Memoirs of the
Queensland Museum.

VOL. VIII, PART I.

ISSUED JANUARY 30th, 1924.

EDITED BY THE DIRECTOR, HEBER A. LONGMAN, F.L.S.

Published by the Authority of the Chief Secretary for Queensland, the Hon. E. G. Theodore.

MEMOIRS OF THE QUEENSLAND MUSEUM, Von. VIII., Part I.

THE ZOOGEOGRAPHY OF MARSUPIALS,

WITH NOTES ON THE ORIGIN OF THE AUSTRALIAN FAUNA.

By Heber A. Longman, F.L.S., Director, Queensland Museum.

Notwithstanding the wealth of literature on the subject, the problem of the
origin of our Australian marsupial fauna is still a fascinating field for research.
It is almost with trepidation that the writer enters this arena of controversy,
wherein so many able authorities have expressed diverse views. None the less,
it seems desirable to try and focus the results of some of our more recently
acquired knowledge, together with a purview of the work of the past, so far as
space and opportunity permit.

For the sake of consecutiveness it is necessary to set out, by means of
extracts, some of the views of authorities. In 1893 H. 0. Forbes published a
formidable list of vertebrates (the majority being birds), invertebrates, and
plants chiefly confined to the southern hemisphere suggesting affinities between
South America, Australia, various Pacific Islands, and South Africa, from which
he deduced the existence of a former circumpolar continent. When analysed
to-day, however, many of the examples quoted by Forbes, with reference to
Australia and South America, have lost their significance, and any affinities that
remain do not afford evidence of lineal relationships requiring southern Tertiary
land bridges.

In 1892 Baldwin Spencer accepted the north-western origin in pre-
Cretaceous time of the monotremes and the primitive polyprotodont fauna
(p. 116), spreading south and then east. Later, in his valuable summary of the
results of the Horn Expedition (1896), Spencer considered that the northern
origin of marsupials was practically negatived by the feeble development of the
more primitive polyprotodonts in north-eastern Australia (p. 185), although he
believed that the characteristic Diprotodonts had mainly developed there. He
considered that Papua was the last and not the first land of the Australian
region to be reached by the marsupial fauna.

Haswell (1914), in his interesting review of the Australian fauna, considers
(p. 217) the connection between Australia and South America “at some time
antecedent to the Pliocene” to be clearly established.

In several papers (1893, 1895, 1899, and 1912) Hedley has argued for the
Antarctic origin of our most characteristic fauna, and his advocacy carries weight.

Hedley (1899) wrote of “ a rich fauna of Antarctic origin, which, entering
by Tasmania, overran the whole continent, crossed Torres Strait into New

Note — The substance of this paper was given before the Zoology Section of the
Pan-Pacific Conference in Sydney, August, 1923.

2

MEMOIRS OF THE QUEENSLAND MUSEUM .

Guinea, and reached its utmost eastern limit in the Solomons. Characteristic
members of it are the marsupials, monotremes, cystignathous frogs, venomous
snakes, and snails of the Order Macroogna.” Using Tate’s term, he calls this
the Euronotian element, and claims it as the most characteristic fauna. In his
address to the Linnean Society of London in 1912, Hedley states : “In Australia
marsupials and monotremes are least developed in the North ; proceeding
southward more groups successively appear till ultimately Tasmania has, as
Professor Spencer expressed it, e a condensation of most, that is noteworthy in
the Australian Region.’ ” This aspect of the Antarctic theory appears to. the
writer to be open to adverse criticism. It is advisable to analyse first living
species of Polyprotodontia.

Dasyuridae — The supposed paucity of polyprotodont forms in Papua and
north-eastern Queensland must be examined in the light of recent work, which
has greatly increased our knowledge of the range of present-day and fossil
marsupials. In this “ Papuan ” (Hedley) or " Torresian ” (Spencer) sub-region,
we find Dasyurus maculatus, geojfroyi, gracilis, hallucatus , albopunctatus , and
dcemonellus (Thomas, 1904), the last two species being Papuan, and there is a
doubtful record of viverrinus for Queensland in our State Museum. The known
species are found to be very strongly represented in north-eastern Australia.
In the genus Phascogale we find that the following species (subspecies omitted),
have been described from the Torresian region during the last twenty years or
so : — lorentzii, noukuyou, naso (three species from Dutch Papua, described by Jentink
in 1911), murex 0. Thos., 1913 (Papua), melanura O. Thos., 1912 (Papua) ; bella O. Thos.,
1904, mimulus 0. Thos., 1906, and ingrami, 0. Thos., 1906, are Northern Territory
species ; godmani O. Thos., 1923, comes from North Queensland. When these
records are added to previous ones, the great majority of species of Phascogale
•are seen to be in the northern region. It is somewhat surprising to find that
the known distribution of the allied genus Sminihopsis has been much more
restricted, although the majority of the species now described are found in the
northern region, and S. rufigenis was recorded for the Aru Islands in 1922 by
Oldfield Thomas. When genera and species are considered, taking fossil as well as
present-day forms, it is found that, instead of the polyprotodont marsupials being
feebly developed in the northern region, it is there that they are most strongly
represented.

In the opinion of Bensley (1903) and of Gregory (1920, p. 159), Myrmecd>ius
is a specialised dasyurid. Wood Jones (1923, p. 126) has remarked that it is greatly
handicapped by its habits : “It does not excavate deep burrows, it does not climb,
it is not fleet of foot.” The gradual restriction of its range in recent years suggests
a far wider distribution in the past.

