I'l E> R.ARY

OF THL

U N IVLRSITY
OF ILLINOIS

590.5

cop.?*

FIELD MUSEUM OF NATURAL HISTORY.

PUBLICATION 207.
ZOOLOGICAL SERIES. VOL. XIV, No. i.

A MONOGRAPHIC STUDY OF THE
AMERICAN MARSUPIAL, C^NOLESTES.

WILFRED H. OSGOOD, £W,t:c f

Assistant Curator of Mammalogy and Qgfe&hology.

r>f Tit-

of Illi

A DESCRIPTION OF THE BRAIN OF

C^ENOLESTES.

C. JUDSON HERRICK,
Professor of Neurology, University of Chicago.

CHARLES B. CORY,
Curator, Department of ZoSlogy.

CHICAGO, U. S. A.
MAY, 1921.

THE LIBRARY OF THE

FE614?
UNIVERSITY Of

OF

FIELD MUSEUM OF NATURAL HISTORY.

PUBLICATION 207.
ZOOLOGICAL SERIES. VOL. XIV, No. i.

A MONOGRAPHIC STUDY OF THE
AMERICAN MARSUPIAL, OENOLESTES.

BY

WILFRED H. OSGOOD,

Assistant Curator of Mammalogy and Ornithology.

WITH

A DESCRIPTION OF THE BRAIN OF
GENOLESTES.

BY

C. JUDSON HERRICK,
Professor of Neurology, University of Chicago.

CHARLES B. CORY,
Curator, Department of ZoQlogy.

CHICAGO, U. S. A.
MAY, 1921.

^

THE LIBRARY OF THE

FE8141938

UNIVERSITY OF ILLINOIS

v.

A MONOGRAPHIC STUDY
OF THE AMERICAN MARSUPIAL, GENOLESTES.

BY WILFRED H. OSGOOD.

CONTENTS.

PAGE

Introduction 4

History 6

Distribution 15

Habits 17

External characters 19

Measurements 21

Myology 22

Muscles of the skin 22

Muscles of the back » 24

Muscles of the head 30

Muscles of the neck 32

Muscles of the thorax 34

Muscles of the abdomen 35

Muscles of the sacrum 36

Muscles of the tail 38

Muscles of the anterior limb 40

Muscles of the forefoot 46

Muscles of the hind limb 47

Muscles of the hind foot 56

Summary of myological characters 59

Urinogenital System 61

Male organs 62

Female organs 66

Alimentary Canal 69

Glands 73

Respiratory System 76

Circulatory System 77

Skeleton 77

Cervical vertebrae 77

Thoracic vertebrae 82

Lumbar vertebrae 84

Sacral and caudal vertebrae •. 84

Ribs and sternum 86

Pectoral girdle 87

Pelvic girdle 88

3

4 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Skeleton — Continued.

Arm and forearm 89

Bones of the hand 92

Thigh and leg 93

Bones of the foot 95

Summary of skeletal characters 97

Skull 99

Summary of cranial characters 109

Dentition in

Number and homologies of teeth 112

Upper incisors • . . . . 116

Lower incisors 118

Canines 118

Upper premolars 119

Upper molars 120

Lower premolars 126

Lower molars 127

Origin of Diprotodonty 128

Relationships of Wynyardia 136

Relationships of Myrmecoboides 139

Phylogeny and taxonomy 140

Dispersal of marsupials • . 145

General summary 150

Literature cited 152

The Brain of Canolestes obscurus 157

INTRODUCTION.

Owing to their many peculiarities of structure and their probable
relationship to ancient and long extinct types, marsupials are among the
most interesting of mammals. Their present distribution, largely in
Australasia but also in South America, adds to the importance they
possess for the morphologist and phylogenist. They have long been
classified in two large groups or suborders, the Polyprotodontia and the
Diprotodontia. These were first recognized by De Blainville in 1816 and
later by Owen in 1866. Broadly speaking, the polyprotodonts include
the carnivorous or insectivorous forms with small and relatively nu-
merous incisors, while the diprotodonts are the herbivorous forms
with the incisors reduced in number and modified much as in rodents.
Without exception, the diprotodonts are characterized also by a syndac-
tylous foot, while the polyprotodonts, with the exception of one family
(Peramelidae), are eleutherodactylous. This makes slightly different
divisions (Diadactyla and Syndactyla) possible", but other considera-
tions have favored the conclusion that the two major groups indicated
by the dentition are the most natural ones.

Until recently, it was supposed that all modern diprotodonts and the

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 5

majority of polyprotodonts were confined to Australasia, only one family
of polyprotodonts being known elsewhere, the Didelphiidae or opossum
family of South and Central America. Under these conditions, it was
a matter of the greatest interest when a small marsupial, widely different
from the opossums, was discovered in South America. This animal,
now known by the generic name Canolestes, presented an apparent
combination of diprotodont and polyprotodont characters, having a
diprotodont dentition and an eleutherodactylous foot. Moreover, it
was found to be closely allied to certain extinct forms known from
fragmentary remains from the Tertiary of Patagonia, being clearly a
surviving member of a highly differentiated group detailed knowledge
of which promised to throw new light on problems connected with the
origin and dispersal of marsupials. These facts were evident from
examination of the few specimens, consisting merely of skulls and skins,
which for many years were the only ones available. Such specimens
had been obtained through aboriginal sources and the exact habitat
of the animal was unknown.

In 1911, while making general collections in the mountains of western
Venezuela (Osgood, 1912), I had the good fortune to discover C&nolestes
living in dense forests at an altitude of about 8000 feet. As a result of
assiduous trapping, a total of eleven animals was captured and, despite
poor equipment for the care of anatomical material, two entire specimens
representing both sexes and two others with some parts mutilated were
preserved in formaldehyde and bichloride of mercury. In addition,
several dry skeletons and a small number of conventional skins and skulls
were saved. This material has formed the basis of the following study.
Its importance seemed to demand the fullest possible treatment and,
although this has not been accomplished, it is hoped that sufficiently
detailed information has been obtained to warrant general conclusions.
Study of particular organs and systems by specialists would have been
exceedingly desirable, but the nature of the material and the conditions
under which it has been studied have not made this feasible except to a
limited extent.1 The cooperation of Dr. C. Judson Herrick has been
greatly appreciated and his description of the brain of Canolestes to-
gether with the figures drawn under his direction which are incorporated
in the present publication will doubtless add greatly to its value.

Comparative material, especially of Australian forms, has been
confined almost exclusively to skeletons and skulls. Fortunately a fairly
representative collection of these is possessed by Field Museum and
this has been supplemented by specimens borrowed from the U. S.

1 It is especially to be regretted that no study of the organ of Jacobson has been
made.

6 FIELD MUSEUM OP NATURAL HISTORY — ZOOLOGY, VOL. XIV.

National Museum, the American Museum of Natural History, and the
Museum of Comparative Zoology. To the officials of these institutions,
especially Dr. J. A. Allen and Dr. G. M. Allen, acknowledgment is grate-
fully made.

The accompanying illustrations are from drawings made by several
different artists. The line drawings of muscles and soft parts (Pis. IV-X)
are the work of Mr. Robert E. Snodgrass, now of the U. S. Bureau of
Entomology; the wash drawings of skeletal parts (Pis. XI-XVIII)
were done by Mr. Kenji Toda of the Department of Zoology, University
of Chicago; the illustrations of the brain (Pis. XXI-XXII) are by Mr.
A. B. Streedain formerly of the Department of Anatomy, University of
Chicago; and the drawings of external parts (Pis. I-II) are by Mr. L. L.
Pray of Field Museum.

HISTORY.

In a brief list of mammals from Ecuador published by R. F. Tomes
in 1860, the following note appeared without reference either to genus

or species:

"A small animal about the size of the Water Shrew (Soreoc fodiens),
with external characters and incisor teeth so much like those of the
Soricidae as to have led in the first instance to the belief that it was a
placental Insectivore, perhaps in some degree resembling the Solenodon
of Cuba. However, the existence of a small and rudimentary pouch
sufficiently attests the implacental nature of the creature, which but for
this must certainly, as far as external appearances go, be regarded as one
of the Soricidae. A more ample account of it will be given on a future

occasion."

This is the first published reference to the animal now known as
Ccenolestes. Three years later (Tomes 1863), the single immature speci-
men was described as to its dentition and external characters and
given the name Hyracodon juliginosus without any reference to its affin-
ities within the order of marsupials. Subsequently it received little
notice for many years. In 1880, Alston referred to it in the following
brief statement, published as a footnote (Alston 1880, p. 195):

"A small Marsupial from Ecuador, named H yracodon fuliginosus by
Mr. Tomes (P. Z. S. 1863, p. 51, pi. VIII), may represent a distinct
family, but it is still only known from the very unsatisfactory original
description. In any case it will require a new title, the name Hyracodon
having been applied to a genus of fossil ungulates by Professor Leidy
in 1856, seven years before its use by Mr. Tomes."

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 7

Again in the concluding paragraph of the very important Catalogue
of Marsupialia and Monotremata in the British Museum (Thomas
1888) it receives only this slight mention:

"And, finally, one animal, referred to this Order, has been so in-
sufficiently described that no idea of its proper position can be gained,
nor have any further specimens of it been collected. This is —

"Hyracodon fuliginosus, Tomes, P. Z. S. 1863, p. 50, pi. viii (animal).

"Hab. Ecuador."

So meager were the data regarding it, that Lydekker (1894) com-
pletely ignored it in his very comprehensive Hand-book of Marsupials
and Monotremes. Thus for more than thirty years after its discovery
the subject of this monograph remained almost as unknown as if no
specimen existed. In 1895, however, a second specimen, consisting of a
skin and skull, was received at the British Museum and in two brief
but important papers (Thomas 1895) was promptly described, tenta-
tively classified, and figured. Its obvious affinity with extinct Patagonian
forms was at once recognized and its importance in connection with
theories of the dispersal of marsupials was noted. The preoccupied
name Hyracodon was discarded and a new name, Canolestes, was pro-
posed as well as another specific name, obscurus, the second specimen
having been obtained in the Bogota region and appearing to represent
a different species of larger size. The following extracts from these
papers of Thomas are of interest:

"In vindication of Mr. Tomes's paper I should like to say, firstly,
that his description, hitherto supposed (from our ignorance of any such
animal) to be imperfect or incorrect, proves to agree, so far as it goes,
very closely with the present specimen; and, secondly, that remarks on
the affinities of the animal must have been at that time more easily
wanted than given, since even now, with infinitely greater material and
the best of advice, I am unable to be at all positive about the exact
position and relationships of the little marsupial described by Mr.
Tomes."

"The rediscovery of this long-lost genus, whose wide distinction
from all other living marsupials its original describer does not appear
to have fully appreciated, is one of the most interesting events in
mammalogy that has happened for many years. A full description of the
animal will be given elsewhere; but it may be here briefly stated (i)
that Canolestes represents among the marsupials a family, and, perhaps,
a suborder, entirely different from any now living; and (2) that it is
closely related to, and evidently a surviving representative of, some of
the fossil marsupials from the Santa Cruz beds of Patagonia."

"Apart from this, the survival to the present day of a member of so

8 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

ancient a group, otherwise wholly extinct, is a fact of the utmost in-
terest, and one whose discovery will be welcomed by every zoologist."

"I have proposed for it ihenameCanolestes,as it is a modern member
of an ancient group."

fl'The affix 'lestes' is connected in mammalogy with small and an-
cient fossil marsupials, e. g., Microlestes, Amphilestes, etc., so that the
above name may be considered to represent an existing animal with
ancient fossil relatives."

"C&nolestes, with its uninteresting exterior, appeals mainly to the
technical mammalogist. To him, however, with its intense palseontolog-
ical and geographical interest, and the added puzzle its structure gives
rise to in the general classification of the order, no animal will appear
more important or more worthy of close and detailed study. That by the
arrival of spirit specimens account of its anatomy may be rendered
possible is very much to be hoped."

In these papers of Thomas, C&nolestes was classified as a diprotodont
and a member of the family Epanorthidae with the reservation that
^ It is, however possible that, in spite of the resemblance of the teeth of
C&nolestes to those of certain Australian Diprotodonts, the study of
further material, including soft parts, skeleton, and milk-teeth, will
bring out differences of such importance as to necessitate its subordinal
separation from them." He says further, "It is clearly a diprotodont,
as not only does it possess the characteristic development of the lower
incisors, but even the molars resemble most closely in structure those of
certain members of the family Phalangeridae, while being wholly unlike
those of the typical Polyprotodonts. From all the existing Diproto-
donts, however, apart from its habitat and numerous detailed differences,
C&nolestes is at once distinguished by not being syndactylous, a character
which is always considered of family rank. It forms, therefore, among
existing marsupials a peculiar family, and one which in America rep-
resents the Diprotodonts of Australia, just as Didelphyidae do the
Polyprotodonts. "

Although some further material was obtained in later years, it was
similar to that used by Thomas and consisted only of a few native-col-
lected skins and skulls. Additional study of these, therefore, added but
little to what was already known and opinions varied as to the inter-
pretation of the peculiar characters of the animal.

The next reference to C&nolestes was a note by Ameghino (1897, p.
95) from which the following may be quoted:

"M. Thomas est venu a La Plata, rapportant avec lui un crane de
Ccenolestes que nous avons soigneusement compare aux formes fossiles
de Patagonie et nous avons pu reconnaitre qu'il presente plus de rap-

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 9

ports avec les Garzonidae qu'avec les Epanorthidae. Pourtant il est
probable que le Coenolestes devra constituer le type d'une famille nou~
velle."

Accepting the suggestion of Ameghino and others, Trouessart
(1898) gave it the rank of a family, the Caenolestidae, including only
the one genus, and assigned it to the suborder Diprotodontia. Ameghino
(1900) later treated it as a family of diprotodonts, and still later (1903),
in his well known rearrangement of marsupials and rodents, he included
in the family the genus Ccenolestes and also the extinct genus Zygolestes.
The family was classified with four others, containing only extinct
forms, in his suborder Paucituberculata, which in turn was referred to
the order Plagiaulacoidea and regarded as directly ancestral to the
Australian diprotodonts. The name Diprotodonta was reserved to
designate a superorder conceived to embrace not only American and
Australian diprotodonts but also the multituberculates and the rodents.
Although Ameghino's ideas have not been generally accepted, especially
those regarding the derivation of rodents, his strong conviction of the
close alliance of American and Australian diprotodonts is noteworthy
in the present connection. While differing from him in other respects,
Weber (1904) agreed with Ameghino in placing Canolestes in neither the
Diprotodontia nor the Polyprodontia but in a third suborder of marsu-
pials for which Ameghino's name Paucituberculata was available. Mean-
while, Bensley (1903), in his very important paper on the evolution of
Australian marsupials, included C&nolestes in his "First Neogseic
Radiation" and in casual references indicated his belief that its dentition
is due to parallel or convergent development rather than to any direct
relationship to Australian diprotodonts closer than that of common
derivation from a didelphid ancestry.

On the other hand, Sinclair (1905, 1906) and Scott (1913) revert to
the opinion of Thomas and Ameghino that Ccenolestes and allied extinct
forms are so closely related to Australian diprotodonts as to furnish
strong evidence of a former land connection between South America
and Australia. Sinclair includes all South American forms with diproto-
dont dentition in the family Casnolestidae which he divides into two sub-
families, of which the Caenolestinae embraces the genera Ccznolestes,
Garzonia, and Halmariphus. Of this family he says (1905): "The
Caenolestidae resemble the primitive phalangers in so many respects
that it is impossible to escape the conclusion that the two families are
related and not merely convergent groups." This statement is some-
what modified in a later paper (1906), as follows: 'While substantially
the same conclusions are still held, it is proper to point out the evidence
in favor of the view that the striking similarity in dental structure

io FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

displayed by the two families may be explained by convergence. Un-
til the upper dentition, skull and feet of the Casnolestidae, and espe-
cially of the primitive members of the family, are fully known, this must
remain an unsettled question. At present the arguments in favor of the
alternatives expressed are about equally balanced."

The next author to give special attention to Ccsnolestes was Miss
Pauline Dederer (1909), who made a study of the skull and pointed out
certain resemblances to polyprotodonts, concluding that "While there
is undeniably a series of forms connecting C&nolestes with the Diproto-
donts in tooth structure, yet C&nolestes itself is so generalized in this
respect that we may perhaps, in the absence of corroborating characters,
question its inclusion within this group. Possibly it may be found to be
an offshoot from the Polyprotodonts, as it appears structurally to be
more generalized than any Diprotodont, and therefore it might well
occupy a separate suborder, as Thomas suggested — the Paucituberculata
of Ameghino."

Gregory (1910, p. 211) refers to Miss Dederer's work and concludes
that "the detailed characters of the skull show no striking Diprotodont
characters and the writer is inclined to regard Canokstes and its allies
as an independent suborder, an offshoot of primitive Polyprotodonts,
which has paralleled Diprotodonts in certain characters of the denti-
tion." Accordingly he places Ccenokstes in the suborder Paucituber-
culata with the parenthetical suggestion that the group might be called
the Casnolestoidea. Gregory also discusses the possible relationship of
the csenolestids to Propolymastodon and allied forms regarded as
multituberculates by Ameghino. He concludes his treatment of the
marsupial group by stating that "The problem of the genetic relations
of the Diprotodontia and the Polyprotodontia is complicated to a cer-
tain extent by the existence of the Csenolestoids; but the opinion may be
expressed that probably the resemblance of certain Casnolestoids to the
Multituberculates is an instance of convergence between related sub-
orders, and that the same is true, but to a less extent, of the resem-
blances of other Caenolestoids to the Diprotodont phalangers." Sub-
sequently (Osborn, 1910, pp. 515-518) the classification proposed by
Gregory was somewhat modified and the caenolestids were included in
the suborder Diprotodontia as a superfamily, the Caenolestoidea.

Broom (1912), in his special paper "On the Affinities of Casnolestes,"
reviewed the evidence adduced by Miss Dederer and pointed out cer-
tain resemblances to polyprotodonts not previously noted. He concluded
that " as Ccenolestes differs from the typical Polyprotodonts only in
tooth specialization, it should not be removed from the Polyprotodontia,
but merely be made the type of a distinct family, or section at most."

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. n

Gidley (1915), in a short footnote, expresses an opinion similar to
that of Broom, saying: "The Casnolestidae have been placed in this
great group [Diprotodontia] apparently on the diprotodont-like de-
velopment of the lower jaw. However, this may be an entirely in-
dependently acquired character. This family more probably belongs
with the Polyprotodonts."

With the exception of casual references in general works, nothing else
of importance has appeared in reference to the systematic or phylo-
genetic position of Canolestes. The occurrence of the animal in several
different localities has been recorded (Osgood, 1912; Stone, 1914) and
the number of species has been increased at least nominally to three, one
of which has been made the type of a second genus, Orolestes (Thomas,
1917), which, as described, seems only slightly different from the original
form.1

Reviewing the foregoing it is seen that, within two decades and upon
a knowledge of its cranial and external characters only, Coenokstes has
been placed in three different suborders and that various competent
authors have disagreed as to its proper position. Thomas, Ameghino
and Scott frankly classify it with Australian diprotodonts; Sinclair
inclines to do the same; Bensley, Miss Dederer and Gregory believe its
diprotodont characters are convergent and place it in a suborder of its
own; while Broom and Gidley would rank it merely as an aberrant
polyprotodont. Beginning with that of the period prior to the dis-
covery of Canolestes, the following are the various classifications of
marsupials concerned in its history, wholly extinct groups being in-
dicated by a dagger:

THOMAS 1888.

Order MARSUPIALIA

Suborder Diprotodontia

Family i. Macropodidae

2. Phalangeridae

3. Phascolomyidae

Suborder Polyprotodontia

Family r. Peramelidae

2. Dasyuridae

3 . Didelphyidae

1 At the time of writing, the privilege of examining specimens of Orolestes is denied
me.

12 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

THOMAS 1895.
Order MARSUPIALIA

I. Suborder Diprotodontia

A. Non-syndactylous. American
Family i. Epanorthidae C&nolestes

•\Epanorthus

•\Decastis

\Paraepanorthus

B. Syndactylous. Australian
Family 2. Phalangeridae

3. Phascolomyidae

4. Macropodidae

II. Suborder Polyprotodontia

A. Syndactylous. Australian
Family 5. Peramelidae

B. Non-syndactylous. American and Australian
Family 6. Didelphyidae

7. Dasyuridae

8. Notoryctidae

TROUESSART 1898.

Order MARSUPIALIA

I. Suborder Diprotodontia

Section i. Syndactylia

Family i. Phalangeridae

2. Phascolomyidae

3. tDiprotodontidae

4. Macropodidae

Section 2. Asyndactylia
Family 5. fAbderitidae

6. fEpanorthidae

7. fGarzonidae

8. Caenolestidae (Ccenolestes only)

II. Suborder Polyprotodontia.

Family 9. Peramelidae

10. fBorhyaenidae

11. Dasyuridae

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 13

12. Notoryctidae

13. fMicrobiotheridae

14. Didelphyidae

15. fAmphitheridae

1 6. fTriconodontidae

17. fDromatheridae

AMEGHINO 1903.
Subclass MARSUPIALIA

Superorder DIPROTODONTA
Order PLAGIAULACOIDEA

Suborder \Allotkeria

Family fPlagiaulacidae
fPolydolopidae
fNeoplagiaulacidae
fPromysopidae
fPolymastodontidae
Suborder Paucitttberculata
Family fAbderitidae
fEpanorthidae
Casnolestidae — Ccenolestes
t — Zygokstes
fGarzonidae
fMicrolestidae
Order HYPSIPRYMNOIDEA

Australian diprotodonts
Order RODENTIA

Rodents. Equivalent to Glires
Superorder POLYPROTODONTA

Australian and American polyprotodonts

WEBER 1904.

SubclaSS;MARSUPIALIA

Order MARSUPIALIA

Suborder Polyprotodontia

Family Didelphyidae
Dasyuridae
Notoryctidae
Peramelidae

14 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Suborder Paucituberculata

Family Epanorthidae (C&nolestes et al)
Suborder Diprotodontia

Family Phascolarctidae
Phalangeridae

SINCLAIR 1906.
Suborder Diprotodontia

Family Ccznolestidae

Subfamily Caenolestinae

Genus fHalmarhiphus
fGarzonia
Caenolestes

Subfamily fPalaeothentinae
Genus fPalaeothentes
fCallomenus
fDecastis

Subfamily fAbderitinae
Genus fAbderites

GREGORY 1910.
Infraclass METATHERIA

Order fTRicoNODONTA
Order JTRITUBERCULATA
Order MARSUPIALIA

Suborder \Alloiheria (Multituberculata)
Suborder Diprotodontia
Suborder Paucituberculata (Casnolestoidea)
Suborder Polyprotodontia

OSBORN 1910.
Order MARSUPIALIA

Suborder Polyprotodontia

Superfamily DIDELPHOIDEA
Family i. Didelphiidae

2. Myrmecobiidae

3. Dasyuridae

4. Thylacinidae
Superfamily PERAMELOIDEA

Family i. Peramelidae

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 15

Super-family NOTORYCTOIDEA
Family i. Notoryctidae

Suborder Diprotodontia

Superfamily C^ENOLESTOIDEA

Family i. Palaeothentidae — ^Pal&oihentes

— ^Abderites
— Caznolestes
2. Garzoniidae

Superfamily PHALANGEROIDEA
Family i. Phalangeridae

2. fThylacoleonidae

3. Macropodidae

4. Phascolomyidae

5. fDiprotodontidae
Suborder Allotheria

Family i. fPlagiaulacidae

SCOTT 1913 1
Order MARSUPIAIIA

Suborder Polyprotodonta
Didelphiidae
Thylacynidas
Suborder Diprotodonia
Ceenolestidas
fGarzoniidae
Suborder Allotheria

fPlagiaulacidse
fPolydolopidae

DISTRIBUTION.

The present known distribution of Canolestes (and Orolestes) is
shown by the accompanying map. Only a few localities are represented,
all Andean, and extending from the Venezuelan border on the north
nearly to the Bolivian boundary on the south. In the extensive regions
between these localities, no specimens have been obtained but it is not
'improbable that the animal has nearly or quite continuous range
throughout the central Andes so far as local conditions meet its needs.

1 American forms only.

1 6 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

so n /-^ 70 of 6<f ss so1

94' 90* 85* BO'

X Records of C&nolestes and Orolestes

The original type of Hyracodon fuliginosus was received from the
well-known collector Louis Fraser in a shipment from Ecuador of which
Tomes (1860, p. 211) says: "The greater portion of these are believed
to have been collected at Pallatanga, on the western slope of the Cor-

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 17

dillera; but the exact locality is not certain, from the specimens having
been unfortunately mixed together." The elevation of Pallatanga is
scarcely more than 1500 meters and, since all exact records of C&n-
olestes are from greater altitudes, it is probable that the first specimen
did not come from there but from the paramos of Mt. Chimborazo or
Mt. Pichincha where Fraser also worked. The type of Ccenolestes ob-
scurus likewise came from an indefinite locality — the vicinity of Bogota,
Colombia. In this case the probable actual locality is at the edge of the
paramos north or east of the city of Bogota in the eastern Cordillera of
the Andes. Exact records are only three in number: Paramo de Tama,
Venezuelan — Colombian Boundary at 7000-9000 ft. (Osgood, 1912);
Hacienda Garzon at 10500 ft., near the Paramo of Mount Pichincha,
Ecuador (Stone, 1914); and Torontoy, Peru, at 1 4000 ft. between Lats.
10° 30' and 13° 30' S. and Longs. 72° and 73° W. (Thomas, I9I7).1

HABITS.

Casnokstes is of terrestrial and crepuscular or nocturnal habits. It
feeds upon insects. These statements compass in a large measure what
is known of its habits. It lives at high altitudes in cool forests not far
from timberline and may be found also in grassy openings in mountain
valleys. On the Paramo de Tama, near the head of the Tachira River,
I obtained it at an altitude of about 7500 feet on both the Venezuelan
and Colombian sides of the international boundary line which is here
formed by the Tachira River. Most of the specimens were secured in the
heart of a dense forest on the Colombian side of the river not far from
a hacienda belonging to Don Mario Gonzales of San Jose de Cucuta,
Colombia. Wishing to collect and observe in such a place under wholly
natural and undisturbed conditions, I had employed men for some days
to cut a new trail directly into the deepest part of this forest for a
distance of several miles. Camp was established at the terminus of this
main trail and short radiating trails were cut thence to various points
of the compass, the forest and undergrowth being so dense that but little
freedom of movement would have been possible otherwise. In the
dense growth along these trails, Casnolesies were caught in small numbers,
rarely more than one specimen to fifty "trap nights." The exact nature
of the places in which individual specimens were taken varied widely,
indicating that the animals move about freely. Some were caught in steel
traps baited with small birds, others in rat traps baited with bacon rind,

1 In a paper published since this was written, Thomas (1920, p. 246) gives three
additional locality records: Machu Picchu, Peru, alt. 12000-13000 ft.; Ocabamba Val-
ley, Peru, alt. 9100 ft.; Gualea, Ecuador, alt. 6000 ft.

1 8 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

and still others in mouse traps with the conventional rolled oats as bait.
In the case of the mouse traps, however, it was probable that the an-
imals had not been attracted by the bait but had merely run across the
lightly set traps. A preference for animal food was quite evident and it
was only by continued effort with "meat baits " that a total of more than
one or two specimens was obtained.

Beneath the heavy canopy of treetops and still further shaded by
masses of low vegetation, the animals of the dark damp forests exercise
their vision only to a limited degree and depend mainly upon the sense
of smell. That this is the case with C&nolestes is indicated by its small
eyes and its highly developed olfactory organs. In fact, although some-
what ratlike in appearance and from its rather long hind legs doubtless
active and roving, C&nolestes is best described as to appearance by saying
that it is a large forest shrew. Like a shrew, it is fond of fresh meat when
such food offers itself, but its steady diet consists only of insects. The
contents of three stomachs were submitted to the U. S. Biological
Survey and as a result of examinations kindly made in their laboratories
the following report was received:

"No. I. Contents: Parts of a weevil, three ants (Dorylidae), skin of
caterpillar, lepidopterous pupa, adult lepidopteron, leg frag-
ments of orthopteron, a Tipulid larva, centipede and spider.
Dipterous and lepidopterous remains form the major portion.

"No. 2. Contents: Fragments of two beetles and a weevil, a cater-
pillar, an orthopteron, a hemipteron (?), spider and one seed of
composite plant. Caterpillar — at least 60 per cent.

"No. 3. Contents: Unidentified insect fragments and parts of
spider."

Other observations on the habits of Canolestes so far as published
are meager. Mr. Geo. D. Child, through whom the type of C. obscurus
was obtained, contributed the following note: "The little animal you
speak of is called 'Raton Runcho,' which means 'Opossum-Rat.'1
It lives in the high brush-wood, and is supposed to feed on birds' eggs
and small birds. It is very rare indeed, and is obtained with much
difficulty." (Thomas, 1895, p. 877.)

Mr. Samuel N. Rhoads, who collected two females on Mt. Pichincha.
Ecuador, writes that "The two specimens were secured in swampy
ground, the edge of a large pasture on the Hacienda Garzon, within £
few feet of a swiftly flowing stream of considerable size. They were
caught in small cyclone mouse traps set in underground runway;
among the thick grass, these runways being about on the level with th<
waterline of the swamp. They were caught on the same day, soon afte:

1 The book name selva was proposed by Lydekker in 1896 (Geog. Hist. Mamm.)
but has seldom been used by others.

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 19

placing the traps in that locality, but although I continued to trap there
for a week longer, having as many as 40 or 50 traps in that place, I
secured no more specimens there, nor in any other similar localities
where trapping was done. The stream alluded to runs over a bed
strewn with volcanic rocks and boulders and is in an open cultivated
valley-head, draining the south slopes of Mount Pichincha, about 8
miles south of Quito and at an elevation of about 10,500 feet, the valley
at this point being about half a mile wide and extending to even greater
widths as far as one can see, in a southerly direction." (Stone, 1914, p. 18.)
The type of C&nolestes (or Orolestes) inca was collected by an ex-
perienced naturalist, Mr. Edmund Heller, whose notes, when published,
will doubtless furnish substantial additions to what is now known of the
animal.1

EXTERNAL CHARACTERS.

General form. — Although frequently referred to as "ratlike," the
general form in Canolestes might with equal propriety be called soricine.
Insectivorous forms such as Nesogale approach it closely and, as else-
where shown (PI. XIX), have practically identical major proportions.
Among marsupials, the species most closely resembling it are perhaps
to be found in the genus Phascologale. Many species of Marmosa ap-
proximate it in size of body, but distinctions in the feet, ears, and tail
are so obvious that the parallel is at once seen to be only of the most
general nature. The legs are of moderate length, the hind pair 20 to
25 per cent longer than the front.

Head. — The head is elongate conical and somewhat shrewlike in
general appearance. The eyes are very small, the distance between the
anterior and posterior canthi measuring only 2.2 mm. in a female speci-
men in alcohol. The rounded ears project well above the pelage but are
relatively small as compared to those of the didelphids. The thickness
of the cartilaginous conch also is greater and it is set with fine short
hairs within and without. The tragus is but little developed and the
considerably larger antitragus is opposed mainly by the cms of the
helix which forms a thin flap directed inward. Two parallel antihelical
folds (metatragus) are prominent and separated from each other by a
deep fossa (PI. II, fig. i). A similar condition is found in Perameles
(cf. Thomas, 1888, PI. XXI, figs. 5, 7). The nostrils are lateral and
bounded by a rhinarium which is extended dorsally to form a small v-
shaped plate. A shallow furrow traverses it anteriorly and dorsally.

1 1 am informed by Mr. Heller that these and other notes on Peruvian mammals
have been prepared for publication by the National Geographic Society, Washington,
B.C.

20 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

The upper lips have a thin coriaceous edge that continues to the
median sulcus which partially divides the rhinarium and forms an
incipient cleft of the lips. The lower lips have a similar narrow edge
which is only very slightly thickened in front. The lower lip is free in
front from the base of the long incisors and has a high internal median
ridge or frenulum which fits into the slight space between the incisors.
Slightly in front of the angle of the mouth and just in advance of the
molariform teeth is a pair of reciprocating labrets or outgrowths of
the lips, one on the upper and the other on the lower lip (PL II, fig. 2).
These appear to be an adaptation for the ejection of undesirable food
particles during mastication. They are directed slightly outward and
downward in such a way that with the lips wholly or partly closed in'
front they might in conjunction with the buccinator muscle serve to
guide hard parts of insects or other matter to the exterior. A quite
different function also is possible, that is, in closing the lips at this point
and assisting in isolating the internal cheek pouch formed by the
buccinator muscle. This pouch is fairly definite and includes a consider-
able space laterad of the molariform teeth.

Tail. — The tail is gently tapering and has no inordinate thickening at
the base. The body fur extends upon it only a very slight distance.
Thence to the tip it is thickly set with short stiff hairs which nearly or
quite conceal the underlying scaly annulations. At the tip there is
usually a slight hairy pencil. Nothing suggests prehensilism in the
slightest degree.1 The contour of the tail is rather distinctly quadri-
lateral instead of cylindrical.

Feet.— The fore feet are pentadactyl, but the outer toes are decidedly
reduced and furnished with small blunt nails instead of claws. The
third or middle toe is slightly the longest and the second and fourth are
subequal, all three bearing sharp curved claws. The reduction of the
lateral toes and especially the absence of claws is in marked contrast to
the condition general among Dasyuridae and Didelphiidae but is almost
exactly paralleled in the Peramelidae in many of which the reduction
of the first and fifth toes has proceeded much farther. In Canolestes,
however, although the elongated middle toes seem to indicate a digiti-
grade adaptation, the plantar surface is carried well backward on the
outer side and supported by an unusually large pisiform bone. There are
five plantar pads well distinguished from each other and having smooth
non-striated surfaces.

The hind foot is relatively long and narrow. Its general proportions

'Thomas (1895), in describing C. obscurus, states that the tail has "its termina
inch below wholly naked" and that "it is therefore presumably prehensile." Nc
such condition has been observed in other specimens and it is probable the tail in th<
above-mentioned one was abnormal.

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 21

might be said to lie somewhere between those of Phascologale and
Antechinomys. The hallux is reduced and bears only a weak nail. It
is set at a slight angle but it scarcely reaches the end of the second
metatarsal and its functional importance is slight. The remaining four
toes are subequal, the third and fourth possibly a trifle longer than the
second and fifth. All are furnished with well-developed curved claws.
The third and fourth digits are connected at the base by an integumen-
tary web which is slightly more extensive than that between the other
digits, but this is scarcely to be regarded as a tendency toward syndac-
tylism. The soles are naked and provided with six plantar pads in
much the same relative positions as in Phascologale. Three of them are
digital, one hallucal, and two metatarsal. All are smooth without trans-
verse striations. See Plate II, figures 3 and 4.

Pelage. — The pelage is soft and thick over the entire body. Owing
to the peculiar distribution of hairs of different quality the coat is not
smooth or velvety but has an appearance of superficial looseness and
unevenness not uncommon in mammals of humid regions. The color is
practically uniform throughout, the underparts being only slightly
lighter than the upper owing to the absence of the fine exserted blackish
hairs which are generally distributed on the upperparts. The claws
and a few hairs at their bases are whitish but the hairiness of the upper
sides of the feet shows no contrast with the body. The tail also is colored
the same above and below and from base to tip except that in four out
of eleven specimens the tip is white. There are no facial markings and
in general it would be difficult to find a more uniformly colored animal.
The color is dull Mars brown, clearer on the underparts and more
mixed with darker and subject to light reflections on the upper parts.

Marsupium. — -No trace of an external pouch can be detected in
adult specimens of either sex and no immature examples have been
examined. Tomes (1863), in describing the original type of Hyracodon
fuliginosus, which was probably immature, states that it had a " small
and rudimentary pouch." Hence it is probable that the pouch in
Canolestes, as in Phascologale and some others, is present in the young
but disappears in the adult.

MEASUREMENTS.

External measurements. — Fresh specimens measured in the field:
Average of 5 males: Total length 240.6 (235-251); head and body 119.2
(113-135); tail vertebrae 121.4 (118-126); hind foot with claw 23.5
(23-24.5). Of 5 females: 223 (209-230); 107.6 (106-113); H5-4 (103-121)

22 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

22.5 (22-23). The ear in a female in alcohol measures 12.5 mm. from the
intertragal notch to the apex. The tail at the base is 3.3 mm. wide and
3 mm. deep ; at ro mm. from the tip it is i mm. x i mm.

Measurements of skull. — Adult male and female, respectively: Great-
est length 32.7, 29.2; basilar length 28.4, 25; zygomatic breadth 15, 13.8;
mastoid breadth 11.4, 10.9; depth of braincase 8.2, 8.1; length of nasals
15.1, 14; greatest breadth of nasals 4.3, 4.3; least interorbital breadth
7.3, 7.3; length of palate from gnathion 18.3, 16.2; anterior palatal fora-
mina 6.3 x 2.5, 5.4 x 2.2; palatal vacuities 6.7 x 3.2, 6 x 3.2; front of
upper canine to back of M4 12.9, 11.5; combined length of four upper
molars 5.7, 5.5; combined length of three lateral incisors 3.1, 2.8; length
of bone of lower jaw from condyle 20.5, 17.5; angle to coronoid 8, 7.2;
depth of jaw at second molar 2.2, 1.7; combined length of four lower
molars 6.6, 6.3; exposed length of median lower incisor 6, 5.

MYOLOGY.

Plates III-VI.
MUSCLES OF THE SKIN.

Panniculus. — Immediately beneath the skin and separable from it
only by careful dissecting is a thin, partly membranous and partly
muscular layer which may be divided into several rather definite parts
but which taken as a whole forms a sac enveloping the entire body and
having relatively few and weak ental attachments. On the fore and hind
legs it becomes exceedingly thin and cannot be traced beyond the middle
of the epipodials. Over the posterior part of the body it is entirely
membranous, thicker over the back and sides and very thin over the
pectoral muscles. It is attached posteriorly on the biceps femoris length-
wise of its caudal fibers; dorsally it becomes slightly muscular and dips
beneath the "cruro-coccygeus" to a definite insertion on the dorso-
ental surface of the tuberosity of the ischium.1 It is almost discontin-
uous over the pectoral and interscapular regions where it is lost in a thin
scarcely perceptible aponeurosis. Over the head it reappears as a thin
sheet chiefly membranous except along the under side of the ear and
forward on the side of the face to the anterior base of the zygoma where
a few muscle fibers run. The principal muscular division proceeding
directly from this outer plane, superficial fascia, or panniculus is that
which enters the axilla forming the so-called cutaneus maximus. Be-
possibly this dorsal part corresponds to the " ischiotergal " slip described by
Macalister in Dasypus and by Wilson (1894, p. 6) in Notoryctes.

MAY, 1921. AMERICAN MARSUPIAL, CJSNOLESTES — OSGOOD. 23

ginning in the middorsal line near the origin of the latissimus dorsi
(and immediately overlying and parallel to it), a thin sheet of converging
fibers runs forward and downward to the axilla. Similar converging
fibers from the mid-thoracic region and from the abdomen join those
from above and unite to form a single muscle which enters the axillary
region dorso-laterally just below the latissimus dorsi and ventrally
parallels the outer edge of the pectoralis major. It makes a turn on itself
and inserts as usual on the proximal two thirds of the inner edge of the
deltoid ridge of the humerus. The abdominal part of this muscle is
somewhat differentiated into a humero-dbdominalis or xiphi-humeralis
(pectoralis quartus). This begins to be fleshy on the middle of the abdo-
men and anteriorly divides to pass on either side of the pectoralis and
parallel to it. On entering the axilla it becomes fused with the cutaneus
maximus and has a common insertion with it. It has similar relations in
most marsupials.

Cervico-auricularis. — This is a thin paired sheet arising from the oc-
ciput and the nuchal crest for a distance of about 10 mm. or to the
fatty deposit in the interscapular space. It is inserted on the cartilage of
the back of the ear and into the ental surface of the superficial fascia be-
low the ear for some 8 mm. lengthwise of the side of the neck. Dorsally
it leaves the base of the ear for a short distance and attaches to the
ectal surface of the muscle which elsewhere underlies it and which is here
called the auriculo-occipitalis.

Auriculo-occipitalis. — Arises from the nuchal crest entad of the
cervico-auricularis, the anterior limit of its origin being even with that
of the ceroico-auricularis and the posterior slightly cephalad of it. Its
two halves separate on the occiput and diverge over the parieto-squa-
mosal region to overly the caudal part of the temporal muscle. Directly
above the ear, it passes into the ental surface of the superficial fascia or
panniculus and thence around the anterior base of the ear the two are
inseparable.

Dorso-cuticularis. — A long slender muscle arising on the side of the
twelfth thoracic vertebra and running forward along the boundary be-
tween the latissimus dorsi and the trapezius to the side of the scapula
where it widens slightly and inserts on the ental surface of the thin outer
sheet of the panniculus. Its width at its insertion is 1.5 mm. I have
been unable to find any trace of this muscle in Didelphis nor any record
of its occurrence in any other marsupial. Its form and relations are
almost exactly as in the Insectivora where it is of frequent occurrence.
Somewhat similar muscles are found also in certain edentates. (See
Plate III.)

24 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

MUSCLES OF THE BACK.

Spino-trapezius. — Origin from the spines of the thoracic vertebrae
from the fifth to the twelfth. Insertion on the middle of the vertebral
third of the scapular spine. A thin rather elongate muscle passing for-
ward superficial to the dorso-cephalic part of the latissimus dorsi.
Dorsally it is separated from the acromio-trapezius by a small fascia
overlying the rhomboideus. A separation has been noted also in Nota-
ry ctes (Wilson, 1894, p. 6). It is the usual condition in the monotremes,
insectivores and certain edentates. In Didelphis, the single muscle is
weaker over the shoulder than elsewhere and in Sarcophilus (Macalister)
also there is a tendency toward division. In Petrogale, Parsons (1896)
found that it had a "continuous fleshy origin except opposite the first
thoracic spine, where it is aponeurotic."

Acromio-trapezius. — Origin from the occipital crest slightly laterad
of the median line and thence along the nuchal crest to the vicinity of
the fifth thoracic spine. Insertion on the front of the scapular spine
and metacromion and thence along a fascia crossing the acromio-
deltoid to the clavicle. A broad thin sheet covering the dorso-caudal
half of the side of the neck and part of the shoulder.

Latissimus dorsi. — Origin from the vicinity of the fifth thoracic
spine along the vertebral column to the middle of the third lumbar
vertebra and to the lumbar fascia. Insertion in common with that of the
teres major entad of the biceps on the side of the bicipital groove. On
its lower side near its insertion is the usual connection of the epitro-
chlearis. It is more definitely connected to the teres at its insertion than
in Didelphis. It has no costal slips and in this agrees with Didelphis,
Dasyurus, Notoryctes, Phascolarctos, etc.

Rhomboideus. — This is divisible into Rhomboideus major (vertebralis)
and R. minor (capitis). R. major has its origin on the nuchal crest from
the occiput slightly caudad of the anterior border of the trapezius to the
fifth thoracic spine. Its insertion is on the inner side of the vertebral
border of the scapula cephalad of the insertion of the serratus magnus.
R. minor has its origin fleshy from the junction of the mastoid, squa-
mosal and supraoccipital bones and thence along the occipital crest a
distance of about four millimeters. It is a thin flat ribbon free from the
vertebral division until a point halfway between the occiput and the
scapula where it unites with its fellow. Its insertion is on the anterior
distal base of the scapular spine dorsad of the insertion of the "atlanto-
scapularis." The rhomboideus in marsupials is usually undivided.

Levator scapulae. — Origin from the transverse processes of the last
five cervical vertebrae just entad of the scalenus. Anteriorly a single slip

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 25

passes ectad over the scalenus from an origin on the superior transverse
process of the third cervical. This arrangement differs from that of
Didelphis in which the scalenus is entirely laterad of the serratus. In-
sertion on the posterior vertebral border of the scapula craniad of
the insertion of the serraius and continuous with it.

Serratus posterior superior. — Insertion by digitations from the
fourth to the ninth ribs inclusive. Dorsally it is mostly aponeurotic
but a broad fleshy prolongation extends anteriorly to the lower border of
the splenius, thence becoming aponeurotic as it crosses that muscle and
continues between the splenius and the rhomboideus to the most anterior
point of its origin on the nuchal crest above the sixth cervical spine.
Thence caudad its origin follows the mid-dorsal line and its wide apo-
neurosis covers the longissimus dorsi and continues into the lumbar
fascia.

Serratus posterior inferior. — Origin on the middorsal line with the
lumbar fascia. Insertion from the tenth to the thirteenth ribs. Practical-
ly as in Didelphis.

Splenius. — Origin from the nuchal crest from the vicinity of the
front of the axial spine to the second thoracic vertebra. Insertion on the
occipital crest from the mastoid process of the squamosal to a point
about 2 mm. from the median line. Its lower border is free as usual and
near its caudal extremity it is somewhat attached to the aponeurosis
of the serratus posterior superior.

Biventer cervicis. — Origin from the transverse processes of the fifth,
sixth, and seventh thoracic vertebrae. Insertion fleshy or very slightly
aponeurotic on the occiput between the median line and the complexus
to which its lateral border is attached for some 4 mm. Anteriorly it
reaches nearly to the parietal bones. It is a larger and thicker muscle
than in Didelphis and lies with its counterpart for a great part of its
length in the space between the elevated axial spine and the second
thoracic spine.

Compkxus. — Origin from the articular processes of and the aponeu-
rotic arches between the last six cervical and the first five thoracic
vertebrae. A large slip somewhat connected also with the longissimus
capitis runs forward to the transverse process of the atlas. Insertion
by a broad tendon on the occipital crest laterad of the biventer to within
about 3 mm. of the mastoid bulla.

Longissimus capitis. — Origin from the articular processes of the last
six cervical and the first four thoracic vertebrae in intimate relation
with the slips of origin of the complexus. Insertion by broad tendon on
the upper cranio-lateral border of the mastoid bulla and to a slight
extent on the adjoining part of the squamosal.

26 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

The biventer, complexus, and longissimus capitis are well distinguished
from each other and in their general relations are as in Didelphis. At
least one distinct tendinous inscription crosses the anterior part of each
some 4 mm. from the insertion.

Rectus capitis posterior superficialis. — Origin from the crest of the
axial spine. Insertion entad of the complexus on the occipital crest. It
runs downward and forward with free upper and lower borders, leaving
a triangle on either side of the median line in which the much smaller
halves of the rectus minor and part of the rectus major are exposed.

This muscle, often found in carnivores, is not generally reported for
marsupials, although Parsons (1896, p. 694) reports the rectus major
as bilaminar in Pelrogale. It is thin and flat and covers a great part of
the rectus capitis posterior major. In Didelphis, it is only partially
separable from the rectus major, in Marmosa it is somewhat freer, and
in Ccenolestes it is quite independent.

Rectus capitis posterior major. — Origin from the craniad surface of
the axial spine and intermuscular septa with the obliquus inferior.
Insertion on the occiput from the base of the paroccipital process dorsad
along the caudal border of the mastoid bulla to a point on the occipital
crest midway to the median line. It meets its fellow of the opposite
side in front of the axial spine and the two have tendinous connection
there and to a slight extent also with the caudal border of the neural
arch of the atlas. Its under surface has relation with the rectus capitis
posterior minor and with the surface of the neural arch of the atlas. It
is a rather thick muscle only slightly smaller than the obliquus inferior
with which its lower border is in apposition.

Rectus capitis posterior minor. — Origin from the straight anterior
border of the neural arch of the atlas. In a female specimen this meas-
ures 3 mm. in length from side to side. Insertion on the occipital from
a point near the median line nearly to the mastoid bulla. Its outer
surface has relation with the biventer, complexus, and rectus posterior
major. At the median line of the occiput it is separated from its counter-
part by i — 2 mm. The two muscles approach each other and have
tendinous union in their caudal halves.

Obliquus capitis superior. — Origin from the lateral surface of the
transverse process of the atlas. Insertion tendinous on the occipital
ridge immediately dorsad of the mastoid bulla. A short flat muscle
crossing nearly at right angles to the large obliquus inferior and rectus
capitis posterior major which has relation with its inner surface.

Obliquus capitis inferior. — Origin from practically the whole lateral
surface of the axial spine. Insertion on the whole caudal surface of the
transverse process of the atlas. It is a thick heavy muscle running

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 27

downward and forward. In Didelphis, this muscle extends to the
third, fourth and fifth cervical spines, but in Marmosa it is much as in
C&nolestes.

Rectus capitis anticus major. — Origin on the transverse processes of
the last five cervical vertebrae. Insertion on the posterior third of the
basisphenoid and the anterior two fifths of the basioccipital slightly
laterad of the median line. Some of its superficial fibers extend caudad
with the longus colli to the ventral side of the posterior thoracic verte-
brae. Anteriorly it forms a good sized bundle and is not peculiar except
for the extent of its attachment to the basisphenoid which is more than
usual.

Rectus capitis anticus minor. — Origin on the middle third of the ven-
tral surface of the neural arch of the atlas midway between the median
tubercle and the transverse process. Insertion on the ventral surface
of the basioccipital in its anterior lateral fourth. A thin flat muscle
lying entad of the rectus anticus major and mediad of the rectus capitis
lateralis but entirely free from both.

