50.^ -I ^r^ftj^ THE HAMILTON FAUNA A LATE PLIOCENE MAMMALIAN FAUNA FROM THE GRANGE BURN, VICTORIA, AUSTRALIA WILLIAM D. TURNBULL AND ERNEST L. LUNDELIUS, Jr. ft* FIELDIANA: GEOLOGY VOLUME 19 Published by FIELD MUSEUM OF NATURAL HISTORY AUGUST 28. 1970 The Libraiy of tbe MAY 1 5 1972 Univwwty of HtlnoJs GEOLOGY LIBRAKX at Urbana^hampiign .\6t^ FIELDIANA: GEOLOGY A continuation of the GEOLOGICAL SERIES of FIELD MUSEUM OF NATURAL HISTORY VOLUME 19 FIELD MUSEUM OF NATURAL HISTORY CHICAGO, U.S.A. 1970 THE HAMILTON FAUNA A LATE PLIOCENE MAMMALIAN FAUNA FROM THE GRANGE BURN, VICTORIA, AUSTRALIA WILLIAM D. TURNBULL Associate Curator, Fossil Mammals, Field Mziseum of Natural History, Chicago AND ERNEST L. LUNDELIUS, Jr. Professor, Department of Geological Sciences, University of Texas at Austin FIELDIANA: GEOLOGY VOLUME 19 Published by FIELD MUSEUM OF NATURAL HISTORY AUGUST 28, 1970 PUBLICATION 1105 Library of Congress Catalog Card Number: 76-129^6^. PRINTED IN THE UNITED STATES OF AMERICA BY FIELD MUSEUM PRESS TABLE OF CONTENTS illus- text tration Introduction 5 Systematic Descriptions Metatheria: Marsupialia Perameloidea: Peramelidae Genus and species indet 11 88 Phalangeroidea: Phalangeridae: Phalangerinae Phalanger cf. gymnotis . 13 90 Trichosurus sp. indet 14 90 ? P/iaianger, Tnc/josMfMS, or unnamed related genus, sp. indet. . 15 90 Incertae sedis: near Phalanger or Trichosurus 19 Incertae sedis (minute form) 20 98 Phalangeroidea: Phalangeridae: Burramyinae Burramys sp 22 98 Phalangeroidea: Phascolarctidae: Pseudocheirinae Pseudokoala erlita new genus and species 26 100 Pseudocheirus stirtoni new species 34 104 Pseudocheirus marshalli new species 40 116 Pseudocheirus sp 44 Phalangeroidea: Macropodidae: Potoroinae Genus and sp. indet., near Aepyprymnus 45 122 Genus and sp. indet., near Hypsiprymnodon 46 122 Sp. indet., with similarity to Potorous and Propleopus ... 47 Dorcopsis sp. indet 47 124 Phalangeroidea: Macropodidae: Macropodinae Thylogale sp. indet 53 132 Incertae sedis: large-sized forms, two taxa near Protemnodon . 63 142 Incertae sedis: small-sized forms, near Dorcopsis and Thylogale .66 136 Phalangeroidea: Diprotodontidae Palorchestes cf. painei 66 146 Eutheria: Chiroptera Microchiroptera: /ncertae sedis at all lower taxonomic levels . 69 144 Class indeterminate, probably Mammalia Incertae sedis at all lower taxonomic levels 70 148 Interpretation of the Hamilton Fauna 74 Summary and General Conclusions 79 Acknowledgements 80 References 81 Appendix Plates 87 Graphs 151 3 INTRODUCTION Australian Tertiary mammals were extremely rare prior to the work of Stirton and his colleagues during the last 20 years. The few specimens known previously were noted and reviewed by Gill in a series of papers between 1952 and 1957. In the last of this series he assessed their significance, stratigraphic relationships, and affin- ities, and he suggested a few other possible Tertiary marsupial locali- ties in Victoria and Tasmania (Gill, 1957). As a result of the efforts of Stirton's group, this meager knowl- edge has been added to, and now a tentative outline of Australian mammalian evolution during middle and late Tertiary time is begin- ning to emerge. Thanks to their concurrent stratigraphic work, the relative geochronologic positions of the various faunae are known (Stirton et al., 1967a, 1968). To date only the New Guinean, Awe fauna, and the Victorian, Hamilton fauna (described here) have been dated isotopically, and thus they will provide a key to late Tertiary faunal correlations in the Australian region as well as give a chrono- logic correlation with Tertiary faunas in other parts of the world. Previous work. — The first find of a fossil marsupial from the Ter- tiary terrestrial deposits of the Grange Bum' was the discovery of a single tooth of a potoroine reported by Gill in 1952 and 1953a, and subsequently. 2 It was first identified as that of a cuscus (Gill, 1957 and earlier; Stirton, 1957a). Ride (1964) restudied and re-illustrated the specimen, concluding that it was either a right M^j or M^ of a potoroo of nearly the size of Propleopus oscillans, but having a mor- phology closest to Potorous gilberti. This tooth was found in the "A" zone of a fossil soil which underlies a basalt flow. Preliminary and progress reports of our work on the fauna from this level appeared subsequently (Tumbull et al., 1965; Lundelius and Turnbull, 1967). 1 The macropodid jaw from Forsyth's Bank reported by Colliver (1933), Single- ton (1935), and Gill (1957 and earlier), and described as a Sthenurine by Stirton (1957) came from shallow water marine beds, stratigraphically beneath the terres- trial deposits. Tedford (1966, p. 57) cast doubt on its unequivocal reference to the Sthenurinae. It does not enter into this report. » Gill 1953 b.c; 1955; 1957; 1965 a,b. FIELDIANA: GEOLOGY, VOLUME 19 Fig. 1. Map of the Hamilton area in western Victoria showing the location on the Grange Burn of the Pliocene (Hamilton Fauna) fossil locality. The reported occurrence of Tertiary macropodid remains from fissure fillings in the Bochara limestone of the same area (Glaessner et al., 1960) has been found to be Pleistocene (Glaessner, 1963). Locality. — The fossils were collected from a deposit at a waterfall on Grange Burn on the George Pelchen property, 4.5 miles west of Hamilton, Victoria (Fig. 1). This is locality no. 10 of Gill (1957). Stratigraphy and Age. — The stratigraphy of the Tertiary rocks of the Hamilton area has been described by a number of people. Gill (1957) has reviewed the literature, has briefly summarized the Ter- tiary stratigraphy of the area, and has shown the critical outcrops on his map. The Tertiary sediments exposed in the natural drainages west of Hamilton consist of marine, shore, and terrestrial deposits. The marine beds are shallow-water limestones, shales, and marls of Mio- cene (Batesfordian and Balcombian) and Pliocene (Kalimnan) ages. These grade upward into shore facies sands with shell fragment lenses which are overlain by continental sandy fossil soils and lacustrine TURNBULL AND LUNDELIUS: HAMILTON FAUNA 7 clay and sand deposits. These have been considered to be of Pho- cene (Kalimnan?) age. The sequence is capped by a basalt. The following sequence is exposed at the fossil locality on the Grange Burn: 1. Basaltic lava flow. At this outcrop the basalt is vesicular at the bottom, denser in the higher portion. Irregular, vertical struc- tures can be seen which appear to be gas bubble tracks. The lower surface of the basalt is irregular with charred tree stumps projecting into it. The sample for the potassium-argon date reported earlier (Turnbull et al., 1965) was taken five feet above the bottom of the basalt in the denser part 6 feet to eroded surface. 2. Tight, non-calcareous silty sand, grey when weathered, blue when fresh. Contains carbonized tree stumps and roots in the living position that show various degrees of charring. The upper two or three inches, which are dark brown to black in color, contain a very large percentage of carbonaceous material. There are many elongate patches of friable, grey to white sand and silt with finely divided car- bonized plant material. These are oriented from vertical to horizon- tal and show branching. They appear to be root channel fillings. The sieve concentrates from which the specimens were picked also contain numerous limonitic and caliche fragments, and basalt chips (which spalled off, or were knocked off from the overlying basalt in procuring the matrix sample). Rarer residues are reddish (contact metamorphosed ?) pieces of clay (seen also on large overturned basalt blocks that had slumped into the streambed), buckshot (goethite magnetite), rounded quartz pebbles, and occasionally fine, hackly, crystalline fragments (some fresh-looking, others weathered) stained with limonite. These contain mainly quartz, feldspar, and mica, and thus appear to be washed in bits of granitic rock. The vast bulk of the sediment is in the clay, silt, and fine sand particle size range. All the fossils discussed here were collected from this unit which ap- pears to be the "A" zone of a soil 1-1.25 feet. 3. Tight, calcareous, grey-gi^een to yellow-green sandy silt con- taining carbonized plant roots, many of which are oriented vertically and appear to be in the living position. There are also numerous cal- careous nodules mainly of two sorts. Most of those in the upper part are about Yo inch in diameter, red to yellow, dense and hard with ferruginous centers. Some of the nodules in the upper part and all in the lower part of this unit are calcareous, have no ferruginous centers, and are less dense than those from the upper part of the unit. 8 FIELDIANA: GEOLOGY, VOLUME 19 Many of the lower nodules are elongate, oriented vertically, and ap- pear to be cemented root channel fillings. A few deeply weathered marine mollusc shell fragments are recognizable in the lower part. This unit is believed to be the "B" zone of a fossil soil. 3) 2 feet exposed. We tentatively accept Gill's soil zone interpretation for units 2 and 3, at least until detailed petrologic analyses of the sediments are available. This soil profile is developed upon and immediately over- lies a calcareous near-shore marine sandstone, the Grange Burn co- quina, of lower Pliocene (Kalimnan) age (Gill, 1957). The stratigraphy indicates a post early Pliocene age for the fossils collected from the "A" zone of the soil. This is confirmed by the K/Ar date of 4.35 ±0.1 m.y. reported by Turnbull et al. (1965). This date is late Pliocene according to the time scale of Evernden et al. (1964). The contemporaneity of the marsupial fauna and the fossil soil with the overlying basalt is indicated by the presence of carbonized tree stumps and roots in the living position in the upper two to three feet of the fossil soil (Fig. 2) . Apparently the lava flow cremated a standing forest. It is unlikely that the fossils are appre- ciably older than the flow because it is doubtful that they would have survived the weathering processes in the soil for any length of time. Methods. — Matrix was removed from the "A" zone, dried, and wet-sieved in the manner described by Hibbard (1949). All residues above 30-mesh were saved and sorted for bones and teeth. The "B" zone was not sampled because of lack of time and the belief that it was likely to be less productive than the "A" zone. No attempt was made to subdivide the "A" zone. A total of three tons of "A" zone matrix was removed and wet-sieved. This produced approximately 500 pounds of size-graded concentrates which has yielded approxi- mately 150 identifiable teeth or significant tooth fragments^ repre- senting 18 mammalian taxa. The material recovered consists almost entirely of isolated teeth and fragments of teeth. Rarely some bone is preserved, mostly in poor condition, but occasionally an unweathered fragment was re- covered. No associations were noted in the process of collecting, but some were made later on the basis of the interdental facets. Many teeth were reassembled from fragments and many teeth are unworn. * Subsequently an additional eight tons of A and B zone matrix were taken, but except for seven specimens (PM 16801-PM 16802; NMV-P 26419-P 26420 and P 26422; and PM 16805 and PM 16807) picked from the sieves at the time of processing this sediment, none of this material is available at this time. ooooooooooooooooooooooooooooooooo • ooooooooooooooooooooooooooooooooo ooooooooooooooooooooooooooooooooo lOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOOOOOOO Q^ o o o o o o o 'Ooooooo BASALT oooooo o/^i^D o o o o o o c OOOOOOOO OOOOOOOOOOOOOO O J^;''^ji^^ o o o o o 'OOOOOOO OOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOOOO Fig. 2. Close-up view of the contact between the Pliocene fossil soil and the overlying basalt with schematic interpretation. Roots or branches, and the bolus of a tree stump, apparently in the living position, may be seen. Scale is approxi- mately X Ve. 10 FIELDIANA: GEOLOGY, VOLUME 19 The teeth exhibit various degrees of weathering which takes the form of solution-pitting on the enamel and often the removal of the roots, the dentine of the crown, or both. Some teeth show no sign of pitting and have the roots well preserved. Others consist only of the solution-pitted enamel caps and the adjacent dentine. This vari- ation in degree of weathering is believed to be a function of the length of time they were subjected to the weathering processes before being sealed off by the lava flow. None of the teeth shows any signs of abrasion such as rounding of the cusps, etc., which would be expected if they had been transported any distance to the site of deposition. The lack of associations of teeth in jaws has created a problem of designation of types for those species which we believe are represented by more than one tooth. In these cases, a single tooth has been desig- nated the type. The only exceptions are those in which a positive association can be made on the basis of well-defined interdental facets. Where the material is too fragmentary to allow the assign- ment of a generic or specific name, the fossils are described, figured and compared to Recent and fossil forms. Abbreviations.— Numbers with the following prefixes denote speci- ments catalogued in the collections of the institutions indicated. PM — Fossil Mammal Collection, Field Museum, Chicago. NMV-P — Paleontological Collection, National Museum of Vic- toria, Melbourne. FM^ — Recent Mammal Collection, Field Museum, Chicago. UT-M— University of Texas, Vertebrate Paleontology compara- tive collection of Recent mammals. AMNH — Recent Mammal Collection, American Museum of Nat- ural History, New York. UCMP — University of Caifornia, Museum of Paleontology. USNM — Recent Mammal Collection, U. S. National Museum, Washington. WAM — Western Australian Museum, Perth. SYSTEMATIC DESCRIPTIONS Class Mammalia Infraclass Metatheria Order Marsupialia Superfamily Perameloidea Family Peramelidae Genus and Species indet. Material. — The anterior half of a right lower molar, probably M^ or M7 (PM 4499) ; part of a trigonid of a right lower molar (NMV-P 26406) ; fragment of a left lower molar consisting of the lingual side of the trigonid and anterior cingulum (PM 4446) ; meta- conid fragment of a right lower molar (PM 4600) ; backside of trigo- nid of a right lower molar (PM 4748). This material is inadequate for either generic or specific assignment. Plate I, Figures A-F. Description.— Most of the description and comparison is based on the best specimen (PM 4499) ; the others conform closely to it except as noted. The following features combine to identify these teeth as belonging to the Permalidae rather than to the Dasyuridae, the only other group with which they might be confused. 1. The anterior cingulum is large and rounded with a shallow but distinct basin. A large anterior cingulum is present in M^_4 of peramelids.^ This structure is generally much less pronounced in the dasyurids where it is almost straight and is closely appressed to the antero-labial surface of the tooth. 2. The protoconid-paraconid and protoconid-metaconid crests lack the sharp, deep clefts such as are found in the bottoms of the "V"s of these crests in the teeth of dasyurids. In this they agree with the known peramelids. 3. The valley between the paraconid and metaconid is much nar- rower than in the dasyurids, and is more like that of the peramelids. 'The following dasyurids and peramelids were studied: Sminthopsis (several species), Phascogale tapoatafa, P. calura, Dasycercus blythi, Parantechinus apicalis, Dasyurus geoffroyi and Dasyurus (Satanellus) hallucatus, Sarcophiltis harrisi, S. laniaris, Thylacinus cynocephalus, Glaucodon ballaratensis, Isoodon, Macrotis, Perameles, Echimypera, Peroryctes and Ischnodon (M^ only, other molars unknown). 11 12 FIELDIANA: GEOLOGY, VOLUME 19 4. The hypoconid ridge (PL I, A-C, arrow h) of PM 4499 joins talonid and trigonid at the middle of the back of the trigonid. This ridge is somewhat variable in its development and morphology in both peramelids and dasyurids, but the usual pattern is clear for both families. In those peramelids with molars in which this ridge joins the trigonid (i.e., most teeth of most of the living species except M^ of Echimypera doreyana) the junction is usually located either in this region or more lingually. In the small dasyurids this junction is lo- cated near the labial side on My-^ (M4S have talonid lingually placed and very reduced in size). In Isoodon ohesulus the junction of this ridge and the trigonid shows the dasyurid pattern; hence it is an exception. The anterior cingulum is large and rounded with a shallow basin very much like that of Echimypera doreyana (FM -54233). It arises on the antero-labial surface of the protoconid, extends antero-lin- gually, and meets the paraconid near its base by turning abruptly posteriorly at the antero-internal corner of the tooth. A tiny cuspule is located on the cingulum just anterior to the paraconid. The cingu- lum does not approach the apex of the paraconid, as in Isoodon, but remains well below it, as in Echimypera, Perameles, and Peroryctes. The antero-lingual end of the cingulum is in line with the paraconid and metaconid. This condition is found in Peroryctes longicaudata and in Mt-tj oi Peroryctes broadbenti and is approached in Parameles gunnii and Echimypera doreyana. The paraconid and metaconid of the Hamilton peramelids are rel- atively farther apart than in living peramelids, except Peroryctes. The protoconid is the largest cusp. It is followed in descending order by the metaconid and paraconid, as in most other peramelids except Macrotis lagotis (which has an enlarged metaconid) and Choeropus (in which the paraconid and metaconid are equal in size and are very close together) . The tooth is less hypsodont than are any of those of living peramelids except Peroryctes. There is only the beginning of cusp hypsodonty (see White, 1959), such as is seen in Perameles. The postero-lingual border of the metaconid has a ridge running from the apex to the base. The lower molars of Peroryctes longicau- data, P. broadbenti, and Echimypera doreyana have a similar ridge which joins the ridge that runs to the entoconid. The ridge on the metaconid is absent in other peramelids. The posterior bolder of the trigonid of PM 4499 is slightly concave (more so than in Peroryctes longicaudata and My_^ of Echimypera doreyana) . TURNBULL AND LUNDELIUS: HAMILTON FAUNA 13 There is a saddle-shaped connection between the remnants of the hypoconid ridge and the metaconid-entoconid ridge in PM 4499. This implies the existence of a basined talonid. In NMV-P26406 the break at the back of the trigonid suggests that the talonid basin, if present, did not extend forward to meet the trigonid for there is no indication of a hypoconid ridge (see PI. I, Fig. F). Instead, the surface is convex and slopes rapidly labially. It is characteristic of the M4S of most peramelids that they lack a talonid basin. There is usually a ridge connecting the metaconid and entoconid. The posi- tion of NMV-P26406 within the molar series is uncertain: it is most likely an M^ but possibly it is an M^ similar to that of Echimypera doreyana (because of the great width of the talonid indicated by the break surface). Measurements. — PM4499 Length of trigonid (paraconid-metaconid length) 1.49 mm. Length of trigonid plus anterior cingulum 1.95 Maximum width of trigonid (parallel to its back edge) . . .1.87 NMV-P 26406 Maximum width of trigonid (parallel to its back edge) . . .1.83 Discussion. — It is impossible to assign these Hamilton specimens to a genus or species or even to be sure that they belong to the same genus or species. They resemble both Peroryctes and Echimypera more than any other peramelid in the spacing of the trigonid cusps, kind and degree of hypsodonty, and a large-basined anterior cingu- lum. The Hamilton specimens show roughly the same degree of divergence from the basic dasynrid lower molar pattern as do Pero- ryctes and Echimypera Superfamily Phalangeroidea Family Phalangeridae Subfamily Phalangerinae Phalanger Storr, 1780 Phalanger cf. gymnotis Material.- — Nearly complete crown of left P4 or right P^ (PM 4457, PI. II, Fig. C). Description. — The tooth is a high, narrow, flat-sided blade with a rounded crest and at least four transverse ridges. In crown view the 14 FIELDIANA: GEOLOGY, VOLUME 19 antero-labial surface is convex and set off by ridges. The anterior ridge extends from the axial crest antero-Hngually toward the base of the crown. The ridge separating the antero-labial surface from the labial surface extends in a weakly sinusoidal curve to the base of the crown. This ridge is continuous with a similar ridge on the lin- gual side of the tooth. The other three ridges do not extend more than 25 per cent of the distance to the base of the crown. They are approximately equally spaced and have distinctive, bluntly pointed apices. The middle two cross ridges are the same size and are slightly larger than the anterior and posterior ridges. All transverse ridges retain their identity where they are folded over the axial crest, rather than blending into it, as is the case in Phalanger orientalis and P. maculatus and in both species of Trichosurus. The axial crest of the tooth reaches the same level between the four transverse ridges. Be- hind the posterior ridge it descends sharply to a broken surface so as to make it appear probable that the greater part of the crown is pre- served. The best comparison is with the fourth premolars of Phalan- ger gymnotis (FM 31861) which show no major differences from this fossil. Discussion. — See p. 20. Trichosurus Lesson, 1828 Trichosurus sp. indet. Materials.— Left P (PM 4563, PI. Ill, Figs. D and E) ; two par- tial premolars, either upper right or lower left preserving the anterior one-third of the tooth (PM 4556 and PM 4557, PI. II, Figs. D and B, respectively) . Descriptions.- — The P is represented by one unworn tooth (PM 4563). The occlusal surface consists of a sharp edge which extends straight to the posterior corner, then upward along the posterior side, and gradually loses its sharp-edged character. At its anterior end this sharp edge merges with ridges which extend short distances above the occlusal surface on the outside and inside of the tooth. The outside one is stronger. The labial anterior ridge extends up- ward and posteriorly, and joins one extending upward from a cuspule located at about the midpoint of the occlusal surface. These two ridges form a shallow triangular-shaped basin on the labial side of the tooth. In the Recent Trichosurus vulpecula wear rapidly elimi- nates this basin. The posterior part of the labial side of the tooth is almost flat with only a trace of the sulcus seen in the Recent species. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 15 The blade-like occlusal surface is turned lingually. The anterior sur- face of the tooth is broad and gently concave. The central cuspule on the occlusal surface also gives off a low ridge on the lingual surface. This ridge extends anteriorly and upward and disappears. There is little or no sign of such a ridge in the Recent form. The fos- sil tooth is the same size as its counterpart in the Recent species of Trichosurus and is markedly larger than the reduced I^ of Phalanger. The two premolar fragments differ only slightly from one another. They are at the lower end of the size range for modern T. vulpecula. Both have the typical three anterior ridges which extend toward the tooth base as in T. vulpecula, the central ridge being the largest and extending nearly in the axial plane. This ridge lacks the sharpness of its counterpart in the Recent form. The labial ridge extends from the apex of the crown to the base in a nearly straight line, rather than being markedly curved in a posterior direction. The lingual ridge is more pronounced in the fossil than in the Recent species of Tricho- surus and it is better developed at the apex in PM 4557 than in PM 4556. It extends to the tooth base in a straight line in the for- mer. In PM 4556 this ridge curves posteriorly at the base of the crown. The antero-labial surface in both teeth is convex except for a slight depression near the apex. The antero-lingual surface is slightly narrower than the antero-labial sui'face and is weakly con- cave in PM 4556. In PM 4557 it is concave at the apex, but flat at the base. These teeth are markedly different from the premolars of Phalanger orientalis in which the antero-lingual ridge is virtually absent. Discussion. — See p. 20. ? Phalanger or Trichosurus or an unnamed, related genus Sp. indet. The collection contains five specimens that appear to represent the sarae species. These teeth share certain features typical of vari- ous species of Phalanger, and other features of species of Trichosurus. We are convinced that they represent a new species, but cannot make a generic assignment because of inadequate material. This species may belong to either of the above-named genera, or to an as yet un- named intermediate genus. Materials.