Review of fossil crocodilians from Australasia

12 Pages·1997·0.67 MB·English

by  P. M. A. Willis

Review of fossil crocodilians from Australasia
'Quinkana Pty Ltd, 3 Wanda Crescent, Berowra Heighrs, New South Wales 2082
ABSTRACT
Recent phylcgenetic studies of fossil crdilian material from Australia, New Guinea and New
Caledonia (e.g., Willis et al. 1993; Willis 1995; Salisbury and Wiliis 1996) indicate most extinct
crocodilians from Australasia form an endemic clade, the Mekosuchinae, and extant crocodilians in
this region are more recent arrivals from Asia. Mekosuchines display a variety of ecomorphs in head
shape ranging from small and deep to iongirostrine to very deep snouted and the significance of these
is discussed.
INTRODUCTION bones of crocodiles in Crinum Creek near Peak
Downs in Queensland. Daintree (1872) mentions
Although fossil crocodilian material has been
crocodile teeth from Maryvale Creek in north-
recovered from Australia for more than 125
eastern Queensland which he named Crocodilus2
years, it has only been within the last decade
australis3. Jack and Etheridge (1892) refer all
or so that this group has received adequate
crocodile material from this site to Crocodilus
attention. This recent interest in crocodilians is
porosus.
the result of a dramatic increase in high-quality
specimens recovered during broad-based Typical of most British colonies of last century,
palaeontological excavations and a consequent important fossil specimens from Australia were
increase in research efforts. This group has sent to the major collections in England, usually
revealed a more complicated and interesting the collections of the British Museum of Natural
history than previously realized. This review History. Thus, when Lydekker (1888) published
compiles both historic and recent literature on the Catalogue of the Fossil Reptilia and Amphibin in
Australasian fossil crocodilians and attempts to the British Mweum (Natural Hivtory), it contained
summarise the current understanding of their the first published records of many Australian
taxonomy, phylogeny and evolutionary history. specimens and localities. Lydekker records
Crocodilur porosus from Clifton Plains, Gowrie
After an opening section on the earliest and the Condamine River in Queensland. These
literature of Australian fossil crocodiles, this specimens were donated by Harkness in 1871,
paper proceeds in a broadly geochronological Daintree in 1872, Bennett in 1872 and Hood in
order looking at the known material for each 1861. Some of this material is being reassessed
taxon during each successive period. Some taxa and may be assigned to Pallimnarchus.
(e.g., Quinkanu and Mehsuchus) are discussed
Jack and Etheridge (1892) record a number of
more than once because they occur in more than
occurrences of crocodilian fossils in Queensland.
one geological period. Summaries of current
They note the occurrences mentioned by Clark
phylogenetic analyses are then given, followed by
(1869), Daintree (1872) and Lydekker (1888)
a consideration of crocodilian evolutionary history
and list both Crocodilus porosur and Pallimnarchw
in Australia and comment on the significance of
pollens from the Darling Downs. In a later section
the broad range of crocodilian ecomorphs present
of the same publication dealing specifically with
in Australia during the Cainozoic. Although the
the fossil vertebrates of Queensland, Jack and
bulk of the crocodilians considered here are
Etheridge list C. porosus from Maryvale Creek,
from Australia, mention is also made of fossil
and at Cowrie and Clifton Plains on the Darling
crocodilian material from New Guinea, New
Downs. They also list P. pollens from Chinchilla,
Caledonia and the Solomon Islands.
Weimhilla Creek, Gowrie Creek, Condamine
Although extensive use of unpublished work is River and Eight-mile Plains (near Brisbane).
made here, it is appropriate because it provides
a more complete picture of work in this area and CRETACEOUS CROCODILIANS
most of this work is nearing publication. More There are now three published records of
information of work cited in preparation can be crocodiles from Cretaceous sediments of Australia
obtained by contacting the referred researcher. (Fig 1). These occurences are fragmentary and
EARLY LITERATURE
'The archaic spellir>g of Cmlodtlur has hcrn rciained nhrrc cmplo!rd h l rhc
The first record of a fossil crocodilian from adtho.,.
Australia is given by Clarke (1869) who refers to 'l'hir clam< has also hcrn applied to Soulll Ameriran lbrml nox recognized as
sprrie, of Prdl;garnr (steel l(li3).
March 1997 Australian Zoologkt 30(3) 287
Cenomanian specimen includes the lumbar,
sacral and proximal caudal vertebrae, the pelvis
and most of the bind limbs. This specimen is
currently being studied by Ralph Molnar and
the author.
EARLY TERTIARY CROCODILES
A number of crocodilian specimens have been
recovered from Paleocene to Oligocene deposits
and all come from Queensland (Fig. 1). Most of
this material is fragmentary and uninformative
but material from the Tingamurra deposits
near Murgon and the Rundle Oil shales near
Q mMdDd . Bt9 Gladstone is more complete and informative.
MH*Y&vs Uetac~ous Most of this material can be referred to the
T 7rddphmvbus .Eocene genus Kambara.
Hf@-P-*rn .E~.ne-Oligm8ne
A Aust&wwa"s .oUgo-Mime
E Unidentil1.d emmian = htioerne Kambara
Figure I. Location of Mesozoic, early and mid Teruary
The early Tertiary genus Kambara was first
crocodile-bearing deposits in Australia. After Hills 1943;
Mchlamara 1993; Megirian et 01. 1991; Megirian 1994; described by Willis et al. (1993), who recognized
Megirian et a1 1993; Molnar 1980; Molnar 1991; Murray a single species (K. murgonensisj from the Eocene
and Meginan 1992; Murray etal. 1993; Rich and Rich 1989; Tingamurra deposits in south-east Queensland.