Rectus capitis lateralis. — Origin on the transverse process of the atlas.
Insertion slightly tendinous on the outer edge of the basioccipital along
the carotid canal and thence forward to the basisphenoid and by tendi-
nous fascia to the presphenoid. Certain lateral fibers attach on the
periotic. It is a well developed muscle and entirely distinct, extending
farther cephalad than in Didelphis.

Longus colli. — The dorsal part of the longus colli is rather distinct,
extending from the transverse process of the sixth cervical along the
bases of the ribs and the sides of the vertebrae to the fourth dorsal
vertebra. Between its two halves are fibers which extend into the
cervical part but they are lateral in position and the median ventral
surfaces of the dorsal vertebrae are mostly covered only with tendinous
fascia. The cervical part consists of various converging slips from the
transverse processes of the cervical vertebrae and caudally from the
sides of the ventral surfaces of the vertebrae. Over the atlas it shows six
rather distinct divisions, two medial ones inserting on the tubercle of the
atlas and two on either side inserting on the body.

Spinalis dorsi.—The spinalis dorsi disengages from the medial side
of the longissimus about opposite the eighth thoracic vertebra, the cau-
dal part of the biventer cervicis forming a wedge between the two. Its
inner fibers run to the spines of the anterior thoracic vertebrae but its
lateral fibers form a practically distinct subvertical ribbon which runs
forward and inserts on the side of the low spines and bodies of the fourth
to the seventh cervicals and the first thoracic. It passes the high spine
of the second thoracic without definite attachment, but between the

28 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

first thoracic and the tip of the spine of the second it is traversed by a
strong tendinous inscription which may be partially attached to either
or both of the vertebrae. Anteriorly it passes between the lateral division
of the semispinalis cervicis and the long interspinal which connects the
spines of the fifth cervical and the second thoracic. In the opossum the
spinalis inserts as high up as the third cervical and the outer surface of
its anterior part is wholly in relation with the complexus.

Semispinalis cervicis. — In the region between the axial spine and the
second thoracic spine are two paired muscles which run ventrad toward
each other and meet at an angle over the sixth and seventh cervicals,
the posterior one passing inside the spinalis between the two halves
of the anterior one. The anterior of these muscles has been regarded
as the semispinalis cervicis. It arises on the bodies of the first thoracic
and the fifth, sixth and seventh cervical vertebrae and converges to insert
on the tip and side of the axial spine in opposition to the obliquus
capitis inferior. Its outer surface has relation with the complexus and
its inner with the spinalis. It is not apparent in Didelphis, but is present
in Marmosa. The posterior paired muscle, above-mentioned, which is
in the nature of an interspinal, may be described here also. Its origin
is on the side of the spine of the fifth cervical and its insertion on the
front of the tip of the spine of the second thoracic vertebra. It stretches
directly between these spines without connection with the intervening
vertebrae. Between its two halves the ligamentum nuchae passes to the
vertebral spines. Beneath its caudal part, but distinct from it, a paired
intervertebral connects the spines of the first and second thoracic
vertebrae,. Its outer surface has relation with the biventer and the
spinalis, its inner with the ligamentum nuchae. A corresponding muscle
is found in the opossum running to the sides of the thickened third
cervical spine, its two sides therefore being widely apart and, although its
actual relations are much as in Canolestes, its appearance is quite
different. In Marmosa the third, fourth and fifth cervical spines are not
enlarged as in Didelphis,. Consequently the arrangement of the neck
muscles is more nearly as in C&nolestes and many other marsupials.

Longissimus. — Anteriorly this is divisible into longissimus dorsi and
longissimus cervicis. Superficially the two divisions appear as one
muscle but the outer part which continues caudad may be partially
separated from an inner mostly concealed part which only extends cau-
dad to the fifth thoracic vertebra and which may be regarded as the
longissimus cervicis. It arises by slips from the laminae of the cervicals
and from the transverse processes of the thoracic vertebrae from the
first to the fifth. It inserts on the transverse processes of the cervical
vertebrae from the second to the fourth or fifth, being inseparable at this

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 29

point from the longissimus dorsi. The longissimus dorsi separates from
the spinalis dorsi about opposite the eighth thoracic vertebra and passes
cranio-ventrad between the iliocostalis and the biventer and longissimus
capitis from which it is entirely distinct. Its insertion is just caudad
of that of the longissimus cervicis on the transverse process of the fifth
and possibly also the fourth cervical vertebra. Its origin is from the
lumbar tendinous sheet from which it becomes fleshy in the vicinity of
the second lumbar vertebra. Passing forward it is conjoined with the
spinalis and consists of slips from the transverse processes of the
vertebrae.

Iliocostalis. — An anterior division or iliocostalis dorsi and a posterior
or iliocostalis lumborum are fairly distinct. The iliocostalis dorsi arises
by slips from the sides of the ribs from the first to the eighth. It passes
cranio-ventrad on the side of the neck laterad of the longissimus and
inserts on the superior transverse processes of the sixth and seventh
cervical vertebrae. For a short distance just before its insertion it is
merged with the longissimus.

The iliocostalis lumborum arises from the complicated lumbar ten-
dons of the sacrospinalis and becomes fleshy opposite the fourth lumbar
vertebra. It inserts by thin flat slips on the sides of the ribs from the
sixth to the thirteenth. Between the ilium and the ribs it forms the
anterior lateral boundary of the mass of muscle filling the space between
the articular and transverse processes of the lumbar vertebrae. Over the
ribs it is thin and the slips are fairly distinct from each other. Between
the sixth and the ninth ribs it overlaps the iliocostalis dorsi which only
reaches the eighth rib.

Sacrospinalis. — Origin from the crest and the ventro-medial surface
of the cranial fourth of the ilium, meeting the origin of the iliacus, and
from the tendinous lumbo-dorsal fascia. Insertion on the dorsal surfaces
and caudal edges of the transverse processes of the lumbar vertebrae
and the twelfth and thirteenth dorsals. A thick rounded bundle ex-
tending from the ilium to the ribs. Easily separable from the multifidae
which lie mediad of it but covered by the tendinous fascia. It fills the
space on the side of the vertebral column between the articular and the
transverse processes of the lumbar vertebrae.

Quadratus lumborum. — Origin from the bodies of the i2th and i3th
dorsal vertebrae and the lumbar vertebrae from the first to the fourth.
Insertion on the tips and ventral surfaces of the transverse processes of
all the lumbar vertebrae except the last. It lies somewhat flattened
between the psoas parvus and the large rounded sacrospinalis and con-
nects the series of transverse processes with each other by series of
overlying tendons, the fibers from the under side of one running to the

30 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

upper side of the one succeeding. Anteriorly it is closely associated
with the psoas parvus.

Psoas parvus. — Arises fleshy and also by tendinous heads from the
first, second, third, fourth and fifth lumbar vertebrae. Thence it
continues caudad as a broad flat tendon crossing the surface of the medial
part of the psoas magnus and inserting dorso-craniad of the marsupial
bone on the iliopectineal tubercle. In its cranial part it is closely con-
nected with the quadratus lumborum which lies laterad of it. The caudal
part of its fleshy origin on the fifth lumbar vertebra is overlapped by the
psoas magnus which intervenes between it and the quadratus lumborum
at this point.

Psoas magnus. — Origin from the sides of the bodies and the ventral
bases of the transverse processes of the fourth, fifth, and sixth lumbar
vertebrae and the cranial half of the first sacral. Anteriorly it reaches
to the posterior epiphysis of the third lumbar. Insertion with the
iliacus on the lesser trochanter and the adjacent inner surface of the
femur nearly as far distad as the insertion of the pectineus.

i

MUSCLES OF THE HEAD.
Plates III-IV.

Masseter. — The superficial masseter has its origin on the lower side
of the zygoma by a short stout tendon nearly opposite and slightly
caudad to that of the inferior zygomaticus and in the vertical plane of
the middle of the eye. It spreads over the lower part of the mandibular
insertion of the deep masseter and covers the ventro-caudal part of
the mandible with its insertion from the base of the coronoid process
just caudad of the buccinator and along the lower side of the mandible
to the angle and the pterygoideus. It does not reach the inner surface of
the mandible and especially in its ventro-caudal portion is relatively
weaker than in Didelphis. The deep masseter lies below it and is closely
connected, but a division is evident. It has its origin on the inner and
lower sides of the zygoma and the fascia over it and is intimately asso-
ciated with the temporalis. It is broadly attached to the whole outer
face of the coronoid fossa.

Temporalis. — This is in two layers. The superficial or outer temporal
arises from the fascia along the middle of the parietal bone and across
the caudal part of the underlying layer. Thence it spreads as a rather
thin sheet over the deep temporal and narrows to pass into the orbital
fossa and insert on the mandible at the anterior base of the ascending
ramus. It is also attached superficially to the membrane between it and
the eye. The deep temporal is more extensive; its origin runs from the

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 31

glenoid fossa across the mastoid bulla along the supraoccipital crest
halfway to the median line and thence directly forward along the
rather indistinct parietal and supraorbital crests to the plane of the
posterior canthus of the eye. Its insertions are extensive on the coronoid
process, mostly on its ental surface and its anterior edge with a few of
the fibers from the mastoid origin going to its ectal surface.

Buccinator. — Relatively extensive and forming a pouch behind the
angle of the mouth. Its origin is along the alveolar border of the maxil-
lary from the vicinity of the infraorbital foramen to the anterior base of
the zygoma and likewise along the ramus of the mandible from the front
and base of the coronoid process forward to the vicinity of the mental
foramen. It is attached to or merged with the orbicularis oris around the
lips.

Pterygoideus internus. — Origin from the whole of the external
pterygoid fossa from the carotid foramen cephalad, passing the tips of
the pterygoids but attached to their bases and thence along the ridge to
the dorsal side of the lateral tuberosity of the palatal ridge. Insertion
broad on the inner surface of the angle of the mandible and by certain
medial fibers running to the fascia over the tympanic. Mediad of the
internal pterygoid the small but distinct tensor iieli palatini runs from the
periotic and the edge of the basisphenoid to the hamular process of the
pterygoid bone.

Pterygoideus externus. — Origin by two heads. The larger ventral one
arises fleshy from the whole of the lateral base of the pterygoid bone
laterad of the pterygoideus internus. The smaller dorsal head arises
from the alisphenoid caudad and laterad of the sphenoidal fissure. The
two heads spread over the convex surface of the alisphenoid between
the sphenoidal fissure and the glenoid fossa and, after uniting as they
pass the front of the tympanic, insert at the base of the mandibular
condyle principally on the stylomandibular ligament.

The muscle is relatively large and broad, bulking fully two-fifths as
much as the internal pterygoid. It is thus quite different from that of
Didelphis in which it is single-headed and very weak. It is described
(Macalister) in Phascolarctos as a "small rudiment crossing and insepa-
rable from the internal" and in Sarcophilus as "exceedingly feeble"; in
Myrmecobius also, it is weak and single (Leche) ; in Choeropus (Parsons,
1903, p. 67), it is unusual and inserts on the inside of the mandibular
ramus; in Petrogale (Parsons, 1896) and other kangaroos, it is small;
for most other marsupials, data are lacking.

Mandibulo-auricularis. — The only prominent ear muscle seems to
correspond to the one for which this name has been used by various
authors. It is a small subcylindrical muscle arising in the fascia between

32 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

the superficial and deep masseters near the mandibular condyle and issu-
ing between the masseters to an insertion on the anterior base of the ear.

Zygomaticus. — This is in three divisions all arising on the zygoma.
Two of them originate from the lateral surface of the posterior half of
the zygoma behind the eye. They are small well-formed fasciculi closely
similar to each other and somewhat distinguished from the third division
which lies below (entad) them. They run forward to the level of the
angle of the mouth where they become tendinous and proceed thus for
some distance to insertions in the tissue on the side of the upper jaw at
the base of the whiskers.

The third division, to which some other name may be applicable,
is similar in appearance but arises from the side of the lower anterior
angle of the zygoma and its slender tendon is inserted into the skin of the
inside of the upper lip in the vicinity of the base of the labret.

Retractor naris. — A slender muscle arising from the fascia between
the eye and the zygoma. It crosses the outer surface of the anterior root
of the zygoma and becomes a very slender threadlike tendon which runs
forward to an insertion on the nasal cartilages.

Digastricus. — Arises on the paroccipital process between the condyle
and the mastoid bulla and runs down around the ear and thence for-
ward to spread over the interramal space and attach along the sides of
the mandible from the base of the angular process to the symphysis.
Anteriorly it is largely aponeurotic. A tendinous slip diverges at a point
ventrad of the masseter and passes dorsad to the caudal median border
of the mylohyoid. It is essentially as in Didelphis.

MUSCLES OF THE NECK.
Plates IV, VI.

Sterno-hyoideus. — This is the principal muscle exposed on removing
the skin from the lower side of the neck. It arises on the ental surface
of the second sternebra and runs directly forward opposite its mate and
inserts on the posterior side of the hyoid bone caudad to the insertion of
the genio-hyoid. Some of its superficial fibers inosculate with those of
the genio-hyoid. Like the other muscles connected with the hyoid and
the tongue, it is not peculiar.

Genio-hyoideus. — A flat paired muscle immediately cephalad of the
sterno-hyoideus and entad of themylo-hyoideus. It arises on the body of
the hyoid bone and inserts on the mandibular symphysis.

Sterno-thyroideus. — A thin, paired muscle lying between the trachea
and the sterno-hyoideus- Laterally it is thin and transparent and
toward the median line it is somewhat thicker, especially in its caudal

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 33

portion where its two sides become fused. It arises on the sides and ental
surface of the second sternebra and inserts on the thyroid cartilage.

Mylo-hyoideus. — A very thin, paired, transverse muscle between the
mandibles and immediately entad of the digastricus. It arises from the
inside of the mandible near the alveolar border and has no bony insertion,
merely meeting its fellow in a raphe in the median line and becoming
aponeurotic anteriorly and posteriorly. Along the dorsal side of this
raphe it is attached to the genio-hyoideus. Its posterior fleshy border
runs from the angle of the mandible forward to the median line opposite
the base of the angular process.

Hyoglossus. — Arises laterad and dorsad of the genio-hyoideus on the
ventrocaudal border of the thyrohyals and basihyal cephalad of the
insertion of the ihyrohyoideus. It passes ventrad of the ceratohyals and
free from them, diverges to enclose the genioglossus, and passes dorsad
and cephalad into the body of the tongue.

Genioglossus. — Well-developed and having the usual relations, aris-
ing on the sides of the mandibles near the symphysis and forming a
wedge between the two halves of the hyoglossus.

Styloglossus. — Runs laterad of the hyoglossus and merges with it
about midway of the mandible. Its origin was not determined, having
been destroyed inadvertently.

Thyrohyoideus. — Origin on the thyrohyal bone entad of the sternohy-
oid and the omohyoid and opposite the hyoglossus. Insertion on the
posterior corner of the bony wing of the thyroid cartilage. A short stout
triangular muscle, relatively well-developed.

Omo-hyoideus. — Origin on the lateral fourth of the posterior side of
the hyoid bone. Insertion on the upper posterior angle of the scapula
just craniad of the insertion of the levator scapulae. On the ental side
near its origin it is joined to the sternohyoid by thin membrane.

Sterno-mastoideus . — Origin on the manubrium of the sternum slightly
overlapped by the edge of the pectoral-is. Insertion divided, one rather
thin sheet inserting by broad aponeurosis on the occipital crest from a
point midway between the median line and the auditory meatus to the
middle of the anterior border of the mastoid bulla. The other division
is a thickened bundle and inserts by a short tendon on the mastoid pro-
cess slightly ectad and cephalad of that of the cleido- mastoid. The two
divisions become fused about midway between origin and insertion.

Cleido-tnastoideus. — Origin on the front of the distal end of the cla-
vicle. Insertion fleshy or only slightly tendinous on the mastoid pro-
cess. Its outer surface has relation with the splenius, the sterno-mastoid-
eus and the cleido-occipitalis; its inner with the digastricus, atlanlo-
acromialis and omo-hyoideus.

34 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Cleido-occipitalis. — Origin on the clavicle just laterad of the sterno-
mastoid. Insertion by aponeurosis on the side of the occiput just laterad
of and in the same plane with the acromio-trapezius. On its outer sur-
face its relations are with the ceruico-auricularis; on its inner with the
rhomboideus, the atlanto-acromialis, and the cleido-mastoideus.

Omo-deido-transversarius . — This includes two distinct muscles as in
Didelphis, a dorsal part or "atlanto-scapularis" and a ventral part or
" atlanto-acromialis" The dorsal part has its origin on the side of the
body of the atlas immediately laterad of the origin of the ventral part.
The insertion is on the front of the scapular spine in its vertebral fourth
just ventrad of the insertion of the rhomboideus minor. It is a weaker
more slender muscle than the ventral part and does not widen so at its
insertion. The ventral part has origin on the side of the hypapophysial
tubercle of the atlas. Its insertion is broad from the caudal side of the
metacromion process of the scapula and for a short distance ventrad
along the deltoid fascia over the acromio-deltoid muscle where it sends
a few fibers ventrad.

MUSCLES OF THE THORAX.

Pectoralis minor. — Origin from the sides of the fourth and fifth
sternebrae. Insertion on the lesser tuberosity of the humerus and the
deltoid ridge by a broad slightly tendinous aponeurosis much wider
than in Didelphis. Its outer surface has relation with the pectoralis
major, its inner with the rectus dbdominis, scalenus, etc.

Pectoralis major. — Origin from the median line of the entire sternum.
Insertion on the distal part of the inner edge of the deltoid ridge. Its
outer edge is curled under and near the insertion it has a slight aponeu-
rotic connection with the ociphihumeralis as in Didelphis.

Supracostalis. — Origin by broad aponeurosis from the second, third,
fourth and fifth sternebrae. Insertion fleshy on the side of the first rib.
This muscle, which Coues called " sterno-costalis" and others designate
transversus costarum, has practically the same relations as in Didelphis.
Its broad aponeurosis for its entire length covers the rectus abdominis.

Subclavius. — Arises from the first rib and inserts on the outer side of
the distal end of the clavicle, not extending to the scapular spine as in
Didelphis. In this it agrees with Phascologale, and some other Australian
forms (Macropus sp., Phalanger sp.).

Serratus magnus. — Origin posteriorly by six very distinct digitations
on the sides of the ribs from the third to the eighth inclusive. Anteriorly
there are no distinct digitations and the origin is from the joints of cos-
tal cartilages with the first, second and third ribs. Insertion on the pos-

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 35

terior border of the scapula where it is continuous with the levator
scapulae.

Scalenus. — This has two fairly distinct divisions, at least posteriorly.
The scalenus medius, or larger division, takes origin from the superior
transverse processes of the third, fourth, and fifth cervical vertebrae
over which it is somewhat narrow and thickened. Thence passing caudad
it becomes thin and flattened and inserts by broad distinct slips on the
middle of the third and fourth ribs, one slip on the third and two on the
fourth, passing beneath and separating digitations of the serratus
magnus in so doing. The scalenus anticus has its origin in coalescence
with the medius, but about midway between the third cervical vertebra
and the first rib it separates and inserts on the front of the first rib
directly opposite the insertion of the supracostalis.

MUSCLES OF THE ABDOMEN.

Rectus abdominis. — Origin from the pubis and marsupial bone. In-
sertion on the sternal part of the first rib and on the adjoining side of the
manubrium of the sternum. It is strong and well developed anteriorly,
broader but thinner abdominally.

Obliquus externus abdominis. — Origin from the middle of the sides of
the ribs from the fourth to the twelfth, interdigitating anteriorly with
the serratus. Thence it follows the lumbar fascia to Poupart's ligament
which is inserted at the inner anterior base of the marsupial bone.
Ventrally it spreads over the abdomen covering the rectus and inserting
by aponeurosis on the linea alba.

Obliquus internus abdominis. — Origin from Poupart's ligament and
the ilium and thence dorsad to the lumbar fascia and craniad to the
posterior borders of the loth to i3th ribs. It is fleshy in its dorsal portion
down to a line from the ventral third of the costal cartilage of the tenth
rib diagonally across the side of the abdomen to the pubic region. Ven-
trad of this line the transversalis appears and no aponeurosis of the
obliquus is demonstrable in the material available.

Transversus abdominis. — Origin from the inner sides of the nth,
1 2th, and i3th ribs and from a coalescence with the diaphragm and
thence along the lumbar fascia to the ilium. Ventrally it is inserted
into the linea alba. Posteriorly it is broadly aponeurotic and anteriorly
mostly fleshy.

Cremaster. — Arises in the male beneath the internal oblique near the
crest of the ilium, and passes ventro-caudad across the iliacus and psoas
major in the plane of the transversalis but separated from it by a slight
space. It passes out of the body cavity through the ring in front of the

36 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

marsupial bone and proceeds to the mid-ventral line and the testes. It
is a well developed strand of muscle about one mm. in width but simply
parallels or covers but does not enclose the spermatic cord until near the
base of the testes where it nearly or quite surrounds the cord before
spreading over the base of the testes and becoming aponeurotic.

MUSCLES OF THE SACRUM.
Plates III, V.

Glutens maximus. — Origin from the crest and the anterior part of the
acetabular border of the ilium and from the articular processes of the
sacral and anterior caudal vertebrae as well as from the tendinous fascia
covering the erector spinae as far caudad as the origin of the femoro-
coccygeus. Insertion somewhat divided, an anterior part on the cranio-
lateral and a posterior one on the caudo-lateral surface of the distal end
of the great trochanter of the femur. It is a broad fan-shaped muscle,
relatively thicker than in Didelphis, being fleshy from the erector
spinae. Its cranial border is thickened and somewhat curled around the
anterior fifth of the ilium and distad lies in the groove between the
glutens medius and the iliacus, continuing ectad of the proximal end of
the vastus externus to the cranial border of the biceps femoris to which it
has a short semitendinous attachment distad and ectad of its insertion
on the great trochanter. Its caudal border parallels the femorococcygeus
but except in the dorsal fascia is distinct from it. No separate tensor
fascia femoris is evident and it is doubtless fused with the glutens as in
various other marsupials.

Glutens medius. — Origin extensive from the dorsal fascia, the
laminar tendons of the caudal muscles, the transverse processes of the
sacral vertebrae, and the crest and ischial border of the ilium. Insertion
on the caudo-lateral proximal part of the great trochanter of the femur
ectad of that of the gluteus minimus which is distinct. It lies entad of
the gluteus maximus and is almost equal to it in superficial extent. It
is imperfectly divisible into two layers, the posterior one overlapping the
anterior. A similar bilaminar arrangement is reported in Phascolarctos ,
Thylacinus, and Dasyurus, but not in Didelphis l and Cuscus (see Leche,
Bronn's Thierr., p. 853). It is likewise in Notoryctes (Thompson and
Hillier, 1905, p. 310). On its ental side it is somewhat connected with
the pyriformis not only near the insertion but also elsewhere except
along the anterior tendinous border of the pyriformis. Anteriorly it is
curled under from the crest of the ilium to the femur but in the distal
half of this extent its edge is free and merely overlies the gluteus minimus.

1Sidebotham (1885, p. 13) reports three layers in Chironectes.

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 37

Glutens minimus. — Origin from the dorsal surface of the ilium from a
point on the acetabular border about midway between the crest and the
acetabulum to a point opposite the middle of the acetabulum. Insertion
broadly covering the antero-internal surface of the great trochanter of
the femur. It is relatively large and completely distinct not only
anteriorly from the glutens maximus and glutens minimus but posteriorly
from the gemelli. In the didelphids and various other marsupials it is
partially joined to the maximus.

Pyriformis. — Origin fleshy from the dorso-lateral surface and the
caudal edge of the crest of the ilium and from the tips of the transverse
processes of the two sacral and the first caudal vertebrae.

Insertion on the apex of the great trochanter of the femur just caudad
of the insertion of the glutens minimus and en tad of the glutens medius.
A very large muscle lying entad of the glutens medius and ecto-caudad of
the glutens minimus. Its anterior border which is full and rounded runs
direct from the trochanter to the iliac crest and at least in its distal half
is tendinous. Fleshy fibers pass dorsad from this broad tendinous edge
to the sacral border. Its outer surface, at least caudally, is somewhat
connected with the glutens medius.

Quadratus femoris. — Origin on the inner side and slightly ventrad of
the tuberosity of the ischium and on the inner edge of the ischium just
craniad of the tuberosity. The fibers curve over the lesser sciatic notch
to pass outward to their insertion on the middle of the ventral edge
of the great trochanter of the femur and on the intertrochanteric ridge.
It is relatively large and well developed and fairly distinguishable from
the gemelli anteriorly.

Gemelli. — No distinction is evident between a gemellus superior and
inferior. The origin is on the body of the ischium dorsad of the origin
of the ischiococcygeus. The insertion is with the tendon of the obturator
internus on the caudal edge of the great trochanter of the femur and in
the digital fossa proximad of the insertion of the quadratus femoris. Its
cranial border is distinct from the caudal border of the glutens minimus.

Iliacus. — Origin from the ventro-lateral or iliac surface and the
acetabular border of the cranial two-thirds of the ilium from the crest
caudad. In its caudal portions its insertion is opposite that of the
iliococcygeus on one side and the glutens minimus on the other with a
part of the psoas magnus lying loosely between. Its general relations are
closely similar to those in Didelphis.

Obturator internus. — Origin from all of the caudal border of the ramus
of the ischium which is ventrad of the quadratus femoris, from the inner
side of the pubic symphysis, and from the caudal border of the obturator
foramen. Insertion by a flat tendon in the digital fossa of the femur.

38 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Obturator externus. — Origin from the side of the ascending ramus of
the ischium from the lower edge of the tuberosity to the symphysis and
thence on the side of the pubic ramus to a point nearly opposite the
anterior base of the marsupial bone. Insertion by a thick tendon in the
digital fossa and on the intertrochanteric ridge of the femur. It is a
broad thickened muscle lying at its origin entad of the adductors. Dor-
sally it meets the edge of the quadratus femoris.

The obturator externus is said by Leche (p. 858) to be "nichts be-
merkenswerthes " in the Marsupialia, but Macalister (1870, p. 167)
states that it is "large and normal in Phascolomys and Sarcophilus, as
well as in Macro-pus giganteus, the Wallaby, Phalanger, and Opossum."
Thompson and Hillier (1905, p. 312) found it well developed inNotoryc-
tes. The so-called differentiation of the obturator, which Leche names
intermedius and reports only for Didelpkis and Philander is doubtless
not peculiar to these forms. It was not specially sought in Ccenolestes
and may have been overlooked and included as an integral part of the
externus. Whether it be called externus or intermedius, it is apparent
that it is not peculiar in C&nolestes.

MUSCLES OF THE TAIL.

All the usual tail muscles of the pubic cavity are present and distinct
but their muscular parts do not extend far beyond the base of the tail.
On removal of the skin and the underlying fascia, the tail appears almost
entirely encased in shining tendons from its base to its tip. On the dorsal
aspect the continuations of the back muscles proceed somewhat farther
caudad but these also soon become entirely tendinous. After all the
tendons have been removed, however, a series of thin muscles connecting
the vertebrae are exposed. These are short muscles, those next the
median line extending from the chevron bone of one vertebra to that of
the one succeeding; those laterad extend from the caudal transverse
process of one vertebra to the cephalic articular process of the second
succeeding vertebra, each muscle thus spanning a vertebra.

Dorsally the tendons from the dorsal extensors cover the angles of
the tail, passing over the articular processes of the vertebrae, but in the
median line the short intervertebral tail muscles are exposed.

Abductor caudae externus. — Origin on the medial side of the iliac
crest and by tendinous connection with the lumbo- dorsal fascia. In-
sertion on the dorsal surfaces and tips of transverse processes of the
caudal vertebrae. This is a spindle-shaped muscle filling the great part
of the space between the transverse and the articular processes of the
anterior caudal vertebrae and passing out on the side of the tail where it
narrows and soon becomes wholly tendinous.

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 39

Extensor caudae lateralis. — Origin by slips from the articular processes
of the sacral and caudal vertebrae. Insertion on the sides and transverse
processes of the caudal vertebrae. Lies between the abductor caudae
externus and the multifidae which are continued to the base of the tail
as an extensor caudae medialis.

Pubococcygeus. — Origin from the whole inner caudal border of the
pubic ramus. It forms a thin sheet over the obturator foramen and
continues fleshy to the front of the fifth caudal vertebra whence it finds
attachment in the aponeurosis on the surface of the tail. In some cases
it is slightly tendinous on its inner side and attached to the chevron bone
of the fifth caudal. Laterally it has slight attachment to the colon.

Sacrococcygeus. — Origin from the ventral surfaces of the transverse
processes of the sacral and the first four caudal vertebrae. Thence by
numerous slender tendons it proceeds along the side of the tail to the
distal caudal vertebrae. These tendons lie slightly laterad of those of
the infracoccygeus and fill the space between the cephalic articular and
transverse processes of the caudal vertebrae. Its muscular heads are
thick and fleshy and lie between the caudal part of the infracoccygeus
and the pubococcygeus and ventrad of the ischiococcygeus.

Ischiococcygeus. — Origin from the inner side of the ilium from a
point opposite and slightly craniad of the caudal border of the gluteus
minimus caudad about halfway to the tuberosity of the ischium. In-
sertion on the tips of the transverse processes of the first four caudal
vertebrae.

Iliococcygeus. — Origin from the ischial border of the ilium from a
point just caudad of the sacrum (posterior inferior spine of ilium) and
thence caudad along the whole inner surface of the ilium (great sacro-
sciatic notch) nearly to the anterior border of the obturator foramen. A
few fibers rise from the tendons of the psoas minor and at this point are
closely connected with those of the pubococcygeus. Insertion by four
distinct tendons which converge to the middle of the ventral surface of
the tail. The outer tendon runs just laterad of the mid-ventral line of
the tail to the chevron bone of the eighth caudal vertebra. The next
one runs similarly to the seventh caudal. The inner pair of tendons run
first mediad, then entad, and finally laterad of the outer pair. Of this
inner pair the outer one inserts on the chevron bone of the ninth and the
other on the tenth caudal. Near their attachments each of the tendons
gives off a slight muscular fasciculus.

Infracoccygeus. — Origin from the medio-ventral surfaces of the
bodies of the fifth and sixth lumbars and the two sacral vertebrae. In-
sertion by long slender tendons running on the ventral side of the tail
next to the median line. The innermost tendon runs to the middle of

40 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

the front of the first caudal. The next tendon laterad runs to the
eleventh caudal, the next to the twelfth and so they continue to the
twentieth caudal. The muscle is broadly spindle shaped and occupies
the sacral basin, its two halves forming an elongate ellipse between the
two halves of the psoas magnus anteriorly and of the iliococcygeus
posteriorly. It consists of superposed laminae each becoming tendi-
nous before issuing from the pubis.

MUSCLES OF THE ANTERIOR LIMB.
Plate V.

Spino-deltoideus. — Origin from the middle third of the scapular
spine. Insertion on the distal end and outer edge of the deltoid ridge of
the humerus. At its insertion it passes beneath and nearly at right angles
to the acromio-deltoideus.

Acromio-deltoideus. — Origin on the acromion and the proximal half
of the clavicle. Insertion on the inner end of the flattened surface of the
deltoid ridge opposite the insertion of the pectoralis major. It covers the
rather broad surface of the deltoid ridge and is a thicker muscle than in
Didelphis and more distinct from the spino-deltoideus and the pectoralis.
It is fan-shaped, the fibers from the clavicle and those from the acromion
converging toward a median raphe and coming to a point ventrally.

Subscapularis. — Origin from practically the whole of the expanded
ental surface of the scapula. Insertion by a short broad tendon on the
lesser tuberosity of the humerus. It is slightly thicker dorsally than
ventrally but in general is thin and flat.

Supraspinatus. — Origin from the surface of the supraspinous fossa.
It passes entad of the clavicle between the coracoid and the acromion
and inserts on the greater tuberosity of the humerus. It is about twice
as large as the infraspinatus and relatively larger than in Didelphis,
agreeing more nearly with Phalanger, Perameles, Phascologale, etc.

Infraspinatus. — Origin from the infraspinous fossa, including the
ental surface of the metacromion and acromion, the axillary border of
the scapula, and from a tendinous raphe between it and the proximal
part of the teres. Insertion partly fleshy and partly tendinous on the
greater tuberosity of the humerus. A weak aponeurosis from the
infraspinous fossa near the base of the acromion may represent the
teres minor, but it is very indefinite. A teres minor is reported for Chiro-
nectes (Sidebotham) but not for Didelphis.

Epitrochlearis. — Origin from the lower border of the latissimus slight-
ly (2 mm.) before it passes beneath the biceps. Insertion on the olecranon
process of the ulna.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 41

Teres major. — Origin from the posterior ventral angle of the scapula
in close relation with the subscapularis. Insertion on the bicipital groove
slightly proximad of and in conjunction with that of the latissimus.

Coracobrachialis brevis. — Origin on the coracoid process entad of the
biceps. Insertion on the humerus just proximad of the insertion of the
teres. It is very slightly developed, almost rudimentary, and in its
fleshy part consists of a very thin layer of fibers instead of a thick belly
as in Didelphis. In one specimen it appeared to insert in the tissue of the
medial head of the triceps rather than on the humerus.

Biceps brachii. — Origin by a single tendon from the coracoid process
of the scapula and adjacent fascia. Soon after becoming fleshy it is
readily divisible into two distinct muscles which run in close apposition
to pass into the forearm between the pronator teres and the extensor
carpi radialis brevis. The outer muscle inserts by a narrow flat tendon
on the tubercle of the radius. The inner muscle is partially attached to
the brachialis near its insertion and then its tendons divide and pass on
either side of the tendon of the brachialis to insertions on the side of the
ulna distad of the lesser sigmoid cavity. In some of the Macropodidae
the biceps is completely divided but in some other diprotodonts and in
polyprotodonts generally its relations are reported essentially as above.

Brachialis anticus. — Origin from the lesser tuberosity and the border
of the head of the humerus passing around the middle of the front of the
humerus and thence distad. At about the middle of the front of the
humerus it becomes attached to the bone; between this and its origin
it is free or nearly so. Thence it continues along the humerus to the elbow
joint where it twists further and has final insertion just distad of the
lesser sigmoid cavity on the inner surface of the ulna by a flattened ten-
don which passes between the divided tendons of the biceps.

Triceps brachii. — The long head has its origin as usual on the axillary
border of the scapula just dorsad of the glenoid cavity. Its insertion by
common tendon with the lateral head is on the outer angle of the
olecranon process of the ulna.

The lateral head arises beneath the posterior border of the head of
the humerus and thence is free from the humerus to its union with the
long head. Its posterior aspect is hollowed to receive the long head for
most of its length and the two are fairly distinct. The median head
arises from the posterior aspect of the humerus for practically its entire
length, reaching dorsally within 2 mm. of the lesser tuberosity and
thence to the olecranon fossa. Its insertion is on the inner angle of the
olecranon process of the ulna quite distinct from that of the other two
heads.

The triceps of Canolestes is characterized by the distinctness of the

42 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

median head which in Didelphis and many other marsupials is united
with the lateral head near its origin.

Anconeus. — Origin from the inner condyle of the humerus. In-
sertion on the inner angle of the olecranon process entad of the epitroch-
learis. A small muscle connected on its lower side by a tendinous raphe
with the extensor carpi ulnaris and on its upper side slightly connected
with the median head of the triceps.

Pronator teres. — Has origin as a thick fleshy bundle from the posterior
side of the inner condyle of the humerus and the epicondylar ridge.
Insertion by broad aponeurosis on the middle third of the front of the
radius.

Flexor carpi radialis. — Origin from the inner humeral condyle and
the septa connecting with the pronator teres between which and the
palmaris longus it lies. Its long slender tendon passes over the extremity
of the radius and thence over the scaphoid to insertion at the bases of
the second and third metacarpals.1

Palmaris longus. — Origin from the septa between it and the flexor
carpi ulnaris, the flexor carpi radialis, and the digital flexors and by a
few fibers from the inner condyle of the humerus. It covers the digital
flexors almost completely and lies between the flexor carpi ulnaris and
the flexor carpi radialis.

In the single specimen dissected it was divided into two parts. The
superficial one (palmaris accessorius?) sends its tendon beside that of the
flexor carpi ulnaris to a slight attachment on the inner side of the fascia
on the pisiform bone and thence superficial to or partly attached to the
flat transverse ligament of the wrist which extends across the wrist from
the pisiform to the base of the first metacarpal. Above this transverse
ligament it extends into the palmar fascia. The deeper division sends
its tendon bound with its fellow to the transverse ligament where the
two separate and the second dips beneath the ligament to reappear in
the facia and send aponeurotic branches to the extremities of the second,
third, and fourth digits. The fleshy parts of the two divisions are
distinct for about one-half their length.

Flexor carpi ulnaris. — Origin from the inner condyle of the humerus,
from the septa between it and the palmaris longus, the extensor digitorum
communis, and the anconeus, and from the outer angle of the olecranon
opposite the insertion of the epitrochlearis. Insertion on the end of the
pisiform bone. A broad fan-shaped muscle with two sets of fibers con-
verging toward its tendon which extends a short distance into its fleshy

1 This is also the case in Didelphis, although Coues states that the insertion is on
the first metacarpal; Cunningham (1882, p. 17) found a similar disposition in Pha-
langer maculata.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 43

part. The fleshy divisions end abruptly and equally about one-third
of the distance from the olecranon to the carpus. The posterior division,
therefore is not longer than the anterior as it is in the opossum and the
tendon is shorter.

Flexor digitorum sublimis. — Origin by three heads in the septa of the
ulnar head of the profundus. These give rise to three distinct slips which
send their tendons loosely bound together until they reach the palm
where they pass beneath the annular ligament and insert ectad of the
large profundus tendons at the bases of the first phalanx of the second,
third, and fourth digits, respectively. Their tendons are slender and
practically of the same size throughout. From their point of origin a
short thick muscle with fibers running diagonally opposite those of the
ulnar head of the profundus extends to an origin from the caudad surface
of the inner humeral condyle and from the proximal edge of the glenoid
cavity of the ulna. Whether it may be one head of the sublimis is doubt-
ful. It is very distinct.

Flexor digitorum profundus. — Origin by four heads each of which
sends a distinct tendon to the palm where all unite in a thick flat tendi-
nous mass which occupies most of the palmar space. From this three
strong tendons radiate to the middle digits. Two others going to the
pollex and the fifth finger are much smaller and arise from the ectal
surface of the common tendon. The heads are as follows :

1. Ulnar head. Origin on the flattened posterior surface of the
proximal two-thirds of the ulna from the olecranon at least to a point
opposite the neck of the radius. Its tendon is very large and strong,
rounded in its proximal part, slightly tapering toward its middle, and
flattening as it joins the outer side of the palmar tendinous mass.

This is the largest fasciculus of the forearm, lying beneath the flexor
ulnaris and giving shape to the limb.

2. The second head lies superficially next to the flexor carpi radialis
and connected with it in its proximal part by intermuscular septa. It is
similarly connected with the fourth head for a distance of nearly half its
fleshy extent. Its origin is from the inner condyle of the humerus entad
of that of the palmaris longus. Its tendon joins the common tendon ectad
of that of the third head, the two being approximately of equal size
and lying midway between the tendons of the first and fourth heads
which are much larger.

3 . The third head is very distinct and wholly free but lies deeper than
the second and is not exposed until that is reflected. It lies between
the flexor ulnaris and the second head of the profundus together with the
conjoined part of the flexor radialis. Its origin is from the inner side
of the lip of the lesser sigmoid cavity of the ulna. Its small round tendon

44 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

runs direct to the common palmar tendon which it joins next to that of
the ulnar head and entad of that of the second head.

4. The fourth head, which is almost as large as the first or ulnar
head, has origin from a distinct slip from the front of the inner condyle
of the humerus and from a broad fleshy expansion arising from the inner
side of the proximal fourth of the radius. The two fleshy parts join to
form a single tendon which runs along the radius to join the common
tendon in the palm.

While obviously very different from that of Didelphis, in which there
is no division above the palm, this muscle is so variable that detailed
comparisons have not been attempted. In most marsupials it sends
tendons only to the four ulnar digits.

Abductor pollicis longus (Extensor ossis metacarpi pollicis). — Lies
in the deep groove in the proximal part of the ulna and takes origin from
the adjacent sides of the ulna and the radius, the fibers converging from
the two sides to the central tendon which extends some distance into
the fleshy part. Its origin (male specimen) is thus some 5 mm. in
length. Its rather broad flat tendon passes beneath the ulnar and lateral
digital extensors crossing the forearm and becoming superficial between
the extensor digitorum communis and the extensor carpi brevis. It then
curves around the radius and inserts on the lateral surface of the first
metacarpal.

Pronator quadratus. — This muscle, which is present in didelphids
and so far as known in all other marsupials except Notary ctes, is not found
in Ccenolestes, or at most is represented by no more than an indefinite
non-muscular fascia. It is reported as poorly developed in Thylacinus,
Sarcophilus, Phalanger and Phascologale. It is lacking in monotremes,
insectivores, bats, and certain edentates (Dasypodidae, Manidae).

Brachioradialis (Supinator longus) . — Origin from the outer epicondy-
lar ridge of the humerus proximad of the extensor carpi radialis longus.
Its tendon passes under that of the abductor pollicis longus and over a
groove in the extremity of the radius to insertion in the external lateral
ligament which overlies the extremity of the radius.

Extensor carpi radialis longus. — Origin from the outer epicondylar
ridge of the humerus distad of the brachioradialis. Insertion on the base
of the second phalanx of the second digit.

Extensor carpi radialis brevis. — Origin from the outer epicondylar
ridge and septa of the extensor digitorum communis, between which and
the extensor carpi radialis longus it lies. Insertion on the outer lateral
surface of the proximal third of the first phalanx of the third digit. It
is somewhat larger than the longus and separate from it as in Thylacinus
and Phalanger. In most other marsupials, the long and short radial

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 45

extensors are reported to be fused at least proximally, but Parsons (1896,
p. 697) expresses the belief that they are normally separate in the
kangaroos.

Extensor digitorum communis. — Origin from the distal part of the
outer epicondylar ridge, the outer condyle and from the septa of the
muscles lying on either side, namely the extensor carpi radialis brevis and
the extensor digitorum lateralis. The fleshy part divides into two slips,
the more superficial of which sends tendons to the second and third
digits and the deeper one to the fourth and fifth. The tendons are
separate but closely parallel until they pass under the transverse liga-
ment and separate to insertions on the terminal phalanges of the second
to the fifth digits. The tendons become somewhat flattened or aponeu-
rotic distally.

Extensor digitorum lateralis. — Origin on the front of the outer con-
dyle of the humerus next to the extensor communis but quite distinct
from it. Connected by septa with the extensor carpi ulnaris. The head
of the radius lies immediately entad of the line of separation between
the two muscles. In the distal two thirds of its fleshy extent it is divided
into three distinct slips from which separate parallel tendons run one to
the fourth and two to the fifth digit. On the fourth digit the insertion
is at the base of the ungual phalanx. On the fifth the insertion in one
case is at the base of the ungual phalanx and in the other in the fascia
covering the second and third phalanges. A double tendon from this
muscle to the fifth digit has been noted in Sarcophilus (Macalister, 1870,
p. 164), but not in other marsupials.

Extensor carpi ulnaris. — Origin from the front of the outer condyle
of the humerus and the adjacent side of the ulna; also connected by
septa with the extensor digitorum lateralis. Its tendon passes over a
groove on the antero-internal surface of the extremity of the radius and
inserts at the base of the fifth metacarpal. According to Macalister
(1. c.) and Wilson (1894, p. 451), this muscle has no ulnar origin in
Sarcophilus, Didelphis, Trichosurus, and Dasyurus, but like Canolestes,
has one in Macropus, Phascolomys, Phascolarctos, and Perameks.

Supinator (brevis). — Origin tendinous on the sesamoid laterad of the
head of the radius. Insertion on the arched anterior surface of the prox-
imal third of the radius. A broad flat muscle. It occupies the upper two-
thirds of the radius in Phascolomys, two-fifths in Phascolarctos, one-
third in Macropus and Sarcophilus and only one-fourth in Didelphis and
Dasyurus (Macalister 1. c.). The presence of a sesamoid in its tendon of
origin appears to be unique among marsupials. It is reported in certain
edentates and bats and an origin from the orbicular ligament of the
radius was found in Dasyurus by MacCormick (fide Wilson, 1. c., p. 48).

46 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Extensor indicis. — Origin from the antero-external edge of the
ulna opposite the greater sigmoid cavity. Its tendon divides into two
about midway of its length. These pass under the digital extensors and
across the carpus to insertions one at the base of the ungual phalanx of
the first digit and one at the same point on the second digit.

MUSCLES OF THE FOREFOOT.

The delicate intrinsic muscles of the forefoot have been worked out as
carefully as possible, but in all the specimens examined some of the
metacarpal bones were found to be broken, making it very difficult to
ascertain exact relations. In general the muscles of the hand differ from
those of Didelphis more than they do from some of the Australian forms.
The adductors are much less highly developed than in Didelphis and
this in connection with the absence of a pronator quadratus indicates a
relatively limited power of action of hand and fingers.

Lumbricals are present but not greatly developed.

Adductors. — Three adductors are readily distinguishable, the
indicis, annularis, and minimus. They arise in the palmar fascia at the
base of the metacarpals and radiate to their insertions as practically
separate muscles without any development of a central raphe. The
indicis arises over the bases of the second and third metacarpals and
inserts on the ulnar side of the distal extremity of the second metacarpal.
It is separable with some difficulty from the ulnar head of the flexor
brevis indicis. The annularis and minimus arise over the bases of the
third and fourth metacarpals and insert on the radial sides of the distal
ends of the fourth and fifth metacarpals respectively. They are sep-
arated from the flexors by a large palmar nerve. An adductor pollicis
could not be distinguished. In general the adductors seem to be very
similar to those of Thylacinus as figured by Cunningham.

Flexors. — Short flexors are well developed on all the fingers. The
flexor brevis pollicis could not be divided into two parts. It arises on
the ulnar side of the base of the first metacarpal and inserts medially
with a sesamoid at the first joint of the pollex. It is a relatively thick
muscle and possibly may include a fused adductor. A rather definite
aponeurotic strand running from the carpus to the ulnar side of the first
joint of the pollex may represent a second head of this flexor. The re-
maining flexors are divided except proximally and lie in the usual posi-
tion on the metacarpals with their tendons of insertion embracing the
metacarpo-phalangeal joints of each of the digits.

Abductors and inter ossei. — Two marginal muscles regarded re-
spectively as abductor pollicis and abductor digiti minimi are present.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 47

There is also an abductor of the index digit which is separately de-
scribed owing to its more palmar position than the other interossei.

The abductor pollicis is a short thick muscle with origin from the
trapezium, the transverse ligament and the base of the first metacarpal.
Its insertion is on the radial side of the distal extremity of the first
metacarpal.

The abductor digiti minimi also is a relatively large muscle. It orig-
inates on the pisiform bone opposite the insertion of the flexor carpi
ulnaris and inserts by a flattened tendon running under the annular
ligament to the outer side of the metacarpo-phalangeal joint of the fifth
finger.

The abductor digiti indicis arises on the trapezium and the base of the
first metacarpal and inserts near the distal end and on the radial side of
the second phalanx in close apposition with the short flexor. It is in
nearly the same plane as the flexors and is visible alongside the abductor
pollicis before removal of the large tendons. It is also visible from the
dorsal aspect of the hand.

Excluding the above described abductor indicis, there are three
dorsal interosseus muscles. These lie over the spaces between the four
outer metacarpals, are almost wholly dorsal in position, and resemble
those described and figured for Phalanger (Cuscus) by Cunningham
(loc. cit., p. 23, pi. II, fig. 2) . The one between the index and the medius
has origin principally on the ulnar base and side of the second metacarpal
and insertion with the tendon of the short flexor on the radial side of the
third metacarpal. That between the medius and the annularis arises
from the bases and the sides of the third and fourth metacarpals and
inserts on the tendon of the short flexor on the ulnar side of the third
metacarpal. That between the annularis and minimus arises from the
bases and sides of the fourth and fifth metacarpals and inserts on the
tendon of the short flexor of the ulnar side of the fourth metacarpal, also
sending a short tendinous slip to the radial side of the fifth metacarpal.

MUSCLES OF THE HIND LIMB.
Plates V-VI.

Adductor longus. — Origin from the side of the ramus of the ischium
from the symphysis pubis to the insertion of the semimembranosus.
Its origin is ectad of the adductor magnus and entad of the gracilis. Its
caudal border parallels the semimembranosus. Insertion on the inner
condyle of the femur and the inner side of the head of the tibia.

Adductor magnus. — Origin from the side of the pubic symphysis and
thence dorso-craniad along the edge of the pubis to the middle of the

48 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

base of the marsupial bone and dorsad along the ramus of the ischium
to the semimembranosus. Insertion on most of the length of the caudal
surface of the femur from the insertion of the adductor brevis near the
lesser trochanter to the insertion of the longus at the base of the inner
condyle. This muscle lies between the brevis and the longus and its
exposed surface lies next to the pectineus. At its origin it lies entad of the
longus which parallels it distad.

