrates of

the Pittsburgh — West Virginia region. Ann. Carn. Mus., 33, Art. 2:
47-110.
. 1958. The Texas Permian redbeds and their vertebrate

fauna. In Studies on fossil vertebrates. Essays presented to D. \L S.

Watson, ed. by T. S. Westoll. London: The Athalone Press, pp. 157-179.

. 1966. Vertebrate paleontology, 3rd ed. Chicago: University

Press, viii + 468 pp.
, AND R. V. Witter. 1941. The skin of the rhachitomous am-

phibian Eryops. Amer. J. Sci., 239: 822-824.

-. 1942. Edops, a primitive rhachitomous

amphibian from the Texas red beds. J. Geol., 50: 925-960.
Sawin, H. J. 1941. The cranial anatomy of Eryops megacephalus. Bull.

Mus. Comp. Zool., 125: 43-107.
Stickler, L. 1899. Ueber den microscopischen Bau der Faltenzahne von

Eryops megacephalus Cope. Palaeontographica, 46: 85-94.

44 BREVIORA No. 428

Thomson, K. S., and K. H. Bossy. 1970. Adaptive trends and relationships
in early Amphibia, forma et functio, 3: 7-31.

, AND P. P. Vaughn. 1968. Vertebral structure in Rhipidistia

(Osteichthyes, Crossopterygii) with description of a new Permian genus.
Postilla No. 127: 1-19.

Vaughn, P. P. 1958. On the geologic range of the labyrinthodont am-
phibian Eryops. J. Paleontol., 32: 918-922.

. 1971. A Platyhystrix-like amphibian with fused vertebrae

from the upper Pennsylvanian of Ohio. J. Paleontol., 45: 464-469.

Watson, D. M. S. 1919. The structure, evolution and origin of the Am-
phibia. — The "orders" Rharhitomi and Stereospondyli. Philos. Trans.
Roy. Soc. Ser. B, 209: 1-73.

Williams. E. E. 1959. Gadow's arcualia and the development of tetrapod
vertebrae. Quart. Rev. Biol., 34: 1-32.

Williston. S. W. 1914. Restorations of some American Permocarboniferous
amphibians and reptiles. J. Geol., 22: 57-70.

. 1915. Trimerorhachis, a Permian temnospondyl am-
phibian. J. Geol., 23: 246-255.

. 1918. The evolution of vertebrae. Contr. Walker Mus..

2: 75-85.

^^ ^-^ L. APR 2 11977

lARX^ARO
ITY

B R E V I 0 R^-^'

Museum of Comparative Zoology

us ISSN 000&-9698

Cambridge, Mass. 27 November 1974 Number 429

AN O LIS RUPINAE NEW SPECIES

A SYNTOPIG SIBLING OF

A. MONTICOLA SHREVE

Ernest E. Williams^

AND

T. Preston Webster^

Abstract. A new species, Anolis rupinae, is distinguished from sibling
A. monticola by its larger size and different coloration. These two species
and A. koopmani, which is smaller than either and somewhat different in
coloration, scale characters, and habitus, constitute a subgroup in the larger
monticola species group. All three occur on the western end of the Tiburon
Peninsula of Haiti and have karyotypes derived by fission of one or more
macrochromosomes from the primitive iguanid complement. The remaining
four species of the monticola species group are morphologically more diverse
and occur north of the Cul de Sac depression. Three have the primitive
karyotype. Anolis rupinae is known only within the habitat of A. monticola,
but is allotopic to A. koopmani. Chromosome change may have been im-
portant in the evolution of this distinctive miniradiation.

As Schwartz ( 1973) has commented, it is clear that the roster
of Hispaniolan Anolis species is not yet complete. Webster and
Bums (1973) have just demonstrated that the A. brevirostris
complex in Haiti must be divided into three species, and Schwartz
(1973, 1974a) has described two striking new species in the
Dominican Republic. In addition Schwartz (1974b) has shown
that the Hispaniolan giant anoles heretofore considered a single
species, A. ricordi, are better interpreted as three species.

The key feature of all the recently described Hispaniolan
anoles is that they are local species. They may or may not be
abundant where they occur but they all have restricted distribu-
tions, often montane, sometimes in very arid regions, often less

^Museum of Comparative Zoology, Cambridge, Massachusetts 02138

2 BREvioRA No. 429

obviously circumscribed. Sometimes they are known from a
single locality, more often from a number of localities relatively
close together. The species of the ricordi complex have the
widest ranges of any of the newly recognized forms, but again
these are allopatric or parapatric, none islandwide.

We here add still another local species, the major peculiarity
of which is that it is syntopic with its closest relative.

Anolis rupinae^ new species

Holotype. MCZ 121740, an aduh male.

Type locality. 1.3 km SSW Castillon, Departement du Sud,
Haiti, T. P. Webster and A. R. Kiester, collectors, 6 September
1969.

Paratypes. All Departement du Sud. From the type locality:
MCZ 121737-39, same data as the type: MCZ 124475-87,
124612-15, 124851, T. P. Webster collector, 2 July 1970.

Diagnosis. Close in all scale characters and counts to Anolis
monticola but differing in larger size and in color.

Head. Head moderate, head scales rugose or keeled. 9 to 15
scales across snout between second canthals. Frontal depression
shallow, scales within it as large or larger than those anterior and
lateral to it. Anterior and ventral nasal scales (or these plus the
anteriormost of the lowest loreal row) in contact with rostral.
7 to 11 scales in contact with rostral posteriorly. Supraorbital
semicircles separated by two rows of scales. 10 to 17 keeled
scales in supraocular disk, which is separated from the supra-
ciliaries by five or more rows of granules. Two elongate supra-
ciliaries ending at about mid-eye, continued by granules. Can-
thals distinct; about 6 to 7 canthals, the first three elongate,
strongly overlapping, first sometimes as long as second. Loreal
rows 6 to 9, lower row slightly larger, supratemporal rows slightly
enlarged. Temporals granular, scales behind interparietal very
slightly enlarged, those anterior and lateral to it markedly larger.
Interparietal smaller than ear, separated from supraorbital semi-
circles by 3 to 6 scales. Suboculars separated from supralabials
by one row of scales. Six supralabials to center of eye. Lower
eyelid with a window of granular scales.

Mental much broader than long, in contact with 4 to 8 scales
between the large sublabials. Only one or two sublabials on each
side clearly defined, posterior to these there are two to three rows

^from the Latin rupina: a rocky chasm.

1974

ANOLIS RUPINAE NEW SPECIES

••H

O
Q

N
U

•--ft

BREVIOKA

No. 429

1974 ANOLIS RUPINAE NEW SPECIES

Figure 3. Anolis rupinae, MCZ 124857. Ventral view of chin.

of enlarged scales alongside the narrow infralabials. Throat
scales smaller, slighdy elongate anteriorly.

Trunk. Two to three middorsal scale rows enlarged, keeled.
Flank granules keeled. Ventrals larger than middorsals, weakly
keeled, imbricate, subimbricate or rarely juxtaposed.

Dewlap. Small, in males only but extending to the level of
the axillae, the largest scales about as large as ventrals, weakly
keeled.

Limbs and digits. Dorsal scales of arm and anterior scales of
thigh and of lower \^g unicarinate. Those of digits and of knee
multicarinate. 16 to 21 lamellae under phalanges ii and iii of
fourth toe.

Tail. Compressed, four middorsal scales per verticil. Postanal
scales large in male.

Size. Males to 56 mm in snout-vent length, females to 42 mm.

Color in life. Webster, 6 September 1969: Adult male type
from Castillon: "Snout uniform olive green above. Neck subtly
mottled with shades of olive and pale green. Five pale green
transverse bands from neck to base of tail. Middorsally the
nuchal and dorsal crests have alternating areas of pale blue-green
and olive. More laterally the transverse bands separate olive
brown blotches with yellowish spots in them. Dorsum of base of

6 BREVIORA No. 429

tail with areas of olive alternating with pale green. Distally, tail
black alternating with greenish cream.

"Side of snout pale dull green. Eyelids yellow-orange. Iris
turquoise. Pupil black. Behind eye very dark green. Below it
pale green. From shoulder along flank a bright green stripe,
broadening where it crosses the transverse bands, which are
lighter green on the lower flanks.

"Below, chin pure bright yellow. Dewlap scales yellow. Skin
sky blue. Chest scales yellow, those of belly not so bright and
with the yellow intermingled with areas of dull orange. Under
tail red orange spots surrounded by yellow scales, the spots be-
coming more diffuse and vanishing toward the tip.

"Limbs dorsally with alternating light yellow green and light
brown. Two green bars on upper and lower arm and tibia but
three such bars on the femur. Hand and foot similarly cross-
barred. Ventrally, limbs mottled yellow brown."

Webster, 1 July 1970 (Castillon) : "All sizes and both sexes
of rupinae can be distinguished from monticola by the red-orange
color on the ventral surface. Males are larger, lack the scapular
patch and have a blue (sky blue) dewlap and a brilliant green
lateral stripe. The edges of the middorsal band in females are
straight without scalloping. In both sexes bright yellow around
the eye."

Color as preserved. The green stripe so conspicuous in life is
usually absent in preserved specimens. The red of the ventral
surfaces also vanishes and the dorsal banding is less vivid. In
preserved male rupinae the most marked difference from A. mon-
ticola is the absence of any scapular spot. Females are more
difficult but the red spots under the base of the tail in life are
seen in preserved specimens as very white spots which may coa-
lesce to an undulating bright line under the first part of the tail.
(In monticola light pigment under the tail is weakly developed
or present as a straight-edged line. )

Karyotype. Diploid chromosome numbers are known for two
male paratypes (MCZ 124612-13). In diakinesis one has five
macrochromosomal bivalents and one trivalent, while the other
has six bivalents and one trivalent. Both have 13 small bodies
interpreted as microchromosomal bivalents. On the basis of this
minimal sampling of the one known population, diploid numbers
in A. rupinae should vary from 38 to 42.

1974

ANOLIS RUPINAE NEW SPECIES

V
CL,

-o

>
u

a,
o

tf3

(73

s:
U

8 BREvioRA No. 429

COMPARISONS

A. rupinae requires comparison only with certain members of
the monticola group, and only nominate monticola is known to
co-occur with it.

In scale characters rupinae (Table 1) is either identical with
monticola or (as the snout scales) overlapping. However, though
both are richly, even gaudily colored, the two species are sharply
distinct in color (Table 2). The bright flank stripe of rupinae
is missing in monticola while the two (nuchal) or four (nuchal
and occipital) black patches with blue ocelli of monticola monti-
cola and monticola quadrisartus are absent in rupinae.

In dewlaps, although small in both species, there is a contrast
also. At the Castillon ravine, monticola has a yellow to reddish
orange dewlap while that of syntopic rupinae is sky blue.

Adult size also distinguishes A. rupinae from syntopic A. mon-
ticola. However, on this point there is a confusion in the litera-
ture. Thomas and Schwartz (1967) cite the maximum size of
monticola males as 55 mm (they do not mention the specimen)
and that of females as 39 mm. They comment on the strong
sexual dimorphism. We have at hand Thomas Schoener's meas-
urements for specimens referred to monticola in Schwartz's col-
lection, the earlier collections of the Museum of Comparative
Zoology and the American Museum of Natural History. A single
specimen is reported by Schoener to reach 52.5 mm (AMNH
49845 from "25 mi N Aux Cayes, Jeremie Road" [corrected by
Thomas and Schwartz to "32 miles from Aux Cayes on the
Jeremie Road" from Hassler's field notes] ) . This locality is well
within the range of monticola monticola and is one of the Hassler
specimens reported by Williams (1962) as A. monticola and so
regarded also by Thornas and Schwartz. It is this specimen that
provided the 55 mm record (Schwartz, personal communica-
tion). It is now clear that this specimen is not monticola {stt
below).

In the relatively large series that the Museum of Comparative
Zoology now possesses from the Castillon ravine and from other
localities no male monticola monticola exceeds 45 mm in snout-
vent length; this size is exceeded by female rupinae (46 mm
snout-vent length) from Castillon. The Schwartz collection of
monticola monticola has no male with a snout-vent length greater
than 42 mm. A. ju. quadrisartus is somewhat larger: Schwartz
(personal communication) reports males of 48 mm snout-vent
length. Thus no veritable specimens of monticola or quadrisartus

1974 ANOLIS RUPINAE NEW SPECIES 9

are known to reach the 55 to 57 mm snout-vent length of Cas-
tillon rupinae or of AMNH 49845.

The latter specimen has been a source of confusion in more
than size. It was cited by Williams (1962) as the basis of a
description in life for male monticola. We quote again the de-
scription which is taken from W. G. Hassler's field notes:

"General dorsal color Hooker's Green. Saddles brown green,
three in number, narrowest middorsally, one across shoulder, two
between fore and hind legs. A light crescent in the temporal
region. Throat and belly dark olive green. Legs barred. Tail
barred. Eyes Antwerp Blue, sometimes changing to greenish.
Edge of orbit yellowish brown. Skin of fan (which is relatively
small) blue, scales light and dark green. Occurring also in a
dark phase almost without pattern."

From the vantage point of present knowledge this description
cannot be matched with either rupinae as known from Castillon
or monticola or quadrisartus. Unmentioned are such diagnostic
elements of color pattern as the flank stripe of rupinae and the
two or the four ocelli of m. monticola and m. quadrisartus. The
specimen itself as now preserved shows no pattern at all.

We may mention here two other difficult specimens (MCZ
124537-38). Both are males (43 mm and 49 mm in snout-vent
length) collected by Webster at Catiche within the range of
quadrisartus. Both are without ocelli and hence are clearly not
m. quadrisartus or m. monticola. However, they were obtained
in a lizard market, along with numbers of m. quadrisartus, and
no detailed notes on color in life exist for them, nothing beyond
the fact that one had a yellow belly and the other a red one.

We cannot on present evidence confidently refer either these
two specimens or AMNH 49845 to rupinae. As preserved, one
of the Catiche specimens shows the subcaudal white spots char-
acteristic for rupinae; the other Catiche specimen does not, nor
does AMNH 49845. Since the one Catiche specimen which had
a red venter has also the white subcaudal spots in preservation,
it .may be truly rupinae. In the case of the other two, we call
attention to the possible existence of still undescribed taxa and
make no assignment of these specimens. We emphasize that our
concept of rupinae rests solely upon the animals from the Castil-
lon ravine.

It is worth noting that for neither rupinae nor the two sub-
species of monticola is sexual dimorphism so marked as Schwartz
assumed it to be for monticola when he included xAMNH 49845

BREVIORA

No. 429

•9

ltd

-g

(X4

1974 ANOLIS RUPINAE NEW SPECIES 11

in that species. His ratio of maximum male size to maximum
female size (55 mm to 39 mm) approximates 1.4. Our revised
data show monticola monticola with maximum cT size 45 mm,
maximum ? size 42 mm, equivalent to a ratio of 1.07. The
comparable data for m. quadrisartus are d 48 mm, $ 42 mm,
ratio 1.14, and for rupinae d 57 mm, ? 46 mm, ratio 1.24.

One other species of the monticola group appears very close
to rupinae. In fact, the resemblances of koopmani to rupinae
seem as close or closer than those to m. monticola or m. quadri-
sartus ( Tables 1 and 2 ) . Particularly striking is the red ventral
coloration found in both rupinae and koopmani, but similar also
is the presence of a flank stripe, the bluish dewlap, the throat
spotting and the yellow or orange chin. The major color differ-
ence between rupinae and koopmani is the absence of trans-
verse banding in koopmani; only in this aspect of color is rupinae
closer to m. monticola and m. quadrisartus.

However, A. koopmani is a grass anole, and the adaptation
has required a body shape different from that of rupinae or
monticola (Fig. 5). In size also A. koopmani is distinctive; it
is the smallest of this group of three related forms ( d maximum
size 42 mm, 9 maximum size 35 mm, with a ratio therefore
of 1.2).

A. koopmani has not before been explicitly referred to the
monticola group. Rand (1961), however, in describing the spe-
cies, did suggest that A. monticola, darlingtoni [now etheridgei],
christophei, hendersoni, baharucoensis, and Xiphocercus [now
Anolis darlingtoni] might, some or all of them, be closely related
and that koopmani s> relationship might lie with these. We would
certainly agree that koopmani has rather close affinities with all
the species Rand named. However, from rupinae's resemblances
on the one hand to monticola and on the other to koopmani, it
is now obvious that these three South Island species are a very
tight group or subgroup of their own. That they are only a little
less close to A. rimarum and related but more distantly to A.
etheridgei, A. christophei and newly described A. fowleri, all
North Island species (Tables 3 and 4), we will also affirm. We
now, however, would distinguish between a monticola group
sensu lato including both North and South Island species and a
monticola group sensu stricto containing only the South Island
species. The intimate relationship of the latter clearly separates
them as a unit, as opposed to the significantly more diverse North
Island set.

12 BREVIORA No. 429

It is clear, therefore, that the description of rupinae (like
Schwartz's recent discoveries of South Island A. sheplani and
North Island A. fowleri) does nothing to diminish the intriguing
differences between North and South Island montane faunas that
were commented on by Williams and Rand (1969).

It is too early to do more than draw attention to a problem
still unsolved. We are in no position to make dogmatic state-
ments about the montane fauna or faunas. To cite only one
example, the genus Chamelinorops, which on reasonable grounds
was thought to be a South Island endemic or even autochthon
(Thomas, 1966) is now known by a single juvenile from the
middle of the Cordillera Central (MCZ 126708 from Limoncito,
southwest of Constanza, La Vega Province, Dominican Repub-
lic) collected by T. P. Webster. In such cases of rare or local
species, no safe judgments will be possible until montane His-
paniola is much better known than it is now. Nevertheless it is
worth noting that at present no parallel is known in the North
Island montane fauna to the South Island close-knit triplet of
monticola, rupinae and koopmani. The North Island set of spe-
cies are each very distinct from one another morphologically and
in color and in ecology. The discontinuities are very sharp, so
sharp that their association as a group is not beyond question.
This is very different from the South Island series.

HABITAT, CONGENERS, ECOLOGY

The type locality.

Castillon is a market place and a diffuse village at a low point
in one of several ridges extending north from the Massif de la
Hotte. The surrounding country is dry and highly disturbed.
Land not used for subsistence agriculture or pasture is covered
by brush. Within this area A. rupinae occurs in the small pocket
of damp and shady habitat in a ravine visible from the hill north
of Castillon. About 200 meters south of the market stalls the
road bends sharply, and at this point there is a well-worn trail
along the side of the ridge. At first it traverses generally open
slopes, but after a little more than one kilometer the trail enters
the ravine near the base of a cliff. Between the cliff and the trail
there is a fairly level area 12 to 15 meters long and 4 to 6 meters
wide, where all specimens of A. rupinae were collected.

Only a few medium to large trees grow within the ravine. The
rocky ravine floor and surrounding slopes are, however, covered

1974 ANOLIS RUPINAE NEW SPECIES 13

by a thick growth of brush. Because of the cliff and the steep
hillsides, the ravine floor is sheltered from the sun. On July 2,
1970 sunlight did not reach the cliff base until 9:30. Water
trickles over and seeps from the base of the rock wall. Mosses
and similar plants flourish on moist and shaded rocks. Below
the trail the ravine is steep and filled with broken rock. It soon
widens and becomes more exposed.

The anoline lizards of the Castillon area.

In September, 1969 and in June, 1970 the fauna of this area
was sampled by organizing lizard markets in the Castillon market
place and by collecting during the day and night around and
within the ravine. Only A. distichus and A. coelestinus are
abundant and generally distributed in the region. Both occur on
the exposed slopes around the ravine but not within it. A third
widespread and essentially lowland species, A. cy botes, is much
less common. It does occur in some numbers on rocks along the
trail entering the ravine and in the ravine itself. Two specimens
of A. ricordii were procured from a lizard market.

Four species occurring here are considered montane, since they
are unknown from coastal localities. (1) A. hendersoni is un-
common around Castillon, at least along the trail to the ravine.
A single specimen was collected near the market place, and two
were taken in trail-edge vegetation near the ravine. (2) On the
ravine floor and along its approaches A. monticola is abundant.
While this species occurs throughout the brush in the ravine up
to the periphery of some bordering garden areas, it is absent from
drier brush patches on the hillsides. (3) A single Chamelinorops
barbouri was found along the trail near the ravine. (4) The
total area inhabited by the population of A. rupinae is probably
quite small. All specimens were collected from a very short seg-
ment of the ravine floor. Perhaps it also occurs on the surround-
ing cliff and slopes of broken rock. It seemed much less common
in 1969 than in 1970.

' Of these anoline species, A. rupinae seems to have the strong-
est requirement for cool, moist conditions. Anolis monticola is
the only other species common on the ravine floor, but it also
occurs on the sides of the ravine. Anolis cy botes occurs within
the ravine close to A. rupinae, but the two species seem to have
exclusive microdistributions. The other species seem to be intol-
erant of the ravine environment or were observed too infrequently
for any statement on co-occurrence with A. rupinae.

14 BREVIORA No. 429

A syntopic sibling?

In one regard A. rupinae appears to be unique among anoles.
This may be a defect of our present information, but rupinae is
currently known only within the immediate habitat of A. monti-
cola.

It is worth emphasizing that, if confirmed, this is a special
situation. A. rupinae is close enough structurally to A. monticola
to be called a sibling of the latter, that is some museum specimens
and perhaps females in the field have been (see above) or could
be confused. Many such sibling pairs are known in the West
Indies, sometimes sibling only in the sense of closest relatives,
sometimes in the more usual sense of both close relatives and
barely distinguishable ( under some, usually museum, conditions ) .
However, such siblings ordinarily are either distinct in climatic
preference and hence allotopic or they are para- or allopatric
(as ^. rupinae appears to be to ^. koopmani) .

Possibly A. rupinae does occur somewhere separately from A.
monticola. Certainly A. monticola is known from a number of
localities at which A. rupinae is not known. However, it can be
pointed out already that the sharply different color patterns of
these two species (and the dewlap difference at Castillon) and
the striking difference in size are the kinds of adaptations — the
color patterns for species recognition, the size difference for
avoidance of competition for food — that syntopic or widely
overlapping anoles have evolved in many instances (the Schoener
rules, Schoener 1970, WiUiams 1972). That rupinae appears to
be even more rigidly tied to shaded and moist situations than is
monticola does not damage the suggestion that rupinae and
monticola are consistently syntopic. On the contrary, this pre-
sumed greater shade and moisture preference of rupinae makes
it all the more likely that its preferred habitat is within the habi-
tat range of monticola. (From the evidence of Castillon rupinae
does not exclude monticola.)

The monticola group sensu strict o — an unusual miniradiation

The status of A. rupinae and A. monticola as unusual siblings
is compounded by the close relationship of both to A. koopmani.
While certainly not a sibling — divergence in scale counts, habi-
tus and size are all reasons for its previously uncertain affinities
— • the presence of this third species in the same small mountain
mass is evidence that the monticola group sensu stricto has

1974

ANOLIS RUPINAE NEW SPECIES

C3 ^

liJ Ixl <i
< < ^

O O
O O

^ £1 Q_ ^ !-

Q. a. o ^ ^

3 3 o O O

q: cr i^ S 2

o ^
c ^

.. o

8 s

XT)
tn I— (

•5 00

V, X!

u Q

4-) C

s-i C

.3 o

<4-l

■i-i <u

o ^

*^ rt

rH >

2 <

^ C/3

C3 O

>^ flj

c
o

J3
P

o s

c
o

»5

g
10

c
o

c/5

I— c «-)

• •* 03 Cd

--V > >

>- OJ «

XI U ^

'J - S

3 3 »

°|§

Z h ^

C/3 "-I .

=^ ^ -^

o a o

O 4;

,n bO «

03 T3 S

S3 <U ►J

16 BREvioRA No. 429

evolved differently from most other anole species groups. All
three are known from moderate elevations (1000' to 3000')
( Fig. 6 ) , although the lower bound is far more meaningful than
the upper one. Even with the inclusion of A. koopmani, climatic
divergence in the monticola group seems relatively limited.

Unfortunately, distributional and ecological information is so
scant that interactions among these species are an open problem.
Only A. monticola has an extensive distribution. A series of
samples taken along the Les Cayes to Jeremie road indicates that
the subspecies monticola and quadrisartus are separated by the
Riviere Glace, a stream originating south of Duchity and flowing
north, disappearing into limestone hills. While A. rupinae as
described is known only from Castillon, the unassigned specimens
from two collections on the Les Cayes to Jeremie road suggest
the possibility of a broader distribution and contact with both
subspecies of monticola.

The scarcity of records for A. rupinae is understandable, since
its deep shade habitat is limited and very patchy in the highly
disturbed mid-elevations of the Massif de la Hotte. The apparent
restriction of A. koopmani to the Les Platons region is more
enigmatic. It can be common in a shaded coffee patch or in the
open brush growing in an abandoned citadel. Even on the Les
Platons plateau it does not always occur in such vegetation and
it appears to be absent also in comparable areas along the Les
Cayes to Jeremie road. T. C. Moermond has studied the anoles
of the Les Platons area and discovered A. monticola, but as yet
it has not been collected syntopically with A. koopm.ani. (Recall
that as A. rupinae is larger than A. monticola, A. koopmani is
smaller (Fig. 5)). Moermond (MS) has documented structural
habitat and foraging differences for the two.

The unusual karyology of the monticola group sensu stricto
has special interest in the context of the morphological, geo-
graphic and ecological relationships of its members. Departures
from the ancestral anoline condition that occur in the comple-
ments of all three can be attributed to centric fission (Webster,
et al., 1972) , a process that is known in few alpha anoles. Of the
six ancestral pairs of metacentric macrochromosomes, five or six
have fissioned in A. monticola to yield diploid numbers of 46
to 48. Two pairs fissioned to produce the diploid number of 40
in A. koopmani. As in A. monticola, in A. rupinae there is
polymorphism for macrochromosomal number with an inferred
range from six to eight pairs (i.e. none to two pairs fissioned).

1974 ANOLIS RUPINAE NEW SPECIES 17

In addition, A. rupinae seems to have one more pair (thirteen)
of microchromosomes than the ancestral complement (twelve),
a condition not previously reported. Whether this additional pair
originated by microchromosomal fission or in the course of change
in macrochromosomal number and morphology is unknown.