— A right M-l (PM 4571, PI. IV); a left upper molar fragment (PM 4728) ; a molar fragment, probably part of a right upper, with part of posterior cingulum, hypocone, metacone, and the 16 FIELDIANA: GEOLOGY, VOLUME 19 metaloph preserved (PM 4735) ; a right My (PM 4555, PI. Ill, Figs. A-C); and a right M^ (NMV-P 26407, PI. V). Descriptions. — The M- (PM 4571) is very similar to that of Tri- chosurus vulpecula and T. caninus (USNM 221121) but differs in having a well-developed parastyle which is well separated from the paracone (PI. IV). This cusp (parastyle) is variable in modern T. vulpecula but does not attain the size or distinctness seen in the fossil specimen. The condition is approached quite closely, however, in a Western Australian specimen, T. vulpecula (UT-M-850), in a juvenile from Victoria (FM-98897), and in a specimen of Phalanger orientalis (FM-54021), all of which have a good parastyle. A small cingular cuspule (mesostyle?) is present on the labial side between the metacone and paracone. This cuspule is variable in the Recent T. vulpecula but is quite like that of the fossil in UT-M-850. The ridges which connect protocone and hypocone form a V-shaped junc- tion in the fossil, a U-shaped one in Recent T. vulpecula from Western Australia, and a U-shaped, but narrower one in the living eastern form. In size the tooth falls at the low end of the range when com- pared with 23 specimens from widely separated localities in Australia (from Northeast Cape, the Southwest and Nullarbor Plain in West- ern Australia; and from South Australia, Victoria, Tasmania, and Northern Queensland). The M- is narrower (absolutely and rela- tively) than in any of the specimens available for comparison. The unworn enamel surfaces are more irregular than in the Recent species but are not crinkled as in Phalanger. A left upper molar, probably an M^ (or M-), is represented by a slightly worn paracone (PM 4728). It is a nearly square fragment, with its most pronounced feature being the anterior cross-loph (pro- toloph). This loph runs horizontally almost straight across the frag- ment, when viewed occlusally in normal orientation. The anterior edge of the procingulum is broken away, but the valley behind the cingulum is well delimited. The sloping anterior and posterior faces of the protoloph are nearly equal, weakly crenulated, and are ori- ented at 90° to one another, each at 45° to the occlusal plane. The labial side of the cusp meets the gently saddled protoloph at the apex of the paracone at an angle of about 90°. It is a flat, gently sloping, plane surface. Crests run forward and backward from the apex of the paracone at the junction of this surface with the inclined faces of the protoloph. The anterior of these crests leads to the anterior cingulum. The posterior leads down toward the transverse valley TURNBULL AND LUNDELIUS: HAMILTON FAUNA 17 where it bends sharply lingually just in front of the deepest part of the valley. A trace of a parastyle is present on the forward crest near the antero-labial corner of the tooth. Wear marks the crest of the protoloph, a distinct anterior facet, and a broad region of the posterior-facing surface of the protoloph. The fragment resembles Phalanger teeth in the subequal proportions of the anterior and pos- terior faces of the protoloph, but is more like Trichosurus teeth in the form and character of its enamel surface sculpture. The right metaloph (PM 4735) has the enamel surface pitted to a greater extent than in T. vulpecula but not so deeply as in Phalanger maculatus or P. orientalis. The metaloph ridge resembles the proto- loph of PM 4728. The anterior face of the metaloph is slightly larger than the posterior one. The difference is less than that seen in upper molars of T. vulpecula but more than in either P. orientalis or P. maculatus. The My (PM 4555), a right, differs from that of most T. vulpecula in a number of ways, from T. caninus in other ways, and from Phalanger in still other ways, but at the same time shows many simi- larities to each of these. The tooth is both laterally compressed and tapered in its anterior half, but is broad and square in outline poste- iorly (PI. Ill, A-C) . The lingual side of the anterior part of the tooth is strongly concave between the anterior ridge of the protoconid and the metaconid. In the middle of the concavity is a small pillar-like, rounded crenulation. It starts near the base and fades out upward. This is absent in the Recent species of Trichosurus, but a similar or more pronounced ridge is preserved in Phalanger orientalis and P. maculatus. The protoconid and metaconid are distinct and subequal, but are close together as in Phalanger. They are at the same height and are joined by a short ridge which has no depression along it. These cusps are more widely separated in all specimens of Tricho- surus available for comparison. T. vulpecula has metaconid relatively smaller than protoconid. The labial side of the protoconid slopes steeply as does the hypoconid with no sign of a cingular bulge as in T. vulpecula. (The base is broken and its absence may be due to this.) A ridge extends downward and posteriorly from the metaconid to the transverse valley, and disappears as it turns toward the center of the tooth. It does not join the anterior ridge from the entoconid. A U-shaped notch is formed as in NMV-P 26407. Also, as in NMV- P 26407, the ridges joining protoconid and hypoconid meet to form a "V" rather than a flat-bottomed "U." The posterior crest of the protoconid is slightly convex labially and the anterior crest of the 18 FIELDIANA: GEOLOGY, VOLUME 19 hypoconid is straight. In the Recent Trichosurus vulpecula the pos- terior crest of the protoconid is concave labially and the anterior crest of the hypoconid is either curved or sharply angled. This difference results in a relatively wider central basin in the Hamilton fossil. The metaloph is broad and continuous, though it is depressed in the mid- dle. The slightly worn enamel is ornamented with irregular rounded pits and ridges as in NMV-P 26407. Very small wear facets appear on some of the crenulations and also on the center of posterior cingulum. The M^ (NMV-P 26407, PI. V) shows no significant difference in proportions from the Recent Trichosurus vulpecula. The posterior and anterior ridges from metaconid and entoconid are well defined as in the eastern Recent form and differ from the Western Australian form in which these ridges are much reduced. The posterior ridge of the metaconid and the anterior ridge of the entoconid do not meet but turn toward the center of the tooth leaving a U-shaped channel- way that becomes rapidly V-shaped as it extends laterally into the central basin of the tooth. This character is variable in the Recent form. It does not appear on any of the University of Texas speci- mens, but wear makes interpretation difficult in most of them. There is no corresponding open channel between the protoconid and hypo- conid. Instead, the posterior ridge of the protoconid extends diag- onally straight toward the entoconid and is joined in the transverse valley by a recurved ridge that runs forward from the hypoconid. These ridges form a V-shaped junction in the fossil, and a flat-bot- tomed U-shaped junction in the Recent T. vulpecula. The cross- lophs are continuous and not broken by clefts. The profiles of the crests are more or less U-shaped although the anterior one is (weakly) V-shaped. These features are variable in the Recent form. A low but distinct cusp is located on the labial side of the anterior cingulum. It is only tenuously joined to the protoconid by the outer part of the cingulum where this structure crosses the valley between the two cusps. In the posterior molars of most specimens of the Recent spe- cies the cingular cusp is also joined to the protoconid by a broad, low ridge which divides the anterior cingular basin into two, a larger inner one and small but deep outer one. A posterior cingulum is present which forms a deep, transversely elongated posterior cingular basin. The posterior cingulum of Trichosurus M^s is highly variable. In some it is pronounced and delimits a well-defined basin, in others it is incomplete lingually so that the basin opens behind the entoconid, and in others it is both incomplete and much shortened. In Phalanger TURNBULL AND LUNDELIUS: HAMILTON FAUNA 19 the posterior cingular basin tends to be open and broad. It is rela- tively shallow and the cingulum is low. The antero-labial corner of the tooth has a very shallow indentation. In Phalanger and most specimens of Trichosurus vulpecula this is a small but definite pit formed by the backward and downward extension of the anterior cingulum. The unworn enamel surfaces are more uneven but do not approach the condition in Phalanger. Measurements. — PM 4571, a right M^ Length 5.16 mm. Anterior width 3,49 Posterior width 3.49 PM 4555, a right My Length (slightly broken) 5.80 Anterior width . . (too broken to measure) Posterior width 3.70 NMV-P26407, a right M^r Length 5.07 Anterior width 3.68 Posterior width 3.71 Discussion. — See p. 20. Incertae sedis: near Phalanger or Trichosurus Sp. indet. In addition to the teeth of large phalangerines described above, there are five tooth fragments that we can only tentatively and dubi- ously place in this grouping. They are PM 4721, PM 4722, PM 4723, PM 4724, and PM 4726. Of these only PM 4722 is illustrated (PI. II, Fig. A). It appears to be the end of the left? upper canine. It very closely resembles the end of the canine of Recent Trichosurus vulpecula from eastern Australia (FM 60935). The cross-section is roughly that of a diamond, with the long axis antero-posterior. The posterior ridge of the Recent form has a small to minute subsidiary cuspule. In the fossil this cuspule is minute and is very closely ap- pressed to the main cusp. It differs from Phalanger which has conical recurved canines. It differs from the potoroines (Bettongia, Potorous, Hypsiprymnodon) which have the upper canine strongly compressed laterally. The broken posterior surface leaves interpretation of this fragment as a canine in some doubt. It shows an inverted "V" shape that can be variously interpreted as a diagonal break along the back of the canine, or as a vertical break running transversely across a small blade-like tooth. 20 FIELDIANA: GEOLOGY, VOLUME 19 Discussion of the large phalangerines. — There seems to be no doubt that at least two taxa are represented by the materials described on pages 13 to 19. It also seems certain on the basis of the P|s that both Phalanger and Trichosurus are represented. The P|s of P. gym- notis are very distinctive, and the close resemblance of the Hamilton specimen to them leaves little doubt as to its relationship. The Tri- chosurus-Mke P|s are evidence that this genus is represented, but probably by a species different from any presently known. This is supported by the I- which is virtually identical to that of the living T. vulpecula. The molars all share various morphologic features of the genera Phalanger and Trichosurus. In the case of the V-shaped junction of protocone-hypocone ridges and protoconid-hypoconid ridges (as well as the near proximity of protoconid and metaconid in My) they are closest to Phalanger. Regarding features such as depth of posterior cingular valley and height of the posterior cingulum, they appear to be closest to some specimens of Trichosurus. In the case of the de- gree of pitting and crenulation of the enamel surface, they are quite intermediate between the two genera. Wyulda squamicaudata (WAM- B1920) was also compared, and it agrees with Trichosurus in this regard. Both Trichosurus and Phalanger have species that show tendencies to bridge the gap. Therefore, in dealing with the large phalangerine teeth, we have no way of telling with certainty which of the front teeth (if any) go with the molars. There are differences between the fossil Hamilton molars and those of our comparative materials of P. gymnotis (FM 31861), while the P|s correspond very closely to those of the modern species. The differences between the Hamilton molars and a broad suite of T. vulpecula is, if anything, even greater, so that their possible correlation with the preserved anterior teeth assigned to Trichosurus sp. indet., seems even less likely. Hence we know that the fauna certainly contains two, and may contain three related, large phalan- gerines. Incertae sedis (minute form) Genus and sp. indet. Material and Description.- — A minute phalangerine approximately the size of Acrohates, Distoechoerus, or Cercartetus, is represented in the Hamilton fauna by two left lower incisors NMV-P26408 and PM 4553 (PI. VI, Fig. B) which are very similar to one another in TURNBULL AND LUNDELIUS: HAMILTON FAUNA 21 almost every detail. We cannot assign them positively to any genus but comparisons with these three are closer than with other genera of minute-to-small phalangerines. The fossils have a flatter, less arched, and more spatulate upper surface than do the lyS of Petaurus (FM 56318) which are also much longer. There is not as well defined an outer ridge as in Petaurus. The fossil lyS are more like those of Cercartetus concinnus (FM 34721 and UT-M 831, 832, and 842) or Acrobates pygmaeus (FM 69898) especially as to size. They have a better defined internal ridge than does C. concinnus, and are more curved and more evenly curved longitudinally in the flattened plane than in C. concinnus. The medial surface shows a trace of a shallow longitudinal trough (more pronounced in NMV-P26408 than in PM 4553), which appar- ently is readily erased by interdental wear. A similar situation exists for C. concinnus. The fossil teeth have a lateral longitudinal ridge as in C. concinnus, but it is set off from the bulk of the tooth by a narrow groove. The tips of the fossil teeth are more rounded and are less symmetrical than in C. concinnus. The Hamilton specimens are slightly less curved and are less tapered when viewed from above (in the horizontal plane) than are the Acrobates pygmaeus specimens available for comparison (FM 60898 and Monash University ,no. 974). A. pygmaeus has a better developed external ridge than does either of the fossils or C. concin- nus. In A. pygmaeus near the back edge of the crown, this ridge swings across the dorsal surface of the tooth and becomes continuous with an equally developed internal ridge. In C concinnus the com- parable ridges die out posteriorly. The posterior ends of the Hamil- ton specimens are broken, and the course of the ridges cannot be determined. In comparison with Distoechoerus pennatus (AMNH-104058) in which the dorsal surface of the ly is convex, the fossil teeth are flatter, and they are also less pointed. They differ from Burr amy s in: 1. Cross-section. — Burramys has a very deep oval section with the long axis of the oval vertical. The Hamilton teeth are more flattened, and the long axis of the oval is more nearly horizontal. 2. Ridge development. — The Buchan Burramys available to us has none, but the tip of the tooth is broken, and B. parvus from Wombeyan Cave has them strongly developed, but only distally 22 FIELDIANA: GEOLOGY, VOLUME 19 on the tooth. The Hamilton fossils are ridged for most of the length of the crown. 3. Torsion. — The Hamilton teeth show torsion, Burramys lyS do not. In comparison with Tarsipes spenceri, the fossils differ markedly in curvature and taper, T. spenceri being straight, with a very gentle taper and with much less torsion. Also, quite unlike the Hamilton teeth, the crown in the lyS is expanded over the root in T. spenceri especially on the dorsal surface (UT-M8294 and FM 34719). Although the closest resemblance is to the lyS of Cercartetus con- cinnus and Acrobates pygmaeus, there are enough differences from both to preclude assignment of the fossils to either, and to suggest that they probably represent an unknown taxon. Subfamily Burramyinae Burramys Broom, 1895 Burramys sp. Material. — Left P^, posterior three-fourths of tooth (NMV- P26409, PL VI, Fig. C) ; right P^, anterior third of tooth (PM 4439, PI. VI, Fig. D) ; right P-*-, antero-internal corner of tooth (PM 4459, PI. VI, Fig. F) ; and left M^ a complete enamel cap (PM 4470, PI. VI, Fig. E) . This material is inadequate to serve as the basis for desig- nating a new species of Burramys, yet we believe that it represents a new form differing slightly from Burramys parvus from the Wom- beyan Caves and from the Buchan Burramys. Description. — The P^ is a high, laterally compressed, serrate, blade-like tooth. There are six grooves preserved on each side of the most complete specimen (NMV-P26409) . The anterior end of this specimen is broken away and it is not possible to be sure of the total number of grooves in the intact tooth. However, the crest of the tooth begins to curve downward at the break, and it is doubtful that there were more than six or seven grooves on the tooth. The P4 of the Buchan specimen has six well-defined grooves plus a minute one in front of these on the lingual surface, seven on the labial. In the unworn P4 of Burramys parvus from Wombeyan (PM 5936) there are seven distinct grooves on both sides, plus a minute antero- lingual one.i In a worn Wombeyan specimen (PM 16185) five 1 Best seen at a magnification of about X 50. This feature also is present on PM 16185. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 23 grooves and a trace of a sixth appear on the labial side, and six with a trace of a seventh may be seen on the lingual side. The antero-internal edge of the P^ of the Hamilton Burramys (PM 4439, PI. VI, Fig. D) has a ridge that is contiguous with the axial crest of the tooth. This ridge, here termed the antero-axial ridge, turns lingually, then somewhat posteriorly, as it descends in a smooth curve to the base of the tooth. It forms a curb which sharply delimits the convex anterior face from the ridged and grooved lingual side of the tooth. The antero-extemal margin has no such sharply defined line of demarcation separating anterior face from the slightly convex labial side with its ridges and grooves. There the anterior three or four sets of ridges and grooves fade out and merge with the bulging anterior face which is progressively swept back ventrally. The first ridge trails off at approximately one-third of the distance from crest to base of tooth. The subsequent ones are progressively longer. In the Buchan and Wombeyan specimens available for com- parison, the first ridge extends approximately one-half and two-thirds this distance, respectively, before disappearing. The anterior surface is more developed and more bulbous in the Hamilton form than in the others. The grooves on the lingual side of the tooth are slightly closer together toward the tooth base than those on the labial side. This is unlike both the Wombeyan and Buchan specimens in which the spac- ing of the grooves is the same on both sides. The first three gi'ooves on the lingual side merge toward the root to form a single broad open groove. The second ridge extends farther toward the root than the first. In the Buchan and Wombeyan specimens, only the first two of the lingual grooves merge, and the first ridge extends farther toward the root than in the Hamilton specimen. The cutting edge of the tooth (best seen in NMV-P26409) is straight with small cusps marking the junction of the ridges on either side of the tooth. The posterior end has a laterally compressed cusp which is two or three times the size of the more anterior cusps. In these two characters it differs from both the Buchan and Wombeyan forms in which the P4S have a crest that is arched and in which the posterior cusp is relatively smaller. The postero-lingual surface of the tooth has two very faint ridges which are only one-half the length of the others (not included in ridge count given above). They converge toward the crown, but do not meet. The anterior ridge of the pair is parallel to the more 24 FIELDIANA: GEOLOGY, VOLUME 19 anterior ridges. There seem to be no comparable ridges on the P^s of the Wombeyan and Buchan specimens. There are wear facets at the top of the outside of the last three ridges and the larger posterior cusp. The posterior end is concave. This concavity is bordered lingually by a ridge which extends down the posterior side of the tooth from the large posterior cusp. The PA is represented by a very small fragment, with a portion of the anterior edge and lingual side of the tooth (PM 4459, PI. VI, Fig. F) . The anterior edge has the same smooth convex surface and sharp, limiting antero-axial ridge as the P4. The antero-axial ridge is less pronounced than in the P^. In addition to the antero-axial ridge, five ridges and six grooves are present on the lingual side. The first two ridges merge with the antero-axial ridge, the first very close to the broken edge near the crest and apparently well above the root. The second extends farther toward the root before it merges with the anterior ridge. The third and fourth ridges disappear near the base of the tooth without joining either the anterior ridge or one another. In the specimens from Buchan and Wombeyan, the first ridge runs parallel and very close to the (weakly developed) antero-axial ridge for about one-half to two-thirds of its length before joining it. The second ridge is parallel to the first and disappears closer to the roots without joining any other ridge. The left M^ of the Hamilton Burramys is known from a single, very low crowned tooth (PM 4470, PI. VI, Fig. E). The tooth con- sists of an unworn enamel cap. It is not quite square in outline, the posterior width is less than the anterior width. Cusp size order is as follows: protocone > paracone '^ hypocone > metacone. Rounded ridges extend both anteriorly and posteriorly from paracone and metacone. The anterior one from the metacone and the posterior one from the paracone do not join, but end close together with a valley between them. Protocone and hypocone are blunt and more rounded than the other cusps. A low, rounded ridge extends from protocone toward paracone, but does not quite join the correspond- ing ridge from the paracone. A very narrow, tiny but distinct, groove located about in the axis of the tooth separates them. A continuous low broad ridge connects the metacone and hypocone. The region between the two transverse ridges is flat. There are anterior and posterior cingulae. The anterior one has a straight anterior border and is flat bottomed, not basined. The posterior cingulum is bulged posteriorly and is slightly basined. The anterior cingulum joins the TURNBULL AND LUNDELIUS: HAMILTON FAUNA 25 anterior crest of the paracone and forms a slightly differentiated cusp- ule (? "parastyle"). The Hamilton molar is quite similar to M^ of Burramys from Buchan and Wombeyan Caves, but differs in the following ways: 1. slightly smaller than the Buchan form (Wombeyan and Grange Burn forms about equal, see measurements), 2. absence of an external stylar cusp (cusp "b" of Bensley, 1903 and "B" of Simpson, 1929), 3. more rounded posterior margin, 4. relatively wider anterior cingulum, 5. lower crowned with lower cusps and ridges, and 6. less basined, more open central portion of tooth. It shows some similarity to M^ of Petaurus breviceps (FM 60939) but differs in the following ways: 1. anterior cingulum relatively wider, 2. transverse ridges show less slope toward center of tooth, 3. lower crowned and less sharply crested, 4. apparent absence of a lingual, ridged bulge on the protocone, such as is characteristic of Petaurus, 5. central part of tooth a broad, flat-bottomed transverse valley rather than a basin, and 6. anterior transverse ridge is broken by a small cleft. Measurements. — See Table 1. Table 1.— C omparison of M-s of various specimens (in millimeters) of Burramys PM 5934 PM 4470 Hamilton Wakefield Buchan Cave Wombeyan Cave Burramys parvus Length M^ 1.2 1.3 1.2 Anterior width 1.2 1.4 1.3 Posterior width 1.0 1.2 1.1 Discussion. — The Hamilton specimens have characters which in- dicate that they represent a species of Burramys distinct from B. par- vus from the Wombeyan Caves and the form from the Buchan Caves. Several of these characters, such as the large posterior cusp on P:^, straight crest of P^, low crown of M^, and absence of a distinct an- tero-external stylar cusp on M^, indicate that the Grange Bum form is more primitive than the other two. Consistent with this inter- pretation is its greater antiquity (Pliocene vs. Pleistocene age for both 26 FIELDIANA: GEOLOGY, VOLUME 19 the Buchan and Wombeyan forms). The characteristic anterior edges of P| may also represent a more primitive stage of blade de- velopment or, perhaps equally probable, a different specialization. Family Phascolarctidae Subfamily Pseudocheirinae Pseudokoalai new genus Diagnosis. — The same as for the species P. erlita until other spe- cies are discovered and a broadened diagnosis is required. Pseudokoala erlita- new species Holotype.— Two separated teeth, L.M^^ (NMV-P26399, PL VII, Figs. A-E and NMV-P26400, PI. VIII, Fig. A; PI. VII, Figs. C, E) which are shown by the interdental facets to have been adjacent teeth, constitute the type. Referred specimens.—A single L. My (PM 4588, PI. VIII, Fig. B). Diagnosis. — An animal with My and M^^, large (see Measure- ments), almost as long but narrower than in Phascolarctos cinereus, and very low crowned. Paracone and metacone of M- elongate an- tero-posteriorly as in Pseudocheirus. Internal ridges of these cusps smaller and weaker in their development than the ectoloph ridges, as in Pseudocheirus. Ectoloph stands nearly vertical, quite unlike that of phascolarctines and in contrast to all other pseudocheirines except the new and yet undescribed Kutjamarpu form.^ Ectoloph ridge forms a "W '-shape, whose angles in both crown and side view are 120° or greater; this is in contrast to other known phascolarctids which have angles approximating 90° or less. Associated with the ectoloph there is a small transverse ridge (modified ornamentation) in the mesostylar region, as in the known phascolarctines. Meso- style well developed unlike that of phascolarctines, and like that of pseudocheirines. Protocone and hypocone low, blunt, and weakly crescentic with broad and flat, nearly horizontal labial faces. This 1 In allusion to the size and similar superficial appearance of these teeth to those of the koala. * The species name is a South Australian aboriginal word meaning ancient. * A cast of this pseudocheirine (UCMP 71664) generously provided byM.O. Woodburne. Its ectoloph stands nearly as vertical as that of M^ of Pseudokoala erlita. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 27 combination of features is unique. My with trigonid more like that in pseudocheirines than that in Phascolarctos cinereus, i.e., paraconid is developed and distinct, and protoconid is connected by a sharp crest to the paraconid-metaconid crest near the front of the meta- conid somewhat as in Pseudocheirus (Pseudocheirops) archeri. My like that of Perikoala palankarinnica in having a well-developed para- conid, but differs in that it has a distinct metalophid which is lack- ing in P. palankarinnica. Ornamentation is coarse, rounded, anast- omosing, and virtually confined to the central basin in M-^^, and is more extensive in M^ than M- or My. NMV-P26399-L M^ PI. VII, Figs. A-E This tooth is more elongate than its homologue (or M-l or M-) in Phascolarctos, or the M^ of Perikoala or the M^ of Litokoala (Stir- ton, Tedford and Woodburne, 1967) . The hypocone is blunt, with a gently rounded internal face, and is not strongly crescentic as in most Pseudocheirus. Apparently the protocone had a similar form. Ridges which give crescentic form are present but are weak, broad, and low. This is different from all living species of Pseudocheirus. The poste- rior ridge of the hypocone is obliquely beveled by a wear facet. It runs directly to the posterior edge of the tooth in a gentle, labially- curved arc, then turns sharply labiad and joins the ectoloph at its base near the postero-labial corner of the tooth to form a shallow, closed posterior basin. The antero-intemal corner of the tooth (in- cluding much of the protocone) is missing so the presence of similar ridge and basin in this region is uncertain but is suggested by what is preserved. The anterior ridge of the hypocone extends in a slightly meandering path to the postero-internal corner of the base of the paracone. There it joins a ridge on the paracone by means of a minute spur that crosses the intervening cleft. The paracone and metacone are low but pointed and the ectoloph ridges are worn slightly: the enamel has been breached only at tips of cusps. Each has a low, rounded ridge on the antero- and postero-internal faces. The anterior one of the paracone extends from the apex of the cusp to the base but does not join the anterior edge of the tooth. The posterior one of the paracone extends from the apex to near the base where it divides. One branch of this division continues in line, to meet the aforementioned ridge and spur from the hypocone, while the other swings medially at the posterior end of the protoconule. The anterior one of the metacone is irregular, gives branches, and 28 FIELDIANA: GEOLOGY, VOLUME 19 joins the metaconule. The posterior one of the metacone joins the posterior end of the metaconule. Most of these internal ridges of paracone and metacone are found in Pseudocheirus archeri, P. lemu- roides, P. herhertensis (posterior one on paracone only), and Schoino- bates volans. They are best developed in the former species and grade roughly to the latter. They are absent or only incipient in P. peregrinus and P. occidentalis. They are well developed in Phasco- larctos cinereus and Litokoala kutjamarpensis. In all living species of Pseudocheirus, Schoinobates, and Phascolarctos the ridges on the para- cone and metacone are nearly straight, although in Pseudocheirus archeri (and occasionally in the others, especially in the anterior mo- lars) they divide near the base of the cusp to contribute to the com- plex enamel ridging characteristic of that species. In Phascolarctos cinereus the posterior ridge of the paracone is the best developed of the four. The parastyle and mesostyle are low and rounded, much as in Pseudocheirus peregrinus but relatively smaller. The mesostyle is not bifurcate as in P. peregrinus, P. archeri, and Schoinobates volans. The parastyle and mesostyle do not have the blade-like wings which in P. archeri extend posterior from the parastyle and both anterior and posterior from the mesostyle. No external cingulum is present such as in P. herbertensis. The central part of the outside surface of both the paracone and metacone has a gently convex central ridge as in Pseudocheirus peregrinus, P. herbertensis, P. occidentalis, and P. convolutor. P. archeri has a well-defined though broad ridge on the outside of each of these cusps. The outside surfaces of these cusps of Schoinobates volans are generally concave, with only traces of the ridge. There is no distinct metastyle. In the stylar area of the tooth the ectoloph ridge stands more nearly vertical than in M^ in Pseudocheirus and Schoinobates. All of these genera differ in this re- spect from Phascolarctos in which the ectoloph is nearly horizontal. The undescribed Kutjamarpu Pseudocheirus is most comparable to Pseudokoala erlita in this regard. The protoconule and metaconule of Pseudokoala erlita are well developed but differ from those of all living Pseudocheirus in the following ways: 1. The protoconule is straight and is oriented in an antero- posterior direction rather than being curved. The transverse (anterior) portion is absent and the protoconule joins the ante- rior edge of the tooth. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 29 2. The metaconule is not smoothly curved but has a right angle turn at its apex with one part (posterior) oriented antero- posteriorly and the other transversely. The metaconule of Pseu- docheirus herbertensis approximates this right angle configuration and the posterior part is oriented straight antero-posteriorly. 3. Both protoconule and metaconule are low and rounded rather than sharp as in all living Pseudocheirus. The enamel of the tooth is coarsely crenulate in the central basin. In this respect it is more like Phascolarctos and Litokoala, in which the crenulation is best developed in the basin between the paracone and metacone, than Pseudocheirus archeri and P. herbertensis. In these latter species the crenulation is found over much of the tooth and tends to radiate from the cusps. There is a short, broad, rounded, irregular ridge on the inside of the ectoloph located half-way between the paracone and the meso- style (at the transverse break in the tooth). It forms part of the crenulation of the crown of the tooth. It arises below the top of the ectoloph and ends in the valley between the paracone and metacone without joining any other ridges. Any of the first three molars of Pseudocheirus archeri has a small ridge (often two or three ridges) in this position, usually best expressed in M^^ or M^ however, and the presence of these ridges and their connections to other ridges is vari- able even from left to right in the same dentition (FM 60919). In Pseudokoala erlita a small elongate cuspule similar to that in Ldtokoala is also present in the median "V" of the ectoloph. It does not join any other ridges. Its long axis is oriented transverse to the tooth. A similarly located minute cuspule is sometimes found in Pseudocheirus archeri but it is variable in its occurrence between individuals, between the molars of one individual, and from side to side for any molar position. In Phascolarctos cinereus the apparent homologues of these crenu- late, ornamental ridges which come off from the ectoloph between the paracone and mesostyle are also well developed. The other cusp- ular development located just medial to the mesostyle in the bottom of the valley between paracone and metacone is more pronounced in P. cinereus in that it generally has the form of an elongate ridge which is usually tied to other more medially situated ridges in a variable manner. 30 FIELDIANA: GEOLOGY, VOLUME 19 NMV-P26W0~L. M^, PI. VIII, Fig. A; PI. VII, Figs. C, E This tooth of Pseudokoala erlita consists of the central basin, and the adjacent portions of protocone, paracone, and metacone. Proto- cone is low, somewhat blunt, and not as strongly crescentic as that in all living Pseudocheirus. The ridges which give the crescentic form to the protocone extend anteriorly and posteriorly and form the internal and antero-internal margins of the tooth. An open groove is present on the anterior side of the protocone. It extends from the region of the apex toward the paracone and grad- ually disappears. The protocone slopes gently toward the center of the tooth and ends in a row of poorly defined "bumps" which are separated from the rest of the tooth by a shallow groove. No living Pseudocheirus shows any of these features. The paracone of M^ has antero-internal and postero-internal ridges, as does the paracone of M^, but they do not reach the base of the cusp. The internal face of the paracone is a flattened surface. The metacone is broken off but was apparently present though reduced. It also appears to have been rotated out of line with the paracone, much as in Pseudocheirus occidentalis, P. peregrinus, and P. herbertensis, and seems to have been about as well developed rela- tive to the other cusps. The hypocone is barely discernible on the postero-internal margin of the tooth. It has about the same relative size as that of P. occi- dentalis. There is a single large central basin. The center of the central basin is raised and the enamel is coarsely crenulate as in the M^. Most of the crenulation is found posterior to a line between the protocone and paracone. This feature gives a strikingly differ- ent appearance to these two areas of the basin. A very small proto- conule is present antero-internal to the paracone. It is about four times longer than wide, oriented antero-posteriorly, and is rounded as in M^. NMV-P26399 (the M^) and this specimen (NMV-P26400) have interdental wear facets which fit together so perfectly that these teeth must be the adjacent teeth of the same individual (PI. VII, Figs. C-E). They were both found early in the processing of the matrix, coming from one of the first large batches of matrix to be worked. PM lt588— Anterior half, left My This specimen preserves paraconid, metaconid, and most of the protoconid. The tooth is broken away postero-lingually along the TURNBULL AND LUNDELIUS: HAMILTON FAUNA 31 axis of the crest that connects metaconid and hypoconid. It is worn and badly broken. The metaconid is only slightly higher than the paraconid, which may be due to wear. The metaconid and paraconid are joined by a strong ridge. The anterior part of the protoconid is joined by a worn ridge (metalophid) to the paraconid-metaconid ridge at about one-third of the distance forward from the metaconid. An anterior cingular ridge arises from a short crest extending anteri- orly from the paraconid. It then extends labially and posteriorly and almost reaches the protoconid near the base of that cusp. This ridge cuts off a basin anterior to the protoconid. A large crenulation arises near the anterior edge of the basin and extends posteriorly to the middle of the basin where it dies out. This secondary ridge is separated from the cingular ridge by a narrow cleft. The antero- lingual side of the tooth has an incipient cingular ridge which de- creases in size anteriorly. It joins an anterior extension of the para- conid-metaconid crest (anterior to the paraconid) at the anterior tip of the tooth as does the external cingulum. The protoconid is weakly crescentic in outline but not blade-like, though wear may be responsible for the rounding that it shows. The valley between the protoconid and the (missing) hypoconid is occu- pied by a coarse, round, pillar-like crenulation that lies close to the hypoconid crest. The posterior internal face of the protoconid has some coarse "bumps." The labial face of the metaconid has a slight ridge which runs a very short distance externally before swinging sharply backward as it descends the posterior edge of the cusp. It is tied to a weak ridge which extends lingually from the protoconid. Both of these joined ridges show beveled wear facets. These two ridges and the metalophid combine to form a somewhat triangularly-shaped small basin. The tooth does not have the deep valley between the paraconid and metaconid such as is present in Perikoala palankarinnica and which was mentioned by Stirton (1957a, p. 73, fig. 10). It resembles the My of Pseudocheirus archeri in all of its basic morphology and in being heavily ornamented; but the details of the ornamentation are different. It shows the pseudocheirine tendency toward lateral com- pression of the trigonid. Also, the paraconid is present and the proto- conid is relatively low. In the MyS of Phascolardos cinereus and Perikoala palankarinnica the paraconid is gi'eatly reduced and the protoconid and metaconid are subequal and well separated. Of the various species of Pseudocheirus, two (both in the sub- genus Pseudochirops) , P. archeri and P. cupreus, are much like 32 FIELDIANA: GEOLOGY, VOLUME 19 Pseudokoala, on the one hand, and the phascolarctines, on the other, than are any of the rest of the species of Pseudocheirus known to us. These similarities include size, basic morphology, and the extent of superficial ornamentation. Such similarities led us to err in our earlier report (Turnbull et al., 1965) where we noted the presence of a koala in the fauna based on NMV-P26399. We now believe this specimen to be a pseudocheirine rather than a phascolarctine (refer to comparisons given on pp. 27-29). Measurements. — See Table 2. Table 2. — Length and width measurements of the teeth of Pseudokoala erlita in the Hamilton fauna (in millimeters). M^ (NMV-P26399) M* (NMV-P26400) Mr (PM 4588) Length 6.9 5.3 (close) — Anterior width — (5.0)* — (4.6)* 2.5 (est. close) Posterior width 4.6 — (3.8)* — * Estimated Discussion.- — The low lingual cusps of M^ suggest that P. erlita was at a more primitive stage of dental evolution than any of the other known phascolarctids. The elongate, rectangular outline of M^ corresponds with that of upper molars of pseudocheirines (which vary from nearly square to elongate) rather than with phascolarc- tines (which are consistently nearly square). The labial cusps also show a slightly more primitive condition than any modern pseudo- cheirine, to judge by the ectoloph development. The nearly vertical orientation of the ectoloph (especially the labial side of the stylar cusps) we take to be a somewhat primitive pseudocheirine feature, since all other known genera of the subfamily have less vertical ecto- lophs than P. erlita. Only the undescribed Kutjamarpu pseudo- cheirine shows as vertical an ectoloph, and it is thought to be older. In all phascolarctines the ectoloph is appressed against the crown surface and shows no tendency to stand off vertically. In the upper molars of pseudocheirines the angle the ectoloph makes with the occlusal surface ranges from a low of 50° for Pseudocheirus archeri to a high of 85-90° for Pseudokoala erlita. In the upper molars of phascolarctines this angle varies from 35-40° for Phascolarctos cine- reus to 45° for Litokoala kutjamarpensis. Cf. Pseudokoala erlita Four other tooth fragments are very tentatively referred to Pseu- dokoala. Two of these, PM 4495 and PM 4496 (PI. VIII, Figs. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 33 C, D), differ from one another somewhat and both appear to be the posterior half of a stylar cusp. Both are unhke P. erlita in that they show distinct cingular mesostylar or metastylar "wings." Perhaps these were developed more on the more anterior molars as seems to be the tendency in most species of pseudocheirines. Both have cren- ulations in the valley that lead out toward the metastyle and run up onto the paracone. The crenulations are smaller and more elon- gate in PM 4496 than in PM 4495 which has large and bluntly rounded crenulations. PM 4729 is such a small bit of tooth that its assignment is ex- tremely dubious, yet it possesses a tantalizing amount of morphol- ogy. It consists of a single broken cusp with crenulations at the base. These crenulations and the apparent size of the cusp are the features that suggest that it may be referrable to Pseudokoala. PM 4589, a postero-lingual corner of a right M^ which shows wear, is placed here with fewer misgivings than are the three last mentioned tooth fragments. It shows but little indication of the sort of crenulate enamel that typifies the best of the P. erlita mate- rials. We believe that this is because the only parts preserved that are unworn are the posterior and labial sides, and that these areas were not crenulate in P. erlita lower molars. In Pseudocheirus (Pseu- dochirops) archeri these same areas lack crenulation although much of the occlusal surface is covered with crenulations. Presumably, lower molars of P. erlita had similarly uncrenulated sidewalls. Al- though not measurable, the tooth is large enough to be in the range for P. erlita, and is far too large to belong to either P. marshalli or P. stirtoni (see below). It has the lingually open valley behind the entoconid, typical for all pseudocheirines, but there is no trace of an entostylid ridge. If we are correct in this assignment, the lower mo- lars of P. erlita apparently lacked an entostylid ridge (whether con- tinuous or interrupted; see pp. 37-38) or else its expression followed a gradient along the tooth row, and the trait was gone on M^. Pseudocheirus Ogilby, 1837 The genus Pseudocheirus is represented by more than 60 teeth or fragments of teeth, most of which fall into two species. One of these shows many similarities to species of Schoinobates. The association of the isolated upper and lower cheek teeth is based upon the corre- lation of characters associated in the upper and lower teeth of Recent species and upon functional and occlusal resemblances. 34 FIELDIANA: GEOLOGY, VOLUME 19 Pseudocheirus stirtoniJ new species Holotype.— Left Mt_t (NMV-P26401-P26404), associated on the basis of morphology, preservation, and the exact fit of the inter- dental facets (Plates IX-XII). Referred material. — Upper Cheek Teeth: Left P^ (PM 4549, PI. XIII, Figs. A-C) ; right P^ (PM 4590) ; right M^s (or M^s) (PM 4541, PM 4543, PI. XIII, Fig. F; PI. XIV, Fig. A); left M^ (or M^) (PM 4542, PL XIII, Figs. D, E) ; right M"s (or M^s) (PM 4422, PI. XIV, Figs. B-D; NMV-P26410, PI. XIV, Figs. E, F); left M^ (or M^) (PM 4500) ; right M^ (or M^) (NMV-P26411) ; left M^ (or M^) (PM 4479) ; fragment of an upper molar with paracone, parastyle, metacone, and metaconule (PM 4480) ; left M^ or M^ frag- ment consisting of metacone only (PM 4493) ; paracone of left M^, M^, or MA (PM 4746) ; paracone of left upper molar (PM 4747) ; protocone of left upper molar (PM 4467) ; left upper molar fragment with parts of paracone, metacone, and metaconule (PM 4749) ; and the protocone of a left upper molar (PM 4759) . Lower Cheek Teeth: Left P^s (PM4426; PM 4477, PL XII, Figs. C-E; PM 4548); left My (PM 4474); right M,, (or M^) (PM 4424) ; left M^ (or M^) (PM 4545) ; anterior half right M^ or M^ (PM 4727) ; midsection of left M^ or M^ (PM 4550) ; labial half of a right M^ (PM 4758) ; a left metaconid (PM 4753) ; and a left pro- toconid (PM 4468). The left P^ (PM 4477) is probably (but not certainly) associated with the type on the basis of the fit of the in- terdental wear facets and preservation. It was recovered from the same lot of matrix as the teeth which constitute the type. Diagnosis.- — An animal near the size of Pseudocheirus pygmaeus, with molars slightly larger than those of that species (see Measure- ments and graphs). P^ triangular with three laterally compressed labial cusps; the crest of the anterior cusp rounded in antero-poste- rior profile; anterior cusp close to center cusp but separated from it by a deep fissure rather than joined to it as in Schoinobates; central cusp with lingual ridge which extends from apex in a sigmoid curve to the back of the tooth near the postero-lingual corner; lingual cingulum present. All cusps of upper molars thin, sharp, and cres- centic. Protoconule and metaconule well developed, sharp, and crescentic as protocone and hypocone. Posterior crest of protocone and anterior crest of hypocone not joined. Postero-intemal ridges ' In honor of the late R. A. Stirton, an ardent student of the Australian Tertiary. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 35 of paracone and metacone much stronger than antero-internal ridges which may be weak or absent. Protocone and hypocone with distinct median labial ridge. Cusps not joined by secondary trans- verse ridges. Enamel smooth and unornamented. The anterior edge of paracone of M-l rounded, and not connected to the prominent para- style by a ridge. Pj very short with two very closely spaced large twinned cusps in center of tooth. These cusps have a deep cleft between them. Posterior large cusp with two posterior ridges, the labial one running either postero-labially, labially, or antero-labially before turning posteriorly and joining a cuspule; the lingual one run- ning posteriorly and lingually and sometimes ending in a cuspule. Mt_7 with well-developed entostylid ridge which is continuous with the entostylid. Anterior ridge of hypoconid wavy on all lower mo- lars. Protoconid of My with distinct anterior and posterior crests. Well-developed metastylid connected to metaconid by a crest. Meta- stylid of My with well-developed posterior crest. My with distinct paraconid. Descriptions.— The P^ (PM 4590 and PM 4549, PI. XIII, Figs. A-C) is a triangular tooth with three high, laterally compressed labial cusps. The anterior of these is rounded in its antero-posterior profile and is separated from the middle one by a deep fissure. The central cusp is more pointed and is joined to the posterior cusp by a thin sharp ridge. The posterior cusp is very compressed laterally and its posterior edge merges with the shelf-like posterior border of the tooth. The central cusp has on its lingual side a ridge which ex- tends in a sigmoid curve dorsally and postero-lingually and fades out near the postero-lingual corner of the tooth. The lingual margin of the tooth is not preserved in either of the Hamilton specimens, but it is clear that there is a cingular basin lingual to the postero-lingual ridge of the central cusp. Thus, there are two posterior valleys as in Schoinobates and P. lemuroides. All other species of Pseudocheirus have only one posterior valley on the P^. The tooth closely resem- bles the PA of Schoinobates. It differs in having the anterior cusp located closer to the central one and in not having a ridge joining them. Most of the species of Pseudocheirus have much more com- plicated P^s than the fossil form. A few species of Pseudocheirus, P. forbesi, P. pygmaeus, P. canescens, P. dahli and P. herbertensis, have small and simple P^s but none has the labial cusp as compressed as do these Hamilton specimens and Schoinobates. The cusps of all upper molars are thin and sharp. The outer ridges of the paracone and metacone are crescentic except in M^-. 