Willis 1993; Willis 1995; Willis rt ol. 1990; Willis et a/. 1993;
Subsequently, a second species of Kambara has
Willis and Molnar 1991a; Willis and Molnar 1991b.
been recognized from this site (Salisbury and
Willis 1996). Material from the Eocene-
uninformative but probably contemporaneous
Oligocene oil shale deposits of Rundle, Eastern
and possibly conspecific. Three other, previously
Queensland, briefly described by Hills (1943)
unpublished Cretaceous crocodilian specimens
and Molnar (1991), may represent a third, as yet
are also recorded in Australia. It is hoped that
undescribed species of Kambara. The anterior
further material will be found to add to our
dentary from Murgon described by Molnar
knowledge of this important group during the
(1982b) has been referred to Kambara (Willis et
Cretaceous.
al. 1993).
Etheridge (1917) referred a small opalized
Kambara is a moderately broad-snouted,
fragment of reptilian mandible from Albian-
generalist crocodile notable for the disparity
Aptian deposits at Lightning Ridge, New South
in size between the smallest and largest teeth
Wales to Crocodylw (Bottosaurwj sehlophensis.
(Fig. 2). This tooth disparity is also seen in other
Molnar (1980) referred a procoelous vertebra
mekosuchines but Kambara is the earliest member
and several other pieces from the same locality
of the group to display it. Distinctive sculptural
to this taxon and concluded that it represented
patterns, such as large sculptured pits between
an eusuchian.
the orbits and the supratemporal fenestrae, and
A crocodilian quadratojugal and two scutes a deep excavation anterior to the internal nares
have been recovered from early Cretaceous also characterize this genus.
sites at Dinosaur Cove, in the Otway Ranges of
A possible fourth species of Kambara has been
Victoria. This undescribed material is fragmentary
recorded by Willis and Molnar (1991a) from
and not very informative but indicates a small
(under 2 m long) crocodile within the Dinosaur Eocene sediments near Brisbane. This material
consists of an anterior dentary fragment and
Cove fauna (Rich and Rich 1989).
an isolated vertebral centrum. This is a more
A recent acquisition by the Australian Museum longirostrine form than the other three species
of opal fossils from Albian-Aptian deposits of but it does display a similar tooth disparity.
Lightning Ridge, New South Wales, includes a
small crocodilian dentary. This dentary is quite An intriguing recent discovery is of Eocene-
different from the dentary referred to Crocodylus Oligocene crocodile scutes from Cape Hills-
(Bottosauu) sehlophensis and is currently being borough, northern Queensland (McNamara
studied (Willis and Molnar, in prep.). Some other 1993). These small, squarish scutes probably
crocodilian material from Lightning Ridge is belong to an advanced neosuchian but otherwise
also being studied by Ralph Molnar in the this material is uninformative.
Queensland Museum (pers. comm.).
Riek (1952) records a piece of skin and some
A posterior section of an articulated crocodilian bone fragments possibly from a crocodile found
skeleton was recovered from central Queensland in the early Tertiary Redbank Plains Series near
in mid 1996 (Molnar and Willis 1996). This Brisbane. The bone fragments cannot be found
288 Australian Zoologist 30(3)
are now recognized in many of the older sites at
Riversleigh and the larger species is restricted to
some of the very earliest sites. Similarly, another
smaller species of Baru may be present at
Bullock Creek (Willis 1995).
Crocodilian material from mid to late Miocene
sediments at Alcoota was originally described by
Woodburne (1967) as Crocodylw sp. Murray and
Megirian (1992) examined new crocodilian
material from the site and assigned it to a species
of Baru, possibly B. darrowi. It is likely that
the material Woodburne (1967) identified as
Crocodylus actually belongs to Baru.
Quinkana
Species of Quinkana were originally recognized
from Plio-Pleistocene deposits of Queensland
(Molnar 1981) but recently prepared material
from Riversleigh and Bullock Creek indicate new
species of @inkana in these earlier deposits (Willis
1992; Willis, in press; Murray and Megirian
1992; Murray el al. 1993; Megirian et al. 1993;
Figure 2. Kambarn murgonensls, reconstrucrion of skull. Top:
Megirian, in prep.). The Quinkana species from
dorsal vie,". Bottom: ventral vie%,. Scale bar represen6 10 cm.
Riversleigh are all quite small in comparison to
previously known species and are distributed
and it is not clear from Riek's paper that these
over a number of sites. Megirian (1994)
fragments were actually collected. Known skin
described the new species of Quznkana present
impressions from this site represent lepidosaurs
in Bullock Creek as Q. timara with a significantly
and lungfish but not crocodilians (Molnar, pers.
narrower snout than that of Q. fortirostrum. The
comm.).
MIDDLE TERTIARY CROCODILIANS
Until quite recently, crocodilian material from
segn a rapid increase in the number of specimens
from Riversleigh in northwestern Queensland,
Bullock Creek and Alcoota in the Northern
Territory and Miocene deposits in the Lake
Eyre Basin of South Australia (Fig. 1). Sites at
Riversleigh and Bullock Creek have produced
particularly diverse and interesting crocodilian
faunas with some sites containing up to four
crocodilian species (Willis 1992; Willis 1995;
Murray and Megirian 1992).
Baru
Relatively complete cranial and mandibular
material from Riversleigh and Bullock Creek
formed the basis of the new genus and species
Baru dawowi Willis et al. (1990). This is a
moderately large (around 4 m long) crocodile
with a particularly massive head, enormous
caninifom teeth that are moderately compressed
and a broad posterior region of the skull allow-
ing large areas for massive jaw musculature
(Fig. 3). Subsequent work has shown that the
Riversleigh material represents a second species
separate &om B. &w&i of Bullock Creek (willis
1992; Willis, in press) and that a smaller, third r&,e 3. Baw dnnowi, reconstruction of TOP: dorsal
species also exists at Riversleigh. Species of Ban view. Bortom: ventral view. Scale bar represents 10 cm.