In addition to supporting the obvious close relationships within
the monticola group sensu strict o, the shared class of chromosome
change — fission — may have been critical in the origin and
di\'ergence of these species. A role for karyotypic differentiation
in the partial or complete genetic isolation of two populations
has been suggested by several authors (see Mayr, 1970; White,
1973; Hall, MS). In addition, chromosomal changes are key
elements in more complex evolutionary scenarios which envision
"cascading revolutionary speciation" (Hall, MS) or a genetic
release that accompanies extensive fissioning and favors adaptive
radiation (Todd, 1970). The components of these more elabo-
rate hypotheses — genetic revolutions, genetic effects of fission,
chance karyotypic change in small populations — are at present
individually such poorly documented phenomena that the larger
constructs are particularly open to criticism (see White, 1973 on
Todd, 1970). We suggest that the derived and complex karyol-
ogy of this small assemblage of anoles merits further study, both
as a possible aid to understanding their miniradiation but more
importantly as a system that may be relevant to larger evolu-
tionary issues.

ACKNOWLEDGMENTS

The discovery and study of Anolis rupinae have been sup-
ported by NSF grants B 019801X and GB 37731X to E. E.
Williams. We owe warm thanks to Lamy Camille of Port-au-
Prince for help with field work and to A. Ross Kiester for com-
panionship and assistance during the trip which first obtained
rupinae.

REFERENCES CITED

Hall, W. P. III. 1973. Comparative population cytogenetics, speciation

and evolution of the iguanid lizard genus Sceloporus. Ph.D. Thesis,

Harvard University.
Mayr, E. 1970. Populations, Species and Evolution, xv + 453 pp. Harvard

University Press.
MoERMOND, T. 1973. Patterns of habitat utilization in Anolis lizards.

Ph.D. Thesis, Harvard University.

18 BREVIORA No. 429

Rand, A. S. 1961. Notes on Hispaniolan herpetology. 4. Anolis koopmani,

new species, from the southwestern peninsula of Haiti. Breviora, No.

137: 1-4.
Rand, A. S. and E. E. Williams. 1969. The anoles of La Palma: aspects

of their ecological relationships. Breviora, No. 237: 1-19.
ScHOENER, T. W. 1970. Size patterns in West Indian Anolis lizards. II.

Correlations with the sizes of particular sympatric species — displacement

and convergence. Am. Nat. 104: 155-174.
Schwartz, A. 1973. A new species of montane Anolis (Sauria, Iguanidae)

from Hispaniola. Ann. Carnegie Mus. 44: 183-195.
. 1974a. A new species of primitive Anolis (Sauria, Iguanidae)

from the Sierra d Aboruco, Hispaniola. Breviora, No. 423: 1-19.

1974b. An analysis of variation in the Hispaniolan giant

anole, Anolis ricordi Dumeril and Bibron. Bull. Mus. Comp. Zool., 146:

89-146.
Thomas, R. 1966. A re-assessment of the fauna of Navassa Island. J. Ohio

Herp. Soc. 5: 73-89.
Thomas, R. and A. Schwartz. 1967. The monticola group of the lizard

genus Anolis in Hispaniola. Breviora, No. 261: 1-27.
Todd, N. 1970. Karyotypic fissioning and canid phylogeny. J. Theor. Biol.

26: 445-480.
Webster, T. P. and J. M. Burns. 1973. Dewlap color variation and elec-

trophoretically detected sibling species in a Haitian lizard, Anolis hrevi-

rostris. Evolution 27: 368-377.
Webster, T. P., W. P. Hall and E. E. Williams. 1972. Fission in the

evolution of a lizard karyotype. Science, N.Y. 177: 611-613.
White, M. J. D. 1973. Animal Cytology and Evolution. Third Edition.

viii + 961 pp. Cambridge University Press.
Williams, E. E. 1962. 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.
. 1972. The origin of faunas. Evolution of lizard congeners

in a complex island fauna. In Dobzhansky, Hecht and Steere, eds..

Evolutionary Biology 6: 47-89.

1974

ANOLIS RUPINAE NEW SPECIES

Ot-

o
o

? V

9 G

TJ
U

1
^

•M

>~^

<-»»

•»*

•<**

•*^

HnA

• •

• 1-1

t-^

-0

-a,
s

4-1

a
8

I V

I— (

I
CO

^—4

4->

CTJ

4-1

»-i

-d

£-

4-1

f-^

• IH

<L)

"cS

S-l

■3 S

2 ^

4-1

ii -T-(

Sh
i-H

OJ
(h

u
O

^ Oj

XJ ««

W3 W

Ih
U

'c3

I
w

4-1 gj

"^ M

•9

c«

CO o

' 4-1

•iH

-d
a>

*|H

F—l

-d
<i>
bO

Sh
CJ

l-H

!h
O

-d
1

-d

I— I
I

l-H

13
§

f-H

ca
+

f-H

4->

4-1

«4-(

•)

4-1

'a;

*fH

I— 1

r»

'%
V

-d

BREVIORA

No. 429

3
h

s
o

•s

o
o

-Si

•a

2 «

f^ ^ ^

a;
bo

o s:
(u ■>

175

>x he

C/3

3 +

b£)

j:3

• F-4

1— (

c
o

+ -5

cvr

fc.

1— H

S-i

1-1

«-)

4-1

-a

flj

1— H

C
u

x:
.£3

+ + I -S

5 I I +

3 11 +

;3
O
CO

• ^H

»-*•

p-H

-a

+ + +

c
>

3
a,

1-1

(J
o

C
a

£3

•-H

4-)

4-1

o
a,

to
Si
O

- a

1974

ANOLIS RUPINAE NEW SPECIES

»n

—

t«

7 a

be -

J2 o

-s:

•^

•*.»

4-1

•**

.s;

• •

•**j

■TS

•wa

••*

eft

<;U

4-1

(4-1
O
CO

T, V

00
I

—. O

I
O

CO
I

o
o

^ 2

I
in

^ o

o
o

o <>^

o
o

7 s
i2 CO

CO c

CTJ

c«

z
h

3
O

•s

C
(J

1— •
-J
CO

c«

•4— t

S-H

173

«2

CO
CO

C4H

1— «

4—)

3
(73

i-(

;-l

i-c

4-1

(L)

4>J

4—1

S-i

-o

C4-I

3
03

• pH

BREVIORA

No. 429

u
O

c
o

l-H

>
3
•<— >

£3

s;
o

>-.

•^H

«J

4-1

•'— >

<»

-3

.:ui

c»

c«

p— 4

4-1

«£4

• ^^

•iH

•>— >

u
O

■?i4

^1*

F-H

•••

£X

4->

•m

t«

C -

">*«

4-1
U-l

•••

-d

a.
a,

fee

4-1

o 2
^-^

3 *-•
w ?
& g

'3 "^

^-H

-3

^
u

• i-i
O

"oJ

•PH

• 1-4

4-1
• —4

3
O
3

o
0-,

3
O
3

3
O

•t

3
O

••-1

"rt

S-i

3
0

4-1

•fH

5«

-3

"rt

S-i

•fH

4-1

33
lU

4-1

4—)

4-*

4-)

_bJO

'a,

3
O
u

-■8

4-»

I-I

•3

4-1

(ti

.&■

'i-i

4-1

4-J

CX3

4-1

•3

t-l

•>H

B R E V I O R A

Vluseiiiii of Comparative "Zoology

us ISSN 0006-9698 /iDn o

JST^

CAMBRroGE, Mass. 28 March 1975 Number 430

—n

A NOUS MARC AN 01 NEW SPEGIESr

SIBLING TO AN O LIS CYBOTES:

DESCRIPTION AND FIELD EVIDENCE

Ernest E. Williams^

Abstract. A new species, Anolis marcanoi, very close to A. cy botes, is
described from the southern slopes of the Cordillera Central in the Domini-
can Republic. Differing from A. cy botes primarily in the species recognition
character of a red rather than a white or grey dewlap, it appears to be
surrounded by populations of A. cy botes and is also sympatric with that
species in a considerable part of its known range. Ecological differences
between the two species are not obvious, and it is possible that neither is
able to displace a resident population of the other.

In December 1966, Joel D. Weintraub, collecting in the
Dominican Republic, brought back from San Jose de Ocoa
two small lizards which in general morphology and in scale
characters appeared to be assignable to the species Anolis cybotes.
These specimens, although clearly juvenile, had rudimentary
reddish dewlaps. The color of the dewlaps immediately at-
tracted attention, since all the then known populations of A.
cybotes, a species widely distributed throughout Haiti and the
Dominican Republic, had yellow or grayish, more rarely orang-
ish pigmentation in the dewlap but never red except in popula-
tions at the extreme end of the Southwest Peninsula of Haiti,
which, while having a reddish dewlap, have also more or less
keeling on chest and belly scales, while the red-dewlapped form
from San Jose de Ocoa had perfectly smooth chest and belly
scales.

A search for the population from which Weintraub took his
specimens began in 1968, and over several subsequent summers
the evidence has built up that the red-dewlapped cybotes-Yikt

^Museum of Comparative Zoology, Cambridge, Mass. 02138

2 BREVIORA No. 430

Anolis is a new species quite distinct from A. cyhotes in electro-
phoretic characters but nearly indistinguishable in squamation
and identical in karyology\ The two species overlap spatially in
a complex way.

The new species is named for Professor Eugenio de Jesus
Marcano F., who helped so much in early investigations in the
Dominican Republic.

Type. MCZ 131837, an adult male from ca 5 km N La
Horma, Peravia Pro\'ince, Dominican Republic, collected by
Jonathan Roughgarden and local inhabitants, 18 July 1972.

Paratypes. All from Peravia Province, Dominican Repub-
lic. Same localitv as tvpe: MCZ 131846-75, J. Roughgarden
and local inhabitants collectors, 18 July 1972; MCZ 143437-43,
P. E. Hertz and R. B. Huey collectors, 2 August 1974. Lizard
markets vicinity of La Horma: MCZ 131824-42, local inhabi-
tants collectors, 19 July 1972; 1 km N Malaqueta on road to
Valle Nue\'o, W. E. Hall, E. J. Marcano and E. E. Williams
collectors, 1 Julv 1969; below pines, Sabana Larga, N of San
Jose de Ocoa: MCZ 117810, W. E. Hall, E. J. Marcano and
E. E. Williams collectors, 1 Julv 1969; San Jose de Ocoa: MCZ
104402-03, J. Weintraub coUector, 21 December 1966; 1.3 mi
S San Jose de Ocoa, 1400 feet: V 34068-79, A. Schwartz and
local inhabitants collectors, 19 November 1971; bridge over the
Rio Ocoa S San Jose de Ocoa: MCZ 107072-76, A. S. Rand,
E. J. Marcano and E. E. WilHams collectors, 27 Julv 1968;
MCZ 117809, 118606, W^ E. Hall, E. J. Marcano and E. E.
Williams collectors, 1 Julv 1969; MCZ 143247, P. E. Hertz and
R. B. Huev collectors, 22 Julv 1974: 3-5 km S San Jose de
Ocoa: V 21392-95, R. K. Bobilin and R. Thomas collectors,
24 July 1969; 16 km N Cruce de Ocoa: MCZ 143246, P.E.
Hertz and R. B. Huev collectors, 21 Julv 1974; 12 km N Cruce
de Ocoa: MCZ 143245, P. E. Hertz and R. B. Huev collectors,
22 Julv 1974; 3 km N Cruce de Ocoa, 500 feet: V 35815, A.
Schwartz and local inhabitants collectors, 27 December 1972;
coconut grove near Las Carreras on road to San Jose de Ocoa,
MCZ 115640, W. E. Hall, E. J. Marcano and E. E. W^illiams col-
lectors, 1 Juh 1969; Las Mayitas, 27 km S San Jose de Ocoa,

'Like A. cyhotes the new species has the 12 niaciochromosoine, 24 micro-
chromosome karyotype that occurs so frequently in iguanids and other
lizards (\V. Hall, personal communication) .

1975 ANOLIS MARCANOI 3

550 feet: V 15645, V 15598, J. K. Lewis collector, 3, 5 August
1968; 6 km N of Bani on road to El Recodo (just S of the first
ford), P. E. Hertz and R. B. Huev collectors, 20 July 1974; La
Jina, 7-8 km N of Bani on road to El Recodo: MCZ 143241^3,
143248-49, 143262, natives for P. E. Hertz and R. B. Huey
collectors, 20 July 1974; MCZ 143244, P. E. Hertz and R. B.
Huey collectors, 2 August 1974; 11 km N of Bani on road to
El Recodo: MCZ 143253-55, P. E. Hertz and R. B. Huey col-
lectors, 20 Julv 1974; 12 km N of Bani on road to El Recodo:
MCZ 143256-61, P. E. Hertz and R. B. Huey collectors, 20 July
1974; 13 km N of Bani on road to El Recodo: MCZ 143250,
P. E. Hertz and R. B. Huey collectors, 20 July 1974.

Head. Head moderately massive, snout to posterior border of
eye about as long as tibia. Head scales mostly smooth. Five to
nine scales across snout between second canthals. A shallow
frontal depression. Naris in front of canthal ridge. Anterior
nasal scale (sometimes divided) in contact with rostral.

Supraorbital semicircles in contact or separated by one scale
row, separated from the supraocular disks by single rows of
granules. Supraocular disks consisting of about six to eighteen
enlarged weakly keeled scales separated by about five rows of
scales and granules from the scales of the supraciliary rows. One
or two elongate supraciliaries continued posteriorly by a double
row of moderately enlarged scales. Canthus distinct, canthal
scales four, the second largest. Loreal rows four to seven, the
lower rows larger. Supratemporal area granular, grading into
moderately enlarged scales surrounding the interparietal. Inter-
parietal slightly larger or slightly smaller than ear, separated
from the supraorbital semicircles by one to three scales.

Suboculars separated from supralabials by one row of scales
or in contact, anteriorly grading into loreals, posteriorly grading
into large scales at the comer of the mouth. Six supralabials to
the center of the eye.

Mentals broad as long, usually in contact posteriorly with four
small throat scales. Infralabials narrow, in contact with two to
three large sublunate sublabials. Throat scales small, swollen,
not keeled; only the anterior ones elongate.

Trunk: Middorsal scales not abruptly larger than flank scales
(Fig. 1, compare also figures in WilHams, 1963). Ventrals much
larger than middorsals, cycloid, smooth. Postanal scales en-
larged, often broken into four.

Figure 1. Dorsal scales. Left: Anolis marcanoi, Paratype, MCZ 107075.
Right: A. cybotes, MCZ 115641. Both from the bridge over the Rio Ocoa
south of San Jose de Ocoa, Peravia Province, Dominican Republic.

1975 ANOLIS MARCANOI 5

Gular Jan. Large, scales smooth, no larger than ventrals.

Limbs and dibits. Hand and foot scales multicarinate. About
15-22 scales under phalanges 2 and 3 of fourth toe. Largest
scales of arm unicarinate, of leg smooth or very weakly multi-
carinate; those of arm smaller, those of thigh larger than ven-
trals.

Tail. Compressed, each verticil surmounted by four sharply
keeled scales, ventrally three pairs of somewhat larger, strongly
keeled scales.

Color. Brown with or without obscure dorsal blotching, head
sometimes distinctly reddish. Dewlap in males distinctly rose-
red at the edge, more orangish anteriorly and posteriorly, but
purplish or even bluish toward the center, the colors grading
into one another. Females with vestigial red-ringed throat fans
and usually distinct longitudinal dark streaks on the white of
chin and throat.

Differential characters. On the classic characters convention-
ally used in Anolis, especially preserved Anolis, A. marcanoi is
a poorly differentiated species. No scale characters will con-
sistently separate marcanoi from cybotes. (From the geographi-
cally adjacent related montane species, A. shrevei as well as A.
whitemani of the arid lowlands to the west, A. marcanoi is amply
distinct by its smooth rather than keeled ventrals.)

Some specimens of marcanoi have almost no enlargement of
the middorsal scales anywhere on the dorsum : in most the sacral
area shows the middorsal scales minimally or not enlarged. How-
ever, some specimens of cybotes and marcanoi — both sexes and
all ages — are impossible to distinguish by this character, i.e. in
these animals of both species the middorsal scales are weakly
enlarged. No other scale characters seen are even as useful as
this.

Color, then, is the major differential character, the male dew-
lap being especially obvious, but the red in the throat of females
is also highly diagnostic.

Distribution. The distribution of A. marcanoi is curiously
complex ( Fig. 2 ) . It is recorded from the area just south of the
first ford on the road to El Recodo, north of Bani (here A.
cybotes is also present) , and from La Jina, the village just beyond
the first ford (no cybotes obtained). Marcanoi is known as far

BREVIORA

No. 430

A shrevei

O cybotes

D marcanoi

Figure 2. The known distribution of Anolis marconoi. Squares: A. mar-
canoi. Circles: A. cybotes. Triangle: A. shrevei.

north on this road as the second ford, and presumably beyond
it, but this ford is impassable in a rental car. The new species
occurs also on the slopes of Loma de Pinos, just east of the road
which connects Constanza in the Cordillera Central via San Jose
de Ocoa with the road west from Santo Domingo to Barahona.
There are only sight records from this area. North of Cruce de
Ocoa on the west road there are occasional records of marcanoi
south of San Jose de Ocoa ; in these instances it is found on fence

1975 ANOLIS MARCANOI 7

posts or in coconut groves, apparently as enclaves with a wider
but sparse distribution of cyhotes in the surrounding acacia. At
the bridge just south of San Jose de Ocoa, and inside or in the
immediate environs of the city, both species occur broadly inter-
mingled. North of the city as far as La Horma, A. cyhotes is
known only from lizard markets in villages, while A. marcanoi
was collected on rocky hillsides, i.e. cyhotes now appears as en-
claves within populations of marcanoi. At lizard markets up to

4 km N of La Horma only ynarcanoi was obtained. A single
specimen of marcanoi is known from 9 km N of La Horma. At
1 3 km N of La Horma cyhotes reappears and, on the evidence of
three specimens of this species and no examples of marcanoi,
appears to separate marcanoi from Anolis shrevei, another cy-
hotes relative living on the peculiar cold plateau of Valle Nuevo.

Many more specimens have been seen and even collected than
have been preserved. Some of the material used for electrophore-
sis was collected by Thomas Jenssen from nine localities within or
near the city of San Jose de Ocoa: 8 km N San Jose de Ocoa
on road to Nizao, 2 km north of the city under the bridge over
the Rio Ocoa, 8 km N on road to La Horma, at the school in
the southwest end of town, 3 km W on road to El Pinar, 2 km

5 at bridge over the Rio Ocoa ( and along the river itself ) , 3 km
S along a small tributary of the Rio Ocoa.

In the vicinity of San Jose de Ocoa the two species occur
almost syntopically but nevertheless with some tendency to ex-
clusion. It is not easy anywhere to define an ecological difference
between the two species. The association with rocky, very open
hillsides is definite for marcanoi in the vicinity of La Horma
(hence at relatively high elevations), but in the lowlands near
the intersection with the west road marcanoi is known from a
shaded coconut grove. Presumably some combinations of tem-
perature and humidity may provide different optima for the two
species, but this is a physiological question not yet worked out.

DISCUSSION

"Sihling species.''

It becomes more and more obvious that, in addition to those
species in which museum taxonomists rejoice because they are
very distinct in terms of the characters conventionally studied,
there are in many groups valid biological species only imperfectly
separable on museum characters, if at all. This phenomenon is

8 BREVIORA No. 430

only interesting in terms of the history of museums, not of biology.
Museum techniques alter as taxonomy progresses. It will not be
necessary in the near future to defend or specially comment on
cases like that here described. Given that species status should
be recognized by any taxonomist on the full suite of characters
known for any population and not on the basis of some subset
selected because of convention or convenience, it is inevitable
that marcanoi be recognized as a full species.

The two juvenile specimens on which the discovery of mar-
canoi was based lack any trace of gular red after eight years in
alcohol. It would be difficult or impossible to separate them as
a distinct taxon now, were they all that was available. But this
is a failure of techniques, a museum failure like the failure of a
library with books printed with impermanent ink.

The biological phenomenon in marcanoi and cybotes that is
interesting is the way in which they overlap. On a large scale
map, marcanoi and mybotes do overlap over a considerable dis-
tance. Macro-geographically they are in part sympatric, but
quite clearly they are rarely syntopic. A. cybotes and A. mar-
canoi are in this regard rather similar to the Cuban homolechis-
allogus-sagrei series. As with marcanoi and cybotes, these color
differences and dewlap differences are more reliable than scale
differences; the latter are in fact few, minor, and usually bridged
by intrapopulational variation. In the Cuban series, as with
marcanoi and cybotes, there may be close physical juxtaposition.
A walk down a path through the woods on a Cuban finca might
find two species on adjacent trees, three species not far from one
another, but close examination would show that one species lived
in deep shade, one in half shade, one in open sun. Where the
environment, at the edges of these different habitats, juxtaposed
the three conditions of shade, half shade and sun, the lizard
species would also be juxtaposed, while where the environment
was homogenous over a larger area, there the lizard populations
would also be homogenous (Ruibal, 1961 ; Ruibal and Williams,
1961).

The relations between marcanoi and cybotes, however, ap-
pears to be subtler than that in the Cuban series. An inadvertent
experiment may demonstrate this point. The first series of mar-
canoi were taken in a grove of trees on the right bank of the
river at the bridge over the Rio Ocoa south of San Jose de Ocoa.
Only marcanoi was taken in this situation. In several subsequent
summers the grove of trees has been occupied by cybotes, never

1975 ANOLIS MARCANOI 9

by rnarcanoi, which instead has been found on rocks and fence-
posts on the open road above the grove. Our latest observations
found the area considerably altered and on the day of observation
neither species was taken in the grove. Our first ecological judg-
ment based on collections in the strove during," the first vear were
that marcanoi preferred shade and cybotes (presumably) sun.
But subsequent multiple observations both at the grove by the
ri\er and elsewhere have demonstrated this conclusion to be
wrong. Apparently cybotes and marcanoi do not respond to the
environment as litmus paper does to acid or base, or as the
Cuban species more nearly seem to do. On the contrary, simple
physical possession seems to be part of the story. By the act of
collecting we cleared an area of marcanoi. Cybotes was the
species that moved in and has held this small area ever since.
There may thus be situations — perhaps many situations — in
which the advantage to either species is so marginal that it can-
not dispossess a population in residence.

By this hypothesis cybotes and marcanoi differ as little physi-
ologically as they do morphologically. If this be true, it is espe-
cially interesting that the electrophoretic evidence presented by
T. P. Webster in Breviora 431 shows that the genetic base for
these very similar morphological and physiological phenotypes is
so sharply different. It is once again a lesson that phenotypic
similarity is an imperfect clue to the continuity of genetic systems.
Clearly no evidence can be neglected if our object is to establish
the reality of genetic discontinuity.

Acknowledgments. Field work has been supported by NSF
GB-37731X and previous grants to E. E. Williams. Thanks are
due to all those who so cheerfully participated.

LITERATURE CITED

RiiBAL, R. 1961. Thermal relations of five species of tropical lizards.

Evolution 15: 98-111.
RuiBAL, R. ANP E. E. Williams. 1961. The taxonomy of the AnoHs homo-

lechis complex of Cuba. Bull. Mus. Comp. Zool. 125: 209-246.
Williams, E. E. 1963. Anolis luhiteniani, new species, from Hispaniola

(Sauria, Iguanidae) . Breviora 197: 1-8.

B R E V I 0 R A

MUS. COMP. ZOOL

Miifeettiw of Comparative Zoology

APR 3 1975 us ISSN 0006-9098

CAMWiroGEvM^ss. 28 March 1975 Number 431

UNlVfcRSiTY

AN ELECTROPHORETIG COMPARISON

OF THE HISPANIOLAN LIZARDS
A NO LIS CY BOTES AND A. MARC AN 01

T. Preston Webster^

Abstract. Samples representing four localities — one for both species,
two for A. marcanoi, and one for A. cybotes — were examined. Results for
24 polypeptides are reported, of which 21 were studied in all individuals.
With each of 10 proteins individual identification is unequivocal or nearly
so. These data confirm the presence of two species in Peravia Province of
the Dominican Republic, verify the recognition of the red-dewlapped form
as the new species A. marcanoi, and indicate that successful hybridization
and introgression must be rare, if they occur at all.

Anolis cybotes and the newly described A. marcanoi (Williams,
1974) are so similar in morphology that no scale character will
consistently separate them. The latter was recognized only be-
cause its red dewlap contrasts with the yellow one of the former.
For anoles such a difference in dewlap color probably is im-
portant for reproductive isolation (Rand and Williams, 1970;
Webster and Burns, 1973). In addition, populations of the two
have been found side by side, but individuals are not known to
mingle freely. This interaction, which is characteristic of closely
related anoles, and the difference in dewlap color together pro-
vide sufficient evidence for the description of A. marcanoi. How-
ever, the great similarity of the two species invites additional
information on the extent to which they have diverged and per-
fected reproductive isolation. I report here a study that used
starch gel electrophoresis to examine some of their enzymes and
nonenzymatic proteins.

^Museum of Comparative Zoology, Harvard University, Cambridge. Mass.
02138

2 BREVIORA No. 431

MATERIALS AND METHODS

Seven samples were examined. Of 62 individuals collected in
October 1970 by T. A. Jenssen in the vicinity of San Jose de
Ocoa, Peravia Province, Dominican Republic, 42 were red-
dewlapped A. niarcanoi (sample 3a) and 20 were yellow-dew-
lapped A. cybotes (sample 4a). In July 1974 E. E. Williams,
R. B. Huey, P. E. Hertz, and R. Holt collected the remaining
Peravia Province samples: additional short series of both species
from San Jose de Ocoa (samples 3b and 4b) and A. marcanoi
from La Gina (sample 1 ) and from the type locality, 5 km N of
La Horma (sample 2). Sample 5 consists of 4 individuals from
Debarasse, Departement du Sud, Haiti, a locality a few kilometers
to the west of Jeremie, the type locality for A. cybotes. The
Jenssen collection was shipped ali\'e to Cambridge where the
lizards were bled and frozen, but all other series were frozen in
the field.

Methods of sample preparation and horizontal starch gel
electrophoresis are derived from Selander et al. (1971). Protein
stains and specific assays are similar to those current in work
with ^'ertebrates. Procedural details such as buffer systems best
suited for each protein and minor modifications to published
assay formulas are available from the author. With the exception
of hemoglobin and a plasma protein, all proteins were examined
in tissue homogenates. For some proteins, particularly indophenol
oxidase, better results were obtained from lizards frozen in Cam-
bridge than from those frozen in Hispaniola.

In many reports on genetic differentiation between vertebrate
populations, including an earlier report on AnoUs species (Web-
ster, Selander, and Yang, 1972), the results are expressed as
values of Rogers' coefficient of genetic similarity, S (Rogers,
1972j. Unfortunately, in some circumstances the effect of this
formula is counterintuitive. When a single locus is considered
and no alleles are shared by two populations, the expected sim-
ilarity is 0. If both populations are polymorphic, however, S is
nonzero. The results of this study are presented as Nei's normal-
ized identity of genes, / (Nei, 1972), which is consistently some-
what (2-7%) larger than S calculated for the same data.