Adductor brevis. — Origin from the side of the dorsal half of the ascend-
ing ramus of the ischium to the side and front of the tuberosity of the
ischium and for a slight distance craniad. Insertion on the distal caudal
edge of the great trochanter of the femur caudo-entad of the insertion of
the gluteus maximus and thence transversely of the femur to a point
slightly distad of the lesser trochanter. The outer surface of this short
broad muscle is wholly covered by the gracilis, the semimembranosus,
and the other adductors.

Caudofemoralis. — Origin by a broad tendinous aponeurosis from the
fascia covering the multifidae and somewhat attached to the articular
processes of the third caudal vertebra. The insertion is double, one
slender slip inserting by a short tendon on the inner edge of the inner
condyle of the femur between the adductor longus and the gastrocnemius .
The other division of the muscle passes on the opposite side of the
adductor longus and inserts by several fasciculi on the caudal surface of
the distal fourth of the shaft of the femur. Near its insertion it is
slightly connected with the adductor. This muscle, which is not present
in the didelphids, is well developed, but shows some variation even in
the small number of specimens dissected. At its origin it lies immediately
entad of the femorococcygeus and slightly distad is separated from it by
the great sciatic nerve. It passes transversely to the bases of the ad-
ductors, but is free from the femur to its insertion. Its ental surface
near its origin has relation with the caudal muscles, the quadratus
femoris, and the obturator internus as it passes distad just craniad of the
biceps femoris. Its presence is reported in Notoryctes, Phalanger, Thy-
lacinus, Myrmecobius, and Dasyurus.

Pectineus. — Origin from the side of the pubis at the anterior base of
the marsupial bone. Insertion on the inner surface of the femur slightly
proximad of the middle. A short thick muscle lying ectad of the
insertion of the psoas major.

Femoro-coccygeus. — Origin aponeurotic from tendinous fascia be-
tween the cephalic and caudal articular processes of the third caudal
vertebra. Insertion by several fasciculi on the distal third of the caudo-
lateral surface of the femur. It is separate from the gluteus maximus
except near the insertion where the fascia join. It has a slight twist

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 49

in passing between the biceps and the glutens and lies "edgewise"
for a short distance, its distal part being covered by the biceps. It is
separate from the gluteus in Sarcophilus (Macalister, 1870, p. 166) and
Notoryctes (Thompson and Hillier, 1905, p. 310), but in other marsupials
is in the same plane and more or less fused with it.

Vastus internus. — Origin from the proximal three-fourths of the
anterior and dorsal surface of the femur, extending from the anterior
edge of .the great trochanter to the lesser trochanter and thence distad.
Insertion with the vastus externus on the inner side of the patella and
thence by aponeurosis to the head of the tibia. A very large muscle
greatly exceeding the vastus externus.

Subcrureus. — Origin on the dorsal surface of the femur just distad of
the vastus internus. Insertion in the tendinous fascia surrounding the
patella. A short flat muscle.

Vastus externus. — Origin from the anterior border of the great tro-
chanter of the femur. Insertion conjoined with the rectus femoris on
the patella. About equal in size and form to the rectus, with which it
forms the anterior border of the thigh. Decidedly smaller than the
vastus internus to which it is closely appressed but from which it is
distinct except for a short distance near its origin.

Biceps femoris. — Origin by a short tendon from the lateral surface of
the tuberosity of the ischium craniad of the inner division of the semi-
tendinosus and entad of the outer division or crurococcygeus. Insertion
by broad aponeurosis from the patella across the proximal third of the
leg to the outer edge of the tibia. This aponeurosis covers a large part
of the outer side of the leg. A large triangular muscle of great width but
moderate thickness. Its anterior border meets the vastus externus and
covers the insertion of the femorococcygeus.

Rectus femoris. — Origin from the dorso-lateral ridge of the ilium im-
mediately craniad of the acetabulum. Insertion with the vasti on the
patella. It passes between the gluteus minimus and the ileopsoas and
forms the inner anterior boundary of the thigh.

Semitendinous. — Origin by two heads, an outer one ("crurococcyg-
eus") from the fascia over the fourth caudal vertebra and an inner one
from the tuberosity of the ischium just caudad of the biceps. Midway
between the tuberosity and the flexed femur these two slips unite and
then divide into three parts, two passing inside the gastrocnemius and
one outside. The inner slips insert by aponeurosis, one with a twist on
the front of the tibia one-third of the distance from the tibial head and
the other (entad) on the same surface of the tibia halfway between
its fellow and the head of the tibia. The outer slip inserts on the outer
surface of the tibia and in the fascia of the biceps femoris. The relations

50 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

of this muscle are essentially as in Didelphis and different from those of
most other marsupials, in which the so-called crurococcygeus is not so
well developed, although tendinous intersections have been found in
Thylacinus, Dasyurus, and Trichosurus (fide Thompson and Hillier,
1905, p. 316) and in Chceropus (Parsons, 1903, p. 73).

Gracilis. — Origin from the pubic symphysis and the ascending ramus
of the ischium nearly half way to the tuberosity. Anteriorly a few fibers
reach to the base of the marsupial bone. Insertion aponeurotic on the
inner surface of the distal half of the proximal third of the tibia.

Semimembranosus. — Origin fleshy from the outer side of the ascend-
ing ramus of the ischium, occupying the middle two-thirds of the space
between the symphysis and the tuberosity. Insertion by a broad ten-
don on the inner side of the head of the tibia entad of the internal lateral
ligament of the knee joint which is strongly developed. A large thick
muscle elliptical in cross section. No -presemimembranosus was dis-
tinguished, but it may be represented in the somewhat anomalous
muscle here called caudofemoraUs although that has no ischial origin.

Sartorius. — Origin on the dorsal border of Poupart's ligament.
Insertion in the fascia extending from the knee joint and the patella to
the proximal third of the front edge of the tibia. It is a thin band con-
sisting of one layer of muscle fibers only and can scarcely be functional.
This muscular part is clearly differentiated from the fascia on either side
of it which cover a great part of the inside of the leg immediately be-
neath (entad) the mass of fat. Between the sartorius and the next layer
of muscle is the crural nerve and its branches, of which the saphenus
innervates the sartorius and runs down the side of the leg (PI. VI, fig. 2).
It is practically imbedded in the sartorius at this point and is visible
through the thin muscle layer. The insertion of the muscle is ectad of
the gracilis.

This much-reduced, almost rudimentary muscle, is easily over-
looked and it was at first thought that a sartorius was entirely absent or
fused with the gluteus maximus. It was possible also that it might be a
second gracilis. It has been demonstrated in two specimens, however,
and its relations with the saphenus nerve seem to justify the conclusion
that it is in reality the sartorius. Its origin from Poupart's ligament is
paralleled in certain edentates, as sloths and armadillos. A partial
origin from Poupart's ligament has been noted in Sarcophilus by
Macalister (1870, p. 170). In the Insectivora, although usually absent,
the sartorius is found in Gymnura partly attached to the gracilis.

Gastrocnemius. — The medial head arises by a short thick tendon from
the inner transverse surface of the inner condyle of the femur and the
intercondyloid notch just distad of the caudofemoraUs and the adductor

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 51

longus. It lies against the lateral head with various nerves intervening
and is joined with it in the distal fourth of its fleshy extent. It thus agrees
with most marsupials and differs from Didelphis in which it is wholly
free from the lateral head. The lateral head has two divisions, one of
which probably represents the soleus. These are:

1. The principal part arises from the extremity of the fibular sesa-
moid opposite the end of the collateral tibial ligament and joins the
second part about midway of its fleshy extent. Except for a small tri-
angular area along its outer proximal edge where the second part is
visible beneath the fascia, it overlies the second part.

2. The second part arises from the caudal surface of the fabella and
from the tendinous edge of the plantaris for about two-fifths of the
fleshy extent of the plantaris. It joins the outer part about midway of
its fleshy extent. This second and more deep-seated part of the muscle
may be the soleus which is not otherwise evident. The conjoined parts
form a rounded tendon which runs to an insertion on the calcaneum.
As a whole the gastrocnemius forms a deep belly on the leg curving
rather abruptly inward to its tendon and occupying in its fleshy part
only forty per cent of the length of the leg, the remainder being tendinous.
Its outer or anterior edge is continuous with the fascia lata which ex-
tends to the patellar region and down the leg entad of and intimately
associated with the fascia of the biceps fenwris. In one specimen a
slight cleft in this outer edge was observed separating a thin slip which
may represent a third or patellar division of the gastrocnemius.

Plantaris. — Origin from the side of the fabella, mediad of the second
division of the lateral head of the gastrocnemius and more or less united
with it. It continues joined with the gastrocnemius for about two-fifths
of its fleshy extent and then becomes free as a slender bundle which
proceeds distad covered by the gastrocnemius. Its slender tendon, which
at first is entad of that of the gastrocnemius, curves around it and passes
over the calcaneum ectad of the gastrocnemius tendon and widens to
merge into the plantar fascia where its prolongations could not be traced.

Flexor digitorumfibularis. — This has two divisions scarcely separable
in their fleshy part. The smaller one, which is on the tibial side, sends
a small tendon beneath the larger one but separate from it until just
before it passes the os calcis when the two merge. The fleshy part of the
muscle is relatively large, lying next below the gastrocnemius and taking
origin from the greater part of the inner and posterior aspects of the
fibula from the inner side of the outer tuberosity distad not only in the
grooved inner expanded part but also on the proximal part of the rounded
posterior aspect. Proximally it is connected by septa with the tibialis
anticus and distally some of its fibers arise from the interosseous mem-

52 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

brane. Its large tendon passes inside the calcaneum where it becomes
broad and heavy and soon divides into four strong tendons bearing the
lumbricals and running to the ends of the outer toes. Just before the
common tendon divides it receives on its ental surface the two flattened
tendons of the flexor digitorum tibialis between which passes the flexor
accessorius.

Flexor accessorius. — Arises on the tuberosity of the distal outer
extremity of the os calcis, spreads over the ventral surface of the distal
part of the calcaneum, passes beneath the plantar ligament and then
becoming tendinous, continues diagonally entad of the common tendon
of the large flexor fibularis. Thence it perforates the tendon of the
flexor tibialis, or at least is somewhat attached to it, and proceeds to
insertion on the distal phalanx of the hallux. Just before passing
through the annular ligament at the first joint of the hallux its tendon
unites with that of the flexor tibialis.

This muscle is present in Didelphis and has been described by Coues
(1872, p. 134) under the name flexor brevis pollicis obliquus. In Didelphis,
however, it passes ectad of the common tendon of the flexor fibularis
and its tendon joins the tendon of the flexor tibialis before it reaches the
hallux. An accessorius is recorded also for Chironectes (Sidebotham,
1885, p. 16).

The accessorius is stated by Leche (Bronn's Thierr., p. 904) to be
generally lacking in marsupials and perhaps the above described muscle
should be regarded as a part of the flexor digitorum communis brevis
which is otherwise rather weak and slightly represented (seepostea, p. 57).
Whatever its name, it is evident that it is more independent in Ccenolestes
than in other marsupials or most other mammals, especially in the
distinctness of its tendon for a considerable distance on the hallux.
A somewhat similar accessorius found in an insectivore, Gymnura, is
regarded as abnormal by Parsons (1898, p. 324). It is of regular occur-
rence in edentates.

Flexor digitorum tibialis. — Origin from the greater part of the proxi-
mal half of the interosseous membrane, from the popliteus, and by a few
fibers from the inner side of the inner tuberosity of the fibula. Its tendon,
closely associated with that of the tibialis posticus, passes the inner mal-
leolus and soon divides into two, one of which runs directly to the ental
surface of the large tendon of the flexor fibularis and the other joins the
small tendon of the accessorius and after a lateral connection with the
main tendon of the fibular flexor it continues to an insertion on the
terminal phalanx1 of the first digit.

Peroneus brevis. — Origin tendinous from the inner tuberosity of the
fibula and proximally by septa with the muscles lying on either side and

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 53

beneath it. Its large tendon passes beneath the annular ligament out-
side the external malleolus and after passing over the front of the
calcaneum turns downward and inserts on the outer side of the tuberos-
ity of the base of the fifth metacarpal. Just above its insertion it is
distinctly increased in size and is perforated by the tendon of the
peroneus tertius. The same arrangement has been noted in Trichosurus
(Thompson and Hillier, 1905, p. 322).

Tibialis anticus. — Origin from the deeply excavated outer surface of
the proximal part of the tibia. Its muscular part lies mostly in this
broad groove but is directly attached to the bone only in its proximal
third. Its principal tendon passes along the fibular edge of the tibia
and after passing over the astragalus turns down entad of the tendon of
the extensor hallucis longus and widening slightly inserts on the under
side of the entocuneiform bone. A second and smaller tendon runs from
an indistinct division of the fleshy part of the muscle and, paralleling
the larger tendon, inserts in front of it on the side of the base of the first
metatarsal.

In Myrmecobius (Leche, Bronn's Thierr., VI, p. 893) it is joined by
the tendon of the extensor hallucis longus, but in this case that muscle
and its tendon are independently developed. An accessory tendon to
the first metatarsal similar to that in C&nolestes was found in Notoryctes
by Thompson and Hillier.

Peroneus tertius. — Origin from the external condyle of the femur and
the under surface of the collateral ligament jointly with the peroneus
longus, from the outer surface of the inner condyle of the fibula and also
by a distinct slip or second head from the surface of the fibula just
distad of the outer tuberosity. It lies entad of the peroneus longus and
although not so completely bicipital there is a distinct correspondence
between the divisions of each. It has only slight connection with
surrounding muscles, being distinct throughout its fleshy extent.
Its tendon passes through the malleolar ligament, perforates the tendon
of the peroneus brevis and continues directly to its insertion in the fascia
on the outer side of the base of the first phalanx of the fifth digit.

This muscle is regarded by Leche as a division of the extensor brevis
digitorum. Its relations in Didelphis are similar to those in C&nolestes
and also in Thylacinus, Myrmecobius, Cuscus, and in Phascolarctos, but
in these latter it has no femoral origin (Leche).

Peroneus longus. — Arises by two heads which are separate proximally
but which join midway of their fleshy extent to form a single tendon.
The inner head lying next to the peroneus brevis takes origin from the
inner tuberosity of the fibula and from the outer edge of a collateral liga-
ment running from the external condyle of the femur to the inner tuber-

54 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

osity of the fibula. A few fibers also arise from the outer base of the
peroneus brews. The outer head arises from the outer tuberosity of the
fibula and the inner surface of the fibular sesamoid. Its tendon lies deep
to that of the peroneus brevis and passes with it through the annular
ligament over the external malleolus. Distad it passes over theperoneal
tubercle of the calcaneum and down the side of the cuboid deeply im-
bedded in tissue. Turning to the under side of the tarsus, it passes over
the cuboid and after being somewhat enlarged and attached near the
base of the fifth metacarpal, it passes the bases of the third and fourth
metacarpals and goes to insertion on the inner base of the hallux. Its
course and insertion are thus much the same as in Phascolarctos , Pha-
langer and Didelphis.

Tibialis posticus. — A distinct muscle with its fleshy part lying be-
tween the flexor digitorum fibularis and the flexor digitorum tibialis.
Its origin is from the inner side of the outer tuberosity of the fibula,
from the expanded inner surface of the fibula and from septa of the
fibular flexor. Its tendon runs beneath that of the flexor digitorum tibialis
and crosses it midway to the ankle finally inserting on the palmar surface
of the scaphoid.

Extensor brevis digitorum. — This is in three distinct divisions two of
which are of fibular origin while the third is on the dorsum of the
foot. The first is a thin flat muscle arising from a small space on the
outer expanded surface of the fibula slightly mediad and distad of the
outer condyle. Its tendon runs through the outer malleolar ligament and
inserts in the fascia on the outer side of the base of the first phalanx of
the third digit. The second division is slightly larger in its muscular
part which arises subjacent to the first and between it and the extensor
hallucis longus. Its origin occupies the distal two-thirds of the proximal
half of the outer expanded surface of the fibula. The proximal part of the
fibula between the tuberosities is without muscular attachments and
directly in relation with the inner surface of the peroneus tertius which
overlies the above-described divisions of the short extensor. The
tendon of the second division passes the outer malleolus with that of the
first and proceeds over the dorsum of the pes to insertion in the fascia
on the outer side of the base of the first phalanx of the fourth digit.
These two divisions obviously belong to the peroneal group of muscles
and are quite as distinct as the others. They are united in Didelphis and
Dasyurus and send tendons to the second, third and fourth toes; in
Thylacinus and Myrmecobius, they are in three divisions (Leche).

A very slender tendon imbedded in fascia crosses the third meta-
tarsal and connects the tendon of the second division with that of the
third or pedal division. The third division or extensor hallucis brevis

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 55

consists of a short flat muscle much involved in fascia and arising chiefly
from the dorsal surface of the fibular side of the distal annular ligament
and the fascia over the os calcis. Its fleshy fibers scarcely reach beyond
the base of the metatarsals and its tendon runs to the fascia on the
fibular side of the base of the second metatarsal. Fascial connection
with the hallux may include a minute tendon but this was not dis-
tinguished. A similar extensor Inallucis brevis is found in Didelphis, but it
is lacking in Dasyurus, Thylacinus and Myrmecobius. Its presence in
C&nolestes appears to be an exclusive resemblance to Didelphis. It is
common, however, in other mammals, as for example, the Dasypodidce
among edentates.

Extensor longus digitorum. — The digital extensors lie next to the
tibialis anticus and take origin from the outer tuberosity of the tibia,
the inner tuberosity of the fibula, and septa joining them to the peroneus
brevis. Their origin is more largely from the fibula than the tibia. They
are united for no more than the proximal third of their fleshy extent and
in the remainder of their length are distinct both as to the distal fleshy
parts and the tendons. Passing over the ankle they are bound down in
common with the tibialis anticus and the extensor hallucis by a broad
annular ligament (ligamentum cruciatum) and also by a short stout
annular ligament (ligamentum fundiforme) which serves for the passage
only of their four ligaments. The most deep-seated of these extensors is
that lying next to the tibialis anticus and its tendon goes to the terminal
phalanx of the second digit, the next to the third digit and the others
to the fourth and fifth, the tendons passing on the inner side of the second
phalangeal joint in each case and thence to the dorsal surfaces of the
third phalanges.

Extensor hallucis longus. — Origin from the inner edge of the fibula
just distad of the inner tuberosity and thence about one fourth of the
way to the junction of the fibula and the tibia. A few fibers also spring
from the popliteal fascia. Its thin flat muscular part passes beneath the
common digital extensor and the peroneus muscles in a diagonal direc-
tion from the head of the fibula to the base of the tibial flexure and thence
its tendon passes down the tibia beneath the large annular ligament and
inserts on the base of the first phalanx of the hallux.

Popliteus. — Origin from the greater part of the inner surface of the
proximal half of the tibia. Proximally it reaches nearly to the inner
condyle of the tibia while distally it extends almost to the union of tibia
and fibula. Its fibers run diagonally across the interosseous space but
do not reach the fibula except proximally. Distally the fibers reach not
more than half way across the interosseous membrane to which they are
attached as a thin nearly transparent sheet. Proximally the muscle has

56 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

tendinous attachment to the posterior side of the inner tuberosity of the
fibula and the tendon continues broad through the muscle, which is
pinnatifid. An indistinct layer lies superficial to this part of the muscle
and is also pinnatifid, but in reverse direction, the tendon attaching to
the shaft of the tibia and the fibers radiating toward the fibula. This
may represent the so-called pronator tibiae, indications of which are
found also in Myrmecobius and Didelphis (Leche, p. 899), forms in
which there is some possibility of rotatory movements between tibia
and fibula, whereas in Ccenolestes such movements are quite precluded
by the closely appressed distal halves of the bones. The retention of
these muscles in C&nolestes, therefore, is possible evidence that the ar-
rangement of the bones is a recent specialization. This is partly nega-
tived, however, by the entire absence of the pronator quadratus in the
foreleg. There is no trace of a distal pronator tibiae such as is described
by Macalister (1870, p. 173) for Phascolomys, Sarcophilus, Phalangista,
and Perameles.

MUSCLES OF THE HIND FOOT.

The number and relations of the muscles of the hind foot, the so-
called intrinsic muscles, are broadly similar to those of various other
pentadactylous marsupials. The size and independence of the hallucal
adductor and the divided abductor indicis are noteworthy, as well as
the sesamoid origin of certain of the short flexors. Although the other
adductors and abductors are present it seems doubtful whether many of
them can be more than very slightly functional on account of the closely
appressed metatarsals. These are almost in contact proximally for
nearly half their length in such a way as to permit of only very slight
lateral motion.

Lumbricales. — These are fairly well developed, but of uneven size,
arising from the sides of the tendons of the flexor digitorum fibularis
and inserting with them at the ends of the toes. Each tendon has two of
these small muscles. Those on the tendons of the second, third and
fourth digits are progressively shorter and rise from the sides of the
tendons a considerable distance distad of their separation from the com-
mon tendon. Those of the tendon of the fifth digit are very unequal, the
outer being short and rising far distad while the inner is long and rises
as high as the point of separation of the tendon from the common
tendon.

A single lumbrical rises on the medial side of the flexor digitorum
tibialis. Another thin muscle rises on the ventral surface of the common
tendon of the fibular flexor just above its digital divisions and sends
weak aponeurotic connection to the fourth digit and possibly also to

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 57

others but this could not be determined. It represents, at least in part,
the flexor digitorum communis brevis, which is similarly but more highly
developed in Didelphis.

Adductors. — All the adductors except the medius are present. Those
of the outer digits arise in the membranous fascia just proximad of the
metatarsals. The indicts goes to the medial or fibular side of the first
joint of the index digit, the annularis goes to the tibial side of the same
joint of the fourth digit, and the minimi digiti to the tibial side of the
fifth digit. In one specimen the indicis and annularis were joined prox-
imally for at least half their length, the minimi digiti being distinct from
the base of the metatarsus. They are rather flat, wide, and quite delicate,
but have small central tendons.

Hie adductor hallucis lies along the lateral surface of the first meta-
tarsal between the abductor hallucis and the short flexors and is the
largest of the intrinsic hallucal muscles. It is not connected with the
other adductors but has independent origin from the scaphoid by a
^relatively stout tendon which passes along the grooved surface of the
ectocuneiform becoming fleshy at the base of the metatarsal. It inserts
with the abductor on the tibial side of the distal extremity of the first
metatarsal. The adductor hallucis is thus quite different from that of
Didelphis and somewhat similar to that of Phascologale as described by
Cunningham (1882, p. 55), but is even more independent, its origin being
completely separate from that of the other adductors.

Abductors. — These consist of an abductor hallucis, abductor minimi
digiti in two divisions, abductor ossis metatarsi minimi digiti, and four
interosseus muscles.

The abductor hallucis is rather poorly developed and muscular only
in its short fan-shaped proximal part. It arises in the fascia surrounding
the small sesamoid laterad of the base of the first metatarsal and sends
a thin tendon to insert on the radial side of the distal extremity of the
same phalanx.

The abductor ossis metatarsi minimi digiti is well developed arising on
the os calcis and running laterad of the plantar tendon to insert on the
tuberosity of the fifth metatarsal.

The abductor minimi digiti is in two divisions. The first arises as a
thin fan-shaped fleshy muscle from the os calcis ectad of the foregoing
and soon becoming a slender tendon continues by a marginal course to
its insertion on the radial side of the distal extremity of the fifth meta-
carpal. The second division, which is muscular throughout, arises from
the sesamoid at the inner (medial) base of the fifth metacarpal and in-
serts on the outer (radial) sesamoid at the distal extremity of the fifth
metacarpal, passing between the divided short flexors. The external

58 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

plantar nerve passes entad of it near its origin and its general relations
are much as described by Cunningham for Dasyurus, Phascologale and
Thylacinus but it is a relatively larger muscle than in these, rather thin
and flat, and it seems not impossible that it may be an opponens minimi
digiti. The usual four dorsal interossei are present but have a tendency
to fusion with the short flexors. With the exception of the first, they
are entirely palmar in position and cannot be seen from the dorsal aspect
of the foot owing to the nearly united basal metatarsal bones.

The abductor indicis is in two divisions. The larger has tendinous
origin as high up as the scaphoid near the origin of the abductor hallucis
and continues tendinous to the medial base of the first metacarpal
where it has slight attachment and becomes a well-formed fleshy belly
running diagonally to insertion on the side of the first joint of the index
digit. The second and smaller division of this abductor arises on the
medial side of the base of the hallux and runs close to the flexor indicis
slightly dorso-mediad of the first division and joins the tendon of the
first division on its medial side slightly before its insertion. This two-
headed arrangement of the abductor indicis is reported for Thylacinus,
a tetradactylous animal, and a somewhat similar condition is found in
Phalanger which is syndactylous. In Didelphis and Dasyurus the ab-
ductor indicis is single-headed.

The third metacarpal has the usual two abductors inserting on each
side of the first joint on the sesamoids. They are dorsad of the flexors
at their extremities but mostly laterad in the middle where they are
visible without disturbance of the flexors. The outer one is rather closely
associated in its proximal half with the flexor brevis of the fourth meta-
carpal. The fourth interosseous is less dorsal in position than the
others and runs nearly laterad of the flexor from origin in the palmar
cartilage to insertion on the outer side of the end of the fourth meta-
carpal. The interossei insert very close to the extensor tendons but do
not appear to be definitely connected with them. Near the insertions of
the second, third and fourth interossei there is a tendinous connection
with the joint opposite the one on which they are directly inserted and
thus the ends of the metacarpals are bound together and necessarily
have but little freedom of movement.

Short flexors. — There are five pairs of short flexors, one for each
metatarsal. They are only slightly connected at their origin in the
palmar fascia just above the bases of the metatarsals and thence to
their insertions on the distal sesamoids the divisions are clearly marked.
Those of the hallux are weaker than the others and the lateral division is
partly aponeurotic and smaller than the median division. Those of the
second and third digits enclose a small flattened sesamoid in the fascia

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 59

where they arise. This sesamoid may not be homologous with that of
Notoryctes, but it is otherwise unique among marsupials and, as noted
by Thompson and Hillier,1 is suggestive of certain edentates. The
flexors of the fifth digit are separated at their insertion by the tendon of
the shorter division of the abductor minimi digiti. It is somewhat doubt-
ful as to whether the outer or radial division is properly to be regarded
as a flexor, especially since it inserts by a slender tendon whereas its
fellow has a broad fleshy insertion extending to both sesamoids. Pos-
sibly this supposed outer division of the fifth flexor should be regarded
as a third division of the abductor minimi digiti.

SUMMARY OF MYOLOGICAL CHARACTERS.

The trunk muscles of Canolestes are, as was to be expected, of the
prevailing marsupial type. It is in the muscles of the limbs and head that
the most significant deviations are found and although these seem quite
numerous when comparison is made only with American forms, there is
such variation among Australian forms that similarity is found with first
one and then another. It is quite evident that some Australian forms,
especially in the Dasyuridae, approach Didelphis more closely than does
C&nolestes. It is also clear that many of the points in which C&nolestes
differs from Didelphis are to be found in Australian forms. Although
some of these resemblances are plainly cases of convergence, others do
not seem so. Therefore, aside from the question as to the nearest re-
lationships of C&nolestes, it is evident that it contributes in a positive
way to the broad general conclusion, already well grounded, that Ameri-
can and Australian marsupials are of common derivation. That is,
many of the points in which Didelphis fails to show resemblance to
Australian forms are supplied by Ccenolestes.

From the standpoint of myology there seem to be few if any distinc-
tions that can be definitely drawn between Polyprotodontia and
Diprotodontia. Nevertheless, in viewing the possibility that C&nolestes
is to some extent transitional between the two large groups, the muscula-
ture can scarcely be adduced as contrary evidence, since it shares many
features with the diprotodonts that are not possessed by all polypro-
todonts. That it shares more than any polyprotodont cannot be said,

Regarding the sesamoid in Notoryctes, these authors say: "In connection with
the plantar sesamoid bone, which as far as we know has not been recorded in any
other marsupial, it is interesting to find that in some edentates there is a very similar
arrangement. Windle and Parsons state that in the Dasypodidae the tibial and
fibular flexors unite in the lower part of the leg, and are inserted into a very large
sesamoid bone in the sole of the foot, which is held in place by a fibrous band from
the calcaneum, the equivalent of the accessorius. From the front of the sesamoid
bone five tendons pass to the terminal phalanges of the five digits." Jour. Anat. &
Phys., XXXIX, p. 326, 1905.

60 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

although better comparisons than are here possible might prove this
to be the case. It is worthy of particular note that it has a considerable
number of points in common with the extraordinary form Notoryctes
and that, as in Notoryctes, some of its most exclusive features as a mar-
supial are suggestive of edentates and insectivores. Among such are
the divided trapezius, the dorso-cuticularis, the absence of a pronator
quadratus, the sesamoid origins of the supinator and the short flexors
of the hind foot, and the peculiar sartorius. It is also to be noted that
some of the resemblances between C&nolestes and Didelphis not shared
by most other marsupials are however to be found in other mammals,
especially edentates, and therefore cannot be taken as certain indications
of close relationship between C&nolestes and Didelphis. The well-de-
veloped pedal extensor hallucis brews and the flexor accessorius pedis
are perhaps examples of this kind.

The muscles of the legs and feet in Ccsnolestes are adapted to a terres-
trial, almost cursorial, life. The leg muscles have short thick fleshy parts
and very long tendinous extensions, relatively longer even than in such
terrestrial forms as Phascologale, and in this respect are perhaps most
similar to those of the saltatorial but otherwise generalized Australian
polyprotodont Antechinomys , the myology of which has not been
thoroughly described. The proportion of tendinous to fleshy parts is
about as 60 to 40 and the outline of the leg thus resembles that of a
digitigrade animal. It is markedly different from that of Didelphis in
which fleshy extensions reach nearly or quite to the carpus and tarsus.
Specialization for terrestrial life has proceeded far beyond that in
Peramys whose habits are presumably not greatly different from those
of C&nolestes but whose musculature is only slightly different from that
of Didelphis.1 The absence of the pronator quadratus is not accompanied
by such reduction of other muscles as might be expected in view of the
facts that otherwise a large number of muscular elements are present in
the limbs and that their distinctness from each other is marked. There
is some reduction of the intrinsic muscles of the feet, especially in the
adductorial sets, and the grasping power is distinctly limited.

All the important muscles have been worked out with considerable
care, but the nature of the material, the small size of the animal, and the
lack of comparative material, have made it impossible to be certain of
relations in all cases. Such comparative notes as are ventured are based
on examination of the scattered literature and the comprehensive work
of Leche in Bronn's Thierreich. Comparison of actual specimens has
only been possible with Didelphis, Marmosa, and to some extent with

1 No careful study of Peramys has been possible, but cursory examination of its
feet and legs indicates close similarity to Didelphis.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 61

Peramys and Macropus. That this is not an ideal method of com-
parative work can be apparent to no one more fully than the author;
but, it is the best that circumstances permit and if small errors have
crept in it is hoped that they will not seriously affect general con-
clusions. The nomenclature adopted has been that of Bronn's Thier-
reich with occasional changes where it seemed feasible to conform to the
BNA terms.

Features in the myology of C&nolestes which seem especially note-
worthy are as follows:

Trapezius divided into spinotrapezius and acromiotrapezius.

Rhomboideus divided.

Dorso-cuticularis present.

Distinct rectus capitis posterior superficialis,

Pterygoideus externus relatively large and bicipital.

Subclavius not extending to scapular spine.

Median head of triceps brachii distinct.

Origin of supinator brews including a sesamoid.

Pronator quadratus absent.

Distinct caudo-femoralis present.

Femoro-coccygeus practically distinct from glutens maximus.

Glutens medius partially divided into two layers.

Sartorius on inner side of thigh, much reduced, and originating from
Poupart's ligament.

Extensor carpi ulnaris with ulnar as well as humeral origin.

Radial extensors of carpus separate from each other.

Median and lateral heads of gastrocnemius partly united.

Fibular divisions of extensor brew's digitorum distinct from each other.

Pedal extensor hallucis brews well-developed.

Flexor accessories present.

Pronator tibiae partly separate from popliteus.

Short flexors of second and third toes arising from a sesamoid.

URINOGENITAL SYSTEM.

The interesting urinogenital system has been studied only by gross
dissection and the use of the binocular microscope. With the scanty
material available and the means at hand, no other course seemed pos-
sible. It is to be regretted that serial sections could not be obtained of
the embryo. One of the imperfect females examined was pregnant,
having three embryos, two in the right uterus, and one in the left. The
uteri were greatly expanded and the foetuses lay surrounded by their

62 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

thin membranes and the thickened walls of the uteri. The largest was
approximately 4 mm. in length and well advanced in development.
By a most unfortunate accident, these embryos, after being removed and
set aside for special study, were totally destroyed.

Mere gross dissection, however, is sufficient to show that the re-
productive system of Canolestes is markedly different from that of the
primitive polyprotodonts.

MALE GENERATIVE ORGANS.
Plates VIII-X.

The prostate gland is exceedingly large, relatively larger than in
any other marsupial known to me and therefore probably larger than in
any other mammal. It is a prominent ovate body occupying practically
the entire space in the posterior part of the body cavity and extending
from the kidneys to the pubis. In a normal adult with a body length
of 113 mm., it measures 16 mm. in length and n mm. in greatest diam-
eter. The bladder, which is small and inconspicuous in comparison, is
situated on the dorsal surface of the prostate somewhat caudad of its
anterior expansion and thus scarcely visible from the ventral aspect.
The gland is not tapering and radiciform as usual but is only slightly
smaller posteriorly than anteriorly and the narrow urethral canal makes
exit from its walls abruptly on its dorso-caudad surface. The prostate is
enclosed in a thin membranous or slightly muscular covering of trans-
verse fibers. The thickened portion consists of numerous slender cylin-
ders of glandular tissue radiating from the urethral canal to the periph-
ery. The urethra in its course through the prostate is quite small and
is longitudinally plicated on its inner surface (PI. X, Fig. 5).

The ureters and vasa deferentia enter a groove between the base
of the bladder and the prostate, the former going to their orifices on the
dorsal side of the bladder near its neck and the latter turning sharply
caudad into a groove in the prostate to a common opening into the ure-
thra a short distance beyond its enclosure by the prostate (PI. X, Fig. i).
The vasa deferentia have the usual course, leaving the base of the testes
and diverging to enter the body cavity and pass dorsally on each side of
the colon to their terminations. Before entering the raphe in the
prostate they are somewhat enlarged and slightly convoluted. There
are no obvious seminal vesicles.

From its exit from the prostate the urethra continues as a slender
tube for about 5 mm. to its bulbous portion where it enlarges and re-
ceives the ducts of the Cowper's glands. This bulbous urethra is short
and wide and has a slight forward projection or cul de sac on its ventral

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 63

side. There are three pairs of Cowper's glands. The anterior pair are
elongate pear shape and open into the anterior part of the bulbous
urethra ventro-laterally. These have been designated in the figures as
Cowper's gland No. i. The two other pairs are situated dorso-laterally
just anterior to the base of the stalk of the corpus spongiosum and
have a common opening (PL IX, Fig. i). In fact the smaller one (No. 2)
appears to be scarcely more than a division of the larger (No. 3) since
they unite at an appreciable distance from the urethra. This gland
No. 3 is very large, several times larger than the others, even larger
than the paired testicles, and more than half the size of the enormous
prostate. It measures 12 mm. x 8 mm. Although not examined his-
tologically, its superficial appearance indicates a structure somewhat
different from that of gland No. i. Gland No. 2 appears to have the
same structure as No. 3, of which, as noted above, it is scarcely more than
a division. In Marmosa and Didelphis there are only two pairs of Cow-
per's glands, but in Peramys I find three as in Ccenolestes, the posterior
pairs being on a common stalk but showing no such discrepancy in size
as in Ccenolestes. The substance of gland No. i, except near the base,
appears as nearly solid, slightly granular tissue of a whitish color, whereas
No. 2 and No. 3 are highly fenestrated and spongy and even in gross
examination are seen to consist of numerous thin-walled tubules which
converge toward the duct and divide into branches toward the periphery
of the gland. In the glands of Marmosa and Didelphis, the same differ-
ence of structure obtains. In a pouch specimen of Macropus there are
three glands, as in C&nolestes, and the posterior ones have a common
stalk, but their structure seems identical with that of the anterior
pair. Histological study of fresh material is necessary before the
significance of the differences in these glands can be understood.1 That
some of them may function as seminal vesicles is perhaps not im-
possible.

The corpus cavernosum has its origin in a bulb invested with
muscles and situated slightly caudad of the first pair of Cowper's
glands. The stalk of this bulb, which is slightly bilobate, expands into
a layer around the ventral wall of the urethra and continues to the
base of the glans penis. The muscle covering the bulb of the corpus
cavernosum turns on itself and attaches on the stalk of the bulb instead
of continuing to form a levator penis as in the opossum. However, there
are three pairs of small thin muscles lying flat on the base of the corpus
cavernosum and the adjoining part of the bulbous urethra, some one of
which may be the homologue of the levator penis, although this seems

1 Such work may have been done already, but the scope of the present study has
not extended to the literature in which it might be recorded.

64 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

doubtful (PI. IX, Fig. 2). In the opossum the levator penis arises as a
direct continuation of the muscles covering the postero-ventral part of
the lateral bulb of the corpus cavernosum and proceeding inward turns
abruptly caudad and continues to the base of the glans penis. There is
no muscle occupying this position in Ccenolestes and none that could
function as a levator penis. There is another paired muscle in the
opossum, of which the function is not apparent, which corresponds
closely in position to one of the small paired muscles in Ccenolestes, the
most anterior and superficial one. These small muscles have been
designated in the figure as Nos. i, 2 and 3. No. i arises along the median
line on the anterior part of the bulbous urethra and continues as a
slender, free, and slightly thickened fascia to the anterior surface of the
muscle on the bulb of the corpus cavernosum. No. 2 lies partly beneath
No. i and overlaps No. 3, which covers the anterior side of the stalk of
the corpus cavernosum. There is no trace of any such muscles in
Macropus. From the ventral side of the corpus cavernosum near the
extremity of the stalk a large flattened tendon extends to an attachment
on the postero-internal surface of the ischium opposite the origin of the
gracilis muscle (PI. IX, fig. 2 , i. t.) . No such tendon is found in Didelphis,
but a similar attachment is normal in Macropus. A plexus of nerves,
blood vessels, and connective tissue emerges between the pubococcygeus
muscle and the ischium and reaches the corpus cavernosum and adjacent
glandular surfaces, assisting in attaching the very large postpubic mass
to the body.

Behind the bulb of the corpus cavernosum but more dorsal in
position is the bulb of the corpus spongiosum, its stalk meeting the
bulbous urethra slightly below (ventrad) the common opening of the
posterior Cowper's glands. It is also enclosed in a thick layer of muscles
but these do not extend beyond its stalk which expands within the dorsal
wall of the urethra in a very thin layer. This continues to the base of the
glans penis where it rapidly thickens and extends to the end of the glans.
The corpus cavernosum continues, forming the central part of the glans,
practically to the pointed tip.

The glans penis is so deeply cleft as to appear like two separate or-
gans. Each half is of acicular shape, narrowing at the tip to a sharp
point. The extreme tip is separated by a circular fold from the main
body of the glans as a sort of secondary glans. In front of the base of the
glans the urethra forks to empty into a deep groove which runs down the
dorsal surface of each division of the glans, terminating at the circular
fold' which marks off the tip of the organ. This groove is expanded in its
ventral part and occupies the center of the glans. In cross section in a
preserved specimen it appears as a transverse opening connected with

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 65

the dorsal surface of the glans by a vertical slit. In life it would doubtless
function almost as efficiently as a closed tube. The surface of the glans
is slightly rugose in its proximal two-thirds and distally it is covered
with small circular papillae. The total length of the cleft part of the
glans is approximately 1 2 mm. The penis when retracted lies in a pouch
formed by the expanded distal part of the urethra and opening into the
cloaca (See PL IX, Fig. 3).

The retractor muscles of the penis are attached one on each side
near the anterior end of the penis pouch. They originate from the
narrow tendons of the infracoccygeus on the ventral surface of the
fifth lumbar vertebra and run for a short distance caudad above the
colon and then turn slightly ventrad, inclosing the colon and continuing
caudad along its latero-ventrad surf ace, and pass dorsad of the Cowper's
glands to their insertion on the sides of the penis pouch.

The cloaca, the penis pouch, and at least certain of the Cowper's
glands are enveloped in a thin layer of muscle, the sphincter cloacae and
its accessory folds. About the cloaca it is thicker and the fibers have a
circular course, but more anteriorly it becomes thinner and follows the
interstices between the organs where it is lost in the connective tissue.
Anteriorly it is attached by the flat tendons to the ventral surface of the
pubic symphysis. These, in crossing the ventral surface of the urethra,
have partial attachment near the base of the corpus cavernosum.

The testes are of relatively large size, broadly elliptical in shape and
slightly flattened on one side where the two are closely appressed in the
scrotum. A single one in a preserved specimen measures 6.5 x 4.5 mm.
The testis is clasped by the epididymis in the usual manner and loosely
attached. The globus major is somewhat thickened and has consider-
able convexity. It is slightly larger than the globus minor which lies
folded upon itself and in its course toward the exit of the vas deferens
has several convolutions which lie against the testis in the inferior
curvature of the epididymis. Part of this is evidently the vas deferens
itself and its position here is evidently somewhat peculiar since the
arrangement in Didelphis and Macropus is not quite the same.

The most noteworthy features of the male generative organs of
C&nolestes are the extraordinary size of the prostate and Cowper's
glands, the very deeply cleft glans penis, and the absence or great re-
duction of the levator penis muscle. It is difficult to see indications of
relationship in any of these. The bifid glans penis occurs in both
polyprotodonts and diprotodonts. It is most pronounced in polyproto-
donts and, according to Owen (Anat. Vert., Ill, p. 648), is associated
with the presence of a levator penis muscle. But in C&nolestes there is
the unusual condition of a deeply bifid glans penis and a rudimentary

66 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

levator muscle. The divided glans and the well-developed levator muscle
are doubtless both to be regarded as generalized conditions and it may
be concluded that if C&nolestes has retained one it is at least in a secon-
dary stage as regards the other. It is by no means certain that any of
the small paired muscles on the cephalo-ventral surface of the bulbous
urethra are homologous with the levator penis, but if not, then that
muscle, which is so well developed in Didelphis and some other polypro-
todonts, is entirely absent in Ccenolestes as it is in the Macropodidae.

FEMALE ORGANS.

Plate VIII.

The female specimen dissected was a virgin. The ovaries were found
lying entirely exposed on the latero-ventral surface of the colon slightly
clasped by the fimbriated funnel of the oviduct. They are smooth and
disklike, about 2 mm. in diameter. The Fallopian tubes are rather
short and heavy as compared to those of the opossum. The uterine
bodies also are relatively short and less flattened than in Didelphis. They
appear simply as gradual expansions of the Fallopian tubes with only
a slight line of division. They are shorter than the outstretched vaginal
canals and measure approximately 6 mm. in length from the end of the
oviduct to the base of the median vaginae. The general relations of the
peritoneal folds forming the uterine fossa and the broad ligament are
as in other marsupials.

The round ligament is well developed. In gross dissection its rela-
tions to the ovary and the uterus could not be determined positively.
Apparently it does not extend directly to the uterus but is lost in the
thickened part of the broad ligament to which the ovary is attached and
from which a well-defined branch runs to the anterior part of the
uterus. Possibly this branch is a continuation of the round ligament, but
this could not be determined. This round ligament is attached by a
membranous fold which extends in directly transverse relation to the
broad ligament and thence to the dorso-parietal wall of the body cavity.
Its antero-yentral border is rounded and about equal in diameter to the
oviduct. Seen from the ventral side it passes at right angles beneath
(dorsad) the folds of the oviduct and the broad ligament and makes a
sharp turn craniad to join the thickened mass upon which the ovary
rests. It is so distinct and well-developed that it does not seem likely
that it can be anything other than the homologue of the round ligament
of higher mammals. Its occurrence is one of the most noteworthy fea-
tures of the female urinogenital system of Canolestes. I am unable to
detect any trace of a similar ligament in Marmosa. A round ligament,

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 67

apparently of similar character, is recorded by Hill (Proc. Linn. Soc.
N. S. W., XXIV, p. 46, 1899) for Perameles and Macropus.

The bladder in its posterior half and the adjoining part of the uro-
genital sinus have their ventral surfaces connected with the abdominal
wall at the linea alba by a vertical membrane, the "Ligamentum vesicae
medium" of Brass and Hill. The bladder is also attached at the base
laterally by a fold on each side running forward into the broad ligament.
The ureters pass dorsad of the uteri and dorso-laterad of the uterine
neck, enveloped in the broad ligament, and enter the base of the bladder
dorso-laterally, passing between the vaginal cul de sac and the lateral
vaginal canals and parallel to them.

The uterine necks are very closely attached to each other along the
median line but may be separated without destruction of their walls
to the beginning of the median vaginae. The os tincae, however, are
completely coalesced and in the virgin female examined they are of
extraordinary size, extending the full length of the vaginal cul de sac
and having thick, slightly rugose walls. They are attached dorsally and
ventrally by strong septa, forming two distinct lateral vaginal chambers.
In an older female they were found to be relatively much smaller.

The thin-walled vaginal cul de sacs extend nearly to the base of
the bladder on its dorsal surface and are attached to it and to the
urogenital sinus by connective tissue, but without direct communication
in the virgin female specimen examined. In a pregnant and probably
multiparous female a possible connection was observed between the
median vaginae and the urogenital sinus. An irregular opening was
found but some destruction of tissue had occurred before the specimen
was examined, so the point needs verification. If this should prove to
be of regular occurrence it would be an important indication of rela-
tionship to the diprotodonts, for it is essentially characteristic of most
of them.

The lateral vaginal canals leave the median vaginae in a dorsal direc-
tion and turn sharply caudad and then craniad in several convolutions,
nearly to their origin, where they again turn and run directly caudad to
their endings on the lateral walls of the urogenital sinus posterior to the
ureters. The length of one of the canals fully extended is 10 mm. or
two-thirds longer than the uteri. In this respect, therefore, C&nolestes
shows some approach to the condition in Macropus, where the vaginal
canals are always longer than the uteri. The urogenital sinus from the
base of the ureters to the cloaca is 9 mm. long in a young female.

The rectum and the urogenital sinus open into the common cloaca
at about the same level, 3 mm. from the lips of the external orifice. The
clitoris is situated in a pocket of membrane on the dorsal wall of the

68 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

sinus and is a highly developed organ, deeply bifid, and attached only
at its base. It consists of two long, smooth, and sharply pointed rods
the apices of which reach nearly or quite to the cloacal opening of the
urogenital sinus. Each of these divisions is sharply and deeply grooved
on the dorsal surface of its basal half. The entire organ is 3.5 mm. in
length and of this each fork is 3 mm. The form of the clitoris, therefore,
is not very unlike that of the divided glans penis of the male. In an old
female, the clitoris appears to be a single organ, hanging free, and having
several longitudinal plications and a transverse cleft behind its rather
blunt apex.

The cloaca is inclosed in an extensive muscle, the sphincter cloacae.
It surrounds the cloacal orifice and spreads over the large rectal glands,
terminating anteriorly in a transverse plexus of nerves and muscle
fibers which is attached to the ventral surface of the urogenital sinus
and which sends fibers and fascia into the pelvic cavity. The relations
of the rectal glands have been described elsewhere (p. 76).

From 'the comparative standpoint, the female reproductive organs
of Canolestes are more significant than those of the male. They are
distinctly of a diprotodont rather than a polyprotodont type. This is
evidenced at once by the deep median vaginae and the long lateral vag-
inal canals. Many polyprotodonts give birth to their young by way of
the lateral vaginae. In fact this seems th3 only method possible in such
forms as Didelphis which have extremely small median vaginae. Cer-
tain others do not conform to this rule, as Perameles, which Hill (1899)
has shown to have a parturition somewhat less perfected but similar to
that of the Macropodidae, in which the foetus has direct passage from
the median vaginae to the urogenital sinus and thence to the exterior.
In certain other respects the reproductive system of Perameles is re-
garded by Hill as primitive. Data are not available as to the phenomena
of parturition in all diprotodonts, but the direct median passage has
been observed, according to Fletcher (fide Hill, 1899, p. 74), in no less
than twelve species of Macropodidae. Hill found it also in Tarsipes, and,
although it is evident that much study of well prepared serial sections is
necessary for a thorough understanding of the subject, the inference is
very strong that those forms having large median vaginae and relatively
long convoluted lateral canals give birth to the young by median
rather than lateral passage, some by an opening from the median vaginae
to the urogenital sinus which becomes permanent after the first parturi-
tion, and others by a rupture at this point and subsequent closing for
each parturition, the connection being made through a "pseudo-vaginal
canal" in the connective tissue. This method of parturition is at least
functionally an approach to the method in the higher mammals having

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 69

a single vagina, but is regarded as a special development not necessarily
leading to nor derived directly from the single eutherian vagina. That
it should have been independently developed in different phyla seems
highly improbable. Hill's statement (p. 74) in regard to this is as fol-
lows: "That this median passage has not been twice independently
acquired within the Marsupial Class I am convinced, and its existence
in Perameles in a condition so obviously unspecialized and in association
with such a persistently embryonic condition of the genital organs,
tends to suggest that its acquisition is of ancient date." Elsewhere he
says: "Finally, as regards parturition, it seems to me that the foregoing
discussion sufficiently upholds the conclusion that Perameles, in respect
to the phenomena connected with that process, in no way stands alone
amongst Marsupials as an aberrant and specialized type, but quite on
the contrary, exhibits more primitive features in the mode of birth of
the young than are shown by any other Marsupial hitherto described
as possessing a direct median passage. That the direct passage in
Perameles is in a much more primitive condition than that of Macropods,
will, I think, be admitted without question. Indeed, the condition of
the passage in Perameles can only, in my opinion, be regarded as the
precursor of the Macropine one and as showing us in use today the
earliest stage in the evolution of that direct median passage which
reaches its highest development in the specialized Macropodidae."