March 1997 Australian Zwlogist 30(3) 289
specific determination of the @inkana in the
Ongeva Local Fauna of Alcoota (Murray et al.
The unusual species Trilophosuchw rackhami
1993; Megirian et al. 1993) cannot be made
was described by Willis (1993) from Ringtail
based on material currently available.
Site at Riversleigh. Similar to species of Mehsuchw
@znkann is a genus of ziphodont crocodilians in having a short deep head and being small,
that are demonstrably mekosuchine and Trilophosuchw bas the striking feature of three
endemic to Australia. Thus their acquisition of longitudinal crests across the skull roof. Low
ziphodont characters is independent from and bumps indicate a poorly developed second
convergent on other ziphodont crocodilians pair of crests lateral to the other three. The
from around the world. They are thought to have supratemporal fenestrae are elongate and
been fully terrestrial based on comparisons with laterally compressed which is either a juvenile
pristichampsines and sebecosuchians and some or paedomorphic feature. An analysis of the
aspects of the taphonomy of the Pleistocene neck muscle insertions (Willis 1993) indicates
species Q. fortirostrum (Molnar 1977; Molnar more similarities with the morphology of the
1978a; Molnar 1978b; Molnar 1981). The early ziphodont Pristichampsus than to more typical,
presence in Australia of ziphodont crocodilians non-ziphodont crocodilians and this has been
and other large reptilian carnivores together taken as tentative evidence for a terrestrial
with the absence of large mammalian carnivores habitat for Trilophosuchw.
indicates that the occupation of the large
carnivore guild noted for Australian Picistocene
faunas (Murray 1984) probably has a much The imperfectly known Harpacochampsa
more extensive history. camfieldensis from Bullock Creek in the Northern
Territory is based on a posterior cranial fragment,
Mekosuchus an anterior snout fragment and a small piece
A species of Mekosuchw was first recorded from of the left dentary, all probably from the same
Pleistocene or Recent deposits of New Caledonia individual, as well as some postcranial material
(Balouet and Buffetaut 1987) so the appearance (Fig. 4). Megirian et al. (1991) recognized certain
of two species in Miocene deposits of Riversleigh longirostrine characters, such as the extremely
(Willis 1992; Willis 1995; Willis, in prep.) was large and cavernous supratemporal fenestrae
surprising. This distinctive genus of small (less and the globose anterior of the snout that
were reminiscent of gavialids, but their
than one metre long) crocodilians is unusual in
phylogenetic analysis indicated mekosuchine or
that the maxilla participates in the orbit thus
crocodyline affinities. More recent phylogenetic
excluding the lacrimal from contact with the jugal.
analyses (Willis 1995; Salisbury and Willis 1996)
Mekosuchw species have very short, deep heads
indicate that Harpacochampsa is likely to be a
and may have been terrestrial or even arboreal (see
mekosuchine or crocodyloid where many key
below). This conclusion is based on postcranial
characters have been masked by the longi-
material from the Pleistocene species and compari-
rostrine condition. Harpacochumpsa is a moderately
son with atoposaurids, libycosuchids and noto-
large crocodilian, probably reaching lengths of
suchids which have a similar cranial morphology.
4 or 5 m and its longirostrine condition indicates
a piscivorous diet.
Australosuchus
The medium sized (around 3 m in length)
generalized crocodilian Awtralosuchw clarkae is
known from abundant material from a number
of sites in the Lake Eyre Basin, South Australia
(Willis and Molnar 1991b). Being a moderately
broad-snouted mekosuchine, Awtralosuchw bas
the unusual condition of the fourth dentary tooth
notch in the snout being almost totally enclosed
(Fig. 5). Complete skulls and mandibles are known
of A. clarkae as well as abundant postcranial
elements, dermal armour and an articulated
partial skeleton. Reports of A. clarkae from
younger sediments (Megirian et al. 1993; Willis
and Molnar 1991b) are based on probably
reworked specimens.
An interesting palaeobiogeographic problem is
presented by the distribution of contemporaneous
Figure 4. HarpIcochnmpa camfielderuis, reconstruction of skull.
Top: doml view. Bottom: ventral view. Scale bar represents crocodilian taxa in Australia during the Oligo-
10 cm, Miocene. Although some of the sites in the Lake
290 Australian Zoologist 30(3)
inhabited the inland drainage basins of Australia
as well as the coastal basins until comparatively
recently (late Pleistocene in most cases). The
demise of crocodilians from inland Australia was
probably coupled with the collapse of the
megafauna and general drying out of this part
of the continent.
Pallimnarchus
The first detailed discussion of crocodilian
fossils from Australia was by De Vis (1886),
when he named Pallimnarchw pollens De Vis
noted that P. pollens was a heavily built, large
crocodilian. Later De Vis (1907) described more
fossil material from the Gulf of Carpentaria and
assigned this to P. pollens. De Vis never formally
defined Pallimnurchw or P. pollens but the names
gained a de Jaclo validity through use in popular
and scientific literature. De Vis was somewhat
equivocal about the affinities of P. pollens variously
comparing it to alligatorines and crocodylines
but favouring crocodyline affinities.