Peramelidse. — During recent years we have added greatly to our knowledge
of the Peramelidse. In 1909 Allen and Barbour ( loc . cit.) described the new
genus Suillomeles for a species ( hispida ) from Dutch Papua. In 1920 the genus
Rhynchomeles was established by Oldfield Thomas for a species ( prattorum ) found

THE ZOOGEOGRAPHY OF MARSUPIALS.--LONGMAN.

3

in Ceram. The genera Echymipera and Peroryctes, now separated from Perameles
and Isoodon, are distinctively Papuan. The writer has no knowledge of the
characters of Anuromeles, Heller (1897), from Papua.

In 1903 Bensley pointed out that “ Of the Peramelidae all the species
representing prototypal forms are Papuan ” (p. 204). In 1923, with the addition
of new genera and species the range and diversity of the Peramelidse are far
more significant. It now appears to be definitely established that the radiation
of the Australian forms emanated from Papua, Thylacomys and Chceropus being
specialised offshoots.

The discovery of new northern genera of peramelids is also of special
interest in view of Osgood’s opinion that this family exhibits characters sugges-
tive of the incipient stages leading from polyprotodont to diprotodont groups.
This is strengthened by their common syndactylism.

Osgood also states (1922, p. 135) that “the conclusion is unavoidable
that among living forms those most suggestive of what the ancestor of
Ccenolestes and other diprotodonts was like, are the Peramelidse.”

T. Thomson Flynn’s recent work on the placenta in Perameles (1923), in
which the genus is stated to be “ a most primitive form ” in this respect
(p. 173), must be noted here.

Notoryctidae.— The description of Notoryctes caurinus, by Oldfield Thomas in
1920, from Wollal, North-west Australia, makes an interesting northern extension
of this curious marsupial, which is evidently a “specialised dasyurid.”

Diprotodontia. — It is of interest to note a few of the more recent records
for existing diprotodont marsupials. Taking the Phalangeridce first, we find that
no less than twelve out of twenty species of Pseudochirus occur in Papua, whilst
the remaining Australian species are so richly represented in the north that
the dominance of the genus there is undoubted. Dactylopsila (with eight
species), Dactylonax and Distcechurus are distinctly Papuan genera, whilst
Acrobates, Petaurus, Eudromicia (Mjoberg, 1915), and Dromieia are represented.
Phalanger, including Ceonix and Wyulda (Alexander, 1918) has a remarkable
distribution, ranging from Celebes to San Christo val in the Solomon Islands,
from Talaut through the Moluccas, Timor, Papua, and New Britain, the Arus
to North Queensland and North-west Australia, with twelve species. This
wide-spread distribution affords striking evidence of northern marsupial extension,
especially in view of the fact, mentioned by Guppy (1887), that cuscus are
preyed upon by dogs which attack them when they descend at nightfall. The
genera of the Phalangeridse are thus seen to be overwhelmingly represented in
northern regions.

Of the macropods, Dendrolagus, Dorcopsis, and Dorcopsulus are obviously
Papuan. Hypsiprymnodon, an annectant type, is confined to North Queensland.
The genera Petrogale, Lagorchestes , JEpyprymnus, and Onychogale are represented
in the northern State. Macropus has a still wider range to Papua, New Britain,

4

MEMOIRS OF THE QUEENSLAND MUSEUM.

and the Aru Islands, but it seems clear that the evolution of the characteristic
saltatorial forms, represented by this genus, has taken place mainly in the
plains and forests of the Australian mainland. But, as in other families, the
majority of the macropodian genera are to be found in the Torresian region.

When fossil representatives are considered, species of Phascolomys are
more numerously represented in northern than in southern Australia.

Tasmanian Fauna — With the possible exception of Wynyardia bassiana
Spencer, dealt with elsewhere, the marsupial fauna of Tasmania presents no
species, neither living nor fossil, that cannot be readily conceived as derivable
from mainland forms. With the palaeontological evidence for Sarcopkilus and
Thylacinus in Queensland, this is surely clear. In the list published by Clive E.
Lord in 1918, twenty species are mentioned. Of these, one, Petaurus breviceps,
is an artificial introduction from the mainland ; there are three varieties of
mainland forms ( Trichosurus vulpecula var. fuliginosus, Macropus ruficollis var.
bennetti and M. giganteus var. fuliginosus ) ; no less than ten are found on the mainland
(Sminthopsis leucopus, Phascogale swainsoni , Dasyurus viverrinus, T). maculatus,
Perameles gunni , P. obesula, Petaurus breviceps , Dromicia nana, Potorous
tridactylus, and Macropus billardieri) ; the remaining seven are confined to
Tasmania {Phascogale minima, Sarcopkilus ursinus, Thylacinus cynocephalus ,
Phascolomys ursinus, Pseudochirus “ cooki,” Dromicia lepida, and Bettongia
cuniculus ) . The present evidence of the marsupial fauna of Tasmania is decidedly
in favour of the northern and not the southern origin of the group.