So far as gross examination indicates, Canolestes is more advanced
than Perameles and, although final proof must await microscopic study
of serial sections, all the present evidence points to the conclusion that
the direct median passage is nearly or quite as well-developed in Coeno-
lestes as it is in various Australian diprotodonts. In a virgin female, the
median cul de sac extends almost to contact with the urogenital sinus
and in an old female, openings were found in the vaginae and the adjacent
surface of the sinus, but the connection between them was imperfect
and possibly due to adventitious circumstances in the history of this
particular specimen.

ALIMENTARY CANAL.

Mouth. — The hard palate has numerous transverse ridges. In-
cluding the one at the posterior edge of the palate, there are nine of
these ridges that extend completely across the palate and five which do
not but merely reach to varying points on either side of the median line
(PI. XVIII, Fig. 3). The ridge between the last molars has 8 to 10
forwardly directed protuberances giving it a somewhat serrate appear-
ance. The ridge between or just behind the canines projects forward in

70 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

a sharp angle. Of the incomplete ridges, the hindmost is scarcely more
than a rounded papule, the next one is posteriorly angled, two others
are straight and another has its ends directed backward. This is the
condition found in a male specimen. A female shows a similar but not
identical arrangement, two of the short ridges being absent and one of
the long ones, that between the third molars, being interrupted in the
middle.

The soft palate and the naso-pharyngeal cavity are relatively long.
The naso-pharyngeal isthmus is guarded on the dorsal side by several
slender filaments directed anteriorly.

Tongue. — The tongue (PI. X, Figs. 3 and 4) is rather slender and
somewhat pointed at the tip. In a male specimen, it measures 22.5 mm.
from the tip to the epiglottis and 4.5 mm. in greatest width. The under
surface is sharply keeled and from the median ridge a longitudinal flap
extends from the basal part of the tongue to the tip. A slight median
raphe is indicated on its upper surface from near the middle almost or
quite to the tip.

The upper surface is densely covered with coronate papillae. No
microscopic preparations were made, but under a strong binocular lens
these coronate papillae appear simpler than those of Didelphis, not
being strongly hooked and including relatively few secondary hairlike
papillae. No differentiation of the anterior from the posterior coronate
papillae was detected (see Poulton, 1883, p. 618).

Fungiform papillae are scattered somewhat irregularly among the
coronate papillae and are rather small and disklike. There is nothing
comparable to the highly developed fringe of fungiform papillae on the
end of the tongue which is found in Didelphis and Marmosa.1

The circumvallate papillae are as usual three in number. They are
about equidistant from each other and the angle formed by the posterior
one is somewhat obtuse. Each has a peculiar retrorse process developed
on its dorso-caudal surface, giving it a form resembling a helmet. Similar
backwardly directed circumvallate papillae were found by Poulton
(1. c., p. 610) in Dasyurus and said by him to be unique in his ob-
servations of the tongues of various marsupials.

Filiform papillae are present in small numbers in the usual position.
No lateral organ could be discerned.

On the whole, the tongue of C&nolestes seems much more similar to

1 This characteristic of these American forms appears not to have been noted
previously, since the specimen (Metachirus) used by Poulton (1. c.) lacked the
extreme tip. The tongues of Didelphis and Marmosa are broad at the tip and thickly
studded with somewhat elongate and pedunculate fungiform papillag, giving the
end of the tongue a fringed or dentate appearance not reported for any other
marsupials.

t
MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 71

that of Australian than American forms. It agrees quite closely with
Poulton's description of the tongue of Dasyurus and that in turn is
said to be of the same type as Halmaturus.

Oesophagus. — The oesophagus from the larynx to the stomach is 30
mm. long. It has membranous attachment to the vertebrae and lies
just dorsad of the aorta until it passes through the diaphragm and al-
most immediately joins the stomach in the middle of the lesser curva-
ture. Its lining has four to six longitudinal plications which continue
into the lumen of the stomach and hang free there. This condition of
the lining of the oesophagus evidently has general similarity to that
reported for most other marsupials except Didelphis. The transverse
folds of the lower oesophagus, which are characteristic of Didelphis* are
not represented in C&nolestes, but they are absent also in Marmosa and
Peramys.1 The lower oesophagus of Marmosa is smooth, however, and
markedly different from that of C&nolestes.

Stomach. — The stomach (PI. VII, Figs. 4 and 5) of a male specimen
in somewhat distended condition measured 24 mm. in longitudinal
diameter and 16.5 mm. in transverse diameter. It is regular in form,
roughly elliptical, with the cardiac expansion only slightly exceeding
the pyloric. In gross examination, the interior shows three marked
divisions which may be called oesophageal, cardiac and pyloric, without
regard to their histological structure. The first or oesophageal division
is the smallest and consists of thin plications or lamellae surrounding
the orifice of the oesophagus and extending along the lesser curvature to
a point nearly midway between the oesophagus and the pylorus. The
cardiac division consists of a strongly differentiated glandular area,
forming a compound gastric gland comparable, at least in certain
respects, to the gastric glands of Phascolarctos, Phascolomys, and Manis.
It is developed principally on the sides of the stomach, being separate
on the two sides toward the pylorus but uniting around the cardiac side
of the oesophagus. In an undistended stomach, it is plainly evident
from the exterior through the rather thin outer muscular walls. From
within it appears as a thick, elevated, and well circumscribed area of
glandular tissue inclosing 40-60 slitlike openings (PI. VII, Fig. 5).
Especially toward the cardiac end of the stomach, these openings have
a distinctly linear arrangement and seem to be the result of an amalga-
mation of the usual rugae of the stomach. In an undistended specimen
there is one prominent fold or ruga which extends three-fourths of
the way to the pylorus and runs about midway between the compound
gland and the greater curvature. The pyloric division of the stomach

1 This is a rather important distinction between Didelphis and Marmosa which,
so far as I am aware, has not been noted previously.

72 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

is relatively smooth and undiff erentiated and is continuous with the area
along the greater curvature to the fundus.

The- differentiation of areas of the stomach in Caznolesies is of course
a specialization and as such is referred to in discussion elsewhere, but
in view of the wide range of the character of the stomach in marsupials,
it is difficult to draw any direct inference of relationship. In fact, the
stomach of C&nolestes is unique and serves to strengthen the general
conclusion that the animal stands by itself quite as independently as
any of the highly specialized Australian forms.

Intestines. — The intestines lie in the abdominal cavity with the main
mass of the small intestines, or Meckel's tract, in ventral position. The
duodenal loop is well differentiated. On leaving the stomach (from
which it is separated by the pyloric sphincter, showing externally as a
distinct band 2 mm. in width), it passes slightly dorsad, then caudad and
turning sharply on itself runs cephalad a short distance ventrad of and
almost in contact with the colon, and thence it passes ventrad to the
center of the mass of Meckel's tract. Throughout this course it appears
to have slightly thicker walls than the distal part of the small intestine
and in an alcoholic specimen is lighter colored. The diameter of the
collapsed duodenum just beyond the stomach is nearly 9 mm. The
duodenum is weakly attached to the dorsal mesentery near the end of
its loop and very close to the colon. The folds of the intestines are
simple and on a continuous mesentery but cannot be straightened in
a preserved specimen without cutting. They seem to be in the condition
figured as stage II by Beddard (1908, p. 592, Fig. 121).

The large intestine is relatively very short and straight and has no
indication of loops. It is scarcely more than 30 mm. in length as com-
pared to 340 mm. for the small intestine from the pylorus to the caecum.
The hindgut is imperfectly divisible into colic and rectal parts. The
colon has the usual longitudinal folds on its inner surface but at its
termination, that is at the beginning of the rectal part for a distance of
about 4 mm., both the large primary folds and the interspaces are crossed
by numerous delicate transverse ridges which in relaxed condition appear
as undulating transverse lamellae. Externally the rectum is covered
with longitudinal muscle fibers, somewhat more numerous dorsally
than ventrally.

The caecum is very small, scarcely more than a vermiform appendix
(PI. VII, Fig. 3). In one specimen it is only 4.5 mm. in length; in another
it is 5.5. It opens into the intestine by a simple aperture and its inner
surface does not differ from that of the ileum. Immediately distad of its
opening the colon begins with a muscular, plicate inner surface in con-
trast to the villous surface of the ileum.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 73

The short colon combined with the small caecum in Ccenolestes
is contrary to the usual condition in marsupials and is most nearly
met in certain of the dasyures. Although the short colon is perhaps
primitive, the small caecum is obviously a secondary condition.

GLANDS.

Liver. — The liver is relatively large in size. It has three primary
divisions, the left lateral, the right central and the right lateral, the last
being subdivided into right lateral proper, caudate and spighelian
(PI. VII, Fig. 2). The largest lobe is the left lateral, although the right
central or cystic lobe almost equals it and the combined bulk of the three
subdivisions of the right lateral is nearly or quite as great as either of the
other primary divisions. In a male specimen hardened by preservative,
the right central lobe is 18 mm. wide; the left lateral has practically the
same width but is slightly more bulky.

The large left lateral is entire without trace of a division, although a
slight projection of its dorsal part is somewhat separated to curve around
the oesophagus. In Didelphis and Marmosa there is at least one cleft
in its lateral boundary partially dividing it into left lateral and left
central.

The right central is entire except for one fissure extending about
halfway to its base and inclosing the gall bladder in a wholly concealed
position similar to that in some other marsupials but unlike that of
Didelfhis in which it is exposed and in contact with the diaphragm.

The right lateral proper lies next to the right central and is regular in
shape although its dorsal border has a slight cleft. It is about half as
large as the right central and its dorso-caudal border is in contact with
the kidney. Although for the most part separate from it by a deep fis-
sure, it is united at its base to the caudate.

The caudate is in ventral position and incloses a considerable part
of the right kidney. It has an anterior projection and two posterior ones
through the inner of which the vena cava passes.

The spighelian lobe arises from the base of the caudate and divides
into two lobes almost at right angles to each other, one appressed to the
left lateral and the other hanging free caudad of it and ventrad of the
caudate lobe. Its peduncle attaches it to the caudate but the portal
vein and the ductus choledochus pass between the free surfaces of the
caudate and the spighelian. The pancreatic duct and the ductus choledo-
chus unite just before entering the duodenum.

The liver of Canolestes differs from that of Didelphis and at least
some other polyprotodonts in its large bipartite spighelian lobe, in its

74 FIELD MUSEUM OP NATURAL HISTORY — ZOOLOGY, VOL. XIV.

small concealed gall bladder, and in its complete fusion of left lateral and
left central lobes. In Dtdelphis, the spighelian lobe is quite small and
undivided. In Antechinomys (Beddard, 1908, p. 563) the spighelian is
entirely absent and left lateral and left central are completely divided.
A spighelian lobe is present in most of the Macropodidae and in one
(Dendrolagus) , at least, it is bilobate (Beddard, 1895, p. 134).

Spleen. — The spleen is suspended from the omentum and the dorsal
mesentery and lies closely appressed to the dorso-caudad surface of the
stomach. It is an elongate organ with a single triangular lobe directed
dorsad from the right hand third of the dorsal border. The spleen thus
has a general shape resembling a figure four or a bisected head of a blunt-
pointed arrow (PL VII, Fig. i).

Pancreas. — The pancreas is an extensive, much divided gland lying
ventrad of the spleen and suspended between it and the intestines over
the surface of which some of its smaller digitations run. In common
with the spleen it has some attachment to the omentum and dorsally
it is connected with the main dorsal mesentery. One long branch of it
lies between the principal longitudinal folds of the duodenum. The
main pancreatic duct, joined with the common bile duct, enters the
duodenum some 10 mm. below the pylorus (PL VII, Fig. 2). The two
ducts meet nearly at right angles a very short distance from or practically
at the surface of the duodenum. Thence through the membranes of the
duodenum there is an appreciable common duct carrying both biliary
and pancreatic products to the common opening on the inner surface of
the duodenum. At a point nearly opposite the union of the main ducts,
the pancreatic duct receives an important branch or accessory duct
derived from that lobe of the pancreas which lies along the duodenal
fold.

Salivary glands. — The parotid gland (PL IV, Fig. 2) is extensive
and invests the base of the external ear except in its dorsal third. It
covers a part of the trapezius muscle, the sternomastoid and all the
superficial muscles of the side of the neck. Three lobules are distinguish-
able, one anterior and two posterior, the dorsal of these latter being the
larger. In a male specimen, the length of the gland is 16 mm. and its
greatest dorso-ventral depth 10 mm. Its duct leads forward crossing the
masseter superficially and then, sinking slightly, opens into the mouth
near the first upper molar tooth.

The submaxillary gland also is large and is intimately associated
with the parotid. It lies in ventro-lateral position relative to the parotid
and has a detached lobe lying between its principal longitudinal lobe
and the parotid. Its duct was not traced with certainty.

A large sublingual gland lies along the inside of the mandibular

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 75

ramus. Posteriorly it lies between the mylohyoid and the styloglossus
muscles and anteriorly it is between the mandible and the genioglossus.
It extends from the vicinity of the base of the angular process to the
symphysis. Its ducts were not sought but it is evidently independent of
the other large salivary glands.

Thyroid. — In a male specimen, the thyroid was found in two elongate
pisiform divisions (PL VI, Fig. i) lying on the lateral surface of the
trachea. The larger anterior one extended from the vicinity of the
thyroid cartilage caudad about 4 mm. The smaller lay just caudad
in the same longitudinal plane and was about one-third the size of its
fellow. In a female specimen, only one thyroid was detected.

Thymus. — All the specimens examined were fully adult and show only
traces of the thoracic thymus. No indication could be found of the
presence of the superficial cervical thymus regarded as characteristic
of the Diprotodontia by Symington (1898) and Johnstone (1901). Four
distinct fatty lobes lying ventrad of the carotids and cephalo-ventrad of
the base of the heart were noted in one specimen as probably represent-
ing the thymus.

Mammae. — There are four mammae, regularly arranged, two on
each side of the abdomen inside the thighs. This is the number present
in the majority of diprotodonts and less than the number usually found
in the polyprotodonts. In comparison with the Didelphudae, however,
the number is perhaps less important than the regularity of arrange-
ment, for members of that family are distinguished from all other mar-
supials by having the mammae present only in odd numbers and more
or less asymmetrically arranged. The number in diprotodonts never
exceeds four; in polyprotodonts it is usually more than six, although
Thylacinus, Myrmecobius and a few species of Phascologale have only
four, while the greatly specialized Notary ctes has but two. The following
are the number found in the principal marsupial genera :

POLYPROTODONTIA:
Peragale

8

DIPROTODONTIA:
Macropus

Perameles

. 6- 8

Acrobates

Chasropus

8

Distoechurus

Thylacinus

4.

Dromicia

Sarcophilus

4.

Petaurus

Dasyurus

. 6- 8

Phascologale

A— IO

Phalanger

Sminthopsis . • . . .

. 8-10

Myrmecobius

4.

Phascolomys

Notoryctes

2

Didelphis

S-I3

Metachirus

7- Q

Lutreolina

0

Philander

7

Marmosa

9-1 S

Peramvs

9-27

76 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Miscellaneous glands. — The smaller sebaceous and lymphatic glands
were not systematically sought but a few have been noted in the course
of dissecting other parts.

Three rather conspicuous axillary glands are found imbedded in a
fatty mass lying laterad of the cutaneus maximus and partially enclosing
the elbow. The largest of these is long ovate pyriform and lies dorso-
laterad of the cutaneous muscle, projecting also over the latissimus. It
is about 4 mm. long and 2 mm. wide. Two others, slightly smaller,
lie together ventro-laterad of the cutaneus maximus.

Two large superficial inguinal lymphatic glands in the male are situat-
ed in the deposit of fat ectad of the cremaster muscle. In one specimen,
the larger of these glands measured 7 mm. in length and the other 5 mm.

A rather prominent precardial lymphatic gland is situated in the
acute angle between the base of the oesophagus and the fundus of the
stomach (PI. VII, Fig. 4). In one male, this gland measured 4 mm. in
length, being approximately equal to the diameter of the oesophagus.

An ovoid lymphatic gland about 3 mm. in greatest diameter is
situated in the angle between the sternomastoid, omohyoid, and di-
gastric muscles (PI. IV, Fig. i). It lies close to the trachea and is sep-
arated from it only by the carotid blood vessels.

In the male, a large paired lymphatic gland (PI. VIII, Fig. 2) lies
laterad of the large Cowper's gland (No. 3). It is about 6 mm. long by
3.5 mm. in width. Its wider end rests in the angle at the base of the penis
pouch between the base of the penis and the bulb of the corpus spongi-
osum. Its smaller end lies in connective tissue and fat on the side of the
large Cowper's gland. It is not directly connected with the rectum
and has no duct that is apparent.

In the female, there is a large ovate pyriform rectal gland lying on
each side of the cloaca with its small end directed dorsally and a duct
leading from near its large end to the ventro-lateral wall of the cloaca
slightly caudad of the opening of the rectum. The gland is 4 mm. long
and has a tough thickened wall somewhat differentiated from the
glandular mass within.

RESPIRATORY SYSTEM.

Plate X, Fig. 2.

Lungs-r- The left lung is entire. In two specimens, the right lung
has two lobes besides the azygos and the anterior of these is slightly
notched at a point corresponding to that where a conplete division is
found in the d'delphids. In a third specimen, the right lung was found
to be trilobate although the anterior division was relatively small. The

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — 'OscooD. 77

evidence is therefore inconclusive as to whether the normal arrange-
ment in Canolestes is bilobate or trilobate. In most marsupials, except
the wombat and some of the Macropodidae, it is said to be trilobate.
In several specimens which I have examined of Didelphis, Marmosa,
and Peramys it is trilobate. In C&nolestes, there is at least a tendency
for it to be bilobate.

Trachea. — The tracheal rings are about nineteen in number and are
not united dorsally but separated by a broad interspace.

Larynx. — The epiglottis is large and elevated, and has a slight
anterior emargination which is most evident in the cartilage after
removal of the soft external membranes. It is attached to the thyroid
cartilage ventrally. The thyroid is continuous medio-ventrally with
the cricoid and its bony wings partially inclose the cricoid laterally.
Its anterior or superior horn has ligamentous attachment to the thyro-
hyal; its posterior or inferior horn attaches similarly to the outer
extremity of the cricoid which is here very slightly ossified. The aryte-
noids are large, subtriangular and partially united dorsally. The
cartilage of Santorini is developed into a prominent curved protuberance
which is continuous with the high muscular crest of the arytenoid
which, meeting its fellow of the opposite side, nearly closes the cavity
of the larynx with the exception of a very small passage on the ventral
side.

CIRCULATORY SYSTEM.

Owing to the condition of the material, no attempt was made to
study the circulatory system in detail. The heart is as usual in mar-
supials. In comparison with that of Didelphis it shows no very important
characters. In ventral view the auricles meet over the base of the
pulmonary artery, entirely concealing it, whereas in Didelphis it is
broadly exposed. The bases of the right subclavian and right carotid
are very closely approximated and both join the left carotid scarcely
higher than its junction with the aorta. In Didelphis there is an appre-
ciable space between the origins of the right subclavian and right
carotid.

SKELETON.

Plates XI-XVI.

CERVICAL VERTEBRAE.
Plate XI.

The seven cervical vertebrae are freely movable and distinct. As
in the majority of marsupials, the only one having a prominent spinous
process is the axis. The neural arches have relatively thin, flattened

78 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

laminae which are weak as compared to the bodies. They are well
separated from each other and have no tendency to imbrication. All
except the first and seventh have vertebrarterial foramina. Only the
first and second have hypapophyses.

The atlas is practically entire, having a body which evidently
ossifies in normal manner and having no large foramina for the passage
of nerves and blood vessels. A very minute perforation (PI. XI, fig. i,
n. /.) at the posterior dorsal base of the transverse process is of doubtful
significance, but its position does not seem to indicate homology with
either of the usual foramina of the atlas. A perforation with exactly
similar relations is found in the extinct Borhyaena (Sinclair, 1906, p. 350,
PI. LIII). The cephalic border of the neural arch of the atlas is nearly
straight for about the width of the neural canal and thence on either
side it is angled latero-caudad to meet the base of the cephalic articular
process. A groove between this angle of the border of the arch and the
articular process serves for the passage of the first cranial nerve and
supplies the office of the foramen found in most other marsupials.
That this foramen is formed by the closing of the points bounding the
mouth of this notch is shown by immature examples of Didelphis in
which the union is slightly cartilaginous. In the adult of Didelphis, the
foramen is well isolated as it is in Philander, Dasyurus, Myrmecobius,
Thylacinus, Sarcophilus, Phascologale, Trichosurus, Petaurus, Petrogale,
Macropus, Phascolomys, and Phascolarctos. The only living marsupial
genera agreeing with Canolestes in this open "intervertebral notch" are
Marmosa, Peramys, and Perameles. The atlas of Perameles is otherwise
similar to that of C&nolestes, but that in Marmosa and Peramys resembles
it still more closely, differing chiefly in the pedunculation of the trans-
verse process and its slight obliquity and in the tendency to the forma-
tion of a vertebrarterial canal with a partly cartilaginous border.

The transverse process of the atlas has decided obliquity to the spinal
axis and is relatively short and broad with no constriction at its base and
with only a shallow groove between it and the posterior zygapophysis
in the position corresponding to the usual course of the vertebrarterial
canal. The only genera examined in which the arterial canal has com-
plete bony boundaries are Philander, Dasyurus, Sarcophilus, Myr-
mecobius, and Petaurus. In all others except Canolestes, however, the
transverse process is pedunculate and in some, as Phascologale and
Peramys (Marmosa and Didelphis to a lesser extent), there is a slight
bony elevation on the pedicel of the arch cephalad or laterad of the
postzygapophysis from Which a cartilage extends to the posterior angle
of the transverse process and forms the arterial canal. So far as material
at hand shows, the gradual formation of the vertebrarterial canal ofjthe

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 79

atlas is illustrated in various stages from the wide open channel to the
heavy bony ring in a series of forms the sequence of which would be
about as follows: C&nolestes, Perameles, Didelphis, Phascologale,
Peramys, Dasyurus, Philander, Petaurus, Myrmecobius. In this series,
Didelphis is about midway between the extremes and various other
forms, including the macropods and phalangers, would occupy a similar
position.

The atlas in Ccenolestes has no trace of a spinous process. The body
of the atlas is directed downward and slightly forward, being almost
vertical and nearly at right angles to the laminae of the arch. Its lower
part is thus a thin triangular plate with a hypapophysis or median
ridge on its posterior aspect and forming a pointed tubercle at its apex.
The odontoid process, therefore, does not have its usual flattened seat
but rests on the sharp edge of this nearly vertical body of the atlas. This
part of the atlas differs from that of practically all other marsupials,
although it is closely approached in Phascologale, Marmosa, and Peramys.
The neural canal of the atlas is slightly wider than deep and measures
4 mm. x 3.4 mm.

The imperforate condition of the atlas in Ccenolestes, both anteriorly
and posteriorly, is doubtless primitive. It is found in the monotremes
and among living marsupials examined only in Marmosa, Peramys, and
Perameles.1 It is characteristic also of the extinct Borhyaena but, so
far as known, of none of the other forms of the Santa Cruz beds. The
atlas, however, is unknown in all the fossil forms placed by Sinclair
(1906) in the family C&nolestidae. In other mammals, including all the
higher groups, no instances have been found in which both the atlanteal
and vertebrarterial foramina are absent.

The axis has a high laminate spinous process similar in general to
that of the majority of marsupials. The anterior wing of this process
is not equal to the posterior and reaches only to the posterior part of the
dorsal surface of the atlas. The posterior wing overhangs the third
cervical. The front border of the neural arch seen from the side is a
continuous curve, and the prominent dentation found in Didelphis,
Peramys, Marmosa, Philander, Trichosurus, Petaurus, and some others
is absent. At the base of this curve is the prezygapophysis which is an
oval facet only very slightly elevated from the pedicel of the arch. The
odontoid process is completely ankylosed in the adult and forms a
simple projecting knob of the usual character. The body of the vertebra
is a fairly thick shield-shaped piece with two relatively large foramina
in the floor of the neural canal. The hypapophysis is peculiar. It^is

lln Notoryctes the atlas is imperforate, but this is obviously due to extreme
specialization and has no significance in this connection. (See Stirling, 1891.)

8o FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

divided into two parts one of which probably appertains to the odontoid
and the other is the usual hypapophysis of the axis. The anterior of
these is a sharp hooked process rising from the base of the odontoid and
projecting backward along the median line. The posterior process is a
ridge rising beneath this hook and gradually becoming more elevated to
terminate in a tubercle which has slight projection caudad beyond the
centrum and thus to some extent effects imbrication with the front
edge of the body of the third vertebra. InMarmosa there is a prominent
axial hypapophysis and in the skeleton of a half -grown Didelpkis with
the odontoid still distinct there is a small median knob but in adults I
find nothing in any marsupial comparable to the hooked process
described above. It very strongly suggests the large recurved hypapo-
physes of Ornithorhynchus. The posterior border of the neural arch in
lateral view presents an even curve from the spine to a prominent angle
above the postzygapophysis. The neural canal has the outline of a
horseshoe and its width and depth are nearly equal.

The transverse process of the axis has a superior and an inferior
division, each a small backwardly directed projection from the caudal
aspect of the pedicel of the arch, ventrad of the postzygapophysis and
caudad of the prezygapophysis. The extremities of these processes are
articulated to the slightly produced basal angles of a small, distinct,
subtriangular pleurapophysis or cervical rib which serves to complete
the boundary of the arterial canal and which is obviously the homologue
of the fully ankylosed processes serving the same purpose in the suc-
ceeding vertebrae. In a fully adult animal this cervical rib has only
cartilaginous union with the vertebra and in the undesiccated specimen
it is freely movable (PI. XI, fig. 7). It is essentially similar to the
persistently distinct cervical ribs in Ornithorhynchus and in reptiles.
It has not been reported among marsupials or higher groups except in
Perameles and Phascologale. In Perameles lagotis, it is noted by Owen
(Trans. Zool. Soc. Lond., II, p. 394, 1841) in a single instance. The only
skeleton of Perameles (P. nasuta) I have been able to examine, has the
axial pleurapophysis fully ankylosed. Whether it is normally so in this
genus, therefore, is uncertain; but its persistence as a distinct part in
C&nolestes is none the less significant. Beddard (1902, p. 119) mentions
a separate axial rib for Phascologale.

The third, fourth and fifth cervicals are not peculiar and resemble
each other in general characters. In the third the neural canal is rela-
tively deeper, its vertical and transverse diameters being nearly equal,
whereas in the fourth and fifth the width considerably exceeds the
depth. In the third the spinous process is very slight; in the fourth and
fifth it is a low but distinct mammillate process. In all three the inferior

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 81

branch of the transverse process is only slightly developed. The superior
branch of the transverse process is directed obliquely backward and
upward and at its base is pierced by the arterial canal. In the fifth it is
slightly thicker than in the third and fourth, has less backward deflection
and is definitely faceted at the tip.

The sixth cervical is similar in general features to that in the majority
of marsupials. The arterial canal perforates the base of the transverse
process as usual but it is to be noted that the dorsal wall of the canal is
very much thicker than the ventral. The inferior branch of the trans-
verse process is expanded into a dolabriform plate practically parallel
with the spinal axis. This has only very slight forward projection but is
extended caudad nearly or quite to the plane of the centrum of the
first thoracic vertebra.

The seventh cervical has a slightly higher spinous process than the
three preceding vertebrae, but the distinction is not marked. It has the
superior branch (diapophysis) of the transverse process with its tubercle
directed almost at right angles to the spinal axis in the horizontal plane
of the body of the vertebra. The inferior branch (parapophysis) and
consequently the arterial canal are obsolescent, that is, they are variable,
being present in some specimens and reduced or absent in others. In
one specimen, the arterial canal on the left side of the vertebra is
enclosed by the merest bony thread representing the parapophysis and
having no tubercle. On the right side of the same vertebra, the condi-
tion is similar but the canal is exceedingly small, scarcely one-tenth as
large as that on the left side. In several other specimens, the seventh
cervical is quite imperforate, the vertebral side of the base of the trans-
verse process being merely grooved (PL XI, figs. 4-6). In a single
specimen of Phascologale, the lower boundary of the arterial canal is
incomplete, indicating a condition very similar to that in Canolestes; in
a single specimen of Sarcophilus, one side is perforate and the other
imperforate. The seventh cervical is imperforate in Trichosurus,
Dasyurus, Perameles, Sarcophilus, Didelphis, Marmosa, and Peramys.
The arterial canal has complete bony boundaries in Thylacinus, Petaurus,
Philander, Phascolomys, Phascolarctos, and the Macropodidae. It is
complete also in the extinct genera Borhyaena and Cladosictis. Thus
in the majority of diprotodonts the seventh cervical is perforated and
in the majority of polyprotodonts it is imperforate, although the
perforate condition is more primitive. In this as in other respects,
therefore, C&nolestes combines the characters of the two groups.

None of the cervicals have foramina piercing the lateral wlls of the
neural arches like those in some other marsupials (Thylacinus, Dasyurus,
Didelphis, etc.).

82 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Considered as a series, the cervical vertebrae of Ccenolestes seem more
similar to those of Perameks than of any other marsupial. The imper-
forate atlas, the persistent cervical rib attached to the axis, the evenly
curved front edge of the neural canal of the axis, and the imperforate
seventh vertebra all are found in Perameles. The inferior lamellae of the
transverse processes of the third, fourth, and fifth cervicals are much
larger in Perameles than in Ccenolestes and this constitutes the principal
distinction between them. In comparison with those of other American
forms, the cervicals of Ccenolestes show a general resemblance to those of
Marmosa and Peramys and, among Australian small polyprotodonts, to
those of Phascologale , but the characters of the atlas and axis are dis-
tinctive. With the highly modified cervicals of Didelphis, comparison is

scarcely- necessary.

»

THORACIC VERTEBRAE.

There are thirteen thoracic vertebrae as in all other marsupials with
the exception of Phascolarctos (n), certain species of Phascolomys (15),
and Notoryctes (i$).1 The spine of the first thoracic is scarcely higher
than that of the last cervical (PL XII, fig. 4). The second spine is
slender and much the highest of the series, measuring about 3.5 mm.
The third to the eighth also are slender and progressively shorter and
more recumbent. The ninth is still shorter, and although directed
backward, it is longer basally and less acicular in form. The first
thoracic differs from the last cervical mainly in the elevation of the
transverse process and in its differentiation for articulation with the
rib. In most marsupials there is an abrupt transition from the last
cervical to the first thoracic, especially in the length of the spinous
process, but in Ccenolestes this break is between the first and second
thoracics. The only form examined which agrees with it in this respect
is Perameles nasuta. In Phascolarctos, the spine of the first thoracic only
slightly exceeds that of the last cervical but it is nevertheless the longest
of the thoracic spines and there is nowhere any marked break in the
series. Owen (1841) in his paper on the osteology of marsupials does not
mention a relatively short spine for the first thoracic of Perameles and
it is therefore possible that it is not found in all species of the genus.

The tenth is the anticlinal vertebra and its spinous process is a low
upright plate with a straight crest occupying practically the entire length

1 In a single specimen of Didelphis paraguayensis from Ceara, Brazil, there are
fourteen ribs instead of the thirteen normal in Didelphis. Moreover, there are
twenty vertebrae in the thoracico-lumbar series, the extra vertebra apparently
being the fourteenth, corresponding to the extra rib. Thfe is probably an abnor-
mality, but the examination of other specimens of this species is desirable, especially
since, as noted elsewhere (p. 86), it has other skeletal peculiarities. A pouch speci-
men of D. p. andinus has the normal thirteen ribs.

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 83

of the dorsal surface of the arch. The eleventh, twelfth and thirteenth
spines are similar, but although only slightly elevated, they are definitely
directed forward instead of backward. The thirteenth is slightly lower
than the eleventh and twelfth and closely resembles the spines of the
succeeding lumbars.

The position of the so-called center of motion sharply distinguishes
Ccenokstes from all living American polyprotodonts and aligns it with
the majority of Australian forms, as well as with most higher mammals.
In Didelphis, Marmosa, Philander, and Peramys there is practically no
anticline since the spines of all the thoracic vertebrae and of all the
lumbars, except the last or at most the last two, are inclined backward.
The only other marsupials having this characteristic are the koala, the
wombats, and certain of the phalangers. In all the Australian poly-
protodonts and at least in the macropods of the diprotodonts the center
of motion is in the posterior thoracic vertebrae and in most of these
forms the anticline is very marked.1 This is also the case in C&nolestes
and in the Patagonian fossil form Cladosictis. Whether the condition in
the didelphids is palaeotelic or not is problematical. Essentially the
same condition is found in the monotremes and edentates, and it is
usual among reptiles, whereas among higher mammals it is rare and
generally correlated, perhaps secondarily, with high specialization, as
in seals and anthropoids. In practically all insectivores, rodents, bats,
carnivores, ungulates, and in Primates, except the higher apes and man,
there is a marked convergence of vertebral spines to a center of motion.
In agreeing with the majority of mammals in this respect, therefore,
C&nolestes at least shows a lack of affinity with the didelphids.

The first nine thoracic vertebrae are essentially similar in form.
The arches and pedicels, although progressively increasing in length,
are relatively narrow and each vertebra is separated from the succeeding
one dorso-medially by a considerable space between the projecting
zygapophyses. Each has a transverse process bearing a facet for the
tubercle of a rib. The tenth begins to assume the character of the
lumbars. It has only a trace of a facet for the tubercle of a rib, its spinous
process is low and ridgelike instead of spinous, its laminae and pedicel
are broad and its body is less compressed. The eleventh has these
characters further developed. There is no trace of a lateral rib attach-
ment and posteriorly there is a tiny pointed accessory process below the
postzygapophysis. In the twelfth and thirteenth this accessory process
is slightly larger and in these the prezygapophyses begin to turn upward,
but there is only a slight rudimentary mammillary process.

1 In the single skeleton of Sarcophtlus examined there is a double anticline, one
at the eleventh thoracic and another at the third lumbar.

84 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

In the character of the first thoracic and essentially in other respects,
the thoracic series of vertebrae, like the cervical, resembles that of
Perameles more closely than that of any other marsupial. It differs
chiefly in the last three thoracics which in Perameles have relatively high
spines and prominent mammillary processes.

LUMBAR VERTEBRAE.
Plate XIII, Figs. 1-2.

The six lumbars have no important peculiarities. The first three
increase slightly in size caudad but aside from the appearance of small
inferior anterior transverse processes they scarcely differ from the last
thoracic. There are still small accessory processes (anapophyses), the
spinous processes are low and ridgelike and the mammillary processes
(metapophyses) scarcely protrude beyond the articular surfaces of the
zygapophyses. The fourth lumbar retains the ridgelike spinous process,
but the transverse process, directed downward and forward, is con-
siderably increased in size. The fifth and sixth lumbars are somewhat
differentiated from the others. The transverse processes are longer, the
spinous processes, which are directed forward, are higher and more
hooklike, and the postzygapophyses are nearer together and turned
slightly upward. There are no accessory processes but the mammillary
processes are more prominent than in preceding vertebrae and the
laminae of the arch are compressed posteriorly for a narrowing neural
canal. The ventral surfaces of the bodies of the lumbars are slightly
differentiated posteriorly into two slight ridges or tubercles for the
psoas muscles. A single median ridge or a long hypapophysis is found
in various forms, in Phascologale, Antechinomys, Perameles and to some
degree in Marmosa. Double ridges on the posterior half of the body of
the vertebra similar to those in Ccenolestes have been observed in
Didelphis, Dasyurus, and Petaurus.

The lumbar series as a whole may perhaps be said to be characterized
by the low and inconspicuous condition of the vertebral processes which
in most other forms are more highly developed. In this respect there is
some resemblance to immature specimens of Didelphis, but even in these
the anterior and posterior zygapophyses are connected by a sharp ridge
and the accessory processes are broad, heavy, and blunt instead of small
and pointed.

SACRAL AND CAUDAL VERTEBRAE.

Plate XIV.

Two vertebrae have been regarded as sacral, although the second
touches the ileum only at the anterior extremity of its transverse process.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 85

The first has a much broader and heavier transverse process, the whole
lateral surface of which is firmly attached to the ileum. The condition
is much like that in Phascologale, Sminthopsis, Philander, and Myrme-
cobius. A slightly greater proportion of the second sacral touches the
ileum in Marmosa, Peramys, Dasyurus and Perameles; a still greater in
Didelphis, Phascolarctos and Phascolomys; and in Trichosurus, Petaurus,
Petrogale and Macropus the second vertebra is almost or quite as fully
attached as the first; while in a single specimen of Thylacinus, the second
sacral, although heavily ankylosed to the first, does not reach the ileum
at all. The two sacrals are ankylosed to each other laterally, but a large
perforation persists between their transverse processes. Their general
character corresponds closely with that of the last lumbar, the chief
differences being in their widened and more horizontal transverse
processes and in the greatly compressed zygapophyses, both anterior
and posterior in the case of the second, and mainly the postzygapophysis
in the case of the first. The first has a well-developed and forwardly
directed spinous process similar to that of the last lumbar. The second
has a smaller, more pointed spinous process directed slightly backward.

The first caudal and sometimes the second is ankylosed to its
predecessor by a narrow bony strip stretching between the extremities
of the transverse processes. The first and second are essentially similar
to the sacrals except that they have no traces of spinous processes and
that the zygapophyses are expanded slightly in the first and greatly in
the second, with short rounded metapophyses. The third and fourth
caudals are similar to the second but their transverse processes are
inclined forward and the first chevron bone is found over their inter-
vertebral space. The fifth caudal is the first to lack the neural canal and
postzygapophyses, but the transverse process is still undivided, although
somewhat reduced in lateral extent and lengthened antero-posteriorly.
In some specimens this transverse process is distinctly emarginate
laterally. The sixth caudal is elongate and subcylindrical with rather
prominent metapophyses and short crescentic anterior and posterior
transverse processes. The succeeding vertebrae are almost exactly
similar up to the sixteenth which has only the merest traces of processes.
The remainder taper gradually to a very slender tip, the total number of
caudals being twenty-seven. The ninth and tenth are the longest of the
caudals each measuring about 7.7 mm. Nineteen chevron bones are
present from the third to the twenty-second caudals. These have one to
four longitudinal ridges on their under surfaces which give them a fluted
appearance. They are nearly as wide as the bodies of the vertebrae.

The characters of the caudals are of course of little or no significance
in the indication of relationships, but it is of some interest to note that

86 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

the absence of spinous processes on the proximal caudals is a more
frequent condition among Australian than American forms. With the
exception of Perameles, which has six caudal spines, and Phascolarctos
and Phascolomys, with practically complete sets, all the Australian
forms examined either have no caudal spines or have them few and
extremely rudimentary. In no form are they so completely absent as in
Canolestes. On the other hand, there are no less than four in Marmosa,
three in Peramys, and three to six in Didelphis.1 Slight caudal spines
are reported in Notary ctes (Stirling 1891).

The angle at which the sacrals are set between the ilia is less in
C&nolestes than in most of the polyprotodonts examined and, as in
Trichosurus and Macropus, the spinal axis is nearly parallel with the
longitudinal axis of the superior outline of the pelvis. Thus, if the
caudals are extended in the axis of the sacrals the transverse processes
of the fourth and fifth caudals lie below the ischial tuberosity instead
of well above it as in Didelphis, Dasyurus, Perameles, Myrmecobius, etc.

RIBS AND STERNUM.
Plate XII.

There are thirteen pairs of ribs as usual, seven pairs articulating
directly with the sternum and six pairs not reaching it. Nine pairs are
doubly articulated with the vertebrae by tubercles and capitulae. The
succeeding four pairs lack tubercles and arise from the bodies of the
vertebrae. The first rib is much heavier than the others and its costal
element is longer than the vertebral which is true of no other marsupials
examined except Trichosurus and Petaurus. Its costal element is fully
ossified, basally flattened and slightly bicipital near its union with the
lateral extension of the manubrium. The vertebral element is similarly
expanded at the head and narrowed toward its union with its fellow.
A strong first thoracic rib, it may be noted, is an edentate character.
The remaining ribs from the second to the thirteenth are very slender
and subcylindrical, especially the vertebral series. The sternal series
are slightly flattened, the seventh being rather noticeably wider than
the others. The sternal ribs from the first to the seventh show a greater
degree of ossification than usual in marsupials. The seventh pair of
costals articulate directly with the caudal end of the last segment of the
mesosternum and therefore lie ventrad of the xiphisternum and nearly
or quite touch each other at this point (PL XII, fig. 2). In Didelphis and
most other polyprotodonts these ribs only reach the sides of the meso-

^Although six caudal spines are distinguishable in Didelphis virginiana, a speci-
men of D. paraguayensis from Ceara, Brazil, has only traces of three and the spine of
the second sacral also is undeveloped.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 87

sternum and are separated at their extremities by the xiphisternum,where-
as in Perameles and most diprotodonts the condition is as in C&nolestes.

The manubrium of the sternum is like that of most of the smaller
claviculate marsupials. It consists of an anterior prong which is some-
what deflected ventrad and slightly faceted dorso-cephalad for the
clavicular union. A small triangular lateral process projects from each
side of the base of this prong and the body of the bone extends thence
caudad as a subcylindrical rod with a slight keel on its ventral surface.
In a male specimen the manubrium is 7 mm. in length.

The four segments of the mesosternum are of nearly uniform width
throughout without pronounced median constriction or lateral expansion
even in the fourth which is scarcely wider posteriorly than anteriorly.
The first is somewhat more rounded than the others which are slightly
flattened but with their ventral surfaces rather more convex than other-
wise.

The xiphisternum is a long slender bony rod with slightly greater
width than thickness, but, in general, rounded in form especially proxi-
mally where it is somewhat expanded. Distally it has a more abrupt and
more flattened slight expansion where it joins the xiphoid cartilage,
which is relatively large and scutate. The xiphisternum is separated
from the last segment of the mesosternum dorsally by a broad epiphysial
cartilage and ventrally by the extremities of the seventh pair of ribs.
Hence, although the plane of the dorsal surface of the xiphisternum is
continuous with that of the mesosternum, its ventral plane is distinctly
higher. This epiphysial cartilage and this relation of the proximal end
of the xiphisternum, although apparently trivial, are nevertheless
characteristic of the diprotodonts as opposed to the polyprotodonts.
In all polyprotodonts examined, except Perameles, both surfaces of the
xiphisternum are practically continuous with the corresponding surfaces
of the mesosternum and the sixth or seventh ribs do not meet ventrad
of the base of the xiphisternum. In all diprotodonts, including Cazno-
lestes, the reverse is the case. Perameles, as judged by one specimen,
seems to occupy an intermediate position in this respect, for although
ribs meet on the end of the mesosternum, the epiphysial cartilage is so
reduced as to be scarcely noticeable in a dried specimen.

PECTORAL GIRDLE.

The clavicle (PI. XII, fig. 5) is well developed and stretches from the
sternum to the acromion in the manner usual among marsupials. Its
lateral or acromial end is compressed and curved upward. In the adult,
the union with the acromion is almost direct, the intervening cartilage
being scarcely evident. It is attached also to the coracoid process by a

88 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

broad flat coraco-clavicular ligament. The medial or sternal end of the
clavicle is somewhat expanded and deeply notched for the short stout
omosternal cartilage1 by which it is attached to the front of the
manubrium.

The scapula (PI. XIII, fig. 3) is less angular than iri the majority of
marsupials, having a slightly elongate, gently curving outline without
sharp prominences. The suprascapular notch, so-called, is not a notch
but a long sweeping curve which does not reach its highest point until
about two-thirds of the distance from the coracoid process to the
vertebral border of the scapula. The axillary border has a slight con-
vexity, instead of being practically straight, and it meets the vertebral
border in a modified curve which in some specimens may be a trifle
angular. The slightly thickened vertebral border is covered by a small,
elongated, cartilaginous, or sometimes partly osseous, suprascapular
epiphysis.2

The spine is relatively low and continuously deflected so that it is
nowhere perpendicular to the fossae on either side of it. Thus the
antero-ventral triangular half of the infraspinous fossa is overhung by
the flat deflected surface of the spine, which continues to the well-
developed acromion, turning sharply forward beyond the triangular
metacromion. The coracoid process is a prominent prong with a slight
notch in its glenoid surface. The superior fossae are nearly equal in
extent, the supraspinous usually having slightly the greater surface.
The inferior surface or subscapular fossa presents a deep longitudinal
channel opposite the base of the spine and on either side of this the
surface is somewhat bulging.

Taken as a whole the scapula of C&nolesies is more similar to that
of Sminihopsis than to that of any other form which is available for
comparison. It agrees with this in its general outline and in its much
deflected spine, whereas practically all other forms have a more angular
shape, a less deflected and usually higher spine, and a shorter more
abrupt suprascapular notch.

PELVIC GIRDLE.

Plate XIII, Fig. 4.

The pelvis is completely ossified and shows no traces of division in
adult specimens. It is relatively long and deep but light in structure.
Although it is not especially peculiar in other respects, it is strikingly
distinguished from that of all other marsupials by its extremely short

1 In a skeleton of Philander, this cartilage is partially ossified.

2 A distinct bony suprascapular epiphysis is especially evident* in Perameles and
Thylacinus.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 89

pubic symphysis. This scarcely equals the diameter of the acetabulum.
The general proportions of the pubes and ischia are thus essentially
similar to those in various edentates and insectivores. The rami of the
pubis and ischium together with the body of the ischium form a nearly
equilateral triangle having its angles respectively at the ischial tuberos-
ity, the pubic symphysis, and the cephalic border of the acetabulum.
The pubic ramus is very slender and is slightly ridged on its lateral
surface. The ischial ramus is slightly wider and more compressed and
the body of the ischium is a trifle more so, but all these bones are
relatively slender and the obturator foramen which they enclose is large
and subtriangular. The ischial rami are not greatly divergent but more
nearly approach the perpendicular than those of any other marsupial
and in this respect show resemblance to certain insectivores. The
ischial tuberosity is fairly well developed and in some specimens forms
a decided protuberance.

The ilium is slender and of the ordinary prismatic shape, being
somewhat flattened and everted at the tip. Its ischial border forms a
high ridge continuous with the inner edge of the ischium and to some
extent overhanging the anterior part of the obturator foramen in shelf-
like manner. A slight lliopectineal tubercle is present. The acetabular
border of the ilium forms a fairly well marked ridge, especially ante-
riorly where it separates the gluteal and iliac surfaces of the ilium. The
acetabulum and cotyloid notch are of the usual character.

The marsupial bones are well developed, their unusually long straight
bases being attached to almost the entire anterior edge of the long
ascending pubic ramus, that is, from the symphysis to a point directly
below the acetabulum and caudad of the iliopectineal tubercle. The
base of the marsupial bone is thus almost as long as the anterior pro-
jecting spine and proportionally longer than in any other marsupial. In
a male specimen, the marsupial bone measures 5.6 mm. in length as
against a base of 4.8 mm., the ratio of base to length being 85.7. This
ratio is approached only in a few forms in which the marsupial bones are
notably rudimentary. The following ratios indicate how widely C&no-
lestes differs from the majority of marsupials in this respect : C&nolestes
85.7; Didelphis, 40; Dasyurus, 34; Myrmecobius, 15; Sminthopsis, 52;
Perameles, 14.2; Trichosurus, 41; Phascolarctos, 37.3; Petrogale, 9.

ARM AND FOREARM.
Plate XV.

In proportion to the' size of the animal, the humerus is large and
heavy, being much the largest single bone in the body. It differs con-
spicuously from that of most marsupials in its great degree of curvature

go FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

and in its prominent flat-faced deltoid ridge. This occupies about three-
eighths of the inner proximal part of the front of the humerus, and instead
of forming a single trenchant ridge it presents an elevated flat rhom-
boidal surface extending downward from the greater tuberosity and
ending bluntly and abruptly. Its surface is slightly roughened and lies
at a considerable angle to the long axis of the humerus. From its distal
extremity the front of the shaft abruptly recedes and forms a long
sweeping curve to the condyles. The deltoid ridge or tuberosity has a
flattened surface in some other forms, as Perameles, Dasyurus, and
Peramys, but in these it is triangular, coming to a point distally. In
the fossil form Garzonia patagonica, of which a humerus is figured by
Sinclair (1906, PI. XL/Ill, fig. n), the deltoid surface is very broad,
and although it is less blunt distally, it seems to approach the condition
in Ccsnolestes more nearly than does that of any recent form.