Figure 5. Reconstruction of the skull of Awlralosuchw clavkae
in dorsal (top) and ventral (bottom) views. The posterior Longman (1925a) described a large crocodilian
part of the cranial surface is unknown (adapted from Willis
snout from Lansdowne Station, central Queens-
and Molnar 199la).
land, referring it to P, pollens. Molnar (1982b)
re-examined this specimen and identified it as
Eyre Basin in South Australia and Rivenleigh in
C7ocudylw porosus. Longman (192513) also
Queensland are demonstrably contemporaneous
described a large, heavily built crocodilian
(Woodburne et al. 1994) and there are large
mandible from Macalister, southeastern Queens-
collections from both localities representing
land, refemng it to P. pollens.
many crocodilian specimens, there are no
common crocodilian taxa shared between the Anderson (1937) described a pair of large
two areas. While this disparity in distribution for mandibular fragments from the Gregory River,
semi-aquatic taxa (Baru and Awtrahuchw) can north Queensland, referring them to P. pollens.
be explained by separate and isolated drainage There is no more precise locality information
basins, this explanation does not extend to taxa and it cannot be determined if they derived from
that are supposedly terrestrial or are at least
competent in terrestrial environments (Qumkana,
Mekosuchw and Tnlophosuchw). Either the
barriers to distribution for small to medium sized
terrestrial crocodilian taxa were formidable or
their abilities to cross terrestrial environments
were not as developed as has been supposed.
Perhaps the "terrestrial" adaptations of QuinRana,
Mekosuchu~ and Trilophosuchw represent the
terrestrial limit of the operating potential of
a mostly aquatic species. For example, the
ziphodont characters of Quihna may well have
allowed it to take prey and process food on land
but they were still tied to permanent water bodies
for protection, reproduction or thennoregulation.
P ~a~~imnomrnn
Q OuiMbnd
PLIO-PLEISTOCENE CROCODILIANS c Crmn*Iu_~ 9
The first records of crocodilians from Australia
Figu~e 6. Location of Pliocene and Pleistocene crocodile-
were from Plio-Pleistocene deposits of Queens-
bearing deposiu in Australia. After Andemon 1937; Anderson
land (Clarke 1869; Daintree 1872; Lydekker and Fletcher 1934; De Vis 1886; De Vis 1907: Dodsonetal
1888). Subsequent excavations have revealed 1993; Corter and Nicoll 1978; Hecht and Archer 1977;
Plio-Pleistocene crocodilians from northern Imngman 19'24; Longman 1925a; Longman 1925b;
McNamara 1990; Megirian etal. 1993; Molnar 1977: Molnar
New South Wales, northern Western Australia
l978a; Molnar 1978b; Molnar 1979; Molnar 1981; Molnar
and central South Australia (Fig. 6). Quite
L982b; Tedford and Wells 1990: Thompson 1980: Willis and
large crocodilians (5 m or more in length) Archer 1990; Willis and Macknesr 1996; Woodburne 1967.
hstraiian Zoologist 30(3) 291
Pleistocene sediments of the kind currently To date only a single species of Pallimnarchw,
known from Riversleigh or some other, unknown P. poilm, has been recognised and described.
deposit. However, new material and an extensive review
of existing Pallimnarchus material reveals that
Molnar (1982b) reviewed the Cainozoic a second species, and possibly more, may be
crocodilians of Queensland with particular represented (Willis and Molnar, in prep).
attention to PaUimnarchus. Although the concepts
Pallimnarchus and P. pollens had received wide- Crocodylus prosus
spread common usage, it was not until Molnar's
Molnar (1979) identified the snout of a small
revision that these concepts were properly defined
crocodilian from Bluff Downs, north Queens-
and lectotype material nominated. Molnar's
land, reported by Archer and Wade (1976), as
generic and specific definition was based on
Crocodylw porosw. This is the first time fossil
features of the anterior dentary, but referred
crocodilian material from Australia was correctly
material included other mandibular elements,
identified as belonging to this extant species. The
premaxillae and quadrates. Molnar also
dating of Bluff Downs is based on WAr dating
described an interorbital fragment and assigned
of an overlying basalt, originally thought to be
it to P. pollens. Molnar (1982b) showed that C.
4-4.5Ma, but recently revised to a significantly
nathani Longman 1924 was an invalid taxon and
younger date within the Pliocene (Mackness,
variously assigned the C. nathani material to
pers. comm.). This makes this specimen the
either C. porosur or P. pollens.
second oldest fossil record of any extant species
of crocodilian (the oldest being for a specimen
Willis and Molnar (in prep.) are devising a
of Gavialis gangetacur).
diagnostic key to aid in the identification of isolated
crocodilian teeth from Plio-Pleistocene sites in Molnar (1982b) made a case to remove the
Australia. The difficulty of identifying isolated Landsdowne snout described by Longman
teeth could mean that teeth attributed to C. (1925b), from P. pollens and place it in Crocodylw
within a fauna actually represent P. pollas. porosus. Recent reexamination of this specimen
~ ~ ~ O S U S
A review of situations where misidentification is by Willis and Molnar (in press) reveal that it does
a possibility is currently being conducted. Based belong to Pallimnarchw but probably represents
on isolated teeth, Pallimnarchus is probably a new species. Molnar (1982b) also showed that
present at Cuddie Springs in northern New South C. nathani was an invalid taxon and variously
Wales (Anderson and Fletcher 1934; Dodson et assigned the C. nathani material to either C.
al. 1993) as well as the Wyandotte Local Fauna porosur or P. pollens. Molnar found that there was
of northern Queensland (McNamara 1990).T he very little fossil material known from Australia
report of Crocodylw from Windjana Gorge, that could be referred to C. porosus.