Syndactylism.— In 1923 Wood Jones analysed the relative age values of
dental and pedal characters in classifying marsupials. After pointing out that
the primitive marsupials were undoubtedly polyprotodont and didactylous, he
shows that syndactylism must be a more primitive feature than diprotodontism
because it is common to both groups ; from a polyprotodont stock a syndactylous
section arose which split up into the diprotodont group of herbivorous
marsupials, on the one hand, and into the polyprotodont peramelicls, on the other,
the still more primitive didactylous dasyurids remaining unchanged. As it
seems most improbable that this distinctive syndactylism could have been independ-
ently acquired, it is almost certain that this pedal character is one of greater
antiquity than the specialised development of two front lower incisors, so
characteristic of our dominant group. Syndactylism is thus a distinctive

character of Australian marsupials, and it has apparently been developed here.
It is absent in Cmnolestes and in all other American marsupials, although in
Marmosa pusilla, figured by Bensley (Plate VII, fig. 7), there is e£ an indication
of the syndactylous condition of the Phalangeridse” (p. 191). It is very
doubtful, however, whether this approximation of the second and third digits
in some species of Marmosa can be compared with the well-marked and
dominant syndactylism of Australian groups. It is remarkable that in the
illustration of Marmosa pusilla given by Oldfield Thomas (PI. XXVIII, fig. 1)
there is no trace of syndactylism. In specimens of Marmosa mitis (0. Bangs,
1898) in the Queensland Museum, examined by the writer, there is no evidence

THE ZOOGEOGRAPHY OF MARSUPIALS. —LON OMAN. 5

of syndactylism, which does not appear to be even a specific character in the
genus. The if. pusilla of the British Museum Catalogue is apparently referable
to the nearly- allied if. marica 0. Thomas, 1898.

Fossil Marsupials. — In tabulating evidence as to the origin of our
marsupials, it is obvious that the index afforded by fossil genera and species
is of even greater value than that of existing representatives. It appears to be
an outstanding fact that our fossil marsupial fauna is much more distinctively
Australian than that of to-day. We have ample evidence of gigantic specialised
marsupials, such as Diprotodon, Nototherium , Euryzygoma, Phascolonus, Sthenurus,
Palorchestes , Procoptodon , and Thylacoleo. There is no evidence whatever for the
derivation of these specialised genera from either living or fossil American
forms.

It is also very obvious that the genera existing to-day were much more
richly represented in the past. The numerous species of Macropus ( sensu lato)
included very large forms, such as M. anak, w T hose trenchant premolar relatively
dwarfs that of the Papuan Dorcopsis. The wombats of to-day are but a few
survivors compared with their Pleistocene predecessors, which ranged from
Clermont (Longman, 1917) to Western Australia. The carnivorous extinct
Thylacoleo carnifex had for associates in Queensland species of Sarcophilus and
Thylacinus, which were larger than Tasmanian species of to-day. A mandibular
fragment of Sarcophilus was recorded from Rockhampton by the writer in 1921,
this being an addition to the many records of the Queensland “devil 5 ’ on the
Darling Downs. The same species has been noted for Western Australia by
L. Glauert, who has also recorded the following genera from his field : — -
Diprotodon, Nototherium , Phascolonus, Phascolomys, Macropus, Sthenurus,
Bettongia , Perameles, Thalacomys, Dasyurus, Thylacinus, and Thylacoleo. Scott
and Lord have published records in Tasmania of Diprotodon, Nototherium,

Palorchestes, Sthenurus, and Macropus .

This wealth of widely spread fauna surely demonstrates that our

marsupials have been long in the land. Bensley (1903, pp. 206-207) expressed
the view that the lowest estimate of the duration of the evolution of the

Australian marsupials is still much too large, and he thought it unlikely that

“ the marsupial radiation could have begun until well on into the middle of the
Tertiary period.” But Bensley 5 s elaborate review dealt mainly with existing
species, and he does not appear to have fully appreciated the wealth and
divergencies of our fossil forms. It seems far more probable that Osgood is
correct when he says (1921, p. 131) “ the view that all living families of
marsupials were well differentiated early in the Tertiary seems to be well
founded.” In the clash of conflicting views as to the age of marsupials in
Australia, it is difficult to obtain a proper perspective. No one, however, can
carefully examine a large and comprehensive collection of our fossils without
recognising that they represent an assemblage resulting from a lengthy process
of evolution, which, in the great majority of genera, cannot be associated with
any forms yet found outside the Australian regions.

The many changes that have taken place in land masses of the Australian

6

MEMOIRS OF THE QUEENSLAND MUSEUM.

region since Cretaceous times, as outlined by Jensen and others, cannot be dealt
with here, but it is evident that a variety of environments was available for the
evolution of our marsupials from remote ancestors which were northern in origin.

Myrmecoboides. — Special interest has been attached to the presence of
Myrmecoboides, with supposed peramelid affinities in the Paleocene of Fort
Union, North America. Gidley’s notable work on this fragment (an incomplete
ramus of a lower jaw) perforcedly deals mainly with dental characters. The
resemblances to Myrmecobius are very suggestive, but are not conclusive, and
may be the result of convergence. Gidley (1915) notes that the name
Myrmecoboides was given on account of its likeness to Myrmecobius rather than
as a positive assumption of real relationship. He does not accept Osborn’s or
Bensley’s view concerning the phylogeny of modern families of marsupials, and
assumes “ a vastly more ancient origin ” than has hitherto been held for them.
He states : “In fact, it seems reasonable to assume that at present nothing is
definitely known regarding the origin of several of the living families of
marsupials, including probably all the diprotodonts, because they are not
represented in our collections from beds older than the Pleistocene, and that in
the lower fossils we have only an incomplete and indefinite history of the
origin and development of a part only of the polyprotodonts ” (p. 401).