For the computation of I, each polypeptide is treated as the
product of a single gene.

1975 ANOLIS MARCANOI 3

RESULTS AND DISCUSSION

Among the polypeptides examined in whole animal homos^e-
nates, the bands representing 21 could be interpreted with suf-
ficient consistency to be used in estimating relationships. Of
these, eight indicate complete or almost complete differentiation
of all populations of A. marcanoi from those of A. cybotes
(Table 1). In addition, samples 3a and 4a apparently do not
share variants of hemoglobin, plasma protein- 1, and indophenol
oxidase. For four of these proteins (hemoglobin, plasma pro-
tein-!, protein A, and lactate dehydrogenase- 1 ) the difference
in electrophoretic mobility is consistent, but so small that an
indi\idual expressing both variants could be confused with one
producing a single variant. The differences for 6-phosphoglu-
conate dehydrogenase, isocitrate dehydrogenase- 1, phosphoglu-
comutase-1, alcohol dehydrogenase, albumin, and peptidase can
be scored unequivocally.

Samples 3b and 4b and the majority of individuals in samples
3a and 4a were collected 2 km S of San Jose de Ocoa, along the
bed and banks of the Rio Ocoa. At this locality the two species
are common and in close contact. In such situations of parapa-
try or sympatry, discrete variation in the electrophoretic mobility
of proteins can be more informative than morphological differ-
entiation. Without genetic analysis or biochemical study of pro-
tein structure, interpretation of observed differences as allelic
variation is generally correct (see Johnson, 1973, for criticism
and enumeration of exceptions). Indeed, the inheritance of
interspecific differences in some proteins has been observed in
natural Anolis hybrids (Gorman et al., 1971; Webster, unpub-
lished) ; and patterns of phenotypic variation in anole popula-
tions can be explained by simple molecular and Mendelian
models. Differences in phenotypic frequencies thus indicate the
presence of reproductive isolation. Detection of isolation does
not depend on absolute separation and could be inferred even
from significant differences in allelic frequencies at a few loci.
For these samples, each of 11 loci indicates an absence of allelic
exchange. Species status for the populations has no reasonable
alternati\'e.

Since codominance is the rule for allelic variation at loci
encoding proteins (it was observed for all of the protein varia-
tion within these samples), electrophoretic data can also be used
to determine whether reproductive isolation is complete and

4 BREVIORA No. 431

whether occasional mismating leads to introgression. Thus the
absence from the San Jose de Ocoa samples of a single individual
heterozygous for one or more of the six clear allelic differences
suggests that introgression between the two species must be rare,
if it occurs at all. The samples are large enough to show that
Fi hybrid individuals must be uncommon but not so large as to
exclude their occurrence. Of course, failure to detect hybrid
individuals does not eliminate the possibility of attempted hy-
bridization, whatever its frequency, if the issue of such unions is
inviable.

A single individual in sample 1 of ^. marcanoi is the exception
to complete divergence of the two species on the basis of 6-
phosphogluconate dehydrogenase variants, A heterozygote for
the common variant of both species, it is not an Fi hybrid (no
A. cyhotes were collected at this locality). This situation cannot
be explained, nor does it require explanation. In extensive com-
parisons of sibling species the characteristic protein variants of
one are often found in low frequencv in the other (e.g., Prakash,
1969; Ayala and Powell, 1972; Webster and Burns, 1973). Had
larger samples and more populations been considered, there prob-
ably would be fewer loci indicating absolute separation.

Conspecific populations are quite similar, both throughout the
small known distribution of A. marcanoi and between A. cybotes
samples separated by 420 kilometers. The unsatisfactory in-
dophenol oxidase results — some individuals in sample 1 have
a variant like that of A. cybotes — provide the only evidence
for significant differentiation within A. marcanoi. Samples 4a
and 4b of ^. cybotes are essentially identical and are similar to
sample 5 for all but one polypeptide (Table 1). If sample 5 is
accepted as representing A. cybotes from the region of the type
locality, then, of the two species around San Jose de Ocoa, that
with the red dewlap has been correctly treated as the new species.
The difference between intraspecific and interspecific levels of
similarity is expressed as values of Nei's I in Table 2.

In nearly all interspecific comparisons involving at least 15
proteins, one or more has allowed an individual to be identified
with complete or almost complete confidence. For instance,
diagnostic proteins giving species assignment with 99% or greater
certainty were found in each of se\eral extensive comparisons of
Drosophila sibling species (Ayala and Powell, 1972). In this
comparison of A. marcanoi and A. cybotes, 10 proteins are diag-
nostic by the same criterion. Joint consideration of several, par-

1975 ANOLIS MARCANOI 5

ticularly the six having very distinct variants, should be sufficient
to assign any individual to either A. cy botes or A. marcanoi.
In fact, while the 1970 sample from San Jose de Ocoa was
divided without error on the basis of dewlap color, for the 1974
sample it was necessary to use the electrophoretic results to cor-
rect some of the casual field identifications of juveniles and fe-
males. Three A. cy botes were misclassed as A. marcanoi and one
A. marcanoi as A. cybotes.

Although unnecessary in the analysis of A. marcanoi and A.
cybotes, the magnitude of a genetic similarity coefficient like
Nei's / can be used arbitrarily to determine whether two allopat-
ric populations merit species status. The proteins merely provide
another class of phenotypic information to be used according to
established taxonomic procedure, but the genetic interpretation
is usually retained. A criterion for species recognition can be
established in the context of several studies of populations at
diverse taxonomic levels, as judged by morphology or observed
reproductive compatibility. Similarity values for conspecific pop-
ulations generally exceed 0.9, and exceptions are often associated
with insular isolates or other distinctive evolutionary situations
(see Selander and Johnson, 1973, for a review of such data).
Infraspecific taxa showing some reproductive isolation differ at
10 to 25% of their loci, which is 10 to 15 times as much diver-
gence as between local populations within those taxa (Ayala
et al., 1974) . I feel that a similarity value of 0.7 or less indicates
so much genie divergence that it is a fairly conserv^ative criterion
for species status. On this basis Anolis marcanoi certainly quali-
fies for recognition as a separate species: in comparisons with

A. cybotes J /is 0.62.

ACKNOWLEDGMENTS

Laboratory work was supported by NSF GB-37731X and
previous NSF grants to E. E. Williams. I thank those who col-
lected the samples for this study and P. Haas for writing a com-
puter program.

BREVIORA

No. 431

Table 1. Polypeptide Variation Within and Between Populations
of Anolis marcanoi and A. cybotes}

Sample
3a 3b

Polypeptide,
Variants-

Albumin

a
b

1.00

Protein A

a
b

1.00

Phosphoglucose
Isomerase

a
b
c
d

1.00

.99
.01

1.00

.22
.75

.02

.38
.62

.88
.12

Lactate

Dehydrogenase-1

a
b
c
d

1.00

.92

.08

.04
.96

1.00

Lactate

Dehydrogenase-2

a
b

1.00

.12

.88

Isocitrate

Dehydrogenase-1

a
b
c
d

1.00

.04
.77
.19

.95
.05

.85
.15

1.00

Malate

Dehydrogenase-1

a
b

1.00

.12

.88

Malate
Dehydrogenase-2

a
b

1.00

.12

.98

1.00

Alcohol

Dehydrogenase

a
b
c

1.00

.12

.88

.04
.96

.12
.88

Glutamic
Oxaloacetic
Transaminase-1

a
b

1.00

.99
.01

1.00

6-Phosphogluconate
Dehydrogenase

a
b
c
d

.06
.94

.98
.02

1.00

.62
.38

Phospho-

glucomutase-1

a
b
c

1.00

.92
.08

1.00

1975

ANOI.IS MARCANOI

Table 1 — Continued

Phospho-
gliiconuitase-2

a
b
c

.06
.83

.83

.07

.83

.85

.92
.02

.81
.19

1.00

.11

.17

.10

.15

.05

Peptidase

a
b
c

.17
.83

1.00

Fumarase

.02

.12

1.00

.99

1.00

.98

.88

1.00

.01

Indophenol
Oxidase

a
b

1.00

Hemoglobin

1.00

Plasma Protein-1

a
b

1.00

Troteins B and C, leucine aminopeptidase, isocitrate dehydrogenase-2,
a-glycerophosphate dehydrogenase, and glutamic oxaloacetic transaminase-2
were invariant.

-Electrophoretic mobility determines order in lists of variants, with 'a' the
most distant from the origin.

Table 2. Normalized identity of genes (/) as computed from 21 genes for
all pairs of samples.

Sample
Number

.996

.998

.997

.622

.611

.624

.998

.999

.618

.608

.623

.999

.614

.604

.619

.617

.606
.996

.622
.958
.951

8 BREvioRA No. 431

LITERATURE CITED

Ayala, F. J., and J. R. Powell. 1972. Allozymes as diagnostic characters
of sibling species of Drosophila. Proc. Nat. Acad. Set. USA 69: 1094-
1096.

, M. L. Trace Y, L. G. Barr, and J. G. Ehrenfeld. 1974. Ge-
netic and reproductive differentiation of the subspecies, Drosophila
equinoxialis caribbensis. Evolution 28: 24-41.

Gorman, G. C. P. Light, H. C. Dessauer, and J. O. Boos. 1971. Repro-
ductive failure among the hybridizing Anolis lizards of Trinidad.
Syst. Zool. 20: 1-18.

Johnson, G. B. 1973. Enzyme polymorphism and biosystematics: the
hypothesis of selective neutrality. Ann. Rev. Ecol. Syst. 4: 93-116.

Nef, M. 1972. Genetic distance between populations. Am. Naturalist 106:
283-292.

Prakash, S. 1969. Genie variation in natural populations of Drosophila
persimilis. Proc. Nat. Acad. Sci. USA 62: 778-84.

Rand, A. S., and E. E. Williams. 1970. An estimation of redundancy and
information content of anole dewlaps. Am. Naturalist 104: 99-103.

Rogers. J. S. 1972. Measures of genetic similarity and genetic distance.
Studies in Genetics VII (Univ. Texas Publ. 7213): 145-153.

Selander. R. K., M. H. Smith, S. Y. Yang, W. E. Johnson, and J. B. Gentry.
1971. Biochemical polymorphism and systematics in the genus Pero-
myscus. I. Variation in the old-field mouse (Peromyscus polionotus) .
Studies in Genetics VI (Univ. Texas Publ. 7103) : 49-90.

, AND W. E. Johnson. 1973. Genetic variation among verte-
brate species. Ann. Rev. Ecol. Syst. 4: 7.5-91.

Webster, T. P., R. K. Selander, and S. Y. Yang. 1972. Genetic variability
and similarity in the Anolis lizards of Bimini. Evolution 26: 523-5.35.

, and J. M. Burns. 1973. Dewlap color variation and electro-

phoretically detected sibling species in a Haitian lizard, Anolis bre-
virostris. Evolution 27: 368-377.

Williams, E. E. 1974. A new Anolis, sibling to A. cy botes. Anolis marcanoi
new species: description and field evidence. Breviora 430.

B R E V I 0 R A

Miiseiiiii of Comparative Zoology

us ISSN 0006-9698 ^^^S. COMp. ZOOL
LtBRARY

Cambridge, Mass. 28 March 1975 , -^Number 432

^^^ APRo 1975

EVOLUTION AND GLASSIFIGATI01\ro
OF PLAGODERM FISHESNjVErsjty

Robert H. Denison

Abstract. The assumption is made that within the Subclass Placodermi
a shoulder girdle that is short anteroposteriorly is primitive. Most orders
retaining this feature show distinctive specializations: thus the Rhenanida
are ray-like, the Ptyctodontida are chimaeroid-like, the Pseudopetalichthyida
have large, dorsal eyes, and many Acanthothoraci have dorsal nostrils. The
Slensioellida show few specializations are are believed to be the most primi-
tive known Placodermi, yet they possess the three characters that distinguish
the subclass: 1) gills anteriorly placed under the neurocranium; 2) a neck
joint between the neurocranium and synarcual; and 3) dermal bones. The
primitively short shoulder girdle becomes lengthened to form a thoracic
shield in several stages. Some Acanthothoraci add posterior lateral and
posterior dorsolateral plates. The Petalichthyida add a long ventral shield.
Primitive Arthrodira lengthen the lateral shield and close it behind the
pectoral fins which then attach through fenestrae. Finally, the Antiarcha
develop a long, boxlike shield and transform the spinal plates into peculiar
pectoral appendages. A phyletic classification of Placodermi is attempted.

INTRODUCTION

The Placodermi are a suborder of fishes whose known history
is practically restricted to the Devonian period unless, as some
think, they were ancestral to the chimaeroids. During that rela-
tively short time span they underwent a considerable radiation
and gave rise to 34 families and about 170 genera. In recent
years they have been the subject of considerable research by
many paleontologists. Yet, in spite of a great advance in our
knowledge of the group, there is still little agreement about their
evolutionary history and classification. This results from widely
different assumptions about what constitutes primitive or derived
characters within the group. Gross (1954) argued that an elon-

2 BREVIORA No. 432

gated thoracic shield such as occurs in early Arthrodira is primi-
tive, and the well-documented reduction of this shield within
Arthrodira may be adduced to support this. Westoll (1945)
likewise placed the long-shielded "Arctolepida" at the base of
his placoderm phylogeny, and Miles ( 1 969 ) has concluded that
the formation of a firm thoracic shield, together with the develop-
ment of a neck joint, was the fundamental placoderm adapta-
tion. On the other hand, Stensio in various works (e.g., 1969-
1971) has based his classification primarily on the pectoral fin
and endoskeletal shoulder girdle; following the fin-fold theory
of paired fin origins, he believes that the primitive state is
long-based pectoral fins together with an elongated endoskeletal
shoulder girdle for their articulation.

CHARACTERS OF PRIMITIVE PLACODERMI

In my opinion, neither of these theories is correct, and my
classification and phylogeny is based on the assumption that
within the Placodermi an anteroposteriorly short shoulder girdle
is primitive. The justification for this assumption is the fact that
a short exoskeletal shoulder girdle occurs in all other groups of
fishes with bony exoskeletons, and a short scapulocoracoid is
characteristic of Chondrichthyes. It is only in certain groups of
Placodermi, the Petalichthyida, Arthrodira, Phyllolepida and
Antiarcha, that the exoskeletal shoulder girdle is elongated to
form a thoracic shield, and this can be taken as an indication
that it is a derived state within Pisces and within Placodermi as
well. On the assumption, then, that a short shoulder girdle is
primitive within Placodermi, we may look at the groups that
possess this character for other primitive states. The classification
used in this discussion is given in the appendix, and is indicated
pictorially in the phylogenetic chart (Fig. 6); some parts of it
will be discussed later.

The following orders have a short exoskeletal shoulder girdle:
Stensioellida ( Stensioella )

Rhenanida {Gemuendina, Asterosteus, Ohioaspis, Jagorina)
Pseudopetalichthyida {Pseudopetalichthys, Paraplesiohatis)

Acanthothoraci {Palaeacanthaspis, Kosoraspis, Radotina,
Kolymaspis, Kimaspis)

Ptyctodontida (8 genera)

1975 PLACODERM FISHES 3

AM of these orders appear in the Lower Devonian; they show
the following characters which may be primitive :

Thoracic region.

1) The ventral shoulder girdle (Figs. 1-2, sh) consists of a
single pair of plates homologous either to the interolaterals or
anterior ventrolaterals of Arthrodira; between them a median
plate has been identified only in Ptyctodontida.

2) The lateral shoulder girdle consists only of anterior laterals
and anterior dorsolaterals, except in some Acanthothoraci (Fig.
IC) where posterior laterals and posterior dorsolaterals are also
present.

3) The spinal plates are absent, or small and doubtfully dis-
tinct, except in Acanthothoraci and some Ptvctodontida (Fig.
IC-D, Sp).

4) A median dorsal plate is probably absent in Stensioellida
and Pseudopetalichthyida.

5) Pectoral fins are narrow-based, even in Rhenanida where
the fins are much expanded distally ( Fig. IB).

6) There is no exoskeletal craniothoracic joint, except in
Ptyctodontida where it is developed differently than in Arthro-
dira and Antiarcha.

7 ) The anterior vertebrae are fused to form a synarcual ( Figs.
1-2, syn) which articulates with the occipital region of the neu-
rocranium (not known in Acanthothoraci).

Skull.

8 ) The neurocranium is long and slender with a long occipital
region, except in Ptyctodontida where it must have been short.

9 ) The dermal cranial roof bone pattern may be variable and
unstable with relationships between bones and sensory canals not
firmly established, except in Ptyctodontida.

10) Dermal cranial roof bones may be small and part of the
rpof may be covered with thin, superficial tesserae in Acantho-
thoraci (Fig. 3, te) and Rhenanida; much of the skull in Sten-
sioellida (Fig. 2A) is covered with denticles or tesserae; the
central part of the cranial roof of Pseudopetalichthyida is cov-
ered with small dermal bones, but there may have been denticles
or tesserae elsewhere. Denticles or tesserae are unknown in
Ptvctodontida, but mav have covered the snout and cheeks where
dermal bones are largely absent (Fig. ID).

Figure 1. Placoderrai with short shoulder girdles: A, Order Pseudope-
talichthyida (ventral view of Pseudopetalichthys problematica, X0.66, from
Gross, 1962) ; B, Order Rhenanida (ventral view of Gemuendina stuertzi,

X0.60, from Gross, 1963) ; C, Order Acanthoihoraci (lateral view of shoulder
girdle of Palaeacanthaspis vasta, X0.94, from Stensio, 1944); D, Order
Ptyctodontida (lateral view of head and shoulder girdle of Rhamphodopsis
thrdplandi, X2.5. fiom Miles, 1967). Adl, anterior dorsolateral plate; Al,
anterior lateral plate; ba, basal elements of pectoral fin; br, branchial arches;
en, endocranium; Ig, lower jaw; It, lower dental plate; Md, median dorsal
plate; mk, Meckel's cartilage; or, orbit; pcf, pectoral fin; Pdl, posterior
dorsolateral plate; PI, posterior lateral plate; pvf, pelvic fin; sh, shoulder

girdle; Sp, spinal plate; syn, synarcual; ut, upper dental plate.

1975 PLACODERM FISHES 5

Jaws and Gills.

1 1 ) The jaws, where known, are more or less transverse and
lack large dermal elements (Fig. lA-B, 2B), except in Ptycto-
dontida where they are directed more anteroposteriorly and carry
large crushing or sectorial tooth plates (Fig. ID, ut, It).

12) Gill covers (submarginals) may be present, though they
are not known in Acanthothoraci and their dermal bones are
small in Ptyctodontida.

Sensory organs.

1 3 ) The orbits are small and lateral in Stensioellida and most
Acanthothoraci, large and dorsolateral in Ptyctodontida, and
dorsal in Pseudopetalichthyida, Rhenanida and one late genus
of Acanthothoraci; the last condition is surely specialized.

14) The nostrils are known only in Rhenanida and Acantho-
thoraci (Fig. 3, no) where they are usually dorsal, a condition
that is surely specialized. In Stensioellida, Pseudopetalichthyida
and primitive Acanthothoraci they are presumed to be anterior
or anteroventral ; there are no clues to their position in Ptycto-
dontida.

Body and fins.

15) The body is depressed and tapers to a diphycercal tail
(not known in Acanthothoraci) .

16) Dorsal fins are little developed except in Ptyctodontida;
there are dorsal ridge scales in Pseudopetalichthyida (Fig. lA)
and Stensioellida and the latter has a small dorsal fin (Fig. 2 A,
df ) at the base of the tail; an enlarged ridge scale forms a small
dorsal spine in Rhenanida. (This region is not known in Acan-
thothoraci.)

17) Pelvic fins (Figs. 1-2, pvf) are long-based and semicircu-
lar in Rhenanida, Stensioellida and Pseudopetalichthyida; they
are specialized by the development of claspers in male Ptycto-
dontida.

Histology.

18) The histology of the Lower Devonian members of the
groups under discussion is practically unknown. There is a pos-
sibihty that the Stensioellida had denticles composed of dentine,
and if so, this would be the only occurrence of this tissue in
Placodermi except for the tooth plates of Ptyctodontida. Typi-
cally in other Placodermi the superficial layer is reduced and the
external part of dermal bones is composed of semidentine or bone.

6 BREVIORA No. 432

PRIMITIVENESS OF PLACODERM ORDERS
WITH SHORT SHOULDER GIRDLES

In reviewing the list of probable primitive characters, it is
clear that the Ptyctodontida (Fig. ID) do not share many of
them. This may be due to the fact that only the shoulder girdle
is known in Lower Devonian ptyctodonts while other characters
are determined from Middle or Upper Devonian genera which
are specialized or advanced in the following ways: the presence
in the shoulder girdle of an anterior medioxentral, a median
dorsal, spinal plates in some, and an exoskeletal craniothoracic
joint; in the shortness of the exo- and endocranium, well-estab-
lished cranial roof pattern without tesserae (except perhaps
anteriorly and on the cheeks), large dorsolateral eyes, large
dermal jaw elements, firm attachment of palatoquadrate to en-
docranium, dorsal fins, and pelvic fins with claspers in males.

It is clear that the ray-hke Rhenanida (Fig. IB) are also spe-
cialized, even in the earliest known Lower Devonian forms. They
have a much flattened body, greatly expanded pectoral fins,
dorsal eyes and nostrils, a median dorsal plate, and a dorsal spine
on the bodv.

The Acanthothoraci, with the exception of the Radotinidae,
are advanced in having the lateral parts of the shoulder girdle
lengthened by the addition of posterior lateral and posterior
dorsolateral plates (Fig. IC) ; well-developed, projecting spinal
plates as well as median dorsal plates are present. The skull in
all members of the order is distinguished by its narrow propor-
tions, subparallel sides, and deeply embayed posterior margin
with strongly projecting paranuchals. Primitively, (Palaeacan-
thaspidae) the eyes were lateral and the nostrils probably ventral,
but the nostrils, or both the nostrils and eves have moved to the
dorsal side in Radotinidae (Fig. 3) and Kolymaspidae, both of
which have a prominent rostrum.

The poorly known Pseudopetalichthyida (Fig. lA) are surely
speciahzed in their relatively large, dorsal eyes, the long preorbital
region, and possibly in the absence of tesserae, at least on the
cranial roof. Their jaws (Fig. lA, Ig), though not well under-
stood, appear to be peculiarly specialized.

This leaves only the Stensioellida, which exhibit very few
characters that can be interpreted as advanced, and are con-
sidered to be the most primitive Placodermi known, even though

1975

PLACODERM FISHES

they are not the earliest members of the subclass. Based on the
two specimens of Stensioella (Fig. 2) from the Hunsriickschiefer
of Germany, the body appears to be somewhat depressed, broad-
est in the head and shoulder regions, and tapering backwards
towards the tail. Flattening after burial spread apart the two
halves of the shoulder girdle (Fig. 2, sh), making it difficult to

— pcf

Figure 2. Stensioella heintzi (the only known reprepresentative of the
Order Stensioellida) , X0.44, from Gross, 1962: A, dorsal; B, ventral, art,
jaw articulation; br, branchial arches; Ce, central plate; df, dorsal fin; dn,
denticles; en, occipital region of endocranium; hy, hyomandibular; mo,
mouth; pcf, pectoral fin; pvf, pelvic fin; sh, shoulder girdle; syn, synarcuaL

8 BREVIORA No. 432

interpret, but the bones are tuberculate and thus largely exo-
skeletal, even though individual dermal bones cannot be identi-
fied. Clearly the shoulder girdle is short anteroposteriorly, lacks
a median dorsal and median ventral, and has no large or pro-
jecting spinal plates. Each half of the shoulder girdle has an
inner or medial lamina which forms a postbranchial wall; such
a wall occurs in many placoderms, but is absent in primitive
Arthrodira, so this mav well be an advanced character in Sten-
sioella. The pectoral fins (Fig. 2, pcf) are narrow-based, scale-
covered and with ceratotrichia distally, but their inner skeleton
is unknown. There is no exoskeletal craniothoracic joint, but
apparently there is a synarcual formed of fused anterior verte-
brae (Fig. 2, syn) that articulates with the occipital region of
the endocranium. The body is covered with denticles (Fig. 2,
dn) which possibly ha\'e pulp ca\ities and thus perhaps were
composed of dentine, and possibly, though not certainly, were
attached to thin tesserae. There are long-based, semicircular
pelvic fins (Fig. 2, pvf), and a small, delicate dorsal fin (Fig. 2,
df ) at the base of the tail, the termination of which is unknown.

Judging from its manner of preservation, the head and body
were depressed dorsoventrally, but only moderately broad. The
neurocranium must have been long and relatively slender, but
was poorly ossified, except in the occipital region where an articu-
lation was developed for the synarcual (Fig. 2, en). The dermal
covering of the head was largely denticles, possibly attached to
tesserae, but there are at least three small bones with radiating
structure — a median postpineal and paired centrals ( Fig. 2A,
Ce). The orbits have not been seen, but must have been lateral,
and the nostrils do not appear on the dorsal surface of the head
so are assumed to be anterior or anteroventral. The supraorbital
sensory canals are bounded by large tubercles and are presumably
quite superficial; from the snout they run subparallel back to the
middle of the skull. Posterior pit lines are shallow grooves on the
central plates. The mouth (Fig. 2B, mo) is ventral, but only a
short distance behind the rostrum. The palatoquadrates and
Meckel's cartilages carry no dermal jaw bones, only small denti-
cles. As interpreted by Gross (1962), the jaw suspension was
hyostylic, but this is not certain. There appear to be five bran-
chial arches (Fig. 2B, br) and these extend far anterior under
the endocranium.

The single species that constitutes the Order Stensioellida has
many characters that are considered primitive within the placo-

1975 PLACODERM FISHES 9

derms, but shows no easily identifiable specializations or unique
derived characters that can be used to distinguish it from other
placoderni orders. Nonetheless, it seems to be a distinct order
occupying an isolated position as an offshoot from the base of
the placoderm stem.

DIAGNOSTIC CHARACTERS OF PLACODERMS

That Stensioellida are placoderms is indicated by their posses-
sion of three characters : 1 ) the gills lie far forward under the
neurocranium ; 2 ) there is a neck joint between the endocranium
and synarcual; and 3) there are dermal bones on the head and
shoulder girdle. The first two characters are shared by the Holo-
cephali which may support, though it does not establish, their
postulated relationship to Placodermi. But the possession of all
three features is unique to Placodermi, and for that reason their
significance requires further consideration.