Western Australia (Gorter and Nicoll 1978), is
Cmodylus johnstoni
based on a premaxilla and scutes and recent
examination of this material by Willis and Being an endemic species, the apparent absence
Molnar (in press) reveal that it should be referred of C. johnstoni from fossil deposits of Australia
to Pallimnarchns. Reports of Pallimnarchw from seemed enigmatic until Willis and Archer (1990)
various Pliocene and Pleistocene sites in central described a dentary from Pleistocene deposits at
Australia (Tedford and Wells 1990; Megirian et Riversleigh, northwestern Queensland. In a note
al. 1993) are based on more substantial frag- added in proof to Willis and Archer (1990), a
ments than isolated teeth. second dentary was identified from Pleistocene
deposits of Floraville Downs in northwestern
Thompson (1980) reported two Pleistocene
Queensland. Both specimens appeared to be
river bank deposits along the Darling River in
somewhat large for this species but they were
western New South Wales, one of which (Myrtle
otherwise unremarkable.
Vale) contained a crocodilian tooth. Thompson
describes the conical tooth as being 15 mm across Oocodylus nathani
the base. This could represent a large C. porosus
Longman (1924) described crocodilian material
but would more likely represent Pallimnnrchw
from Tara Creek, north Queensland, as Crocodylw
sp. because of its size. The present location of the
nathni. He defined this taxon based on the
fossils Thompson collected is unknown and the
position of the posterior extent of the mandibular
Myrtle Vale site has been submerged under the
symphysis and the supposed sequence of
waters of a weir. Thus there appears to be little
tooth enlargement that he inferred from two,
chance of correctly determining the identity of
unassociated mandibular fragments. This taxon
this crocodilian or of collecting more material.
is now recognized as invalid (Molnar 1982b).
However, similar deposits may occur elsewhere
along the Darling River (Jeannette Hope, pers.
Crocodylus sp.
comm.). Myrtle Vale and Cuddie Springs are the
most southern records of Pleistocene crocodilians Some recent identifications of material as being
in Australia. Crocodylw sp. should, on review, be assigned to
various mekosuchine taxa. Swinton (1924) festooning (completely suppressed in other
identified crocodilian remains collected from the species of Quinkana) and is known from a partial
Lake Eyre Basin, South Australia, by Gregory maxilla and isolated teeth.
and referred them variously to Crocodilus or Isolated ziphodont crocodilian teeth from
C. porosus. This material may be referred to Pleistocene localities in Meganesia should
species of Pallimnarchus (pers. obs.) which would tentatively be refelred to Quinkam sp cf. fios-
be consistent with the findings of Megirian et al.
tmm. This includes the ziphodont teeth referred
(1993). Megirian et al. (1993) do record a
to by Plane (1967) from the Awe Fauna of Papua
Crocodylm-like specimen from the Kutjitara
New Guinea, as well as teeth from Croydon,
Formation of the Lake Eyre Basin which has north Queensland and various sites on the
possibly been reworked from other deposits. A Queensland Darling Downs reported by Molnar
jugal possibly referrable to Crocodylm is also (1981) and Cuddie Springs in New South Wales
reported from the Katipiri Formation of the Lake as reported by Dodson el al. (1993).
Eyre Basin (Megirian et al. 1993). However,
Megirian (pers. comm.) cautions that designations Gavialis papuensis
from these sites are based on fragmentary
De Vis (1905) described fragmentary man-
material with few diagnostic features.
dibular material from Woodlark lsland in the
Gorter and Nicoll (1978) referred a poorly Solomon Sea as the new species GavialiF papwnm.
preserved premaxillary fragment and some scutes Molnar (1982a) reviewed this material and
from Winjana Gorge, Western Australia, to assigned it to a Euthecodon-like crocodile. Sub-
Crocodylus sp. but this material should be referred sequently Aoki (1988) and Molnar (1993)
to Pall2mnarchus (see above). Woodburne (1967) suggested that it could belong to a malformed
described crocodilian fossils from Alcoota, individual of Gauialw bengawanicus. A popular
Northern Territory, and referred them to account of this taxon is given by Willis (1987).
Crocodylus sp. but this material should probably
be referred to Baru. Plane and Gatehouse (1967) Mekosuchus inexpectatus
noted the presence of "nondescript" crocodile Balouet and Buffetaut (1987) described a small,
teeth in the Awe Fauna from New Guinea. These deep-snouted crocodile from cave deposits in
can probably be referred to a species of Crocodylus New Caledonia as the new genus and species
but may represent a species of Pallimnarchus. A4ekosuchm inexpectatw. These deposits were
Resolution of this matter requires a x-examination initially dated as Recent (Balouet 1991) but there
of the material. is some doubt about these dates and the deposits
are more likely to be Pleistocene (Tim Flannery,
Quinkana
pers. comm.). Being an unusual, short-faced,
The first reference to ziphodont crocodilian deep-headed form, initial phylogenetic analyses
material from Meganesia is by Plane (1967) erroneously suggested that Mekoruchw formed a
who recorded ziphodont teeth in the Awe Fauna sister group relationship to all other eusuchian
from New Guinea. Hecht and Archer (1977) crocodilians (Balouet and Buffetaut 1987). More
described two maxillary fragments, one from the recent analyses have interpreted some of the
Lake Eyre Basin, South Australia, the other
from Texas Caves, Queensland, and referred
them to an unknown sebecosuchian. Meanwhile complex, similar to the morphology of Osteolaemus
Molnar (1977), Molnar (1978a) and Molnar and Paleosuchus (Willis 1994, 1995).