Osgood (1921, p. 140) states : “ The hypothesis can scarcely be avoided
that, provided Myrmecoboides is a marsupial, it may well be ancestral to the
csenolestids.”

South American Fossils. — The Sparassodontia of Ameghino from the
Miocene of Patagonia are placed by W. T. Sinclair in the Family Thylacinkke.
But it is significant that in Borhycena the canine is actually replaced and in
Prothylacinus the canine and two premolars are replaced. This marked
tendency to a diphyodont dentition is unparalleled in Australian forms. Palatal
vacuities and epipubic bones are lacking, and It. Lydekker, in 1907, in analysing
these and other features, expressed the opinion that the Patagonian fossils were
Creodonts and not marsupials. W. K; Gregory (1910, p. 207) states that these
fossils show “ a remarkable series of resemblances in general proportions and

dental characters to some of the Creodonts among Placentals The

family Thylacinidse is thus seen to be a specialized offshoot either of very
primitive Dasyuridae or more directly from the Didelphiidae.”

Gidley (1915, p. 401) points out that in the Miocene the Thylacinidae had
reached almost their present state of specialization. They certainly do not
represent a primitive type suggesting an ancestral source for Australian forms.
Tate Regan (1914) adversely criticises the suggested relationship between the
two groups.

It is by no means improbable that the resemblances between the Patagonian
fossils and Thylacinus are due to convergence. But even if the lineal relation-
ship were definitely proved, which Sinclair does not suggest, the South American
origin of our Diprotodontia would still be without a satisfactory basis.

The position of the remarkable American marsupial Ccenolestes has been

THE ZOOGEOGRAPHY OF MARSUPIALS.— LONGMAN.

discussed in great detail by several authors, and it has been placed in no less
than three sub-orders. Whilst many of its characters suggest that it may

represent a prototypical generalized form, from the ancestors of which dipro-

todonts may have been derived, it seems to the writer that far weightier
evidence is required before we associate the varied and manifold Diprotodontia
of Australia with this marsupial from the Andes.

Among the many interesting points raised by Osgood in his elaborate
memoir on Ccenolestes is the fossil evidence that the diprotodont dentition of the
csenolestid group goes back far in geological time, “ perhaps in the Mesozoic,”
as the didelphids and csenolestids were already well distinguished in the
Patagonian Miocene (Osgood, 1921, p. 129). Among his conclusions he states :
“ The ancestor of the csenolestids was probably a northern form which had
already separated from the generalized polyprotodont stock. . . . The

North American ancestor of the csenolestids possibly extended throughout
Holarctica and therefore may also have given rise to the Australian dipro-

todonts. This assumes that the main lines of divergence between Diprotodonts
and Polyprotodonts were established in the north prior to the Australian and
South American radiations. Hence an Antarctic land connection is not necessary
to explain the resemblances between American and Australian marsupials ”
(p. 151).

W. K. Gregory (1910) points out that the pes is “ entirely eleuthero-
dactylous,” and he is inclined to regard Ccenolestes and its allies as an
independent sub-order, an offshoot of primitive Polyprotodonts, which has
paralleled the Diprotodonts in certain characters of the dentition” (p. 211).

Gidley (1915, p. 401) states that the Csenolestidse “ more probably belongs
wdth the Polyprotodonts.”

In his analyses of the affinities of Ccenolestes in 1911, R. Broom considers
that it “ was much more nearly allied to the American Polyprotodonts than to
the Australian Diprotodonts ” (p. 315).

Assuming that Wood Jones’s argument for the special primitiveness of
syndactylism is correct, the fact that modern csenolestids are diadactyl is an
important point. It is also significant that a diprotodont dentition has
independently appeared in several orders. It seems most probable that the
resemblances between csenolestids and Australian diprotodonts are mainly due
to convergence, but partly patterned by remote ancestral inheritance. There is
now manifold evidence for the phenomena of convergence in obviously unrelated
orders ; if may be urged, a fortiori, that convergence is still more likely to
appear in groups distantly related owing to the working out of similar ancestral
potentialities.

Gregory (1910, pp. 212-213) writes : “ Marsupials in general, and especially
South American families of many orders, seem to abound in examples of
homoplastic and convergent resemblances to animals of other orders.”

Ameghino, who has described so many of the South American fossils,
necessarily relies mainly on dental characters. In his phylogenetic diagram of

s

MEMOIRS OF THE QUEENSLAND MUSEUM.

diprotodonts (1903, p. 188) lie derives his Hypsiprymnoidea (“ diprotodontes

australianos ”) from the ancestral Garzonidse through the primitive Caenolestidse.

After referring to the affinities of northern forms of Multituberculates with

Patagonian fossils, he says : “ Ademas, la transicion entre los paucituber-

culados de la Argentina (Epanorthidae, Caenolestidae), y los diprotodontes actuales
de Australia, es igualmente tan completa, que ya no es discutible la conclusion
de que todos en conjuntc constituyen un solo gran grupo.” But the unitary
grouping of the Caenolestidae and Diprotodontia is not accepted by many
authorities, and it seems a colossal assumption to derive the varied families and
genera represented in Australia from the American species.