36 FIELDIANA: GEOLOGY, VOLUME 19 On either side of their junction (at the mesostyle) is a recurved, wing-Hke cingular ridge. The anterior one is usually the better de- veloped of the two. The anterior edge of the paracone of M^- is rounded and not joined to the well-developed parastyle by a ridge (PI. XIII, Figs. D-F; PI. XIV, Fig. A). In this it is like the M^- of Schoinobates volans. The postero-internal ridges on the paracone and metacone are strong. The one on the paracone is nearly straight until it reaches the bottom of the cusp where it may turn sharply posteriorly, then labially before dying out. The ridge on the meta- cone is more curved and usually ends abruptly at the base of the cusp, but sometimes joins the posterior crest of the hypocone at the posterior edge of the tooth. The antero-internal ridges on the para- cone and metacone are more reduced than in most species of the genus Pseudocheirus or are absent, especially on M^. This is very similar to Schoinobates. The protocone is strongly crescentic. Its anterior crest forms the anterior edge of the tooth, and joins the parastyle in M-L"^. A small ridge is present on the anterior crest of the protocone, which forms a very rudimentary secondary selene. This is best developed on the MJ-. A similarly-positioned, well- developed, additional selene characterizes the upper molars of all Recent species of the subgenus Pseudochirops. It is weakly devel- oped in Schoinobates and many other species of the other subgenera of Pseudocheirus. The posterior crest of the protocone extends pos- teriorly and labially and ends in a valley between protocone and hypocone near the posterior end of the protoconule without joining another cusp. The hypocone is strongly crescentic. Its anterior crest runs forward and swings labially into the median valley where it disappears near the postero-lingual ridge of the paracone. The posterior crest of the hypocone forms the posterior border of the tooth and joins the posterior end of the ectoloph. The labial faces of the protocone and hypocone have distinct ridges which extend from the apices of the cusps to their bases where they disappear without joining the protoconule or metaconule. This is also the con- dition in Schoinobates. The protoconule and metaconule are well developed, and their antero-lingual faces are angular or ridged. The protoconule has its anterior end joined to the parastyle or paracone in MJ-, and to the anterior edge of the tooth in M^"^ (PI. XIV, Figs. D, F). It terminates posteriorly at (but not joined to) the anterior crest of the hypocone. Its apex is lined up with those of the para- cone and the incipient additional selene on the anterior crest of the protocone. This is the usual pattern in those Recent species of pseu- docheirines in which the additional selene is present (Schoinobates TURNBULL AND LUNDELIUS: HAMILTON FAUNA 37 and all species of P. (Pseudochirops)) . The anterior end of the metaconule runs into the valley between the paracone and metacone and ends without joining any other ridge or cusp. The posterior end dies out labial to the postero-intemal ridge of the metacone in the valley, usually near the posterior margin of the tooth. It does not join any other cusp or ridge. The apex of the metaconule is anterior to a line connecting metacone and hypocone. A small ridge, variable in length, is present at the posterior side of the base of the paracone of M-L -. It parallels the posterior ectoloph ridge of the paracone and does not join any other ridge. A small lingual cingular cusp may be present between the protocone and hypocone of M^"^ (PI. XIV, Figs. D, F). It is variable in size and exact location; it may lie ex- actly between those cusps, or more anteriorly against the base of the protocone. This cuspule shows similar variations in several Recent species of pseudocheirines. The P4 is a very short, triangular tooth with two large closely- spaced cusps in the center of the tooth. There is some variation in their spacing but they are always twinned and have a deep transverse cleft between them. The anterior large cusp is bulbous and rounded anteriorly. The posterior one is triangular with three ridges diverg- ing from the apex. The antero-lingual ridge extends from the apex of the cusp to its base where it disappears. The postero-labial ridge runs postero-labially in PM 4426, labially in PM 4548, and antero- labially for a short distance then turns sharply posteriorly in PM 4477 (PL XII, Figs. D, E). In all three specimens this ridge ends in a cuspule at the postero-labial corner of the tooth. The postero-lingual ridge runs posteriorly and lingually and sometimes ends in a tiny cuspule. The P4 is one of the most distinctive teeth in Pseudocheirus stirtoni. It is most like the P^ of Schoinobates and Pseudocheirus peregrinus, but differs in being relatively smaller and shorter with its crowded, twinned, main cusps. Within Schoinobates it is more like S. volans than S. minor in the presence of two well-developed main central cusps and a developed postero-labial cusp. In S. minor the posterior of the main cusps, and the postero-labial cusp, are both very weakly developed being little more than sharp crests. In the reduction of the anterior cuspule it is more like Pseudocheirus pere- grinus. The lower molars have the general structure of Pseudocheirus molars. They resemble P. (Pseudochirops) molars in having a well- developed entostylid ridge in the valley between entoconid and hypo- 38 FIELDIANA: GEOLOGY, VOLUME 19 conid. This ridge is continuous with the entostyhd. It is present on all lower molars except M^. In the living species of Pseudocheirus only P. archeri has this ridge on the M4. Schoinohates has a small entostylid ridge on Mj.t^, which is not continuous with the entostylid, and none on M^ t- The enamel is smooth and unornamented. The My has a prominent protoconid which is not joined to any other cusp. The protoconid has well-developed anterior and posterior crests of which the posterior (Arrow #1, Plate IX, Fig. A) is the larger. Schoinohates and several Recent species of Pseudocheirus (P. occi- dentalis, P. herbertensis, P. convolutor, and P. lemuroides) have a com- parable ridged independent protoconid on the My. The metaconid is the largest cusp on the trigonid, and it is connected to the para- conid by a sharp ridge. A prominent metastylid is present which is almost as high as the metaconid and is joined to the metaconid by a ridge. It has a strong posterior crest which extends about half the distance to the entoconid and which overlaps lingually the anterior crest of the entoconid. The crest that connects the hypoconid and metastylid is wavy (Arrow #2, PI. IX, Figs. A, C). It is interrupted in the middle by a small cleft in NMV-P26401 but not in PM 4474. The posterior end of the tooth is formed by a slightly wavy ridge which connects the hypoconid and the hypoconulid. The entostylid ridge is straight and extends from the entostylid to an imaginary line connecting the hypoconid and entoconid (Arrow #3, PI. IX, Fig. A). There is a variably developed, procumbent anterior cingulum. Mij and M^ resemble the My in the form of the talonid except for the terminal connection (s) of the anterior crest of the hypoconid (see area anterior to Arrow #2, PL IX, Fig. D; PL X, Fig. D). The anterior part of the teeth is somewhat different; there is no para- conid, and the protoconid has two crests. The anterior one extends straight anteriorly then turns about 45° lingually and almost joins the anterior crest of the metaconid at the anterior end of the tooth. The posterior crest is oriented transverse to the long axis of the tooth and ends at the posterior side of the base of the metaconid and either does not join it or joins by a very weak ridge. The ante- rior crest of the hypoconid of M^ (Arrow #2, PL IX, Fig. D; PL X, Fig. C) is wavy and is angled in much the same way as the anterior crest of the protoconid. It does not join the metastylid but joins the posterior crest of the protoconid close to the midline of the tooth. That of the M^ (PL X, Fig. D; PL XI, Fig. A) is only weakly wavy. The posterior crest of the metaconid on M^^ - My (PL IX, Fig. D; PL X, Figs. A, D; PL XI, Figs. A, D, E) slopes downward and joins a distinct but low metastylid lingual to the anterior crest of the ento- TURNBULL AND LUNDELIUS: HAMILTON FAUNA 39 conid. The metaconid and entoconid have rounded ridges which ex- tend from the apices to the bases on both the lingual and labial sides. The entostylid ridge of M^ is large while that of M^ is short (compare Arrow #3, PI. IX, Fig. D; PL X, Fig. D). The M4 (PI. XI, Figs. D, E) lacks the entostylid ridge as in all Recent Pseudocheirus, except P. (Pseudochirops) archeri in which it is well developed. The anterior crest of the protoconid is curved, not angled. The posterior end of the tooth is not flattened but is rounded. The junction of the anterior crest of the hypoconid with the posterior crest of the protoconid is located farther labially than in the other molars, as is generally true in the M4 of Pseudocheirus. Measurements. — See Table 3. Table 3. — Measurements of teeth of Pseudocheirus stirtoni (in millimeters) Anterior Posterior Length Width Width P^ PM 454t ) 2.2 >1.2 >1.3 PM 4590 2.2 >1.1 — M^^ (or M^) PM 4541 >2.7 — 2.7 PM 4542 3.2 ~2.5 ~2.4 PM 4543 3.2 2.5 2.5 M^ (or M^) PM 4422 2.9 2.7 2.5 PM 4500 3.0 ~2.7 2.5 NMV-P26410 3.1 2.8 2.5 M^ (or M^) NMV-P26411 2.9 2.6 2.1 PM 4479 3.0 2.5 2.2 P? PM 4426 2.0 1.0 1.3 PM 4477 2.2 1.1 1.3 PM 4548 2.1 1.1 1.2 Ml NMV-P26401 3.3 1.8 1.8 PM 4474 3.2 1.8 1.8 Mj NMV-P26402 3.3 1.9 1.9 M2 (or Mg) PM 4424 3.1 1.6 1.9 PM 4545 3.2 1.6 1.7 PM 4727 — 1.7 — M^ NMV-P26403 3.1 1.8 1.9 M5 NMV-P26404 3.5 1.8 1.5 PM 4758 3.3 — — Discussion. — The genera Schoinohates and Pseudocheirus cannot be distinguished from one another solely on the basis of individual dental features. Even tooth proportions are inadequate for this, as can be seen from Appendix graphs A-D. However, Schoinohates does have a composite set of dental features that distinguish it. P. stir- toni shows many of the Schoinohates characterizing features (strong postero-intemal and weak antero-internal ridges on paracone and metacone; rounded anterior edge of paracone on M^; double valley 40 FIELDIANA: GEOLOGY, VOLUME 19 on posterior end of P^; and a short, morphologically distinct form of Px) but it combines them with three other quite different features or sets of features. The first of these is the extensive continuous entostylid ridge development, such as is characteristic of all species of P. (Pseudochirops) . (Schoinohates has the ridge, too, but it is separated from the entostylid.) Second is the failure of hypocone and protocone to join in M^"^. The third of these different sorts of fea- tures is a new one, unique to P. stirtoni and the other new species of Pseudocheirus described below — the wavy form of the anterior ridge of the hypoconid. Thus P. stirtoni is intermediate in dental morphology between Pseudocheirus and Schoinohates. Because of this, and because the most striking feature of Schoinohates (the gliding membrane) cannot be demonstrated for the fossil P. stirtoni, it is assigned conservatively to Pseudocheirus. Perhaps the strongest dental argument for an as- signment to Schoinohates that can be made is one that relates to the pattern of change in tooth proportions from tooth to tooth (P^-M^), as can be seen in Graph E of the Appendix. In this regard P. stir- toni fits the Schoinohates pattern more closely than that of any of the subgenera of Pseudocheirus (Petropseudes, Pseudochirops, Pseu- docheirus or Hemihelideus) , although several species of P. (Pseudo- cheirus) also have a somewhat similar proportion pattern. Unfor- tunately, the sample sizes are small and until stronger evidence is forthcoming, based upon better samples, we prefer the more conserv- ative assignment of the species to the broader genus Pseudocheirus. Pseudocheirus tnarshallii new species Holotype.—heft M^ (NMV-P26405), a nearly complete tooth showing moderate wear. Plate XV, Fig. B. Referred specimens. — Fragment of a slightly worn left upper mo- lar, probably an M-l, preserving metacone, metaconule and part of protocone (PM 4423, PI. XVI, Fig. A) ; slightly worn left M-^ (PM 4587, PL XV, Fig. C) ; five upper molar fragments, three of which are each a left metacone (PM 4460, PM 4750, PM 4751), one a left hypo- cone and one a right hypocone (NMV-P26412, PM 4449) ; right P^ nearly complete and showing a trace of wear (PM 4591, PI. XV, Fig. A) ; unworn right My (PM 4455, PI. XVI, Fig. B) ; fragment of right My, including parts of protoconid and metaconid (PM 4461) ; slightly worn right Mtj (or My) fragment preserving hypoconid, ento- * In honor of the late A. J. Marshall, founding professor of Zoology and Com- parative Physiology, Monash University. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 41 stylid and its ridge, and half of the entoconid (PM 4752) ; two par- tial left molars, either M^ or M^, one a talonid (PM 4425), the other a talonid with posterior half of trigonid (PM 4453, PI. XVII, Fig. A) ; slightly worn right M^ (or M^) (NMV-P26413, PI. XVI, Fig. C) ; a right M4 showing the beginning of wear on the anterior ridge of pro- toconid (PM 4476, PI. XVII, Fig. B) ; and two fragments of lower molars, a right entoconid (NMV-P26414) and a left metaconid (PM 4462). It is probable that the M^ (NMV-P26413) and the M4 (PM 4476) belong to the same individual on the basis of wear and preservation, but the teeth had not been functional long enough for interdental facets to have formed. Diagnosis. — An animal near the size of P. stirtoni or P. pygmaeus, with unornamented teeth about the size of those of the latter species. Primary cusps of trigon, trigonid, and also the hypoeone, entoconid, and hypoconid are all more attenuated toward their apices, have sharper primary cresting and lack or have less secondary cresting than in P. stirtoni. M^- with low, well-developed, pjrramidal proto- conule and metaconule, each with three straight ridges running from its apex. Paracone of M^- rounded anteriorly, laterally compressed posteriorly to a blade-like form, and lacking internal ridges. Para- style prominent, and metacone compressed and rounded posteriorly and lingually and with only traces of internal ridges. P^ elongate, triangular, and with two prominent midline cusps; the anterior large and rounded, the posterior smaller and ridged antero-posteriorly. A third cusp on the postero-labial corner connected to posterior end of ridged cusp by a diagonal ridge. The posterior end of tooth defined by a straight transverse ridge that extends from postero-labial cusp to postero-lingual corner of tooth. Entostylid ridge prominent on lower My but interrupted (i.e., not connected to entostylid); vari- able, but weak if present, in M^; absent in M^-j. Mij and M^ with small entostylid and occasionally a trace of a hypoconulid. My with rounded, isolated protoconid. Posterior molars with thin blade-like metaconid and entoconid; posterior ridge of protoconid oriented perpendicular to line connecting those cusps. It is not angled or wavy as it usually is in the posterior molars of P. stirtoni. Anterior crest of hypoconid wavy but less so than in P. stirtoni. Descriptions.— [Jffer molars: A left M^ (NMV-P26405, PL XV, Fig. B) consisting of about three-fourths of a tooth with protocone and anterior half of hypoeone missing, and one other upper molar fragment (PM 4423, PI. XVI, Fig. A) which is probably an M^ are 42 FIELDIANA: GEOLOGY, VOLUME 19 the only anterior molars of P. marshalli in the collection. The major cusps are attentuated toward the apex. The anterior end of the para- cone is rounded and the posterior part is laterally compressed to a blade-like structure somewhat like that of the M-l of Pseudocheirus occidentalis. There are no internal ridges on the paracone. The meta- cone is rounded on its lingual face and shows only traces of the inter- nal ridges. There is a pronounced wing-like recurved cingular ridge, a part of a mesostyle, just anterior to the middle of the ectoloph. As in the MJ- of most species the ectoloph shows more flexure of its middle than would be expected in M^ or M^. The parastyle is large and separated from the paracone by a rounded groove that shows wear. The protoconule and metaconule are pyramidal in form. Each has a straight posterior crest oriented antero-posteriorly, and an an- terior, labial crest oriented transversely. Each has a short straight ridge on the antero-lingual face. The anterior crest of the proto- conule is very short and connects it with the base of the paracone. The anterior crest of the metaconule extends well into the valley between paracone and metacone. There is no secondary ridging in the valleys, and the enamel is smooth and unornamented. The left M^ (PM 4587, PI. XV, Fig. C) is eloangted antero-poste- riorly and is sub-triangular in outline. The three major trigon cusps are attenuated. The protocone is crescentic; its anterior ridge runs from the apex of the cusp diagonally forward and labially in an even arc which flattens out before reaching half-way across the tooth, and then runs straight labially along the front of the tooth to a small para- stylar cusp. Paracone and metacone are laterally compressed and lack any internal cresting. Most of the ectoloph ridging is concen- trated in the area between these cusps, where the ridges give off cingular wings which delimit a broad mesostyle. Lingual to the metacone at the postero-labial end of the posterior ridge of the proto- cone, there is a small hypocone. It is tied labially at its base to the base of the metacone, and more posteriorly by a continuation of the ridge that forms its axis to the base of the metacone near its posterior edge. Thus a minute basin is found between metacone and hypo- cone. There is a large and quite smooth-bottomed central basin with few features in it. The arm of this basin which extends broadly be- tween the spread central ectoloph crests, has a slight valley or pit just lingual to the mesostyle. A similar valley lies immediately in front of the hypocone. A weak, straight antero-posteriorly oriented protoconule is present and ties straight to the front edge of the tooth. There is no trace of a metaconule. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 43 Lower teeth: The P^ is represented by an almost complete tooth (PM 4591, PI. XV, Fig. A) with only the anterior portion miss- ing. It is elongate, triangular and has three cusps preserved. The anterior of these (1 in PI. XV, Fig. A) is the largest and appears to be the homologue of the main cusp of the P4 of P. occidentalis (PM 4765) in which it is located just anterior to the center of the tooth. It is rounded with two weak ridges on its anterior face that extend anteriorly to the broken edge of the tooth. There was probably a small anterior cusp as on the P^ of most living species of Pseudo- cheirus. Posterior to the large rounded cusp is a smaller median cusp (2 in PL XV, Fig. A) with antero-posterior ridging. This ridge turns abruptly postero-labially at the posterior end of the cusp and joins another cusp (3 in PI. XV, Fig. A) on the postero-labial comer of the tooth. From this cusp a ridge extends across the posterior end of the tooth and ends at the postero-lingual corner. The tooth is quite similar to the P4 of P. occidentalis and P. lemuroides in the basic arrangement of the cusps. However, the ridge that connects the two posterior cusps runs down the posterior side of the compressed me- dian cusp rather than off from its labial side. Also, the posterior end of the tooth is oriented more transversely with respect to the long axis of the tooth. The My (PM4455, PL XVI, Fig. B), is similar to, and is only slightly smaller than, that of P. stirtoni, but differs in two funda- mental ways. The protoconid, which is not joined to any other cusp, is not compressed laterally as in P. stirtoni, but is a simple oval pillar. Also, the entostylid ridge is not continuous with the entostylid (a Schoinobates characteristic) as it is in P. stirtoni. Neither of these two diagonostic features shows variation in the Mj of any of the living species, and thus it seems unlikely that PM 4455 could be simply a variant of P. stirtoni. Because of the slight size difference, the major cusp height and attenuation, and the character of the de- velopment of the entostylid and its associated ridge which shows a gradient descending from My to M4 (as in Schoinobates) , P . marshalli is distinguished from P. stirtoni. The posterior lower molars (PM 4425, PM 4453, NMV-P26413, PM 4476) are also similar to those of Pseudocheirus stirtoni but have no more than a trace of an entostylid ridge. The M^ (NMV-P26413, PL XVI, Fig. C) typifies them all. The metaconid and entoconid are very thin and blade-like and lack prominent rounded ridges on the labial faces. There is a weak entostylid, and while this specimen lacks a hypoconulid, a trace of one can be seen on PM 4425 and PM 4453. The posterior crest of the protoconid (except in M^) is 44 FIELDIANA: GEOLOGY, VOLUME 19 oriented perpendicular to a line connecting metaconid and entoconid instead of slightly obliquely as is usually the case in P. stirtoni. It does not join either the metaconid, the anterior crest of the ento- conid, or the metastylid. The anterior crest of the hypoconid has a less pronounced labial loop where it joins the posterior crest of the protoconid than it does in P. stirtoni, and is somewhat less wavy than in P. stirtoni. The M4 of P. marshalli (PM 4476, PL XVII, Fig. B) has a thin- ner metaconid and entoconid than does that of P. stirtoni. It is slightly smaller than the M4 of P. stirtoni and lacks the low ridges on the labial face of the metaconid. Measurements. — See Table 4. Table 4. — Measurements of teeth of Pseudocheirus marshalli (in millimeters) Anterior Posterior Length width width M^ NMV-P26405 3.4 — — M* PM 4587 2.8 2.3 1.4 P5 PM 4591 — — 1.5 Mt PM 4455 3.0 1.4 1.5 M2 (or M3) PM 4425 — — 1.9 PM 4453 — 1 . 7 est. 1.8 Mj (or M^) NMV-P26413 3.0 1.6 1.7 Mi PM 4476 3.1 1.7 1.5 Discussion.- — P. marshalli shows fewer resemblances to Schoino- hates than does P. stirtoni. It resembles Schoinobates in the rounded anterior edge of the paracone of M- and in the interruption of the entostylid ridge in My and M^, and its absence in M4. This ridge shows a steeper gradient in P. marshalli in its degree of expression from My to M4 than in Schoinobates. In other features it resembles one or more living species of Pseudocheirus. Neither P. stirtoni nor P. marshalli shows any decidedly generalized characters such as might suggest either one to be the ancestor of a living species. Neither one appears to be more primitive or more specialized than most of the Recent species. P. marshalli fits best into the subgenus P. (Pseudocheirus) (Tate, 1945B). Pseudocheirus sp. (incisors and other fragments) Material.~^Two left I^s (PM 4472, PI. XVII, Fig. C; NMV- P26415, PI. XVIII, Fig. A) ; tips of two right lyS (PM 4427, PM TURNBULL AND LUNDELIUS: HAMILTON FAUNA 45 4428) ; section of a right ly at base of enamel (PM 4584) ; distal part of left It (PM 4551) ; root of ly (PM 4585) ; and the following cheek tooth fragments : a protocone or hypocone (PM 4488) ; a metaconid (PM 4463) ; eight isolated fragments (PM 4596) ; an entoconid, prob- ably an Mt (PM 4464) ; a left metaconid (NMV-P26416) ; a worn, hypoconid? (PM 4594, PL XXIX, Fig. F); a hypoconid (PM 4754). Description. — The lyS are very close to the size of the Ij of Pseu- docheirus pygmaeus and show approximately the same longitudinal curvature and dorso-ventral flattening. They differ from the ly of P. pygmaeus and all other living species of Pseudocheirus in being rela- tively narrower and in having a more pronounced bulge on the dorsal surface at the enamel line. The lateral cutting edge is thin and blade- like as in all living species of Pseudocheirus. The profile of the lateral cutting edge is straight, not strongly sigmoid as in most species of the subgenus Pseudochirops. The medial interdental facets tend to truncate the medial dorsal ridge at the distal end of the tooth. This is less pronounced than in the living species of Pseudocheirus and more like the condition found in the ly of Petaurus. There is no way to assign the incisors to either of the named species of Pseudocheirus of the Hamilton fauna, but it seems likely that they probably belong to one of those species. There are no differences between the lys to indicate that more than one species is represented. These incisors show in their shape and the lesser truncation of the medial dorsal ridge, a condition more primitive than is seen in any Recent species of Pseudocheirus. Family Macropodidae Subfamily Potoroinae Genus and sp. indet., near Aepyprymnus Material. — Metacone of a left upper molar (PM 4492), with the posterior cingulum and a small part of the hypocone preserved. Plate XVIII, Fig. B. Description.- — This specimen is about the size of the metacone in the upper molars of Aepyprymnus rufescens. The cusp is broader and more bulbous than in A. rufescens, particularly the lingual part. There is a poorly defined wear facet on the anterior side which pro- duces what little cresting there is to the low transverse ridge, just posterior to the center of the cusp. The same pattern of wear is found on metacones of A. rufescens. The posterior cingulum (arrow, PI. XVIII, Fig. B) is short, heavy, and rounded, and shows a small 46 FIELDIANA: GEOLOGY, VOLUME 19 wear facet near the broken edge of the tooth fragment. The break surface truncates the extreme postero-labial edge of the hypocone and its pulp canal where the cingulum joins that cusp. The cingu- lum joins the metacone more lingually along its posterior face than in A. rufescens in which it extends onto the labial side of that cusp. The posterior cingular basin thus formed is almost circular (slightly crescentic) and is located at the postero-lingual corner of the meta- cone. This basin in A. rufescens is elongated transversely to the tooth and extends across the breadth of the tooth. PM 4492 differs from the large potoroo ("cuscus") of Gill (1953b, 1957), which was described by Stirton (1957b) and has since been redescribed and iden- tified by Ride (1964), in its smaller size and more bulbous metacone, and the less prominent, less ridged metaloph. Discussion. — This specimen is closer to Aepyprymnus than to any other known genus of the Potoroinae, but the differences listed above suggest that better material may necessitate the establishment of a new genus. Genus and sp. indet., near Hypsiprymnodon Ma^mai.— Anterior half, right M^- (PM 4575, PI. XVIII, Fig. C). Description. — This Hamilton tooth is unworn, sharply crested, and is quite similar to, and about the size of, the M^ of Hypsiprym- nodon moschatus (FM-60953) . It has a rounded, columnar ridge run- ning the full height of the protocone, and sharp angular ones at the antero-lingual (parastyle) and antero-labial corners of the procingu- lum. The procingulum is thus squared off anteriorly; it is connected to the protocone by an uninterrupted angular ridge, and to the para- cone by a similarly sharp ridge which is nicked by a noticeable notch. Rootward from this notch, on the labial side of the tooth, is a de- pressed area. As a result, the antero-labial end of the procingulum has the appearance of a parastylar cusp. There is a short lingually- curved posterior ridge to the paracone which fades out without join- ing a crest from the metacone. A sharp protoloph is a prominent feature of the tooth. It has a minute cleft at its midpoint and it is weakly expanded just labial to this cleft. Postero-labial to the para- cone, on the labial side of the tooth near its base, there is another low ridge that is sharpest away from the apex of the cusp and dis- appears about half-way from base to apex. The posterior crest of the protocone must have reached to, and, judging by the break surface, was probably connected with a crest from the hypocone. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 47 The overall shape of the tooth in crown view is very similar to that of Hypsiprymnodon moschatus because of the extensive develop- ment of the procingulum, with a straight anterior edge. It is also similar in having the well-developed parastyle. It differs in having a postero-labial ridge on the labial side of the paracone, in lacking the posterior and anterior bulges on the labial segment of the protoloph, and in lacking a pitted enamel surface. The only other genus of potoroine in which there is a tooth at all comparable is Aepyprymnus. In this genus the M^ shows most of these features, but the ridges are less sharp and the procingulum is more rounded. Sp. indet., with similarities to Potorous and Propleopus No additional representatives of the original potoroine find (Gill's and Stirton's "cuscus") have turned up in our collection. Hence each of these potoroine taxa is presently represented by a single tooth or fragment. By contrast, Dorcopsis is common. It usually is put in the Macropodinae, but here following authority (Wood- bourne, 1967; Stirton et al., 1968) we put it in Potoroinae. Dorcopsis Schlegel and Mtiller, 1839-1844 Dorcopsis sp. indet. Material.