(1978b), working with more complete ziphodont
The biogeographic distribution of species of
material from Chillagoe Caves, north Queens-
Mekosuchus (Miocene species in north Queens-
land, recognised that this ziphodont form was
land, Pleistocene species in New Caledonia)
eusuchiin with similarities to the pristichampsines.
presents an intriguing problem. Postcranial
Molnar (1981) described and defined this form
material known for M. inexpectatus indicates a
as the new genus and species @inkana Jortirostrum
fully terrestrial habit. Indeed, the form of the
and referred to this genus the material described
humerus is so similar to that of species of Varanus
by Hecht and Archer (1977) from Texas Caves.
it is even possible that this was an arboreal taxon.
Unfortunately the specimen from the Lake Eyre
Such a terrestrial species is unlikely to have
Basin has been lost and its refenal to, or exclusion
swum between Australia and New Caledonia. It
from, Qunikana cannot be tested. However, until
is also unlikely that the distribution of Meko.mchw
evidence to the contrary can be produced,
species is the result of a vicariant event, the last
previous claims for sebecosuchians in Australia
land bridge between the two landmasses occurred
(Hecht and Archer 1977) should be discounted.
80 million years ago in the late Cretaceous
A Pliocene species of Quznkana has recently (Balouet 1991). Being a small (no more than 1 m
been recognised from the Spring Park locality, snout-vent length) and possibly an arboreal
north Queensland (Willis and Mackness 1996). species, the distribution ol species of Mekosuchus
The new species, Q. babarra, differs from Q. may have been a dispersal by rafting from Anst-
fo~tirostmm in having a shorter snout with mild ralia to New Caledonia.
Australian Zoologist 30(3) 293
PEWLOGENETIC SYSTEMATICS EVOLUTIONARY HISTORY
Early work on crocodilian taxa from Australia The phylogenetia of the Mekosuchinae appear
recognized various taxa as either crocodyline to neatly fit into the zoogeographic history of
or alligatorine (e.g., De Vis 1886) although Australia. The earliest representatives of the
crocodyline affinities were favoured (e.g., group (Kambara spp.) were present in Australia
Longman 1925a; Longman 1925b). Quinkunu prior to the rifting of Australia from Antarctica
was initially thought to be sehecosuchian (Hecht and while Antarctica was still attached to South
and Archer 1977) or eusuchian, possibly pristi- America. Such a land bridge may have offered
champsine (Molnar 1981). It was not until more a potential dispersal route for mekosuchines into
recent work that similarities were recorded for or out of Australia. Similarly, alligatorines and
Australia's endemic crocodilian taxa (Willis 1986; other crocodilians present in South America at or
Willis et al. 1990; Willis and Molnar 199lb). prior to this time (e.g., Gasparini 1981; Gasparini
Initially, this group was termed the Australian 1984; Gasparini and Buffetaut 1980; Gasparini
Tertiary radiation of crocodiles (Willis el al. 1971; Langston 1965; Simpson 1937a; Simpson
1990). Cladistic analyses supported the 1937b) were also presented with dispersal
monophyly of these taxa (Willis 1993; Willis opportunities into Australia. However, although
1995). The subfamily Mekosuchinae was erected some mammal taxa appear to have migrated via
by Willis el al. (1993) to encompass the pre- this route (Archer et al. 1993,1994), neither the
existing concept of an Australian Tertiary mekosuchines nor South American crocodilian
radiation of crocodilians, redefining the taxa appear to have dispersed through Antarctica.
taxonomic concept of Mekosuchidae originally
Whatever the origin of the mekosuchines,
erected by Balouet and Buffetaut (1987).
they were present in Australia before its isola-
Mekosuchine characters include a great tion from Antarctica and, once isolated, provided
disparity of alveolar size between the smallest the rootstock of subsequent crocodilian faunas
and largest alveoli, the development of a wedge
throughout the Tertiary until the Pliocene.
of the supraoccipital on the dorsal surface of
Possibly as a response to the nutrient-poor
the skull and reduction or loss of the anterior
environments of Australia favouring large
process of the palatines. The Mekosuchinae
ectotbermic predators over endothermic
includes Bam, Quinkana, Mekosuchw, Trilo-
predators (as suggested by Flannery 1994), the
phosuchw, Pallimnarchus, Kambara and
mekosuchines evolved diverse crocodilian
Awh-alosuchw. The very derived, longirostrine
faunas throughout the Tertiary including the
genus Harpocochampsa (Megirian et al. 1991) is a
occupation of terrestrial habitats (see below).
problematic taxon previously of uncertain phylo-
The decline of mekosuchine diversity and their
genetic affinities. However, Willis (1995)
eventual extinction is probably related to the
presents an argument for its inclusion in the
collapse of the associated Australian megafauna.
Mekosuchinae.
Competition between mekosuchines, varanids,
The Mekosuchinae forms a sister g~oupto the madtsoiids, marsupial carnivores and newly
Crocodylinae which indudes species of Crocodylw arrived crocodylines could also have contributed
and Osteohm. Character distributions are to the demise of mekosuchine taxa.
shown in Table 1 and the wider affinities of the
Two species of Crocodylw are currently endemic
Mekosuchinae are shown in Figure 7.
to Australia and they are not mekosuchines.
T.L. 23, 1 tree YetosncBus The sudden appearance of the crocodylines
C.I. 0.783 (Crocodylw porosw) in the early Pliocene record
rr~~opbosncbus of Australia (Molnar 1979) suggests a dispersal
event, probably from south-east Asia, sometime
in the early Pliocene or possibly the latest
Miocene, as suggested by Willis (1986) and Willis
et al. (1990). The salt tolerance of this species
allows it to disperse across oceanic barriers (Taplin
1984). The subsequent appearance of C. johtoni
in Australia suggests a speciation from a C.
pmosus ancestor (Willis and Archer 1990).