Wynyardia. — Baldwin Spencer’s description of Wynyardia bassiana (1900)
forms a notable contribution to the records of Australian vertebrate fossils.
These remains, which have been attributed to the Eocene, were unaccompanied
by teeth, but the describer instituted a large number of comparisons, based on
available characters, and came to the conclusion that Wynyardia represented a
form distinct from present marsupials, but allied to both existing Polyproto-
dontia and Diprotodontia. Osgood (1921) has criticised the diagnostic value of
many of the characters set out by Spencer in his comprehensive analysis, but,
whilst agreeing with the view that Wynyardia combines many features of the
two groups, classifies it with the diprotodonts and makes it the type of a new
family — Wynyardiidce. He then compares this fossil with Ccenolestes, as suggested
by Gregory in 1910.

It appears to the writer that Wynyardia has no real relationship with
Ccenolestes, and that the differences are far more marked than the eight
resemblances noted by Osgood (p. 139). The proportions of the nasals and the
slope of the sutures between premaxillae and maxillae, mentioned by him, are
most variable in modern marsupials of the same genus, and are of no real
diagnostic value.

Among the many distinctions between Wynyardia and Ccenolestes which
can be noted by comparing the two forms, I have drawn up the following,
which seem to be significant : —

The squamosal is extremely large (“a very remarkable feature” — Spencer)
in the Tasmanian fossil, and excludes the alisphenoid from union with the
parietal, whilst in Ccenolestes the alisphenoid has a wide union with the upper
element.

The zygomata are massive in Wynyardia and set relatively high ; in Ccenolestes
fhey are slender and are set lowx

The lachrymal in Wynyardia is so w r ell developed that it actually prevents
contact between the maxillary and nasal bones, being in striking contrast to the
American form.

The occipital regions in the two fossils are quite unlike.

So far as can be judged from the mandibular fragment of Wynyardia, it
is utterly dissimilar from the lower jaw of Ccenolestes.

THE ZOOGEOGRAPHY OF MARSUPIALS,— LONGMAN.

9

The tibia of Wynyardia is of a very different type, and the fibula “is a
-complete separate bone.”

It should be stated that Osgood recognised that his comparison between
the two forms was “rather unsatisfactory as a basis for positive conclusions.”

Osgood also compares Wynyardia with extinct American forms, Palceothentes
and Abderites, allied to Ccenolestes. One of the most remarkable features of the
Tasmanian fossil is the presence in the incomplete ramus of the lower jaw of
“ a deep depression across the upper surface, which must apparently have been
situated in the area occupied by the molar series. It extends in a slanting
direction from side to side across the jaw, and its smooth, rounded floor-surface
and margins . show clearly that it is not a mere break in the jaw. With this is
associated a swelling on the internal face. . . . ” (Spencer, p. 784). The

describer suggests that it may have been associated with the presence of a large
sectorial tooth in the molar or premolar series such as is developed in Abderites .
Gregory (1910, p. 215) instances an alveolar groove in Ameghino’s Garzonia
minima, and Osgood, in 1921, writes that the structure of the lower sectorial
tooth in Abderites “ seems closely comparable to the condition in Wynyardia
When Ameghino’s figures of Abderites meridionalis and of Garzonia minima are
consulted, it seems to the writer that there is no evidence whatever of relation-
ship. In these American fossils we see no sign of the remarkable contours
figured for Wynyardia. Assuming that the depression represents diastemata
between molar series and incisors, then there is no similarity between the two
groups of fossils. Should this smooth transverse furrow be interpreted as
representing the site of a large sectorial tooth that had been accidentally lost
long before the death of the animal, then comparisons for this special feature
might quite as well be instituted between Hypsiprymnodon or even Phalanger,
present Australian forms with prominent oblique premolars. To suggest a theory
of relationship on the possible shape of a tooth that is not even represented by
an alveolus is, however, far beyond the scope of this paper.

In the light of our present knowledge, there appears to be no valid
reason for suggesting that Wynyardia bassiana is related to the csenolestids of
South America.

Two fossils from eastern Australia are significant because they demonstrate
combinations of characters. Burramys parvus from Taralga, N.S.W., described
by R. Broom in 1895, has a large grooved premolar, and is probably an
annectant form between the Macropodidse and Phalangeridse. Triclis oscillans,
from King’s Creek, Darling Downs, Queensland, has somewhat similar characters,
and was compared by De Vis (1888) to the existing Hypsiprymnodon.

In 1916 W. D. Matthew described, under the name of Eodelphis broumi ,
a primitive opossum from the upper Cretaceous of Alberta. Referring to this
form, Gregory (1920, p. 156), in reviewing the early polyprotodont marsupials,
isays : “ The more primitive members {Eodelphis) seem to be directly ancestral to

10

MEMOIRS OF THE QUEENSLAND MUSEUM.

the modern Didelphiidse of North and South America, and were also probably
related to the Upper Cretaceous European and Asiatic polyprotodonts that were
ancestral to the later Australian members of this group.”

BRIEF NOTES ON THE NON-MARSUPIAL FAUNA.

Monotremes. — The absence of monotremes from South America and their
special development in Australia is evidence of their northern origin, probably in
pre-Tertiary times. The development of Zaglossus and Prozaglossus (Herbert,
1913) in Papua is significant. Pliocene monotremes have been recorded by
W. S. Dun (1895) from New South Wales. Species of Zaglossus have also been
recorded by L. Glauert from West Australian deposits, and by Scott and Lord
from Tasmania. Two genera are represented in Darling Downs deposits.