The head of the humerus is rather large and has considerable back-
ward extension beyond the axis of the shaft. The external or so-called
greater tuberosity is much smaller and less marked than the internal.
It is greatly compressed and its outer surface is continuous with the
deltoid ridge. Its upper surface is continuous medially with the surface
of the head of the humerus. The internal tuberosity is prominent and
distinctly separated from the head by a deep groove as well as from the
external tuberosity by the bicipital groove. A well-marked ridge ex-
tends distad from it parallel with the internal border of the deltoid
ridge and terminating in a roughened area for the insertion of the
latissimus muscle.

The middle of the shaft of the humerus for about one-third the length
of the bone is relatively smooth, slightly compressed, subcylindrical,
with only traces distally of the upper continuation of the external
epicondylar or supinator ridge. The caudal outline from the head of
the humerus to the supinator ridge is a deep sigmoid curve. The distal
end of the humerus is broadly and rather abruptly expanded. The
internal condyle is prominent and the sulcus between it and the ulnar
trochlea is developed as an articular surface for the inner upper lip of
the sigmoid cavity of the very obliquely situated ulna. The trochlea
is somewhat smaller than the capitellum with which it is continuous.
Thus there are three articular surfaces at the distal end of the humerus,
a very unusual condition. The third or internal articular surface is fully
developed only on the caudal aspect of the bone and does not continue
to the front as the trochlea and capitellum do. Although partly con-
fluent, these three articular surfaces are sufficiently distinct from each
other, but of course the third or inner one could be regarded as an
extension of the trochlea which might be developed without difficulty.

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 91

Nevertheless, I have been unable to find a case of a similar extension in
any other marsupial or in other mammals. Its effect is a much wider
articulation with the ulna than usual and, as described below, there is a
corresponding increase of the articular surfaces of the ulna. Func-
tionally it may have some correlation with the distinctness of the
median head of the triceps muscle, and although it might not be re-
markable in an animal of fossorial habits, it seems quite incompatible
with the distal elements of the arm and leg which are adapted to cur-
sorial life.

A small oval epicondylar foramen is present and the ridge bounding
it has a smooth rounded surface directed with a spiral curvature toward
the distal end of the deltoid crest but widely separated from it by the
smooth shaft. The outer epicondylar ridge is similarly twisted and
gradually merges with the surface of the shaft without any abrupt angle
or hooked process. Its edge distally is broadly curled so that it presents
a relatively broad flattened lateral aspect instead of the sharp edge of
most other marsupials. The external epicondyle is little more than the
extremity of this ridge. It has a decidedly concave surface for the seat
of the sesamoid of the tendon of the supinator brews muscle. This
sesamoid, as mentioned elsewhere, is not present in other marsupials.
It is sufficiently distinct to be obvious to the preparator and is usually
found in situ in first class ligamentary skeletons.

The ulna is broad, heavy and compressed proximally but rapidly
tapers to its distal third which is subcylindrical and scarcely half the
diameter of the radius. The outer side of its proximal half is deeply
channelled. Its inner side is smooth and convex except an excavated
area on the side of the olecranon. The upper lip of the greater sigmoid '
cavity is unusually broad, especially its inner division which extends
well beyond the inner base of the olecranon from which it stands out
prominently. Its inner division has the usual position opposite the
lesser sigmoid cavity. This broad upper lip articulates with the whole
of the posterior articular surface of the humerus and in doing so brings
the ulna into a very oblique or nearly transverse position with its
broad grooved outer side directed nearly forward. Except at its distal
end the ulna is caudad of the radius. The olecranon is relatively thick
and heavy with a rhomboidal roughened extremity for various muscle
attachments.

The radius has a rounded ovoid head supported on a distinct neck
below which is a prominent tubercle having a rounded roughened face
quite two-thirds as large as that of the head. The proximal third of the
radius is rather abruptly flexed forward leaving a considerable space
between it and the ulna. For the remainder of its length it is in close

92 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

contact with the ulna, so close that more than a slight degree of rotation
is not probable.

While the heavy humerus and the broad elbow joint of C&nolestes
seem adapted for digging or scratching habits, the character of the fore-
arm distinctly tends toward the cursorial type of structure. Aside from
a few detailed peculiarities, it may be said that the humerus has con-
siderable general resemblance to that of Peramys and Marmosa while
the radius and ulna are clearly nearer those of Antechinomys and
Sminthopsis or even Perameles.

BONES OF THE HAND.
Plate XVI, Fig. i.

The carpal bones are eight in number, four in the distal row and four
in the proximal row, including the pisiform. In the distal row the
trapezium forms the seat of the first metacarpal and is in contact also
with the inner proximal extremity of the second metacarpal. Its inner
surface articulates with the trapezoid and its inner palmar surface with
the distal projection of the scaphoid. It is a short longitudinally com-
pressed bone with broad grooves on its distal and proximal surfaces
respectively. No sesamoids could be detected in relation with it. The
trapezoid and magnum are about equal in size. The former is rounded
distally for articulation with the concavity of the proximal end of the
second metacarpal. It articulates also with the trapezium, magnum,
and scaphoid. The magnum is subtriangular in its anterior or dorsal
aspect and articulates with the third metacarpal, the trapezoid, scaphoid
and unciform. The unciform is large and roughly pentagonal in outline.
It supports the fourth and fifth metacarpals. Its contact with the fifth
however is limited to its radial side. The unciform articulates with five
bones, viz., the two metacarpals, the magnum, lunar, and cuneiform.
Its dorsal surface has two deep pits. The cuneiform receives the head
of the ulna in its deeply concave proximal surf a6e. Distally it articulates
with the unciform and medially with the lunar. The lunar is quite
distinct and presents a triangular dorsal surface about half as extensive
as that of the magnum. With the scaphoid it supports the broad head
of the radius. The scaphoid is nearly as large as the unciform. It lies
between the head of the radius and the trapezoid and sends a curving
process beneath the palmar surface of the trapezium. Its medial or
ulnar surface is in contact with the lunar. No sesamoids were found
associated with it. The pisiform is exceedingly well developed and
broadly expanded distally. It is nearly as long as the fifth metacarpal
and relatively longer than in any other marsupial which has been

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 93

examined. It articulates principally with the outer lateral surface of the
cuneiform but also has slight contact with the ulna.

The metacarpals are slender and rounded with somewhat expanded
heads and bases, the former slightly the larger. The medius is the
largest. The indicis and annularis are slightly shorter and of about
equal length. The pollicis is about three-fifths as long as the indicis
and the minimus is scarcely more than half as long as the annularis.
The first three are of approximately equal diameter, the annularis is very
slightly more slender, and the minimus is slightly broader. The outer
pairs are evenly and slightly divergent from the middle. The pha-
langes are regular in number and form. The basal ones of the second,
third and fourth fingers are nearly equal in length; that of the fifth
finger is slightly shorter, and that of the pollex is shortest. The claw
of the middle finger slightly exceeds the others. The claw of the
pollex reaches to the middle of the first phalanx of the index finger and
the claw of the fifth finger extends to the first phalangeal joint of the
fourth finger. The ungual phalanges of the first and fifth fingers are
pointed but less curved and clawlike than those of the middle fingers.
The usual paired sesamoids are present on the palmar sides of the
metacarpo-phalangeal joints.

THIGH AND LEG.

Plate XII, Figs. I, 3.

The femur is slightly shorter and decidedly more slender than the
humerus. The shaft is cylindrical and nearly straight but the extremities
are considerably expanded. The head is supported on a slight neck and
is definitely directed forward as well as inward. Its smooth articular
surface is partially incomplete behind. The outer or greater trochanter
rises to the level of the top of the head and is inclined slightly inward so
that its apex is practically in the long axis of the shaft. Laterally it is
much compressed and it is only superficially separated from a broad
ridge extending distad a short distance to a small rounded protuberance
having a roughened surface and clearly representing a third trochanter.
This has approximately the same relations as the third trochanter in
Phascolomys, the only other marsupial except the extinct Wynyardia in
which there is any obvious development of the kind. The great tro-
chanter is bounded on the inner side by a deep slitlike digital fossa
which has as its other boundary a high thickened intertrochanteric ridge
almost as prominent as the great trochanter. This ridge runs from the
apex of the great trochanter distad with slight inclination toward the
lesser trochanter and then turning sharply and losing most of its eleva-

94 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

tion joins the lesser trochanter. A considerable depression lies between
its lower half and the base of the head of the femur. In many marsupials,
this ridge is only represented by a tubercle between the great trochanter
and the head. It is fairly well developed in Didelphis and in Perameles.
The lesser trochanter forms a prominent triangular projection on the
inner side of the femur separated from the head by a short rounded
notch.

The distal end of the femur is of the usual character. The condyles
are produced backward to a marked degree and the intercondylar notch
is narrow and deep. The anterior rotular groove is broad and shallow
being scarcely more marked than in forms with a ligamentary patella.
The patella itself is rather short and broad and well ossified. This is an
important peculiarity of C&nolestes, for among living marsupials the
only others having an osseous patella are Tarsipes, Perameles and
Notoryctes. It is recorded also as present in the extinct genera Prothy-
lacynus and Amphiproviverra (Sinclair).

The tibia is very long and much curved, resembling that of the
macropods and certain small saltatorial polyprotodonts, as Sminihopsis
and Antechinomys. Its proximal half is bowed inward and forward in a
long curve and is widely separate from the fibula but sweeps downward
into the same longitudinal axis for its distal half. The proximal half is
of the usual trihedral form, its wide inner surface being longitudinally
channelled and its antero internal surfaces convex. The distal half is
nearly cylindrical. The internal malleolus is highly developed, only
being equalled among forms examined in the macropods. Its perpendicu-
lar articular surface exceeds the horizontal surface of the end of the
tibia.

The fibula is very slender, especially in its distal three-fourths.
Proximally it is compressed and somewhat expanded antero-posteriorly,
its inner surface being somewhat grooved. It is surmounted by a rather
large ovoid fabella. The tibia and fibula are closely appressed but not
fused in the distal half of their length. The hind leg of Ccenolestes is thus
of a cursorial or saltatorial type and closely resembles that of Antechi-
nomys, although even in this notably saltatorial form the internal
malleolus is relatively small. A leg of similar character with tibia and
fibula practically joined distally is found in Sminthopsis, Perameles, and
in the Macropodidae. In almost all other marsupials the tibia and
fibula are widely separate throughout.

The relative length of the lower leg in C&nolestes is very great. In
proportion to the length of the femur, the length of the tibia in C&nolestes
is only exceeded, among forms examined, by that of the Giant Kangaroo.
The proportions in various genera are indicated by the following table,

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 95

the ratio given being that of the femur to the tibia. From this table it
is seen that this ratio is less in Ccenolestes than in any other forms ex-
amined except the Giant Kangaroo and that it is approximated only by
saltatorial forms.

Length of Length of Ratio of Femur
Femur Tibia to Tibia

Macropus giganteus 265 465 57

Csenolestes obscurus 14.3 22.5 63.5

Sminthopsis crassicaudata ... 13.6 19.8 68

Antechinomys laniger .... 19 27 70

Petrogale xanthopus 153 195 78

Phascologale cristicauda .... 23.3 28 82

Marmosa mexicana 23 26 . 8 85 . 8

Dasyurus viveninus 65 73 89

Myrmecobius fasciatus .... 45.5 51 89

Perameles nasuta 64.5 69 93-4

Peramys domesticus 21 22 95-4

Philander philander 35-4 37 95-6

Didelphis virginianus 76.8 78 98.4

Sarcophilus harrisi 91-5 9^-5 loo

Trichosurus vulpecula .... 94-5 92-5 102.1

Thylacinus cynocephalus . ... 195 190 102.6

Phascolomys ursinus 175 133 131

Phascolarctos cinereus .... 155 115 134

BONES OF THE FOOT.
Plate XVI, Figs. 2-5.

The foot is long and narrow, the tarsus being relatively short and
the metatarsus very long. It has general resemblance to the foot in
Dasyurus, Sminthopsis , Antechinomys and similar forms, but its structure
shows decided approach to that of Perameles and the macropods.
Although the hallux is present and relatively well developed (longer
than in Dasyurus, for example) the relations of the ankle joint are more
adapted to cursorial life than in some of the forms in which the hallux
has nearly or quite disappeared. This is shown especially by the astra-
galus which is somewhat intermediate between that of the pedimanous
marsupials and that of the highly saltatorial forms. Although it retains
general similarity to the astragalus of Didelphis, Sarcophilus, Myrme-
cobius, Phascologale and various other polyprotodonts,it shows a distinct
approach to the condition in the Macropodidae in which there is one
large trochlea and facets on either side, nearly at right angles, for the
outer malleolar parts of the tibia and the fibula. The body of the
astragalus stands at right angles to the neck and head and scarcely
exceeds them in size. The neck is not constricted to differentiate it

96 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

from the head. The inner malleolar facet is extensive and sharply
deflected from the medial tibial facet from which it is separated by a
high trenchant ridge which extends over the neck to the head at the
navicular facet. In all other forms examined, this ridge is discontinuous
or confined to the trochlea. The formation of a depression in this ridge
between the trochlea and the head would seem to be the first step in
the separation of the head from the trochlea. Hence, in this respect the
astragalus of Canolestes may be primitive. The tibial facet is very
slightly concave and merges imperceptibly with the fibular facet which
is narrow and considerably deflected. The sharp division of the two
principal tibial facets virtually permits the tibia and astragalus to be
locked together almost as effectively as in the macropods. Lateral
movement of the tibia is practically precluded and the general freedom
of the ankle joint found in most polyprotodonts is quite impossible.
Even such forms as Sminihopsis and Antechinomys have much looser
ankle joints than Ccenolestes.

The inferior surface of the astragalus in C&nolestes is largely occupied
by the sustentacular facet which is unusually broad and convex. It
is nearly at right angles but confluent with the large navicular facet
which covers the head of the bone and runs up the inner side for a
considerable distance. The ectal facet is crescentic in outline and is
separated from the sustentacular by a deep groove. The inner anterior
margin of the navicular facet has slight contact with the cuboid bone
but there is no distinct development of a cuboid facet. A small sesamoid
is situated between the astragalus and the posterior distal extremities of
the tibia and fibula.

The os calcis has the tuber laterally compressed and very slightly
flexed downward and inward. The anterior sustentacular part of the
bone is relatively large and broad. In addition to the usual ectal,
sustentacular, and cuboid facets it has a prominent triangular projec-
tion on its antero-external surface for the attachment of the tarso-
metatarsal and perhaps other ligaments.

The navicular is somewhat anvil-shaped with a smaller dorsal head
articulating with the astragalus and slightly constricted from the larger
plantar part which spreads inward to the middle of the sole, its plantar
surface being broad and flat instead of narrow and ridgelike. Distally
it articulates with the ectocuneiform, entocuneiform and mesocuneiform.
The cuboid articulates distally with the two outer metatarsals and with
the proximo-external surface of the mesocuneiform. Its plantar surface
has the usual groove for the tendon of the peroneus longus muscle. The
ectocuneiform, entocuneiform and mesocuneiform are rhomboidal bones,
the first supporting the hallux, and the others the second and third

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 97

metatarsals respectively. The ectocuneiform is about twice as large as
the mesocuneiform and nearly four times the size of the entocuneiform.
The hallux is supported on the distal end of the ectocuneiform more
especially toward its plantar side and has considerable freedom of
movement, although it is but slightly divergent. It is much more slender
than the other digits and its terminal phalanx reaches only to the distal
extremity of the second metatarsals. The proximal bases of the meta-
tarsals form a very uneven row. The fifth reaches farthest proximad,
the second and third are considerably short of it, and the first and
fourth are still more so. The irregularity in this respect is greater than
in any other forms examined. The second and fifth extend distad about
the same distance and are slightly exceeded by the third and fourth
which also end evenly. The fourth however is somewhat longer than
the third by actual measurement since the uneven bases give it greater
extension proximad. In a male specimen it is 9 mm. long. The first and
also the second phalanges are approximately equal in length from the
second to the fifth digits. The paired sesamoids on the metatarso-
phalangeal joints are as usual. There is no prehallux.

SUMMARY OF SKELETAL CHARACTERS.

The skeleton of C&nolestes has the broad general features found in
most other marsupials. With the exception of modifications correlated
with cursorial limbs, most of its peculiarities are of a primitive or
generalized nature rather than of high specialization. The movable
cervical rib and the very short pubic symphysis are examples. The
general form of the pelvis is remarkably similar to that of some insecti-
vores. In fact the whole skeleton in its general aspects is so closely
paralleled by that of one species of insectivore (Nesogale dobsoni), which
chances to be available, that the two have been illustrated together
(PI. XIX).

Skeletons of the following marsupials have been available for com-
parison: Antechinomys laniger, Sminthopsis crassicaudata, Phascologale
(Chcetocercus) cristicauda, Dasyurus viverrinus, Thylacinus cynocephalus,
Sarcophilus harrisi, Didelphis virginianus, Marmosa marica, Peramys
domesticus, Philander laniger, Myrmecobius fasciatus, Perameles nasuta,
Petaurus australis, Trichosurus vulpecula, Phascolarctos cinereus, Mac-
ropus giganteus, and Phascolomys mitchelli. A larger representation
would have been desirable and it is possible that some conclusions will
need modification when wider comparison is made. Numerous minor
peculiarities of general characters of the skeleton have been discussed
in preceding pages.

98 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

The features which seem most noteworthy, although of widely vary-
ing significance, are the following:

1 . Atlas with open grooves or notches instead of closed foramina
for passage of cranial nerves and vertebral arteries; agreeing
only with Perameles, Marmosa, and Peramys.

2. Axis with movable cervical rib; agreeing only with certain
species (or individuals?) of Perameles and Phascologale.

3. Seventh cervical tending to become imperf orate; inter-
mediate between diprotodonts and polyprotodonts.

4. Spine of first thoracic vertebra scarcely exceeding that of
last cervical but greatly exceeded by that of second thoracic;
agreeing only with Perameles.

5. Vertebral spines converging to a center of motion; disagree-
ing with Didelphiidae.

6. Costal element of first rib longer than vertebral; agreeing
only with Trichosurus and Petaurus.

7. Ventral surface of xiphisternum not continuous with that of
mesosternum; agreeing only with diprotodonts.

8. Pubic symphysis exceedingly short; unique among mar-
supials.

9. Ischial rami not flaring.

10. Marsupial bone with very long base; unequalled among
marsupials.

11. A large heavy humerus combined with a relatively slender
forearm; unique among marsupials.

12. Trochlea of humerus extended inward to form a partially
separate articular surface; unique among marsupials.

13. A sesamoid in the supinator tendon at the external epicondyle
of the humerus; unique among marsupials.

14. Radius flexed forward in its proximal third, in contact with
ulna distally; agreeing with saltatorial or semi-saltatorial
forms.

15. Pisiform large, relatively larger than in other marsupials
examined.

1 6. Patella ossified; agreeing with Perameles, Notary ctes and
Tarsipes; also with extinct forms Amphiproviverra and Pro-
ihylacynus. .

17. Femur with a rudimentary third trochanter; agreeing only
with Phascolomys and Wynyardia.

1 8. Tibia and fibula relatively long and in close contact distally;
agreeing only with saltatorial forms.

19. Internal malleolus of tibia large ; agreeing with Macropodidae.

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 99

20. Astragalus with a large inner malleolar facet nearly per-
pendicular to the medial tibial facet; agreeing with saltatorial
forms.

21. Prehallux not present; agreeing with Macropus, Trichosurus,
and Phascolomys; disagreeing with Didelphis.

SKULL.

Plates XVII, XX.

General form. — The skull of C&nolestes is characterized superficially
by the great elongation of the facial region, by the deep and extremely
thin- walled braincase, the swollen interorbital and lacrymal regions, the
general absence of crests or ridges, the slender zygomata, and by the
modifications of the maxillary by which the anterior teeth are at a
different level from the posterior.

The general resemblance to the skull of Perameles was immediately
noticed by Thomas, who says (1895, p. 872): "Skull in its general
proportions something like that of a Perameles, although thinner and
more delicately built,- with a similarly elongated muzzle, smooth and
rounded brain-case, and obsolete supraorbital and cranial crest and
ridges; the zygomata are, however, so much more boldly expanded as
somewhat to spoil the resemblance, which in any case does not apply
to details." A still stronger statement of this resemblance to the skull
of Perameles might be made and, as appears later, it does apply at least
to a certain number of details. The zygomata are more spreading than
in some of the peramelids but their sweep is practically identical with
that of others (e. g. Isoodon}. Certainly the general form of the skull
is nearer that seen in the peramelids than that of any other living mar-
supial. The resemblances of general shape to the smaller dasyurids
noted by Miss Dederer (1909) seem mostly to be associated merely with
small size and are found also in Acr abates and Dromicia.

The ratio of zygomatic breadth to total length of skull in various
marsupials is as follows: Perameles nasuta 39.7, Perameles bougainvittei
43.2, C&nolestes 47.7, Isoodon barrowensis 51, Philander 54, Phasco-
larctos 54.3, Myrmecobius 56.4, Peramys 56.9, Didelphis 57, Chironecles
57.1, Phascologale (Ch&tocercus) cristicauda 63.7, Petaurus australis 66.4,
Dactylopsila 67.5, Dasyurus viverrinus 69.7, Trichosurus vulpecula 71.7,
Sarcophilus 81.2. In this series, C&nolestes falls with the peramelids.
The didelphids occupy a position midway between these dolichocephalic
forms and the broader types presented among the phalangers and
dasyurids.

Braincase. — The braincase is relatively large and its walls are

ioo FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

exceedingly thin and papery. Viewed in longitudinal section (PI. XVII)
its depth is seen to be proportionately greater than in other marsupials.
The ratio of the depth to the length measured inside the braincase in
several forms is found to be as follows: Ccenolestes 53.9 ; Perameles nasuta
48.8; Petaurus australis 47; Myrmecobius 44.4; Trichosurus 42.2; Phas-
cologale flavipes 41.7; Marmosa cinerea 41.6; Didelphis virginiana 39.7;
Peramys domesticus 38.6.

The dorsal outline of the braincase is broadly arched, agreeing in
this respect with Perameles and with most diprotodonts and contrasting
with the didelphids and dasyurids in which the dorsal outline is more
nearly straight. The olfactory fossa is large and full, especially in its
lateral parts, but it is relatively short antero-posteriorly as compared
to didelphids and dasyurids. In these polyprotodonts its median section
is subtriangular, often with an acute angle in front, whereas in Ccenolestes
and Perameles, and the several diprotodonts examined, it is broadly
rounded anteriorly. It is separated from the cerebral fossa by a light
ridge, the position of which is plainly indicated externally by the furrow
running upward and slightly backward from the ethmoid foramen.

The cerebral fossa is very full, especially in its lower anterior part
where it is extended forward on either side of the median part of the
cribriform plate. The floor of the braincase is thus wider and flatter
anteriorly than in other marsupials and it meets the cribriform plate
more abruptly. This is in the region occupied by the tuberculum olfac-
torium which, as elsewhere noted, is unusually large (see p. 159). In
front of the occipito-sphenoid suture is a shallow pituitary fossa. The
sphenoidal fissure and the foramen rotundum open posteriorly on the
floor of the braincase and have no lateral exposure like that in other
marsupials. Externally the sphenoidal fissure and the foramen rotundum
have what is practically a common opening, the bony septum between
them being quite internal and not visible in lateral view. In this, the
condition seems nearer that of the diprotodonts than of the polyproto-
donts in which the foramen rotundum opens decidedly posterior to the
sphenoidal fissure with the septum between them broadly exposed in
lateral view. •

The cerebellar fossa is smooth and without obvious division into
lateral and median parts. It is separated from the cerebral fossa by an
exceedingly slight tentorial ridge. The outline of the cribriform plate
as seen in section is more similar to that of Perameles than to that of any
other form examined. It is more nearly vertical than in polyprotodonts
generally and more abruptly divided between its upright and its posterior
basal extension in the presphenoid region. This posterior extension is
broader and higher than in any other marsupial examined.

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 101

The foramina leaving the braincase include the sphenoidal fissure
and foramen rotundum, which already have been mentioned, the fora-
men ovale, transverse canal, entocarotid foramen, and ethmoid foramen,
in all of which the relations are generally similar to those in other mar-
supials although there is much variation, especially in diprotodonts.

Ethmoid and turbinals. — The cribriform plate of the ethmoid, as
previously stated, is more nearly upright than in the didelphids and
resembles that of Perameles. Its exact relations with the orbitosphenoid
and presphenoid are not distinguishable in the adult, but the bony floor
of the olfactory fossa is high and has a more rounded surface than in
other marsupials. It extends exactly parallel to the palate and hence
nearly at right angles to the anterior wall of the braincase. The prin-
cipal ethmo-turbinals visible in a simple section (PI. XVII, et.) are four
in number. The lowermost is relatively larger than in any other form
examined and is most nearly approached in Perameles nasuta. The
uppermost and longest of the ethmo-turbinals is relatively short,
extending forward only to the plane of the front of the last premolar.
This seems to be a resemblance to diprotodonts since it agrees with the
condition in Trichosurus and Petaurus. In the didelphids and in Phas-
cologale this upper ethmo-turbinal is very long, reaching as far forward
as the canines and almost completely inclosing the fenestrated part of
the maxillo-turbinals. In Perameles and Myrmecobius an intermediate
condition appears.

The naso-turbinal offers no especial peculiarities. The maxillo-
turbinal springs from the nasal wall at a somewhat higher level than in
polyprotodonts and the nasal floor is narrower.1 In these respects,
there is some variation among polyprotodonts but in general the
resemblance seems to be to diprotodonts. The maxillo-turbinal, as in
diprotodonts, is not largely enclosed by the anterior extension of the
ethmo-turbinal. The mesethmoid is a thin vertical plate of the usual
character.

Nasals. — These are long and slender, coming nearly to a point
anteriorly and being abruptly expanded posteriorly. Their anterior
extremities exceed the dorsal edges of the premaxillae and overhang the
anterior nares, reaching nearly as far forward as the lower anterior
extension of the premaxillae. In this they differ markedly from the
Dasyuridae, Thylacynidae, and Myrmecobiidae but are approached by
most diprotodonts, by the didelphids and especially by the peramelids.
The posterior expansions of the nasals include a slender curved prong
on each side projecting across the extraordinary vacuity which lies

1 This is contrary to the observation of Broom (191 1, p. 318), whose material may
have been imperfect.

io2 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

between the nasals, frontals and maxillary. They do not exclude the
maxillaries from contact with the frontals, however, and despite their
great length their posterior endings are in advance of the orbits.

Premaxillae. — The premaxillae are longer than high and reach quite
to the bases of the canines. The maxillary suture approaches the per-
pendicular as in Perameles and most diprotodonts. A narrow posterior
ascending branch extends backward between the nasals and the maxil-
lary to the plane of the front of the middle premolar. The palatine
processes of the premaxillae do not extend the full length of the anterior
palatine vacuities as in most other marsupials, but meet similar processes
of the maxillary which extend forward. The bony division of the
anterior vacuities thus consists of premaxillary elements only in its
anterior three-fifths. The only other marsupials in which I have found
this arrangement are the peramelids; in all others the premaxillary
processes extend nearly or quite to the posterior boundary of the
vacuities. In every respect, therefore, the premaxillae are more similar
to those of the peramelids than of any other marsupial.

Maxillae. — The slender processes of the maxillaries projecting
forward in the anterior palatine vacuities are only paralleled among the
peramelids. The posterior fenestration of the palatal aspect of the
maxillae also resembles closely that of Perameles. In their upward
extension, the maxillae are not greatly narrowed between the antorbital
vacuities and the lacrymals, but meet the frontals broadly at a point
about even with the front of the orbits and somewhat caudad of the
posterior nasal endings. In these respects they are conspicuously
different from those of the didelphids.

In uniting with the jugals, the maxillaries send out a slender prong
which takes a larger part in the formation of the anterior root of the
zygoma than is usual in marsupials and contrasts especially with the con-
dition in most polyprotodonts in which the jugal is expanded on the
face below the lacrymals. In general, however, the relation of the an-
terior root of the zygoma to the maxillary is as in polyprotodonts, that
is, the principal part of the maxilla, including all the molars except the
last two, is anterior to the zygoma. This is in decided contrast to the
diprotodonts in which that part of the maxillary bearing the molars is
largely posterior to the anterior root of the zygoma. In some forms,
especially among the Phalangeridae, the root of the zygoma rises from a
point as far forward as the last premolar and consequently the entire
molar series is posterior to it. In polyprotodonts, on the contrary, it is
rare for more than the last reduced molar to lie posterior to the zygoma.
This constitutes one of the most striking resemblances to polyprotodonts
shown by the skull of Canolestes.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 103

Frontals. — The frontals are nearly as wide anteriorly as posteriorly.
They are smoothly rounded at the orbital border and have neither ridges
nor post-orbital processes, thus differing from most polyprotodonts.
There is no post-orbital process in Phascologale and some allied forms but
in these the frontals are more flattened and the orbital border is more
angular. Within the orbit, the frontals of Ccenolestes are less deflected
toward the midventral line than in other marsupials and thus provide
for the unusual space required by the large olfactory lobes of the brain.
A distinct venous foramen perforates the orbital border in its upper
anterior part.

Lacrymals. — The lacrymal is of moderate size and not largely
extended beyond the orbit. It is slightly gibbous but not ridged between
its facial and orbital parts. It is well separated from the nasals by the
frontal process of the maxillary. The lacrymal canal has a single open-
ing, thus differing from the didelphids and many other marsupials. It
opens on the orbital border, not within the orbit, but not so definitely
outside it as in other forms.1

Parietals. — The parietals are broad and smooth with no suggestion
of a sagittal ridge. The suture between the two parietals and that
between them and the supraoccipital is practically obliterated in adults
although easily visible in most other marsupials except the peramelids.
Apparently the parietals reach laterally to the vicinity of the lambdoid
ridge as in other forms and are not shortened as in the figure published
by Miss Dederer (1909). A minute postparietal foramen is usually
present.

Zygomata. — The anterior root of the zygoma, as previously stated,
rises at the level of the third molar in marked contrast to the position
general among diprotodonts. The jugal is sharply narrowed anteriorly
and comes to a point just below the lacrymal canal. The arch has a
general sweep somewhat as in Phascologale and the bone is similarly
flattened but the anterior base is narrower and there is no trace of a
postorbital process, this last character being a resemblance to Perameles.
The jugal reaches the glenoid fossa as usual and its ending is truncate to
form a slight anterior boundary for the fossa, thus agreeing more nearly
with polyprotodonts than diprotodonts. The posterior root of the
zygoma is not inflated as in most diprotodonts, but it is more slender
than in dasyurids and didelphids and more nearly resembles the condi-
tion in the peramelids.

Squamosal. — The cranial part of the squamosal is not peculiar, being

1 In Gregory's exhaustive comparative study of the lacrymal (Bull. Am. Mus. Nat.
Hist., XLII, p. 235, 1920), made since the above was written, it is concluded that
"The evidence of the lacrymal region is in harmony with the view that the caenoles-
toids are allied with the Peramelidae."

104 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

smaller than in Perameles and Sarcophilus, smaller also than in some
diprotodonts, but relatively quite as large as in others. Its proportions
are nearly as in the didelphids and smaller dasyurids. The glenoid
fossa is considerably above the basicranial axis and is distinctly longer
than wide. Except on its extreme outer edge where the end of the
jugal bounds it, it is open in front. Posteriorly it is bounded on the
outer side by a papillate bony elevation which is unconnected with the
tympanic and has only very slight transverse extent. The general
features of the glenoid fossa are thus somewhat intermediate between
those in polyprotodonts and in diprotodonts. The transverse extent
and the broad platelike posterior boundary usually found in polyproto-
donts are changed in the direction of the narrower more open condition
of the diprotodonts which allows of greater antero-posterior motion of
the jaws in conformity with the requirements of their dentition. The
arrangement of the foramina at the base of the squamosal is slightly
different from that of the Didelphiidae and similar to that in the smaller
Dasyuridae and the Peramelidae. As in these, the postzygomatic fora-
men is lacking and the postglenoid and subsquamosal are closely asso-
ciated. The subsquamosal foramen opens laterally just behind the
postglenoid process and the postglenoid foramen antero-ventrally
between the postglenoid process and the superior end of the tympanic.
A thin bony plate bounds the subsquamosal foramen ventrally and
extends from the postglenoid process to a short flat mastoid process of
the squamosal.

Occipitals. — Contrary to the usual condition in marsupials, supra-
occipital, exoccipitals and basioccipital are completely ankylosed in
adult skulls making it impossible to determine their exact boundaries.
In other forms, especially among polyprotodonts, the sutures separating
these bones are usually obvious. The supraoccipital shows no clear
indication of a forward extension to the dorsal surface of the skull and
the median occipital region is smooth without any sharp angle between
the dorsal and occipital surfaces. Laterally a slight ridge rises on each
side suggesting a lambdoid crest at the union of the supraoccipital and
the parietals.

The exoccipitals are small and extend but little beyond the condyles.
Probably they do not meet above the foramen magnum. Small paroc-
cipital processes lie prostrate between the mastoid and the periotic and
do not form projecting points. Their condition is nearly the same as in
Perameles and Myrmecobius but different from other forms examined.
The exoccipitals are perforated by two small condylar foramina near the
apparent line of union with the basioccipital.

The basioccipital is deeply cleft by the inferior lip of the foramen

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 105

magnum. Its ventral surface is smooth or even slightly concave medially
without suggestion of a central ridge. This constitutes another slight
and perhaps unimportant, but still practically exclusive, resemblance to
the peramelids.

Sphenoid bones. — The basisphenoid is decidedly concave on its
exposed surface but has no sharp lateral ridges. It is perforated as usual
by the transverse canal and the entocarotid foramen. The suture
between it and the basioccipital is closed in adults although usually
indicated by a faint transverse ridge. On its superior or cranial surface
it is excavated posteriorly for the reception of the hypophysis but there
is no distinct sella turcica nor clinoid process.

The presphenoid is only slightly exposed ventrally, being largely
inclosed by the thin irregular extensions of the palatine, vomer, and
pterygoids. The suture between it and the basisphenoid, usually so
distinctly open, is but faintly indicated. Within, the vertical part of the
presphenoid is slightly thickened basally but soon becomes exceedingly
thin. The orbitosphenoid occupies a small area between the lacrymal
and the sphenoidal fissure, its exact boundaries being uncertain. The
alisphenoids show no peculiarities beyond those connected with the
unusual fullness of the braincase in the region covered by them. As
stated elsewhere, the sphenoidal fissure and foramen rotundum are
practically confluent just before reaching the exterior but those of the
two sides of the skull are more distinctly separated from each other
than in any other form examined. The alisphenoid bulla forms a
simple subhemispherical capsule similar to that of the didelphids and
dasyurids.

Palatine. — Owing to the very large posterior vacuities the palatine
bones are relatively small. In the median line they send a pair of slender
processes forward to the plane of the middle of the second molar where
they meet similar processes from the maxillaries. Posteriorly they are
bounded by a slight transverse ridge which is perforated on each side
by a small postero-lateral foramen. In close proximity to this foramen is
the slightly larger spheno-palatine foramen piercing the upright plate.
This plate forms the side walls of the posterior part of the nasal passage,
but anterior to the spheno-palatine foramen the walls become dis-
continuous since the turbinals are fenestrated and exposed much farther
caudad than usual and the narrow vomer does not reach the side walls.
Posteriorly the palatines form part of the sloping ridge which supports
the pterygoids. The pterygoid processes are essentially similar to those
of polyprotodonts although they take the form of slender rounded
prongs rather than flattened subtriangular plates.

Vomer. — The vomer is narrow in front and expanded posteriorly

io6 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

enclosing somewhat more of the ventral surface of the presphenoid than
usual. It is not emarginate posteriorly and extends slightly beyond the
posterior nares. Its ventral surface has a low median ridge but this has
no contact with the palatines nor with the maxillaries except anteriorly.

Tympanic. — The tympanic is a simple ring open behind and attached
to the alisphenoid bulla by cartilage. It is wider and heavier than in the
didelphids but is essentially the same as in Australian polyprotodonts.

Periotic. — The petrous part of the periotic is exposed ventrally as a
subtriangular bone between the basioccipital, the tympanic and the
alisphenoid bulla. It is thus almost wholly exposed as in polyprotodonts
and not partly or almost completely covered by the alisphenoid bulla
and the tympanic as in diprotodonts. It is less inflated than in the
dasyurids but projects below the surface of the basioccipital slightly
more than in the didelphids. Internally it shows the usual openings,
including a large floccular fossa. The foramina surrounding it are much
as in the other marsupials except for the so-called carotid canal which
forms a long open slit between the petrous periotic and the basioccipital
which are actually in contact scarcely more than enough to separate
this opening from the jugular foramen behind and the entocarotid fora-
men in front. Such an opening has not been found in any other mar-
supial examined, although in polyprotodonts it may be represented by
the small foramen which Gregory (1910, p. 223) has called the posterior
carotid foramen.

The mastoid is relatively larger than in any other form examined and
is especially characterized by its great lateral extent. Its occipital sur-
face is not more than one-third its lateral and between them there is
only a gentle curve or an exceedingly feeble ridge. The lateral surface
is slightly convex and meets the squamosal evenly. The short flattened
paroccipital process and the mastoid process of the squamosal clasp it
on either side making the entire mastoid region relatively smooth and
continuous with the general outlines of the skull. This lack of promin-
ences is in marked contrast to the condition in the didelphids and is
nearer to that of Antechinomys and other primitive dasyurids. A very
large mastoid foramen is situated at the upper end of the occipital part
of the mastoid separating it from the supraoccipital for a considerable
distance. A mastoid foramen occurs in the didelphids but in adults it
is very small or practically closed, whereas in C&nolestes it is as large in
the adult as in the very small pouch young of Didelphis.

Hyoid. — The hyoid apparatus is of the usual marsupial type. The
basihyal is flat and subcircular with its posterior edge free but its
anterior and lateral boundaries united with the ceratohyals and thyro-
hyals. The ceratohyals are slightly arched forward and only ossified

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 107

basally, becoming cartilaginous and slender in their distal two- thirds.
The thyrohyals are flattened bars standing nearly at right angles to
the general axis and ossified to their tips which are united by a short
cartilage to the osseous wing of the thyroid cartilage.

Vacuities. — The palate is highly fenestrate, its bony floor being
relatively less extensive than in any other living marsupials with the
possible exceptions of some of the peramelids. There are two pairs of
vacuities of constant occurrence and fairly regular form. The anterior
pair extend from the front of the second lateral incisor to the front of
the middle premolar and occupy nearly all the space between the tooth-
rows. They are divided by the median processes of the maxillaries and
premaxillaries which are flattened ventrally to an extent about equalling
one-third the width across the paired vacuities. These vacuities are thus
about equal to the combined length of the two separate pairs found in
peramelids and they are much longer than the single pair of most other
polyprotodonts although they are rather closely approached by Smin-
ihopsis. The posterior vacuities begin at the front of the last premolar
and reach nearly to the transverse ridge bounding the palate. Their
endings are almost exactly at the level of the back of the last molars.
They are divided by a narrow median extension of the maxillaries and a
similar one from the palatine, the two meeting midway of the length of
the vacuities and forming a slight expansion at their union. Laterally
the boundaries of the vacuities are slightly irregular but anteriorly they
are always evenly rounded and the general shape and extent in different
specimens is very constant.

The palatal vacuities are regarded as secondary developments and,
since they vary so widely throughout the marsupial group, it is not
probable that their character in Canolestes has any special significance.
The palate in extinct caenolestids is only imperfectly known.

A prominent vacuity on each side of the face directly above the
infraorbital foramen is one of the most unusual features of the skull of
Caznolestes. It is bounded by the nasal, frontal and maxillary bones
and opens into the large sinus between the naso-turbinal and the
maxillary. Its relations to the overlying dermal tissues are simple and
no glandular or other specialized development is apparent. As noted
by Thomas (1895), a vacuity in this part of the skull is found in other
mammals only among ungulates.

Auditory ossicles. — The malleus has the general features usual in
marsupials. The neck is relatively short, the lamina very thin, and the
processus muscutaris is but feebly indicated. The manubrium is about
two-thirds as long as the processus gracilis and forms a thin flattened
blade with a pointed apex and a slightly falcate outline when seen from

io8 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

the flat side. The processus brevis at the base of the manubrium is
relatively insignificant. The processus gracilis has two thin longitudinal
ridges bounding a long v-shaped groove which seats the tympanic ring.
The extremity of the process is somewhat expanded and obtuse. The
union with the tympanic is close and considerable care is required in
forcing the parts away from each other. The malleus differs from that
of Perameles (P. bougainvillei} principally in its shorter neck; from that
of Didelphis it is distinguished by its generally frailer construction and
its closer union with the tympanic.

The incus is practically identical with that of Perameles, differing
mainly in the processus brevis which is somewhat compressed and
truncate instead of long and conical. The processus longus is very
slender distad of the sharp right angle at which it is bent. The Sylvian
apophysis is elliptical. The incus in adults is rather firmly attached to
the malleus and does not separate except after slight pressure.

The stapes is strictly columelliform without trace of an opening and
there is not even an appreciable difference in its diameter from the head
to the base. The head is very small and the base relatively large and
elliptical. In adults, the stapes is well ossified.

The auditory ossicles of Ccenolestes are clearly of a type closely
approaching that of Perameles, which was regarded by Doran (1879) as
the lowest among marsupials. They also show resemblance to some
of the dasyures, which, as described by Doran, have the stapes columel-
liform. Whether this simple stapes is in truth a primitive character or
not, it serves at least to distinguish Ccenolestes sharply from the didel-
phids, in which the stapes is always bicrurate. So far as the auditory
ossicles are concerned, therefore, Ccenolestes resembles Australian rather
than American forms and especially approaches the peramelids and
dasyurids.

Mandible. — The mandible is slender and relatively straight ante-
riorly, its greatest curvature being just below the base of the coronoid
process. The symphysis is long but weak. The rami have exceptionally
little divergence. The coronoid is large, broad, and nearly upright. Its
anterior border is gently curved but approaches the vertical more closely
than in any polyprotodonts. The broad masseteric fossa, is rather shal-
low and is indistinctly divided into an anterior and a posterior portion
by a weak elevation between the tips of the coronoid and the condyle.
Near the lower edge of the masseteric fossa is a tiny foramen, indistinct
in some specimens, but obvious in others. A similar foramen is found
in many Australian diprotodonts but does not occur in polyprotodonts.
Its presence in Abderites and Garzonia has been noted by Ameghino and
Sinclair. Mental foramina of the usual character are present, a larger

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 109

one below the posterior root of the middle premolar and a smaller below
the anterior root of the first molar.

On the lingual side of the mandible the dental foramen is situated as
usual at the base of the angular process. In some specimens a shallow
groove runs along the inner side of the proximal half of the jaw but in
no case does its position or character seem to warrant any assumption
that it is a true Meckelian groove.

The angular process is moderately inflected and subtriangular in
shape with a fairly wide base and a rounded but not decidedly obtuse
tip. The inflection is slightly less than in the majority of polyprotodonts
and the process is not produced into a slender prong like that in the
smaller dasyurids. The condyle is relatively high in position, slightly
more than halfway from the angle to the tip of the coronoid and extend-
ing well beyond a line drawn between them. In this respect, therefore,
it is more like Phascologale and other dasyurids than like the didelphids
in which the condyle is seated on a relatively short base.

The length of the bony lower jaw is proportionately less than in most
polyprotodonts and this is especially true if it is considered in relation to
the upper jaw. In two specimens of nearly equal size, the length of the
bony part of the jaw in Ccenolestes is 77 per cent of the total length (in-
cluding terminal incisors) and that in Phascologale is 95 per cent.

Owing to the wide variation in the lower jaws of marsupials, it is
difficult to draw any conclusions from the findings in regard to Cceno-
lestes. Miss Dederer (1909) has regarded the jaw of Ccenolestes as indic-
ative of polyprotodont affinity, believing it to be very similar especially
to Antechinomys and Sminthopsis. It differs from these in many respects,
as in the broader more upright coronoid, the less inflected and less
attenuated angle, the less divergent rami, and in the presence of a
foramen in the masseteric fossa, this last being a diprotodont character.

Among the didelphids, the mandible shows considerable variation
and comparison is difficult. It might be said that as respects the angle,
C&nolestes is somewhat intermediate between Didelphis and Philander
and as respects other characters it shows no especial indications of
relationship to the didelphids more than to other marsupials.

SUMMARY OF CRANIAL CHARACTERS.

The great range of variation among both polyprotodonts and
diprotodonts leaves relatively few characters that are diagnostic of the
two groups. Without a complete representation of all known forms it
can only be said that certain characters are regarded as polyprotodont
or diprotodont because they are mutually exclusive in the majority of
forms examined. On this basis, it is clear that the skull of C&nolestes

no FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

shows more polyprotodont than diprotodont characters. These are as
follows:

1. Anterior root of zygoma only slightly in advance of end of
toothrow.

2. Posterior root of zygoma not inflated.

3. Tympanic narrow and annular.

4. Petrous part of periotic largely exposed ventrally.

5. Alisphenoid bulla not fused with paroccipital process.

6. Foramina of squamosal arranged in general as in Dasyuridae.
Practically all of these characters relate to the otic bones or to the

backward extension of the palate with relation to the zygomata. In the
extinct Pal&othentes, a close relative of C&nolestes, the zygomata and
palate are practically as in diprotodonts, so the polyprotodont char-
acters of the skull in casnolestids are reduced to those of the tympanic
region. The additional characters supposed to indicate polyprotodont
affinity and cited by Miss Dederer (1909) and Broom (1911) are found
in the smaller diprotodonts, as Dromicia and Tarsipes. Hence their
importance is doubtful.

A few characters are at least in agreement with those of some
diprotodonts, and with no polyprotodonts so far as known. These are
the following:

1. Upper anterior ethmo-turbinal short and enclosing but little of
the maxillo-turbinal.

2. Maxillo-turbinal arising relatively high up on the nasal wall.

3. Glenoid fossa relatively long and with a small postglenoid
process which is not extended transversely.

4. A small foramen in masseteric fossa of mandible.
Distinctions in cranial characters between Ccenolestes and the

didelphids are numerous; in fact the only characters in common are of
quite a general nature.1 On the other hand, resemblances to Australian
forms, especially peramelids, are abundant. In the following respects
there is agreement with some or all Peramelidae or Dasyuridae and
disagreement with all Didelphyidae:

1. Rostrum slender and elongate.

2. Palatine processes of premaxillae not extending full length of
incisive foramina.

3. Basioccipital concave medially.

4. Maxillo-premaxillary suture nearly upright anteriorly.

5. Cribriform plate approaching the vertical.

6. Fronto-maxillary suture broad.

1 The mastoid foramen may be an exclusive resemblance, but in the absence of
immature specimens I am unable to learn whether or not it occurs in dasyurids.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. in

7. Sagittal and lambdoid crests absent.

8. Postorbital processes absent.

9. Paroccipital and mastoid processes small and prostrate.

10. Lacrymal canal with a single opening.

11. Postzygomatic foramen lacking.

1 2 . Stapes colurnellif orm.

The first five of these characters are exclusive resemblances to the
peramelids while the others are common to the peramelids and certain
of the dasyurids. The total number of resemblances to the peramelids
therefore, is very large. Moreover, important distinctions between
C&nolestes and Perameles are comparatively few, being confined mainly
to the region between the sphenoidal fissures and to the proportions of
the processes of the mandible.

In a number of respects the skull of Canolesles differs from all other
living marsupials. The most important of these are as follows:

1. A preorbital vacuity between the nasals, maxillary, and frontal.

2. A large mastoid foramen persisting in adults.

3. A long narrow carotid canal between the petrous periotic and the
basioccipital.

4. Floor of braincase very wide between sphenoidal fissures.

5. Mastoid large and broadly exposed laterally (approached by
peramelids),

6. Olfactory fossa relatively large and wide.

DENTITION.

Plate XVIII.

The dentition and the cranium of Ccenolestes have been known since
the original discovery of the animal and have been variously described,
figured, and subjected to comment by different authors, notably
Thomas, Bensley, Sinclair, Dederer, and Gregory. The diprotodont
modification of its lower incisors combined with a polyprotodont upper
incisor formula and quadrate bunolophodont molars at once furnished
conditions not found elsewhere among living marsupials and not
typically representing either the diprotodont or the polyprotodont
group. The teeth have been well described in considerable detail by
Thomas (1895), but his specimen was somewhat affected by wear,
which obscured a few important points. For this reason and for the sake
of completeness and convenience of discussion full description of the
teeth is given in the following pages. Unfortunately it is still impossible
to give any information as to the succession of the teeth, since all the
specimens at hand are fully adult.

ii2 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

NUMBER AND HOMOLOGIES OF TEETH.