ECOMORPHIC DIVERSITY
Mekosuchines display a broad range of
ecomorphs from small taxa such as Mekosuchus
and T~hosuchu(sp robably less than 1 m long)
to comparative giants such as Pdimnnrchw
(reaching lengths in excess of 5 metres). Even
Figurr 7. Cladopm showing the relationships of the
more striking are the differences in head shape
rnekosuchines based on an analysis of 15 charnuem.
Characten unweighted. (After Willis 1995). from small deep "box-heads".(Mekosuchw spp.
TnOle 1. Character distribution for analysis of the Mekosucharine.
Number of
Character 1 2 3 4 5 6 7 8 9 1 0 11 12 13 14 15 unknowns
iMekoruchui 1 2 1 1 1 1 2 1 1 ? 1 0 1 1 0 I
T~lophorucl~u.~ 0 1 ? 1 1 ? 2 1 I ? 1 0 2 1 0 3
Quinlmna 1 2 1 0 ? I 2 1 I I 1 1 ? 0 ? 3
Bnnr 1 2 1 1 ? 1 1 1 I I 0 1 0 0 1 I
Kambaro 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 I
Aucrolaruchw 1 1 1 1 I 0 0 0 0 0 0 0 0 0 0 0
I'nllimnarcht~r 1 1 1 1 1 0 0 0 0 0 0 0 0 0 1 0
Hnrpacochompro ? 1 1 1 0 0 0 ? 0 0 0 0 0 0 0 2
Crocodylw 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Brochyuronochompsa 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Unknowns 1 0 1 0 2 I 0 1 0 2 0 0 1 0 1
Character states
(1) Presence or absence of an anterior process of thepaliltines. (9) Locarion of rhe orbi~s.
Present (0) Posteriorly locared orbits (0)
Absent (1) Anleriorly located orbirs (I)
(2) Presence or absence of an alveolar process. (10) Form of the snout nocch.
Absent (0)
Symmetrical notch (0)
Law alveolar process (I)
Inclined notch (1)
Well developed alveolar process (2)
(I I) Direction of the orbits.
(3) Size of the fifth premaxillary alveolus relative to the hrst
maxillary alveolus. Dorsally facing orl~irs( 0)
Fifth premaxillary roorh smaller than rhe first maxillary Laterally facing orbits (I)
tooth (0) (12) Degree of alveolar compression.
Fifth premaxillary tooth larger than the first maxillary Rounded section (0)
tooth (I) Compressed section (I)
(4) Presence or absence of a great disparity in 100th she. (13) Degree of lareral compression of the supratemporal
Srnall tooth disparity (0)
fenestrae.
Large tooth disparity (1)
Uncompressed supratemporal fenestrae (0)
(5) Presence or absence of a supraoccipital wedge. Moderately compressed (I)
Absent (0) Strongly compressed (2)
Present (I)
(14) Degree of participation of rhe prerygoids in the narial
(6) Shape of the ex~ernafn ares. passage.
Oval or round nares (0)
Restricted pterygoid participation (0)
Trapezoidal nares (1)
Exrensive participation (I)
(7) Depth of the snout.
(15) Size of the supratemporal fenesrrae.
Shallow snout (0)
Moderately deep snout (I) hloderate sized supratemporal fenestrae (0)
Very deep snout (2) Small fenestrae (1)
(8) Localion of the palatal fenestrae.
Posterior to 8th alveoli (0)
Reach or anterior ro rhe 8rh alveoli (I)
and Trilophosuchw rackhami), to long but broad- these species are similar in size and aspects of
snouted ziphodonts (Quinkana spp.), to massive- their cranial morphology. Either the differences
headed, "hatchet-mouthed taxa (Baru spp.), to in skull morphology that exist between these two
moderately broad snouted, generalized forms species reflect enough diversity of resource use
(Awtralosuchus clarkae and Kambara spp.) to to allow the two to be sympatnc or the two
large, very broad snouts (Pallimnarchuss pp.) and species are found together only in a death
the longirostrine condition (Ha+cochamp.~a assemblage and were not syrnpatric in life. The
camfieldensis). Meyer (1984) demonstrated that recognition of a second species of Kambara in
more than one species of crocodilian can live the deposits of Murgon, identical in size to
syrnpatrically if they have different head shapes. K. murgonenses and with a very similar skull
The difference in head shape reflects differing morphology, raises a similar problem.
diets and thus resources can be divided by
CONCLUSIONS
syrnpatric species. This rule holds true for most
Australian localities where more than one species Despite a large collection of fossil crocodile
of crocodile has been found. For example in material having been compiled from all over
Bullock Creek there is a large species of Bam and Australia and Meganesia prior to the 1980s,
possibly a second, smaller species, a ziphodont this material received little attention. As a
species of Quinkana and a longirostrine species consequence the history of crocodilians in this
of Harpuochampsa. More difficult to explain is area was poorly understood. A dramatic increase
Ringtail Site at Riversleigh which has a species of in the amount and quality of material that
Mekomchus and Trilophosuhw rackhami. Both became available during the 1980s and the
consequent increase in interest in this group has Balouet, J. C. and Buffetaut, E., 1987. Mekosuchur Gn+xcfdus
provided a more complete understanding of n, g n. rp. Cmcodilian nouveau de 1'Holocene be
crocodilians in Australia and has produced Nouvelle Caledonie. C. R. Acod. Sc. Park. 304: 85.3-57.
some interesting and unexpected results. There
Clarke, W. B., 1869. Dinmir an Australian genus. Gcol. Mag.
is no reason to suggest that ongoing research 6: 38344.
on fossil crocodilians from Australia will not
Daintree, R., 1872. Notes on the geology of the Colony of
continue to be productive. Of particular interest
Queeensland. Q.J . Geol Soc. Lrmdm 28: 271-317.
is the origin of the mekosuchines and their
associated dispersal into Australia. On-going De Vls, C. W., 1886. On the remains of an extinct saurian
research into the fossil crocodilians of Australia Proc. Rq.S oc. Qld 2: 181-91.
is being conducted by myself and Steve Salisbury De VIS,C . W., 1905. FOSSILv ertebrates from New Gumea.
at the University of New South Wales, Ralph Ann. Qld Mus 6: 2631
Molnar at the Queensland Museum, Dirk
De Vs, C W., 1907 Fossllsof the Gulf M'atershed. Ann Qld
Megirian at the Northern Territory Museum of
Mur 7: 1-7.