Rodentia. — The evidence of other mammals is distinctly favourable to the
hypothesis of lengthy and extensive northern land connections. In 1916 the
writer remarked on the significance of about fifty autochthonous species of
rodents in Australia and Tasmania. Since then the list has considerably increased.
So far as I am aware, no alternative to a northern origin of the ancestors of
these rodents has ever been suggested.

Reptilia. — -With regard to the reptilian fauna, Thomas Barbour, in his
valuable study of the Zoogeography of the East Indian Islands (1912) states :
“The remarkable development of Ophidia proteroglypha in Australia can
hardly be explained by assuming that they came from South America.” He
holds, on the contrary, that they originated in Asia and spread through Australia
to South America (p. 155). He refers elsewhere to the strong Australian and
perhaps a Papuasian tinge in the fauna of Timor, instancing Chdodina and
Liasis.

If we eliminate such genera as Typhlops and Natrix, which are cosmo-
politan and presumably palaeogenic forms, we find that the affinities of our
ophidian fauna are certainly not with America. There are no true pythons in
America, and our genera of constricting snakes are evidently northern in origin.
Of the Acrochordinae, Acrochordus and Chersydrus are northern, and this also
applies to the genera of Homalopsinae and to Dendrophis. Compared with the
rich development of Elapine snakes in Asia, Africa, and Australia, this group is
sparsely represented in America. The Hydrophiinae have special means of
distribution, but if they are entitled to consideration they tell the same story.
The paramount affinities of the Australian snakes as a whole are undeniably
northern.

Many species of Ophidia, Lacertilia, and Chelonia are common to North
Queensland and Papua.

Of the Lacertilia, the Gecko nidae and Scincidse are so cosmopolitan that no
special significance can be attached to their distribution, although Australia has
such distinctive genera as Nephrurus and Trachysaurus , to mention but two.
The Pygopodidse are confined to Australia and Papua. The Agamidae and

THE ZOOGEOGRAPHY OF MARSUPIALS.— LONGMAN. n

Varan ida? are African, Asiatic, and Australian in range. There are no Iguanidae
in our region. The lacertilian fauna affords no evidence whatever for an
American invasion, and its affinities are decidedly northern.

Amphibia. — Much has been made of the presence of Cystignathidse in
America and in Australia, but it seems exceedingly doubtful whether the many
genera of frogs placed in this family in these regions constitute a natural
assemblage.' Gadow writes : “ The numerous genera not only run into each
other, but also get entangled with those of other families. In fact the whole
family is ill defined.” (1909, p. 209). Amongst the amphibians the phenomena
of convergence are very well known.

Van Kampen has pointed out that many Indian forms of Amphibia must
have reached Papua by ancient land connections (1909).

Fishes. — The evidence of certain fresh-water fishes (Galaxiidae and Aplo-
ehitonidae) found in Australia, New Zealand, and South America has been
repeatedly stressed by the exponents of the Antarctic connection. Tate Regan,
however, insists that these are of marine origin (1914) and A. R. McCulloch
1915) actually discovered young specimens of Galaxias attenuatus making their
way from the coast into fresh water. Regan ( loc . tit.) points out that the only
true fresh- water fishes common to these regions are the Osteoglossidse, and that
our well-known representative of this ancient group (the Barramundi, Scleropages
leichJiardti ) has greater affinities with Asiatic forms. He concludes that “ neither
marine nor fresh-water fishes support the theory that the Antarctic continent
connected America with Australia during the Tertiary Period.” (p. 41).

Max Weber and de Beaufort have found an Australian element in the
fresh- water fishes of the Aru Islands and Waigen, and an interesting summary
of their views appears in “ The Zoology of the Indo- Australian Archipelago ”
(1923.)

Flora. — The present endemic nature of the great bulk of the Australian
fauna is borne out by the testimony of the flora. The “ original Australian
flora ” (to use Maiden’s term) is the dominant group on the western side, where
it apparently developed. On the east, however, the Austro-Malayan element is
the more prominent, and one may reasonably suggest that only an unfavourable
environment prevented its extension into temperate regions towards the south
and eremian regions to the west. There appears to be no absolute need to
postulate a land bridge, suitable for emigrating mammals, with South America
to account for the third, or so-called Antarctic element.

The writer is in no way competent to comment on the geological evidence,
but, judging from R. N. R. Brown’s review of the problems of Antarctica
(1923), further exploration is needed before the relationships of the eastern and
western portions can be made clear. W. N. Benson, in his able study of Palaeozoic
and Mesozoic seas in Australasia, gives strong evidence of the land connection
of Australasia with Antarctica and South America at the close of Mesozoic

12

MEMOIRS OF THE QUEENSLAND MUSEUM.

times, but it must be noted that, at the same time, there were land connections
to the north and north-east linking Australia with the remains of the ancient
continent called iEquinoctia, by E. C. Abendanon (1919, p. 577). Similar
northern extensions are a salient feature of Hedley’s map, published in 1909
(p. 333). The probable extent of this northern continental mass, as it may
have existed in early Mesozoic times, is shown by A. B. Walkom (1918, text-
figure 2, p. 105). The shallow submarine plateau connecting Australia with
Papua, the Aru, and some other islands is called “the Sahul flat” by
Molengraaf.