Miles (1967, 1969) attempted to show that the neck joint
arose to compensate for the rigidity of the anterior part of the
body when it became enclosed within a thoracic shield. How-
ever, this joint occurs in the placoderm orders discussed above
which have a short shoulder girdle and no rigid thoracic armor.
The same is true in chimaeroids so one may question whether it
was the evolution of a stiff armor that led to the development
of the neck joint. The joint permits largely vertical movement
between the head and shoulder girdle and functions in three main
ways (Miles, 1967) : 1 ) to aid in locomotion by control of pitch-
ing equiHbrium ; 2 ) to aid in feeding by permitting a wider gape
and by helping to force food into the esophagus; and 3) to aid
in respiration by forcing water through the gills. The first was
probably of only minor importance to early placoderms which
were slow-swimming, benthonic forms. The second may have
been important to some later, predaceous placoderms, but the
early ones had small mouths and surely ate small food that did
not require a wide gape. However, the neck joint may have been
necessary for respiration when the gills became crowded under
the neurocranium; then, a raising and lowering of the head
would help to force a stream of water through the gills. Thus
ithe neck joint may have been related to the anterior position of
the gills under the head; instead of being a response to the
rigidity of the thoracic region, it may have permitted the later
development in some groups of a stiff trunk armor.

BREVIORA

No. 432

Dermal bones are characteristic of Placodernii, and typically a
superficial layer of dentine is absent and their surface is formed of
semidentine or bone. In the Lower Devonian groups with a short
shoulder girdle, specimens are either unavailable or unsuitable
for histologic study so superficial tissues have not been identified.
Primitive or ancestral Placodermi might be expected to retain
dentine in teeth, denticles or tubercles, and Gross (1962) has
recognized what may be pulp cavities in the denticles of Sten-
sioella, suggesting that they were made of dentine. Lower
Devonian Rhenanida have not been studied histologically, but
the Middle Devonian members have semidentine superficially.
Ptyctodontida have dentine in their tooth plates (0rvig, 1957),
the only occurrence of this tissue in later Placodermi.

The problem of dermal bone origins in placoderms is com-
plicated by the presence of tesserae in certain groups — the
Rhenanida, Acanthothoraci, Lower Devonian Petalichthyida,
and possibly Stensioellida. Since tesserae occur mostly in early

Figure 3. Radotina kosorensis, dorsal view of incomplete cranial roof,
X0.9, from Gross, 1958. Ce, central plate; 11, main lateral line; no, nasal
opening; or, orbit; pp, posterior pit line; Pro, preorbital plates; Pto, post-
orbital plate; Ro, anterior plate perhaps homologous to rostral or pre-
median; soc, supraorbital sensory canal; te, tesserae.

1975 PLACODERM FISHES 11

forms, they are probably a primitive character, as has been
maintained by Gross (1959). He has shown in Radotina (Fig.
3, te) that tesserae are thin, superficial structures that occur for
the most part between bones, and that do not fuse together to
form bones or even their superficial parts. In Rhenanida they
are homologous to the scales that cover the body (Gross, 1963).
They may be considered remnants of the dermal scales that were
the only exoskeleton of ancestral placoderms, and as such are
comparable in general to chondrichthyan scales. When bones
first appeared in placoderms, they apparently arose deeper in
the dermis quite independently of the tesserae and also of the
lateral line system. The depth of their formation may account
for the absence of any true dentine on the bones of typical Pla-
codermi, and also for the fact that the course of the lateral line
canals in Rhenanida and Acanthothoraci is not dependent on
the dermal bones. Presumably the close relationship between
the dermal bones and lateral line canals was secondary and, as
suggested by Parrington ( 1 949 ) , the precursors of dermal bones
may later have come to influence the direction of growth of
lateral line primordia.

The pattern of dermal bones on the skull differs in the various
groups of Placodermi yet shows enough similarities to suggest
that, in most cases at least, it was derived from a common an-
cestral pattern. In Stensioellida the pattern is hardly developed
for in the cranial roof there is onlv a median bone identified as
a postpineal and paired bones that resemble centrals (Fig. 2 A,
Ce). Likewise in the Lower Devonian Rhenanida the cranial
roof largely lacks dermal bones, though laterally there are sub-
orbitals, submarginals and possibly paranuchals. In all other
groups, except perhaps the poorly known Pseudopetalichthyida,
the skull bones are developed according to a similar pattern.
This pattern includes some or all of the following : 1 ) median
nuchal, postpineal, pineal and rostral; 2) paired centrals over
the otic region; 3) paired paranuchals and marginals carrying
the main lateral line forward; 4) paired pre- and postorbitals
o^^er the orbits; 5) paired postnasals beside the nostrils; and
6) paired suborbitals, postsuborbitals, postmarginals and sub-
marginals in the cheek and opercular region. Much of this pat-
tern is becoming established in the Acanthothoraci (Fig. 3),
while in Ptyctodontida, Arthrodira, Phyllolepida and Antiarcha
there are relatively stable cranial bone patterns, though with
characteristic modifications in the various subgroups (Figs. 4-5).

12 BREVIORA No. 432

PHYLETIC HISTORY OF PLACODERMI

In my theor\' of placoderm evolution, as presented pictorially
in the phylogenetic chart (Fig. 6), particular emphasis is given
to the dermal shoulder girdle. This remains short in Stensioellida,
Pseudopetalichthyida, Rhenanida and Ptyctodontida, while the
first steps towards lengthening it to form a thoracic shield are
seen in some Acanthothoraci (Palaeacanthaspidae and Koly-
maspidae), where posterior laterals and posterior dorsolaterals
are added (Fig. IC, PI, Pdl). The second stage is the develop-
ment of a ventral shield composed, in addition to interolaterals,
of anterior and posterior ventralaterals and anterior and posterior
medioventrals ; this is seen in Petalichthyida and Arthrodira.
Early members of the latter group go one step further in uniting
the posterior parts of the ventral and lateral shields behind the
pectoral fins to enclose pectoral fenestrae (Fig. 5B-F, pf). The
Antiarcha have the longest thoracic shield and have a posterior
median dorsal incorporated in it (Fig. 5K, Pmd).

There are three major phylogenetic problems that require
special mention, the first involving the Petalichthyida (Fig. 5A).
Their thoracic shield might have evolved quite independently
from that of Arthrodira, in which case a relationship to Pseudo-
petaUchthyida should be considered. However, since the latter
group is so poorly known and the petalichthyid thoracic shield is
so similar to that of Arthrodira, this theorv has little to recom-
mend it. Secondly, the petalichthyid thoracic shield may have
arisen as a result of a posterior reduction of the lateral parts of
the arthrodire shield. There is no evidence to support this, and
in fact it is quite unlikely that the petalichthyid cranial roof was
derived from the arthrodire type, so this theory is rejected. The
third theory is that the petalichthyid thoracic shield represents
an intermediate evolutionary stage, more advanced than in Acan-
thothoraci in the possession of a ventral shield, but less advanced
than early Arthrodira as the pectoral fins are completely behind
the shield. This theory seems most probable and is supported by
the retention of certain primitive characters in Petalichthyida,
such as the two pairs of paranuchals and tesserae on the cheeks.

The evolutionary position of Phyllolepis (Fig. 4) is also con-
troversial because, though it has a moderately long thoracic
shield, it lacks posterior laterals and posterior dorsolaterals. Is
the absence of these plates the result of a phyletic reduction, or
did Phyllolepis branch off the arthrodiran ancestral line before

1975

PLACODERM FISHES

?Ptn

Figure 4. Phyllolepis o-rvini, dorsal view of cranial and thoracic shields,
X0.2, modified from Stensio, 1936. Adl, anterior dorsolateral plate; Al,
anterior lateral plate; cc, central sensory canal; ioc, infraorbital sensory
canal; Ic, main lateral line; Md, median dorsal plate; Mg, marginal plate;
Nu, nuchal plate; Pnu, paranuchal plate; pp, posterior pit line; Pro, pre-
orbital plate; ? Ptn, possible postnasal plate; Pto, postorbital plate.

these plates were acquired? Since this genus is known only from
the late Famennian there is little evidence to decide this question.
However, the genus Antarctaspis, known only from an imperfect
cranial roof, seems in some ways to bridge the gap between Phyl-
lolepis and primitive Actinolepina, which suggests that Phyllole-
pifia were derived from the latter by a reduction of the thoracic
shield, and, of course, by considerable modification of the cranial
roof.

The Ptyctodontida (Fig. ID) are a third phyletic problem.
If it is accepted that their short dermal shoulder girdle is a prim-
itive character and not the result of reduction, they cannot be
derived from Arthrodira, Petalichthvida or some Acanthothoraci.

14 BREVIORA No. 432

Yet in their dermal cranial bones thev show many resemblances
to these groups, so they probably had an ancestor with a short
shoulder girdle and the basic placoderm cranial bone pattern.
The Radotinidae are the only known group that satisfies these
conditions, but because of their elongated skull and dorsal nos-
trils (Fig. 3, no), their relationship to Ptyctodontida will be
questioned, particularly by those who belie\e the ptyctodonts
were ancestral to chimaeroids. However, it must be pointed out
that nothing is known about the position of the nostrils in ptycto-
donts.

The Order Arthrodira is the best known and most varied
group, including currently 121 genera or 72% of known placo-
derm genera, yet the classifications that have been proposed for
it have been largely by level of organization, rather than phylo-
genetic. This is true of the commonly used major subdivisions,
the Arctolepida (or Dolichothoraci ) and Brachythoraci (and its
two subgroups, the Coccosteomorphi and Pachyosteomorphi) .
It appears to be worthwhile to attempt a phyletic classification,
even though our incomplete knowledge will make this provisional
and certainly subject to future corrections and additions. Instead
of the two to four gradal subdivisions of current usage, the 21
arthrodiran families are grouped according to their probable
common ancestry in 8 suborders.

Figure 5. Cranial and thoracic shields of Placodermi with elongated
shoulder girles, lateral views except I. A. Order Petalichthyida (Lunaspis
herohU, after Stensio, 1963) ; B, Suborder Actinolepina (Sigaspis lepidopJiom,
after Miles, 1973) ; C, Suborder Phlyctaeniina {Phlyctaenius acadica, after
Heintz. 1934 and ^Vestoll and Miles, 1963) ; D, Suborder Wuttagoonaspina
(Wiittagoonaspis fletcheri, attempted restoration based on figures of Ritchie,
1969 and 1973) ; E, Suborder Holonematina {Holoyiema xvestolli, after Miles.
1971) ; F, Suborder Coccosteina (Coccosteus nispi'datiis, after Miles and
Westoll, 1968) ; G, Suborder Pachyosteina (Rhinosteus parvulus, after Stensio.
1963) : H, Suborder Brachydeirina (Leptosteiis hickensis, after Stensio.
1963) ; I-J, Suborder Heterostiina (Heterostins ingeyjs) ; I. dorsal view of
cranial and thoracic shields, after Heintz. 1929; J. lateral view of thoracic
shield, after Heintz, 1929; K, Order Antiarcha (Pterichihyodes milleri, after
Traquair, 1914) . Adl. anterior dorsolateral plate; Al. anterior lateral plate;
Amd. anterior median dorsal plate; art, cranio-thoracic joint; Ce, central
plate; Md, median dorsal plate; ng. nuchal gap; Nu, nuchal plate; or, orbit;
pa, pectoral appendage; Pdl, posterior dorsolateral plate; pe, pectoral
emargination; ]:)f, pectoral fenestra; PI, posterior lateral plate; Pmd. pos-
terior median dorsal plate; Pnu, paranuchal plate; So, suborbital plate; Sp,
spinal plate.

1975

PLACODERM FISHES

The first suborder to appear and surely the most primitive is
the Actinolepina (Fig. 5B), with a single family, the Actinolepi-
dae. It has the elongated thoracic shield that typifies early
Arthrodira, and it is closed behind the pectoral fins to form
pectoral fenestrae (Fig. 5B, pf), as is characteristic of primitive
members of the order. The spinal plates are well de\'eloped and

16 BREVIORA No. 432

projecting but not greatly elongated (Fig. 5B, Sp), the pectoral
fins are narrow-based, the median dorsal short and broad (Fig.
5B, Md), the endocranium platybasic, the orbits small and an-
terior, and the rostral region containing the nasal capsules some-
times separately ossified. All of these characters are primitive
within Arthrodira, though some are advanced for Placodermi.
However, Actinolepina are distinguished from other Arthrodira
by one feature that is clearly derived: there is a sliding joint
between the cranial and thoracic shields formed by smooth,
anterior flanges on the anterior dorsolaterals that are overlapped
by the underside of the paranuchals. No doubt there were a
number of phyletic Hues within the Actinolepidae ; one of them,
represented by Baringaspis and Aethaspis, shows a tendency to
reduce the centrals and elongate the nuchal, sometimes by fusion
with the postpineal. It is from this line that the Antarctaspidae
and Phyllolepidae may have been derived.

Another line retained a typical actinolepid thoracic shield
with a sliding craniothoracic joint (if Ritchie's 1969 restoration
is correct), yet modified the cranial roof so greatly that it has
been placed in its own suborder, the Wuttagoonaspina (Fig. 5D),
The cranial modifications resulted from great enlargement of
the nuchal plate and a migration of the eyes backwards.

Though Miles has recently (1973) expressed a contrary opin-
ion, it seems probable that in some Actinolepidae the sliding type
of neck joint evolved into a more complicated and efficient
ginglymoid articulation, with condyles developed on the anterior
dorsolaterals and glenoid fossae on the paranuchals. It is the
acquisition of this joint (Fig. 5C, art) that particularly distin-
guishes the Phlyctaeniina from their ancestors among the Actino-
lepina, and the joint is retained, with one exception, in all the
many descendants of the Phlyctaeniina. The dominant family,
the Phlyctaeniidae, showed a tendency to elongate the median
dorsal plate (Fig. 5C, Md), though one genus retained the short,
broad type of Actinolepidae, and many of the known genera
became specialized in their excessively long spinal plates (Fig.
5C, Sp). The Williamsaspidae may be a differently specialized
side-branch of Phlyctaeniina, but this is uncertain since their
skull and dorsal part of the thoracic shield is unknown.

An early and distinctive branch from the Phlyctaeniina is the
Holonematina with the single family Holonematidae (Fig. 5E).
Their skulls are distinguished by the large pineal plate lying

1975 PLACODERM FISHES 17

between the preorbitals, the orbits that deeply notch the cranial
roof, and the moderately small, subtriangular nuchal. The tho-
racic shield remains long or is even lengthened, and retains the
contacts between the lateral and ventral shields behind the pec-
toral fins. The anterior laterals (Fig. 5E Al) tend to lengthen,
crowding the pectoral fins backwards. The posterior laterals are
large (Fig. 5E, PI), and there is a large anterior medioventral.
Characteristically the main lateral line extends towards the pos-
teroventral corner of the anterior dorsolateral and has a strong
flexure on the posterior dorsolateral. Primitive members of the
suborder have previously been referred to the Groenlandaspididae
which, until the recent discoveries of Ritchie (1974), have been
of uncertain affinities.

The Suborder Coccosteina (Fig. 5F), the most important
derivative of the Phlyctaeniina, may be recognized by the nuchal
plate which is trapezoidal in shape and widened posteriorly, by
the paranuchals which are narrow posteriorly except for strong
postnuchal processes, and by the centrals which tend to be
divided into anterior, lateral and posterior lobes. The orbits
typically are directed more laterally than in Phlyctaeniidae, and
the pineal comes to lie posteriorly between the preorbitals. In
the thoracic shield, the median dorsal, which is primitively rather
long, tends to be shortened ; the pectoral fenestrae are lengthened
though usually remain closed posteriorly (Fig. 5F, pf). The
spinals tend to be reduced (Fig. 5F, Sp), and the ventral shield
is typically lengthened. These characters are well displayed by
the Family Coccosteidae, which is also distinguished by the post-
branchial laminae projecting from the mesial faces of the anterior
laterals, by the course of the main lateral lines parallel to the
ventral exposed edges of the anterior dorsolaterals, and by the
long, slender suborbital processes of the suborbital plates. The
Gemuendenaspidae show their relationship to the Coccosteina
in the shape of the dermal bones of the posterior part of the
cranial roof, but retain a number of primitive characters, such
as the broad, depressed shield, the long, narrow median dorsal,
and the short, deep suborbital processes on the suborbital plates.
The Buchanosteidae also have the characteristic nuchal and
paranuchal plates of Coccosteina, but show a peculiar mixture
of primitive and specialized characters: they are primitive in not
having the rostral capsule fused to the rest of the skull, in the
forwardly directed orbits, in the short, deep suborbital processes,
and in the short, wide preorbitals; but they are distinctively spe-

18 BREVIORA No. 432

cialized in the long postmarginals, the unusually shaped anterior
laterals which bend inwards to form postbranchial laminae, and
in the short, nonprojecting spinals. A specialized family known
only in the Frasnian, the Pholidosteidae, is distinguished by its
enlarged eyes and elongated orbitotemporal region, by having
the cheek bones rigidly sutured to the cranial roof, and by their
long, laterally projecting spinal plates carried by protruding
wings of the anterior laterals and anterior ventrolaterals. This
family must have diverged early from the Coccosteidae before
the reduction of the spinals. The Homostiidae (including both
typical Homostiidae and Euleptaspidae) show a relationship to
the Coccosteina in the characteristically shaped nuchal, para-
nuchals, and centrals, and their appearance in the Siegenian
suggests an origin from early members of the suborder. The
family includes large forms with a broad, depressed head and
body, and is characterized particularly by the great elongation
of the bones of the posterior half of the cranial roof. The ad-
vanced Homostiidae are highly specialized in the dorsal position
of the eyes and in the great shortening of the thoracic shield, but
retain some primitive characters such as a narrow nuchal gap
and tuberculated dermal bones, Finallv, the Rachiosteidae are
shown to be Coccosteina by the shape and proportions of the
nuchal, paranuchals and centrals, but have reduced the lateral
and ventral thoracic shields even more than in some advanced
Pachyosteina, and have also lost the ornamentation on their
dermal bones.

The Pachyosteina (Fig. v5G), the dominant placoderms of the
Upper Devonian, are probably, though not certainly, a mono-
phyletic group derived from the Coccosteidae. They are char-
acterized particularly by a thoracic shield shortened dorsally and
laterally, anterior laterals reduced ventrally to slender bones
(Fig. 5G, Al), reduced or lost spinals, and pectoral fenestrae
opened behind so that the bases of the pectoral fins could be
lengthened. These trends were initiated in their coccosteid an-
cestors and are paralleled in some specialized families of Cocco-
steina. They differ from Coccosteina in having the posterior
margin of the skull roof embayed, in the wider nuchal gap be-
tween the cranial and thoracic shields (Fig. 5G, ng), in the
shorter nuchal plate with a pointed or rounded anterior margin
and a concave posterior margin, and generally in the absence of
prominent lobes on the central plates. They also show a tendency
to lose tuberculation on the dermal bones.

1975 PLACODERM FISHES 19

Manv Pachyosteina retain primitive, coccosteid-like characters
among which are small orbits, long, loosely attached cheeks, a
small nuchal gap, a relatively long median dorsal, rudimentary
spinal plates, and tuberculated dermal bones. Another primitive
character is an anteroventrally sloping neck-slit between the head
and thoracic shield. This sloping neck-slit is retained by the
Selenosteidae (Fig. 5G) which indicates that they were an early
side-branch of the suborder, even though they do not appear
until the Upper Frasnian. In many other respects the family was
highly specialized, especially in the weak jaws, and in the orbits
which had enlarged so much that the marginal plates formed
their posterior boundaries and the cheeks were greatly shortened.

The Bungartiidae (new family), known only from a single
Upper Famennian genus, Bungartius, is another family that
retains the sloping neck-sHt, but is pecuHarly specialized in other
ways. The preorbital part of the skull is greatly elongate, the
nuchal gap is much enlarged due to the posterior projection of
the paranuchal plates, and the jaws are shearing.

The Mylostomatidae are among the most specialized of Ar-
throdira with their durophagous jaws and their short, broad, flat
shield. Their origin is obscure; they show some resemblances to
Selenosteidae, but if Tafilalichthys is correctly referred here, it
is possible that they were independently derived from primitive
Pachyosteina.

Three families of Pachyosteina are distinguished by having the
cheeks and gill covers extended posteriorly, resulting in a nearly
vertical neck-slit. This may also give rise to a sharp angulation
in the anterior lateral plates where they bend around and under
the posterior edges of the gill covers. The first to appear, and in
fact the earliest Pachyosteina, are the Dinichthyidae, which are
mostly very large, broad-skulled forms with powerful, trenchant
jaws bearing strong anterior cusps on the anterior supragnathals
and infragnathals. The Leiosteidae are smaller forms with nar-
rower skulls that are deeply embayed behind, and with crushing
jaws. The third family, the Trematosteidae, has rather large
orbits, long preorbital and short central plates, a postpineal fen-
estra, strong shearing jaws, and a tendency to deepen the cheeks
and lower the jaw articulations. They are possibly related to
Leiosteidae, but could not have been derived from known genera.

The last family referred to the Pachyosteina is the Titanich-
thyidae, which were highly specialized giants known only from
the Famennian. Their shield is broad and depressed, and their

20 BREVIORA No. 432

jaws are long and slender, without teeth, cusps or shearing edges.
Their origin is obscure but possibly lies in the primitive Dinich-
thyidae.

The two remaining suborders of Arthrodira include forms that
have generally been referred to Brachythoraci or Pachyosteo-
morphi. The Heterostiina, including the single family Hetero-
stiidae (Fig. 5I-J), would at first sight appear to belong to
Pachyosteina. Like the Homostiidae and Titanichthyidae, it in-
cludes large forms with a broad, depressed head and body, but
is distinguished by a characteristic posterior widening of the
cranial roof. The latest forms have a very short thoracic shield
(Fig. 5J) in which the anterior laterals send a long, tusklike
process to meet the ventral shield, the latter a single plate lying
far anterior under the head. Since the Heterostiidae occur in
the Middle Devonian, it is not surprising to find that they retain
a number of primitive characters. Among these are a relatively
unspecialized cranial roof, a small nuchal gap, small anteriorly
placed orbits that face anterolaterally (Fig. 51, or), suborbital
plates with short suborbital processes and long blades, and tuber-
culated dermal bones. However, in spite of their early appear-
ance, they show no coccosteid characters and this, together with
their phlyctaeniid orbits and suborbital plates, suggests for them
a precoccosteid origin. If this is true, they are parallel to Pachyo-
steina, and thus referable to their own suborder.

The last suborder, the Brachydeirina (Fig. 5H), includes four
genera grouped in two families, the Leptosteidae and Brachy-
deiridae, though the three genera of the second family are so
distinctively specialized that each is commonly placed in a family
of its own. In contrast to all other Arthrodira, the head and
body are laterally compressed, high and elongate. In contrast
to Pachyosteina, the lateral walls of the thoracic shield are not
greatly reduced and large posterior laterals and posterior dorso-
laterals are retained (Fig. 5H, PI, Pdl). In spite of the long
thoracic shield, deep pectoral emarginations (Fig. 5H, pe) sep-
arate the lateral and ventral shields except anteriorly, indicating
probably that the pectoral fins were long-based. The nuchal
gap is never enlarged and in one genus, Synauchenia, the cranial
and thoracic shields have become sutured together, eliminating
the neck joint completely. The Leptosteidae (Fig. 5H) have
smaller orbits bounded posteriorly by postorbitals and suborbitals,
and a very long, slender thoracic shield. The Brachydeiridae
have larger orbits bounded posteriorly by marginals, and a

1975 PLACODERM FISHES 21

shorter thoracic shield in which the ventral part may be reduced.
The long thoracic shield of Brachydeirina indicates a derivation
from a very primitive Coccosteina or perhaps even from one of
the Phlyctaeniidae.

The last order, the Antiarcha (Fig. 5K), includes probably
the most highly specialized of Placodermi. The thoracic shield
is greatly elongated and has incorporated a second median dorsal
plate (Fig. 5K, Pmd) behind the anterior one. Instead of pec-
toral fins, they have peculiar, usually jointed appendages, cov-
ered with small dermal plates (Fig. 5K, pa). Though often
considered to be modified fins, these appendages were more
probably derived from arthrodiran spinal plates. Their skulls,
with their dorsal eyes and nostrils and large anterior premedian
plate, are so modified that it is difficult to compare them with
those of Arthrodira. Although antiarchs have been reported in
China from beds that are supposed to be Lower Devonian, their
first certain record is Eifelian. The first to appear are typical
members of the order and so there are no intermediate forms to
relate them to more typical placoderms. The elongate thoracic
shield suggests an origin from primitive Arthrodira, and since
their exoskeletal craniothoracic joint was certainly independendv
acquired, their ancestors probably are to be sought among
Actinolepidae.

22 BREVIORA No. 432

LITERATURE CITED

Gross, W. 1954. /iir Phylogenie des Schultergiirtels. Pal. Zeit., 28: 20-40.

. 1959. Arthrodiren aus dem Obersilur der Prager Mulde.

Palaeoiuogr., Abt. A, 113: 1-35.

1962. Neuuntersuchung der Stensioellida (Arthrodira, Unter-

devon) . Notizbl. Hessischen Landesamtes f. Bodenfoischung, 90: 48-86.
1963. Gemuendina stuertzi Traquair. Neuuntersuchung.

Notizbl. Hessischen Landesamtes £. Bodenforschung, 91: 36-73.
Miles, R. 1967. The cervical joint and some aspects of the origin of the

Placodermi. Colloq. Intemat. Cent. Nation. Recherch. Sci., 163: 4f^71.
. 1969. Features of placoderm diversification and the evolution

of the arthrodire feeding mechanism. Trans. Roy. Soc. Edinburgh, 68:

23-170.
. 1973. An actinolepid arthrodire from the Lower Devonian

Peel Sound formation, Prince of Wales Island. Palaeontogr., Abt. A,

143: 109-118.
0RVIG, T. 1957. Notes on some Paleozoic lower vertebrates from Spits-
bergen and North America. Norsk Geol. Tidsskr., 37: 285-353.
Parrington, F. R. 1949. A theory of the relations of lateral lines tO) dermal

bones. Proc. Zool. Soc. London, 119: 65-78.
Ritchie, A. 1969. Ancient fish of Australia. Australian Nat. Hist., 16(7):

218-223.
. 1974. From Greenland's icy mountains ... A detective

story in stone. Australian Nat. Hist.. 18(1): 28-35.
Stensio, E. 1969-1971. Anatomic des Arthrodires dans leur cadre systema-

tique. Ann. Paleont., Vertebres, 55: 151-192; 57: 45-83, 158-186.
Westoll, T. S. 1945. The paired fins of placoderms. Trans. Roy. Soc.