— Upper teeth: Posterior third of a right P^ (PM 4436, PI. XIX, Figs. A-D) ; complete unworn right molar (NMV-P26417, PI. XXI, Figs. A-D); complete left molar with good interdental facets, and showing enough wear that the enamel of protocone and paracone is breached, exposing dentine (PM 4433, PI. XX, Figs. A- E) ; complete unworn right molar (PM 4434, PI. XXII, Figs. A-D) ; the anterior half of a right molar (PM 16807) ; two fragments of left molars that may have been adjacent teeth in the same mouth to judge by fit of interdental facets, preservation, and wear — (PM 4485) a worn left metacone, and (PM 4431) a left anterior half tooth with procingulum, protocone, paracone, and protoloph; anterior quarter of a left molar, somewhat worn, with well-developed interdental facet and with procingulum and anterior half of protocone, paracone and protoloph preserved (PM 4484) ; anterior half right molar showing beginning of wear on protoloph, and possessing an unusual, small accessory cuspule postero-labially to the paracone (PM 4570) ; two molar fragments, each consisting of the labial side of the median val- ley-— a right (PM 4572), which also has most of the paracone and procingulum preserved, and a left (PM 4573) ; a fragment of the an- 48 FIELDIANA: GEOLOGY, VOLUME 19 terior face of a left metacone (PM 4755) ; and a fragment consisting of parts of a protocone and procingulum of a right upper molar (PM4713). Lower teeth : Nearly complete slightly worn right molar (prob- ably M^ or M4) lacking tip of protoconid and half the hypoconid (PM 4565, PI. XXIII, Fig. A) ; anterior half right molar (NMV- P26418); anterior half right molar (PM 4482). Descriptions.— The P^ fragment (PM 4436, PI. XIX, Figs. A-D) closely resembles the posterior part of the P^ of Dorcopsis mulleri (FM-8379) from New Guinea in both size and morphology. The outer blade is thin with a flat labial surface up to the posterior ridge- let, where it is broken. There is a postero-lingual cusp which is joined to the outer blade by two ridges. The anterior of these ex- tends antero-labially to the external blade. The posterior ridge extends posteriorly, then turns labially, decreases in height and joins the outer blade at the base of the tooth. These two ridges delimit a posterior basin which has a narrow opening posteriorly. A third ridge extends lingually, then anteriorly from the postero-lingual cusp. It decreases in height and forms a low shelf along the lingual side of the tooth. A small tubercle is located on this shelf at its anterior end, opposite the posterior ridgelet of the outer blade. The P- of D. mUlleri lacks this tubercle. The Hamilton specimen is not as high-crowned as the P^ of the Recent Dorcopsis. Wallabia bicolor has the posterior basin elongated and oriented more or less parallel to the long axis of the tooth, and has a relatively more massive external blade. Dendro- lagus differs from the Hamilton specimen in having no posterior basin, and in its more massive and irregular external blade. The upper molars are sufficiently alike one another that a general description of them may be given and the variations noted. The protoloph and metaloph are sharp when unworn, and are V-shaped or curved with the concavity facing posteriorly. The paracone and metacone have well-developed anterior and posterior crests. The anterior crest of the paracone joins the procingulum. The posterior crest of the paracone is always straight and extends in a slightly lin- gual angle from the apex to the base of the cusp. The anterior crest of the metacone shows considerable variation: it may be straight (NMV-P26417) or curved (PM 4433, PM 4434, PM 4485) and may (NMV-P26417, PM 4433, PM 4434) or may not (PM 4485) reach the base of the cusp. In those teeth in which these crests between paracone and metacone reach the central valley, they may (PM 4433) or may not (NMV-P26417 and PM 4434) be joined. In those in TURNBULL AND LUNDELIUS: HAMILTON FAUNA 49 which the crests do not reach the central valley there is a U-shaped notch between the cusps. Variations similar to these are found as a graded series (from connected crests anteriorly to interrupted crests posteriorly) along the tooth rows of many Recent macropodids. The shape, relative size, and orientation of these crests are more like the Recent Dorcop- sis than any other macropodid, but the gradient along the tooth row is somewhat steeper. The paracone and metacone of Wallahia bi- color, Dendrolagus lumholtzi, Thylogale brunii, T. stigmatica, T. billi- ardieri, and Setonyx brachyurus have these crests or ridges, but they differ from those of Dorcopsis and the Hamilton species in various ways. In Dendrolagus they are much reduced. In W. bicolor these two crests of M-l and M- are joined in the median valley and have a short segment parallel to the long axis of the teeth. In M- and M^ of that species the crests are absent or weak. In all species of Thylo- gale the anterior crest of the metacone is either absent or is reduced to a low ridge. In Onychogale the metacone lacks an anterior crest. In Setonyx these ridges have a smooth curve from apex to central valley, and those of the anterior molars join while those of the pos- terior molars do not. The midlink of the upper molars in the Hamilton species arises on the posterior side of the protocone below the apex and extends postero-labially to the central valley, then turns posteriorly and joins the metaloph at its midpoint. It does not reach the crest of the meta- loph. A short labial spur may be present on the midlink in the cen- tral valley (PM 4434, PL XXII, Fig. C). This is either absent or only slightly developed in the Recent species of Dorcopsis. Low on the labial side of the crown, postero-labial to the protocone, there is a diagonal bulge (in all but one specimen) or ridge (PM 4570) as in upper molars of the Recent Dorcopsis and Dorcopsulus. In these Re- cent forms there is a marked gradient to the expression of this trait, from strongest on the anterior molars to weakest on the posterior. The procingulum is prominent with the straight anterior edge parallel to the protoloph. The anterior crest of the paracone joins the anterior ridge of the cingulum. The cingulum has a square an- tero-labial comer, and does not reach the lingual edge of the teeth. A forelink is absent, being represented by a slight bulge. It is present though much reduced in the Recent species of Dorcopsis. The post- cingulum is crescentic. It leaves the apex of the hypocone, extends lingually and rootward, and then swings up onto the apex of the metacone. This is similar to the postcingulum in Recent Dorcopsis. 50 FIELDIANA: GEOLOGY, VOLUME 19 The postcingulum is greatly reduced in W. bicolor, and in M^-- M^ in Thylogale, W. eugenii, and Onychogale frenata. There is a well-de- veloped cingular shelf labially at the central valley, as in D. miilleri. (Dorcopsulus lacks this shelf.) There is no sign of the pitting such as is found in the central valley and anterior and posterior cingular basins of D. hageni. The one nearly complete lower molar of Dorcopsis sp., in the Hamilton fauna (PM 4565, PL XXIII, Fig. A) is relatively wide, like the molars of the Recent D. hageni and D. miilleri, in contrast to most species of Wallabia. The protolophid and hypolophid either have the form of very flat V's with the concave sides anterior or are straight. The metaconid and entoconid each have a prominent an- terior crest oriented at a right angle to the corresponding protolophid or hypolophid, as in the molars of Recent Dorcopsis. These crests are between one-quarter and one-third the length of the cross lophs. They die out on the lower parts of the cusps. The crests are devel- oped to about the same degree in D. miilleri (FM-8379) and D. hageni (FM-31859). They are present but are relatively much shorter in Dendrolagus lumholzi and Thylogale hilliardieri. The apex of the hypoconid is not preserved in PM 4565, so the nature of its connec- tion with the midlink is uncertain. The anterior part of the midlink is located at the middle of the tooth and is parallel to the long axis of the tooth. It seems to have turned sharply lingually then anteriorly after leaving the hypoconid. The midlink is interrupted by a nar- row cleft at the middle of the transverse valley in PM 4565. The anterior quarter of the midlink appears to be a posterior spur from the protolophid. The midlink is interrupted in D. miilleri and D. hageni but in the latter species it is more like the fossil. The midlinks in both Recent species of Dorcopsis are located more labially than in the fossil and are not parallel to the long axis of the tooth. The junction of the forelink (paralophid) with protoconid is not preserved in PM 4565 as it is in NMV-P26418 and PM 4482— in both of which the link is located labial to the mid-axis of the tooth. In the latter it runs as a crest diagonally forward from the protoconid apex for nearly half its length before swinging to an antero-posterior alignment. In NMV-P26418 it runs almost straight diagonally to the procingulum. In PM 4482 the procingulum is rounded and more procumbent than in the others. In D. hageni and Dorcopsulus (AMNH 157267) the anterior cingulae of the anterior lower molars are more procumbent than those of the posterior molars. The an- terior part of the forelink in PM 4565 is located close to the midline TURNBULL AND LUNDELIUS: HAMILTON FAUNA 51 of the tooth and is nearly, but not quite, parallel to the long axis of the tooth. It joins an anterior cingular ridge at the mid-point of the anterior edge of the tooth. This junction is located more labially in both NMV-P26418 and PM 4482 and in the Recent species of Dor- copsis. The anterior cingular ridge in PM 4565 and NMV-P26418 swings labially in a smooth curve and merges with the base of the protoconid before reaching the labial edge of the tooth. In PM 4482 there is a small vertical groove between the protoconid and the lab- ial end of the procingulum. The base of the protoconid has a swelling adjacent to this groove. The lingual segment of the pro- cingulum extends lingually, then turns abruptly posteriorly and joins the base of the metaconid before reaching the lingual edge of the tooth. There are no pits in the enamel at the bases of the anterior faces of the protolophid and hypolophid as in D. hageni. The Hamilton Dorcopsis material shows many resemblances to Dorcopsoides fossilis from the Alcoota fauna of central Australia (Woodbume, 1967), particularly in the parastylar ridge and small lingual cingulae of the upper molars, the vertical lophids, similar structure and position of fore-link and mid-link of the lower molars, the absence of a lingual cingulum, and a slight development of a lab- ial cingulum in the lower molars. It is also similar in the relatively low crown height of the teeth. The metacone height of M^ of Dor- copsoides ranges from 2.8-3.8 mm. ; that of one M^- is 2.8 mm. Meta- cone heights of the two measurable upper molars of the Hamilton Dorcopsis are 2.3 and 2.5 mm. The Dorcopsis from Hamilton differs from the Alcoota species in: smaller size; the presence of a posterior basin and a lingual tubercle ahead of the postero-internal cusp on F^; absence on upper molars of mesostyle and of facets on paracone and metacone; absence on lower molars of longitudinal crests on protoconid and hypoconid; metaconid and entoconid not conical; and the anterior crests on meta- conid and entoconid are better developed. The Hamilton Dorcopsis also shows similarities to a specimen from the Awe fauna of New Guinea which was referred by Plane (1967) to the genus Dorcopsis. The Hamilton material differs in: smaller size; absence of spurs on the paracone and metacone which meet at the labial end of the median valley; absence of mesostyle; relatively narrower upper molars; and possibly weaker mid-links. In addition, the Hamilton specimens do not show strong posterior tapering, although this may result from the lack of an M^^ or M^ in the Hamilton material. Woodburne (1967) has suggested that this 52 FIELDIANA: GEOLOGY, VOLUME 19 New Guinea species may be closer to Dendrolagus than to the Dor- copsis-Dorcopsoides group. Measurements. — See Table 5. Table 5. — Measurements of teeth of Dorcopsis sp. from the Pliocene, Grange Burn, Hamilton fauna (in millimeters) Anterior Posterior Length width width Right pi (PM 4436) — — 3.9 Right upper molar (NMV-P26417) 5.5 4.7 4.2 Left upper molar (PM 4433) 5.4 4.4 4.5 Right upper molar (PM 4434) 5.8 4.3 4.4 One-half left upper molar (PM 4431) — 4.4 — One-half right upper molar (PM 4570) — 4.1 — One-half right upper molar (PM 16807) — 4.3 — Right lower molar (PM 4565) 5.4 3.9 3.8 Anterior one-half right lower molar (PM 4482) — 3.7 — Anterior one-half right lower molar CNMV-P26418) — 3.8 — Discussion. — We believe that the Dorcopsis sp. materials de- scribed above represent a new species, but in spite of the fact that four of them are complete teeth and that several others are very in- formative partial teeth, we are not naming a new taxon for we cannot be certain about associations. Most of the diagnostic features, which appear to be consistent for the Hamilton specimens may be found individually, or in other combinations, elsewhere within the macropo- dines. As examples of these diagnostic features, all of the upper molars preserving the pertinent parts have: 1. Paracone and metacone with antero-posterior cresting (except one which shows the other features the batch has in common) ; 2. Straight procingulum with a strong tendency to square off the antero-labial corner of the tooth so that the outer corner of the cingulum is even with the paracone and metacone; 3. Straight posterior crest of paracone; 4. Tendency in some teeth (presumably in the anterior molars) for paracone and metacone crests to join, forming a labial auxilliary "midlink"; 5. Absence of forelink. Fewer diagnostic features are to be found in the lower molars. In them: a. the metaconid and entoconid both have strong anterior crests that hook forward from the lingual end of the lophid, and b. the procingulum is notably broad rather than being narrowly sharp or pointed. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 53 On the basis of the parts of the animal available to us, the Hamil- ton Dorcopsis species appears to be just as advanced as any Recent species of the genus. See also further discussion in section on Thylogale (pp. 59-62). Subfamily Macropodinae There are 33 teeth or tooth fragments which are clearly assign- able to this subfamily. Eighteen of these teeth can be assigned to the genus Thylogale. On the basis of size and general morphology, the remainder (15 teeth) probably represent two or three taxa. These include Protemnodon as well as one or more other taxa whose generic affinities are uncertain. Thylogale Gray, 1837 Thylogale sp. indet. Materials. — Upper teeth: A worn partial left I^ (PM 4733, PI. XXV, Fig. C); an unworn partial right I^ (PM 4731, PI. XXV, Figs. A, B); a complete, unworn right molar, probably M^, M- or M^ (PM 16802) ; metaloph of an unworn left molar, probably M^, M^ or M^ (NMV-P26421) ; the unworn hypocone of a left molar, prob- ably M^ or MA (PM 4452); and a worn left hypocone (PM 4574). Lower teeth: Three left lyS, each only slightly worn (PM 4444, PL XXV, Fig. D; NMV-P26419; PM 16805); posterior two-thirds of a worn right P^ or P^ (PM 4562, PL XXVII, Fig. A) ; a complete unworn right dP^ (PM 4438, PL XXVI); the antero-labial quarter of a right dP^ (PM 4586) ; the posterior half of a left dP^ (PM 4483) ; an unworn, complete right My (PM 4564, PL XXIII, Figs. B-E); two moderately worn complete teeth, probably M^s (or M^s), (NMV-P26420 and PM 4576, PL XXIV, Figs. A-D);'^the unworn, anterior half of an M^ or M^ (PM 4432, PL XXIV, Figs. E and F) ; and a worn antero-lingual corner of an M^ or M^ (PM 4451). Descriptions.- — Upper teeth : The incisors are represented by the lateral sides of two teeth, a right U (PM 4731) and a left I^ (PM 4733). The I^ has a triangular lateral surface which is gently con- vex from root to crown. Very gentle rounded pillars are present at the anterior and posterior ends of the tooth. It is similar to the I^ of T. billiardieri but the pillars are less obvious and the tooth is slightly smaller. It is at about the same stage of wear as a specimen of T. billiardieri (FM 81526). 54 FIELDIANA: GEOLOGY, VOLUME 19 The right I^ preserves the lateral side of the tooth anterior to the notch on the outside of the tooth. The portion of the tooth which is preserved is gently convex from root to crown. The labial surface has a broad, low, rounded pillar running from the crown toward the root in the anterior quarter of the specimen. It is very similar to the anterior part of the I^ of Thylogale billiardieri in which the notch is located near the posterior end of the tooth. The blade, which is labial to the median groove on the crown of the tooth, is proportion- ately lower than in T. billiardieri. The complete right upper molar (PM 16802) is about the size and shape of the M^ or M^ of T. billiardieri or those teeth and M^ of T. stigmatica. The lophs are concave posteriorly and are gently rounded. The procingulum is broad and is connected to the para- cone by the sharp anterior crest of that cusp. There is no trace of a forelink. The paracone also has a sharp posterior ridge that extends rootward and slightly lingually almost to the bottom of the trans- verse valley where it swings abruptly lingually and disappears. The metacone is ridged in much the same way as the paracone. Its an- terior ridge is very prominent near the apex of the cusp and gradually dies out as it approaches the transverse valley, thus failing to form a complete labial link. The bottom is tightly V-shaped where it runs between this ridge and the posterior ridge of the paracone. The pos- terior ridge of the metacone is sharp near the apex of the cusp, but becomes rapidly reduced midway to the base of the tooth. In crown view it is a smooth continuation of the arc of the metaloph. The tip of the protocone is broken away. The midlink arises on the pos- tero-labial side of the protocone at a point about one-third the dis- tance from the apex to the base and extends to the middle of the tooth. There it is met by a nearly equal contribution from the meta- loph. The central part of the midlink rises to about one-half of the height of the lophs and is clefted in its upper part. The postcingulum is sharp, higher on the hypocone than the metacone and isolates a narrow V-bottomed cingular basin that opens labially. This tooth resembles the M^ and M^ of T. billiardieri in morphol- ogy more than those of any other macropodid, but differs in several minor ways. The procingulum is relatively broader than those of the lower molars of T. billiardieri, the midlink is less kinked, and the metaloph is more curved. It differs from M^ and M- of T. brunii and T. stigmatica in being more elongate and in having no trace of a forelink. It also differs from those teeth in T. billiardieri, as well as in the above-named spe- TURNBULL AND LUNDELIUS: HAMILTON FAUNA 55 cies, in the extensive development of the antero-labial crest of the metacone. Another specimen (NMV-P26421) consists of the metaloph. It is similar to PM 16802 in every respect except the extent of the ridg- ing on the posterior face of the metacone and the minor development of pitting on the posterior face of the metaloph. The posterior ridge of the metacone extends to the cingulum. The other two upper molar fragments are left hypocones. The larger (PM 4452) is slightly larger than its counterpart in the poste- rior molars of T. hilliardieri, and resembles them in the lateral taper- ing of the lophs, a weak crest to the postcingulum, and a midlink that almost reaches the crest of the metaloph. The hypocone of the smaller upper molar fragment (PM 4574) is about two-thirds the size of the other. The loph naiTows laterally toward its crest. It is quite worn so the original height of the con- nection of the midlink and the metaloph is uncertain. It shows no trace of a posterior ridge leading to a postcingulum. In this last character it is like the M^ or M^ of Thylogale stigmatica. The best two lower incisors are virtually identical to those of T. hilliardieri (FM 81526). They are lanceolate, broad, and gently curved along the long axis. The transverse curvature on the ventro- labial side extends through an arc of almost 90°. The central axial bulge is low and rounded, and is flanked on either side by a slightly recurved ridge. The ventro-medial ridge is slightly broader than the dorso-lateral one. The line that marks the rootward limit of the enamel on the lingual surface is W-shaped, with the enamel extend- ing farther rootward along the edges and with a spur of enamel run- ning rootward along the center of the tooth. On the labial side of NMV-P26419 near the dorsal ridge there is a U-shaped line marking the rootward extent of the enamel. The pattern of the enamel line on the lingual surface is like that in T. hilliardieri (FM 81526). PM 16805 adds nothing to our knowledge that is not conveyed by the others. The lower P^ or P^- (PM 4562, PI. XXVII, Fig. A) consists of the posterior two-thirds of a worn blade-like tooth that has a decided but gentle lingual turn at its posterior end just in front of the last cus- pule. There are three cuspules preserved, but it is doubtful that the blade was truly serrate in character when unworn, for the ridges and grooves are broad. There is a small, well-defined triangular inter- dental facet on the posterior end. This facet is situated labial to the sharp postero-lingual ridge that continues on from the axis of the 56 FIELDIANA: GEOLOGY, VOLUME 19 tooth at the apex of the last cuspule. The ridge runs steeply down to the base of the tooth. The tooth compares best with P^ or P^ of Thylogale billiardieri. It is larger than the Pt^ of that species (FM 81526) and the posterior inflected turn of the crest is slightly less marked. It resembles these teeth in T. stigmatica, too, but is nothing like those of T. hrunii with their high, well-formed distinct cusps. There are some similarities to the P^ and P^ of Setonyx hrachyurus and Dorcopsis mUlleri but these are not nearly as close as Thylogale bil- liardieri. The description of the dP^s (PM 4438, PM 4483, PM 4586) is based on the complete tooth (PM 4438, PI. XXVI) and, except as otherwise noted, applies to the other two. The tooth is very elon- gate, narrow anteriorly, with an elongate, flat but not procumbent anterior cingular shelf which is inclined lingually at about 45°. The metalophid is sharp, almost straight, and flat across the top. The hypolophid is slightly V-shaped with its concavity anterior and with the midpoint lower than either of the cusps. The protoconid has anterior and posterior crests. The anterior crest descends with pro- gressively increasing steepness anteriorly, flattens out suddenly, and goes forth and joins the anterior cingular ridge. The cingular ridge curves antero-lingually then postero-lingually and dies out anterior to, and slightly labial to the metaconid. It is separated from the metaconid by a short, straight antero-medially-directed valley which extends from the center of the base of the anterior side of the protolo- phid to the antero-labial corner of the tooth. In PM 4586 the antero- labial corner of the protoconid is sharply angled from its apex to its base rather than being rounded as in the complete tooth. The meta- conid has sharp anterior and posterior crests. The anterior one flexes slightly as it descends the cusp, and ends in a small cuspule located postero-lingual to the opening of the cingular valley. The posterior crest of the metaconid is straight, extends from the apex of that cusp to near the transverse valley, and swings labially as it disappears. The entoconid has an anterior crest which is continuous with the hypolophid. The crest descends the front edge of the cusp, swings labially, and dies out before reaching the central valley. This crest and the posterior one of the metalophid leave an open symmetrical U-shaped notch. The midlink is a low but sharp crest that extends postero-labially, then postero-lingually, from the apex of the protoconid. It descends the posterior face of the protoconid to the median valley, where it TURNBULL AND LUNDELIUS: HAMILTON FAUNA 57 turns slightly labially and climbs the anterior face of the hypoconid. It disappears about half-way up that cusp. In PM 4483 the mid- link appears to have joined the hypoconid at its apex. The hypoconid has an anterior crest which extends antero-labi- ally. The crest is labial to the midlink, which it does not join. The crest dies out about half-way to the base of the cusp. When viewed from the labial side, the metaconid and entoconid have a columnar appearance rather than a tapered one because of the crests. These specimens resemble the lower molars and dP^ of Dorcopsis and Thylogale in having crests on the cusps. The dP4 of Dorcop- sis has the protoconid and metaconid fused, and thus lacks a protolo- phid. In Wallahia eugenii the dP^ and all lower molars have a protolophid, but the cresting is less developed. In W. parryi the dP4 also has a protolophid, but the crests on the metaconid