The teeth are 46 to 48 in number. In the upper jaw there are one
pair of terminal incisors and three pairs of lateral ones, a pair of sharp,
well-differentiated canines, three pairs of more or less triconodont pre-
molars, and four pairs of molars, the first and second of which are
quadrate and the third and fourth subtriangular. In the lower jaw there
is a pair of long terminal chisel-like incisors resembling those of the
most pronounced diprotodont dentitions. Behind these in lateral
position are four and sometimes five pairs of small unicuspids, practically
undifferentiated from each other, which are interpreted as two (or three)
incisors, one canine, and one premolar. Next are two pairs of double-
rooted premolars and four pairs of molars. The dentition, therefore, is
a rather highly modified one but retains approximate numerical agree-
ment (at least in certain specimens) with the polyprotodont dentition
usually regarded as generalized. With the exception of the didelphids,
which have five upper incisors, and Myrmecobius, which has super-
numerary molars, no living polyprotodont has more teeth than C&no-
lestes. The formula may be written:

I. ~ C.j;Pm.|;M. -^5x2=46-48.

This is essentially the classification made by Thomas (1895) who
divided the four unicuspids behind the median incisor into two incisors,
one canine, and one premolar. "Any other determination, " he says,
"would involve the presence of four incisors or four premolars, each
equally unlikely." It now appears that, although the normal number of
lower incisors is three as decided by Thomas, the presence of four would
by no means have been unlikely. The number of lower unicuspids or
intermediate teeth is normally four, but a variation in which there are
five is not infrequent. In one specimen (Field Museum No. 18603),
there are five on the right side of the jaw and four on the left. Bensley
(1903, PI. 5, fig. 38) has figured a specimen belonging to the British
Museum in which there are five on each side of the jaw. Others, so far
as examined, have four unicuspids on each side, but since the total
number of specimens is very small the percentage of variation is high.
It is. possible, therefore, that these aberrant specimens represent a
condition once normal and in which the lower antemolar formula was
equal to that of the didelphids and in excess of other living marsupials.
This is further suggested by the occurrence of the same lower antemolar
formula in the extinct genus Halmarhiphus, regarded by Sinclair as the
direct ancestor of C&nolestes. A closely allied form, Garzonia, has as

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 113

many as six unicuspids in the lower jaw, giving it a total antemolar
formula of nine, exceeding that of any living mammal. While this
formula, as observed in only one specimen, may be abnormal, as sug-
gested by Sinclair, it is evident that the formula of C&nolestes, large
though it be, has already undergone a reduction. A further and gradual
reduction may be traced in Australian diprotodonts to those in which
only the median or terminal pair of incisors and one pair of differentiated
premolars remain, two pairs of antemolar teeth in all, as in Phascolarctos
and Phascolomys.

Representatives in this series with the number of antemolar
teeth in each would be as follows: Garzonia, g; Halmarhiphus, 8;
Ccenokstes, 7-8; Phalanger, 6-7 -,1 Acr abates, 5; Dromicia, 4-5; Dis-
taechurus, 4; Trichosurus, 3-4; Macropus, 3; Phascolarctos, 2. When the
form and character of the teeth as well as their number is considered,
the probability that the series is an expression of homologies seems very
great. Among the living forms of the Phalangeridae the reduction is
going on at present, as indicated by the great instability of the number
of the "small intermediate teeth" in various species. This was shown
especially in the genus Phalanger by Jentink (1885) who examined a
considerable series and found great variation. The subject has been
further elaborated by Bateson (1894). In the extinct forms Pal&othentes,
Acdestis, and Callomenus, a reduction also has occurred, these having
only five or six antemolar teeth. Hence, C&nolesles is more primitive
than these. In fact, so far as mere number of teeth is concerned, Ccsno-
lestes and the allied extinct forms like Halmarhiphus and Garzonia are
almost as primitive as the didelphids. Assuming their common ancestry,
it is necessary to believe that the diprotodont modification preceded any
numerical reduction. Moreover, it is not impossible that all living and
extinct polyprotodonts are more reduced than these primitive dipro-
todonts. The case of Garzonia with its nine lower antemolar teeth is of
interest in this connection. At least five of these must be regarded as
incisors unless reduplication with no reference to homologies be assumed.
This is a larger number of lower incisors than is possessed by any known
polyprotodont. Further, if the specialized median incisor is not the first
but the second as shown by embryology in the macropods (Woodward
1 893) , we must assume the ancestral formula to be six. Such an assump-
tion is further justified by Woodward's (1. c.) discovery of vestiges of six
upper incisors in the macropods and by the occurrence as an abnormality
of six upper incisors in Didelphis (Bateson 1894, p. 247; Allen 1901, p.
158). If there were six in the upper jaw there may well have been six
in the lower, although of course it is not yet known that the casnolestids
1 In abnormal cases only.

ii4 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

agree with the macropods in the ontogeny of the upper jaw. The dental
formula of the primitive diprotodont would then have been as follows:

6 i

The earliest known mammals, whether they were pro-marsupials or
not, furnish no indication that such a large mammalian incisor formula
ever existed and it must be admitted that the evidence in favor of it is
little more than suggestive. Still, if early mammals only were con-
sidered, it would be almost as difficult to believe in an incisor formula of
$ as one of I, for the usual number of lower incisors in the Triconodonta
and Trituberculata was four. Both Thomas (1888) and Winge (1895)
have presented evidence that the primitive formula was at least f and
the case for it must be regarded as considerably stronger than that for
one of i so to that extent the testimony of the little known "pro-
marsupials " is weakened. In fact, in this as in some other respects, it
is evident that the long known Mesozoic mammals are extremely liable
to misinterpretation. Evidence is accumulating to show that even
these early triconodonts were not altogether generalized and many
connectant forms are needed before the main mammalian stem will be
definitely revealed. The highly specialized nature of the dentition in
some of the theriodont reptiles, including both secodont and crushing
types and even showing a mammalian succession, suggest that some
basic lines of the general evolution of mammalian teeth may have been
laid down in these early reptiles and, although lost in the triconodonts,
persisted in still later forms the immediate ancestors of which are
unknown. Although most of these theriodonts had only four upper
incisors on a side, there were some (e. g. Pristerognathus) which had as
many as six. Hence a mammal with six is theoretically open to no serious
objection. In any case it is clear that as early as Miocene times caeno-
lestids were at least as primitive as polyprotodonts in the number of the
teeth in the lower jaw. The number in the upper jaw is one less than in
the didelphids and certain of the peramelids, which have five upper
incisors, most other polyprotodonts having four.

The lateral lower incisors of Ccenolestes are small and while not wholly
functionless are apparently approaching that condition, so variations
in their number are more to be expected than would be the case with
fully functional teeth. It is possible that the occasional appearance of
an extra unicuspid has no reference to ancestral conditions and therefore
that all the specimens of the extinct forms Halmarhiphus and Garzonia
are abnormal. Evidently this would be the conclusion of Bateson (1894)
who places no confidence in "reversion" or ancestral influences as the

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 115

explanation of such variations. However, he regards "the variability of
a form as much a part of its specific character as any other feature of its
organization." Without dissenting from this, it may be added that
specific and especially generic characters (regardless of how they were
initiated or established) are themselves the result or sequence of ances-
tral conditions. Whether or not characters are "presence or absence"
characters, that is, "meristic" rather than "substantive," matters
little. In specialized dentitions the loss of a series of several teeth can
plainly be traced from species or genera in which the whole series is
constantly present to those in which all of it is constantly absent. In
such a series, the intermediate forms, or those in which the process of
reduction is going on, are notoriously variable. How else, it may be
asked, could a reduction take place, especially if discontinuous variation
be assumed? No better example of a variable intermediate type could
be had than Phalanger, which is treated by Bateson at considerable
length. In the artificial series from the extinct casnolestid Garzonia
to the modern Phascolarctos the lower antemolar teeth are reduced from
nine to two and the "small intermediate" teeth from seven to none.
Phalanger lies between the two extremes. Among 76 specimens of one
species (P. orientalis) examined (Bateson, p. 253), 57 had three inter-
mediate teeth on each side of the mandible. The remaining 19 varied
from those having only one on each side to one specimen having five on
one side and four on the other. Of these, 13 had at least three inter-
mediate teeth on one side of the jaw. This variability is what might
be expected as the result of interbreeding after the occurrence of a
discontinuous variation in certain individuals. Or, as seems more
probable to a taxonomist and comparative anatomist, it may be the
result of the interbreeding of individuals which have in varying degrees
gradually developed tendencies toward the loss of certain teeth. Such
tendencies are illustrated by the familiar example of the tardily devel-
oped and frequently absent third molar or wisdom tooth of Homo. The
vestigial incisors (many of which are partly calcified) which have been
found in all marsupials investigated by embryologists would doubtless
be admitted by everyone as indications of ancestral conditions. These
vestigial incisors are shown by Woodward and Wilson and Hill to be,
at least in some cases, not members of a pre-existing series (lacteals or
prelacteals) but persistent rudiments of suppressed members of the
existing series. Hence it seems not improbable that there may be cases
in which these vestiges arise in place and become functional as "super-
numerary" incisors. An embryological study of an extremely variable
form like Phalanger should yield interesting results bearing on this
point.

n6 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Therefore, while it is undeniable that some cases of supernumerary
teeth are teratological with a basis that may be largely physiological and
mechanical, it cannot be admitted that there are no cases of which
phylogenetic influence is the chief determining factor. The objectionable
term reversion need not be applied to them but it might be justified to
refer to them as examples of persistence of ancestral characters. Aside
from what has been said above, it is evident to anyone familiar with the
dentition of many species of mammals, that cases are innumerable in
which a nonfunctional or slightly functional tooth is present in one
species and normally absent but occasionally present in a closely related
species. Among bats and rodents instances are particularly abundant.
The homology of such a tooth is invariably too clear to be questioned
for a moment.

That the fifth lower unicuspid sometimes found in Ccenolestes is
evidence of an ancestor in which this tooth was normally developed may
not now be proved conclusively but the evidence from paleontology,
from embryology, and from phylogenetic series is sufficient to render it
exceedingly probable. That the primitive diprotodont incisor formula
was at least £ is also strongly indicated. Hence the main points in the
present connection seem fairly certain, that is, that primitive diproto-
donts had numerous teeth and that the diprotodont modification may
have arisen prior to any reduction from the primitive number.

UPPER INCISORS.

Description. — The median incisors are set near the end of the pre-
maxillae in nearly terminal position, only a slight shelf of bone extending
in front of them. Their alveoli are more elevated than those of the suc-
ceeding teeth and they are separated from the next incisors by a slight
space. They are inclined forward slightly away from the perpendicular
and their points exceed the lateral incisors. They are separated at the
base sufficiently to permit the insertion of a needle or bristle but their
points are closely in contact. Their cutting edges are beveled and receive
the extreme tips of the lower incisors. They are closely similar in shape
to corresponding teeth in the Australian diprotodonts, especially some of
the smaller macropods. They have relatively broad faces which suggest
the undifferentiated median incisors of Perameles quite as much as they
do those of the didelphids. The median upper incisors of typical poly-
protodonts (dasyures and didelphids) are usually slender and peglike
and in most cases the tips are not in contact although occasional speci-
mens may be found having some resemblance to those of C&nolestes.

The three succeeding incisors (second, third, and fourth) are wholly
lateral in position and the series on one side is almost exactly parallel

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 117

with that on the other, a condition approached in no polyprotodonts
except Myrmecobius and Perameles. The second and third are in contact
with each other, the second being somewhat larger than the third.
They are laterally compressed into relatively broad blades, their cutting
edges are truncate and their bases narrow. The blades, which are dis-
tinctly hatchet-shaped, are somewhat produced anteriorly. These teeth,
although small, are completely functional since they engage the sharp,
beveled, outer edges of the median lower incisors with which they have a
very effective shearing action. Similar action is seen between the
incisors of some of the macropods, and especially of certain of the
Phalangeridae, as for example, Dactylopsila, in which it is very pro-
nounced. Somewhat similar relations occur in Perameles.

The fourth upper incisor is abruptly smaller than the preceding ones
and is usually separated from them by a slight space. In superficial
view it appears to be pointed but closer examination shows that this is
due to a partial forward rotation which has elevated the posterior angle
of its cutting edge. This cutting edge is thus inclined forward so that
it is parallel to the unbeveled edge of the lower incisor and in this rela-
tion it is possible for it to be slightly functional.

Discussion. — All the upper incisors are definitely adapted for relation
with the long procumbent lower incisors. The result is the modification
of the lateral upper incisors from simple seizing or grasping teeth to
cutting or shearing blades. Such a modification, as shown byBensley,is
easily derived from the conical tooth by lateral compression accom-
panied by the movement of its anterior edge to a horizontal position and
its posterior edge to a vertical one continuous with the root. The
original apex of the tooth thus becomes the posterior angle. In less
marked degree this change has been noted and figured in Myrmecobius
and Perameles by Bensley (1903, p. 105) who speaks of it as a "curious
appearance, due to a subcaniniform modification of their tips," and
makes a rather obscure comparison with a condition in the degenerate
teeth of the Madagascan viverrine Eupleres.

In viewing these teeth comparatively in Myrmecobius, Perameles,
and Ccsnolestes, one is at once impressed with the practical identity in
form and function in all three forms and it is evident that the modifica-
tion of the lateral upper incisors is correlated with the elongation and
specialization of the median lower incisors. In other words, this form
of lateral upper incisor is an indication of beginning diprotodonty and
is only found in the Australian diprotodonts and forms conceivably
ancestral to them. This subject is discussed elsewhere. It may only
be said further that the reciprocal relations of the median lower incisors
and the lateral upper ones are exceedingly well-developed in C&nolestes.

n8 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

In fact they appear so efficient for cutting that their relation to the
animal's insectivorous habits is not obvious. Functionally, they are
quite as well adapted for cutting vegetation as those of strictly her-
bivorous forms and the possibility that they may have passed through
a herbivorous or partly herbivorous stage is to be considered.

LOWER INCISORS.

The median pair of lower incisors are long, slender, and only slightly
curved, merely continuing the inferior outline of the mandibular ramus.
They are strongly beveled on the upper side anteriorly and their outer
edges are sharp and almost bladelike. They are separated from each
other by a considerable space and in this and in the other foregoing
respects they are closely similar to the lower incisors in the macropods.
Their inner edges are not highly developed for cutting and seem to
show a slight tendency to imitate even this character of the macropods.
The lower incisors of most phalangers are close-set and decidedly up-
curved. Among those examined, the only form showing general resem-
blance to C&nolestes is Petauroides.

Of the small unicuspidate, single-rooted, "intermediate" teeth
situated between the terminal incisors and the two-rooted premolars,
there are two, and in some cases three, which are to be regarded as
incisors. They are indistinguishable by size or shape from the supposed
canines and anterior premolars except that the foremost one shows an
extreme of pronation which is slightly less pronounced in the succeeding
ones. Their crowns are turned forward and lie overlapping each other,
the first being in contact with the base of the median incisor and the
second with the upper exposed surface of the root of the first. Thus the
original apex of each tooth is directed forward and the posterior side of
the tooth including a part of the exposed root has come into horizontal
position. This arrangement appears to be unique among living mar-
supials although slight approaches to it are seen in some of the Phalan-
geridae. It is duplicated in every feature, however, in the extinct
Patagonian caenolestids.

CANINES.

The upper canines in C&nolestes obscurus and apparently in most
specimens of C. Juliginosus are simple slender prongs slightly more
curved but otherwise similar to the canines of other omnivorous or
carnivorous marsupials. In the male they are much more prolonged
than in the female, the exposed part measuring 2.5 mm. in a male and
only 1.3 mm. in a female. In a specimen of C. fuliginosus figured by
Bensley (1903, pi. V, fig. 38) the canine is shown to have a deep lateral

MAY, 1921. AMERICAN MARSUPIAL, CJENOLESTES — OSGOOD. 119

groove extending from the middle of its outer surf ace to the broad single
root. In the closely allied genus Orolestes recently described by Thomas,
the canine is said to be double-rooted, and shaped like a premolar.
Hence we have in the caenolestids now living a graded series showing a
change in the character of the canines from a two-rooted premolarif orm
condition to a single-rooted caniniform condition. That this is the
direction of the change and not vice versa seems fairly clear. In the first
place, the two-rooted condition is the more primitive. Moreover,
although it might be reacquired through "retrogression," as held by
Bensley, it is very unlikely that the reversal would be represented by
these particular stages. Thus, while the grooved root is the obvious
intermediate state in change from a double to a single root, it is scarcely
what would be expected if the direction of change were reversed. The
logical intermediate stage, if we were dealing with a retrogression from
a single to a double-rooted condition, would be a lengthening of the
crown with possible appearance of antero-posterior cusps before tend-
encies to division of the root began. Hence it seems that the change in
this case is toward rather than away from a long single-rooted piercing
type of canine. It is plainly a very recent specialization and is quite in
accord with the animal's present insectivorous and predaceous habits.
Of still further interest in this connection is the fact that this change
has progressed farthest in the northernmost form, that is, the one
presumably most distant from the parent stock.

The lower canines have no intrinsically distinctive characters but
are quite like the incisors preceding them upon which they lie in semi-
prostrate position.

UPPER PREMOLARS.

The anterior premolar is small, greatly compressed laterally, and its
crown In side view is nearly triangular, its anterior outline being prac-
tically straight and its posterior slightly concave. In wholly unworn
condition, traces of a posterior cusp are evident but none of an anterior
one. It is situated about midway between the canine and the middle
premolar and is separated from each of these by a space equal to approxi-
mately one and one-half times its own length.

The middle premolar is somewhat larger than the anterior, but is
similarly compressed. Its posterior outline is concave and sweeps back-
ward to a definite posterior cusp. It is separated from the posterior
premolar by a slight space about half as long as that between it and the
anterior premolar.

The posterior premolar, which is probably a replacing tooth, is
separated by a slight space from the middle premolar. It is situated

120 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

somewhat obliquely with its small anterior cusp slightly internal to its
high triangular posterior blade. Posteriorly it stands closely against the
first molar and, although its principal cusp is higher than any of the
molars, it is functionally a member of the molar series and fairly well
separated from the anterior teeth as in diprotodonts generally. It is
slightly thickened posteriorly but has no internal cusps although the
cingulum is well-developed. Of its two roots, the anterior one beneath
the small antero-internal cusp is the longer and heavier.

UPPER MOLARS.

Description. — The first and second molars are essentially alike in
structure and size, the first being only very slightly longer relative to its
width. They are quadrate and have four principal cusps, two inner and
two outer, i. e., paracone, metacone, protocone, and hypocone.1 At the
inner base of the metacone is a small but high and very distinct inter-
mediate conule which in worn teeth is sometimes obliterated. In lateral
view, the paracone and metacone appear as nearly equilateral triangles
rising from the cingulum. The angle between these two cusps is more
acute in the first molar than in the second and both cusps of the first are
slightly higher than those of the second. The protocone is well developed
and more thickened but is somewhat lower than the outer cusps. At its
antero-internal base it is distinctly bulging. A distinct internal cingulum
bounds its base and extends to the anterior base of the hypocone. The
hypocone is about half as high as the protocone and its apex scarcely
reaches the height of the intermediate conule, so it is much the smallest
of the four principal cusps. Protocone and paracone are connected
anteriorly by a commissure which forms the front boundary of the deep
central depression of the tooth. A low commissure connects the proto-
cone and the hypocone and near the middle of this is a slight ridge some-
what imperfectly connecting the posterior base of the protocone with
the intermediate conule. Between the hypocone and the intermediate
conule there is an open channel to the back of the tooth.

The third upper molar is subtriangular and the hypocone is prac-
tically undeveloped, although a very slight shelf exists between and
slightly posterior to the intermediate conule and the protocone. The
first, second, and third molars have three strong roots each and these
are situated respectively beneath the protocone, paracone, and meta-
cone. The external cingulum in all these teeth is well-marked, but shows
no differentiation into styles. The fourth molar is very small, scarcely

1 These terms and others are used without reference to homologies to designate
the cusps to which they have been commonly applied in connection with the Cope-
Osborn tritubercular theory.

MAY, 1921. AMERICAN MARSUPIAL, CJENOLESTES — OSGOOD. 121

one-sixth as large as the third, and it is turned slightly inward from the
toothrow. It is roughly triangular and carries three small cusps appar-
ently representing paracone, protocone, and metacone. They are
scarcely higher than the commissures which connect them and inclose a
central depression. There is still retained an external cingulum, a
minute anterior shelf, and two roots, one of which is larger and obviously
formed by coalescence. This tooth is plainly undergoing reduction and
is in about the same stage as the fourth molar of Dromicia. In Acrobates
and Distaechurus it has quite disappeared.

Discussion. — The upper molars of Ccenolestes show an obvious resem-
blance to those of the diprotodont family Phalangeridae. Two principal
types of molars are found in this family: (i) the bunoid and (2) the
selenoid, respectively characterizing the two subfamilies Phalangerinae
and Phascolarctinae. In general, the bunoid group are without styles
or intermediate conules while the selenoid group usually have at least
traces of both styles and conules, but in these respects there are excep-
tions in both groups.1 Several theories have been advanced as to the
origin and mutual relationships of the two groups. Winge considered
them derivative the one from the other and regarded Phascolarctos
ancestral to the bunoid Phalangerinae with Pseudochirus as a connecting
form. On the other hand, students of placental molars have usually
inclined to the belief that, among ungulates especially, bunoid types were
ancestral to selenoid. Bensley, however, comes to the conclusion that
"the molars of the Phascolarctinae and Phalangerinae have been de-
rived by a divergent evolution from a common insectivorous secodont
type," that is, independently from a tritubercular or post-tritubercular
type similar to that seen in the Peramelidae. This divergent evolution
took place, according to Bensley, as follows :

1. The selenoid type evolved by the simple addition of a hypocone
and the reduction of the styles, no change in the shape of the principal
cusps being required since they are already selenoid in the tritubercular
molar. This method was recognized also by Winge. For the production
of the molar of Pseudochirus the addition of the intermediate conules was
further required and that these were secondary elements which arose
de novo seems to have been taken for granted, although, as shown later,
there is some evidence that they are not.

2. The bunoid type evolved also by reduction of styles but accom-
panied by a lowering of the principal cusps and a gradual change in their
form from crescentic to subconical. These and other changes involved

1 In the Phalangerinae, traces of a metaconule occur in Petaurus and a large
persistent style is present in Dactylopsila. On the other hand, the Phascolarctinae
include some species of Pseudochirus in which styles are lost and Phascolarctos in
which conules are scarcely evident.

122 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

with them are foreshadowed in the molar of Philander. As in the
selenoid group the quadrate form was produced by the addition of a
hypocone.

It is in their relation to these theories of the origin of bunodont and
selenodont molars and also to some of the still more debatable theories
of the general evolution of mammalian teeth that the molars of C&no-
lestes must be viewed. The essential difference between the bunoid and
selenoid groups of Phalangeridae is in the shape of the outer labial cusps
of the upper molars, these being in the one convex in their external
aspect and in the other concave. Judged by this character alone, the
molar of C&nolestes at once classifies itself with those of the bunoid
Phalangerinae. Among these, however, it stands out as a relatively
primitive type. This is evidenced by (a) the high lanceolate labial cusps,
(b) the difference in height between the labial and lingual cusps, (c) the
practical absence of a hypocone in the third molar (d) the presence of a
distinct metaconule in the first, second and third molars.1 When, in
addition to these characters of the molars themselves, the primitive
nature of the antemolar formula is considered, it is clear that there are
strong reasons for considering the molar of Canolestes as possibly
prototypal to those of the Phalangerinae. It is of course conceivable that
it is now becoming specialized in an insectivorous direction secondary
to a previous bunodont stage. This possibility is suggested by the
cursorial adaptations in the skeleton of Canolestes, but, as shown by the
Peramelidae, these may accompany relatively slight modifications of
molar structure.

Assuming that the molar of C&nolestes represents a late stage in the
development of bunodonty, it becomes difficult to harmonize it with
current theories. Such a molar would scarcely be produced by the lower-
ing and thickening of cusps which appear in Philander. These changes
are only relatively slight deviations from the conditions prevailing
among all Didelphiidae. They are seen to some extent also in Didelphis
where, as in Philander, they are probably related to habits which are
more f rugivorous and omnivorous than those of Peramys and Marmosa.
It is conceivable that they might have led either to the bunoid or the
selenoid section of the Phalangeridae or to some of the conditions seen in
the Peramelidae. But they cannot be accepted as any indication of the
genesis of the quadrate buno-lophodont molar of C&nolestes.

How, then, was this molar derived? Considerable evidence points
to a possible evolution by persistence of stylar elements rather than by
their reduction and total elimination. The theories of Bensley and
others, with the exception of Winge (1882), invariably give a minor role

1 The interpretation of the metaconule as primitive is discussed on a later page.

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 123

to the styles, although they are so highly developed in the tritubercular
molar. They are supposed to grow up from the cingulum and then to
shrink back into it to complete extinction. In some cases this may
occur, especially in placentals, but in the marsupials there is much to
show that the styles have played an important part and that they may
be the principal elements preserved in the outer cusps of the molars of
C&nolestes and the bunoid Phalangerinae. In some cases they may
have wholly supplanted the paracone and metacone and in others they
may have fused wholly or partly with them, the resulting cusp usually
preserving more of the character of the style than of the so-called
primary cusps. This hypothesis favors Winge's theory that the styles
are the original elements of the molar crown persistent from the tri-
conodont stage, but it is only facts bearing on the evolution of the
bunoid molar from the tritubercular that are of present importance.
Some of these are the following:

1. The stylar elements of the molar crown are deep-seated struc-
tures, being recognizable in many early mammals and in developmental
stages of recent ones.

2. They occupy the outer or labial side of the teeth which is phylo-
genetically older than the inner or lingual.

3. They are well developed in polyprotodont marsupials and show
a high degree of consistency in general relations (Bensley, 1906).

4. Morphologically they are bunoid, at least conical or compressed
conical in section.

5. The cusps of the trigon in the tritubercular molar are selenoid,
essentially alike and essentially different from the styles.

6. Hence the crown of this molar consists of two elements one of
which is fundamentally bunoid and probably primitive while the other
is selenoid and secondary.

7. The styles are situated directly over the buccal roots and in line
with the primitive longitudinal axes of the premolars in the identical
position of the outer cusps of the quadrate bunoid molar.

8. On the contrary, the paracone and metacone are midway between
the lingual and buccal roots and in order to become the outer cusps in
the bunoid molar would require not only a change in form but a migra-
tion from their position.

9. Aside from the confusing factors introduced by minor adapta-
tions, therefore, the general relations in the tritubercular molar are
favorable for positive participation of stylar elements in the formation of
the bunoid molar.

With these broader facts in view, evidence may be sought as to
detailed methods of change other than that of downward reduction

124 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

and elimination of styles. The high bladelike cusps in the molar of
Ccenolestes suggest an evolution along lines similar to those which have
produced the carnivorous dentitions. In the Dasyuridae, the carnivorous
evolution obviously has proceeded by lateral compression and fusion of
styles and cusps. In this process it is difficult in all cases to demonstrate
to what extent one element dominates the other, but the styles are
quite as consistent as the so-called cusps of the trigon. In Sarcophilus,
the style dominates the paracone in the first and second molars and the
two are evenly divided in the third. Metacone and style, on the other
hand, are evenly divided in the first molar, while the metacone dominates
in the third. The character of the outer aspect of the teeth in all cases
seems determined by that of the styles. In Thylacmus, stylar elements
persist in reduced condition anteriorly and posteriorly but in the inter-
vening region have completely fused with the metacone, doubtless some-
what after the manner indicated in Sarcophilus. Even after these
extensive changes, the resulting cusp, which is derived principally from
the metacone, is situated in the same relation to the roots as the meta-
cone of the tritubercular molar of the didelphids and dasyurids. In
the carnivorous evolution, therefore, although the tendency to con-
solidation of cusps is very great, the styles have been reduced only
when their corresponding cusps were reduced and in cases of fusion they
have given their character to a large part of the combined product.

In the formation of the molar of Ccenolestes with its decidedly convex
labial surfaces it seems probable that fusion of styles has played quite
as large a part as in carnivorous forms. In the bunoid Phalangerinae it
is possible that fusion and reduction may have been combined in some
cases. As compared with the upper molars of other living marsupials,
those of Ccenolestes are unique in the possession of a marked undiffer-
entiated external cingulum. It forms a continuous faintly crenulate
ledge along the bases of all the molars. A similar structure is found in
various degrees of perfection in the extinct casnolestids but nothing
exactly comparable occurs in modern forms. It is not necessarily
evidence of reduction of styles and may easily have been developed
independently. In fact an exactly analogous structure is found at the
inner base of the protocone of the first upper molar and cingula of the
same character appear on the lower molars to which no suspicion of
reduced styles attaches. The irregular swelling at the bases of the
upper molars in primitive Phalangeridae, as Dromicia and Petaurus, is
regarded by Bensley as a vestigial external cingulum which "appears to
be equivalent to that bearing the external styles in polyprotodont forms."
In a certain sense this is apparent enough, but I believe that at most
only the minor stylar elements are represented and that the principal

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 125

ones (styles 6 and c} are no more reduced than they are in the anterior
molars of Sarcophilus in which there is a similar slight basal ridge com-
bined with large obvious styles.

Evidence of combined fusion and reduction is seen in the first upper
molar of Dactylopsila. In this the quadrate form is complete so far as the
principal cusps are concerned, but there remains on the antero-external
border of the tooth a single well-marked style, which is well distinguished
from the paracone. Apparently the other styles have fused, but this
one has changed only little and that by reduction. It is to be noted also
that this style retains a primitive relation to the root. Most interesting
is the fact that the antero-external surface of the paracone, that is, the
part opposite the style, is concave or semi-selenoid while its postero-
external surface is convex like that of the succeeding molars. In other
words, it seems that where fusion takes place, the outer surface of the
style is added to the cusp making it convex and where fusion does not
take place, the outer surface remains concave as at first.

To produce a bunoid molar from a tritubercular one by this method
requires less change than by reduction of styles. In the primitive
Dasyurinae and in Marmosa and Peramys the styles are very well-
developed and their convex external surfaces and sublanceolate form
closely parallel the conditions in Ccenolestes and some of the primitive
Phalangerinae. Moreover, the paracone in these forms is obviously
derived from the styles, in most cases being subordinate to them and in
none reaching a stage of development which overshadows them. In the
first upper molar of Phascologale cristicauda the paracone is not yet
differentiated and in the corresponding tooth in Dasyurus viverrinus
nearly the same condition obtains. This tooth, on account of its anterior
position, is probably more primitive than those posterior to it. In these
forms, its antero-external cusp has all the characteristics of a style
(style 6). A very slightly bifid tip indicates the beginning of the separa-
tion of the paracone and its complete differentiation is seen progressively
emphasized in the succeeding third and fourth molars. This first upper
molar in the dasyurids is at least superficially similar to the deciduous
molar of the didelphids in which, as suggested by Bensley, the antero-
external cusp appears to be the result of fusion of style and paracone
and may be retrogressive. However this may be, the tendency to
persistence of the styles is evident and the conclusion seems justified
that the same elements are still represented in the quadrate bunoid
molars of diprotodonts. If this be true it is possible that the small
metaconule in the molar of C&nolestes is not a new element but a per-
sistent unfused vestige of the metacone of the ancestral tritubercular
tooth. This at first seems improbable on account of the large size and

126 FIELD MUSEUM OF NATURAL HISTORY — ZOSLOGY, VOL. XIV.

functional importance of the metacone in polyprotodonts. Gregory
(1916, p. 248) has suggested even that the metacone in these forms
represents the apex of the original two-rooted tooth. It is to be noted,
however, that with the reduction of the metacone spur and the growth
of the hypocone the relative importance of the metacone decreases and
in some cases (e. g. Dactylopsila) the postero-external cusp, presumably
the metacone element, is exceeded by the hypocone which was non-
existent when the metacone dominated the tritubercular crown. In
Phascologale, Antechinomys and other generalized dasyurids, the stylar
element (style c) is very well-developed and only slightly exceeded by the
metacone. According to Sinclair, style c is well-developed also in
Microbioiherium tehuelchum of the Patagonian Miocene. To totally
eliminate this style and change the selenoid form of the metacone to
bunodont, meanwhile transposing the relations of the two to the root,
would be a much more extensive process than fusion of style and meta-
cone or even more than increased growth of the style and elimination
of the metacone. Moreover, it would be contrary to general evidence on
cusp migration, most of which is negative (Gidley, 1906; Tims, 1903).

LOWER PREMOLARS.

Description. — The single-rooted, unicuspid, anterior lower premolar
lies in series with the canine and the lateral incisors and is essentially like
them in size, form, and implantation. It is separated from the middle
premolar by a slight space. In the abnormal specimen having five
unicuspids, this space is unaffected, indicating that the extra tooth is
an anterior rather than a posterior one.

The middle premolar is two-rooted and tricuspid but it is in a
rather advanced stage of reduction toward the unicuspid condition.
The principal or middle cusp has taken an anterior position directly
above the anterior root leaving a long sweeping curve between it and
the posterior cusp. The anterior cusp is practically eliminated, although
its position is still indicated by the angle between the root and the front
of the main cone. A further reduction involving the loss of the small
posterior cusp and root would produce a tooth like the unicuspid
anterior premolar.

The posterior lower premolar, like the corresponding upper tooth,
is functionally associated with the molars rather than with the an-
terior premolars. A slight space separates it in front from the middle
premolar but posteriorly it is closely engaged with the front of the
first molar. It rises considerably above the anterior teeth and even
exceeds slightly the level of the molars. It has one large upright cusp
rising above its anterior root and a short depressed heel which is over-

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 127

lapped by the anterior shelf of the first molar. Thus, although higher,
it is slightly shorter than the middle premolar. On its antero-external
surface is a faint indication of a cingulum.

Discussion. — The lower premolars, like the upper ones, show tenden-
cies toward the condition prevalent in diprotodonts. This is evidenced
principally by the reduction of the anterior teeth and the separation of
the posterior one from them to a position at the head of the row of
grinding teeth. The only approach to this among polyprotodonts is
seen in the Peramelidae.

LOWER MOLARS.

Description. — The first lower molar has a relatively small trigonid
and an enlarged talonid. The protoconid, which is only very slightly
larger and higher than the other cusps of the trigonid, is connected by
crests with the paraconid and metaconid. The talonid is broad and
extended and the lateral angle between the protoconid and the entoconid
is relatively shallow. No hypoconulid is distinguishable. A distinct
external cingulum is present and continues forward to form a slight
anterior shelf. The second lower molar is similar in general to the first,
but is larger and its trigon is relatively wider, although still somewhat
narrower than the talonid. There is a slight indication of a hypoconulid
and the paraconid is flattened or even emarginate anteriorly where it
meets the entoconid of the preceding tooth. The third molar is smaller
than the second and approximately equal to the first. The talonid
exceeds the trigon in width only very slightly and the general outline of
the tooth from above is rectangular. The flattening of the paraconid is
more pronounced and almost divides it into two distinct tubercles.

The fourth molar is compressed and greatly reduced in size, being
narrower and only slightly longer than the trigon of the preceding tooth.
The metaconid persists nearly as large and well-differentiated as in the
other molars but the other cusps are rudimentary. The entoconid is
slightly elevated and distinguishable; the hypoconid is represented
merely by the postero-internal angle of the short blunt talonid; the
protoconid has disappeared; and the imperfectly divided paraconid is
vestigial. Despite its small size, this tooth is two-rooted, agreeing in
this respect with the extinct form Haltnariphus, but disagreeing with
Garzonia.

Discussion. — The lower molars, as compared with those of primitive
polyprotodonts, show three conspicuous tendencies.

(i) The reduction of the protoconid and the general reduction in
size and height of the outer cusps as compared to theinnerones. (2) The
elevation of crests connecting all the cusps. (3) The production of a

128 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

subrectangular tooth by the extension of the talonid and the antero-
posterior compression of the paraconid. These characteristics are seen
to a somewhat lesser extent in the Peramelidae and in general they lead
to conditions found among diprotodonts. The talonid still remains
slightly wider than the trigon, the individual cusps are still distinctly
elevated above the crests, and the molars as a whole are relatively long
and narrow. These appear to be primitive features and it is evident that
the lower molars are less advanced than the upper. Among the extinct
csenolestids, as noted by Sinclair, some are even slightly more primitive
in these respects than Ccenolestes while others show increased develop-
ment of transverse crests making the evidence quite clear for the
derivation of buno-lophodont types from tuberculo-sectorial.

THE ORIGIN OF DIPROTODONTY.

The diprotodont modification of the antemolar teeth has arisen in
so many diverse groups of mammals that it may well have had different
causes and courses of development in different cases. In some, however,
at least the general factors involved must have been the same. It occurs
in multituberculates, modern marsupials, insectivores, primitive eden-
tates, rodents, ungulates, bats and the lemuroids. In its most specialized
condition it consists in the complete suppression of all the antemolar
teeth except one pair of incisors in each jaw, these being functionally
modified and situated terminally. Among living marsupials, the pro-
gressive steps of this modification are illustrated by a series which is so
nearly complete that, except for the incipient stages, it could scarcely
be more convincing if it were in reality a linear phylogenetic series,
which of course it is not. Of living forms in which diprotodonty is
unmistakably developed, Ccenolestes undoubtedly has the most gen-
eralized dentition. From Ccenolestes to the more primitive phalangers,
as Dromicia and Distcechurus, the step is very slight; thence to the
herbivorous phalangers of the genus Trichosurus and on through
Hypsiprymnodon and Bettongia we come to the more typical macropods;
and in Phascolomys, although in other respects it is not a true terminal
member of this series, we see diprotodonty highly developed. This has
been traced in great detail and described by Bensley (1903) in the
important paper to which reference is so frequently necessary. Begin-
ning with Ccenolestes, the gradation from one form to another is almost
as complete as could be desired.

The idea that Ccenolestes is not a true diprotodont, but has only
paralleled the diprotodonts in dentition, need not here be considered
at length, since it is only dentition that is important in this connection

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 129

and so far as that alone is concerned there could be no possible objection
to considering C&nolestes directly prototypal to the Australian diproto-
donts. It may be said, however, that the ancestor of diprotodonts by
merely a priori reasoning would at some point have the general charac-
ters of C&nolestes, that is, it would have a diprotodont dentition while
retaining a number of polyprotodont characters. This will be discussed
elsewhere, but in the series of diprotodont dentitions we can at least
place that of Canokstes as the first and most generalized among living
forms.

The uncertainty lies in the steps preceding the development of the
caenolestid type of dentition, for the gradations between generalized
polyprotodont dentitions and the caenolestid stage of diprotodonty are
not recognizable with certainty among known living or extinct forms.
Hence the most constructive stages in the development of diprotodonty
in this case must be supplied theoretically. Bensley, whose opinion is
most important, states that "the diprotodont modification, although
characteristic of the herbivorous section of the Marsupials, is the result
of an insectivorous adaptation which must have been developed in the
minute ancestors of the Phalangeridae during the incipient stages of the
omnivorous evolution, but after the separation of the peramelid stem."
He therefore finds the beginnings of diprotodont tendencies in small
polyprotodonts such as Phascologale and Peramys which have undiffer-
entiated lower incisors but show a slight differentiation of the median
upper ones. Since primitive diprotodonts like Dromicia have upper
incisors of this same character he concludes that such incisors are part of
the course toward diprotodonty. With enlargement of the median
lower incisors in a form like Phascologale1 he would expect a shortening
of the lower jaw and consequent disturbance and reduction of the
"intermediate" antemolar teeth. These conditions he finds exemplified
in Dromicia and C&nolestes. This explanation by Bensley may be the
true one, for it is difficult to gainsay; but if applied to Casnolestes it
practically assumes a direct didelphid ancestor and therefore needs
careful examination. Before granting that diprotodonty arose ex-
clusively in this manner, it may be inquired whether there is any other
way in which it might have originated among marsupials. The didel-
phids and caenolestids were- already well distinguished in the Patagonian
Miocene. Hence the origin of the diprotodonty of Canolestes is still
farther back, perhaps in the Mesozoic, and it is certain that the ancestral
form was not the present day didelphid but an ancient generalized type,
all the characters of which are not preserved in any recent forms.

1 A much greater development of the anterior incisors is found in Oxygomphius
frequens, a European Miocene form usually referred to the Didelphiidae.

130 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Besides the didelphids and dasyurids there are at least two other
groups which may have retained some of the characters of the generalized
marsupial and which may throw some light on the origin of diprotodonty .
These are the Myrmecobiidae and the Peramelidae, the former group
with only one living species and the latter with a considerable number.
Myrmecobius, as is well known, has a very remarkable dentition com-
bining an excessive number of molars and a relative degeneration of all
its teeth. This apparently incompatible condition has been explained in
at least three ways. Owen, Thomas, Leche and others believed it to be a
primitive survival from ancestral forms having numerous post-canine
teeth such as Dryolestes of the North American upper Jurassic. Winge
has advanced the hypothesis that it is due to the retention of teeth
originally belonging to the deciduous series. In this he is followed by
Gidley (1915), who points out that "it requires but the addition of two
more permanently retained deciduous molars to equal the greatest
number of post-canine teeth found in this species, namely, nine." A
third explanation is offered by Bensley who believes that the extra
molars of Myrmecobius are due to a "simple reduplication of teeth from
the posterior position of the dental lamina" induced by "the favorable
conditions of increased space in the molar region." This conclusion is
obviously influenced by the reduced incisor formula of Myrmecobius
which he believes is derived through the dasyurids from the didelphids
which have a more primitive number of incisors. But he fails to take into
account the fact that similar reductions of incisors have taken place in
the early history of marsupials in Tertiary or even Mesozoic times.

To the writer, these different explanations seem equally theoretical
and the original one of Owen, if stated in general terms, is not less probable
than the others; that is, the resemblances between Myrmecobius and
Mesozoic forms seem quite as great as those between it and the modern
forms and, since other considerations point to great antiquity for several
different types of marsupials, there would seem to be no great improb-
ability in the belief that the extra molars are a survival rather than a
modern development.

At the present time Myrmecobius is obviously in process of losing
rather than gaining teeth. This has been noted by Gidley who says:
"The variability in the molar series in Myrmecobius seems due to the
presence or absence of the last molar, probably a disappearing tooth.
In the upper jaw the last molar seems to be normally wanting, while the
second is apparently in the process of disappearing, being sometimes
present and sometimes wanting." It seems highly improbable that the
causes which originally brought about an increase in the tooth formula
are the same as those now producing a decrease. A more reasonable

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 131

assumption would be that the increase took place before degeneration
of individual teeth began and for a different reason. The process of
reduction now beginning in Myrmecobius may perhaps be compared to
the more advanced condition in Tarsipes which retains but three or at
most four post-canine teeth and is evidently well on the way to an
edentulous state.

Gidley makes the unequivocal statement that "There is in our
present knowledge nothing to support Owen's hypothesis regarding the
derivation of Myrmecobius." But in succeeding pages of the same paper
he argues for an ancient origin of the Myrmecobiidae, which he supposes
to have been well differentiated from the didelphids and dasyurids as
early as the Paleocene. Evidently he believes in a Mesozoic ancestor for
Myrmecobius and to that extent at least agrees with Owen. Whatever
the case may have been with respect to the history of the extra molar
teeth of Myrmecobius, the view that all the living families of marsupials
were well differentiated early in the Tertigry seems to be well founded.
Therefore, without reference to possibly archaic characters other than
teeth, it is still possible to believe in an early predidelphid origin for
Myrmecobius. As a possible forerunner of diprotodonts, the Myrme-
cobius line is thus entitled to some consideration. It has the same
antemolar formula as C&nolestes, its upper canines are somewhat
reduced, and its median lower incisors are greatly specialized and
relatively larger than in any other polyprotodont. These incisors are
completely terminal in position and, although slightly recurved at
the tips, are essentially proclivous with their roots nearly in the principal
longitudinal axis of the mandible and set at a decidedly different angle
from the lateral incisors. Thus the median lower incisors of Myrme-
cobius are essentially like those of diprotodonts. If these teeth are as
much reduced as the rest of the dentition, they must have been derived
from a type which might easily have led to the diprotodonts. The
median upper incisors of Myrmecobius are reduced in size and slightly
smaller than the lateral ones. They are set very far apart and are
slightly procumbent, obviously in a secondary condition which differs
from that of polyprotodonts and diprotodonts alike. Hence no con-
clusion is to be drawn from them although it is to be noted that the
median upper incisors are widely separated at the base in diprotodonts
more generally than in polyprotodonts. The first pair of upper lateral
incisors are modified to meet the median lower ones as in diprotodonts
and all the lateral incisors are laterally compressed. The differentiation
of the median upper incisors is scarcely to be regarded as a diprotodont
character for it is not pronounced except in a few extreme cases, as
Phascolarctos and Phascolomys, and in many it is distinctly less than in

132 FIELD MUSEUM OF NATURAL HISTORY — ZodLOGY, VOL. XIV.

some polyprotodonts. In Trichosurus, for example, and in C&nolestes,
the median upper incisors are less differentiated than in some species of
Phascologale.

For the development of diprotodonty from such a form as Marmosa
the following are among the changes which would be necessary: Reduc-
tion in number of incisors; enlargement of median lower incisors; reduc-
tion in size of canines and anterior premolars; decrease in the depth of
the mandible; increase in craniofacial length; formation of short diastem-
mata between at least some of the teeth ; assumption of proclivous and
fully terminal position by median lower incisors; assumption of wholly
lateral position by posterior incisors; lateral compression of incisors and
development of long and f orwardly directed cutting edges, especially in
those of the upper jaw. All of these modifications are found to a con-
siderable degree in Myrmecobius and in these respects this form re-
sembles the diprotodonts more closely than do any of the didelphids or
dasyurids. Whether these characteristics were more or less pronounced
before the present degeneration of teeth began cannot be said, but at
least it must be conceded that Myrmecobius has some claims to a position
in or near the line of the diprotodonts whether it be regarded as an
ancient type derived from Jurassic forms or a more recent one proceeding
from the dasyurids.

Leaving Myrmecobius for further discussion later, it may now be
inquired whether the peramelids could have furnished the beginnings
of diprotodonty. Although Bensley finds in the Peramelidae many
indications of an ancestral relation to the modern diprotodonts and an
advance over the primitive dasyurids and didelphids, he is not inclined
to regard the diprotodont dentition as directly derived from them. In
this he was influenced largely by the undifferentiated condition of the
median upper incisors in the peramelids; by the retrogressive (?) char-
acter of the upper canines and premolars, the existence of primitive and
advanced stages of the hypocone in both groups, and the hypsodonty
of the peramelids. The development of the hypocone and the hypsodont
modifications in the peramelids are progressions beyond the most
primitive stages of the known diprotodonts. But they need not be
considered wholly homoplastic and their inception may have preceded
the divergence of the peramelid stem. In any case they do not preclude
a belief in a common ancestry for the peramelids and the diprotodonts
and one which was not very remote. Such an ancestry at most need only
be removed to a stage in which the hypocone and the hypsodont^ were
approximately at the most primitive stage common to both peramelids
and diprotodonts. This would carry us back to a brachyodont condition
but not to a total absence of the hypocone. So of the two characters

MAY, 1921. AMERICAN MARSUPIAL, C.ENOLESTES — OSGOOD. 133

involved, only one — hypsodonty — needs to be accounted for. This may
well be regarded as a recent specialization for it is of a very peculiar
nature, a sort of half -hypsodonty confined to one side of the teeth and
not fairly comparable with the full hypsodonty of the most advanced
diprotodonts.

The undifferentiated median upper incisors of peramelids and the
supposed retrogressive canines and premolars, in the light of a study of
Coenokstes, do not seem at all unfavorable to a hypothesis that a dipro-
todont dentition might have proceeded from the peramelid stem. The
incisors, while not differentiated for seizing as in the dasyures and
didelphids, are unquestionably in a secondary condition of which no
further development is seen elsewhere, unless it be among diprotodonts.
These incisors have become broad and somewhat bladelike with their
cutting edges in the same horizontal plane as those of the adjacent
lateral incisors. In fact they are practically indistinguishable in shape
from the pair of lateral incisors situated next to them. Hence it seems
that the same factors which have led to modification of the lateral teeth,
may have affected the median ones. The principal of these factors is the
functional interaction with the lower incisors. In Perameles the three
pairs of lower incisors are collectively so modified in form and especially
in position that their functional relation to the upper incisors is much
like that of the elongated single pair of lower incisors in diprotodonts,
certainly much more so than in any other polyprotodont. They are set
so closely together that their cutting edges are practically continuous.
In addition they are exceedingly proclivous with their roots almost as
near the long axis of the mandible as in diprotodonts. Thus the func-
tional effect of the three incisors in Perameles is practically the same as
that of the one in diprotodonts and it is probable that specialization in
the lower jaw has conditioned that in the upper rather than vice versa.

Another interesting character correlated with the proclivity of the
lower incisors in Perameles is found in their exposed roots, those of the
first and third especially being without bony covering in front. The
second incisor is more fully rooted, but the other two, although well-
developed, are very weakly attached to the jaw. Perhaps enlargement
of any one of the three might lead to the loss or reduction of the others
which would result in a diprotodont dentition. This is uncertain, but it
is probably fair to assume that the proclivity and the exposed roots of
these teeth are indications at least of some instability, possibly following
the loss of the fourth pair or foreshadowing a changed relation of those
remaining. The reduction that has already taken place in Perameles
corresponds to the reduction in the diprotodonts, that is, the first pair
of lower incisors has been lost and thus one of the stages preliminary to

134 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

diprotodonty has been passed. Embryological evidence so far adduced
points to the conclusion that this is the case. Woodward's (1893) studies
of developing teeth in the Macropodidae led him to believe that the
large functional lower incisor is a persistent member of a series in which
it formerly stood in second place. Wilson and Hill (i 897) , from extensive
work on the development of Perameles, find that the first lower incisor
has been lost and that the one now found in terminal position is homo-
logous with the large lower incisor of the macropods. These authors say
(p. 509) : "It is interesting to compare the condition in Perameles with
that described by Mr. Woodward for Petrogale. In that form he found
two small calcified vestigial incisors. The anterior of these was entirely
in front of the large permanent lower incisor, which latter he therefore
regarded as really * 2. • • • And may it not be that Woodward's
second vestigial incisor (his *3) to which he was unable to recognize any
ancestor, is really the homologue of our dii whose legitimate successor
is Woodward's iz, the almost certain homologue of our li\ ' of the adult
Perameles?"