There is some diversity of opinion as to the age of the connections
denoted by the great shelves in what Molengraaf calls “ the Australasian
Mediterranean ” (1921), and it is not improbable on zoological evidence that the
first connections may have been long anterior to the relatively recent variations,
which may have been indirectly effected by glacial and interglacial periods.

Valuable references to the views of authorities on the relationships of the
Asiatic and Australian faunas are given in “ The Zoology of the Indo- Australian
Archipelago ” (1923) ; whilst Chilton (1909) gives an interesting summary of the
relationships of circum-austral forms from the New Zealand standpoint.

The assumption of a continent in the Antarctic in early Tertiary times,

in order to account for related elements in the faunas and floras of South

America, Australia, New Zealand, various Pacific Islands, Madagascar, and
South Africa, has proved a veritable Pandora’s box to biologists. In order to
explain many of the difficulties raised, it is suggested that this circumpolar area
had radiating land bridges, inconstant in direction and dimension, apparently
reaching out on the one hand to deposit certain types in Pacific archipelagoes ;
elsewhere stretching to Madagascar to transfer iguanas ; extending to New
Zealand to allow other forms to gain new land ; and independently reaching
Tasmania to enable the ancestors of our marsupials to complete a trans-polar
journey and find a congenial refuge.

The presence of more extensive land areas in the Antarctic in Mesozoic

times is apparently demonstrated, but it seems improbable that these areas were

faunal land bridges for vertebrates. On the other hand, the presence of
“ ancient continental ” islands to the north-west, as shown by Russel Wallace
in “ Island Life” and the “ Geographical Distribution of Animals,” is undoubted.
That the evidence of these “ ancient continental ” islands is incomplete is
doubtless due to partial or entire submergence. In this connection it is

interesting to note that Weber and Beaufort (1912) in their analysis of the
fresh- water fishes of Timor found that this fish fauna was “ almost totally
composed of immigrants from the sea,” the inference being that the island
had been largely submerged in Tertiary times.

Possibly fossils from the Austro- Malayan- Asiatic regions will one day be
forthcoming to strengthen the view that our marsupials had a northern origin
probably before “ Wallace’s Line” had significance as a partial barrier. But
even on existing evidence, it seems to the "writer that the alternative route of
Antarctic land bridges involves us in far greater difficulties and anomalies.

THE ZOOGEOGRAPHY OF MARSUPIALS.— LONGMAN.

13

CONCLUSIONS.

A fundamental consideration regarding the origin of Australian marsupials
is the undoubted fact that the group was once widely distributed in Holarctic
or northern regions.

A second consideration is the present-day knowledge that North Australia
and Papua maintain a wealth of diverse marsupial genera, the significance of
which could not be realised in earlier discussions.

A third consideration is the demonstration that our extinct marsupial fauna
was even more distinctly Australian than that of to-day ; the evolution of the
group has largely taken place here.

A fourth consideration is that the differences between Wynyardia bassiana,
our oldest marsupial fossil, and the South American Csenolestids far outweigh
the resemblances.

A fifth consideration is that the most characteristic group of Australian
marsupials, the Diprotodontia, has little or nothing in common with South
American forms ; the slight resemblance between Ccenolestes and our
diprotodonts is probably due to convergence.

A sixth consideration is the demonstration by Osgood that “ didelphids
and caenolcstids were already well distinguished in the Patagonian Miocene,” and
that c< the view that all the living families of marsupials were well differentiated
early in the Tertiary seems to be well founded ” (p. 131).

A seventh consideration is the rich development of the primitive family
Peramelidse in the Austro-Malayan region.

* An eighth consideration is that distinctive syndactylism is a special
Australian feature ; Wood Jones has analysed the alternative values of dental
and pedal characters, demonstrating that syndactylism is the more primitive
feature.

These and other considerations are supported by evidence, briefly outlined
in this paper, dealing with the probable relationships of several non- marsupial
groups (monotremes, rodents, etc.) of the Australian fauna, and the work o i
zoogeographers on the Indo- Australian Archipelago is considered of specia J
importance.

14

MEMOIRS OF THE QUEENSLAND MUSEUM .

Abendanon, E. C.

LITERATURE CITED.

. . 1919, Jonm. Geol., 27, p. 577.

Alexander, W. B.

. . 1918, Journ. Roy. Soc. West. Aust., IV., pp. 31-36.

At. den and Barbour . .

1909, Proc. New Eng. Tool. Club, 4 ( Suillomeles ).

Ameghino, F.

. . 1903, Los Diprotodontes, An. Mus. Nat., Buenos Aires, ser. Ill,

tom. 11, pp. 81-192.

Bangs, 0.

.. 1898, Proc. Biol. Soc. Wash., XII., pp. 161-165.

Barbour, T.

1912, Mem. Mus. Comp. Zool. Harv., XLIV., No. 1.

Bensley, B. A. . .

1903, Trans. Lin . Soc. Lond (2), IX., pp. 83-217.

Benson, W. N.

1923, Trans. N. Zeal. Inst., vol. 54, pp. 1-62.

Broom, R.

1895, Proc. Linn. Soc. N.S.W. (2), X., pp. 563-567 ( Burramys )

Ibid.