Edinburgh, 61: 381-398.

1975 PLACODERM FISHES 23

APPENDIX ~ CLASSIFICATION OF PLACODERMI

Class Pisces

Subclass Placoclcrmi

Order Stensioellida l-'ainily SLensioellidac

Order Rhenanida Family Asterosteidae

Order Pseudopetalichthyida Family Paraplesiobatidae

Order Ptyctodontida Family Ptyctodontidae

Order Acanthothoraci Family Palaeacanthaspidae

Family Radotinidae
Family Kolymaspidae

Order Petalichthyida Family Macropetalichthyidae

Order Arthrodira

Suborder Actinolepina nov Family Actinolepidae

Suborder Wuttagoonaspina Family Wuttagoonaspidae

Suborder Phlyctaeniina Family Phlyctaeniidae

Family Williamsaspidae

Suborder Holonematina Family Holonematidae

(including Groenlandaspididae)

Suborder Coccosteina Family Gemuendenaspidae

Family Buchanosteidae
Family Coccosteidae
Family Pholidosteidae
Family Homostiidae

(including Euleptaspidae)
Family Rachiosteidae

Suborder Pachyosteina Family Dinichthyidae

Family Titanichthyidae
Family Leiosteidae
Family Trematosteidae
Family Mylostomatidae
Family Selenosteidae
Family Bungartiidae nov.

Suborder Heterostiina Family Heterostiidae

Suborder Brachydeirina nov Family Brachydeiridae

Family Leptosteidae

Order Phyllolepida

Suborder Antarctaspina Family Antarctaspidae

Suborder Phyllolepina Family Phyllolepidae

Order Antiarcha Family Bothriolepididae

Family Asterolepidae
Family Sinolepidae

24 BREVIORA No. 432

Figure 6, Phylogenetic chart of Placodermi. Each branch represents a
family except in Acanthothoraci which includes three families. The width
of the branches is determined by the number of genera. — >

ANTI- PHYLLO-
ARCHA LEPIDA

ARTHRODIRA

Pachyosteina

[\A

APR 2 1 1977

MARVARO
UNI\AERSiT\

B R E V I 0 W'K

Museum of Comparative Zoology

us ISSN 0006-9698

Cambridge, Mass. 19 September 1975 Number 433

SOUTH AMERICAN A NOUS:
ANOLIS IB AGUE. NEW SPECIES OF THE
PENTAPRION GROUP FROM COLOMBIA

Ernest E. Williams^

Abstract. Anolis ihague, new species, is described on the basis of a single
juvenile female. It is regarded as a distinctive peripheral member of the
Anolis pentaprion group.

In a series of Anolis antonii received from the Vienna Museum
is a single small female anole with quite distinctive head and
dorsal scalation. It is clearly new and I name it after the locality
at which it was collected :

Anolis ibague, new species

Holotype: Vienna 18942:38; a juvenile female.
Type locality: Ibague, Dto Tolima, Colombia.

Head. Head scales smooth, imbricate, those in frontal depres-
sion larger than any on the snout. Scales across snout between
second canthals 8. 8 scales border rostral posteriorly. Anterior
and inferior nasal scales in contact with rostral. Six swollen but
narrow scales between supranasals.

Scales of supraorbital semicircles very broadly in contact, all
very large, the second and third pair relatively larger, the third
pair in contact with the enormous interparietal. Scales of supra-
ocular disk about 16 in number, smooth, in contact with supra-
orbital semicircles. Supraciliaries elongate, single, followed by
granular scales. Six canthal scales, canthus falling well short of
nostril, separated by swollen subgranular scales. Five loreal rows,

iMuseum of Comparative Zoology, Harvard University, Cambridge, Mas-
sachusetts 02138.

2 BREvioRA No. 433

uppermost and lowermost largest. Temporal and supratemporal
scales subgranular, not swollen. No differentiated supratemporal
line. Supratemporal scales gradually enlarging toward the inter-
parietal, with the scales immediately lateral and anterolateral to
the interparietal very large. One row of large scales posterior to
the interparietal immediately followed by scales similar to those
of the back.

Suboculars in contact with supralabials. 6-7 supralabials to
the center of the eve.

Mental wider than long, in contact with only two small scales
between the very large sublabials. Four sublabials on each side
in contact with the infralabials.

Throat and anterior chin scales between the sublabials laterally
large, becoming smaller centrally and posteriorly.

Trunk. Middorsal scales slightly larger than the lateral gran-
ules. Lateral granules becoming larger, merging into the much
larger smooth and imbricate ventrals.

Dewlap (juv. $). Absent. The merest indication in a very
small central fold, the scales not enlarged.

Limbs and digits. Scales of upper afm, front of thigh and
lower leg smooth. Those of lower affti unicarinate. Those of
digits weakly multicarinate. 19 lamellae under phalanges ii and
iii of 4th toe.

Tail. Compressed. No enlarg&d postanals. No tail crest, a
double line of weakly keeled scales middorsally. Most ventral
tail scales more distinctly keeled but scales immediately behind
vent smooth.

Color. A white middorsal zone diminishing to a point on the
occiput but continuing on tail. Head dark, vaguely marked with
lighter. Flanks light purpHsh, spotted and flecked with darker
purple. Belly and throat lighter, the throat spotted, the belly
more indistinctly tinged with darker.

COMPARISONS

The affinities of Anolis ibague would appear to lie with those
beta anoles with smooth ventrals, suboculars in contact with
supralabials and counts of fourth toe lamellae between 15 and 20.

On th^ one hand this would appear to ally ibague with the
fuscoauratus complex, and it is in fact sympatric, perhaps syn-
topic, with one member of this series-^ — antonii. Not surprisingly,
A. ibague more closely resembles a species not sympatric with it,

1975 ANOLIS IBAGUE 3

A. orton'u a species widely distributed throughout Amazonia. A.
ortoni approaches A. ibague in its large interparietal and its su-
praorbital semicircles in contact. It differs in having small scales,
like those of the dorsum, behind the interparietal. A. ortoni
resembles A. ibague in the presence of a middorsal light stripe
in the female. (This, however, is a character frequently present
in female anoles, even in very distantly related species. ) It differs
in a tendency to a higher number of loreal rows and in having
the scales immediately behind the interparietal small like the
dorsals. Neither the resemblances nor the differences are unique
or special.

There appear to be greater resemblances to the pentaprion
group which has now been described in some detail by C. W.
Myers (1971) with the description of two new species and the
restoration from synonymy of a third.

Myers has defined the pentaprion group in the following terms:
"Beta anoles of small to moderately large size, relatively short
legs (appressed hind limbs usually failing to reach ear, never
reaching eye) ; digital pads dilated, with distal phalanx raised
from the dilated pad; low loreal region (maximum of 2-5 hori-
zontal scale rows) ; black throat lining and parietal peritoneum;
a bluish gray or blue-covered sliver of tissue at the corner of the
mouth; few rows of scales on dewlap of relatively persistent (i.e.
fade resistant in preservative) red or purple coloration; tendency
for lichenose or fungous color pattern (in two of three species) ;
no vertebral stripe; tendency for smooth scales over most of the
head and body; relatively small dorsal and ventral trunk gran-
ules; ventral granules tending to obliquely conical (ontogenetic
change to flat and imbricate in one species)."

Some of these characters cannot be determined in the unique
preserved type, and others do not apply. However, Myers has
already been forced to acknowledge occasional exceptions to his
character list, and some characters such as the absence of a verte-
bral stripe in the female are the sort of characters that are pro-
visionally accepted as part of a group definition in a small sample
of species but are discarded without hesitation if the ensemble of
characters proves that a species belongs in a group. The Hght
vertebral streak has apparently been evolved many times within
the genus Anolis, and its appearance in yet another species, what-
ever its relationships, causes no surprise.

I would place especial reliance on some of Myers' characters
and add certain others. Thus, smooth scales on head and body

4 BREVIORA No. 433

are at one end of a spectrum that in the genus as a whole varies
from completely smooth to rugose and heavily keeled. In any
small set of closely related species, smooth scales are likely to be
consistent. Similarly likely to be good group characters are low
loreal counts (lower than 6) and short limbs.

Quite as useful — ordinarily — are contact between suboculars
and supralabials and low counts across the snout between pos-
terior canthals (<10). In some species there is considerable
variability in these regards; more often these two conditions are
reliable group characters.

In these features in which I would place considerable confi-
dence — they are more distinctive within the beta section of
Anolis than in alphas — A. ibague fits the pentaprion group.

DISCUSSION

The single individual described above seems to be a juvenile
female. As such it will not appear to be the best material on
which to base a new taxon. Barbour (1934) has commented:
"It is most unfortunate to describe Anolis from single female
specimens as also Boulenger did on all too many occasions."

Barbour's philosophy, widely shared, rests upon the general
proposition that male Anolis are often more distinctive in both
scale and color characters than females of their species. This is
undoubtedly true. Underwood and Williams (1959), speaking
of Jamaican anoles, said: "The males of the various forms are
far more clearly differentiated than the females. The possession
of a fan by the male contributes to this, but the color of pattern
of the males is always more distinctive. In some cases females
are almost impossible to distinguish . . . Descriptions of species
founded only on female material are of limited value."

Again the truth of this for Jamaican animals would be difficult
to deny, but they represent a small radiation that, despite sig-
nificant differences in ecology and size, is still remarkably close
knit. In similar mini-radiations of anoles it is often true that the
color patterns and the spectacular dewlaps of males may be, like
the voices of male frogs, the major way in which the species tell
themselves apart.

However, in this, as in so many cases, no rigid rules apply.
The \'ariability of each group and subgroup is peculiar to itself
alone and must be empirically determined. Males are in anoles
the sex of choice for species descriptions, but sex dimorphism in

1975 ANOLIS IBAGUE 5

anoles does not go so far that valid species cannot be recognized
on females alone. Sexual dimorphism in Anolis is most often
evident in color and size, much more rarely in the general char-
acters of scalation. Aspects of morphology most probably asso-
ciated with social interaction and display — dewlaps, the pro-
bosci of proboscis anoles, tail crests, etc. ■ — are apt to be sexually
dimorphic. Sometimes there may be differences in head scales
but these are minor, e.g., greater keeling of all head scales in
females than in males, as in females of the Anolis homolechis
series of Cuba. In no case are scale differences of the kind that
would permit belief that male and female are quite distinct spe-
cies; at most they are differences of the kind that could be ex-
pected to occur between males of very closely related, doubtfully
distinct species.

Color differences are often more radical, but here in anoles
sharp differences may occur as morphs within well-understood
species or even, not at all unusually, between phases in the same
individual.

In any case, the problem of Anolis ibague is not that it is
rather characterless or differs only in subtleties from any other
anole. On the contrary, its characters are extreme for its group
and relati\'ely extreme within anoles.

The characters of A. ibague that are extreme are the great
size of the interparietal, of certain of the supraorbital scales, and
of the sublabials.

The size of these scales in the juvenile type specimen may well
be more extreme than in adults of the species. Some head scales
are often relatively larger in very young specimens of any species.
But, although the enlargement of certain head scales is greater in
ibague than in any related species, and these scales are at one
end of the curve of head scale variation for the genus Anolis as a
whole, they are, however, nearer the taxonomic norm for such
iguanid genera as sceloporines or tropidurines, for which a huge
interparietal and large supraorbital scales are in fact partly diag-
nostic. There is nothing anomalous about these conditions: they
are merely highly derived character states.

The discussion of relationship above has suggested that ibague
is a local representative in central Colombia of a group — the
pentaprion group — otherwise unknown there. Special peculi-
arity in a peripheral isolate is not unusual ; it seems the preferable
explanation of the exceptional features of this species.

BREVIORA

No. 433

1 1
1 1
1 1

I (

I I
1 1

1«
ll

ll
II
ll

ll
n
V

•4—1
•4-)

!h
(U

4-)

lb
s

bo
«

1975

ANOLIS IBAGUE

Fig. 2. Anolis ibague Type. Dorsal view of head scales.

Fig. 3. Anolis ibague Type. Lateral view of head scales.

BREVIORA

No. 433

Fig. 4. Anolis ibague Type. Ventral view of chin scales.

1975

ANOLIS IBAGUE

-^10'

Fig. 5. Asterisk indicates type locality of A. ibagiie.

10 BREvioRA No. 433

Table 1, Scale characters of A. ibague compared.

ihague sulci frons fungosiis vociferans pentaprion

scales across snout

7-13

7-14

scales between semicircles

1-2

0-2

loreal rows

3-5

2-5

interparietal/ear

scales between interparietal

and semicircles 0 1 1-3 2 1-3

scales between suboculars

and supralabials 0

supralabials to center of eye 6

7-8

6-8

7-10

fourth toe lamellae 17

19-24

ACKNOWLEDGMENTS

The study of South American anoles of which this is a part
has been supported by National Science Foundation Grant GB
3 7 73 IX and previous grants. Dr. Joseph Eiselt of the Vienna
Museum generously loaned the material in which A. ihague
was discovered.

LITERATURE CITED

Barbour, T. 1934. The anoles II. The mainland species from Mexico

southward. Bull. Mus. Comp. Zool. 77: 119-155.
MvERS, C. W. 1971. Central American lizards related to Anolis pentaprion:

Two new species from the Cordillera de Talamanca. Amer. Mus. Novi-

tates No. 2471: 1-40.
Underwood, G. and E. E. Williams. 1959. The anoline lizards of Jamaica.

Bull. Inst. Jamaica Sci. Ser. No. 9: 1-48.

APR 2 1 1977

B R E V I 0 W-A

"Miiseiim of Comparative Zoology

us ISSN 0006-9698

Cambridge, Mass. 19 September 1975 Number 434

SOUTH AMERICAN ANOLIS:

AN O LIS PARI LIS, NEW SPECIES,

NEAR A. MIR US WILLIAMS

Ernest E. Williams^

Abstract. Anolis parilis is described as the west Ecuadorian representa-
tive of A. minis from the Rio San Juan, Colombia, A. parilis differs from
A. mirus in a number of ways, all individually minor, but sufficient in sum
to indicate species status.

The species Anolis mirus was described (Williams, 1963)
from a single specimen with the imprecise locality "Rio San
Juan Colombia." No further specimens have been collected in
the intervening years.

However, another single specimen, obviously related, has come
to hand from intermediate elevations in Ecuador. Despite its
closeness to A. mirus, even in characters quite special to that
species, it appears to differ enough to deserve description as a
new species which I name because of its similarity as:

Anolis parilis n. sp.

Type. UIMNH 82901, an apparently adult male.

Type locality. Rio Baba, 2.4 km S Sto Domingo de los
Colorados, Pichincha, Ecuador. George Key, collector. Novem-
ber, 1965.

.Diagnosis. Very close to A. mirus but differing in color, in
smooth rather than keeled ventrals and in other minor scale
characters. Perhaps also different in size.

Head. Head scales small, weakly keeled. About 17 scales
across snout at level of second canthals. Six scales bordering

^Museum of Comparative Zoology, Harvard University, Cambridge, Mas-
sachusetts 02138.

2 BREvioRA No. 434

rostral posteriorly. Anterior nasal separated from rostral by one
scale. Seven scales between supranasals.

At least 4 scales between supraorbital semicircles, the scales of
which are not much enlarged. Supraocular disk not differen-
tiated. A short supraciliary on each side followed by granules.
Canthus distinct, 9 canthal scales, the fourth largest. Seven
loreal rows below third canthal (2nd canthal behind level of
loreal rows on the rise of the orbit). Uppermost and lowermost
loreal rows largest.

Temporal and supratemporal scales granular. An indistinct
double line of enlarged granules at margin between supratem-
poral and temporal areas. Scales around interparietal larger.
Interparietal about equal to ear opening, separated from supra-
orbital semicircles by six scales.

Suboculars narrowly in contact with supralabials, posteriorly
grading into upper temporal granules, anteriorly separated by
one scale from canthal ridge. Nine supralabials to below center
of eye.

Mentals wider than deep, in contact with eight scales between
infralabials. No differentiated sublabials. Central throat scales
smallest, grading laterally into larger distinctly keeled scales.

Trunk. Two middorsal rows tending, especially on nape, to
be conical, enlarged, smooth, subimbricate. Ventrals larger than
dorsals, subquadrate, smooth.

Dewlap. Large, extending onto first third of belly. Edge
scales about equal to ventrals. Lateral scales much smaller than
ventrals, in rows, widely separated by naked skin. Above dewlap
on sides of neck complex folding between ear and shoulder.

Limbs and digits. Largest arm and leg scales about equal to
ventrals and weakly unicarinate except those of elbows and knee
larger and multicarinate. Supradigital scales multicarinate. Fif-
teen scales under phalanges ii and iii of fourth toe; distal pha-
lanx not raised.

Tail. Compressed, without crest. Dorsalmost scale row sin-
gle, keeled. Ventralmost scales larger, strongly keeled. Postanals
irregularly enlarged.

Color (as preserved). Red-brown with a narrow black mid-
dorsal line. Black mottling tending to transverse banding on
side of neck and lower flanks.

Size. 81 mm, snout-vent length.

Discussion. The resemblances and differences between A.
parilis and A. mirus are made clear in Table 1. The differences

1975 ANOLIS PARILIS 3

are just sufficient to imply species distinction given that there
are only two specimens before us. Size appears to differ but it
is precisely in the larger species of Anolis that there is a long
period of growth after sexual maturity. The color and pattern
of the two are radically different as preserv^ed, but neither are
known from life. It is improbable but not impossible in a genus
such as Anolis that a difference as great as seen here could exist
in the color repertoire of a single species. No single one of the
scale differences — smooth versus keeled ventrals, suboculars in
contact with supralabials rather than separated by one scale row,
the greater number of scales across the snout, the different
rostral-nasal relationship, etc. — are quite outside the possibility
of intraspecific variation. Taken together, however, they point
to a high probability of specific difference, i.e., genetic discon-
tinuitv.

Nothing is known of the ecology of either of these species.
The few suggestions that can be made are inferences from struc-
ture only. The narrow toe pads without a raised anterior margin
(the condition described as the diagnostic character of the in-
valid genus Norops) are characteristic of some anoles that are
not arboreal but are grass or ground dwellers; this is a derived
condition within anoles that has been evolved repeatedly. Most
Norops-Yik^ anoles are small (less than 60 mm snout-vent length) ,
but the South American group to which parilis and mirus seem
to belong — the eulaemus species group — verges on giant size
(arbitrarily defined for Anolis as 100 mm snout-vent length).
Within the eulaemus group two subgroups may be distinguished,
one of which has the toe pads narrow but with a "raised" distal
edge — the eulaemus group s. str. — and another with the toe
pads Norops-Yikt. The latter is the subgroup to which parilis
and mirus belong along with A. aequatorialis (the ecology of
which again is quite unknown ) . A combination of giant size
and toe pads that are poorly differentiated would suggest a
ground dweller. The artist who drew mirus in fact showed the
animal on a rocky substrate — on no evidence whatever (Fig. 2,
Williams, 1963). In fact, however, both parilis and mirus have
the first phalanx of each digit enlarged and strengthened (shown
well in mirus in Fig. 1, Williams, 1963), a fact that probably
does imply climbing propensities but with claws not pads. No
more can be said until observations on the live animals are re-
ported.

BREVIORA

No. 434

llr''^-'-

> f\ ' iS'tf -■•;■;

't-.. " --;.-■

'- ■) ^.. -■'■%,^y

|; ,;^:'"_-

-. . ; 'U-' -i^-", ■'-

■{; 3^.-

.;•'.' i-. afc.

"u ■'■-

'^..•-

':"r" i- "<i^: , . -^

.fx-: '■■- -ir;' J-

•t-- -■

'■ -'A::'^-?;.- r^' ^

»\

t\
M

1 1
' I

Oh
O

1975

ANOLIS PARILIS

Fig. 2. A. parilis Type. Dorsal view of head scales.

Fig. 3. A. parilis Type. Lateral view of head scales.

BREVIORA

No. 434

Fig. 4. A. mirus Type. Dorsal view of head scales.

Fig. 5. A. minis Type. Lateral view of head scales.

1975

ANOLIS PARILIS

60°

■10

h/0'

Fig. 6. Dark circle = type locality of Anolis mirus. Dark square = type
locality of Anolis parilis.

ACKNOWLEDGMENTS

Description of A. parilis was made possible by the studies of
South American anoles that continue under National Science
Foundation Grant GB 3773 IX and previous grants. My thanks
go also to the authorities at the University of Illinois who made
the unique type available to me.

LITERATURE CITED

Williams, E. E. 1963. Studies on South American anoles. Description of
Anolis mirus, new species, from Rio San Juan, Colombia, with comment
on digital dilation and dewlap as generic and specific characters in the
anoles. Bull. Mus. Comp. Zool. 129: 463-480.

BREVIORA

No. 434

scales across snout
rostral/nasal

scales between supra-
orbital semicircles

supraciliaries

TABLE 1

Comparison of A. parilis and mirus

parilis
17

one scale between
nasal and rostral

one (short) followed
by granules

temporal line

scales around
interparietal

rows between inter-
parietal and semicircles

rows between suboculars
and supralabials

supralabials to center
of eye

mental

scales in contact with
mental between infra-
labials

sul)labials

dewlap

a very indistinct
double line

gradually larger than
dorsals or temporals

wider than deep

7711 rus

two scales between nasal
and rostral

on one side the same; on the
other one (short) and
granules in the middle of the
supraciliary margin and
enlarged scales posteriorly

a triangle of distinctly
enlarged scales

abruptly larger than
dorsals or temporals

one inteirupted row

10
wider than deep

adhesive pad

lamellae under
phalanges ii and iii
of fourth toe

snout-vent length

not dilferentiatcd

large, scales in weakly
defined rows, edge
scales ca = ventrals,
complex folding be-
tween ear and shoidder

not set off from first
phalanx
(NoroJ)s condition)

81 mm

same

large, scales in ivell defi7ied
rows, edge scales ca =
ventrals, complex folding
between ear and shoulder

same

15
116 mm

L^^

'^H'imUO. ^ APR211977

B R E V I O R A

Mmseiiiii of CoiiiparatiYC Zoology

us ISSN 0006-9698

Cambridge, Mass. 8 April 1976 Number 435

TWO NEW SPECIES OF CHELUS
(TESTUDINES: PLEURODIRA)

FROM THE LATE TERTIARY OF
NORTHERN SOUTH AMERICA

Roger Conant Wood^

Abstract: Two new species of the pleurodiran turtle Chelus are de-
scribed from the late Tertiary of northern South America. These are the
first valid extinct species of the genus to be described. Both occur outside
the present range of the single living species, C. fimhriatus, and neither
appears to have been directly ancestral to it. Observations on variability
in a sample of C. fimhriatus shells are recorded to facilitate comparisons with
the fossils.

INTRODUCTION

Of the world's living turtles, the most bizarre in appearance
is unquestionably the mata-mata, Chelus fimhriatus. Its shell
is gnarled and serrated, while its broad and extraordinarily flat
head, festooned weirdly with fleshy tendrils, looks as if it were
the product of some science fiction writer's fevered imagination.
This species at present enjoys a widespread distribution through-
out the Amazon and Orinoco River basins of tropical South
America. Yet surprisingly little is known about the behavior,
ecology, or intra- and interpopulational variation of this peculiar
creature, and virtually nothing is known about its ancestry.

The purpose of this paper is to put on record two species of
the genus from Tertiary sediments of northern South America.
These are the first fossil remains of Chelus well enough pre-
serv^ed to permit determination of diagnostic characters, and
knowledge of them provides the first, albeit imperfect, glimpse
into the evolutionary history of the genus. One of these species,

iFaculty of Science and Mathematics, Stockton State College, Pomona, N. J.
08240.

2 BREVIORA No. 435

from the late Miocene of Colombia, was discovered during the
mid-1940's by the late Dr. R. A. Stirton and his associates from
the University of California Museum of Paleontology and the
Geological Survey of Colombia. Occasional reference to the
existence of this material has been made over the years (e.g.,
Royo y Gomez, 1945; Stirton, 1953; Medem, 1968), but until
now it has not actually been described. Remains of the other
new species, from beds of Huayquerian age in northern Vene-
zuela, were collected during the summer of 1972 by the author
and colleagues from Harvard's Museum of Comparative Zool-
ogy.

The following abbreviations are used:

AMNH - — • American Museum of Natural History

BMNH — British Museum (Natural History)

GMB — Museum of the Geological Survey of Colombia, Bogota

MCNC — Museo de Ciencias Naturales, Caracas

MCZ — Museum of Comparative Zoology, Harvard University

MZUSP — Museo do Zoologia, Universidade de Sao Paulo

PCHP — personal collection, Dr. P. C. H. Pritchard

UCMP — University of California Museum of Paleontology

USNM — United States National Museum of Natural History

CLASSIFICATION AND DESCRIPTION

Order Testudines
Suborder Pleurodira
Family Chelidae
Genus Chelus

Chelus colombianus sp. nov.
Plates 1-3, Figure 1

Type. UCMP 78762, a nearly complete shell.

Hypodigm. The type, and GMB 2045 A, an incomplete shell
lacking part of the right side of the carapace and the anterior
plastral lobe; GMB 2049, a partial shell, completely disarticu-
lated; GMB 2446, carapace fragments; GMB 2085, left epi-
plastron; GMB 2242, left hyoplastron; GMB 2042, GMB 2089,
and UCMP 38851, neurals; and UCMP 38838, a peripheral;
all from the vicinity of Villavieja.

GMB 1844, a left xiphiplastron ; UCMP 39014 and UCMP
39024, neurals; all from the vicinity of Carmen de Apicala.

GMB 1885 and GMB 1891, left xiphiplastra ; GMB 1934,
right hyoplastron; all from the vicinity of Coyaima.

1976 TWO NEW SPECIES OF CHELUS 3

GMB unnumbered, posterior left quadrant of a carapace,
locality unknown.

Horizon and localities. Villavieja Formation (late Miocene),
upper Magdalena River Valley, Colombia.

The specimens making up the hypodigm were collected in
the vicinity of three settlements, Coyaima, Carmen de Apicala,
and Villavieja, the majority cotning from the last (see above).
Stirton (1953) designated the fossil vertebrates from these three
different localities as, respectively, the Coyaima, Carmen de
Apicala, and La Venta faunas. The first of these he regarded
as being of late Oligocene (Colhuehuapian) age while to the
latter two he assigned a late Miocene date. Subsequently, Fields
(1957) suggested that the Coyaima fauna was of the same age
as the others. Bryan Patterson (personal communication) in-
forms me that he tends to agree with Fields, being unable to
see anything diagnostically Colhuehuapian in the published
account of the scanty and fragmentary Coyaima mammalian
faunule. As regards the hypodigm of the species here described,
it is not possible to differentiate the few Coyaima fragments
from the remainder.