and ento- conid are joined to form a lingual link. The Hamilton specimens are most like Thylogale billiardieri in having a protolophid, a similarly tapered procingulum with the antero-lingual valley, and crested cusps. However, the posterior end of the dP4 of T. billiardieri is flatter than that of the Hamilton specimens. The My (PM 4564, PI. XXIII, Figs. B-E) is elongate, but not narrowed anteriorly. It has a large procumbent procingulum and the metalophid is rounded and nearly level, with a small notch at its midpoint. The hypolophid is broadly V-shaped and is slightly de- pressed at its midpoint. The anterior crest of the protoconid arises below the apex of the cusp, runs antero-lingually, then anteriorly and joins the procingulum. Lingual to the forelink is a broad, flat, cingu- lar shelf that is inclined posteriorly at about 20°. The cingulum itself extends lingually and postero-lingually but does not join the meta- conid. Labial to the forelink there is a narrow cingular shelf which extends from the anterior end of the tooth, in a rootward and poste- rior direction, to the base of the protoconid. The outer edge of this shelf does not merge directly with the protoconid. Instead, it is notched and then expanded into a cingular bulge on the antero- labial and labial sides of the protoconid. The metaconid has an an- terior crest which extends rootward and labially and disappears abruptly at about two-thirds of the distance to the base of the cusp. The protoconid and metaconid lack posterior crests. The entoconid and hypoconid have weak anterior crests. The midlink arises near the apex of the hypoconid, extends anteriorly and 58 FIELDIANA: GEOLOGY, VOLUME 19 lingually, then turns anteriroly and joins the metalophid labial to the midline of the tooth. At its lowest point the midlink rises about halfway from the base of the tooth to the crests of the lophs. There is a small transverse spur near the low point on the midlink. There is a small, low, horizontal postcingular ridge. The enamel is smooth. The two M^s that show wear (PM 4576, PI. XXIV, Figs. A-D; NMV-P26420) are so much alike that a single description will serve for both. They are relatively broad and rectangular in outline. The metalophid and hypolophid of both teeth are gently curved. The procingulum is large, procumbent, and rounded in outline and is sep- arated from the metaconid as in the My. The forelink arises at the apex of the protoconid. In NMV-P26420 it extends in a nearly straight line antero-lingually and joins the anterior edge of the cingu- lum. In PM 4576 the forelink extends antero-lingually, then ante- riorly and joins the cingulum. The position of the forelinks of both teeth is located closer to the midline of the tooth than in the My. Hence, the labial cingular shelf is larger than in the My, and the labial cingulum bulges. The lingual portion of the anterior cingular shelf is flat and is inclined posteriorly at about 20° as in the My. The mid- links of the two teeth differ from one another in the same way as the forelinks, that of NMV-P26420 being nearly straight and that of PM 4576 being bent lingually first, then running forward to the metalo- phid. The metaconid and entoconid have weak anterior crests which disappear about one-third of the way down the cusps. NMV-P26420 has a very small postcingular ridge as in the My, while PM 4576 has only a weak bulge in that region. The two fragmentary posterior molars (PM 4451 and PM 4432, PI. XXIV, Figs. E, F) consist of the anterior half of a tooth in the latter and the antero-lingual quarter of a tooth in PM 4451. The metalophids of both are curved and in PM 4432 the crest is wavy. The forelink is continuous with the metalophid and extends in a sig- moid curve antero-lingually then anteriorly, and joins the anterior edge of the tooth. The forelink is located close to the labial side of the tooth, and the labial cingular shelf is much reduced from the con- dition in the M^s discussed above. The procingulum is procumbent and broad, and the cingular shelf is flat and inclined posteriorly as in the other lower molars. It does not reach the metaconid. The meta- conid has a rounded anterior crest which gradually fades out half- way down the cusp. The junction of the midlink with the metalophid is located slightly labial to the midline of the tooth, and rises only about one-third of the distance up the back of the lophid. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 59 Measurements. — -See Table 6. Table 6. — Dimensions of the Hamilton teeth assigned to Thylogale sp. indet. and of a Recent specimen of T. billiardieri for comparison (in millimeters) Occlusal Anterior Posterior length crown crown (outer surface) height height Left I^ (PM 4733) 4.1 6.4 5.2 Right 13 (PM 4731) 4.8 (to notch) — — Length , tip to enamel Maximum line (labial side) width Left It (NMV-P26419) 16.2 6.4 Left It (FM 81526) — (not exposed) 6 . 6 Anterior Posterior Length width width M* (PM 16802) 6.8 5.2 4.6 M* (NMV-P26421) — — 4.5 dP, (PM 4438) 4.5 2.1 2.8 dP, (PM 4483) — — 2.8 Mt (PM 4564) 5.6 3.0 3.3 M, (PM 4576) 6.1 4.0 4.1 M^ (NMV-P26420) 5.8 3.9 4.0 Mg or M^ (PM 4432) — 4.8 — Discussion. — We believe that the materials referred to the genus Thylogale as Thylogale sp. indet., represent a single taxon which is new. As in the case of the Dorcopsis material, no formal name is being proposed because of some uncertainty regarding the association of some of the cheek teeth. The closest resemblances of the Hamilton Thylogale materials among Recent species are to Thylogale billiardieri, but they share many features with other species of the genus and with those of Dor- copsis. The cheek teeth, especially the molars, of these two genera have many characters in common and consistent differences are few and subtle, particularly in the lower teeth. They appear to be as follows: Upper cheek teeth of Dorcopsis are more nearly square than those of Thylogale which are more elongate. The dP^ of Dorcopsis has a parastyle which is larger relative to the size of the tooth than is that of Thylogale. Forelinks are present in all upper molars of the living species of Dorcopsis and virtually absent in Thyogale. 60 FIELDIANA: GEOLOGY, VOLUME 19 The procingulae of the molars of Dorcopsis extend broadly across nearly the entire front edges of the upper teeth, while in Thylogale they occupy only the central part of the anterior edges of the teeth. The dPi of Thylogale has a definite metalophid, while that of Dorcopsis has a single laterally compressed cusp in this position. The procingulae of all lower cheek teeth of Thylogale have flat- tened or only slightly basined occlusal surfaces which are not bounded lingually by a ridge, while in Dorcopsis they are basined and bounded by a lingual ridge. The lower cheek teeth of Dorcopsis usually have stronger ante- rior crests on their lingual cusps than are found in those teeth in Thylogale, but wear may quickly obscure this difference. Because so many dental characters in these genera show inter- species variation and variation within the tooth row, and because the size ranges of the various species overlap broadly, the identification of individual teeth is difficult. In an attempt to solve this problem, bivariate scatter diagrams have been made (Appendix Graphs F, G, H, I). These plot length vs. anterior width for the molariform teeth (dP| through M|) for D. mulleri, D. hageni, Dorcopsulus sp. (prob- ably D. macleayi), Thylogale hrunii, T. stigmatica, and T. billiardieri. These scatter diagrams show the progressive changes in size and pro- portions of the teeth along the tooth row for the small samples avail- able to us. For T. stigmatica, the largest species sample with nine individuals, the amount of variation and overlap gives a suggestion of what can be expected in these genera of wallabies. As a means of further broadening the comparison tooth measures of two other sam- ples are also presented (eight individuals of Dorcopsoides fossilis (Woodbume, 1967) and eight of Setonyx brachyurus (lower teeth only) . These scatter diagrams very clearly show that in the living species of Thylogale and Dorcopsis (and Dorcopsulus sp.), there is a progres- sive increase in the length of the molar teeth from M\ to M|. In D. hageni the M^ may be about equal to M^. A similar, but somewhat less marked, increase in anterior widths also extends back to Mf in the three genera. In Thylogale these trends continue to M^, but in Dorcopsis only the width measurements show increase of M^ over Mj. In Setonyx brachyurus M^ is shorter than M^, and widths are equal. In the rear of the upper molar series the trend changes: M^ and M^^ are nearly the same size in Thylogale. The one Dorcopsis available having M^, has that tooth somewhat longer and decidedly narrower than its M^. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 61 Placement of the measurements of the Hamilton specimens on these scatter diagrams provides no clearcut indication of their affin- ities. However, it does give us one more means of comparison by which we can both 1) assess the reasonableness of the assignments that were based in part upon other criteria, and 2) gain a better idea of the dentitions they represent. Looking first at the upper molariform teeth, we see that the three which we have assigned to Dorcopsis sp. (PM 4433, PM 4434, NMV- P26417) plot to form a cluster that can be interpreted in different ways. The cluster is tight enough that the specimens all could be examples of the same tooth, but at the same time it is open enough, and intermediate enough, that they could also represent either the dP^ or M^ of a species close to D. mulleri or D. hageni. Tedford (personal communication, 1968) has suggested that PM 4433 may be a dP^ or M-L, but we consider that the great breadth of the procingu- lum makes the dP^ interpretation quite unlikely. We therefore con- sider this tooth and PM 4434 (for the same reason) to be molars, probably MJ^s. NMV-P26417 with its broad, but tightly-appressed (i.e., not procumbent) procingulum, appears to be one of the more posterior molars, yet its length and width proportions fit an MJ- in- terpretation best. One other possible interpretation is that these teeth may be the M^ or M^ of a form near Dorcopsulus. As a Dor- copsis near D. mulleri or D. hageni, the Hamilton species would be characterized by having relatively narrow upper molars, but if its ties should be with Dorcopsulus, then they would be characterized by having more elongate proportions. Whatever the case, they are somewhat intermediate in terms of proportions. The one complete upper molar referred by us to Thylogale sp. as probably M^, M^, or M^ (PM 16802) compares quite well in its length-width proportions with the M^ of T. stigmatica. Unfortunately, the M^ of T. billiardi- eri is imknown to us, but if the trends seen in the pattern of propor- tion changes along the tooth row follow those of the other species, PM 16802 should compare equally well with it, too. In fact, the comparison is reasonably good with the M^ in those species. Turning to the lower teeth, it may be seen that the only complete Hamilton specimen which we have assigned to Dorcopsis sp. (PM 4565) is shorter than the others (PM 4564, NMV-P26420, PM 4576) which are identified as Thylogale sp., and in size, proportions, and morphology it is closest to the M^ of D. mulleri. The three Thylogale molars and the dP^ (PM 4438) are all proportioned so that they 62 FIELDIANA: GEOLOGY, VOLUME 19 combine to make a pattern that is most reasonably interpreted as that of a species near to T. billiardieri. The morphology of PM 4438 and PM 4564 indicates that both are teeth belonging to the anterior portion of a cheek tooth series, but the presence of a distinct metalophid on PM 4438 appears to rule out any possibility of its being the dP^ of a Dorcopsis, in which genus this lophid is not developed. It could well be referred to Thylogale, however, because in the species of that genus the dP4 not only has a definite metalophid, but has a broad and well-developed one. The relatively very long and narrow proportions of PM 4564 and its labi- ally open, flat-bottomed procingulum make it hard to consider it to be anything other than the My of a species of Thylogale. The dP4 or Mx of Dorcopsis both have a more rounded valley floor that is cut off (without a lingual exit) by a ridge. The other two teeth (NMV-P26420 and PM 4576) are of a size and proportion that fits with either the M^ (or possibly M^) of a Thylogale, or the My or Mg of a Dorcopsis, but their weak cresting fits the Thylogale pattern better than that of Dorcopsis. Thus the morphology and size of all of the complete lower molari- form teeth, except PM 4565, appear to be consistent with their be- longing to one taxon, and that to be a species of the genus Thylogale. The Hamilton Thylogale material resembles T. billiardieri in a large number of characters such as general size, structure of the up- per and lower incisors (so far as they are known), and the structure of the P^. It shows some differences which indicate that it is not assignable to that species. They are: the labial inclination of the procingulum shelf of dPj, the lesser development in breadth of the metalophid of the dP^, the slightly larger (absolutely, but not rela- tively) size of the molars, and the relatively broader upper procingu- lae, which are somewhat intermediate in their development between those in the extant species of Dorcopsis and Thylogale. Macropodinae incertae sedis There are a number of isolated macropodine teeth and fragments of teeth which cannot be identified at present. They fall roughly into two size groups, large and small. The large represent animals in the Macropus size range and the small are from animals in the wallaby size range. They are grouped on this basis for description, but this does not imply close taxonomic affinity. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 63 Large-sized macropodines Material. — A complete, nearly unworn left molariform tooth, probably a dP^ (PM 4429, PI. XXVIII, Figs. A-E) ; antero-lingual comer, worn right upper molar (PM 4498, PI. XXVIII, Fig. F) ; an- tero-lingual comer, worn left lower molar (PM 4441) ; posterior half of a slightly worn right lower molar (NMV-P26422) ; posterior edge of a lower molar showing wear facet of hypolophid and a large inter- dental facet (PM 4469) ; two insignificant molar fragments (PM 4435; PM 4732), and the postero-lingual corner of a very worn hypolophid (PM4715). Description. — The complete upper molar or dP^ (PM 4429) is somewhat larger and lower crowned than the upper molars of Wal- labia agilis but is relatively wider. It is similar in proportion to those of W. bicolor, but is lower crowned. The protoloph and metaloph are curved. The paracone and metacone have distinct, sharp labial ridges on the anterior and posterior sides. The anterior ridge of the paracone joins the labial edge of the procingulum. The posterior ridge of the paracone extends in a broad curve to the base of the cusp where it branches. The lingual branch quickly dies out and the labial one meets but does not join an independent short spur on the base of the metacone. The anterior ridge of the metacone curves lingually toward the base of the cusp and dies out before reaching the central valley. Several genera of Recent macropodines have upper molars with these labial crests or ridges (W. bicolor, Setonyx brachyurus, Thylogale, Dorcopsis) and at least three extinct forms do also (Sthenu- rus tindalei (Tedford, 1966), and the Dorcopsis and Thylogale in this fauna), but none shows quite the development of the condition as pre- sented by PM 4429. Upper molars of the extinct Protemnodon have these ridges reduced and Prionotemnus lacks them (Stirton, 1955). The midlink is joined to the apex of the protocone. It extends in a straight line postero-labially to the center of the tooth, then turns posteriorly and joins the metaloph. It does not reach the crest of the metaloph. The highly crested part of the midlink is about equally formed by the protoconal spur and the basal anterior metaloph spur as in Setonyx, Dorcopsis, Dendrolagus, and Lagorchestes. In the cen- tral valley the midlink has a lingually-directed kink. The labial con- cavity in the midlink formed by this kink is filled by a small pillar. The midlink and the postero-labial crest of the paracone form a basin in the central part of the tooth between the cross lophs. Such a basin is a feature of the dP- of Protemnodon and is found in upper molars of W. bicolor, Setonyx, Dorcopsis, Thylogale billiardieri. Den- 64 FIELDIANA: GEOLOGY, VOLUME 19 drolagus, and Petrogale, but in these species the labial border of the basin is formed by the labial crests of the paracone and metacone and is often incomplete. The procingulum is slightly basined. The anterior edge is con- nected labially to an anterior crest of the paracone. The widest part of the cingulum is located near its lingual end. The anterior border joins the protocone before reaching the lingual edge of the tooth. There is only a trace of a forelink. A small postlink extends rootward and labially from the hypo- cone, becomes horizontal across the base of the tooth and ends at the base of the metacone. It forms a small pit-like cingular basin, A short ridge extends rootward on the back of the metacone but does not quite reach the postlink. The tooth shows the beginning of wear. There is a small trans- versely elongated interdental facet on the middle of the anterior edge of the procingulum. There is another equal-sized abrasion facet ad- jacent to this on the occlusal surface of the anterior edge of the pro- cingulum. The crest of the protoloph has a slightly oblique (nearly vertical) facet on its anterior edge. The metaloph shows no sign of wear. This complete upper tooth has many resemblances to the upper molars of Dorcopsis, Dendrolagus, W. bicolor, and Thylogale hilliardi- eri. It differs from all of them in its much greater size and in the combination of characters, many of which are found in one or more of the Recent species. The best comparisons are with the dP^ of Pro- temnodon hrehus (Tedford 1967, fig. 25B, table 36) and Protemnodon sp. (Stirton 1963, fig. 6A and p. 120) in terms of form and proportions as well as size. The dP-^ of Protemnodon is very molariform as is the Hamilton specimen, and these close resemblances suggest that PM 4424 is probably a dP^ of a species of Protemnodon. Measurements. — The measurements in millimeters for PM 4429 are: length 11.3, anterior width 10.1, posterior width 10.6. Description. — The half lower molar (NMV-P26422) has a hypolo- phid that is as large as that of the lower molars of Macropus or Pro- temnodon. In crown view it is broadly curved and lowest in the cen- ter, and lacks the strong root to crest taper that is characteristic of Macropus. The midlink arises just below the apex of the hypoconid and is not confluent with the hypolophid. The portion of the mid- link that is preserved extends rootward and changes its direction along a gentle curve from antero-labially to anteriorly at its broken TURNBULL AND LUNDELIUS: HAMILTON FAUNA 65 edge. There is a very faint, gently rounded ridge that extends from the entoconid antero-lingually to the broken edge of the tooth. It is confluent with the hypolophid, but does not join the midlink. There is a prominent rounded bulge in the position of the postcingulum. This tooth differs from the lower molars of Macropus and Wal- lahia in lacking the strong lateral taper toward the crest of the lophid in unworn teeth and in the confluence of a strong midlink and the hypolophid. It is as large as that in some species of Protemnodon, but differs in having a slightly more curved hypolophid. It is very similar in size (posterior width 8.5 mm.) and general appearance to the lower molars figured by Bartholomai (1967) and referred by him to Troposodon minor. Features responsible for the similarity of ap- pearance are: the degree of curvature of the hypolophid, the height and curvature of the midlink, and the presence on the anterior face of the hypolophid of a faint accessory ridge which ends close to, but not joined to the midlink. This last character is apparently better developed in Troposodon than in the Hamilton tooth. No orna- mented shelf is formed in the median valley of the latter. The remaining fragmentary large macropodine teeth convey only scant information. The worn right upper molar (PM 4498) consists of the antero-lingual corner of a tooth that is about the same size as PM 4429. It may have been slightly more massive and higher crowned when unworn than the complete tooth (PI. XXVIII, Fig. F). In the worn left lower molar (PM 4441) the enamel of the metaconid has been breached, and its procingulum forms a transverse "lophid" which is connected to the protolophid by a low broad forelink. Discussion. — The large indeterminate macropodine material prob- ably represents two taxa. One, represented by the large upper mo- lariform tooth (PM 4429) , has many dental resemblances to upper molars of Dorcopsis and Dendrolagus, but shows much more similarity to the dP^ of Protemnodon. Its affinities probably lie with that genus. The other, represented by the hypolophid of a large lower molar (NMV-P26422), has many characters in common with the extinct genera Protemnodon and Troposodon. It is doubtful that these two large teeth could represent the same taxon. In general, upper molariform teeth that have the kind of cresting seen in PM 4429 should be accompanied by lower molars with stronger cresting on the labial cuspids than is seen in NMV- P26422. 66 FIELDIANA: GEOLOGY, VOLUME 19 Small-sized macropodines Material and descriptions. — This category consists of wallaby tooth fragments identifiable only to the subfamily level. They are simply listed here with a few brief comments given in some cases. Several are illustrated. PM 4559- — anterior edge of a small, very highly crested, narrow pre- molar blade somewhat like that of the P^ of Dorcopsis mulleri (AMNH 108044), and in other ways like that of the deeply clef ted blade of Thylogale hrunii (PM 31864), (PI. VI, Fig. A). PM 4558 — posterior half of a small, high crested, serrate premolar blade. This could be the back half of PM 4559, but there is no good contact. If such were the case, a tooth quite like the P^ of the men- tioned D. mulleri would result, but it would be smaller. PM 4583 — posterior edge of a hypolophid of a small wallaby, pos- sibly Thylogale. PM 4734- — fragment of a lateral upper incisor. PM 4578^ — piece of a lower incisor blade, possibly Thylogale. PM 4448- — a worn ?protocone, possibly referable to Thylogale. PM 4593- — posterior end of a premolar blade similar to the P^ or P^ of Dorcopsis and Thylogale, but not exactly comparable (PL XXIX, Fig. A). Family Diprotodontidae Palorchestes Owen, 1874 Palorchestes cf. painei Woodburne, 1967 Material.— A left lower molar, probably Mj or M^ (PM 16801, PI. XXX). Description.