The dasyurids have the same number of lower incisors as Perameles,
but whether they also have lost the first pair has not as yet been demon-
strated. Embryological evidence in their case seems lacking. It was
formerly supposed (Cf. Thomas, 1887) that the dasyures had lost the
fourth incisor and therefore that the maximum of three incisors in higher
mammals represented the first, second, and third of the primitive series.
The evidence for this supposition adduced by Thomas now seems
inadequate, but in connection with theories of the origin of diprotodonty
it would be important to know whether the dasyurids have lost the same
incisor as the peramelids.

The canines and premolars, which Bensley interprets as retrogressive,
constitute a similarity between Perameles and Ccznolestes and would seem
to favor a common ancestry rather than otherwise. Whether they are in
reality retrogressive is another question. Retrogression in this sense is
a sort of reversed homoplasy which can be recognized only if the imme-
diate ancestry of the form in question is known. That is, in the case of
Perameles, a didelphid ancestry is essential to a belief in the retrogressive
nature of its canines. On the other hand, if Perameles had no distinct
didelphid ancestry but descended from a more primitive form also
ancestral to the didelphids, it may have retained a primitive premolari-
form canine. Highly specialized canines appeared very early, in fact
they were well differentiated among some early reptiles, and various
Mesozoic mammals had simple single-rooted canines, but others, like
the Jurassic Triconodon, possessed double-rooted or grooved canines
similar to those now found in Perameles, Ccenolestes, and Myrmecobius —

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 135

all of which it should be noted are forms having certain primitive char-
acters not retained by the didelphids. Double-rooted canines are not
unknown in other groups than marsupials but are very rare except in
insectivores, another primitive group. Among these, premolariform
canines show no evidences of retrogression. Double-rooted canines are
said to occur also in the Miocene pig, Hyotherium (Beddard). In
Antechinomys laniger, one of the most primitive polyprotodonts, the
canine is single-rooted but has its crown distinctly premolariform. To
interpret this as primitive seems quite as reasonable as to call it retro-
gressive (see p. 119). Canines with lateral or antero-posterior cusps are
found also in certain bats (e. g. Cynopterus, Harpionycteris). In the
Miocene caenolestids the canine is mostly unknown, but in Parae-
panorthus, as figured by Ameghino (1903, fig. 62), the canines are small
and single-rooted, not unlike those of many phalangers. These teeth
may have been derived from double-rooted ones but indicate no tendency
toward them. Such a form as Paraepanorthus, however, is evidently
farther advanced in some respects than C&nolestes, for its incisors are
more reduced in number.

On the whole, it appears that double-rooted, groove-rooted, or
premolariform canines are found almost exclusively among forms
exhibiting various primitive characters. The Peramelidae are primitive
in certain other respects and, although it is not unlikely that their
canines may have passed through various changes prior to recent times,
it seems quite as possible that they may have retained an early condi-
tion like that shown in the Jurassic Triconodon. The same may be true
of Myrmecobius and of C&nolestes.

The objections to a diprotodont evolution from the peramelids
therefore, are not insuperable. In fact it is clear that many of the
changes necessary to a transition from a didelphid to a phalanger or
caenolestid are already present in Perameles and it does not seem demon-
strable that these changes took place subsequent to the separation of the
peramelid and diprotodont stems. In other words, the early peramelids
may have become considerably differentiated from the generalized
didelphoid marsupial and later divided into one line leading to the
modern peramelids and another to the diprotodonts. At least, the con-
clusion is unavoidable that among living forms those most suggestive of
what the ancestor of C&nolestes and other diprotodonts was like, are the
Peramelidae. Moreover, if the early Tertiary form Myrmecoboides is
in reality a "properamelid," as suggested on another page (p.!i42), belief
in the origin of diprotodonty from this group is greatly strengthened.
The following resemblances between Ccenolestes and Perameles are to
be noted: Six or more exclusive cranial characters (see p. no), tooth

136 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

formula, molariform teeth with hypocone, lower incisors proclivous,
lateral upper incisors compressed, reduced canines and premolars,
specialized posterior premolar, movable cervical rib, osseous patella,
large median vagina, cursorial structure of legs, first and fifth toes of
forefeet reduced, caecum small, tail not prehensile. Many of these
characters, while not diagnostic of the diprotodonts as a group are
present in the majority of that group and absent in the majority of
polyprotodonts.

It may be concluded, therefore, that diprotodonty in the ceenolestids
may have originated in one of three ways:

(1) from didelphid pre Miocene types through undiscovered inter-
mediate forms existing prior to the differentiation of the peramelids.

(2) from an ancient line now represented by Myrmecobius.

(3) from primitive peramelids after their separation from the
dasyurid-didelphid stem.

The first of these is the theory of Bensley and in general of those
who have followed him. The second is perhaps less entitled to con-
sideration, but is still not wholly baseless. The third is to the writer
more probable than either of the others.

RELATIONSHIPS OF WYNYARDIA.

The extinct mareupials known from the Australian region are nearly
all of Pleistocene age and unequivocably diprotodont or polyprotodont.
The most conspicuous exception to this is Wynyardia bassiana Spencer
(1900) from beds at Table Cape, Tasmania, which are regarded as late
Eocene or early Oligocene. This form, the remains of which include
parts of limb bones and an imperfect skull without teeth, exhibits such a
curious mixture of characters that it has so far been omitted from formal
classifications. Spencer summarizes his study of it as follows: "A
consideration of all the features would appear to lead to the conclusion
that the fossil is the representative of a now extinct series of forms which
were more nearly allied to ancestral Polyprotodonts than are any of the
existing Diprotodont forms. It may, in fact, be regarded as inter-
mediate between the former and the latter, and as indicative of a stage
in the development of Australian marsupials when the ancestors of the
recent diprotodontia were beginning to diverge from the original
Polyprotodontid stock from which they have been developed within the
limits of the Australian region."

Bensley (1904, p. 200) is somewhat more definite, saying: "While it
would be difficult to add to the excellent comparisons presented by
Spencer, it is probable that the reference to the animal as an inter-

MAY, 1921. AMERICAN MARSUPIAL, CENOLESTES — OSGOOD. 137

mediate form must be excluded, if for no other reason, on account of its
large size. . . . All the forms which approach the hypothetical inter-
mediate type are of small size. Not only this, but the diprotodont
modification itself may, as already explained, be shown to represent an
insectivorous adaptation which could only have taken place in com-
paratively small animals. The relations of Wynyardia are more probably
with one of the advanced genera, such as Pseudochirus, Phascolarctos,
or Phalanger."

Osborn (1907) says: "Among fossil forms the gap between the two
suborders is largely bridged over by the extraordinary genus Wynyardia
of Baldwin Spencer, which presents a perfect melange of characters seen
elsewhere only in the Opossums and Dasyures (Polyprotodonts), and
in the Phalangers and Kangaroos (Diprotodonts)." Gregory (1910,
p. 214) speaks of Wynyardia as "another form that helps to bridge over
the structural gap between the Poly protodontia and theDiprotodontia."
Later (Osborn, 1910) it fails to receive formal classification.

Spencer, in summarizing the important characters of Wynyardia
finds five which he interprets as indicating polyprotodont affinities.
These, as numbered and stated by him are as follows:

"(i) Proportionate length to breadth of the skull, 100:67. This
approximates most nearly to Dasyurus, and shows a decidedly
greater width than in the Phalangeridae.

" (2) Lambdoidal crest well developed, as in Dasyurus.

" (3) Sagittal crest strongly developed, resembling that of Dasyuridae
and species of Didelphys.

" (5) The wide sweep and upward curvature of the zygomatic arches,
as in Dasyuridae.

" (7) The transverse elongation of the glenoid cavity, the downward
produced plate of bone which forms the boundary is not con-
nected with any structure forming part of the auditory passage.
In this respect it agrees with Dasyuridae and Perameles, and
differs markedly from the Phalangeridae, amongst which it
forms the anterior part of a bony auditory canal."

At least the first three of these characters cannot be regarded as
exclusively polyprotodont. Taking them up numerically, it is found as
to (i) that, although Dasyurus and Sarcophilus have very wide skulls,
polyprotodonts in general have relatively narrow ones. The ratio runs
from only 39.7 in Perameles to 81.2 in Sarcophilus (see p. 99). On the
other hand the average relative width is greater in diprotodonts than
in polyprotodonts, the only very narrow type seen among diprotodonts
being in the prhpiitive Ccenolestes.

As to (2), it is to be noted that a high lambdoid crest is no more than
a generic character among polyprotodonts. Within the limits of the

138 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Dasyuridae and the Didelphiidae it is extremely variable. Various
diprotodonts, including most of the macropods and some phalangers
have considerable indications of a lambdoid crest and in Phascolarctos it
is highly developed. Thylacoleo also is described as having a high crest.

As to (3), much the same is true of the sagittal as of the lambdoid
crest. In the Didelphiidae we find a high sagittal crest in Didelphis and
none in the closely related Peramys and Philander. A sagittal crest of
some prominence is seen in the skull of Trichosurus. Its presence is less
frequent than its absence in both polyprotodonts and diprotodonts and
cannot be regarded as of great significance.

Spencer's characters (i), (2), and (3), therefore, cannot be taken as
evidence of the polyprotodont affinity of Wynyardia.

Characters (5) and (7) seem somewhat correlated and it is evident
that in the region of the glenoid fossa and the posterior base of the
zygoma we have a more pronounced resemblance to the condition in
polyprotodonts than in any living diprotodont with the exception of
C&nolestes. In the relatively high position of the glenoid fossa, however,
there is approach to the phalangers. It seems, therefore, that only one
of the supposed resemblances to polyprotodonts is important and this
is approximated in the primitive diprotodont C&nolestes.

The resemblances to diprotodonts are almost as variable and as
difficult to regard as diagnostic, but there are more of them and on the
whole they seem more convincing of relationship. This is especially true
if comparisons are made only with typical polyprotodonts (as didelphids
and dasyurids) and with typical diprotodonts (phalangers), leaving
aberrant forms, as Perameles on the one hand and C&nolestes on the
other, out of consideration. Thus the large squamosal is found in
Phascolarctos much as in Wynyardia completely excluding the alisphe-
noid from contact with the parietals; and in general the squamosal tends
to forward extension among diprotodonts but, while this offers con-
siderable contrast to the condition in the Dasyuridae and Didelphidae,
it is fully equalled or even outdone in the Peramelidae. One of the
characters suggesting diprotodont relationship of Wynyardia is the
large premaxilla but here again the peramelids must be ruled out. The
relatively large cranium, the high position of the glenoid fossa, the
elevated mandibular condyle, the position and shape of the anterior
root of the zygoma, the massive ischium, and the proportionately long
hind limbs all seem to point to diprotodont affinity in Wynyardia.
Hence, while it may be agreed with Spencer and others that it combines
many features of the two groups, it seems best for practical purposes to
classify it with the diprotodonts and in recognition of its various unique
features to regard it as the type of a distinct family, the WYNYARDIIDAE.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 139

A comparison of C&nolestes with Spencer's figures and description of
Wynyardia, although rather unsatisfactory as a basis for positive con-
clusions, reveals a number of points in common. These may be enu-
merated as follows: (i) Cranial part of skull large as compared to facial.
(2) Relations of glenoid fossa to auditory meatus similar. (3) Squamosal
not inflated and posterior base of zygoma somewhat Dasyurine.
(4) Glenoid fossa relatively high. (5) Nasals extended anteriorly and
expanded posteriorly. (6) Premaxillae large and separated from the
maxillae by a suture which is only slightly oblique. (7) Mandibular
condyle relatively high. (8) Femoral trochanters similar and a third
tuberosity present as in Phascolomys.

Comparison of Wynyardia with extinct American forms allied to
C&nolestes, as Palceoihentes and Abderites, also shows several interesting
points. Thus Palceoihentes has lambdoid and sagittal crests combined
with zygomata of a type similar to those of Wynyardia and the frontals
are widest in their cranial part. In Abderites, as previously noted by
Spencer, the seat of the large lower sectorial tooth forms a deep trans-
verse furrow in the mandible which seems closely comparable to the
condition in Wynyardia.

Present material is not adequate for the positive conclusion that
Wynyardia is intimately allied to the American forms but the above
similarities are most suggestive and must be regarded at least in some
degree as strengthening belief in extensive community of relationship
between Australian and American marsupials, both polyprotodont and
diprotodont.

RELATIONSHIPS OF MYRMECOBOIDES.

The genus Myrmecoboides was established by Gidley in 1915 for an
incomplete lower jaw from the Paleocene (Fort Union) of Montana.
It was tentatively classified as belonging to the family Myrmecobiidae
and characterized as follows: "Canine semipremolariform, being
irregularly triangular in cross section and but slightly curved; canine
and the three simple premolars evenly spaced with short intervening
diastemae. There is also a short diastema between the canine and i3
(the position of the other incisors is not known) . Fourth tooth behind
the canine (probably dp 4 retained) completely molariform; true molars
tritubercular, with well-developed basin heel, but with inner cusps of
trigonid (paraconid and metaconid) as high or higher than main outer
cusp (protoconid)."

The principal characters regarded by Gidley as suggesting possible
affinity with Myrmecobius are (i) the diastemata between the premolars,

140 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

(2) the indicated straightness of the jaw anterior to the canine, (3) the
premolariform canine, (4) the compressed premolars, (5) the absence of
cingula, (6) the relatively high inner cusps (metaconid and entoconid).
He also points out that the extinct form differs from Myrmecobius in at
least three important particulars (i) the possession of the normal num-
ber of post-canine teeth, (2) the approximation of the paraconid to the
metaconid, (3) the more elevated and better defined trigonid.

Comparison of Gidley's figures and description with several species
of Peramelidae shows at once that practically all the above enumerated
points of resemblance are to be found in that family and none of the
important points of difference. This is indicated by the following
parallel:

Myrmecoboides and Perameles Myrmecobius

Canine equal to or lower than pre- Canine higher than premolars.
molars.

Protoconid and hypoconid well Protoconid and hypoconid small or

developed. rudimentary.

Paraconid and metaconid approxi- Paraconid and metaconid not approxi-
mated, mated.

Trigonid elevated and distinct. Trigonid not differentiated.

Post-canine teeth seven. Post-canine teeth more than seven.

The important points of difference between Myrmecoboides and
Perameles are (a) the unicuspid anterior premolar in the extinct form,
(b) the large last molar with its well-developed hypoconulid, (c) the
exceptionally large paraconid of the first molar, and (d) the somewhat
more brachyodont condition of the molar series. The unicuspid pre-
molar is a specialization such as might be expected in an ancestral
diprotodont but the other characters are primitive and doubtless hark
back to a tritubercular stage. The resemblances to Perameles are
pronounced, however, and cannot be overlooked, although the classifica-
tion of such an ancient form on fragmentary material has much possi-
bility of error. In view of the many similarities to Perameles shown by
C&nolestes, the occurrence in the early American Tertiary of a form with
distinct leanings to Perameles becomes especially significant. The
hypothesis can scarcely be avoided that, provided Myrmecoboides is a
marsupial, it may well be ancestral to the caenolestids. A point of espe-
cial interest and significance is the unicuspid premolar which is paralleled
among marsupials only in caenolestids and diprotodonts.

PHYLOGENY AND TAXONOMY.

The great range of variation in structure and adaptive characters
shown by existing marsupials makes it possible without reference to
extinct forms to construct linear series in which the sequence from

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 141

generalized to specialized types is remarkably free from interruptions.
If their present distribution also be disregarded, that is, if both American
and Australian forms be included, the problem is simplified. As worked
out in great detail by Bensley (1903), the American Didelphiidae stand
in a prototypal morphogenetic relation to the -remaining groups of
marsupials. This conclusion is based upon their possession of a com-
bination of various generalized features which are distributed over at
least three different Australian groups. They have the primitive
tritubercular molar as in the Dasyuridae, the upper incisor formula and
upper molar stylar characters as in the Peramelidae, and the pedimanous
foot-structure (including incipient syndactyly in one form) as in the
Phalangeridae.

As the probable ancestor of the modern Didelphiidae, Bensley finds
the Oligocene Peratherium fulfilling practically all theoretical require-
ments including a distribution in both Europe and North America
which might permit its dispersal either to South America or Australia or
to both. This dispersal he supposes to have taken place in at least three
radiations, the first being represented by the Australian fauna, the
second by the Miocene fauna of South America and the third by the
present day Didelphiidae of South America, this last being in its incipi-
ency. The evolutionary series so thoroughly expounded by Bensley is
an exceedingly plausible one in spite of its having been based largely
upon existing forms. The various modern polyprotodonts are obviously
reducible to a common type and, although the diprotodonts offer more
difficulty, they may at least be traced from one to another. Between the
two groups is a rather definite hiatus but there is little room for doubt
that diprotodonts proceeded from polyprotodont ancestors. Indications
of some of the steps leading from one group to the other are seen prin-
cipally in the Peramelidae, the Wynyardiidae, the Myrmecoboididae and
the Palaeothentidae. No one of these aberrant groups is free from
objection as an ideal transitional type, yet taken together they serve to
bridge quite thoroughly the gap between the generalized polyprotodonts
and the more primitive diprotodonts. Bensley dealt chiefly with modern
Australian forms, giving relatively little attention to Wynyardia or
C&nolestes, and Myrmecoboides was unknown to him. But the Pera-
melidae he regarded as proceeding from a hypothetical stem-group called
Properamelidae, one branch of which led to modern peramelids and
another to the diprotodonts. This group presumably had begun to
acquire subquadrate upper molars with incipient hypocones as in
modern peramelids but had not developed a terrestrial foot structure.
The Properamelidae themselves were supposed to have evolved from
the generalized polyprotodont having the characters of the Didelphiidae

142 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

and giving rise through another line to the Dasyuridae and modern
polyprotodonts. Interesting in this connection is the fact noted by
Bensley (1903, p. in) that the modern peramelids agree with the
Oligocene Peraiherium more closely in the character of the important
styles of the molars than they do with the Australian dasyurids.

The fossil record is very incomplete but its testimony accords in a
surprising degree with the hypothetical requirements of Bensley's
scheme. Myrmecoboides might without serious objection be looked upon
as in the direct line between the " Properamelids " and the diprotodonts,
since it has begun to reduce its premolars, the first being small, unicuspid,
and single-rooted as in many diprotodonts. Furthermore, Canolestes
shows numerous resemblances to Perameks, so many that a common
ancestry subsequent to the generalized polyprotodont stage is by no
means impossible. The material representing Myrmecoboides is too
imperfect to warrant more than a provisional conclusion that it may
be an early stage in the line of the caenolestids. Nevertheless the evi-
dences of affinity between Myrmecoboides, Perameles, and Canolestes
cannot be overlooked. Moreover, if Myrmecoboides is in truth a repre-
sentative of the early differentiation of a diprotodont stem then it is
the first form of the kind to be discovered outside the southern continents
and goes far toward explaining the present day occurrence of diprotodont
types in both Australia and South America.

Further points of interest in connection with the resemblances of
Ccenolestes to the peramelids are found in the slight indications of com-
mon relationship to Notoryctes. This peculiar form is so highly spe-
cialized that its affinities are very problematical, but Bensley is inclined
to regard it as an offshoot of the "properamelids." Certain features of
the myology (see p. 60) and osteology of Canolestes and Notoryctes are
similar and favor belief in a common (properamelid) derivation rather •
than otherwise. The actual course of evolution must be inferred to a
considerable extent even under the best of circumstances. The view
that Canolestes is a primitive diprotodont is not proved but is strongly
supported by its resemblance to the peramelids which of all polyproto-
donts are the ones most suggestive of the incipient stages leading from
one large group to the other. It is probably not too much to say that if
the caenolestids had been discovered in Australia instead of South
America they would have been accepted without question as ancestral
diprotodonts. This geographical difficulty is at least partly overcome
by the presence of Myrmecoboides ', also with peramelid affinities, in
North America. Hence it is a hypothesis of considerable probability
that the caenolestids, like the Australian diprotodonts, developed

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 143

from "properamelid" ancestors and that the characters or tendencies
fundamentally distinguishing them from polyprotodonts were a heritage
which they had in common with the Australian diprotodonts.

Increased knowledge frequently, perhaps usually, leads to increased
difficulty in defining taxonomic concepts. This is especially true if it
be attempted to express both morphogenetic^ and phylogenetic relation-
ships in one and the same scheme of nomenclature. The division of
modern marsupials into polyprotodonts and diprotodonts was made
originally upon a morphological basis and until the discovery of extinct
forms, which not only were connectant themselves but drew attention
to connectant characters in certain modern forms, the two groups were
clearly defined. In the case of the peramelids, a position with the
polyprotodonts was generally conceded, notwithstanding their syn-
dactyly which was otherwise characteristic of diprotodonts; Wynyardia
was left without definite allocation; and Canolestes, as detailed in pre-
ceding pages (pp. 6-15), was regarded by some authors as a diproto-
dont and by others as a polyprotodont.

The foregoing study of Canolestes, while adding much to knowledge
of fact, does not make classification easier. It is believed that the
caenolestids and the peramelids, possibly with Notoryctes also, proceeded
from the same stem as the phalangeroid diprotodonts. Hence these
might be brigaded as one group. To characterize such a group on the
basis of the morphological characters exhibited by modern forms, espe-
cially with relation to degree of specialization, however, seems practically
impossible. After trying various alternatives and considering the present
state of knowledge, no more satisfactory primary division of marsupials
offers itself than the old one into Polyprotodontia and Diprotodontia.
In such a classification, as already shown, C&nolesies and Wynyardia
may be placed with the diprotodonts. The line must be drawn some-
where and the least objectionable place for it seems to be between the
caenolestids and the modern peramelids. The position of Myrmecoboides
must remain doubtful until better material is forthcoming, but its close
alliance with Perameles is scarcely to be doubted unless it should prove
to have placenta! affinities. The recognition of a superfamily Caenoles-
toidea, as in the classification of Osborn and Gregory, serves to divide
American and Australian diprotodonts and has its advantages as a
matter of convenience. The modern genus Canolestes, however, need not
take separate family rank but may be included in the Palaothentidae
with the very closely allied extinct forms from the Santa Cruz beds
of Patagonia. To rank it as a subfamily, Caenolestinae, would be
convenient.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 145

The accompanying diagram (p. 144), like others of its kind, has many
shortcomings and some inconsistencies. It is offered only provisionally
as an aid in indicating the relative position of Ccenolestes. Various prob-
lems in the phylogeny of marsupials must somehow be met in such a
diagram even though they seem properly beyond the scope of the
present study. So far as any conclusions are indicated, they are to be
regarded mainly as suggestive rather than decisive, as indeed they would
be in any case. Myrmecobius is separated from the main polyprotodont
stem and from the dasyurids in order to indicate its possible direct
descent from mesozoic types, but a position nearer the dasyurids is
almost equally probable.1 The problem of the relationship of Propoly-
mastodon has not been studied and its uncertain position is merely
suggested (see Gregory, 1910, p. 212). The same is true of Pediomys
and Didelphops, while the multituberculates (Allotheria), which Broom
contends are monotremes rather than marsupials, are indicated as
probably proceeding from the unknown forms indefinitely termed
"generalized marsupials" and in reality meaning very little.

DISPERSAL OF MARSUPIALS.

Broadly speaking there are only two theories of the dispersal of
marsupials, one that the group originated in Holarctica and spread
southward, the other that it had its beginnings in Antarctica, the
hypothetical southern continent which formerly may have connected
part of present-day South America and Australia. Advancing knowl-
edge has from time to time favored one or the other hypothesis but
opinions are still divided. A review of the subject with many references
is given by Osborn (1910, pp. 64-80).

Before the discovery of the extensive extinct fauna of Patagonia,
the distribution of marsupials offered no serious objection to a theory of
northern origin for the group. Practically the only fossils known were
didelphids from the Upper Miocene and Oligocene of Europe and North
America, and the conclusion was fairly obvious that these might repre-
sent the ancestral stock from which both Australia and South America
received their marsupial populations. The best known advocate of this
theory was Wallace. A reference to his work on the Geographical Dis-
tribution of Animals (1876), however, shows at once that the information

1 Although this disposition of Myrmecobius is somewhat heretical, I am led to it
through the cumulative effect of encountering in this study and in the published
work of others the continual necessity of special theory or explanation to account
for first one and then another peculiarity of this animal. Much of this would be
obviated by carrying its origin farther back. The possibilities for longer lines of
descent than usually assumed is evidenced by Canolestes itself in its persistence
from Miocene times to the present in practically unchanged condition.

146 FIELD MUSEUM OP NATURAL HISTORY — ZOOLOGY, VOL. XIV.

on which his conclusions were based was vastly inferior to that now
available.

Although it may be admitted that some of the Mesozoic mammals
were generalized marsupials, it is hardly possible to consider them in
efforts to follow the course of differentiation and dispersal of the modern
groups. For present purposes the multituberculates also may be dis-
regarded. From the early Tertiary we now know relatively primitive
types of marsupials from North America, Europe and South America.
Correlation of these different forms has been a matter of much difficulty
but according to the most recent authorities, there is little difference in
age between the northern and the southern types. Thus in North
America and Europe we have the didelphoid forms collectively placed
in the genus Peraikerium and regarded as of upper Eocene and Oligocene
age. We also have in North America Myrmecoboides, which, although a
less primitive form, is found in the Basal Eocene (Fort Union) . Of still
earlier age, but more doubtful relationship, are Pediomys and Didelphops
from the North American Upper Cretaceous. From South America the
most primitive type known is Proteodidelphys from the Notostylops
beds the age of which has been variously placed from Lower Cretaceous
to Upper Eocene. From Australia the earliest known form is Wynyardia
which is of at least Oligocene age and, although not primitive, has a
combination of diprotodont and polyprotodont characters which at this
early period is significant.

Subsequent to these primitive types there are three large marsupial
faunas the members of which show various similarities to each other
and any or all of which may have been developed from the early gen-
eralized forms. These faunas are represented by (i) the Miocene fossils
from South America including the caenolestids, the borhyaenids and the
so-called Microbiotheridae which have didelphid characteristics;
(2) the present day Australian fauna with its diversity of specialization;
and (3) the present day American fauna all referred to the single family
Didelphiidae. The first and second of these faunas, broadly speaking,
are mutually representative, that is, the South American caenolestids
are at least to some degree counterparts of certain Australian diproto-
donts; the borhyaenids stand in the same or even closer relation to the
Australian thylacine; and probably it is not too much to say that the
Microbiotheridae, in spite of didelphid similarities, are not far removed,
structurally, from the generalized Australian polyprotodonts.

This correspondence between the South American Miocene forms
and the Australian recent forms has not unnaturally led to the con-
clusion that a direct interchange between the faunas actually took place

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 147

perhaps in Cretaceous times by means of an Antarctic land bridge.
Further evidence of the former existence of such a bridge is afforded by
plants, mollusks, crustaceans, insects, amphibians, reptiles, and other
forms of life of which closely allied species are now found inhabiting
both Australia and southern South America. Enumeration of these
forms has been made at great length by various authors and need not
be repeated. It is sufficient to admit that the evidence is very extensive
and if the southern continent be supposed to have been one of no great
permanence, but, as Hedley (1895) has indicated, may have been attenu-
ated, irregular and unstable at many points, belief in its existence is
greatly strengthened and the objections based upon geological and
isostatic data do not weigh so heavily against it. To cite only a few
of the believers in Antarctica, it may be said that among its most
recent adherents have been Sinclair (1906), Ortmann (1905), Scott
(1913), and Osborn (1910). One of its earliest and most distinguished
advocates was Huxley. In general it seems to have been conceded that
with so much and so varied evidence there could be little doubt that
some sort of connection must have existed. The northern origin of many
and very important groups, however, could scarcely be denied and the
general theory of mammalian dispersal was thus a bipolar theory.
Nevertheless the Argentine paleontologist Ameghino went so far as
to regard South America as the original home and center of dispersal
of practically all the larger groups of mammals.

In contradistinction to these views and in reversal of his own previ-
ous acquiescence to belief in Antarctica, Matthew (1915) has presented
a powerful argument for the dispersal of mammalian life exclusively
from Holarctic centers.1 His conclusions are based upon a broad and
comprehensive survey of geological and biological data and the similari-
ties between certain southern faunas are regarded as exceptions to the
general rule to be explained either by the imperfection of the paleonto-
logical record, by natural raft transportation, or by convergent evolution.
The case of Ceenokstes is mentioned by Matthew and treated as one of
convergence, the views of Broom and Dederer as to its polyprotodont
affinities being accepted. A similar view of Ccenolestes and its allies
has been taken by Gregory. The present study does not favor a great
degree of convergence, but it tends to substantiate the opinion that an
Antarctic land connection is not required to explain the presence of
C&nolestes in South America.

The many resemblances which C&nolestes shows to the Australian

north polar theory had been presented also by Haacke in 1887 and by
Haseman in 1912.

148 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

peramelids would be difficult to interpret from the distributional stand-
point were it not for the occurrence in the Eocene of North America of
Myrmecoboides, the imperfect remains of which are so suggestive of
peramelid relationship. This form seems to show that differentiation
along diprotodont lines was well advanced in North America during
the basal Eocene at an earlier period than the appearance of any of the
South American csenolestids, all of which are of Miocene age. Since
all modern peramelids are Australian while Canolestes and Myrmecoboides
are American, the simplest explanation of their distribution is found in
a common northern origin subsequent to the generalized polyprotodont
stage; in other words, it may be assumed that the main lines of diver-
gence between polyprotodonts and diprotodonts were laid down in the
northern hemisphere. In this case the supposed convergence or parallel-
ism exhibited by C&nolestes may be merely of a superficial nature, all its
fundamental characters having been derived directly from its northern
ancestor.

The absence of fossil marsupials from Asia leaves us without direct
evidence of the history of the group in that region, but a wide distribu-
tion of the early Tertiary fauna throughout the Holarctic region is
generally admitted. Hence it may be assumed that if a division of the
marsupial stem had taken place in the early Tertiary of North America
the same division or one arising in a similar manner may have extended
to Eurasia and thence to Australia. The lack of ancient fossil types from
Australia is almost as complete as that from Asia. Still it is important
to note that Wynyardia, from the Upper Eocene or at latest Oligocene,
is the oldest known Australian form and has many diprotodont char-
acters. Moreover, the diprotodonts exhibit various primitive features,
in some respects (e. g. auditory ossicles) being more primitive than the
didelphids. They reached a very high degree of specialization in the
Pleistocene of Australia when such forms as Diprotodon and Thylacoleo
were abundant. At the present time the group is decidedly the dominant
one in Australia.

The time of the introduction of marsupials into Australia has been
variously estimated from the Jurassic to the Eocene, recent authors
being inclined to the later period. Bensley concludes that it is "unlikely
that the marsupial radiation could have begun until well on into the
middle of the Tertiary period." Elsewhere (1903, p. 86) he speaks of the
possibility "that the ancestors of the Australian fauna passed the
iacipient phases of their evolution in another country, such as Asia or
even South America." The same idea is expressed by Matthew (1915,
p. 267) as follows:

MAY, 1921. AMERICAN MARSUPIAL, QENOLESTES — OSGOOD. 149

"There is nothing unlikely in the view that they (Diprotodonta)
originated primarily in the North like their polyprotodont and allo-
therian relatives and were driven southward with the former group and
somewhat more thoroughly extinguished in the north, while in Aus-
tralia they blossomed out into a great adaptive expansion paralleling
the absent ungulate mammals."

Therefore, the conclusion that the South American and Australian
marsupials had a common Holarctic origin is in accord with the conten-
tions of Matthew as to the origin of mammalian life in general. In fact
the case of Coenolestes might be regarded as exemplifying the manner in
which he has assumed that forms with irregular distribution are to be
explained. Thus it is not only a case in which some convergent develop-
ment may have occurred, but also one in which a slight addition to the
fossil record (i. e. Myrmecoboides) is of great importance.

The evidence in this case may be divided as follows:

1. The many primitive features of Canolestes indicating the
possible development of diprotodont characters very early
in the history of the order Marsupialia.

2. The existence of very primitive characters among Australian
diprotodonts (see Gregory, 1910, p. 227).

3. The large number of minor resemblances between C&nolestes
and the peramelids.

4. The probable affinity of Myrmecoboides to the modern
peramelids.

5. The correspondence of Myrmecoboides with the hypothetical
ancestor of diprotodonts, that is, the " properamelid " of
Bensley.

6. The diprotodont characters shown by the earliest known
Australian fossil, Wynyardia.

7. The lack of any special affinity between the caenolestids and
the didelphids.

8. The probability of a more ancient origin for various mar-
supial groups than has been supposed (see Gidley, 1915,
p. 400).

9. The fact that the Miocene caenolestids were specialized to a
high degree and coeval with polyprotodont types which
differed from them in essentially the same characters as those
which separate modern diprotodonts and polyprotodonts.

10. The presumption of a northern origin of all mammalian
groups as indicated by Matthew.

150 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

GENERAL SUMMARY.

The most significant results of the foregoing study may be sum-
marized as follows:

1 . C&nolestes is a surviving member of an ancient group and retains
many primitive characters.

Among these may be mentioned: (a) small size and insec-
tivorous habits; (b) in the brain — the very large olfactory bulbs
and tubercles and the small simple cerebellum; (c) in the myology
— the agreement in certain respects with edentates and insec-
tivores, as in the divided trapezius, the dorso-cuticularis , the
reduced sartorius, the sesarnoid origin of the supinator and the
absence of the pronator quadratus; (d) in the reproductive system
— the possession of a median vagina of similar primitive type to
that of Perameles; (e) in the alimentary canal — the short colon;
(f ) in the skeleton — the movable cervical rib and the short pubic
symphysis; (g) in the skull — the smooth braincase, large olfactory
fossa, proximally expanded nasals, broad frontals, small paroc-
cipital processes, annular tympanic, large persistent mastoid
foramen, and columelliform stapes; (h) in the dentition — the
large upper incisor formula only exceeded by certain peramelids
and didelphids, the large number of lower antemolar teeth, the
premolariform canines.

2. It has numerous resemblances to modern peramelids.

These are not of equal importance, but collectively seem
significant of affinity. Among them are the following: (a) in
external characters — first and fifth toes of forefeet reduced and
having nails instead of claws, ear with double antihelical folds,
limbs cursorial, tail not prehensile; (b) in the brain — general
primitive structure; (c) in the myology1 — general cursorial
adaptations; (d) in the skeleton — movable cervical rib, imper-
f orate atlas, short spine of first thoracic vertebra, ossified patella;
(e) in the skull — elongate rostrum, short palatine processes of
premaxillae, concave basioccipital, nearly perpendicular maxillo-
premaxillary suture, long fronto-maxillary union, absence of
postorbital processes, lateral exposure of mastoid, small paroc-
cipital and mastoid processes, lacrymal canal with single opening,
postzygomatic foramen lacking, stapes columelliform; (f) in the
dentition — large incisor formula, quadrate molars with hypocone,
proclivous lower incisors, compressed lateral upper incisors,

1 Adequate comparison of the myology of C&nolestes and Perameles has not been
made.

MAY, 1921. AMERICAN MARSUPIAL, C-ENOLESTES — OSGOOD. 151

premolariform canines, specialized posterior premolar; (g) in the
urinogenital system — presence of round ligament, median vaginae
indicating parturition by direct median passage; (h) in the
alimentary canal — small caecum.

3. It has few non-marsupial characters and no great degree of
specialization.

The important characters not shared with at least some other
marsupials are: (a) the preorbital vacuity found elsewhere only
among ungulates; (b) the large persistent mastoid foramen; (c)
the large carotid canal; (d) the very short pubic symphysis;

(e) the extended articular surface of the trochlea of the humerus;

(f) the sesamoid in the supinator tendon.

The cardiac gland of the stomach has a counterpart in Phas-
colarctos and Phascolomys; otherwise it is unique among mar-
supials.

4. It has no especial affinity to the American Didelphiidae.
Except for a few relatively unimportant myological char-
acters concerning which data are not available for many Australian
forms, Ccenolestes shows practically no exclusive resemblances to
the didelphids. On the other hand it differs from them in numer-
ous characters including nearly all the above-mentioned resem-
blances to the Peramelidae.

5. The ancestor of the canolestids was probably a northern form
which had already separated from the generalized polyprotodont
stock.

This conclusion is based chiefly upon the probable affinity of
C&nolestes to modern peramelids and of both to the Eocene
Myrmecoboides which partially fulfils the hypothetical require-
ments of the ancestral diprotodont. Many considerations favor
this assumption.

6. The North American ancestor of the canolestids possibly extended
throughout Holarctica and therefore may also have given rise to
the Australian diprotodonts.

This assumes that the main lines of divergence between dipro-
todonts and polyprotodonts were established in the north prior to
the Australian and South American radiations. Hence an Antarc-
tic land connection is not necessary to explain the resemblances
between American and Australian marsupials.

7. The phylogenetic and morphological relations of C&nolestes are
best expressed by classifying it in the suborder Diprotodontia,
family Pal&othentidae, subfamily C&nolestinae.

152 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

This does not emphasize its supposed phylogenetic relationship to
the peramelids, but recognizes its advance beyond them to greater
morphological similarity to the specialized diprotodonts. On the basis
of present findings, the only logical alternative would be to remove
peramelids from the Polyprotodontia and unite them with the caeno-
lestids in a group co-ordinate with (i) the Australian Diprotodontia and
(2) the remaining Polyprotodontia. This, however, could not be done
consistently without further division, especially of polyprotodonts,
which does not seem advisable in the present state of knowledge.

LITERATURE CITED.

ALLEN, J. A.

1901. A Preliminary Study of the North American Opossums of the Genus
Didelphis. Bull. Am. Mus. Nat. Hist., XIV, pp. 149-188, June 15, 1901.

ALSTON, E. R.

1880. Biologia Centrali- Americana, Mammalia. 1879-1882.

1880. On Antechinomys and its Allies. Proc. Zool. Soc. Lond., pp. 454-461,

1880.
AMEGHINO, FLORENTINO.

1897. Mammiferes cretace"s de 1'Argentine. Deuxieme contribution a la
connaissance de la faune mammalogique des couches a Pyrotherium. Bol.
Inst. Geog. Argent., XVIII, pp. 1-117, 1897.
1899. On the Primitive Type of Plexodont Molars of Mammals. Proc. Zool.

Soc. Lond., pp. 555-571, 1899.
1903. Los Diprotodontes del Orden de los Plagiaulacoideos. Anales del Museo

Nac. Buenos Aires, ser. 3, vol. II, pp. 81-192, July, 1903.
BARDELEBEN, KARL VON.

1894. On the Bones and Muscles of the Mammalian Hand and Foot. Proc.

Zool. Soc. Lond., pp. 354-376, 1894.
BATESON, WILLIAM.

1894. Materials for the Study of Variation. Pp. 1-598.
BEDDARD, F. E.

1902. Mammalia. Camb. Nat. Hist. Pp. 1-590. London, 1902.

1895. The Visceral Anatomy and Brain of Dendrolagus bennetti. Proc. Zool.
Soc. Lond., pp. 131-137, figs. 1-6, 1895.

1908. On the Anatomy of Antechinomys and Some Other Marsupials with Special

Reference to the Intestinal Tract, etc. Proc. Zool. Soc. Lond., p. 561, 1908.
BENSLEY, B. ARTHUR.

1903. On the Evolution of the Australian Marsupialia; with Remarks on the
Relationships of the Marsupials in General. Trans. Linn. Soc. London (2), IX,
pp. 83-217, 3 pis., 6 figs., 1903.

1906. The Homologies of the Stylar Cusps of the Upper Molars of the Didel-

phyidae. University of Toronto Biological Series, No. 5, pp. 149-159, 1906.
BROOM, R.

1911. On the Affinities of Canolestes. Proc. Linn. Soc. N. S. W., XXXVI,
p. 315, 1911.

MAY, 1921. AMERICAN MARSUPIAL, C^NOLESTES — OSGOOD. 153

COUES, ELLIOTT.

1872. The Osteology and Myology of Didelphys virginiana. Mem. Bost. Soc.

Nat. Hist., II, pp. 41-154, figs. 1-35, 1872.
CUNNINGHAM, D. J.

1882. Report on Some Points in the Anatomy of the Thylacine (Thylacinus
cynocephalus) , Cuscus (Phalangista maculata), and Phascogale (Phascogale
calura), collected during the Voyage of H. M. S. Challenger in the years 1873-
1876; with an account of the Comparative Anatomy of the Intrinsic Muscles
and Nerves of the Mammalian Pes. Voy. H. M. S. Challenger, Zool., V, pt. XVI,
pp. 1-192, pis. I-XIII, i8>82.
DEDERER, PAULINE H.

1909. Comparison of Ccenolestes with Polyprotodonta and Diprotodonta.

Amer. Nat., XL/Ill, pp. 614-618, figs. 1-2, Oct., 1909.
DORAN, A. H. G.

1879. Morphology of the Mammalian Ossicula auditus. Trans. Linn. Soc.

Lond., (2), I, Zool., pp. 371-497, pis. 58-64, 1879.
GIDLEY, JAMES WILLIAMS.

1906. Evidence Bearing on Tooth-Cusp Development. Proc. Wash. Aoad.
Sci., VIII, pp. 91-1 to, July 10, 1906.

1909. Notes on the Fossil Mammalian Genus Ptilodus with Descriptions of
New Species. Proc. U. S. Nat. Mus., XXXVI, pp. 611-626, I pi., 9 figs., 1909.

1915. An Extinct Marsupial from the Fort Union with Notes on the Myr-
mecobidae and other Families of this Group. Proc. U. S. Nat. Mus., XLVIII,
PP- 395-402, pi. 23, Jan. 28, 1915.

GREGORY, WILLIAM K.

1910. The Orders of Mammals. Bull. Amer. Mus. Nat. Hist., XXVII, pp. i-
524, Feb., 1910.

1916. Studies on the Evolution of Primates. Bull. Am. Mus. Nat. Hist., XXXV,
pp. 239-355, June 16, 1916.

1920. Studies in Comparative Myology and Osteology; No. IV. A Review of
the Evolution of the Lacrymal Bone of Vertebrates with Special Reference to
that of Mammals. Bull. Am. Mus. Nat. Hist., XLII, pp. 95-263, Dec. 4, 1920.
HAACKE, W.

1887. Der Nordpol als Sch&pfungscentrum der Landfauna. Biol. Centralblatt,

VI, pp. 363-370, 1887.
HASEMAN, J. D.

1912. Some Factors of Geographical Distribution in South America. Ann.

N. Y. Acad. Sci., XXII, pp. 9-112, May 31, 1912.
HEDLEY, C.

1895. Considerations on the Surviving Refugees in Austral Lands of Ancient

Antarctic Life. Jour. & Proc. Roy. Soc. N. S. W., XXIX, pp. 278-286, 1895.
HILL, JAMES P.

1899. Contributions to the Morphology and Development of the Female
Urogenital Organs in the Marsupialia. I. On the Female Urogenital Organs of
Perameles, with an Account of the Phenomena of Parturition. Proc. Linn. Soc.
N. S. W., XXIV, pp. 42-82, 12 pi., 3 figs., 1899-

1900. Contributions to the Morphology and Development of the Female
Urogenital Organs in the MarsupiaKa. II. On the Female Urogenital Organs of
Myrmecobius fasfiatus. Proc. Linn. Soc. N. S. W., XXV, pp. 519-532, pis.
XVII-XIX, 1900.

154 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

HUXLEY, T. H.

1880. On the Application of the Laws of Evolution to the Arrangement of the
Vertebrata, and More Particularly of the Mammalia. Proc. Zool. Soc. Lond.,
pp. 649-661, 1880.
JENTINK, F. A.

1885. A Monograph of the Genus Cuscus. Notes Leyden Mus., VII, pp. 87-1 19.
JOHNSTONE, JAMES.

1898. The Thymus in the Marsupials. Jour. Linn. Soc. Lond., XXVI, pp.
537-557, 3 pl- 2 figs., 1898.

1899. On the Gastric Glands of the Marsupialia. Jour. Linn. Soc. Lond., Zool.,
XXVII, pp. 1-14, i pl., i fig., 1899.

1901. The Neck Glands of the Marsupialia. Trans. Liverpool Biol. Soc., XV,
PP- 354-362, 3 figs., 1901.

LECHE, W.

1874-1900. Bronn's Klassen u. Ordn. Thierreichs, Bd. VI, Abth. V, Mammalia,

Bd. I., Leipzig, 1874-1900.
LYDEKKER, RICHARD.

1894. A Hand-Book to the Marsupialia and Monotremata. Allen's Nat. Lib.,

London, 1894.
1899. The Dental Formula of the Marsupial and Placental Carnivora. Proc.

Zool. Soc. Lond., pp. 922-928, pl. LXII, 1899.
MACALISTER, ALEXANDER.

1870. On the Myology of the Wombat (Phascolomys wombata) and the Tas-
manian Devil (Sarcophilus ursinus). Ann. & Mag. Nat. Hist., (4), V, pp. 153-
173, March, 1870.
1872. Further Observations on the Myology of Sarcophilus ursinus. Ann. &

Mag. Nat. Hist., (4), X, pp. 17-20, July, 1872.
1872. The Muscular Anatomy of the Koala (Phascolarctos cinereus). Ann. &

Mag. Nat. Hist., (4), X, pp. 127-134, Aug., 1872.
MATTHEW, W. D.

1915. Climate and Evolution. Ann. N. Y. Acad. Sci., XXIV, pp. 171-318,

Feb. 18, 1915.
METCALF, M. M.

1920. Upon an Important Method of Studying Problems of Relationship and

of Geographical Distribution. Proc. Nat. Acad. Sci., VI, pp. 432-433, 1920.
MITCHELL, P. CHALMERS.

1905. On the Intestinal Tract of Mammals. Trans. Zool. Soc. Lond., XVII,

Pt. V, pp. 437-535, 1905-
ORTMANN, A. E.

1902. Tertiary Invertebrates. Repts. Princeton Univ. Exped. to Patagonia,
IV, Paleon., pp. 45-332, pis. XI-XXXIV, April 19, 1902.

OSBORN, HENRY FAIRFIELD.

1907. Evolution of Mammalian Molar Teeth. Pp. 1—250. New York, 1907.
1910. The Age of Mammals in Europe, Asia and North America. Pp. XVIII —

635. New York, October, 1910.
OSGOOD, WILFRED H.

1912. Mammals from Western Venezuela and Eastern Colombia. Field Mus.
Nat. Hist., Zool., X, pp. 33-66, Jan. 10, 1912.

MAY, 1921. AMERICAN MARSUPIAL, C^ENOLESTES — OSGOOD. 155

PARSONS, P. G.

1896. On the Anatomy of Petrogale xanthopus, Compared with that of Other

Kangaroos. Proc. Zool. Soc. Lond., 1896, pt. 3, pp. 683-714, 10 figs., 1896.
1898. The Muscles of Mammals with Special Relation to Human Myology.

Journ. Anat. & Phys., XXXII, pp. 428-450, 721-752, 1898.
1898. The Limb Myology of Gymnura rafflesii. Journ. Anat. & Phys., XXXII,

pp. 312-324, 1898.

1903. On the Anatomy of the Pig-footed Bandicoot (Char opus castanotis).
Journ. Linn. Soc. Lond., Zool., XXIX, pp. 64-80, 10 figs., 1903.

POULTON, E. B.

1883. On the Tongues of the Marsupialia. Proc. Zool. Soc. Lond., pp. 599-628,

pis. LIV-LV, 1883.
RUGE, GEORG.

1878. Untersuchung uber die Extensorengruppe am Unterschenkel und Fiisse

der Saugethiere. Morph. Jahrb., IV, pp. 592-643, pis. XXXII-XXXV, 1878.
1878. Zur vergleichenden Anatomic der tiefen Muskeln in der Fusssohle.

Morph. Jahrb., IV, pp. 644-659, pis. XXXIV-XXXV, 1878.
SCOTT, W. B.

1913. A History of Land Mammals in the Western Hemisphere. Pp. 1-693.

SlDEBOTHAM, E. J.

1885. On the Myology of the Water-Opossum. Proc. Zool. Soc. Lond., pp.

6-22, 1885.

SINCLAIR, W. J.

1905. The Marsupial Fauna of the Santa Cruz Beds. Proc. Amer. Philos. Soc.,
XLIV, pp. 73-81, 2 pis., 1905.

1906. Mammalia of the Santa Cruz Beds. Marsupialia. Repts. Princeton Univ.
Expeds. to Patagonia, IV, Paleon., I, pp. 333-460, Sept. 14, 1906.

SPENCER, BALDWIN.

1900. A Description of Wynyardia bassiana, a Fossil Marsupial from the
Tertiary Beds of Table Cape, Tasmania. Proc. Zool. Soc. Lond., pp. 776-794,
pis. XLIX-L, 1900.
STIRLING, E. C.