The strata containing these fossils all belong to the Honda
Group. These rocks have recently been subdivided into two
for'mations, the lower termed the La Dorada and the upper the
Villavieja (Wellman, 1970). The vertebrate-bearing sediments
are apparently confined to the Villavieja Formation (Van
Houten and Travis, 1968:696).

Diagnosis. Differing from all other South American chelids in
having intergular scute withdrawn from anterior margin of
carapace, and in hexagonal to octagonal shape of intergular;
seven or eight pairs of scutes (in addition to an unpaired gular)
on plastron, rather than six pairs; shell between fifty and one
hundred per cent larger than that of C. fimbriatus; median ridge
of carapace not increasing in prominence toward posterior end.

Description. Most of the specimens that I identify as C. colom-
bianus are isolated shell elements. Owing to the distinctive shell
morphology of Chelus, however, there is no doubt about the
propriety of the generic identification. Only two of the speci-
mens (UCMP 78762 and GMB 2085) actually preserve evi-
dence of the diagnostic scute position and pattern, but nearly
all are from large individuals. Because the beds in which they
were found are all of essentially the same age, there is no reason
to suspect that more than one species is represented.

4 BREVIORA No. 435

Of the type, little is missing except at the anterior margin of
the anterior plastral lobe. There has been some dorsoventral
compaction of the shell, which has produced numerous cracks
in the bone, especially on the carapace. Bone sutures can be
clearly discerned on the plastron and, to a lesser extent, the
boundaries of the peripheral bones can be delimited. So badly
cracked is the central part of the carapace, however, that all
traces of sutures have been obliterated from this sector. None
of the other specimens, however, reveal any peculiarities in the
pattern of bone sutures for this part of the carapace. No grooves
demarcating the boundaries between adjacent scutes have been
preserved on the dorsal surface of the carapace of the type, but
most of the vertebral outlines can be detected in another speci-
men (GMB 2045 A). No striking differences in vertebral pro-
portions are evident. The proportions of the midline ridge
provide the only possibily diagnostic character of the carapace;
this ridge does not become increasingly prominent toward the
rear of the shell, as is typically the case in the single living spe-
cies. This feature by itself, however, would be insufficient to
persuade me to recognize a new species, particularly in view of
the fact that the limits of variation in the shell structure of the
living species are so poorly known. In fact, except for the for-
tunate circumstance that parts of the anterior plastral lobe have
been preserved in two specimens (the type and GMB 2085),
there would be no compelling reason to suspect that the Colom-
bian fossils represented anything other than overgrown examples
of C. fimbriatus.

The scute pattern of the anterior plastral lobe is unique among
chelonians in that one or two extra pairs of scutes were clearly
present (Fig. 1). As the standard and heretofore invariable
number of paired plastral scutes is six, these extra scutes have
no counterpart elsewhere within the order. ^ The existence of
these scutes does not seem to represent an abnormality as they
were clearly present on both of the only two remains of anterior
plastral lobes in the hypodigm. The derivation of these novel
scutes is problematical. They may have grown in to fill the void
left by the intergular scute as it withdrew from the forward edge
of the plastron. If so, they might be termed the pre- or ante-

II know, however, of one example of C. fimbriatus (PCHP 38) in which
the humeral scutes have nearly been fully subdivided into anterior and
posterior portions (Fig. 2) . Of all the chelonian specimens that have ever
come to my attention, this is the only one I have seen exhibiting such a
tendency. Perhaps it is atavistic.

1976 TWO NEW SPECIES OF CHELUS 5

gulars. But why supernumerary scutes should develop here in
the case of C. colombianus but not in the case of the various
living species of Chelodina, in which the intergular is similarly
withdrawn, is not readily explicable. The extra pair of scutes
might equally well have resulted from the anteroposterior sub-
division of the humeral, pectoral, or abdominal scutes, in which
case some other name would be more appropriate. Because of
my uncertainty as to the homologies of the scutes on the front
half of the plastron, I refrain from proposing a new name for
the extra pair characteristic of this species. Disagreement about
the nomenclature for the bones and scutes of chelonian shells is
already widespread; the publication of almost every new mono-
graph or book on turtles is usually an occasion for proposing
a new name for some bone or scute, reviving one long dis-
regarded, or reshuffling the standard terms to apply to elements
not previously so named. This unsatisfactory situation can
hardly be improved by introducing a new name arbitrarily
assigned to any one of four pairs of scutes on the anterior half
of the plastron. What is important is not the name of this pair
of scutes, but their existence.

Intergular scutes that do not enter into the anterior margin
of the plastron are found elsewhere among the chelonians only
in the related genus Chelodina, whose distribution is limited to
parts of Austraha and New Guinea (Goode, 1967:24, 36).
Except in occasional specimens of Chelodina siebenrocki (sensu
Goode, 1967:44), in which the forward tip of the intergular
may appear truncated by reaching the plastral margin,^ the
intergular scute is invariably hexagonal in shape. Of the two
specimens of Chelus colombianus in which the shape of the
intergular can be determined, one (the type) displays the typical
hexagonal configuration seen in Chelodina while the other
(GMB 2085) is octagonal. Apparently the shape of the in-
tergular in C colombianus was somewhat variable, but in any
case it differs from that of the living species C. fimbriatus, which
is also characterized by a variably shaped intergular, but one
that is usually either triangular or pentagonal (see, for example,
Schmidt, 1966, fig. 2). The octagonal intergular shape is, to the
best of my knowledge, unique. The scute furrows radiating out
toward the margin of the plastron on the specimen having the
octagonal intergular (GMB 2085) indicate that at least in some
instances C. colombianus had an eighth pair of plastral scutes,

iln a series of fifteen specimens examined by my colleague A. Rhodin,
one exhibited this atypical scute pattern (personal communication) .

6 BREvioRA No. 435

again a unique condition among turtles. Clearly the number of
pairs of supernumerary scutes (one or two) depended on the
shape of the intergular and likewise was an individually variable
feature.

Accurate measurement of overall shell dimensions is possible
only for the type. But a second, fairly complete and undistorted
specimen (GMB 2045 A) has been well restored and a reason-
ably reliable determination of its length and width is obtainable.
When compared to a sample of shells of C. fimbriatus (Table 1),
those of C. colombianus are obviously larger. The two measur-
able shells of the latter species were not exceptional representa-
tives of the taxon since most of the fragmentary remains included
in the hypodigm are of more or less the same size as comparable
parts of the whole shells. Typical adult specimens of C. colom-
bianus evidently were much larger than are those of its living
congener.

Chehis lewisi sp. nov.^
Plates 4-5, Figure 3

Type. MCNC 239, a complete shell.

Hypodigm. The tvpe, and MCZ 4337 and MCZ 4338, complete
shells; MCNC 240, a pleural; MCNC 241, posterior half of a
carapace and plastron; and MCNC 242, a crushed vertebra
(probably a cervical) associated with a right xiphiplastron.

Horizon and locality. Urumaco Formation (Huayquerian),
from several localities in the vicinity of the town of Urumaco,
northwestern Falcon, Venezuela.

Specimens were collected at three different localities. The
type was found just south of the oil pipe line running from
Punta Gorda to the Paraguana Peninsula, about .6 kilometer
SW of where this conduit crosses the highway leading westward
from Urumaco toward Maracaibo (National Route 3). A
single specimen (MCNC 240) was encountered 3.5 kilometers
NW of a hill known as El Picacho on the up side of the Chi-
guaje fault. The remaining material was all confined to a small
area of exposures .4 kilometer SSW of Cerro Bacunare between
the Valle de la Paz and Bacunare Faults.

II take great pleasure in naming this species for my good friend, Arnold
D. Lewis, not only because he discovered the type specimen, but also in
recognition of his many and varied contributions to the science of vertebrate
paleontology over the years.

1976 TWO NEW SPECIES OF CHELUS 7

DiagJiosis. Differing from other species of Chelus in marked
posterior widening of carapace and in square rather than rec-
tangular shape of the first neural bone. Shell 15 to 20 per cent
larger than that of adult C. fimbriatus.

Description. Like most of the vertebrates from the Urumaco
Formation, the specimens of C. lewisi are covered with a gyp-
siferous encrustation that has damasfed the bone surface. The
scute sulci have mostly been obliterated and it is possible to
determine the full bone suture pattern in the type alone, and
this only after weeks of painstaking preparation in the labora-
torv\ Dimensions of the three complete shells are given in
Table 1.

The distinctive shape of the carapace leaves no doubt about
the validity of this taxon. The shells of both of the other species
of Chelus are parallel-sided or nearly so,^ whereas in all three
of the complete shells of this species the width increases markedly
from front to rear. Although each of these shells has undergone
a varying degree of dorsoventral compaction, with the type
showing the least amount of crushing, there is no evidence of
significant lateral deformation and the present outline of the
carapace is, I think, an accurate reflection of its true propor-
tions in life.

The shape of the first neural bone also appears to be a dis-
tinctive feature of C. lewisi. Its length only slightly exceeds its
width, and it is subrounded in outline, whereas in C. fimbriatus
the length of this bone is generally much greater than its width,
giving it a rectangular appearance (see Table 1 for measure-
ments). The width/length ratio for C. lewisi (.92) is outside
the range of values (.52-.84) for my sample of C. fimbriatus
and well above the mean value (.69) for this species." The six
succeeding neural bones are indistinguishable from their counter-
parts in C. fimbriatus. The neurals are arranged in an uninter-
rupted sequence. Part of the seventh as well as all of the eighth
pairs of pleurals meet in the midline of the carapace between
the last neural and the suprapygal. This is the typical condition

iln some specimens of C. fimbriatus the sides of the carapace are actually
bowed inwards slightly in the bridge region between the axial and inguinal
notches.

20ne specimen in my C. fimbriatus sample (PCHP 39) has a W/L ratio
of 1.10. I have deliberately excluded this from consideration because its
first neural has been transversely subdivided, thus resulting in clearly anom-
alous proportions (Fig. 4) .

8 BREVIORA No. 435

in many specimens of C. fimbriatus. Some variation, however,
does occur in the Hving species ( Table 1 ) . A relatively small
proportion of the carapaces in my sample (four out of nineteen)
had eight rather than seven neurals. In all but one of these
four specimens, the eighth neural abuts directly against the
suprapygal, thus preventing any of the pleurals from meeting in
the midline. In one specimen with only seven neurals, the
neural series also extends continuously from the nuchal to the
suprapygal so that no pleurals meet in the midline.

Outlines of three vertebral scute sulci (the second through
fourth) can be detected on the carapace of the type specimen
(MCNC 239), but otherwise none have been preserved on this
or any of the other specimens referred to C. lewisi. The verte-
brals are all proportionately broader than in a somewhat smaller
specimen of C. fimbriatus (MCZ 4028; Table 2), but this may
in part or entirely be due to dorsoventral compaction of the
fossil, which is most pronounced in the middle of the carapace.

The smallest of the three shells of C. lewisi is nearly five
centimeters longer than the largest of the available shells of
C. fimbriatus, while the largest is slightly more than nine centi-
meters longer (Table 1). Hence it appears that typical in-
dividuals of C. lewisi were somewhat larger (15-20 percent)
than their living congeners.

Aside from the proportions of the entoplastron, there is
nothing exceptional about the shape or arrangement of the
plastral bones. Entoplastral dimensions can only be determined
for one specimen of C. lewisi, the type. For this individual, the
greatest width of the entoplastron is only slightly less than its
midline length, the width/length ratio being .93 (Table 1). Its
proportions are such that it barely falls within the upper limits
of the range recorded in Table 1 for similar measurements of
C. fimbriatus (.50-93). It may be that this bone tended to be
relatively broader in C. lewisi than in the living species. If so,
its proportions may prove to be a useful diagnostic character.
However, until the rans^e of variabilitv in the dimensions of the
entoplastron of C. lewisi is better known, judgment must be
reserv^ed regarding its diagnostic utility.

Another feature that may serve to differentiate C. lewisi from
C fimbriatus is the extent to which the three anteroposterior
ridges on the carapace are developed. For all three of the com-
plete shells of C. lewisi, the median ridge tends to be rather thin
and only moderately undulating and, to a lesser extent, the same
seems to be true of the lateral ridges. In the living species, the

1976 TWO NEW SPECIES OF CHELUS 9

thicker median ridge becomes increasingly prominent toward
the rear, whereas this does not appear to be the case for the
\^enezuelan fossils. These differences may to some degree be
artifacts resulting from the dorsoventral compaction that all of
the shells of C. lewisi have undergone. Although I suspect that
they are indeed real, they are not vital for establishing the valid-
ity of the new species and therefore have not been mentioned
in the diagnosis.

DISCUSSION

Up to now, the known fossil record for Chelus has been
almost nonexistent. Although regrettable, this fact is hardly
surprising, as the fossil record for South American chelids is in
general abysmal. This rarity is somewhat puzzling, as the re-
lated pelomedusid turtles, forms that apparently have generally
similar ecological requirements, are reasonably well represented
in the vertebrate-bearing fossil deposits of the continent.

Fossilized remains were first referred to Chelus more than
eighty years ago; these consisted of two shell fragments from
the Amazon Basin of Brazil (Barbosa Rodrigues, 1892:48-49
and plates 12-15). They were recovered from beds that are of
Pliocene or Pleistocene age along the course of the Rio Purus,
probably not far downstream from the Peruvian border. The
museum in which the specimens were apparently deposited no
longer exists (Patterson, 1936:50) and the present disposition of
these remains is unknown. Of the two fragments described by
Barbosa Rodrigues, the more notable is a portion of the left
xiphiplastron in which the distinctive elongation of the posterior
tip, so characteristic of Chelus, has been preserved. Attribution
of the two fragments to this genus was certainly justified. No
species-specific characters are evident, however, and Barbosa
Rodrigues showed commendable (and somewhat unusual) re-
straint for his times by simply designating them as Chelys (sic).
Unfortunately, these specimens tell us nothing about the evolu-
tion of the genus, as they cannot be differentiated from compa-
rable parts of the shell of the living species.

Subsequently, Wieland (1923:12-14) described a small por-
tion of a carapace as representing a supposedly new species,
"Chelys{?) patagonica." This specimen was of uncertain age
and vague provenance — "Patagonian Tertiary beds (Mio-
cene?)." Originally catalogued as part of the collections of the
Peabody Museum of Natural History at Yale University, the

10 BREVIORA No. 435

fragment is now e\idently misplaced or lost. Wieland should
never have formally proposed a name for it. He was actually
uncertain as to its proper generic allocation, suggesting that it
might well belong to ''Testudo [Geochelone in current terminol-
ogy] or its allies," which I suspect might actually be the case
since tortoise remains have been recovered from the Miocene of
Patagonia (Simpson, 1942). Nor were any specific characters
given for "Chelys{?) patagonica," which Wieland stated was
"... a purely arbitrary name of convenience." By modern
taxonomic standards it can only be regarded as a nomen nudum
(see Simpson, 1942:2), and the specimen is of no further in-
terest to the present study.

In 1956, while on a paleontological expedition to the upper
reaches of the Rio Jurua, Dr. L. I. Price of the Geological
Survey of Brazil discovered a very large plastron as well as a
quantity of unassociated fragments that are all clearly referable
to Chelus. These specimens are probably Plio-Pleistocene in
age. None have so far been formally described, so I am uncer-
tain whether they possess any distinctive characters other than
exceptional size. It is possible that these remains represent a
new species.

The two new taxa described in this paper represent the first
valid extinct species of Chelus yet described. (They are also
among the best preserved fossil chelids of any sort recorded from
South x\merica.) UnHke the Brazihan fossils mentioned above,
both occur outside the present range of the living species.

Chelus at present occupies an enormous expanse of territory
and yet remains a monotypic genus, much as the side-necked
turtle Pelomedusa in sub-Saharan Africa. Unfortunately, rela-
tively little is known about its ecology. Brief anecdotal com-
ments have occasionally been published, but none of these, to
my knowledge, are based on detailed or prolonged study of a
single population or series of populations. The species is ap-
parently not uniformly abundant throughout its range, nor does
it appear to be especially common. Instead, populations seem
to be scattered in ox-bow lakes and swamps along the banks of
rivers. (This information was gleaned during the course of a
paleontological expedition to Peru in the summer of 1974.
Probably the species is distributed in a hke manner throughout
its range, but I cannot verify this.) Geological and faunal e\7-
dence associated with the discoveries of C. colombianus indicate
that the ecology of this species was similar or even identical to
that of C. fimbriatus. During Miocene times the area covered

1976 TWO NEW SPECIES OF CHELUS 11

by the Villavieja Formation was a flood plain through which
broad ri\'ers and their tributaries meandered. Swamps, mud
flats, and ox-bow lakes dotted the flood plain, which was peri-
odically inundated. In general appearance, the area probably
would not have differed appreciably from the wet, tropical zone
of the present-day upper Amazon basin (Fields, 1957:279,
389-393). The habitat of C. lewisi is more difficult to recon-
struct. This species is part of a fauna that consists predomi-
nantly of a variety of aquatic reptiles whose remains were buried
in both continental and near-shore marine deposits (Wood and
Patterson, 1973:2). Most or possibly all components of the
fauna, however, were clearly nonmarine forms. Thus, it seems
likely that lewisi was a nonmarine form, but it is unfortunately
not possible at present to determine its habitat more precisely.
Both fossil species of Chelus possess characters that, I think,
preclude them from the direct ancestry of C. fimbriatus. The
intergular of fimbriatus borders on the lip of the anterior plas-
tral lobe, as is the case for most turtles. But the recessed
intergular scute of colombianus is an atypical chelonian feature,
seen elsewhere only in certain Australian chelids. Hence, it is
probably a derived rather than a primitive character for the
genus. Since it is unlikely that a species with a derived char-
acter would later revert to the primitive condition, I suspect that
these species are members of two distinct lineages. Both colom-
bianus and fimbriatus have a carapace that is essentially parallel-
sided; because this is characteristic of the oldest known and also
of the only surviving species of Chelus, it seems to be the typical
carapace shape for the genus. Thus I suspect that lewisi, with
its posteriorly flaring carapace, represents a lineage divergent
from that which gave rise to fimbriatus. Just as in the case of
colombianus, it seems improbable that, in the course of evolu-
tion, a parallel-sided ancestral form could give rise to a flare-
shelled species such as lewisi and then re-evolve the parallel-
sided shell shape to give rise to the living species. It is con-
ceivable that colombianus could have been ancestral to lewisi,
but there are at present no compelling reasons to believe this.
Whatever the relationships between these two species inay have
been, it now appears that there must have been a greater species
diversity within the genus in the past, with several distinct
lineages evolving in different directions at one time or another,
only one of which has survived. Although its fossil record is
still woefully fragmentary, it seems probable that Chelus has
not always been a monotypic genus.

12 BREVIORA No. 435

Part of the problem in dealing with fossil remains of Chelus
is that so little is known about morphological variation in the
living species. Certain variable features — the number of neu-
rals and whether or not pleurals intervene between the last
neural and the suprapygal — have already been noted. Other
character variants of potential taxonomic importance also exist,
notably the scute pattern on the anterior plastral lobe^ and
possibly also the proportions of the entoplastron. Schmidt (1966)
has recorded additional ones: color patterns of the shell and
extremities; shape of the intergular scute; morphology of the
head; and relative width of the anterior plastral lobe. Accord-
ing to Schmidt, it is possible to recognize several subspecies of
C. fimbriatus although he did not formally do so in his paper.
This was just as well, as his sample was miniscule (five speci-
mens) and the associated locality data were vague (e.g., "Bra-
zil?", "Colombia", "Peru"). Nevertheless, it may indeed be
possible to distinguish valid subspecies using some or all of the
characters cited above, and perhaps others too. To do so, how-
ever, would require better collections than exist at present in
museums, for several reasons. First, population samples from a
single locality do not seem to exist, so that there is no basis for
estimating the extent of intrapopulational variability. Second,
the total number of specimens available for study appears to be
rather small. And, third, variation in recent shells cannot be
correlated with different parts of the species' range owing to the
generally poor locality data associated with most museum speci-
mens. For instance, ten of the nineteen specimens listed in
Table 1 were obtained from zoos, identified simply as being from
"South America," or were accompanied by no locality data
whatsoever. Several others were labelled as being from the
vicinity of Leticia or Manaus. These cities (as well as Iquitos)
have long been the headquarters of professional animal collec-
tors, and specimens brought to them may actually have been
found far away. Only two had data good enough to permit
identification of the river system in which they were captured,
and even this is not entirely satisfactory as many tributaries of
the Amazon and Orinoco Rivers are themselves hundreds of

lEleven of the nineteen specimens of C. fimbriatus recorded in Table 1
have the intergular completely separating the gulars. In an additional
sample of eleven specimens, consisting of live individuals, juveniles, or shells
for which I only have information about the relative positions of the scutes
on the anterior plastral lobe, nine have the intergular fully intervening
between the gulars.

1976 TWO NEW SPECIES OF CHELUS 13

miles long. Considerable field work will therefore be necessary
before it will be possible to determine convincingly whether or
not valid subspecies of C fimbriatus can be distinguished. Such
field work would also have the added benefit of providing for
the first time adequate knowledge about the ecology of this
species.

opB>i i/m

-t-» gj ■•->

«j

O -v

qipiM. ninuiixBjv

4^ c3 «

'4-1

■

Dh ^ -O

S 2:^ ^

OJ <U C)
d vi ^

qjSua^ auTipijv

ISO

siB|no SuijBiBdas

.<u o

Xpjaiduio3 jB|nS.i35ui

• S "t^ a;

S) Oj ^

opB^ I/M

O g:S

^-1

if S c

S '-^

qiPIM

:^ — o

4J
CO

^'o'^

S-i

^1 ^

O G S-,
<o F cj

qjSu3T

■»— 1

t_ W (U

auTipira UT SuTjaani

;[onaij 30 sjred ipg -^0

he shells of
centimeters
have not b

qU 9^5 JO 11^ ^^ s^^^d
siEjna^ JO -o^j

mSuaq 9DBdcjB3 aunpiM

o t) o

4:; b ^

d u

tn cd

g c o

•

e c ^

S-i l!^ D
P ^ O

C
3

«^ ^' t::

dj ^ 'r^

^ <U 03

C g Oh

•o>i uarapads

CJ ^^ Cj

(73

.^ 5 '*H

^ id S

(—1

—

o ^ c

2^ F^

Some
deteri

-t-J
CJ

sapads

CM CM

0-. 0-.

rjH CO

<
m
-^

lO cm ^-4
CD CD lO

r^ CD t^

r.. r.. o.. ^ 0

0-. 0-.

\* "<•

LO r>-

CD r^

CO CO

O '—I CO

CD CD lO

f>^ {->-. 0-. r^ r^

LO O CD O^ ^
Tf LO '=^ CO CO

S?;S

^ lO

^ LO

^S ??

CD CD

0 •* ^

^ >i

;< N N
000

SSS

p p

•COOT^COCOCMCTiCOCMCOCXJ'-HO^r^COOCM

T^LO^r:icococr5t^c^cr)r^cocoooLoc^cr)r--

tDioOrjHLOLOLOi^^LnOLOcoLO'^crjo^

<>jOCOcot£5coO'-^'--^LnOOTiHC^CNjLqco

2> o o o
>. c c c

C/2

Tf)

^OCOCTiCMCOCTJCOCri'—iO^CM'^CO^CMr^O

-r^cDLOcor--LO<X)c^r^r--cr5coi>-i>-LO^'-H

xfixnxriViviKnvivivi-.

_ _ W C» CT! C/3 -.

O O <u (U QJ (U o
C C >- >^ >. >> G

r^t^cor--r^r^r^r^i>-cot^coi>-r--t^t^co

COCMCOCOCOCOCOCO

^ CO ^

^ CO CO
CO CO CO

• t^

CO CO

. CM

CM CM CO

''"' cr^

O CO CM

CO ^:

CO LO o

O) '-I

CO r^ T-i

^ CO

CvJ

s s s

ffi ffi

^ ^ ^

^ ^

N
U

(/) C/D C/D

>; >i

ID ^ p

pq PP

>;'

CO m

CT) o

CM CO

CO CO

< <

CO T-H CO

CT) ^ CD

LO CT) O

CO LO r--

ffi ffi ffi

^ ^ ^

< < <

Tt; CO o CO CM r-.
O CM ^ 1-3 r^ CO

"^ CO CO CO CO CO

o c

^-^ -^ CO r-- CO CT)

Pl, pL| CO CO CO CO

^ c^ Ph Ph Plh Ph

g g ffi III K HI

a ^ U U U U

^ r^ Ph pL^ p^ PLh

>--

CTl

■4^

-d

C
«

1— H

•S

'fcj

1— H

M-l

'c«

4-1

• »iH

4— >

C
0

<-t-i

«4-l

c«

faO

■1^

-^J

4-1

t— 4

'rt

M-l

r-"

<4-l

'y?

(—

•fH

-£3

-4—1

#-i

•

•4—1

1—1

rt
■1-^

^iH

4-1

■4— «

4—1

3
OJ
(U

-4^

(/3

t«

'^■f

«M

BREVIORA

No. 435

TABLE 2

Dimensions (in centimeters) of the three vertebral scutes
preserv^ed on the type carapace of Chelus lewisi (MCNC 239)
compared with those of comparable scutes of a specimen of
C. fimbriatus (MCZ 4028).

Specimen number

Vertebral number 2

Midline length 8.5

Greatest width 12.4

W/L ratio .69

MCNC 239

MCZ 4028

3 4

8.9 8.0

6.6

5.9 5.1

13.1 10.4

7.1

6.9 6.3

1 .68 .77

.93

.86 .81

Figure 1. Scute patterns on the anterior plastral lobes of two specimens
of Chelus colombianus, GMB 2085 (top) and UCMP 78762 (bottom) .
Compare with Figure 2 for an example of the typical scute pattern in
Chelus fimbriatus. The right epiplastron of GMB 2085 has been restored
as a mirror image of the left side. The specimens are not to the same scale
but have been drawn so that the entoplastron in each is of the same length.

1976

TWO NEW SPECIES OF CHELUS

Figure 2. The plastron of a typical specimen of Chelus fimbriatus (MCZ
4028; left) and one (PCHP 38; right) in which the humeral scutes have
nearly been fully subdivided into anterior and posterior portions. Both
plastra are drawn to the same midline length.

BREVIORA

No. 435

Figure 3. Sketch of the carapace of Chelus lewisi to show the pattern
of bone sutures as well as those scute sulci that can be detected. Some
compensation has been made for distortions resulting from the dorsoventral
compaction of the specimen (compare with Plate 4) .