. . 1911, Proc. Linn. Soc. N.S.W., XXXVI., pp. 315-320.

Brown, R. N. B.

1923, Scottish Geographical Mag., XXXIX., pp. 172-181.

Chilton, C.

. . 1909, Subantarctic Islands of New Zealand, II., Article XXXV,

De Vis, C. W.

1888, Proc. Linn. Soc., N.S.W. (2), III., pp. 5-8 ( Triclis ).

Dun, W. S.

. . 1895, Rec. Geol. Surv., N.S.W., IV.

Flyn p, T. Thomson

.. 1923, Q.J.M.S., vol. 67, pt. 1, April.

Forbes, H. 0.

1893, Natural Science, III., pp. 54-57.

Gadow, H.

1909, Amphibia and Reptiles, Camb. Nat. Hist.

Gidley, J. W.

.. 1915, Proc. U.S. Nat. Mus., vol. 48, pp. 395-402.

Glauert, L.

1910-1914, Records West Aust. Mus., I., pts. 1, 2, and 3.

Ibid

1920, Jr. & Proc. Roy. Soc., West. Aust., VII.

Gregory, W. K.

. . 1910, The Orders of Mammals, Bull. Amer. Mus. Nat. Hist.,

XXVII.

Ibid

1920, The Origin and Evolution of the Human Dentition,
Journ. Dental Research, vol. II.

Guppy, H. B. . .

1887, The Solomon Islands, p. 160.

Haswell, W. A.

1914, Fed. Hdbk. Brit. Assn. Mtg., Aus.

Hedley, C.

. . 1893, Natural Science, III., pp. 187-191.

Ibid

. . 1893, Austr. Assn. Adv. Science, V., pp, 444-446.

Ibid

.. 1895, Jr. & Pr. Roy. Soc., N.S.W., XXIX., pp. 278-286.

Ibid,

1899, Proc. Linn. Soc., N.S.W., XXIV., pp. 391-417.

Ibid

1909, Austr. Assn. Adv. Sci., XII., p. 333.

Ibid

1912, Proc. Linn. Soc. Lond., 124th Session, pp. 80-90.

Jensen, H. I. . .

.. 1912, Proc. Roy. Soc., Qld., XXIII., pp. 149-198.

Jentink:

.. 1911, Notes Mus. Jentink, 33, pp. 233-238.

Kamteu, P. 1ST., van . .

1909, Bijblad Nat. Tijdschr., Ned. Indie.

Herbert, C.

. . 1913, Zool Anz., pp. 162-167.

THE ZOOGEOGRAPHY OF MARSUPIALS. — LONGMAN.

15

Longman, H. A.

Ibid

Ibid

Ibid

Lord, C. E.
Lydekkee, R.
Maiden, J. H.
McCulloch, A. R.
Mjoberg, E.
Molengraaf, G. A. F
Osgood, W. H.

Regan, C. Tate
Scott, H. H.

Ibid

Scott & Lord
Sinclair, W. J.
Spencer, Baldwin

Ibid

Ibid

Thomas, Oldfield
Ibid. . .

1916, Mem. Qld. Mus., Vol. V. ( Muridce ).

1917, Proc. Roy. Soc. Qld., vol. XXIX., p. xii. ( Phascolomys ).
1921, Proc. Roy. Soc. Qld., XXXIII., p. xiv. ( Sarcophilus ).
1921, Mem. Qld. Mus. VII., pt. 2 ( Euryzygoma ).

1918, Papers Royal Society, Tasmania, pp. 16-52.

1907, “Nature,” vol. 75, p. 499.

1914, Fed. Hdbk., Brit. Assn. Mtg., Aus.

1915, Austr. Zool. I., pp. 47 49.

1915, Vet. Ak. Hd., Bd. 52, pp. 13-20.

1921, The Geographical Journal, LVIL, February.

1921, Monograph of Csenolestes, Field Mus. Nat. Hist., Zool.
Ser. XIV., No. 1.

1914, Brit. Antarct. “Terra Nova” Exped., Zool. I., pp. 1-54,

1915, Geol. Sur. Rec. Tas., No. 4 (Nototherium).

1917, Mus. Brochure, No. 6, Launceston.

1920, 1921, 1922, Papers Royal Society, Tasmania.

1906, Repts. Princeton Univ. Exped. Patagonia, IV., Pal.
1892, Austr. Assn. Ad. Sci. IV., pp. 82-124.

1896, Horn Exped. Rep., pt. 1, pp. 171-199.

1900, Wynyardia, Proc., Zool. Soc., pp. 776 794, plates XLIX.
& L.

1888, Brit. Mus. Catal. Marsupialia.

1895, On Ccenolestes, P.Z.S., pp. 870-878.

(Genera and species of recent marsupials described by Oldfield Thomas are mentioned
in the text with dates ; these have mostly appeared in the Annals and Magazine of Natural History.)

Walkom, A. B. .. .. 1918, Proc. Linn. Soc. N.S.W., vol. 43, pp. 37-115.

Wood Jones, F. . . . . 1923, The Mammals of South Australia, pt. 1.

Weber & Beaufort . . . . 1912, Kon. Akad. van. Wetenschappen, Amsterdam.

1923, The Zoology of the Indo -Australian Archipelago (Published by the Section
Zoology of the I.C.O. Committee, Kon. Akad. van Wetenschappen, Amsterdam./