- — The tooth is elongate, rounded in front, and rounded but somewhat flattened in back. It is weakly but noticeably con- stricted in the middle when viewed in the occlusal aspect. The crown is relatively low: the ratio of protolophid height to tooth length is 0.65. Comparative values of 0.74 or 0.73 were found for a slightly worn M^ and an unworn Mj (?), respectively, for some Wellington specimens of P. azael (PM 1580; PM 1665). The more worn MjS in the Wellington specimens gave values of 0.68 and 0.60. The crests of the lophids are concave forward and they cross the tooth somewhat diagonally; the labial edge is more anterior than the lingual. The lingual sides of the lophids rise almost straight up from the tooth TURNBULL AND LUNDELIUS: HAMILTON FAUNA 67 base, but they are not flattened in this plane as is often the case with those of P. azael. The labial sides of the lophids are rounded, and they do not stand quite as straight up as the lingual sides. There is a massive, rugose, weakly crested anterior ridge that runs from the apex of the protoconid down to the level of the cingulum. The pro- cingulum bulges up rapidly on the labial side of the front edge of the tooth to meet this ridge, and continues in a lingual direction, gradu- ally approaching the crown base. The junction of the "forelink" ridge from the protoconid with the cingulum is weak and the cingu- lum protrudes slightly where it crosses in front of the ridge. Thus, the two structures do not merge fully. The protolophid shows the faint beginning of wear across its en- tire width in a facet inclined at about 45° up from the occlusal plane of the tooth. The hypolophid is narrower apically than the protolo- phid and its ridge turns smoothly forward as it crosses the hypoconid. The ridge then swings diagonally, forward, and lingually, so as to descend into the transverse valley as a stout midlink. It joins a much weaker, straight, rapidly descending ridge from the protolophid above the midpoint of the floor of the valley. The disproportionately small contribution of this protolophid part of the midlink to the whole is a notable feature of the tooth. The postlink, which is char- acteristic for the diprotodonts, is well developed and joints a stout postcingulum. The postlink begins as a faint ridge that runs down from the middle of the hypolophid to the level of the cingulum where it rapidly enlarges as it levels out and swings backward to meet the cingulum. The cingulum rises rapidly from each side toward this junction. The postcingulum is perhaps a little less well developed than in P. azael. Measurements. — See Table 7. Discussion. — The smaller size, relatively low crown height and somewhat weaker links and pro- and postcingulae suggest a primi- tiveness for this specimen in comparison with P. azael and indi- cate a close relationship to P. parvus and P. painei. The Hamilton specimen is only slightly larger than the largest My of P. painei re- ported by Woodburne (1967). It also resembles the Mj of P. painei in the lack of a well-developed labial cingulum. The transverse val- ley has an open V-shaped profile when viewed from the labial side. The calculation of ratios of protolophid height/tooth length for P. painei from Woodbuvne's (1987) data is impossible for Mj and M.2, and not entirely satisfactory for M^ and M^, but the results suggest c o 1-1 1 ^ 05 <>] 00 1 (M CO 1— 1 X X X X X fO o o o i-H (M o i-H OR i-H 00 '-' CO «3 T— 1 t •^ c-. o tXD S >> Eh ' ^-^ ,— , Oi O T-l CO ?D CO ;o o o 00 00 OS oi CO CO t- t~ t- (M CO CO (iH Ech EiH EiH Em P^ O lO 00 i;d in «D ^1^ 68 TURNBULL AND LUNDELIUS: HAMILTON FAUNA 69 a maximum expected value of about .52. This is somewhat lower than the value of .65 for the Hamilton specimen. Infraclass Eutheria Order Chiroptera Suborder Microchiroptera Incertae sedis at all lower taxonomic levels Material. — One left lower molar in a mandibular fragment (PM 4458, PI. XXIX, Fig. H). Description. — This specimen is the only representative of the pla- cental mammals known from the Hamilton fauna. It was mistakenly considered to be a phascogaline in our earlier report (Tumbull et al., 1965) but our detailed comparisons indicate chiropteran affinities. The molar is probably an Mij or M^ and is about the size of these teeth in the living Tadarida (Nyctinomus) australis or Nyctophilus goeffroyi. Its length is 1.41 mm. ; anterior width, 0.82 mm. ; and pos- terior width, 0.80 mm. The trigonid is higher than the talonid, but not as much higher as it is in the phascogalines. The three cusps of the trigonid form a triangle in which the paraconid and metaconid are closer to each other than ei ther is to the protoconid. The profiles of the ridges from protoconid to paraconid, and from protoconid to metaconid are V- shaped, but lack the deep narrow clefts seen in the teeth of phascoga- lines. The three cusps of the trigonid are not as high above the trigo- nid basin (because basin is high relative to crown base) as are those of the phascogalines. The talonid basin is bounded lingually by a blade-like ridge that connects the entoconid with the metaconid. The entoconid is rela- tively larger than in the lower molars of the phascogalines. There is a well-developed hypoconid, and a small but distinct hypoconulid which lies close to the entoconid at its base. Talonid and trigonid are very close to the same size. There is a well-developed and con- tinuous cingulum on the anterior, labial, and posterior sides of the tooth. In this it contrasts with the phascogalines, in which the cingu- lum is reduced labially. Dental features such as those mentioned in the preceding para- graphs serve adequately to eliminate the Phascogalinae from con- sideration, and to aflEirm the microchiropteran affinities of the speci- men. However, within the Microchiroptera we've not been able to 70 FIELDIANA: GEOLOGY, VOLUME 19 determine with any certainty which of several famihes of the sub- order the specimen belongs to. We suspect that there are too many genera in which the lower molars have the same basic form, for an identification based upon a single lower tooth to be clearcut. The two mentioned species, Tadarida (Nyctinomus) australis and Nyc- tophilus goeffroyi, both have lower molars that are very similar, and they are classified into two different families, Molossidae and Ves- pertilionidae (Nyctophilinae), respectively. Also several other gen- era of vespertilines (Myotis, Pipistrellus, Eptesicus, Nycticeus, and Chalinolobus) all have lower molars with a morphology that is very similar to one another and to the Hamilton tooth. (Specimens of Myotis and Pipistrellus available for comparison all have the hypo- conulid either too well developed or too set apart from the entoconid. This would appear to rule out those genera, but the sample is too small to be certain of this.) The latter three vespertiline genera at least, and one other genus, Hipposideros — this from yet another fam- ily (Hipposideridae), also have lower molars with this same basic morphology. Thus, we concluded that we cannot make a family assignment for the Hamilton specimen. Class indeterminate, probably Mammalia Incertae sedis at all lower taxonomic levels Material— Five enamel caps (NMV-P26425, PM 4468, PM 4466, PM 4465, and PM 4599, PI. XXXI, Figs. A-F). Descriptions. — Four of the five specimens consist of what appear to be complete enamel caps. The fifth specimen is incomplete and has been sectioned to allow an examination of the structure of the enamel. The teeth consist of elliptical enamel caps each with a closed cen- tral basin. The basins are shallow with smooth rounded bottoms. They fall reasonably well into three gi^oups on the basis of size and morphology. The two middle-sized specimens (NMV-P26425 and PM 4465) each have a predominant, low, rounded cusp on one end. They have a variable number of smaller low cusps on the rim that borders the central basin. The smallest of the complete teeth (PM 4466) has its sole cusp on one end and has a smooth rim. One of the two middle-sized ones (NMV-P26425) has a relatively large cusp on each end, the larger of which appears to have had the enamel breached by wear at the apex. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 71 The largest tooth of the four (PM 4468) has no bulbous cusps at its ends, but has a higher pointed one at one end and a number of small cusps on the rim of the central basin. None of the teeth shows any sign of an interdental facet which would give some indication of the orientation in the jaw. No indi- cation of the orientation can be obtained from what appears to be wear on one of the medium-sized teeth (N WM-P26425) . The sections of the incomplete tooth show an enamel structure which appears to be most like that seen in mammals. There are elongated enamel prisms to be seen in the transverse tooth section, oriented so that their full length is exposed. In the tangential tooth section one area on the extreme left side shows others similarly cut, but most of the section shows the prisms cut across nearly perpen- dicular to their axes. The enamel prisms are relatively small; prism diameter being of the order of 3 to 5 micra. Most prisms are moderately twisted along the long axis, and there is a slight tendency for them to intertwine weakly. The twisting is most pronounced near the base (i.e., near the dentine-enamel boundary). In this respect the prism histology conforms to that of the relatively uncomplicated, untwisted pattern as is shown by Peyer (1968, Pis. 77A, B) for a mole and a bat. This would indicate, according to Peyer's interpretation, a relatively prim- itive type of mammalian enamel. In cross-section the prisms show much the same appearance as those of Homo as figured by Keil (1966, Fig. 51) or Lehner and Plenk (1936, Fig. 42). However, there seems to be a poorer alignment than is seen in the first of these examples. None of the lower vertebrates shows a comparable enamel struc- ture. Prismatic enamel is known for reptiles and amphibians, but in them it is not nearly as highly organized as in the mammals. Within the fishes, enamel occurs in crossopterygians and lungfish, but it does not have a prismatic structure. In some holosteans and in teleosteans enamel occurs as a collar or ring surrounding the tooth crown without forming the tip of the crown which consists of a highly specialized kind of dentine. The combination of gross morphology and enamel structure indi- cates that the five enamel caps are most probably mammalian teeth, but any assignment to a lower taxonomic group would be purely con- jectural at this time. They look most like teeth of microcleptids, some bats, or some small phalangerids with very atrophied last mo- lars. They bear very little resemblance to the true teeth of Ornitho- 72 FIELDIANA: GEOLOGY, VOLUME 19 rhynchus. The teeth of the latter are multicuspate with ridges break- ing the teeth up into two or more basins. Recognition of their affin- ities must await the discovery of additional, and probably more com- plete material. Miscellaneous Scraps and Fragments This category consists of the final residue of tooth fragments. Most of them are probably too fragmentary ever to be identified, but some show morphological features which may make them inter- pretable at some future date. Several hundred smaller chips with no structural features (grouped under PM 4597 and PM 4598) are not considered. The listed materials carry comments as to our tentative identifications. PM 4595 — a conical chip bearing four longitudinal ridges (PI. XXIX, Fig. D). PM 4560 and NMV-P26424- — pieces from the center of a high serrate premolar blade (PI. XXIX, Figs. E, G), possibly of a potoroine or macropodine. PM 4730^ — two cusps, presumably the anterior-most ones of a minute premolar blade, possibly a potoroine (PI. XXIX, Fig. C). PM 4740 — fragment that seems to have been a posterior end of a premolar blade, badly worn and with a good interdental facet. PM 4447 — indeterminate mammal tooth fragment. PM 4491 — indeterminate mammal tooth fragment. PM 4494 — indeterminate mammal tooth fragment. PM 4497 — indeterminate mammal tooth fragment. PM 4577- — indeterminate mammal tooth fragment. PM 4580 — indeterminate mammal tooth fragment (PL XXVII, Fig. C). PM 4581- — upper incisor, possibly potoroine or macropodine. PM 4582 — very worn tooth, possibly macropodid molar (PI. XXIX, Fig. B). PM 4714 — fragment of a lower incisor, possibly macropodid or pha- langerid. PM 4716 — indeterminate molar fragment, probably pseudocheirine. PM 4736 — fragment from the center of a premolar blade, possibly potoroine. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 73 PM 4737^ — indeterminate tooth fragment, probably macropodid. PM 4738— indeterminate mammal tooth fragment. PM 4739— indeterminate mammal tooth fragment. PM 4741- — indeterminate mammal tooth fragment. PM 4742^ — indeterminate mammal tooth fragment, probably macrop- odid. PM 4760A-C^ — indeterminate mammal tooth fragment. PM 4450 — complete toebone, well preserved, probably mammalian. PM 4743 — scrap of bone, well preserved. PM 4744 — scrap of bone, well preserved, possibly fish. PM 4757 — indeterminate small mammal, proximal end ulna. INTERPRETATION OF THE HAMILTON FAUNA Faunal Composition The Hamilton fauna as presently known is composed almost en- tirely of teeth and fragments of teeth belonging to the Mammalia. Bone is rarely preserved in the sediment containing the fauna and light skeletons would not be likely to preserve. This accounts for most of the lack of lower vertebrates since many have light, fragile skeletons. The lack of resistant structures, such as teeth of fish and reptiles, is probably either an indication that the wrong niches are being sampled, or that their structures are less resistant than thought. The absence of larger reptiles, such as varanids, is probably an acci- dent of sampling. Despite the presence of a considerable number of taxa (18) show- ing a moderate ecological breadth, there remain several obvious de- ficiencies in the representation of the faunal elements. Large-sized species are very poorly represented or are absent, and dasyurids and rodents are totally lacking. The dasyurid absence is surely an acci- dent of sampling. Table 8 shows that a small number of individuals (43) could account for the materials collected to date. Thylacoleo- nids, dasyurids, and large herbivores probably had smaller popula- tion sizes than did the small herbivores, and they would be more likely to be missed in small samples. It is also possible, but more difficult, to attribute rodent absence to an accident of sampling. Whenever comparable collecting tech- niques have been used on Pleistocene sediments in Australia and on Pleistocene or Tertiary sediments on other continents, rodent re- mains are almost always present if small-sized animal remains are found. In addition, rodents make up 37-53 per cent of the taxa of small mammals of Pleistocene faunas from Western Australian caves and 83-91 per cent of the individuals of those faunas where such data are known (Lundelius, 1960). The comparable figures for the small dasyurids are 27-29 per cent and 6-11 per cent, respectively. Thus, even allowing for considerable environmental differences between the 74 Table 8. — Composition of the Hamilton Fauna Taxon Class: Subclass: Order: Minimum no. Superfamily: Family: Subfamily: No. of no. of Genus and species specimens individuals 1 Mammalia: Metatheria: O. Marsupialia: Perameloidea: Paramelidae: Genus and species indet. Phalangeroidea: Phalangeridae: Phalangerinae: 2 Phalanger cf. gymnotis 3 Trichosurus sp. indet. - IPhalanger, or 1 Trichosurus, or genus indet. - Incertae sedis, near Phalanger or Trichosurus 4 Incertae sedis — genus and sp. indet. (minute) Phalangeroidea: Phalangeridae: Burramyinae: 5 Burramys sp. indet. Phascolarctidae : Pseud ocheirinae : 6 Pseudokoala erlita n. genus and sp. cf. Pseudokoala erlita 7 Pseudocheirus stirtoni n. sp. 8 Pseudocheirus marshalli n. sp. - Pseudocheirus sp. indet. Macropodidae: Potoroinae: 9 Incertae sedis near Aepyprymnus 10 Incertae sedis near Hypsiprymnodon 11 Incertae sedis near Propleopus oscillans 12 Dorcopsis sp. indet. Macropididae: Macropodinae: 13 Thylogale sp. indet. 14 Incertae sedis (large form-A near Protemnodon) 15 Incertae sedis (large form-B) — Incertae sedis (?large A, or B, or ?) — Incertae sedis (? Dorcopsis or Thylogale, or ?) Diprotodontidae: Palorchestinae: 16 Palorchestes near P. painei Mammalia: Eutheria: O. Chiroptera: M icr ochir op ter a : 17 Incertae sedis, genus and sp. indet. ?Mammalia: Subclass uncertain: 18 Incertae sedis at all taxonomic levels Totals 18 mammalian species Miscellaneous scraps 1 2 2 or 3 1 or 2 2 8 2 4 1 32 3 18 2 or 3 14 4 1 1 1 1 1 1 16 3 or 4 18 3 or 4 2 1 3 2 3 ?2 7 2 or 3 1 1 5 ?1 i 155 43 5 27 6-8 182 49 75 76 FIELDIANA: GEOLOGY, VOLUME 19 Western Australian caves with their probable owl pellet bias (see Schram and Turnbull, in press) and the deposit in the Hamilton area, it would be unlikely that the rodents would be missed by a sampling accident. The absence may be real, and this possibility deserves consideration. If real, the absence of rodents can be interpreted in at least two ways. It is possible that at the time of accumulation of the Hamilton sample rodents were excluded from the particular environmental situation from which the fauna was drawn. This seems unlikely be- cause rodents generally have representatives in most environments. To date no rodents have been reported from pre-Pleistocene de- posits in continental Australia, although one has recently been re- ported from Awe fauna of New Guinea (Plane, 1967). The Awe fauna has been dated radiometrically and is somewhat older than the Hamilton fauna. Although it is now certain that some rodents were in New Guinea by Late Pliocene time, it is possible that they had not entered continental Australia. The Torres Strait connection may not have been continuously available and their entrance into Australia may have been delayed. This would mean that the greater part, at least, of the Australian rodent radiation has taken place after Late Pliocene time (in 4.35 million years), rather than in post Late Miocene time (Simpson, 1961; Tate, 1951). Comparisons with Other Faunas Detailed comparisons of the Hamilton fauna with other fossil mammal faunas from Australia will be deferred until the material collected in 1966-67 is studied. However, because the Hamilton fauna has been dated radiometrically, and thus offers a key to the correlation of other Australian faunas with the Lyellian epochs, some preliminary comparisons are given here. A comparison of the Hamilton Dorcopsis and Palorchestes mate- rial with Pleistocene representatives of these genera shows the Hamil- ton material to be clearly more primitive in both size and crown height. The Hamilton Palorchestes is distinctly smaller and lower crowned than P. parvus from the Chinchilla fauna of Queensland. The Chinchilla fauna has been considered to be Late Pliocene in age (Woods, 1956, 1960; Plane, 1967; Woodburne, 1967). Tedford (1966) questioned the age assignment of the Chinchilla to the Pliocene. His use of "later Pliocene or early Pleistocene" reflects this uncertainty. He has also pointed out that in southwestern Victoria, P. parvus lived well into the Pleistocene as evidenced by its occurrence in the TURNBULL AND LUNDELIUS: HAMILTON FAUNA 77 Strathdownie fissures. The more primitive condition of the Hamil- ton specimen also relates to the age assignment of the Chinchilla fauna. The radiometric date 4.35 x 10^ years for the Hamilton fauna places it in the Late Pliocene (Evernden et al., 1964). At the present time it is uncertain as to whether the difference in the Palorchestes material indicates a younger age for the Chinchilla fauna, perhaps early Pleistocene, or an ecological difference. The Hamilton Palorchestes is very similar to P. painei from the Alcoota fauna, and the Hamilton Dorcopsis is similar to the Alcoota Dorcopsoides in crown heights of the molars. Also, when sizes and proportions of the cheek teeth in these species are compared (Ap- pendix Graphs G, I, J) it may be seen that the Palorchestes materials from the two faunas are nearly identical, while the Hamilton Dor- copsis shows somewhat greater differences from Dorcopsoides. These similarities suggest that the Hamilton and Alcoota faunas may be closer together in age than was originally thought (Woodbume, 1967; Stirton et al., 1968; and see also Stirton et al., 1961) and that the Alcoota fauna may be somewhat younger than late Miocene or early Pliocene. The Hamilton fauna has been shown to be younger than the Awe fauna of New Guinea, but, unfortunately, there is very little in com- mon taxonomically between the two faunas which would allow some evaluation of the amount of ecological control on persistence of more primitive types in the lower latitudes (Plane, 1967). A comparison of the Hamilton Dorcopsis with the Awe form is not particularly helpful, because of the size discrepancy between the two species and because the best Awe specimen is at a more advanced wear stage than are most of the Hamilton teeth. Ecological Interpretation Of the 18 mammalian taxa in the Hamilton fauna, four (Dorcop- sis sp., Phalanger sp., the potoroine similar to Hypsiprymnodon, and the peramelid) are closely related to forms which today are found in wet sclerophyll or rain forest areas in northern Australia or New Guinea, or both (Tate, 1945a; 1948a; 1948b; Keast, 1961; Trough- ton, 1962; Marlow, 1962). The presence of three species that are either assignable to Pseudocheirus, or to a closely related genus (Pseu- dokoala), is also suggestive of a wet forest environment for the fauna. The only places today where more than two species of Pseudocheirus are found are wet forest areas in northern Queensland and New Guinea (Tate, 1945b). Three other taxa (Thylogale sp., Burramys 78 FIELDIANA: GEOLOGY, VOLUME 19 sp., and the Cercartetus-Yike phalanger) probably indicate forest con- ditions, too, although one species of the latter genus is now distrib- uted through dry scrub forest (Wakefield, 1963). The remainder of the fauna is made up of forms whose Recent relatives are widely dis- tributed in Australia or, in the case of the extinct forms, whose eco- logical requirements are unknown. At least one is related to a Re- cent species (Aepyprymnus rufescens) that is reported to inhabit open woodland (Marlow, 1957, 1958). It may indicate the presence of local open areas or it may have had ecological requirements different from the modem form. Thus the fauna appears to be a forest fauna, and apparently the forest was considerably more humid than those in Western Victoria today. This conclusion thus reinforces the earlier one (Gill, 1957) regard- ing the general environment of the Hamilton area at the time of formation of the fossil soil with its fauna. SUMMARY AND GENERAL CONCLUSIONS The Hamilton fauna consisting of 18 mammalian taxa (mainly marsupial) has been collected from a fossil soil developed on the Kalimnan limestone of Pliocene age on the Grange Bum, four miles west of Hamilton, Victoria. The material consists almost entirely of teeth collected by wet sieving. The soil is overlain by a basaltic flow which has been radiometrically dated at 4.35 ±0.1 MY. The presence of charred stumps under the basalt indicates contemporane- ity of the fossils and the flow. Large-sized species are either poorly represented or are absent, and lower vertebrates, dasyurids, and rodents are absent. With the exception of the rodents, these deficiencies in the fauna are believed to be the result of either destructive burial environment or of sam- pling accidents. The absence of rodents is believed to be real be- cause of their abundance in Pleistocene deposits. All taxa represented, except the bat and the five small enamel caps described on pages 70-72, fall into known Recent or Pleistocene marsupial subfamilies. Most are clearly related to known Recent or Pleistocene genera, although some may require generic separation when they become better known. Three new species and a new genus of pseudocheirines are named, and probably many of the unassigned and unnamed specimens of other families represent undescribed spe- cies, but the inadequacies of the materials militated against naming of new taxa. Only a few, such as Burramys and Palorchestes, show differences from Recent or Pleistocene relatives that might indicate primitiveness. The Hamilton Burramys sp. has characters of its P|s which could be regarded as primitive. The lower molar of Palor- chestes is smaller and slightly more generalized than any of the teeth of the small Pliocene-Pleistocene species, P. parvus from Queens- land, and is very similar to the ? Miocene-Pliocene Alcoota, P. painei. Taxa of the two best represented groups, Pseudocheirinae and Macro- podidae, show no obvious indications that would mark them as being either more primitive or more advanced than the known fossil or Recent close relatives. 79 80 FIELDIANA: GEOLOGY, VOLUME 19 As a whole, the fauna appears to be a wet forest fauna with many indications that the environment was a wetter one than presently occurs in the Hamilton region. In composition, except for its lack of dasyurids and rodents, and in evolutionary grade with its strong similarities to Pleistocene and Recent faunas, it is about what one would expect of a Late Pliocene fauna. It thus supports broadly the late Tertiary faunal correlations that have been made during the past two decades, but it provides suggestive evidence that the Al- coota fauna may be slightly younger than has been thought. ACKNOWLEDGEMENTS The following people most generously provided us with facilities, work space, and information: Mr. John McNally, Mr. Edmund Gill, Mrs. J. Hope Macpherson Black, and Mr. Thomas Darragh of the National Museum of Victoria; the late Professor A. J. Marshall, Professor James Warren, Dr. E. H. M. Ealey, and Mr. James Guthrie of the Department of Zoology and Comparative Physiology at Monash University; Dr. Murray Littlejohn of the Department of Zoology, University of Melbourne. The George Pelchen family, on whose property on the Grange Burn, the Hamilton fauna was col- lected have welcomed us, provided assistance and opened their home to us. This work was supported by the National Science Foundation (Grants GB 975, GB 3729), by Field Museum and the University of Texas. The following assisted in sorting the matrix concentrates or helped in other ways : Frederick Schram and Rodger Podewell (NSF- supported graduate student assistants); Charles Walker and Mrs. Janet Adejunmobi (NSF-supported undergraduate student assist- ants); Mrs. Priscilla Turnbull (Field Museum, Dee Fellow); Mrs. Judith Lundelius, Mrs. Eleanore Pringle, and Mr. Clarence Hope (volunteers) ; and Joan Samuelson, Mark Woolsey, and Robert Lewis (Field Museum -Antioch College co-operative project students). We acknowledge help with the illustrations by the following Mu- seum personnel: Marion Pahl, Staff Illustrator; Dr. Tibor Perenyi, Staff Artist; and John Baylis, Homer Holdren, and Ferdinand Huys- mans of the Division of Photography. We are grateful to Dr. Rainer Zangerl, Chief Curator of Geology, Dr. Richard Tedford of the Amer- ican Museum, and Professor E. C. Olson of the University of Chicago for critically reading the manuscript and for related discussions. We also acknowledge Dr. Zangerl's help with histological interpretations. REFERENCES Bartholomai, a. 1967. Troposodon, a new genus of fossil Macropodinae (Marsupialia). Mem. Queensland Mus., 15, I, pp. 21-33. Bensley, B. a. 1903. On the evolution of the Australian Marsupialia; with remarks on the relationships of the marsupials in general. Trans. Linnean Soc, London, Ser. 2, 9, Zool., pp. 83-216, figs. 1-6, pis. 5-7. Broom, R. 1896. On a small fossil marsupial with large grooved premolars. Proc. Linnean Soc, N.S.W., Ser. 2, 10 (reg. vol. 20), pp. 563-567, pis. XXV and XLV. COLLIVER, F. S. 1933. Some interesting fossils from the Tertiary deposits of the Grange Burn. Victorian Nat., 50, pp. 71-73. EvERNDEN, J. F., Savage, D. E., Curtis, G. H., and James, G. T. 1964. Potassium-argon dates and the Cenozoic mammalian chronology of North America. Amer. Jour. Sci., 262, pp. 145-198, 1 fig., 7 tables. Gill, E. D. 1952. Range in time of the Australian Tertiary flora. Austral. Jour. Sci., 15, no. 2, pp. 47-49, 1 fig. 1953a. Palaeoecological interpretation of some Victorian fossil diatom floras. Mem. Nat. Mus. Victoria, 18, pp. 141-153, figs. 1-3. 1953b. Fluorine tests in Australia on the Keilor skull and a Tertiary marsupial. Nature, 172, no. 4374, pp. 409-410. 