1976

TWO NEW SPECIES OF CHELUS

MCZ 4028

PCHP 39

' Figure 4. Anterior ends of two recent Chelus fimbriatus carapaces (not
drawn to same scale) to show typical proportions of the first neural (top)
and its abnormal subdivision (bottom) .

BREVIORA

No. 435

Plate 1. Carapace of the type specimen of Chelus colombianus (UCMP
78762) . Its midline length is 54.8 cm.

1976

TWO NEW SPECIES OF CHELUS

, --^ . , \>v^ ; 5"^ - "

Plate 2. Plastron of the type specimen of Chelus colomhianus (UCMP
78762) .

BREVIORA

No. 435

Plate 3. Left entoplastron (in external view) of a specimen (GMR
2085) referred to Chelus colombianus. The scale is in centimeters.

1976

TWO NEW SPECIES OF CHELUS

Plate 4. Carapace of the type specimen of Chelus lewisi (MCNC 239) .
Its midline length is 45.5 cm.

BREVIORA

No. 435

Plate 5. Plastron of the type specimen of Chelus leivisi (MCNC 239)

1976 TWO NEW SPECIES OF CHELUS 25

ACKNOWLEDGMENTS

My visits to Bogota in 1970 and to Berkeley in 1971 for the
purpose of studying the Colombian fossils were made possible
by a grant from the National Geographic Society. Field work
in Venezuela during the summer of 1972 was made possible by
National Science Foundation grant no. GB-32489X to Prof.
Bryan Patterson, and by the cooperation of colleagues at the
Escuela de Geologia, Universidad Central de Venezuela (espe-
cially Profesora Lourdes de Gamero) and the Ministerio de
Minas e Hidrocarburos. Carmen Julia Medina, Kevin Maley,
and Robert Repenning assisted Arnold D. Lewis in the prepara-
tion of the Venezuelan specimens of Chelus. For their hospital-
ity and assistance, I am particularly grateful to Drs. Luis Felipe
Rincon Saenz and Andres Jimeno Vega, directors, respectively,
of the Museo Geologico and the Instituto Nacional de Investi-
gaciones Geologico Mineras, Bogota, Colombia. For access to
or information about specimens in their care, I am also indebted
to: A. G. C. Grandison; J. T. Gregory; J. H. Ostrom; L. L
Price; P. C. H. Pritchard; A. Rhodin; A. F. Stimson; P. E.
Vanzolini; E. E. Williams; R. Zweifel; and G. Zug. I thank
Prof. Patterson for critically reviewing this manuscript, A. Cole-
man for photographs of C. lewisi, and J. T. Gregory for photo-
graphs of C. colombianus.

26 BREVIORA No. 435

REFERENCES CITED

Barbosa Rodrigues, J. 1892. Les reptiles fossiles de la vallee de I'Amazone.
Vellosia, 2: 41-56.

Fields, R. \V. 1957. Hystricomorph rodents from the late Miocene of
Colombia, South America. Univ. Calif. Publ. Geol. Sci., 32: 273-404.

GooDE, J. 1967. Freshwater Tortoises of Australia and New Guinea (in
the Family Chelidae) . Melbourne, Lansdowne Press. 154 pp.

Medem, F. 1968. El desarrollo de la herpetologia en Colombia. Rev. Acad.
Colomb. Cien. Exact., Fis., Nat., 13: 149-199.

Patterson, B. 1936. Caiman latirostris from the Pleistocene of Argentina,
and a summary of the South American Cenozoic Crocodilia. Herpeto-
logia, 1: 43-54.

RoYO Y Gomez, J. 1945. Los vertebrados del Terciario continental Colom-
biano. Rev. Acad. Colomb. Cien. Exact., Fis., Nat., 4: 496-511.

ScHMmT, A. A. 1966. Morphologische Unterschiede bei Chelus fimbriatus
verschiedener Herkunft. Salamandra, 2: 74-78.

Simpson, G. G. 1942. A Miocene tortoise from Patagonia. Am. Mus. Novit.,
no. 1209: 1-6.

Stirton, R. a. 1953. Vertebrate paleontology and continental stratigraphy

in Colombia. Bull. Geol. Soc. Amer., 64: 603-622.
Van Houten, F. B. and R. B. Travis. 1968. Cenozoic deposits, upper

Magdalena Valley, Colombia. Bull. Amer. Ass. Pet. Geol., 52: 675-702.

Wellman, S. S. 1970. Stratigraphy and petrology of the nonmarine Honda
Group (Miocene) , upper Magdalena Valley, Colombia. Bull. Geol. Soc.
Amer., 81: 2353-2374.

VViELAND, G. R. 1923. A new Parana pleuiodiran. Amer. Jour. Sci., (5) ,
5: 1-14.

Wood, R. C. and B. Patterson. 1973. A fossil trionychid turtle from South
America. Breviora, no. 405: 1-10.

o /\(/xl u ur APR 2 11977

B R E V I O R A^

Miuseiim of Comparative Zoology

us ISSN 0006-9698

CambridCxE, Mass. 8 April 1976 Number 436

STUPENDEMYS GEOGRAPHICUS,
THE WORLD'S LARGEST TURTLE

Roger Conant Wood^

Abstract: Stupendemys geographicus, a gigantic fossil pelomedusid turtle
from the late Tertiary (Huayquerian) Urumaco Formation of northern
Venezuela is described. Stupendemys was evidently a highly aquatic form.
Whether it was a fresh water or marine turtle, however, cannot be determined
with certainty on the present evidence. One or perhaps even both pairs of
limbs may have been modified into flippers, and the head may not have
been fully retractable in the usual pleurodiran manner. Comparisons with
records of other enormous chelonians reveal that the carapace of Stupen-
demys is larger than that of any other turtle, fossil or recent.

INTRODUCTION

Paleontologists are occasionally fortunate enough to make
totally unexpected discoveries. Such was the case during the
summer of 1972, when a Harvard paleontological expedition
working in late Tertiary deposits of northern Venezuela un-
earthed the remains of several huge fossil turtles. These cer-
tainly attained greater size than any other extinct chelonians yet
known; they also appear to be larger than any living ones and
hence the largest turtles that ever existed. The purpose of this
paper is to describe these gargantuan creatures.

The following abbreviations are used :
AMNH: American Museum of Natural History, herpetological

collections
MCNC: Museo de Ciencias Naturales, Caracas
MCZ: Museum of Comparative Zoology: (H), herpetological

collections; (P), paleontological collections
PU: Geology Museum, Princeton University

iFaculty of Science and Mathematics, Stockton State College, Pomona, N. J.
08240.

2 BREVIORA No. 436

SYSTEMATICS

Order Testudines

Suborder Pleurodira
Family Pelomedusidae

Sttipendemys^ gen. nov.
Plate 1 and Figures 1-3, 5, 6, and 9

Type species. S. geographicus^ sp. nov.

Distribution. Huayquerian, Venezuela

Diagnosis. Shell gigantic; carapace depressed, with irregular
nodular contours on external surface and deep median notch at
front; anterior border of nuchal bone thickened and moderately
to strongly upturned; posterior peripheral bones moderately
scalloped along margins; neurals arranged in uninterrupted
sequence, numbers two through six hexagonal, the seventh
pentagonal. Mesoplastra hexagonal to subcircular, largely con-
fined to bridge; lateral ends of pectoral-abdominal scute sulci
terminating just anterior to axial notches of shell.

Cervical vertebrae (probably seventh and eighth) with saddle-
shaped articulations; neural arches relatively high in relation
to anteroposterior lengths of centra; angle of neural arch of
presumed eighth cervical with horizontal plane greater than in
any other pelomedusid; articular facets of postzygapophyses of
both cer\dcals forming acute angle of less than ninety degrees
with respect to each other; prezygapophyses of presumed eighth
cervical directed more perpendicularly than in other pelomedu-
sids; thin, bladelike spine on anterior face of eighth neural arch;
no ventral keel on eighth centrum.

Angle of divergence between two ventral processes of scapu-
locoracoid roughly ninety degrees; ventromedial process of scap-
ula dorsoventrally flattened; coracoid greatly thickened along
medial edge; glenoid socket facing forward rather than laterally.

Humerus squat, massive, lacking ectepicondylar groove of
foramen; deep bicipital fossa between radial and ulnar articular
facets on ventral surface; prominent ridge traversing ventral
surface of shaft from ulnar process to distal end, terminating

iThe generic name alludes to the astonishing size of this turtle, and the
species is named in honor of the National Geographic Society in recognition
of its generous support of my research on turtles.

1976 world's largest turtle 3

just above radial condyle; ulnar condyle broadest at anterior
end; ulnar and radial condyles facing somewhat more ventrally
than in other pelomedusids; entepicondyle and supinator process
strongly developed, resulting in distal expansion of humerus
almost as great as that of proximal end; shaft triangular in
cross-section rather than circular.

Femur squat, massive, greatly flattened dorsoventrally ;
breadth of tibial condyle approximately one-third total length
of bone.

Stupendemys geographictis sp. nov.

Type. MCNC 244, medial portion of the carapace with asso-
ciated left femur, frag'ments of a scapulocoracoid and a cervical
vertebra, probably the eighth.

Hypodigm. The type, and MCZ(P) 4376, much of the cara-
pace, fragments of the plastron, a cervical vertebra (probably
the seventh), both scapulocoracoids and a caudal vertebra;
MCNC 245, a plastron lacking the epiplastra and entoplastron,
two nearly complete pleurals, several peripherals, and one neu-
ral, all from the same individual; MCZ(P) 4377, a cervical,
probably the eighth; and MCZ(P) 4378, a left humerus.

Horizon and localities. "Capa de huesos" (also known as
"Capa de tortugas"), upper member of the Urumaco Forma-
tion, Huayquerian (which is probably of Pliocene age; see, for
example, Simpson, 1974: 5).

Outcrops of the Urumaco Formation are restricted to a rela-
tively small area in the northwestern part of the state of Falcon,
centering around the now-abandoned El Mamon oil field (lat.
11°13'N, long. 70°16'W), just north of the town of Urumaco.
The type was found immediately west of Ouebrado Tio Gre-
gorio, near its mouth. Other specimens were found as follows:
MCZ(P) 4376 — one-half km north of Ouebrado Picacho and
50' m east of the Chiguaje fault; MCZ(P) 4377 — three and
one-half km north 30° west of El Picacho on the up side of the
Chiguaje fault; MCZ(P) 4378 — as for the previous specimen,
but "about 15 m higher in the section; MCNC 245 — three-
quarters km north of Kilometer 153 on the oil pipeline running
from Punta Gorda to the Paraguana Peninsula (same locality
as MCNC 238, a trionychid; Wood and Patterson, 1973).

Diagnosis. As for the genus.

4 BREVIORA No. 436

DESCRIPTION

Shell. The most complete carapace is that of MCZ(P) 4376
( Plate 1 ) , which lacks some of the anterior peripherals on the
right side, as well as peripherals from the bridge region on both
sides. Scute sulci are deeply impressed onto the external surface
but, as in many giant chelonians, most of the bone sutures have
become largely fused and the pattern of these cannot be traced
with any degree of certainty. The carapace is low-arched in
the manner typical of aquatic turtles, and its dorsal surface,
rather than being smooth, is somewhat nodose. There is a
strong median indentation at the anterior margin of the cara-
pace that is unique among pelomedusids (and perhaps even
among turtles in general) in having the bone of this region
curled up into a thickened, collarlike structure. Posterior to the
bridges, the peripheral bones have mildly scalloped margins.
The sacral region of this specimen is fairly well preser\'ed. There
are four sacral ribs abutting against the attachments of the ilia
onto the visceral surfaces of the eighth pair of peripherals; the
distal ends of the last two of these are fused together. This is
essentially the same pattern as reported by Zangerl ( 1 948 :
30-31 and pi. 4, fig. 3) for the largest living South American
pelomedusid, Podocnemis expansa. There is a slight postero-
medial overlap of the iliac scars onto the suprapygal. Whether
these also extended forward onto the under surface of the
seventh pair of pleurals (and if so, to what extent) is uncertain
because the course of the suture between the seventh and eighth
pairs of pleurals cannot be determined. Measurements of this
carapace are given in Table 1.

The carapace of the type specimen ( Fig. 1 ) differs in several
respects from that of the one just described and moreover pro-
vides information about the shape and arrangement of the neu-
ral bones not revealed by the more complete specimen. Meas-
urements of the vertebral scutes of the two carapaces (Table 1)
indicate that the type was somewhat larger, roughly by five
per cent. Its midline length, therefore, would have been in the
neighborhood of ten to twelve centimeters longer, giving an
estimated midline length of as much as 230 centimeters. The
curling and thickening of bone at the anteromedian indentation
is less pronounced in the type than in MCZ(P) 4376. The out-
lines of six neural bones can be traced on this specimen. The
pattern revealed is typical for South American pelomedusids;
the last neural, which I believe to be the seventh, is pentagonal

1976 world's largest turtle 5

while those anterior to it (presumably the second through sixth)
are hexagonal. The neurals, again typically, tend to become
progressively broader in relation to their anteroposterior length
toward the rear of the series (Table 2). As far as can be de-
termined, the neurals were arranged in an uninterrupted se-
quence. Behind the last neural, part of the seventh and all of
the eighth pair of pleurals meet in the midline.

An isolated neural bone from another specimen (MCNC
245 ) adds further information about the structure of the median
part of the carapace. The bone is hexagonal and somewhat
longer than broad (Table 2), indicating that it comes from the
anterior part of the series. Because the first neural of pelomedu-
sids is usually elongate and rectangular or oval, it seems reason-
able to assume that the specimen in question is either the second
or third. The bone was obviously in direct contact with neurals
both to the front and rear. This reinforces the impression al-
ready given by the type carapace that the neural series was
continuous, and, in fact, if the neural is actually the second
rather than the third, proves the point. A notable feature of
this neural is its exceptional thickness in proportion to its length
and width; at various places around the periphery the bone
measures 2.8, 2.6^ and 2.4 centimeters dorsoventrally. In gen-
eral, pelomedusid neurals tend to be proportionately much
thinner. Although it is not feasible to measure the thickness of
the individual neurals of the type carapace, it is possible to state
that the carapacial bone does appear to be disproportionately
thick, even for a turtle of such exceptional size. Perhaps the
unusual thickness of the shell should be considered a diagnostic
character of the taxon.

There is nothing remarkable about the carapace scute pattern
of S. geographicus. It is virtually indistinguishable from that of
any of the living South American pelomedusids which, except
for minor variations, are all very similar.

No identifiable plastral remains are associated with the type
specimen. However, the mesoplastra, hyoplastra, and right
hypoplastron of MCZ(P) 4376 were recovered; these had been
crushed down into and molded against the shallow bowl-shaped
depression formed by the visceral surface of the carapace (the
shell was found lying upside down) and unfortunately preserve
little in the way of detail. Nevertheless, the presence of meso-
plastra in conjunction with pelves that were clearly fused to the
shell leaves no doubt that these gigantic turtles are pelomedusids.

6 BREvioRA No. 436

The mesoplastra are relatively small, hexagonal to subcircular
elements, laterally positioned and confined largely to the bridge.
This is the standard configuration for all known living and
fossil South American pelomedusids. On the basis of size and
thickness, I have referred a fairly complete plastron and some
miscellaneous carapacial fragments ( MCNC 245 ; Fig. 2 ) to
Stupendeynys. Although very large by ordinarv' pelomedusid
standards (Table 3), this plastron is relatively small in com-
parison to the carapaces described above. Presumably it repre-
sents a young adult. The forward portion of the anterior lobe is
missing. This is regrettable because it is this part of the pelo-
medusid shell that is generalh' the most useful for taxonomic
purposes. Nevertheless, some interesting characteristics are evi-
dent. The bridge is considerably longer at its base than the
posterior plastral lobe (Table 3). The bone is exceptionally
thick in proportion to its length and breadth. And, most notably,
the lateral ends of the pectoral-abdominal scute sulci terminate
just in front of the bases of the shell's axial notches, on the
edges of the anterior plastral lobe. This position is in contrast
to other South American fossil and recent pelomedusids in which
these sulci typically meet marginal scute sulci on the forward
third of the bridge, usually just in front of the anterior meso-
plastral bone sutures. The plastral formula, insofar as it can
be determined, is: femoral > abdominal > anal.

Axial skeleton. The three cervical vertebrae that have been
recovered (MCZ[P] 4376, MCZ[P] 4377, and MCNC 244)
belong to three different individuals and represent only two of
the eight bones in the series. Measurements of these are given
in Table 4. Because of the unique morphology of these verte-
brae, it is difficult to be certain as to their positions in the series.
In the cervicals of living pelomedusids, the neural arches be-
come increasingly prominent from front to rear, that of the
eighth always having the greatest height in relation to the length
of the centrum (Table 4). The two morphologicallv identical
fossil cervicals (MCZ[P1 4377 and MCNC 244) have neural
spines that are, relatively, even more prominent than that of
the eighth cervical in living pelomedusids, while the third
(MCZ[P] 4376) has an arch only slightly less prominent
(Table 4). On this basis it would seem likely that we are
dealing with cervicals at the posterior end of the series, pre-
sumably the seventh (MCZ[P] 4376) and eighth (MCZ[P]
4377 and MCNC 244).

1976 world's largest turtle 7

However, examination of the central articulations furnishes
contradictory evidence. Cervicals four, five, and six of all living
South American pelomedusids have saddle-shaped articulations,
the seventh is similarly shaped anteriorly but convex posteriorly,
and the eighth is concave in front and convex behind (Williams,
1950: 528, 532, 552, and fig. 11). The three known cervicals
of Stupendemys have saddle-shaped articulations, and hence
compare in this feature to the fourth through sixth cervicals of
the extant South American pelomedusids, rather than to the
seventh or eighth. (Undescribed fossil pelomedusid cervicals
from the late Cretaceous of Brazil, which I have been able to
examine through the courtesy of Dr. L. I. Price, are indistin-
guishable from those of living South American representatives
of the family.) In living African pelomedusids, the centra of
cervicals three through eight are uniformly procoelous (Williams,
ibid. ) . Cervicals are known for only one African fossil pelo-
medusid (Wood, 1971), and these differ from both living Afri-
can and South American forms in having articular surfaces
intermediate in shape between the saddle joints of the latter and
the procoelous condition of the former. No cervicals have been
reported for fossil pelomedusids from continents other than
Africa and South America, the only regions, together with
Madagascar, where the family still survives. The cervical artic-
ulations of Stupendemys are therefore most closely comparable
to those of its South American relatives.

Because the trend of anteroposteriorly increasing neural spine
height seems to be consistent in all pelomedusids, whereas the
pattern of cervical articulation varies somewhat, I am inclined
to place more reliance in the former feature as a means for
determining the relative position of the Stupendemys neck
vertebrae in the cervical series. As Table 4 shows, the height/
length ratio of the eighth cervical is always the greatest for any
individual. Moreover, as shell size increases, the height/length
ratio also increases, so that it is greater for the eighth cervical
of Podocnemis expansa than for that of the much smaller Pelo-
rhedusa subrufa. Given these observations, and in view of the
fact that the height/length ratios of MCZ(P) 4376 and MCNC
244 are considerablv s^reater than those recorded for anv of the
Recent species, while that of MCZ(P) 4376 is about the same
as the greatest ratio for the largest Recent specimen measured,
it seems that the cervicals of Stupendemys are from the pos-
terior part of the series, probably representing the seventh and
eighth.

8 BREvioRA No. 436

If the cervicals of Stupendemys are, in fact, the seventh and
eighth, then they are unique among known pelomedusids by
virtue of their saddle-shaped articulations. There are, in addi-
tion, several other features of these vertebrae that reinforce this
impression. One of the most obvious is that the neural arch of
the eighth cervical of Stupendemys makes a much less acute
angle with the anteroposterior axis of the centrum than do those
of the comparable cervical in other pelomedusids. (In the
cervical series of Recent pelomedusids that I have examined, the
neural arch of the eighth cervical always makes the greatest
angle to the horizontal plane.) In posterior view, the articular
facets of the postzygapophyses form an acute angle of less than
ninety degrees with each other. Those of other pelomedusids
are nearly horizontal to the dorsoventral axis of the vertebrae
(fig. 4; see also WilHams, 1950, fig. 11). Viewed laterally,
the shafts of the prezygapophyses of the presumed eighth cervi-
cals of Stupendemys are directed much more perpendicularly
than those of other pelomedusids. Although impossible to meas-
ure precisely, the angle made with the horizontal plane in the
specimens of Stupendemys seems to be roughly sixty to seventy
degrees, whereas in others it is closer to thirty degrees (cf.
figs. 3 and 4). The thin, median, bladelike spine on the anterior
face of the neural arch of the presumed eighth cervical of
Stupendemys is also unlike anything seen on comparable parts
of other pelomedusid cervicals. In most pelomedusids, the
ventral surfaces of the cervical centra are typically bowed up-
wards, sometimes quite strongly, along the anteroposterior axis.
The one exception known to me is the eighth cervicals of
South American representatives of Podocnemis. In these, a flat
blade of bone projects downward from the ventral surface
( Fig. 4 ) . But in both examples of the presumed eighth cervical
of Stupendemys, the ventral surface is neither bowed upwards
nor downwards; it is, instead, flat. Unfortunately, the bottom of
the presumed seventh cervical vertebra (MCZ[P] 4376) is too
badly damaged to determine its original shape.

A single, small caudal vertebra was found in association with
one of the shells (MCZ[P] 4376). It is poorly preserved and
reveals no features of special interest.

Appendicular skeleton. Much of both scapulocoracoids have
been preserved for MCZ(P) 4376, as well as fragments of one
belonging to the type. It is not possible to determine with cer-
tainty the relative lengths of the three prongs making up the

1976 world's largest turtle 9

shoulder girdle. The medial tips of the ventromedial portions
of the scapulae are broken ofT. The dorsal processes of this
same bone ha\'e been broken at their bases and flattened into
the same plane as the other two elements. Since their basal
contacts have been obliterated, it is impossible to determine how
much (if any) of these processes is lacking. The coracoids,
however, appear to be complete. Both the left and right ones
are of essentially the same lengths in MCZ(P) 4376 and are
considerably lons^er than what remains of the ventromedial
processes of the scapula, but slightly shorter than the more
complete of the two dorsal scapular processes that have been
preserved ( Table 5 ) . These proportions are in accord with
those of Recent pelomedusids, in which the ventromedial process
of the scapula is much shorter than the dorsal one, while the
coracoid is intermediate in length, generally somewhat flattened
dorsoventrally, and moderately to greatly expanded towards its
extremity. Despite this incompleteness a number of distinctive
features are evident. The glenoid socket faces almost directly
forward in Stupendemys, whereas in typical pelomedusids it
tends to face in a lateral direction ( Fig. 5 ) . The angle at which
the two ventral prongs of the scapulocoracoid diverge is con-
siderably less acute in Stupendemys than in any other known
pelomedusid (Fig. 5). The shoulder girdle of Stupendemys
further differs from those of typical Recent South American
pelomedusids in that the ventromedial process of the scapula is
dorsoventrally flattened. In specimens of Podocnemis dumerili-
ana, P. expansa, and P. unifilis that I have examined, this bone
is anteroposteriorly flattened. The medial side of the coracoid
of Stupendemys is greatly thickened. This is not true of the
coracoids in living xA.frican representatives of the family, which
are uniformly thin, flat, and greatly expanded. In typical South
American pelomedusids as well as in Podocnemis madagascarien-
sis, the coracoid is not so expanded but is transversely arched,
with the apex of the arch on the dorsal side. (The one excep-
tion of which I am aware is Podocnemis erythrocephala [Mit-
t'ermeier and Wilson, 1974]; the coracoid of this species does
not expand at all towards its tip but remains uniformly oval
along its entire length [e.g., MCZ(H) 10096].) The coracoid
of Stupendemys may have been similarly arched, if the dorso-
ventral crushing of this element is taken into account. The
thickness of bone along its medial edge, however, still seems to
set it apart from the other South American forms. The dorsal
scapular process in Stupendemys appears somewhat flattened,

10 BREvioRA No. 436

whereas in Recent pelomedusids it is more oval in cross-section.
This flatness, however, may result from crushing in the hori-
zontal plane; because of my uncertainty about this feature I
have refrained from listing it as a diagnostic character.

A nearly complete left humerus (MCZ[P] 4378) is all that
is known of the forelimb. This specimen is of great interest in
that it is totally unlike the humerus of any other known chelon-
ian — let alone pelomedusid — living or fossil. The head as
well as the terminal portions of the radial and ulnar processes
are missing, but otherwise the bone is complete ( Fig. 6 ) . This
humerus is extraordinarily massive, with distal and proximal
ends both markedly expanded, the latter slightly more so than
the former (see Table 5 for measurements). The curvature of
the shaft does not appear to differ appreciably from that of
living pelomedusids. There is no trace of an ectepicondylar
groove or foramen on the dorsal surface, a feature present in
all other pelomedusids (and, indeed, chelonians in general).
Between the radial and ulnar processes, on the ventral side, is a
very deep, semicircular depression, the bicipital fossa. This is
more prominent than in the fossil pelomedusid Bothremys
barberi ( Zangerl, 1 948 : 34 and fig. 1 3 ; Gaffney and Zangerl,
1968) or Podocnemis but is developed to about the same
extent as in Pelomedusa or Pelusios. Immediately above the
articular facets on the ventral surface at the distal end of the
shaft is a very deep, triangular fossa. This seems to be a natural
depression rather than the result of poor preservation of the
bone and has no equivalent, so far as I have been able to de-
termine, elsewhere within the order. A thick, prominent ridge
extends transversely across the ventral surface from the base
of the ulnar process to a point adjacent to the radial condyle.
Such ridges are absent in living pelomedusids, although less
pronounced ones have been reported in fossil pelomedusids,
Bothremys (Zangerl, 1948) and Taphrosphys (Gaffney, 1975;
Fig. 8, this paper). Typically, the ulnar condyle in pelomedu-
sids has a spool-shaped outline, equally expanded at both ends.
The ulnar condyle of Stupendemys, however, is markedly
broader at its anterior end than at its posterior limit. A further
distinctive feature of Stupendemys is that the trochlea extends
farther onto the ventral surface than in other pelomedusids.
To either side of the trochlea, the supinator process and
entepicondyle bulge outwards, the latter especially. Only in
Taphrosphys is the distal end of the humerus expanded to
such an extent (distal width over total length equals 0.47 in

1976 world's largest turtle 11

Taphrosphys [Gaffney, 1975, p. 16], 0.44 in Stupendemys).
In cross-section, midway between the ends, the shaft is triangular
rather than circular or oval, as is typically the case for pelo-
medusids.