Arts and Sciences and Neville Pledge of the
South Australian Museum. Dodson, J., Fullagar, R., Furby, J.,Jones, R. and Prosser, I.,
1993. Human and megafauna in a Late Pleistocene
environment from Cuddle Springs, northwestern New
ACKNOWLEDGEMENTS
South Wales. Archeeology in Oceanin 28: 9G99.
I would like to thank Ralph E. Molnar, Dirk
Etheridge. R., 1917. Reptilian Notes; Megnlonia @ca,
Megirian and John Scanlon for reading and
Owen, and Noliosounrr dmlohu, Owen, lacenian dermal
commenting on early drafts of this manuscript, armaur: opalized remains from Lightning Ridge. Proc.
their comments were most supportive and R. Soc. Vict. 49: 127-33.
beneficial.
Flannery, T. F., 1991. The Future Enters. Reed Books:
This review provides an overview of my Sydney.
doctoral thesis and would not have been possible
Gasparini, Z., 1971. Los Notosuchia del Cretaciw de America
without the assistance of many people and del sur coma un nuevo infraden de los Mesosuchia
organizations. Primarily I would like to thank (Crocodilia). Ameghiniane 8: 83-103.
my supervisor, Professor Michael Archer for his
Gasparini. Z., 1981. Los Crocodylia fosiles de la Argentina
help and assistance over the last eight years. AmrghinCa~2 8: 177-205.
Financial support from the Dean of the Faculty
of Biological and Behavioural Sciences (University Gasparini, Z., 1984. New Tertiary Sebecosuchia (Crocodylia:
Mesosuchia) from Argentina. J. Vwl. Paleont. 4: 85-95.
of New South Wales), the Australian Museum,
the Peter Rankin Trust Fund for Herpetology, Gasparini, 2. and Buffetaut, E., 1980. Dol~hochampsaminima,
the Linnean Society of New South Wales, the n. g. n. sp., a representative of a new family of eusuchian
crocodiles from the late Cretaceous of northern
Friedrich Wilhelms Universit%t (Bonn), and the
Argentina. N. Jb. Geol. Paht.M h. 1980: 257-71.
Hessische Landes Museum all contributed to the
completion of my thesis and I extend my sincere Gorter, J D. and Nicoll, R S., 1978. Reptilian fossils from
thanks to them all. Windjana Gorge, Western Australia. Proc. R. Soc. W.A.
60: 97-104.
REFERENCES Hecht, M. K. and Archer, M., 1977. Presence of xiphodont
crocodilians in the Tertiary and Pleistocene of Australia.
Anderson, C., 1937. Palaeontological notes No. 1V. Rec.
Alchminga 1: 38345.
Aul. Mu. 20: 73-77.
Hills, E. S., 1943. Tertiary fresh-water fishes and crocodilian
Anderson, C. and Fletcher, H. O., 1934. The Cuddie
mmains from Gladstone and Duaringa, Queensland.
Springs Bone Bed. A d .M u. Mag. 5: 152-58.
Mem. Qld Mu. 12: 96100.
Aoki, R., 1988. Protruded toothed fossil crocodilian from
the Solomon Islands - Gouinlh paptanus (sic). Kiiyoh to Jack, R. L. and Etheridge, R., 1892. Thegeologynndpalaeon-
Seibulu (Agubiology) 10: 60. tology of QwecmIand and New Guinea. Government
Printer: Brisbane.
Archer, M., Hand, S. J. and Godthelp, H., 1993. Early -
Eocene marsupial from Australia. Kaupio Donnrk?dtw Laneston. W.. 1965. Fossil crocodilians from Colombia and
Bdriige iur Nalurgeschrchfe 3: 193-200. the Cenozoic hisrory of the crocodile in South America.
Uni. Calfornia p1161. Geol Sci. 52: l57Pp.
Archer, M., Hand, S. J. and Godthelp. H., 1994. Tertiary
environmental and biotic change in Australia. In Longman, H. A., 1924. Some Queensland fossil vertebrates.
En~~mmalrahlo nge and molulton ed by E. Vrba. Yale Afm. Qld Mw. 8: 1624.
University Press: New Haven.
Longman, H. A,, 1925a. A crocodilian fossil from
Archer, M. and Wade, M., 1976. The results of the Ray E. Lansdowne Station. Mm. Qld Mw. 8: 103-08.
Lemley Expedition Part 1. The Allingham Formation
and a new Pliocene vertebrate fauna from northern Longman, H. A., 1925b. A largejaw ofPdl&mrchw pollens.
Queensland. Mm. Qld Mu. 17: 379-97. Mem. Qld Ma. 9: 158-59.
Balouet. 1. C.. 1991. The fossil vertebrate record of New Lydekker, R., 1888. Cornlop of the Fosil Replilia and
Amphibin m 1h.e Brinrh Museum (Natural Hirtory). Part 1.
Trustees of the British Museum (Natural History):
London.
296 Australian Zoologist 30(3)