1891. Description of a New Genus and Species of Marsupialia, Notoryctes

typhlops. Trans. Roy. Soc. South Aust., XIV, pp. 154-187, pis. II-IX, 1891.
STONE, WITHER.

1914. On a Collection of Mammals from Ecuador. Proc. Acad. Nat. Sci. Phila.,
pp. 17-19, Mch. 31, 1914.

SWEET, GEORGIANA.

1904. Contributions to our Knowledge of the Anatomy of Notoryctes typhlops
Stirling. Parts I and II. Proc. Roy. Soc. Victoria, N. S., XVII, pp. 76-111,
4 pis., 1904.

1907. The Skin, Hair, and Reproductive Organs of Notoryctes. Contributions
to our Knowledge of the Anatomy of Notoryctes typhlops Stirling. Parts IV-V.
Quart. Jour. Micr. Sc., LI, pp. 325-344, 2 pis. I fig., 1907.

SYMINGTON, JOHNSON.

1898. The Thymus Gland in the Marsupialia. Jour. Anat. & Phys., XXXII,

N. S. XII, pp. 278-291, figs. 1-4, 1898.
THOMAS, OLDFIELD.

1887. On the Homologies and Succession of the Teeth in the Dasyuridae.
Philos. Trans. Roy. Soc. Lond., CLXXVIII, B, pp. 443-462, pis. 27-28, 1887.

156 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

1888. Catalogue of Marsupialia and Monotremata in the British Museum.

Pp. 1-401, pis. I-XXVIII, 1888.
l895a. Descriptions of Pour Small Mammals from South America, Including One

Belonging to the Peculiar Marsupial Genus "Hyracodon" Tomes. Ann. & Mag.

Nat. Hist., (6), XVI, pp. 367-370, Nov., 1895.
18950. On Canolestes, a Still Existing Survivor of the Epanorthidae of Ame-

ghino. Proc. Zool. Soc. Lond., p. 870, 1895.

1896. On Ccenolestes n. g. of Marsupials. (Abstr.) Zool. Anz., XIX, no. 493,
p. 31, 1896.

1917. Preliminary Diagnoses of New Mammals Obtained by the Yale-National
Geographic Society Peruvian Expedition. Smiths. Misc. Coll., LXVIII, p. 3,
Apr. 10, 1917.

1920. Report on the Mammalia Collected by Mr. Edmund Heller during the
Peruvian Expedition of 1915 under the Auspices of Yale University and
the National Geographic Society. Proc. U. S. Nat. Mus., LVIII, pp. 217-
249, pis. 14-15, 1920.
THOMPSON, PETER, and HILLIER, W. T.

1905. The Myology of the Hind Limb of the Marsupial Mole (Notoryctes
typhlops). Jour. Anat. & Phys., XXXIX, pp. 308-331, pis. XXXVIII-XXXIX.
TIMS, H. W. M.

1903. Evolution of the Teeth in the Mammalia. Jour. Anat. & Phys., XXXVII,
pp. 132-140, 1903-

TOMES, R. F.

1860. Notes on a Second Collection of Mammalia Made by Mr. Fraser in the
Republic of Ecuador. Proc. Zool. Soc. Lond., p. 213, 1860.

1863. Notice of a New American Form of Marsupial. Proc. Zool. Soc. Lond.,

PP- 50-51. pl- VIII (col.), 1863.
TROUESSART, E. L.

1898. Catalogus Mammalium, V, p. 1205, 1898.
WATERHOUSE, G. R.

1846. Natural History of the Mammalia, Vol. I., Marsupiata, 1846.
WEBER, MAX.

1904. Die Saugethiere. Pp. 1-866. Jena, 1904.
WILSON, J. T.

1894. On the Myology of Notoryctes typhlops, with Comparative Notes. Trans.

Roy. Soc. South Aust., XVIII, pp. 3-74, pis. II-XV, 1894.
WILSON, J. T., and HILL, J. P.

1897. Observations upon the Development and Succession of the Teeth in
Perameles; Together with a Contribution to the Discussion of the Homologies
of the Teeth in Marsupial Animals. Quart. Jour. Micr. Sci., XXXIX, N. S.,
pp. 427-588, pis. 25-32, 1897.

WINDLE, BERTRAM C. A., and PARSONS, F. G.

1898. On the Anatomy of Macropus rufus. Jour. Anat. & Phys., XXXII, N. S.,
XII, pp. 119-134, 1898.

WINGE, HERLUF.

1882. Om Pattedyrenes Tandskifte, isaer med Hensyn til Tasndernes Former.

Vid. Medd. f. d. Naturh. Foren., Kjobenhavn, pp. 15-67, 1882.
WOODWARD, M. F.

1893. On the Development of the Teeth in the Macropodidae. Proc. Zool.

Soc. Lond., pp. 450-473, 1893.
1896. On the Teeth of Certain Insectivora. Proc. Zool. Soc. Lond., pp. 557-594.

MAY, 1921. BRAIN OF C^ENOLESTES — HERRICK 157

THE BRAIN OF C&NOLESTES OBSCURUS.
BY C. JUDSON HERRICK.

Plates XXI-XXII.

These notes are based on a single female specimen which had been
preserved in formalin and later transferred to alcohol. Since the cranial
cavity had not been opened before the specimen was put into the harden-
ing fluid, the preservation of the brain is not as perfect as might be
desired. In particular, the swelling of the brain tissue produced by the
formalin has caused considerable compression of the brain within the
endocranial cavity, thus possibly exaggerating somewhat the superficial
relief, so far as this conforms to the sculpturing of the endocranial cavity,
and obscuring some other features.

The brain was very skillfully removed from the skull of the alcoholic
specimen in the Anatomical Laboratory of the University of Chicago
by Dr. G. W. Bartlemez and drawn under the Zeiss stereo-binocular
microscope by Mr. A. B. Streedain. Since it seemed desirable to pre-
serve the brain intact in the hope of a later opportunity to prepare it
for microscopic examination, this report must of necessity be limited
to the superficial anatomy.

Measurements. — The dimensions as measured on the alcoholic
specimen' are as follows: —

Total length, tip of olfactory bulb to first spinal nerve . 14. 1 mm.

Length, tip of olfactory bulb to rostral end of cerebral hem-
isphere 2.6 mm.

Length of cerebral hemisphere 10.0 mm.

Length of cerebellum on longitudinal axis of brain in me-
dian plane 3.0 mm.

Greatest width of both olfactory bulbs 7.6 mm.

Greatest width of both cerebral hemispheres . . . . 1 1 . 8 mm.

Total width of cerebellum and flocculi 1 1 . o mm.

Width of cerebellum exclusive of flocculi 8.8 mm.

For our most precise knowledge of the brains of lower mammals
we are indebted to the researches of Elliot Smith. Throughout this
account we shall make frequent references to his papers and base our
interpretations to a large extent upon them. His papers cited in the
appended bibliography give references to the other relevant literature.

The general form of the brain of this little marsupial is evident from
the figures. It is apparent that we have here before us one of the
simplest types of mammalian brain hitherto described. The brain is
strongly macrosmatic, as shown by the enormous size of the olfactory
bulbs and secondary olfactory area.

158 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV.

The olfactory bulbs are closely appressed to each other in the me-
dian plane, convex above and concave below. They are very slightly
overlapped dorsally by the cerebral hemispheres. From the widely
flaring lateral borders of the bulb the lateral olfactory tract passes spinal-
ward and ventralward along the lateral border of the very large tuber-
culum olfactorium.

The cerebral hemispheres extend backward quite to the cerebellum,
these two structures being in intimate contact for the entire width of the
body of the cerebellum. There is a very slight divergence of the posterior
borders of the hemispheres from the median plane, which is somewhat
exaggerated in Figure i (PL XXI), so that even if the meninges were
entirely removed from the median longitudinal fissure (as has not been
done) , but little if any of the mesencephalon would be visible from the
dorsal surface. This is in contrast to the usual marsupial arrangement,
for the corpora quadrigemina are in most cases well exposed dorsally.
(Petaurus is another exception; see Elliot Smith, '95, p. 168.)

Superficially the cerebral hemisphere, exclusive of the olfactory
bulb, exhibits three chief regions: — (i) the dorsal convexity; (2) the
lateral convexity, or pyriform lobe; (3) the ventral convexity, or
tuberculum olfactorium.

The dorsal convexity of the hemisphere is purely neopallial; that is,
it is non-olfactory cortex. About one-fifth of the distance backward
from the frontal to the posterior pole of the hemisphere there is a
distinct, though shallow, transverse sulcus which probably represents
the sulcus orbitalis of Elliot Smith's descriptions. Otherwise the dorsal
convexity is smooth.

The dorsal convexity is bounded laterally by an imperfectly devel-
oped fissura rhinalis. This begins anteriorly as a sharp sulcus in the
transverse fissure between the olfactory bulb and the cerebral hemisphere
at the dorso-lateral angle of the tuberculum olfactorium (PI. XXI, fig. 2)
and extends backward. On the lateral aspect of the hemisphere it is
obscurely confluent with the orbital sulcus, and behind this level it
disappears in a depressed area on the lateral wall of the hemisphere.
Two-thirds of the distance back from the anterior to the posterior pole of
the hemisphere this depressed area is slightly deepened, thus marking
more precisely the location of the fissura rhinalis in this region; and at
the posterior end of the hemisphere there is a wide, shallow notch which
marks the posterior end of this fissure (PI. XXI, figs, i and 2).

The pyriform lobe (lobus piriformis) comprises the larger part of
the lateral and ventral aspects of the hemisphere. As we have seen
above, it is very imperfectly separated from the dorsal neopallial cortex,
though the location of this boundary, the fissura rhinalis, is evident.

MAY, 1921. BRAIN OF C^ENOLESTES — HERRICK 159

Antero-ventrally it is separated from the tuberculum olfactorium by a
very sharp fissura endorhinalis. The ventral surface of the pyriform
lobe shows two shallow depressions separated by an elevated ridge run-
ning obliquely backward and lateralward from the posterior margin of
the tuberculum olfactorium. Further information regarding the internal
structure is desirable before the signification of this sculpturing is ex-
plained. It may be due merely to the conformation of the brain to the
wall of the cranium.

The tuberculum olfactorium is very large and strongly convex
ventrally. The medial borders of the two tubercula are divaricated
posteriorly, exposing a portion of the anterior perforated space in front
of the optic chiasma which probably includes the diagonal band of
Broca (PL XXII).

The lateral olfactory tract arises from the ventrolateral border of
the olfactory bulb and can readily be seen as a clear white stripe accom-
panying the fissura endorhinalis. This band of fibers lies distinctly on
the ventromedial side of the fissura, that is, within the tuberculum
olfactorium, though we may infer by analogy with other mammals that
many of the fibers are distributed within the pyriform lobe on the other
side of the fissure. Microscopic examination will probably show that a
portion of the tract lies in the floor and walls of the fissure.

The tuberculum olfactorium in mammals generally is a basal, that
is, subcortical, reflex center, receiving olfactory fibers of the second
order from the olfactory bulb. The pyriform lobe, on the other hand,
is a structure of transitional type. So far as it and the underlying
amygdala receive secondary olfactory fibers from the lateral olfactory
tract it should be considered as a part of the basal secondary olfactory
area (nucleus olf actorius lateralis) . So far as its differentiated cortex is
in physiological connection with this secondary olfactory area it should
be regarded as archipaliial in type, i. e., olfactory cortex of the same type
as the hippocampal cortex. So far as its cortex is in physiological con-
nection with the non-olfactory thalamic projection fibers it is neopallial,
i. e., of the same type as the dorsal cortex. The surface characteristics
suggest that in Ccenolestes these three components of the pyriform lobe
are very incompletely differentiated, thus resembling the still more
generalized reptilian condition (cf. Elliot Smith, '10, and Crosby, '17).
In higher mammals, on the other hand, these components of the pyriform
lobe attain much more distinct spatial localization in the gyrus hippo-
campi formation, though even in the human brain we find the same
transitional type of structure in the uncus region.

The cerebellum. — The cerebellum is smaller and simpler than in any
hitherto described mammals with the exception of Notoryctes and

160 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

Perameles, though the elephant-shrew, Macroscelides, has a cerebellum
but little more complex than that of Perameles (Elliot Smith, 'oab).

The cerebellum is composed of a simply organized transverse body
and the two lateral floccular lobes. (In the preparation of the brain the
left flocculus was destroyed. In the figures it has been restored after the
one on the opposite side.) Each flocculus is mushroom-shaped with a
slender pedicle and a widely expanded cap or pileus, whose surface is
somewhat lacerated but apparently was smooth or nearly so.

The body of the cerebellum, so far as it is visible from the surface, is
a transverse bar with two shorter convolutions under its posterior border.
The main bar consists of a plump median lobe and two lateral lobes,
each of which is somewhat less than half the length in th ± transverse
plane of the medial lobe. The median lobe and the two smaller posterior
convolutions apparently correspond with the vermicular portion of
higher cerebella and the lateral lobes with the hemispheres.

Since only a part of the cerebellum is visible in the undissected
specimen, it is impossible to determine with certainty the homologies
of the exposed structures. But comparison with the very similar
cerebella of Notoryctes and Perameles as described and figured by Elliot
Smith ('osa and 'o3b) suggests that in Ccenolestes the lobus anterior and
fissura prima are entirely concealed in the transverse fissure between
the cerebellum and the cerebral hemispheres, the transverse bar is the
lobus medius, the fissure limiting it posteriorly is the fissura secunda, the
first of the two shorter convolutions is the uvula, the fissure behind the
latter is the fissure postnodularis, and the posterior one of the shorter
convolutions is the nodulus, these names being used as defined by
Elliot Smith ('oaa).

The brain stem. — On the ventral surface (PI. XXII) the optic nerves
are seen to be very slender. Behind the optic chiasma is a rather wide
tuber cinereum whose surface is somewhat obscured by meninges which
in view of the poor state of preservation of the specimen it seemed
inexpedient to attempt to remove. The pituitary body is mushroom-
shaped and elevated on a very short infundibular stalk. On account
of the flexure of the brain in the isthmus region the cerebral peduncles
are entirely concealed. Stumps of the third and fourth cranial nerves
are seen in the usual relations.

The medulla oblongata under the cerebellum is very wide. The
sculpturing^ of the ventral surface is obscured by pressure against the
floor of the cranial cavity and posteriorly by a blood clot in the meninges.
Nevertheless the roots of all of the cranial nerves were identified. The
trigeminus is large, arising from the posterior border of the pons, which is
very slender. The abducens is very minute and is recognized (with

MAY, 1921. BRAIN OF C^ENOLESTES, — HERRICK 161

some uncertainty) at the posterior border of the pons. Immediately
below the pons is a broad, flat, transverse band which is probably the
trapezoid body and at whose lateral ends are the stumps of the VIII
nerves. The facialis root arises from the infero-lateral surface of the
trapezoid body; the IX, X and XI nerves arise from the lateral surface
of the medulla oblongata; and medially of these are the XII roots. The
pyramidal tracts are visible as light colored bands near the mid-ventral
line extending downward from the posterior surface of the pons. At the
level of the first spinal roots there is a strong cervical flexure.

General considerations. — Comparing the brain as a whole with those
of other mammals, it is seen to resemble most closely those of Notoryctes
and Perameles. In all three cases the rhinencephalon is enormously
developed, the olfactory bulbs and tubercles being very large. The
cerebral cortex is nearly smooth and apparently relative to the total
size of the brain less extensive than in any other mammals hitherto
described. Elliot Smith's figures of Notoryctes ('95) indicate that in this
genus the cerebral hemispheres are relatively smaller than in Ccenolestes;
his figure of the ventral surface of Perameles ('02, p. 171, fig. 52) suggests
that here the hemisphere is relatively as large as in C&nolestes or larger.
In absolute dimensions the brain of Notoryctes is about the same size as
that of C&nolestes, while that of Perameles is more than three times as
long. The simplicity of these brains cannot, therefore, be correlated
directly with the size of the animals. In fact, our figure of the ventral
view of the brain of C&nolestes resembles more closely Elliot Smith's
figure of the ventral surface of the larger Perameles than of the
Notoryctes of equal size. Both Notoryctes and Perameles belong to
the Polyprotodontia. The larger members of the Diprotodontia, such
as the kangaroos, have larger and more highly convoluted brains than
any of the Polyprotodontia. The brain of C&nolestes is more simply
organized than that of any Australian diprotodont.

LITERATURE.

CROSBY, ELIZABETH CAROLINE. 1917. The Forebrain of Alligator mississippiensis.

Jour. Comp. Neur., Vol. 27, pp. 325-402.

SMITH, G. ELLIOT. 1894. A Preliminary Communication upon the Cerebral Com-
missures of the Mammalia, with special Reference to the Monotremata and
Marsupialia. Proc. Linnean Soc. New South Wales, Series 2, Vol. 9, pp. 635-657.
1895. The Comparative Anatomy of the Cerebrum of Notoryctes typhlops.
Trans. Roy, Soc. South Australia, 1895, pp. 167-193.

1 899. Further Observations on the Anatomy of the Brain in the Monotremata.
Jour. Anat. and Physiol., Vol. 33, pp. 309-342.

162 FIELD MUSEUM or NATURAL HISTORY — ZOOLOGY, VOL. XIV.

18993. The Brain in the Edentata. Trans. Linnean Soc. London, Series 2,
Zoology, Vol. 7, Part 7, pp. 277-394.

1902. Descriptive and Illustrated Catalogue of the Physiological Series of
Comparative Anatomy contained in the Museum of the Royal College of Sur-
geons of England, 2 Ed., Vol. 2.

I902a. On a Peculiarity of the Cerebral Commissures in Certain Marsupialia,
not hitherto Recognized as a Distinctive Feature of the Diprotodontia. Proc.
Roy. Soc. London, Vol. 70, pp. 226-231.

19020. Notes on the Brain of Macroscelides and other Insectivora. Jour.
Linnean Soc. London, Zoology, Vol. 28, pp. 443-448.

I902C. The Primary Subdivision of the Mammalian Cerebellum. Jour.
Anat. and Physiol., Vol. 36, pp. 381-385.

1903. Notes on the Morphology of the Cerebellum. Jour. Anat. and Physiol. ,
Vol. 37, pp. 329-332.

19033. Further Observations on the Natural Mode of Subdivision of the
Mammalian Cerebellum. Anat. Anzeiger, Bd. 23, pp. 368-384.

igosb. On the Morphology of the Brain in the Mammalia, with special
Reference to that of the Lemurs, Recent and Extinct. Trans. Linnean Soc.
London, Series 2, Zoology, Vol. 8, Part 10, pp. 319-432.

1910. The Arris and Gale Lectures on Some Problems Relating to the Evolu-
tion of the Brain. The Lancet, Jan. i, 15, and 22, 1910.

.11 3TAJS

JAX33T7.3

an 29mit

.ratrhfiniih

.tool bnirf lo 3lo8 .£.
t aiol io s'

PLATE II.

EXTERNAL CHARACTERS.

Three times natural size.
Fig. i. Head.

Fig. 2. Dorsal aspect of rhinarium.
Fig. 3. Sole of hind foot.
Fig. 4. Sole of fore foot.

FIELD MUSEUM OF NATURAL HISTORY.

PLATE II, ZOOLOGY.

EXTERNAL CHARACTERS.
Three times natural size.

T;;: •.••RAW

OF THE
UNIVERSITY Of ILLINOIS

ki x->

=

- H
i

»

i T S

~ £

..

to

1 %

u
to

, pectoralis;
., rectus abdominis

., rectal gland;
,, scalenus;

,, scrotum;
, serratus magnus;

, semitendinosus;
., sphincter cloacae

, spino-deltoideus;

, spino-trapezius;
, sterno-hyoideus;

, sterno-mastoideu
, tibialis anticus;

-JJ -t-3

& a

rt rt
y y

to" to"

On G,

y tu
y y
'C 'C
•»-> -u

teres;
, temporalis;
, vastus externus;

, zygomaticus.

tn

*, "•

<xc w

iS fc

* £

•a

r1 "**

£ «

"w -2.

•s 1 »•

**l

J

O

** -•

"H

£

X

a

E

• - in

D

•§ "S

U2

5fi 3

fa

C g

.*.

0

~ 1

la
3

E/3

•c

^

x S

9

•9

1

.w

fe

^ e

'-^ H

8

.. v>

CS 0

JSCLES AFTER REMI

to

13

1

c
$

, extensor carpi rad
, extensor digitorui

and lateralis;
, fascia over rhomb

, femoro-coccygeus
, flexor carpi ulnari

, flexor digitorum f
, gastrocnemius;

, gluteus maximus;

, iliacus;
, latissimus dorsi;

, mandibulo-auricu
, ramus of mandibl

, inferior masseter;
, superior masseter

obliquus externus
, obliquus internus;
omohyoideus;

, peronei;
, psoas magnus;

M

o

-*! -a

**->

u- 8.

•*> «w

s

^ -a'

-2 ^

•» <*>

^ *** g

v. g

V,

&

'w

**"•» <0

is

3,

si

**

a

£ **
$ §

•%•% i

•a ^

til

^f^

H

D

0

,

fa

it

O

<u

la

H

is

'£

a

H

%

£

1

a

CLi

o

n

_to

rrt

0

y

s

cd

K

a
z
w
O

, anconeus;
, acromio-deltoideu!

, acromio-trapezius;
, brachialis anticus;

, biceps brachii ;
, biventer cervicis;

, biceps femoris;
, brachioradialis;

, buccinator;

, cruro-coccygeus;
, muscular bulb of c

nosum ;
, cleido-occipitalis;

,, Cowper's glands;
, muscular bulb of

giosum;
, dorso-cuticularis 0
, digastricus;

, extensor carpi radi

a •«

"G e.

5

S g

«T

S ^

•°.

U

0

a o

*0 .0

<0

'•

^. 'i

"*

fts

i;;: :

OF THt

OF ILLINOIS

*p
£

rsariu:

tn

o

•£, D

* *

U

W

Q

;moval of trapezius

d

N

"55

omo-cleido-transv

ventralis;

, omohyoideus;

, rhomboideus capi
, rhomboideus vert

, retractor naris;

, splenius;
, sterno-hyoideus;
, sterno-mastoideus
, temporalis;

, zygomaticus.

.2
'55

rt

ibmaxillary gland.

Q

B

S

s»

1

S ^

K

•ft.-* g "»

%

C

B

vl

03

W

CD

-4-J

•

W

%^ **•

V)

„ *

CD

^*

X

"rt

C

0

•

.6

|

>

X

^

u

N

^H

—

H

jg

_E

-»J

UJ

fe.

a

9

F

0

*T3

,..

0

_i

to

Q

*rf

*4-<

Q.

Q

cfl

J5

(L)

fc

'O

O

O

|

o

1

G

Q

*C

c

|

c3

w

a]

fZ

<u

Cu

*/r

a

. In CD

C

MUSCLES A

'£

•8

o

1

0

c

, acromio-deltoideus

, buccinator;

, cleido-mastoideus;

, cleido-occipitalis;
, digastricus;

, mandibulo-auricul

, inferior masseter;
, superior masseter;
, tendon of massete
, omo-cleido-transv

dorsalis;

Salivary glands. O

u

"c

5

"o
i
p<

^

s

...

0 "«

Q

•<* <<i •«' "«

r

?

*o

»^

,^

t^j

•§

% to' t^ •^-

ti

.

v/

*"*

Q

Q

S S S «

•<x

•*x

M

ei

S

S

T;;Z LIBRARY

OF THE
UNIVERSITY OF ILLINOIS

.

: • . '

'.).'«

•at ,.a.<j jguimsnooi^gjr

:•! mcni gateiol irfgtH .s

;au9noon£ ,.»

;aK
• iioj-ohbb

'• ' '"
iztst

• •'•

PLATE V.

MUSCLES OF THE LEGS.

Two and one-half times natural size.

Fig. i. Left hind leg from outside.

b.f., biceps femoris;
c.c., cruro-coccygeus ;
cf., caudo-femoralis;
e.c., extensor-caudae;
f.c., femoro-coccygeus;
J.d.f., flexor digitorum fibularis;
g., gastrocnemius ;

g.m., gluteus maximus;

p., peronei;
r.f., rectus femoris;
sm., semimembranosus;
st., semitendinosus;
t.a., tibialis anticus;
v.e., vastus externus.

Fig. 2. Right foreleg from inside.

a.

b.a.

b.b.

br.

d.

cl.m.

cl.o.

c.m.

d.

e.c.r.
et.

f.c.r.
f.c.u.
f.d.p.

l.s.

0.

anconeus;

brachialis anticus;

biceps brachii;

brachioradialis;

clavicle;

cleido-mastoideus ;

cleido-occipitalis; '

cutaneus maximus;

deltoideus;

extensor carpi radialis brevis;

epitrochlearis ;

flexor carpi radialis;

flexor carpi ulnaris;

flexor digitorum profundus;

latissimus dorsi;

levator scapulae;

omohyoideus;

o.c.t.d., omo-cleido-transversarius

dorsalis (atlanto-scapularis) ;
o.c.t.v., omo-cleido-transversarius

ventralis (atlanto-acromialis) ;
p. I., palmaris longus;
p. mi., pectoralis minor;
p.mj., pectoralis major;
p.t., pronator teres;
sc., subclavius;
s.m., serratus magnus;
ss., subscapularis;
tr., trapezius;

t.c.L, triceps brachii, caput laterale;
t.c.lo., triceps, caput longum;
t.c.m., triceps, caput mediale;
t.m., teres major.

FIELD MUSEUM OF NATURAL HISTORY.

PLATE V, ZOOLOGY.

p.t. \ f.d.p- p.i \ a. t.c.lo

\ , I

e-c.r. f.c.r. f.c.u-

MUSCLES OF THE LEGS.
Two and one-half times natural size.

OF THt
DIVERSITY OF ILLINOIS

cL m, —
>a J ax

t _/.»iC
Ml

:.-::V .I..!' , •.! -MI-

;au9bioix^J-om9Ja ,.\.\?
;abrul3 bioryrfi ,.\s.\
.M»wo^r!-<nYrfJ ,.A.\

\

•_*-*=— ,J A. '! '

;air^

;ah(

-

'

!u ,.V>

;ar.

;bn£lg oitcflqrrrif ,
;au ,.o

•ino t.b.\.^.

•-ia isnnT .£ .:

PLATE VI.

MUSCLES OF THE THROAT AND INNER SIDE OF HIND LEG.
Three times natural size.

Fig. i. Muscles of the throat.
d.m., cleido-mastoideus;
d.o., cleido-occipitalis;

d., digastricus;
gh., geniohyoideus;
hg., hyoglossus;
l.gl., lymphatic gland;

o., omohyoideus;
o.c.t.d., omo-cleido-transversarius
dorsalis;

o.cJ.v., omo-cleido-transversarius

ventralis;
sc., scalenus;
st.h., sterno-hyoideus;
st.m., stemo-mastoideus;
st.t., sterno-thyroideus;
t.gl., thyroid glands;
t.h., thyro-hyoideus.

Fig. 2. Inner side of hind leg showing innervation of sartorius muscle by

saphenus nerve.

a.L, adductor longus;
a.m., adductor magnus;

c.n., crural nerve;

g . , gastrocnemius ;
g.m., gluteus maximus;

gr., gracilis;

l.p., ligamentum poupartii;

o.e., obliquus externus;

o.i., obliquus internus;

p., pectineus;
ps.m., psoas major;
f./., rectus femoris;
sa.t sartorius;
sm., semimembranosus;
s.n., saphenus nerve;
St., semitendinosus;
/.a., tibialis anticus;
v.i., vastus internus.

Fig. 3. Same as Fig. 2 with sartorius muscle in situ.

FIELD MUSEUM OF NATURAL HISTORY.

PLATE VI, ZOOLOGY.

d.

ps....

MUSCLES OF THROAT AND INNER SIDE OF THIGH.
Enlarged.

,

OF W

»LUHQIS

naJnA .«r

mm-.

;niav leiioq ,.«.<
al IfiiJnso Jrfgh ,.a.

'mp. to noiaivrb LsbirBo ,.oa.^i
[fqg 'lo noiaivib Isineio ,.Ti.4i

.£V60 snsv ..

;

0.5

':-.

'< • ;l3bba!d Hag ,.<i.s
;}Di;b. oiisqsrl ,-b-A

adol

• i)M3

.nsv TO cnooofiO .£ .gt

.(sx) rfoarnoi?; .g-

^. : :ba

/fliw ••

J i

PLATE VI I.

PARTS OF ALIMENTARY CANAL.
Two to three times natural size.
Fig. i. Spleen (xa).

Fig. 2. Ventral view of liver and adjoining organs. Anterior part of liver ele-
vated and posterior lobes spread apart. Duodenum reverted and displaced to left

c.d., cystic duct; p.v., portal vein;

c.l., caudate lobe of liver; r.c., right central lobe of liver;

d., duodenum; r.l., right lateral;

d.ch., ductus choledochus; r.v., renal veins;

g.b., gall bladder; sp.ca., caudal division of spighelian lobe;

h.d., hepatic duct; sp.cr., cranial division of spighelian lobe;

/./., left lateral lobe of liver; st., stomach;

oe., oesophagus; v.c., vena cava.
p.d., pancreatic duct;

Fig. 3. Caecum or vermiform appendix (x3).
Figs. 4-5. Stomach (xa).

e.g., cardiac gland; gl., glandular structure of stomach

c.s., cardiac division of stomach, with after removal of outer coat;

specialized glands; oe., oesophagus;

d., duodenum; oe.s., oesophageal division of stomach;

p.s., pyloric division of stomach.

FIELD MUSEUM OF NATURAL HISTORY.

PLATE VII. ZOOLOGY.

PARTS OF ALIMENTARY SYSTEM.
Enlarged.

Ti,I 13RARY
OF THt

OF ILLINOIS

~*s

/

f\5

.«*

2

a,

y/

,.,i
^ -|

JT

•x

ASTR qG^GtGUi
AGO3 C3A3!

iti lonojaoq

srofsuoi. nu
ntcfGi.;

3.
a

~ix <~,

w S

jTStie I final ,
m lot

i !

-I'
' £,

C 5

i. £

1 "^ a ? F

••: § E
3 g

tw

» s

<4

£ 2

^i

S

^

. rr

Is -*• •

S

O

g

\ ,-

r

'V* a
. • / &>

O

n

v

' S <S

w

>^ o

<

CD

t ^

^

u

' Cr

• «rf

£+'

5

t*'

o

s

m

• **

i

o

,<

L\>

|

s

8

Q

•s

Vj

w

• •

f

•^ .

•

9

a

8

X!*

i

o

*-^

(J>

* *

*_ •

C*1
!W

1

» ^S

§1

g

-•

to 5

•j- — a 3

S "a.

K

»

"S ^

aJ

V

fc o 3 ^

^c

*•»

•5'

S "§

8 w-

cw

Q.

<; H_-

: «

Jt

o S'

"c H,

|

5" ST

Q. g

ll

rl

£*, vJ

0;a

g;r

«.

•4

'i
§

i.J-

S" 3
3 <£

B !L *; «! .

3' 8 * a c£-

life ^

^

^ g

:/

H
0

X. 4>

rt S

03 l-«

>?

c §

9

rt

NN *O

u

"3

lH

"c

(U

° o
G -o

03 .
.*? CJ

11

o a

J-< C/3

, ureter;
, uterine body

, uterine neck;

, urogenital sir

, lateral vaginj

7l •- 3

x >i S

« r. G

£ o •- tJ o '3

.2 ta S 3 ° >
r* TO C o
;> +j ^< .— i :i ,_ ,

•a lilli

w •? a) <u a> <u
^ CX i-, ^ w, u

•- VH

M <U

rl -t^>
1 g

, anterior uretl
, posterior urel

, vena cava;

, vas deferens.

S

w

4J

6

^ I

S1

Q

;

VJ

a) J" ^ " • r

**•* «s

s a

^

"jj

rt

a

(D

G

t/5

"c

w

&

a

0

K*

O

w

*-

U

4>

0

f

1

'-2
o

i

g

J

"S

5

H

rt

1
a

0)

i

«H M fO

S
O

1

So

^

i

E

o 6 o

o

-^T

O

o

C

*~2

• ».

aj ^ «

oJ

a.

«^

g
<U
fo

.S5

ladder;
road ligament;

loacal opening;
dlopian tube;

emphatic gland

ledian vagina;

£

S
>

1 £ 11
w «2 ^ ,w

_ti *3 .. o3 <u t>

E § ^ ^§ §* §"
'S o ^3 ° °

ulb of corpus c
loacal opening;

a ..

'— '/;

s s

"o ^

^c

1
?|

1 1

cL to
§ g

E

JQ J3

0 „_,

— '

a

o

OJ 03 X2 O O^

~— u

J3 0)

*

PH O,

•*

Ti,Z '-"WARY
OF THt

f ILUNOIS

AC*

.hajr

07/7 .,;-T-': ,-!;; :-:K>:f" ,-nus alisq te waiv fsgioQ .1 .g

•(&•' :af rnoi^ iavomsi

.(dx) atieq gnkii- :' o : ;-;, cpwsomaveo auqioo lo Jlud 5o wsiv teinaV .s .31'?
iniocba ridiv/ BII: i »^f*nib^i§no^>l£ .si*?

•.'ud ,.».i
w.<$

. .->.-*

.

.

1

.

'

'

•

•^.

./^.W'

rgq ,.<^

'

PLATE IX.

DISSECTIONS OF MALE ORGANS.
Enlarged.

Fig. i. Dorsal view of parts surrounding bulbous urethra. Two Cowper's
glands removed from left side

Fig. 2. Ventral view of bulb of corpus cavernosum and adjoining parts (x6).
Fig. 3. Longitudinal section of urethra with adjoining parts (xs).
b.u., bulbous urethra;
b.w., body wall;
c.c., corpus cavernosum;
c.c.m., muscular bulb of corpus cavernosum;

c.s., corpus spongiosum;
c.s.m., muscular bulb of corpus spongiosum;

gl., lymphatic gland;
g.p., glans penis;
i.t., tendon attaching corpus cavernosum to ischium;

I., median line;

ml,mz,m?, small paired muscles at anterior end of bulbous urethra;
p.p., penis pouch;
pr., prepuce;
p.t., tendon running to pubis;

r., rectum;
r.m., rectal muscle;
r.p., rectractor muscle of penis;
sp.f., fascia to covering of bulb and to sphincter;

t.f., tendinous fascia connecting bulb and muscle in front of bulbous urethra;
u.a., anterior urethra.

FIELD MUSEUM OF NATURAL HISTORY.

Cp'

cp

g-p

DISSECTIONS OF MALE ORGANS.
Enlarged.

.

OF TH£
HHIVtRSlTY OF ILLINOIS

OJ I .X 3TAJW
CKfi ,YflAT>J3Hl

bn£ 279J911! to egrribng gaiworie ic>

bns 8§nnl lo naqas
•;x)9t,'3no;
ix) 9vod£ moil auyiot lo 92^9 .^ .gi^

0819V2I1B1T .g .§i1

-igiqs ,.<

lo adol t
lanmol

;§ar;i lo gdol Jrigh io noisivib
;§nul io sdol id§h lo noir

Re

PLATE X.

PARTS OF REPRODUCTIVE, ALIMENTARY, AND RESPIRATORY SYSTEMS.

Enlarged.

Fig. i. Dissection of base of bladder showing endings of ureters and vasa
deferentia (x5).

Fig. 2. Ventral aspect of lungs and trachea (x2j^).
Fig. 3. Lateral view of tongue (x3^).
Fig. 4. Base of tongue from above (xio).
Fig. 5. Transverse section of prostate gland (xs).
az., azygos lobe of lung;
bl., bladder;
c.p., coronate papillae;
c.v.p., circumvallate papillae ;

ep., epiglottis;
fg.p., fungiform papillae;
/./., left lobe of lung;
p.t., columnar tissue of prostate gland;
r.l.a., anterior division of right lobe of lung;
r.l.p., posterior division of right lobe of lung;
u.a., anterior urethra;
v.d., vas deferens.

FIELD MUSEUM OF NATURAL HISTORY.

PLATE X, ZOOLOGY.

r. I. a

fs-P-

f.p.

PARTS OF REPRODUCTIVE, ALIMENTARY, AND RESPIRATORY SYSTEMS.
Enlarged.

OF THt
HMIVF.BSITY Of ILLINOIS

.IX 3TAJ9

,BBliA .s
Isaiob ,ZB!JA .£
onoj at noiiahfiv sarworte alsubivibai JnawBib moil taggr T 2 .d-|, .

jaloisdiri Ifiiay

;rio}on
rzsaocnq auoniqa lo eni

,.\.w
bioinobo ,.<\.o

,.*.*

,/rt.
.svooig B ot baoubei ,.S-^

PLATE XI.

CERVICAL VERTEBRAE.

Five times natural size.
Fig. i. Atlas, posterior aspect. *
Fig. 2. Atlas, anterior aspect.
Fig. 3. Atlas, dorsal aspect.

Figs. 4-6. Seventh cervical from different individuals showing variation in form
and occurrence of vertebrarterial foramina.

Fig. 7. Axis, left lateral aspect.

a.z., anterior zygapophysis;
h.t., hypapophysial tubercle;
i.v., intervertebral notch;
l.s., lamina of spinous process;
n.f., nutrient foramen;
o.p., odontoid process;
p.p., pleurapophysis or cervical rib;
p.z., postzygapophysis;
tr., transverse process;
v.c., vertebrarterial canal;
v.c.g., vertebrarterial canal reduced to a groove.

FIELD MUSEUM OF NATJRAL HISTORY.

PLATE XI, ZOOLOGY.

n.f.

li.t.

- tr.

v.c.g.

o.p.

CERVICAL VERTEBRAE.
Five times natural size.

p.p.

OF THt

OF ILLINOIS

flp nxiH

^mpoqz^ lohateoq .ir/mol JdjjLJI

:Jfi bos
!t;cl3 bits fircfrj

. --:!: rfDOT} 19ifi9i3 ,-i-S
T b£9f{ ,.A

jJni ,.i.»

oBifooiJ isaaal ,.i.V
njHtnala lo mnhdrrnfiin ,.nt
) ,.<\

,»H

«.t

d| bti i ,.\.\

• - • : -

;riqix ,.x
..^.x

,-V^

2

PLATE XII.

BONES OF STERNUM, HIND LEG, AND THORACIC REGION.

Five times natural size.
Fig. I. Right femur, posterior aspect.
Fig. 2. Sternum and attachment of ribs.

Fig. 3. Right tibia and fibula with adjoining part of femur, outer aspect.
Fig. 4. First and second ribs and attachments, outer aspect.
Fig. 5. Clavicle.

c1., seventh cervical vertebra;
/., fibula;
fb., fabella;

g.t., greater trochanter;
h., head of femur;
i.r., intertrochantenc ridge;
l.t., lesser trochanter;
m., manubrium of sternum;
p., osseous patella;
r1., first rib;
s., shaft of femur;
/., tibia;

t.t., rudimentary third trochanter of femur;
t1., first thoracic vertebra;
i2., second thoracic vertebra;
x., xiphisternum;
x.c., xiphoid cartilage;
2-7., costal attachments of second to seventh ribs.

FIELD MUSEUM OF NATURAL HISTORY.

PLATE XII, ZOOLOGY.

h.

it.— —4

BONES OF STERNUM, HIND LEG, AND THORACIC REGION.
Five times natural size.

OF THt

'DIVERSITY OF ILLINOIS

.IIIX 3TAJR
HAh

JJo

I;

;E..

i
.vr ..\.i "

PLATE XIII.

SCAPULA, PELVIS, AND LUMBAR VERTEBRAE.

Five times natural size.

Fig. i. Lumbar vertebrae, left lateral aspect.
Fig. 2. Lumbar vertebrae, ventral aspect.
Fig. 3. Scapula, right lateral aspect.

Fig. 4. Pelvis, right lateral aspect.
ac., acromion;
co., coracoid process;

i., ilium;

ip.t., iliopectineal tubercle;
i.t., ischial tuberosity.

FIELD MUSEUM OF NATURAL HISTORY.

PLATE XIII, ZOOLOGY.

SCAPULA, PELVIS, AND LUMBAR VERTEBRAE.
Five times natural size.

OF THt

>JS

*>.D av
aomiJ
iv I.cbjj

Lfibueo dimti oi bno

8^ 8« «^ 2-»

PLATE XIV.

SACRAL AND CAUDAL VERTEBRAE.

Five times natural size.

Fig. i. Sacral and first six caudal vertebrae, dorsal aspect.
Fig. 2. Second to ninth caudal vertebrae, ventral aspect.
A,e*.,c'.,A; caudal vertebrae;
ch., chevron bone;
s1., first sacral vertebra.

FIELD MUSEUM OF NATURAL HISTORY.

PLATE XIV, ZOOLOGY.

SACRAL AND CAUDAL VERTEBRAE.
Five times natural size.

Of THt

OF ILUlUfc

iH
I -6"''

!
'

.

-A

,

.u

PLATE XV.

HUMERUS, ULNA, AND RADIUS.

Five times natural size.

Fig. i. Right humerus, posterior or caudal aspect.
Fig. 2. Right humerus, anterior or cranial aspect.
Fig. 3. Right radius and ulna, outer lateral aspect.
Fig. 4. Right ulna, anterior aspect.
ca., capitellum;
d.r., deltoid ridge;
e.f., epicondylar foramen;
g.s., greater sigmoid cavity;
g.t., greater tuberosity;

h., head of humerus;

i.a., inner articular surface of distal end of humerus;
*'.£., inner condyle;
/./., lesser tuberosity;
ol., olecranon and fossa;
r., radius;

s.r., supinator ridge;
tr., trochlea;
«., ulna.

FIELD MUSEUM OF NATURAL HISTORY.
k.

PLATE XV, ZOOLOGY.

ol.

HUMERUS, ULNA, AND RADIUS.
Five times natural size.

, ..

OF TKt
DIVERSITY nF ILLINOIS

.(-. . . '

.
:•:) l3$qZB r

.

Jen ,.OR
rsa ,.\.«

PLATE XVI.

MANUS AND PES.

Enlarged.

Fig. i. Right manus, anterior aspect (x5).
Fig. 2. Right pes, anterior aspect (xs).
Fig. 3. Left astragalus, posterior aspect (xio).
Fig. 4. Left astragalus, anterior aspect (xio).
Fig. 5. Right os calcis, anterior aspect (xio).

a.L, facet for ligament to astragalus; na., navicular;

as., astragalus; «./., navicular facet;

c., cuneiform; o.c., os calcis;

cu., cuboid; p., pisiform;

ec., ectocuneiform; r., radius;

e.f., ectal,facet; . sc., scaphoid;

en., entocuneiform; s.f., sustentacular facet;

/./. , fibular facet ; t.f. , tibial facet ;

i.m,, inner malleolar facet; tr., trapezium;

/., lunar; tz., trapezoid;

m., magnum; «., ulna.

me., mesocuneiform;

FIELD MUSEUM OF NATURAL HISTORY.

PLATE XVI, ZOOLOGY.

4 »./

MANUS AND PES.
Enlarged.

OF THt
HI5WERS1TY OF ILLIHOIS

•

.

c
ft

'

o

tn

&

•4

•

3

S
a

u

bl

w

B

3

>

"

&'

r*

q

CD

c

•

5. EL ri ^
JT fc;

cr, — • CT

-

: premaxillary

c"

0)

• *,

o

eS

J

D
U!
W

*

o

(U
'w

13

, mastoid foramen
, maxillo-turbinal ;

cT

1

V£>
.. $

13 13

% 3
a c

, nasoturbinal;
, olfactory fossa;

% IT

si

13 o3
a a

, pituitary fossa;
, palatine process >

li,
•

a

, pterygoid;
, subsquamosal for

, transverse canal ;

, vomer.

5

H

a

"^ "i

8 S

"§' ^

•^ £

S §

£

^ «*^

vj

H

X

3

1

§

8

<=>'

*

•^ -S,

•»

"*»

UJ

w

c

•^

5

J

CJ

OL

^

.s

^

'-C

o
5

OJ

3

£

^

o

, basioccipital;
, basisphenoid;

, carotid canal;
, cerebellar fossa;

, condyloid foram
, cerebral fossa;

, cribriform plate:
, ethmoturbinal;

, frontal;
, carotid foramen

, floccular forame

, foramen ovale;
, frontal sinus;

, jugular foramen

, maxilla;
, mesethmoid;

S.S £ S

OF THt

nF II I MfllS

' ;

i

.' r •

PLATE XVIII.

DENTITION AND HARD PALATE.

Enlarged.

Fig. i. First upper molar, inner lateral aspect (xio).
Fig. 2. Left lower molar series from above (xio).
Fig. 3. Hard palate (xs).
Fig. 4. Left upper molar series (xio).
Fig. 5. Left lower molar series, outer lateral aspect (xio).

FIELD MUSEUM OF NATURAL HISTORY.

PLATE XVIII, ZOOLOGY.

4

DENTITION AND HARD PALATE.
Enlarged.

OF Trtt

.XIX

•''A lo rtotete}13 LaJnool/L .1 .gi1?
to note:

PLATE XIX.

Fig. I. Mounted Skeleton of Nesogale dobsoni.
Fig. 2. Mounted Skeleton of Ccenolesfes obscurus.

FIELD MUSEUM OF NATURAL HISTORY.

PLATE XIX, ZOOLOGY.

1. MOUNTED SKELETON OF NESOGALE DOBSONI.

2. MOUNTED SKELETON OF CAENOLESTES OBSCURUS.

OF Ut

tf ILLINOIS

.XX

' .'!

;nor

;bi-
; nsmmoil ohsio

;ellix£n:

;bio.

; Ir-
jotjohaq , ••

.

,.\.'

,.t.\<^

x*n.

;l;i

' fftfn.

.•'

•

HJ

' ionoioo

-

,.:HM.\^.O

,.•*
,.n.\.\

j

PLATE XX.

SKULL OF CENOLESTES.

Adapted from Miss Dederer.
Fig. i. Palatal aspect.

Fig. 2.

ang., angular process of mandible
a.o.v., antorbital vacuity;
a.pl.vac., anterior palatine vacuity;

As., alisphenoid;
As.bl., alisphenoid bulla;
Bo., basioccipital;
Bs., basisphenoid ;

c., mandibular condyle;
car. can., carotid canal;
car./., carotid foramen;

c.f., condyloid foramen;
cor.pr., coronoid process;
eust.f., eustachian opening;
Exo., exoccipital;
/./.a., sphenorbital foramen;
f.l.p., foramen lacerum posterum;
J.ov., foramen ovale;

Fr., frontal;

f.v.f., venous foramen in frontal;
L., lachrymal;
l.d., lachrymal duct;
Ma., malar;

Lateral aspect.

w./., mental foramen;
Ms., mastoid;
ms.f., masseteric foramen;
MX., maxilla;
Na., nasal;
Os., orbitosphenoid;
Pa., parietal;
Per., periotic;
P-gl-f-, post glenoid foramen;

PI., palatine;

p.l.f., postero-lateral foramen;
pl.r., palatal ridge;
Pmx., premaxilla;

p. pi. vac., posterior palatine vacuity;
Ps., presphenoid;
Pt., pterygoid;

pt.p., pterygoid process of palatine;
s.sq.f., subsquamosal foramen;
st.m.f., stylomastoid foramen;
tr.can., transverse canal;
Ty., tympanic;
F2., infraorbital foramen.

FIELD MUSEUM OF NATURAL HISTORY.

a.fl.iac

PLATE XX, ZOOLOGY.

Pm*

Ex*

SKULL OF CAENOLESTES.
Adapted from Miss Dederer.

I;,; :::>RARY

OF THt

OF

.dbia Jrlgh srfj moil nund odJ '1.

•
;ra; •' ,.<b

'

PLATE XXI.
DORSAL AND LATERAL VIEWS OF BRAIN.

Five times natural size.

Fig. i. View of the dorsal surface of the brain.
b.ol., bulbus olfactorius;

cb., body of the cerebellum;
floe., flocculus;
f.rh., fissura rhinalis;
s.orb., sulcus orbitalis.

Fig. 2. View of the brain from the right side.
b.ol., bulbus olfactorius;

cb., body of the cerebellum;
f.erh., fissura endorhinalis;
floe., flocculus;
f.rh., fissura rhinalis;
l.pir., lobus piriformis;
s.orb., sulcus orbitalis;

t.o., tuberculum olfactorium;
tr.ol.lat., tractus olfactorius lateralis.

FIELD MUSEUM OF NATURAL HISTORY.

PLATE XXI, ZOOLOGY.

floe.

DORSAL AND LATERAL SURFACES OF BRAIN.
Five times natural size.

OF THt
HMIVF.RSJTY OF ILLINOIS

PLATE XXII.
VENTRAL SURFACE OF BRAIN.

Five times natural size.
b.ol., bulbus olfactorius;
f.erh., fissura endorhinalis;
floe., flocculus;

hy., hypophysis;
// to XII., cranial nerves;
l.pir., lobus piriformis;
pyr., pyramid;
t.o., tuberculum olfactorium;
tr.ol.lat., tractus olfactorius lateralis;
trap., corpus trapezoideum.

FIELD MUSEUM OF NATURAL HISTORY. PLATE XXII, ZOOLOGY.

VENTRAL VIEW OF BRAIN.
Five time> natural size.

OF THc
OF