A left femur (Fig. 9) was found associated with the type
shell. The head and terminal portions of both trochanters are
missing, as well as some bone from an area at the distal end
of the dorsal surface. The distal articular surfaces, however,
have been largely preserved. If complete, the femur would
have been of essentially the same length as the only known
humerus (Table 5). Like the humerus, the femur of Stu-
pendemys is massive. Its shaft is oval in cross-section and
greatly flattened dorsoventrally. The shaft of Podocnemis ex-
pansa is also oval in coss-section but is instead flattened antero-
posteriorly. As for the humerus of Stupendemys, the curvature
of its femur does not seem to differ significantly from that of
living pelomedusids. The distal end of the shaft is markedly
expanded, much more so than in Podocnemis expansa (distal
width over total length equals 0.47 in Stupendemys, 0.29 in
P. expansa [MCZ(H) 4469]).

DISCUSSION

Stupendemys has many very unusual anatomical features.
No modern chelonian is at all comparable to it, nor does it
closely resemble any of the better known fossil turtles.

Its systematic position, at least, is clear: it is an aberrant
member of the Pelomedusidae. This is conclusively demonstrated
by several characters : 1 ) the presence of mesoplastra ; 2 ) fusion
of the pelvis to carapace and plastron; and 3) shape of the
cervical articulations.

It is when one strives to understand Stupendemys as a living
animal that difficulties arise. In the following pages I attempt
a functional analysis of the known parts of the skeleton, search-
ing for clues to behavior and habitat.

The relatively low-arched carapace of Stupendemys indicates
that it was almost certainly a highly aquatic form, as are all
living pelomedusids and most fossil ones. Pelomedusids (not
yet formally described) from two different African fossil local-
ities, one of Oligocene and the other of Miocene age, are the
only terrestrial members of the family yet known (Wood, 1971).
These forms had extremely high-domed shells, superficially very
tortoiselike in appearance. Conversely, the only strictly terres-

12 BREVIORA No. 436

trial, flat-shelled turtle is the exotic pancake tortoise of East
Africa, Malacochersus, and its shell structure represents an
adaptation to most unusual habits. Shell shape thus seems to
be a nearly infallible indicator as to whether a chelonian was
aquatic or terrestrial, and Stupendemys clearly falls into the
former catesrorv.

The strong median indentation at the front end of the cara-
pace is not characteristic of pelomedusids in general, but is
reminiscent of the condition seen in the unrelated, big-headed
turtle, Platysternon, of southeast Asia. Platysternon has a xtry
large head in proportion to the size of its shell; consequently,
indi\iduals of this genus are not able to withdraw their heads
into the shell in the typical cryptodiran manner. But the an-
terior embrasure of the carapace provides a notch into which
the back of the head fits when retracted to the maximum extent
possible. The heavily boned dorsal roof of the skull then acts,
in effect, as an anterior continuation of the carapace and evi-
dently ser^^es as a reasonably efTective deterrent to predators.
Stupendemys, too, may have had a proportionately large, heavily
armored skull which did not have to be swung under the cara-
pace for protection in the usual pleurodiran fashion, but instead
was simply lodged against its anterior border when danger was
imminent.

I cannot readily account for the significance of the thickened,
curled-up bone at the anterior margin of the carapace. It might
represent a variably-expressed secondary sexual character if the
two carapaces in the available sample represent opposite genders.
It has, so far as I am aware, no structural equivalent elsewhere
within the order.

South American pelomedusids are the only chelonians having
saddle joints on the articular surfaces of their cervical centra
(Williams, 1950, appendix 1). But, as pointed out (p. 8),
the cervical vertebrae of Stupendemys, although possessing the
characteristic saddle joints, are in detail very different from
those of any pelomedusid known from that continent or else-
where. This fact supports the supposition that neck retraction
in the genus was fundamentally different from that of other
pleurodires. But if, as suggested above, Stupendemys was com-
parable to Platysternon in its ability to retract its skull only
partially, then the similarities in behavior were not paralleled
by structural resemblances of even the most superficial kind.
The articular surfaces of the fifth throus^h eis^hth cervical centra
in Platysternon are generally doubled, the centra themselves are

1976 world's largest turtle 13

very broad and flat, the neural arches lack spines, and so on.
In sum, while it is clear that the cervicals of Slupendemys are
markedly different from those of any other known turtle, the
significance of these differences is not readily apparent.

Re2:rettablv, the relative sizes of the humerus and femur in
Stupendernys cannot be determined with any degree of cer-
tainty. This is unfortunate because, for turdes in general, the
proportions of the fore and hind limbs are good indicators of
the customary mode of progression. Pelomedusids and most
aquatic cryptodires rely primarily on their hand limbs for pro-
pulsion while swimming, hence their femora are larger than
their humeri. But in tortoises and marine turtles, the opposite
is true. Thus, for example, if it were possible to establish that
the humerus of Stupendemys was larger than its femur, this
might be taken as reasonably good presumptive evidence that
this peculiar pelomedusid swam in a different way from all
other pelomedusids ^ perhaps even with flipperlike appendages,
as in the modern marine turtles. But direct comparisons between
the humerus and femur of a single specimen of Stupendemys are
impossible. Moreover, the only known humerus of Stupendemys
was an isolated find, which therefore cannot be tied to sheU size,
so that even indirect comparisons (in which limb size is related
to shell length) cannot readily be made.

Normally, limb structure is also a good index to the loco-
motory capabilities of turtles. The highly modified, flippered
forelimbs of marine cryptodires have a humerus that tends to
be broad, flat, and relatively straight-shafted. In aquatic (or
largely aquatic) forms, such as the pleurodires and emydines,
it is much more gracile, ordinarily more or less circular in cross-
section, and with a moderate curvature of the shaft. Tortoise
humeri are stout and often have a strongly bowed shaft. The
humerus of Stupendemys does not fall satisfactorily into any of
these broad categories. It is considerably more massive even
than that of a tortoise, fairly straight in the shaft, but more
circular than flat in cross-section. The heavy ridge across the
ventral surface of the shaft almost surely provided an increased
area for the attachment of hypertrophied antebrachial muscula-
ture. Such muscles would only be required if the distal ex-
tremity of the forelimb were for some reason disproportionately
large, as in marine turtles. While admittedly tenuous, this line
of reasoning leads me to suspect that the foreHmb of Stupen-
demys was modified into a paddle, a structure highly efficient
for swimming but ill adapted to a terrestrial existence of any

14 BREvioRA No. 436

sort. Gh'en the absence of direct fossil evidence, however, this
can only be a very tentati\'e suggestion.

The humerus of the fossil pelomedusid Taphrosphys (Fig. 8;
Gaffney, 1975, fig. 12) appears to be intermediate in structure
between that of Stupendemys and those of typical representa-
tives of the family. Unfortunately, the humerus is the only part
of the forelimb of Taphrosphys so far known, so that this taxon
provides no further insight into the structure and function of the
Stupendemys forelimb.

Forms intermediate in femoral structure between Stupen-
demys and the typical pelomedusids (or turtles in general,
for that matter) do not exist. Had the femur not been found in
association with pelomedusid shell remains, its familial allocation
would have been impossible. Differences between the femur of
Stupendemys and that of a representative pelomedusid [Podoc-
nemis expansa) hax^e already been enumerated (p. 11). The
strongly projecting trochanters, broad intertrochanteric fossa and
flattened shaft of Stupendemys distinguish it readily from both
marine cryptodires and tortoises, while the massiveness of the
bone and the broad, flat shaft together differentiate it from that
of the other aquatic forms. In these characters, in fact, together
with the relative straightness of the shaft, the femur of Stupen-
demys is more like the forelimb of marine turtles than anything
else. For this reason it is tempting to speculate that the hind
limbs of Stupendemys may have been modified into paddling
flippers as large as those possibly present on its forelimb.

In sum, the available anatomical evidence demonstrates that
Stupendemys was an aquatic form. In all likelihood, one or
perhaps even both pairs of limbs were modified as flippers. The
very size of its shell suggests that Stupendemys must have in-
habited large, permanent bodies of water which it probably left
only to lay eggs. Among living aquatic turtles in general, the
larger the species, the less likely it is to come out of the water
except for nesting. Size alone probably prevented Stupendemys
from basking along shores. Flippers, if it had them, would have
made such an undertaking even more awkward. I suspect that
Stupendemys was largely if not entirely herbivorous, again
simply because of its size ; all of the largest living turtles - — land
tortoises as well as the marine forms — are totally ( or nearly
totally ) herbivorous.

Geological evidence, although often helpful in attempting to
determine the habitat of a fossil, is, in the present case, equivo-
cal. A variety of different facies are represented in the upper

1976 world's largest turtle 15

member of the Urumaco Formation, including near-shore ma-
rine, brackish, and fresh water deposits. Some of these fresh
water facies consist largely of platy concretion zones, which are
probably best interpreted as representing small ephemeral ponds.
Root casts and locally abundant leaf impressions are also char-
acteristic of these deposits. Mammalian remains (especially very
large rodents) tend to be more abundant here, as are certain
of the reptiles (e.g., Chelus, nettosuchids ) . Other fresh water
deposits probably represent stream channels and, in some cases,
swampy areas (as evidenced by localized accumulations of veg-
etable debris). In general, the vertebrate-bearing sediments
were evidently laid down in a coastal area over which the posi-
tion of the shoreline fluctuated back and forth repeatedly.
Stupendemys could thus have been a marine form that washed
up on a barrier beach or was stranded in the lagoonal waters
behind one. Or it may have been a fresh water form carried to
the delta of a large river system and buried there. Since the
associated fossil fauna has strong Amazonian affinities and is
deficient in typical marine components, the latter possibility
seems strong. But all of the largest known aquatic turtles, both
living and fossil, are marine forms. This fact, coupled with the
fairly convincing presumptive evidence that a number of other
fossil pelomedusids were marine forms,^ prevents categorical
rejection of the idea that Stupendemys may have been a marine
turtle.

The largest of the living pelomedusids (all of which are fresh
water forms) is Podocnemis expansa, which has a wide distribu-
tion throughout much of the Amazon and Orinoco River basins
of South America. This species is sexually dimorphic, the fe-
males growing to much larger adult size than males (Ojasti,
1971 ). In a large sample taken from the Orinoco River over a
period of several years, the maximum carapace length for a male
was 51 centimeters whereas that for a female was 81 centimeters
(J. Ojasti, personal communication). The largest shell of this
species yet reported is 82 centimeters long (Williams, 1954:
293). Presumably this record is of a female, although the sex
of this particular specimen was not indicated. With the excep-

Jlncluded among these are several species of Taphrosphys (Schmidt, 1931;
Gaffney, 1975; Wood, 1975) , Bothremys (Zangerl, 1948; Gaffney and Zangerl,
1968) , and a generically indeterminate form from Puerto Rico (Wood,
1972) . All of these were found in near-shore marine sediments, generally
under circumstances such that tPicy cannot reasonably be regarded as exotic
elements washed in from a nonmarine environment.

16 BREVIORA No. 436

tion of Stupendemys, no known fossil pelomedusids exceed
Podocnemis expansa in size, nor do representatives of the only
other known family of side-necked (pleurodiran) turtles, the
CheHdae, ever approach P. expansa in size. Thus Stupendemys
is by far the largest pleurodire, living or fossil, yet known.

A few species of living fresh water cryptodiran turtles attain
greater carapace lengths than P. expansa, but none are reliably
known to approach the size of Stupendemys. A length of nearly
130 centimeters has been recorded for the carapace of the
Asiatic trionychid Pelochelys bibroni (Pope, 1935). Another
Asiatic soft-shelled turtle, Chitra indica, is generally believed to
have a maximum carapace length of approximately 90 centi-
meters. One unsubstantiated report indicates that Chitra may
occasionally reach a carapace length of roughly 180 centimeters
(Pritchard, 1967:211). No other living or ifossil fresh water
cryptodires as large as either of these recent trionychids are
known.

Some other fossil cryptodiran turtles of enormous size have
been described, but none of these had shells as large as those of
Stupendemys. Archelon ischyros, from the Cretaceous of North
America, is the largest of the fossil marine turtles; its straight-
line carapace length is 193 centimeters (Wieland, 1909), When
first described, Geochelone atlas (originally and rather appro-
priately named Colossochelys) was believed to reach twelve feet
in carapace length (Falconer and Cautley, 1844). This estimate
was based on composite reconstructions of fragmentary material
and has subsequently been modified to a maximum of six feet
(roughly 180 cm; see Lydekker, 1889, and Auffenberg, 1974:
173). None of the specimens that have since been referred to
G. atlas, which is now known from the Pleistocene of India,
Burma, Java, Celebes, and Timor (Hooijer, 1971; Auffenberg,
ibid.), appears to have reached or exceeded this length. One
or more species of Geochelone from the Pleistocene of Florida
and Texas may also have attained similarly gigantic dimensions
( W. Auffenberg, personal communication ) . However, no tor-
toises — li\'ing or fossil — ever seem to have grown any larger.

In fact, of all known turtles, only the anatomically peculiar
marine turtle Dermochelys coriacea may rival Stupendemys in
size. Dermochelys, commonly referred to as the leatherback, is
reputedly the largest of all turtles, living or fossil. x\dults con-
sistently attain carapace lengths of over 150 centimeters (Pritch-
ard, 1971 ). In the only large series of measurements ever made,
involving 1500 mature female specimens encountered laying eggs

1976

WORLD S LARGEST TURTLE

on the beaches of French Guiana over several field seasons, the
maximum length recorded was 180 centimeters (three individu-
als; P. C. H. Pritchard, personal communication). Larger speci-
mens have occasionally been reported, up to a supposed length of
3.35 meters, but these are unusual and suspect because they are
probablv based on estimates rather than actual measurements
(Carr, 1952:446), and, as Brongersma (1968:38-39) has
noted, estimates of the sizes of free-swinging marine creatures
generally tend to be greatly exaggerated. Thus, there do not
seem to be any reliable records of leatherbacks that equal or
exceed Stupendemys in carapace length. On the average, cer-
tainly, carapace lengths of Dermochelys are significantly shorter
than those of Stupendemys. Moreover, if the known specimens
are typical representatives of Stupendemys, then adult popula-
tions evidently tended to be significantly larger than those of
Dermochelys are today. In sum, it is clear that Stupendemys
is unquestionably larger than any other previously described
fossil turtle and it also appears to be larger than any living spe-
cies. Stupendemys, therefore, is the largest turtle that ever lived.

TABLE 1

Measurements (in cm) for carapaces of Stupendemys geographiciis. Di-
mensions are given as straight-line distances rather than over the curvatures
of the shells,

MCNC 244 MCZ(P) 4376

midline length (as preserved)
total midline length
maximum width (estimated)
maximum parasagittal length

first vertebral
second vertebral
third vertebral

fourth vertebral

fifth vertebral

184

218

approx. 230

218

190-195

185

250

235

(length
1 width

37.1
approx. 26

34.5
approx. 24

^ength
i width

33.5
36.4

34.0

32.7

(length

33.3

32.4

) width

39.3

34.4

Clength
J width

39.3
approx. 34

37.8
28.1

^ength
) width

52.4
51.7

BREVIORA

No. 436

TABLE 2

Neural bone measurements (in cm) for specimens of Stupendemys geogra-
phicus.

Midline Maximum Width/

Specimen No. Neural No. Length Width Length

MCNC 244

16.3

14.8

.91

16.6

19.2

1.16

15.5

18.0

1.16

*»

11.7

19.0

1.62

11.4

14.9

1.30

MCNC 245

2 or 3

7.7

6.5

.84

TABLE 3

Measurements (in cm) of the plastron (MCNC 245) referred to Stupendemys
geographicus.

midline length (as preserved)
total midline length (estimated)
width at axial notch
width at inguinal notch

anteroposterior length of bridge

midline length of posterior lobe
parasagittal length of posterior lobe
(to tips of xiphiplastra)

Cleft side
j right side

fleft side
) right side

57.2

34.0

35.3

35.2

36.2

21.0

25.2

25.5

1976

WORLD S LARGEST TURTLE

TABLE 4

Measurements (in cm) of the cervical vertebrae of Stupendemys compared
with those of adult representatives of each of the three living pelomedusid
genera, (MCZ [H]44G9, Podocnemis expansa; AMNH 10065, Pelusios sub-
niger; MCZ[H]1 46146, Pelomedusa subrufn) .

Height of Neural
Midline No. in Midline Arch Spine above

Carapace Cervical Length of Base of Posterior Height/
Length Series Centrum End of Centrum Length

Specimen No.

MCZ(P)4376

218

7(?)

9.0

13.41

1.49

MCZ(P)4377

8(?)

9.0

15.1

1.67

MCNC 244

230 .

8(?)

10.8

18.7

1.73

MCZ(H)4469

72.2

3.1

2.8

0.90

3.5

3.3

0.94

«>

3.6

4.1

1.14

»f

»»

2.7

3.9

1.44

AMNH 10065

24.2

1.3

1.1

0.85

1.3

1.2

0.92

1.6

1.5

0.94

1.5

1.00

MCZ(H)146146

12.8

1.0

0.6

0.60

1.0

0.7

0.70

>»

1.1

0.8

0.73

i»

1.0

0.9

0.90

iThe bottom of the posterior end of this centrum is somewhat damaged so
that a precise measurement is impossible; the figure recorded here is an
estimate.

20 BREVIORA No. 436

TABLE 5

Measurements (in era) of the known appendicular skeletal elements of
Sttipendemys geographicus.

SCAPULOCORACOID (MCZ[P]4376)

Cleft: 36.2

lengths (as preserved) of dorsal processes of scapulae -: . , oq ^

lengths (as preserved, along anterior edge, start-
ing from lateral side of glenoid fossa) of ventro- ^eft: 25.3
medial prongs of scapulae bright: 26.9

Cleft: 37.0

lengths of covacoids j right: 36.9

HUMERUS (MCZ[PJ4378)

length (as preserved) 31.0

estimated total length 34

maximum width of proximal expansion (as preserved) 18.0

maximum width of distal expansion 15.0
dorsoventral width at middle of shaft 8.3

anteroposterior width at middle of shaft 6,4

combined widths of ulnar and radial condyles on ventral surface 10.1

FEMUR (MCNC 244)

length (as preserved) 29.5

estimated total length 33-34

maximum width of distal expansion 15.7
dorsoventral width at middle of shaft 6.5

anteroposterior width at middle of shaft 8.0

1976

WORLD S LARGEST TURTLE

Plate 1. The carapace of Stupendemys geographinis (MCZ[P]4376) , in
dorsal view. Note especially the strongly curled bone at the base of the
antero-median indentation. Midline length of this specimen is 218 cm.
Peripheral bones in the region of the bridge on both sides, some of the more
anterior peripherals on the right, and the lateral ends of some of the
pleurals have been restored.

BREVIORA

No. 436

50
1

Figure 1. Carapace of the type of Stupendemys geographicus (MCNC
244) showing the shapes and positions of the second through seventh neural
bones.

1976

WORLD S LARGEST TURTLE

Figure 2. Sketch of a plastron (MCNC 245) referred to Stupendemys
geographicus, showing the unusual position of the pectoral-abdominal scute
sulcus. The full extent of the abdominal -femoral scute sulci cannot be
traced.

BREVIORA

No. 43&

Figure 3. The seventh (bottom; MCZ[P]4376) and eighth (top; a com-
posite based on MCNC 244 and MCZ[P]4377) cervical vertebrae of Stu-
pendemys geographicus in left lateral (left) , anterior (center) , and posterior
(right) views.

Figure 4. The fifth through eighth cervical vertebrae of Podocnemis ex-
pansa (MCZ[H]4469) in left lateral view. The arrow points toward the-
anterior end of the neck. Compare with the lateral views of Figure 3.

1976

WORLD S LARGEST TURTLE

I L_

for AMNH 13582 & MCZ(H) 4467

scapula

0 cm 10

for MCZ(P) 4376

Figure 5. The ventral elements of the left scapulocoracoid of Stupendemys
geographicus (MCZ[P]4376; bottom) juxtaposed with comparable bones of
the Recent pelomedusids Podocnemis unifilis (MCZ[H]4467; middle) and
Pelusios castaneus (AMNH 13582; top) . The midline axis of the specimens
to which they belong would be toward the left margin of the page. The
arrow points anteriorly. The glenoid socket of the fossil faces forward while
those of the Recent specimens are directed laterally. For clarity, the dorsal
prong of the scapula and the suture between the scapula and coracoid
have been omitted.

BREVIORA

No. 436

0 3 civT^iSiSs^P

Mil

Figure 6. The left humerus of Stupendemys geographicus (MCZ[P]4378)
m dorsal (left) and ventral (right) views.

1976

WORLD S LARGEST TURTLE

3 CM

Figure 7. The left humerus (top) and left femur (bottom) of Podocnemis
expansa (MCZ[H]4469) in dorsal (left) and ventral (right) views. Compare
with Figures 6 and 9.

BREVIORA

No. 436

Figure 8. The right humerus of Taphrosphys sulcatus (PU 18707) in
ventral view, showing the prominent ridge extending from the base of the
ulnar process to just above the radial condyle. Compare with Figure 6.

1976

WORLD S LARGEST TURTLE

0 3 CM

I I I I

Figure 9. The left femur of Stupendemys geographicus (MCNC 244) in
dorsal (left) and ventral (right) views.

30 BREvioRA No. 436

ACKNOWLEDGMENTS

My thanks go first to my colleagues in the field during the
summer of 1972, Messrs. Bryan Patterson, Arnold Lewis, Daniel
Fisher, Robert Repenning, and Michael Stanford, all of whom
helped to collect the various specimens of Stupendemys found
by the expedition. Without the splendid cooperation of our
Venezuelan colleagues from the Escuela de Geologia, Universi-
dad Central de Venezuela (especially Profesora Lourdes de
Gamero) and the Ministerio de Minas e Hidrocarburos, our
fossil collecting in Venezuela would have been impossible.
Funds for our field work were provided by National Science
Foundation grant no. GB-32489X to Professor Patterson. I
am, in addition, grateful to the National Geographic Society
for its support of my research on South American turtles.
Mr. Arnold Lewis supervised the monumental task of preparing
the specimens of Stupendemys for study and exhibition in his
usual capable manner, and I am particularly indebted to him.
The considerable talents of Messrs. Al Coleman and Laszlo
Meszoly are responsible, respectively, for plate 1 and figures
3, 4, and 6-9. For information, access to or loan of specimens
in their care, I am grateful to: W. AufTenberg; D. Baird;
D. Fisher; J. Ojasti; L. Price; P. Pritchard; E. Williams; and
R. Zweifel. Finally, my special thanks go to Professor Patterson
for critically reading several manuscript versions of this paper.

REFERENCES CITED

AuFFENBERG, W. 1974. Checklist of fossil land tortoises (Testudinidae) .
Bull. Fla. State Mus., Biol. Sci., 18: 121-251.

Brongersma, L. D. 1968. The Soay Beast. Beaufortia, 15: 33-46.

Carr, a. 1952. Handbook of Turtles: the Turtles of the United States,
Canada, and Baja California, Ithaca, Comstock Publishing Associates.
542 pp.

Falconer, H. and P. T. Cautley. 1844. On the osteological characters
and paleontological history of the Colossochelys atlas, a fossil tortoise
of enormous size from the Tertiary strata of the Siwalik Hills in the
north of India. Proc. Zool. Soc. London, 12: 54-55.

Gaffney, E. S. 1975. A revision of the side-necked turtle Taphrosphys
sulcatus (Leidy) from the Cretaceous of New Jersey. Am. Mus. Novit.,
no. 2571: 1-24.

Gaffney, E. S. and R. Zangerl. 1968. A revision of the chelonian genus
Boihremys (PIcurodira: Pelomedusidae) . Fieldiana: Geol., 16: 193-239.

1976 world's largest turtle 31

HooijER, D. A. 197L A giant land tortoise, Geochelone atlas (Falconer
and Caullcy) , from the Pleistocene of Timor. Proc. Koninkl. Nederl.
Akad. Wet., (B) , 74: 504-525.
Lydekkfr. R. 1889. On the land tortoises of the Siwaliks. Rec. Geol. Surv.

India, 22: 209-212.
MiTTERMEiER. R. A. AND R. A. WiLsoN. 1974. Rcdescriplion of Podocnemis
ery throe ephala (Spix. 1824) , an Amazonian pelomedusid turtle. Papeis
Avulsos Zool., 28: 147-162.
OjASTi, J. 1971. La tortuga arrau del Orinoco. Defensa de la Naturaleza,

no. 2: 3-9.
Pope, C. H. 1935. The Reptiles of China: Turtles, Crocodilians, Snakes,
Lizards. Natural History of Central Asia, 10 (Amer. Mus. Nat. Hist.) :
604 pp.
PRITCHARD, p. C. H. 1967. Living Turtles of the World. Neptune City,
N. J.. T. F. H. Publications, Inc. 288 pp.

1971. The leatherback or leathery turtle, Dermochelys coria-

cea. International Union for the Conservation of Nature Monograph,
1: 1-39.
Schmidt, K. P. 1931. A fossil turtle from Peru. Field Mus. Nat. Hist.,

Geol. Ser., 4: 251-254.
Simpson, G. G. 1974. Notes on Didelphidae (Mammalia, Marsupalia)
from the Huayquerian (Pliocene) of Argentina. Amer. Mus. Novit.,
no. 2559: 1-15.
WiELAND, G. R. 1909. Revision of the Protostegidae. Amer. Jour. Sci., 27:

101-130.
Williams, E. E. 1950. Variation and selection in the cervical central articu-
lations of living turtles. Bull. Amer. Mus. Nat. Hist., 94: 505-562.

. 1954. A key and description of the living species of the

genus Podocnemis {sensu Boulenger) (Testudines, Pelomedusidae) .
Bull. Mus. Comp. Zool., Ill: 279-295.
Wood, R. C. 1971. The fossil Pelomedusidae (Testudines, Pleurodira) of
Africa. Ph.D. thesis. Harvard University.

. 1972. A fossil pelomedusid turtle from Puerto Rico. Breviora,

no. 392: 1-13.

1975. Redescription of "Bantuehelys" eongolensis, a fossil

pelomedusid turtle from the Paleocene of Africa. Rev. Zool. Africaine,
' 89: 127-144.

Wood, R. C. and B. Patterson. 1973. A fossil trionychid turtle from
South America. Breviora, no. 405: 1-10.

Zangerl, R. 1948. The vertebrate fauna of the Selma Formation of Ala-
bama. Part 2. The pleurodiran turtles. Fieldiana: Geol. Mem., 3:
17-56.

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Breviora"

See other formats