1953c. Australian Tertiary marsupials. Austral. Jour. Sci., 16, no. 3, pp. 106- 108. 1955, Fluorine-phosphate ratios in relation to the age of the Keilor skull, a Tertiary marsupial, and other fossils from western Victoria. Mem. Nat. Mus. Victoria, 19, pp. 106-125, figs. 1-3, pi. 1. 1957. The stratigraphical occurrence and palaeoecology of some Australian Tertiary marsupials. Mem. Nat. Mus. Victoria, 21, pp. 135-203, figs. 1-18, pis. 1-4. 1965a. Paleontology of Victoria. Victorian Year Book, no. 79, pp. 1-23, 1 fig., pis. 1-8. 1965b. The paleogeography of Australia in relation to the migrations of mar- supials and men. Trans. N. Y. Acad. Sci., Ser. II, 28, no. 1, pp. 5-14. Glaessner, M. F. 1963. "Intrusive" vertebrate fossils in the Miocene of Victoria; a correction. Austral. Jour. Sci., 25, no. 9, pp. 411-412. Glaessner, M. F., McGowran, B., and Wade, M. 1960. Discovery of a kangaroo bone in the middle Miocene of Victoria. Aus- tral. Jour. Sci., 22, pp. 484-485. 81 82 FIELDIANA: GEOLOGY, VOLUME 19 Gray, J, E. 1837. Description of some new or little-known Mammalia, principally in the British Museum Collections. Mag. Nat. Hist. (Charlesworth) New Ser., 1, pp. 577-587. HiBBARD, C. W. 1949. Techniques of collecting microvertebrate fossils. Contrib. Mus. Paleon- tol., Univ. Mich., 8, no. 2, pp. 7-19, pis. 1-4. Keast, a. 1961. Vertebrate speciation in Australia: some comparisons between birds, marsupials and reptiles. The Evolution of Living Organisms, a symposium of the Royal Soc. Victoria held in Melbourne, December 1959. No. 32, pp. 380-407, figs. 1-8. Keil, a. 1966. Grundziige der Odontologie. Gebriider Borntraeger. Berlin-Nikolassee. i-x, 278 pp., 251 figs. Lehner, J. and Plenk, H. 1936. Die Zahne. Handbuch der Mikroskop. Anatomie V, no. 3, pp. 449- 708, figs. 1-119. Lesson, R. P. 1828. In Vol. 13, p. 333. Dictionnaire classique d'histoire naturelle. Boudouin Freres, Paris. LuNDELius, E. L., Jr. 1960. Post-pleistocene faunal succession in Western Australia and its climatic interpretation. Int. Geol. Cong., XXI session, Norden, 1960. Part IV, chronology and climatology of the Quaternary. Copenhagen. Pp. 142-153. LuNDELius, E. L., Jr. and Turnbull, W. D. 1967. Pliocene mammals from Victoria, Australia. Austr. New Zeal. Assn. Adv. Sci. 39th Congress, Melbourne, Jan. 1967. Abstracts, Sec. G: K-9. Marlow, B. J. 1957. A recent record from New South Wales of the rufous rat-kangaroo, Aepyrymnus rufescens (Gray) [Macropodidae]. Commonwealth Sci. Indust. Res. Org. Wildlife Res., 2, no. 2, pp. 166-167, fig. 1. 1958. A survey of the marsupials of New South Wales. Commonwealth Sci. Indust. Res. Org. Wildlife Res., 3, no. 2, pp. 71-114, figs. 1-34. 1962. Marsupials of Australia. Jacaranda Pocket Guide. 141 pp. Ogilby, W. 1837. Observations on the opposable power of the thumb in certain mammals, considered as a zoological character; and on the natural affinities which sub- sist between the Bimana, Quadrumana, and Pedimana. Mag. Nat. Hist. (Charlesworth), New Ser., 1, pp. 449-459; 517-525. Owen, R. 1874. On the fossil mammals of Australia. Part IX. Macropodidae: genera Macropus, Osphranter, Phascolagus, Sthenurus and Protemnodon. Trans. Roy. Soc, London, 164, no. 23, pp. 783-803, pis. LXXVI-LXXXIII. Peyer, B. (R. Zangerl translation) 1968. Comparative odontology. Univ. Chicago Press, Chicago and London, i-xiv, 347 pp., figs. 1-220, pis. 1-88 and 1-8 in color. TURNBULL AND LUNDELIUS: HAMILTON FAUNA 83 Plane, M. 1967. Stratigraphy and vertebrate fauna of the Otibanda formation, New Guinea. Bull. Bur. Mineral Res., 86, i-vi, 64 pp., figs. 1-14, pis. 1-6. Ride, W. D. L. 1964. A review of Australian fossil marsupials. Jour. Royal Soc. W. A., Perth., 47, no. 4, pp. 97-131, figs. 1-13. ScHLEGEL, H. and Muller, S. 1839-1844. Over drie Buideldieren uit de familie der Kengoeroes. In Temminck, C. J., Verhandelingen over de Natuurlijke geschiednis der Nederlandsche overzeesche besitzigen door den Leden der Natuurkunde Commissie in Indie en ondere Schrijvers. Zoologie. Leiden. I, pp. 129-148, pis. 19-24. ScHRAM, F. R. and Turnbull, W. D. 1970. Structural composition and dental variation in the murids of the Broom Cave Fauna, Late Pleistocene, Wombeyan Caves area, N. S. W. Rec. Austral. Mus., 28, no. 1, pp. 1-24, pis. 1-3, figs. 1-5. Simpson, G. G. 1929. American Mesozoic Mammalia. Mem. Peabody Mus., Yale Univ., 3, Part I: i-xv, 235 pp., figs. 1-62, pis. I-XXXII. 1961. Historical zoogeography of Australian mammals. Evolution, 15, no. 4, pp. 431-446, 1 fig. Singleton, F. A. (fide Gill, 1957) 1935. "Cainozoic." In Outline of the physiology and geology of Victoria. Austr. New Zeal. Assn. Adv. Sci. Handbk. Victoria-Melbourne meeting 1935, pp. 128-135. Stirton, R. a. 1955. Late Tertiary marsupials from South Australia. Rec. S. Austral. Mus., Adelaide, 11, no. 3, pp. 247-268, figs. 1-11. 1957a. A new koala from the Pliocene Palankarinna fauna of South Australia. Rec. So. Austral. Mus., Adelaide, 13, no. 1, pp. 71-81, figs. 1-2. 1957b. Tertiary marsupials from Victoria. Mem. Nat. Mus. Victoria, Mel- bourne, 21, pp. 121-134, figs. 1-6. 1963. A review of the macropodid genus Protemnodon. Univ. Calif. Publ. Geol. Sci., 44, no. 2, pp. 97-162, figs. 1-15. Stirton, R. A., Tedford, R. H., and Miller, A. H. 1961. Cenozoic stratigraphy and vertebrate paleontology of the Tirari desert. South Australia. Rec. So. Austral. Mus., Adelaide, 14, no. 1, pp. 19-61, figs. 1-4. Stirton, R. A., Tedford, R. H., and Woodburne, M. 0. 1967a. Review of Tertiary mammal-bearing deposits in Australia. Austr. New Zeal. Assn. Adv. Sci. 1967 (39th) Congress Melbourne. Abstracts Sec. C: K 6-7. 1967b. A new Tertiary formation and fauna from the Tirari desert. South Australia. Rec. S. Austral. Mus., Adelaide, 15, no. 3, pp. 427-462, figs. 1-12. 1968. Australian Tertiary deposits containing terrestrial mammals. Univ. Calif. Publ. Geol. Sci., 77, pp. 1-30, figs. 1-3 and chart. Storr, G. C. C. 1780. Prodromus methodi mammalium. Tubingen. Friderice Wolffer. 43 pp., pis. 1-4. Tate, G. H. H. 1945a. The marsupial genus Phalanger. Amer. Mus. Nov., no. 1283, pp. 1-41, figs. 1-9. (Results Archbold Expedition, no. 52.) 84 FIELDIANA: GEOLOGY, VOLUME 19 1945b. The marsupial genus Pseudocheirus and its subgenera. Amer. Mus. Nov., no, 1287, pp. 1-30, figs. 1-5. (Results Archbold Expedition, no. 54.) 1948a. Studies on the anatomy and phylogeny of the Macropodidae (Marsupi- alia). Bull. Amer. Mus. Nat. Hist., 91 , no. 2, pp. 233-352, figs. 1-3, (Results Archbold Expedition, no, 59,) 1948b, Studies in the Peramelidae (Marsupialia). Bull. Amer, Mus, Nat, Hist,, 92, no, 6, pp. 313-346, 1 fig. (Results Archbold Expedition, no. 60.) 1951. The rodents of Australia and New Guinea, Bull, Amer. Mus. Nat. Hist., 97, no. 4, pp. 183-430, figs, 1-4, (Results Archbold Expedition, no, 65,) Tedford, R. H, 1966. A review of the macropodid genus Sthenurus. Univ. Calif. Publ. Geol. Sci., 57, pp. 1-72, figs. 1-20. 1967. The fossil Macropodidae from Lake Menindee, N,S,W, Univ. Calif. Publ. Geol. Sci., 64, i-v, 156 pp., figs. 1-32, pis. 1-4. Troughton, E. 1962. Furred animals of Australia. Angus and Robertson (7th ed.). i-xxxii, 376 pp. TuRNBULL, W. D., LuNDELius, E. L., Jr., and McDougall, I. 1965. A potassium-argon dated Pliocene marsupial fauna from Victoria, Aus- tralia. Nature, 206, no. 4986, p. 816. Wakefield, N, A, 1963, The Australian pigmie-possums, Victorian Nat,, 80, pp, 99-116, figs. 1-4, 2 maps. White, T. E, 1959, The endocrine glands and evolution, no, 3 : os cementum, hypsodonty, and diet, Mus, PaleontoL, Univ, Mich,, Ann Arbor, 13, no, 9, pp, 211-265, Woodburne, M, O. 1967. The Alcoota fauna. Central Australia, Bull. Bur. Min. Res., 87, i-ix, 187 pp., figs. 1-34. Woods, J. T. 1956. The skull of Thylacoleo carnifex. Mem. Queensland Mus., 13, no. 2, pp. 125-140, figs. 1-6. 1958. The extinct marsupial genus Palorchestes Owen. Mem. Queensland Mus., 13, no. 4, pp. 177-193, figs. 1-5. 1960. Fossiliferous fluviatile and cave deposits. In D. Hill and A. K. Den- mead, eds. The Geology of Queensland, 7, pp. 393-403, fig. 55. PLATES AND APPENDIX PLATES Plates I-XXXI are comprised almost entirely of steroscopic pho- topairs (The exceptions are figures D and E of Plate VII and figures E and F of Plate XXXI) illustrating all of the taxa of mammals in the Pliocene, Hamilton fauna from the Grange Burn. Most of the better specimens are shown. The marsupial families are covered as follows: Peramelidae — PI. I; Phalangeridae — Pis. II-VI; Phascolarc- tidae— Pis. VII-XVIII; Macropodidae— Pis. XVIII-XXVIII; Dip- rotodontidae — Pis. XXX. The few non-marsupialian remains, and some miscellaneous materials are illustrated on Pis. XXVII, XXIX, XXXI. Abbreviatons for major cusps are as follows: end=entoconid; esd=entostylid; hy=hypocone; hyd=hypoconid; hyld=hypoconulid; me=metacone; med=metaconid; mele=metaconule; mes=mesostyle; msd = metastylid ; pa = paracone; pad = paraconid ; pas = parastyle ; pr=protocone; prd=protoconid; prle= protoconule. 87 Plate I . — Peramelid teeth. Arrows mark certain other features discussed in the text. e= ridge connecting entoconid to the posterior ridge of metaconid. h=ridge connecting hypoconid to the back edge of trigonid. A-D. PM 4499. Trigonid of a right molar, lateral, crown, posterior, and medial views. Scale approx. X 13. E-F. NMV-P26406. Partial right trigonid, crown, and posterior views. Scale approx. X 15. J 88 Plate II. — Teeth of various phalangerines. Scale approx. X 13. A. Incertae sedis, near Phalanger or Trichosurus, PM 4722, tip of left upper canine, crown, and lateral views. B. Trichosurus sp. indet., PM 4557, fragment of a left P^, crown view. C. Phalanger cf. gymnotis, PM 4457, a heavily ribbed left P4, lateral and crown views. D. Trichosurus sp. indet., PM 4556, fragment of an anterior edge of a left P4, crown view. 90 Plate III. — Teeth of two phalangerines. Scale approx. X 123/^. Stereo relief slightly exaggerated. A-C. A tooth of uncertain generic affinities, ? Phalanger, or Trichosurus, or an unnamed related genus, PM 4555, a right My, crown, antero- lateral, and medial aspects. D, E. Trichosurus sp. indet., PM 4563, a left I^, lateral and crown views. 92 Plate IV. — Tooth of a phalangerine of uncertain generic affinities, ? Phalan- ger, or Trichosurus, or an unnamed related genus, PM 4571, a right M^. Scale approx. X 13. A, B. Lateral and crown views. C, D. Anterior and medial views. ^^ 94 Plate V. — Right M5 of a phalangerine of uncertain generic affinities, ? Pha- anger, or Trichosurus, or an unnamed related genus, NMV-P 26407. Scale approx. X 12. A, B. Medial and anterior views. C, D. Crown and lateral views. 96 Plate VI. — Teeth of several phalangerids and an indet. macropodine. A. In- det. macropodine, PM 4559, a partial left lower premolar, labial view. B. Indet. minute phalangerine, PM 4553, a left ly in dorso-mesial (dorso-medial) view. C. Burramys sp., NMV-P26409, posterior three-fourths of a left P4, medial view. D. Burramys sp., PM 4439, anterior third to half of a right P4, lingual (top right), anterior (middle), and labial (bottom left) views. E. Burramys sp., PM 4470, a complete left M-, crown view. Scale A-E approx. X 13. F. Burramys sp., PM 4459, fragment of a right P- preserving the antero-lingual corner of the tooth, lingual view. Scale F approx. X 18. 98 Plate VII. — Teeth of Pseudokoala erlita, the large pseudocheirine from the Hamilton fauna. Scale approx. X 73^. A, B, NMV-P26399 (part of type). Left M^, lateral (buccal) and crown views. C. NMV-P26399 (left), posterior view, and NMV-P26400 (also part of type), anterior view showing the matched interdental facets. D, E. Sketches of these facets as an aid to interpretation of the stereo photos shown in C. which proves the teeth to belong together. 100 Plate VIII. — Teeth of pseudocheirines from the Hamilton fauna. Scale approx. X 11. A. and B. Pseudokoala erlita. A. NMV-P26400, the left M^ (part of type) in occlusal view. B. PM 4588, anterior half of the left My, dor- sally oblique, antero-lateral view. C. & D. ? P. erlita. C. PM 4495, possibly the posterior half of a paracone, or metacone of an anterior molar, or more likely the back of a P^. The view is a postero-medial oblique one for each of three alter- native interpretations. D. PM 4496, the posterior half of a paracone seen in a comparable angle as in C. 102 ^.^^ Plate IX. — Teeth of Pseudocheirus stirtoni. Scale approx. X 13. A-C. Left Mj, NMV-26401 (part of type), crown (A), medial (B), and lateral (C) views. D. Left M2, NMV-P26402 (part of type), crown view. Numbered arrows cor- respond to the following scheme: (1) posterior crest of protoconid of My; (2) hy- polophid ridge; (3) entostylid ridge. 104 Plate X.— Teeth of Pseudocheirus stirtoni. A-C. Left Mg, NMV-P 26402 (part of type), medial (A), posterior (B), and lateral views. Scale approx. X 13. D. Left Mg, NMV-P26403 (part of type), crown view. Arrow 4 indicates an interdental facet. Abbreviations 1, 2, 3 as in Plate IX. Scale approx. X 12. 106 iJV^4 iJ^^ end ^ ^6.' d A li ^\.*«V > 2 m m 'M- ir.ii:'^ % ■ f f ^ end • ' » '■ K*^' V .?J5r'-^fc?»^ ■^^■■IP^^r., med , • -^Tv-w . prd ■••T^ med end p r d ^ * ^^^"^ h y d Plate XI.— Teeth of Pseudocheirus stirtoni. A-C. Left M5, NMV-P26403 (part of type), lateral (A), anterior (B), and posterior (C) views. Scale approx. X 12. D-E. Left M^, NMV-P26404 (part of type), crown (D), and medial (E) views. Scale approx. X 13. Abbreviations 1, 2, 3 as in Plate IX. \.^ 'A 108 Plate XII. — Teeth of Pseudocheirus stirtoni. Scale approx. X 13. A, B. Left M4, NMV-P26404 (part of type), lateral (A) and posterior (B) views. C-E. Left P4, PM 4477, medial (C), lateral (D), and crown views. Arrows in C-E point to cleft between the major cusps discussed in text. 110 Plate XIII. — Three upper cheek teeth of Pseudocheirus stirtoni. A-C. A left P^ (PM 4549), lingual (A), crown (B), and labial (C) views. Scale approx. X 10. D, E. A left Mi (or M^) (PM 4542), labial (D) and crown views. Scale approx. X 13. F. A right M-"^ (or M^^) (PM 4543), labial view. Scale approx. X123^. 112 2 m m 4 jt ^ir- ■[m'L ♦# 4 1 ._>«IR.^-*">-r^--'<. '^liSBP^ . --Of -^■'-' 'S'l'i^^W^ •• ■ ■•..i:_> .!►♦•' .„■■ ^KUt^KMBf" j^M^Pv '^^^^^■S^^^BI pa \ 2 / me m m ^^^HJ^^^^^^H ^^^^^^^^^.. » f \. Plate XIV. — Three upper molars of Pseudocheirus stirtoni. A. A right M^ (or M^) (PM 4543), crown view. Scale approx. X 13. B-D. A right M^ (or M^) (PM 4422), posterior (B), labial (C), and crown views. Scale approx. X 9. E, F. A right M^ (or M^) (NMV-P26410), labial (E), and crown views. Scale approx. X13. ^/ H4 Plate XV. — Three teeth of Pseudocheirus marshalli. Scale approx. X 13. A. PM 4591. A right P4, crown view, B. NMV-P26405, the type. A left M^ lacking the paracone, labial (above), and occlusal views. C. PM 4587. A left M-, labial (above) and occlusal views. 116 Plate XVI. — Three teeth of Pseudocheirus marshalli. A. PM 4423. A frag- ment of a left molar, probably M^-, crown view. Scale approx. X 9. B. PM 4455. Right My, crown, lateral, and medial views. Arrow indicates separation of "ento- stylid ridge" from entostylid, one of the features that distinguishes P. marshalli from P. stirtoni. G. NMV-P26413. Right Mg (or Mj), crown, lateral, and medial views. Scale of B and C approx. X 10. 118 Plate XVII. — Teeth of Pseudocheirus marshalli and Pseudocheirus sp. indet. from the Hamilton fauna. A. PM 4453. Posterior two-thirds of a left lower molar (either M^ or Mg) of P. marshalli preserving part of the trigonid and all of the talonid, occlusal (above, left), and posterior views. Arrow 1 points to the entostylid and arrow 2 to the minute hypoconulid. B. PM 4476. A right M4 or P. marshalli, occlusal (above, left) and medial (-lingual) views. A and B shown at approx. X 12. C. PM 4472. A left I^ of Pseudocheirus sp. indet., dorsal and medial views. Scale approx. X 9. 120 -2 mm Plate XVIII. — Teeth of Pseudocheirus sp. indet., Aepyrymnus sp. indet., and Hypsiprymnodon sp. indet. A, NMV-P26415. A left ly of Pseudocheirus sp,, dorsal (left) and near-medial views. B. PM 4492. Fragment of left upper molar referred to Aepyprymnus sp. preserving the metacone and part of the hypocone. The fragment is shown in its posterior (top, right) and occlusal aspects. Arrows point to the posterior "cingulum." Scale of A and B approx. X 10. C. PM 4575. Anterior half of a rightM^ referred to Hypsiprymnodon sp., labial (top) and occlu- sal views. Arrows indicate anterior cingulum. Scale approx. X 123^. 122 Plate XIX. — Partial premolar of a macropod, Dorcopsis sp. indet., from the Pliocene, Hamilton fauna, PM 4436. Posterior third of a right P^. Scale approx. X 12. A. Occlusal view. B. Lingual view. C. Posterior view. D. Labial view. 124 2 mm- Plate XX. — An upper molar of the macropod, Dorcopsis sp. indet., PM 4433. Scale approx. X 8. A. Anterior view. B. Labial view. G. Occlusal view. D. Lingual view. E. Posterior view. Numbers 1 and 2 indicate the protoloph and metaloph. 126 Plate XXI. — NMV-P26417, a right upper molar of the macropod, Dorcopsis sp. indet. Scale approx. X 9. A. Anterior view. B. Labial view. G. Crown view. D. Posterior view. The arrow indicates the procingulum. 128 Plate XXII. — Right upper molar of Dorcopsis sp. indet., PM 4434. Scale approx. X 10. A. Anterior view. B. Labial view. C. Crown view. D. Pos- terior view. The arrow indicates the procingulum. 130 Plate XXIII. — Two lower molars referred to two genera of macropods in the Hamilton fauna. A. Dorcopsis sp., PM 4565, a nearly complete right molar lack- ing the tip of the protoconid and most of the hypoconid, crown view. Scale approx. X 9. B-E. Thylogale sp., PM 4564, a right molar, anterior (B), crown (C), lin- gual (D), and labial (E) views, ant cing=anterior cingulum. Scale approx. X 7H- 132 Plate XXIV. — Two macropod teeth referred to Thylogale sp. indet. of the Hamilton fauna. Scale approx. X 7. A-D. PM 4576, a right lower molar, an- terior (A), crown (B), lingual (C), and labial (D) views. E, F. PM 4432, the ante- rior half of a right lower molar, anterior (E) and crown (F) views. 134 Plate XXV. — Three macropod incisor teeth referred to Thylogale sp. indet? from the Hamilton fauna. Two are shown compared directly with corresponding teeth of the Recent T. billiardieri. A. Comparison of right I^^'s of the Recent T. billiardieri, FM 81526 (left), and PM 4731 (right), ventro-lateral oblique views. B. Comparison of left I^ of T. billiardieri (left) with a partial right I^, PM 4731 (right), oblique view of occlusal surfaces. C. PM 4733, a left I^, antero-lateral, oblique view. Scale of A-C approx. X 7. D. (Right) PM 4444, a left ly, ventro- lateral view, compared with the left Ix of T. billiardieri (left), FM 81526. Scale approx. X 21^. 136 Plate XXVI. — A deciduous lower molar tooth, PM 4438, referred to Thylo- gale sp. indet.: crown (A), posterior (B), lingual (C), and labial (D) views. Scale approx. X 11. 138 Plate XXVII. — Two macropod teeth and one tooth of an undetermined mar- supial (C) from the Hamilton fauna. Scale approx. X 11. A, PM 4562. The posterior half of a premolar blade of Thylogale sp. indet., labial, lingual, and crown view. Identification is uncertain, specimen thought to be a P3 or P4. B. PM 4558. Posterior half of a premolar blade of an indet. macropodine, crown and side views. G. PM 4580. Fragment of a tooth as yet undetermined. 140 2 m m \ Plate XXVIII. — Two teeth referred to Macropodinae, Incertae sedis (large- sized form, probably a Protemnodon) from the Hamilton fauna. A-E. PM 4429, a left upper molar or a dP- in anterior (A), crown (B), labial (C), lingual (D), and posterior (E) views. F. PM 4498, a fragment of a worn right upper molar, an- terior view. Scale approx. X 4}^. 142 Plate XXIX. — Various tooth fragments referred to the Marsupialia are shown in Figures A-G, and the one placental speciman in the fauna, a bat, is shown in Figure H. Specimens in A-G all appear to be referrable to either Macropodinae, Potoroinae, or Phalangeridae. They have sufficient morphology that eventually more positive identification may be possible. Scale approx. X 11. A. PM 4593 B. PM4582. C. PM4730. D. PM 4595. E. PM 4560. F. PM 4594. G. NMV-P26424. The bat (PM 4458), a microchiropteran, consists of a left ramus fragment with one molar tooth in place. Scale approx. X 12. 144 mm Plate XXX. — A left lower molar, probably My (or M^) of a Palorchestes c. f. painei (PM 16801), the only diprotodont recovered from the Pliocene, Hamilton fauna; anterior (A), crown (B), posterior (C), lingual (D), and labial (E) views. Scale approx. X 23^. 146 med Plate XXXI. — Five mammal teeth from the Grange Burn, Pliocene, Hamil- ton fauna are shown. They are indeterminate as to subclass and lower taxo- nomic levels. A. (NMV-P26425), crown and side views; B-D, crown views B. PM4468. C. PM4466. D. PM 4465. Scale for A-D approx. X 15. E, F. Photo micrographs of PM 4599, E is a X-section showing peculiar enamel prisms cut lengthwise. F. A section approximately tangential to surface of the tooth which shows the enamel prisms cut across over most of the field, and cut obliquely, at the extreme left. Scale of E and F approx. X 940. 148 GRAPHS pq ^ _flj X) _2 ^ .. , -tJ ' — c C S • — c< *^ ^ XI -£3 !i Ki kJ^w < aoiy 3XNV CQ .2 Sk g 73 <^ 2 (3s 0! 35 Q +J O t^ » -C +i o o -1^ -< x: o a! 5^ be — Is. ^ OJ t< o C ° x: rt •^ c ^H •ro s- ;j 0! w c a 0) .« rr 2 S3 0) c J^ .~ a C si 0) c3 X! O T3 >8 ^1 = m a ^ ^ < 0) O t-, xn O ocu.s 152 . -1 2 9 £ 1 5 5 O S IS « 2 3 2 2 2 B D » < 1 1 < O y O 1 I i X ♦ O ^ « ^ D 1 0 t> «o •• «« •01 X O D o t 5 o! 1 *4- I * 1 J I S ' 6 5 * IsHl 1 a! s n s 0 i 1 1 1 > < o i -j-kM HJLQIM yoiyaxNV Graph C. — Bivariate plot of length vs. anterior width of M-^ of various species of pseudocheirines. The position of Pseudocheirus stirtoni and P. marshalli shows that in the size and proportions of the M^ they are like the M^ of P. pygmaeus, P. forbesi, and P. canescens and in contrast to the PJ less like Schoinobates volans minor. 153 Graph D. — Bivariate plot of length vs. anterior width of My of various species of pseudocheirines. The position of Pseudocheirus stirioni and P. marshalli indicates that in the proportions of the Mx P. stirtoni is similar to P.forbesi and Schoinobates polans minor, and P. marshalli is more similar to P. pygmaeus and P. canescens. See Graph C for detailed explanations of symbols. 154 B i" m r »- 0 z -J s ! s H S as s i B a <] ( B, Ox^J X ««x «>x «> ( > ) o ^ 1 ' *■ -^\n H±aiM yoiy3±NV S 155 Graph E. — Bivariate plots of average dimensions of length vs. anterior width of the P4 through M^: of various species of pseudocheirines. The symbol for each species appears at the P4 position and the line extends from there through Myi M2. M3, to M:j. The sample sizes are given adjacent to each species symbol. The plots show graphically the change in proportions of the teeth from anterior to posterior. The resulting patterns are variable but show reasonably consistent patterns at the subgeneric level. The subgenus Pseudocheirus consists of two groups of species with different proportions of the P4 and the Mt - Mg. The plots show that in the way in which the proportions of the teeth change P. stirtoni is similar to Schoinobates volans minor and P. marshalli is similar to the small-sized group of species in the subgenus Pseudocheirus (P. pygmaeus, P. forbesi, and P. canescens). See Graph C for detailed explanation of the species symbols. 156 Graph E / e n g t h 157 Graph F 8.0 9.0 10.0 11.0 12.0 Graphs F, G. — Bivariate plots of length (abscissa) vs. anterior width (ordi- nate) of upper cheek teeth of various species of Recent Dorcopsis, Dorcopsulus, and Thylogale, and of the fossil Dorcopsoides fossilis compared with those of the Hamil- ton Dorcopsis and Thylogale teeth. Key to symbols is shown on G. Tooth symbols are indicated at the joint mean of length and anterior width measures. Point clouds or lines show the dispersion of the samples. The positions of the Hamilton specimens of Dorcopsis indicates their closer affinity to the Recent Dorcopsis than to either the Recent Dorcopsulus or the fossil Dorcopsoides. The Hamilton Thylo- gale is similar in size and proportions to the posterior molars of several extant spe- cies of Thylogale. Discussion is given on pages 60 to 62. Dorcopsoides data from Woodburne, 1967. 158 Graph G 159 Graph H 5.0 4.0 3.5 3.0 4.0 3.5 3.0 f- PM4438 ■"PM4564 Key to symbols Various Living Species: Hamilton Specinnenst PMorNMV - No Si^ o 4.0 3.5 3.0 2.0 3.0 4.0 5.0 9.0 11.0 Graphs H, I. — Bivariate plots similar to those in F and G, showing length vs. anterior width (abscissa and ordinate, respectively) for the lower cheek teeth of the same species samples as shown in F and G. In addition, a sample of Recent Setonyx is also shown. The Hamilton Dorcopsis and Thylogale teeth are plotted for comparison. Symbols are keyed on H, and as further indicated on caption to F. The positions of the Hamilton Thylogale specimens indicate that they are all sim- ilar in size and proportions to those of several living species of the genus. The Hamilton Dorcopsis is most similar in size and proportion to the M^ of D. mulleri. Dorcopsoides data from Woodburne, 1967. 160 Graph I 5.0 60 7.0 80 9.0 100 12.0 161 21 20 19 18 16 15 Q q: O _^ UJ I- z < Palorches tes parvus -Q- 13 Pa/orc/ie^tes painei D. II IIIIHIili H illjlItU :: T O ••i-H:; 10 iiiiiiiii'iiiiiiiii i2(Jr Hami I ton near P. p Palorche ainei ?i»m«|[ick O [] Jl [] O ii ::»:: 5/^5, ie 17 18 io 20 22 23 Graph J. — Bivariate plot of length and anterior width of lower molars of t species of Palorchestes. The Hamilton specimen is clearly similar in size and pr portions to those of P. painei. Most of the data is from Woods (1958). 162 ■■ • • • • 7 ■ ? 7 • • • 7 • • • • • • • 7 oo • • : 7 Palorch azael estes 1 1 ■° 1^ o KEY TO ( ALL S SYMBOL REGIES) s # =My O =M-2 ■ -M3 LENG TH 25 26 27 28 29 30 31 32 33 3* 163 ffilU.GEOLOGVCHGO 19 1970 Publication 1105