Acrobates pulchellus Rothschild, 1892

P. Aplin, Kenneth, N. Armstrong, Kyle, M. Aplin, Lucy, Jenkins, Paula, Ingleby, Sandra & Donnellan, Stephen C., 2025, Hidden diversity in an ecologically specialized genus of Australian marsupials, the feather-tailed gliders, Acrobates (Diprotodontia, Acrobatidae), Zootaxa 5566 (3), pp. 535-564 : 546-553

Acrobates pulchellus Rothschild, 1892

Type specimens. [designation by Thomas (1922)]: The lectotype ( NHMUK 1888.3.17.1; an adult male) and paralectotype ( NHMUK 1888.3.17.2; an adult female) of A. pygmaeus ( Shaw, 1794) are spirit preserved bodies, the lectotype with the skull extracted and in good condition, the paralectotype with cranium in situ .

Diagnosis. Tail furred ventrally to the end, tipped with a small nubbin of naked skin ( Fig. 3A ); terminal pads of pedal digits 4 and 5 rounded, sub-equal in length and width, and with narrow distal groove; base of hallux with two distinct raised pads, both bearing striae ( Fig. 3C ); tail fringe hairs usually with a 1–2 mm wide cream emargination; ventral fur grey-based with cream tips, most individuals with entirely cream or white fur on throat and upper chest, rarely extending to lower abdomen ( Fig. 4 ). The ear canal disc, in external view, occupies between 60 and 70% of the width of the tympanic canal, has a raised narrow central ridge from the base that projects between 50 and 80% of the width of the disc. The disc in the internal view lacks sculpturing ( Fig. 5 ).

Description. Body measurements based on mature adults are presented in Table 3 . Small body size, up to 20 g and 85 mm head-body length. Flanks with a narrow patagium, running along the limbs and flank from the wrist to the knee.

Soft and silky fur, greyish brown dorsally ( Fig. 4 F-I). Most individuals with entirely cream or white fur on throat and upper chest, rarely extending to lower abdomen, otherwise ventral fur grey-based with cream tips ( Fig. 4 F-I). Colouring sharply demarcated at the edge of the patagium.

Tail equal in length to the head and body (mean = 1.08). Tail markedly distichous and oval in cross-section, with very short ashy-brown fur on the upper and lower surfaces, paler below than above. Tail furred ventrally to the end, tipped with a small nubbin of naked skin. Broad fringe of stiff long hairs along each side (about 8 mm in length), tail fringe hairs usually with a 1-2 mm wide cream emargination.

The eyes are large, dark and encircled with black fur. The rhinarium is naked, flesh coloured, and deeply cleft; the nostrils are lateral.

The ears are moderately large, flattened, oval, and moderately furred externally.

Long vibrissae, unusually numerous, extend from the snout, cheeks, and base of each ear with shorter vibrissae extending from the chin.

Hands and feet brown above, feet well furred, hands thinly furred. Terminal pads of pedal digits 4 and 5 rounded, subequal in length and width, and with narrow distal groove; base of hallux with two distinct raised pads both bearing striae ( Fig. 3C ).

Manual digits of relatively similar length in the following order 4>3>2>5>1. Palmar surfaces hairless, with four broad low pads, three at the base of the fingers and one at the base of the palm. Length of pedal digits in the following order 4>5>3>2>1. Plantar surfaces hairless with five small pads, four at the base of the toes and one at the base of the foot. Tips of the toes with flattened and striated pads, those on the syndactylous toes 2 and 3 separate and well developed, but smaller than the others. Syndactylous toes 2 and 3 united only to the ends of the first phalanx. Claws sharp and well developed, small in comparison with finger and toepads.

Dentition diprotodont. The homology of the anterior dentition of the mandible is not yet determined (Aplin 2015; Kelt and Patton 2020). The dental formula is currently described as I 3/1, C 1/1?, P 3/3? M 3/3 = 20+16=36 ( Fig. 6 ), based on examination of five vouchers (ANWC M3745, M6184, M24072, NHMUK 88.3.17.1, NHMUK 1939.2988). We note that the dental formula we have observed is at variance with previous descriptions in Harris (2015), Nowak (2018), Thomas (1888), but consistent with the dental formula in an illustration of the cranium and dentary of AMS M 3294 in Harris (2015).

Variation in cranial and dental morphology is illustrated in Fig. 6 and Supplementary Fig. S2 . We note that the description of cranial and dental morphology for Acrobates in Thomas’ (1888) account is likely based on a composite series of both species of Acrobates given the collection locations he listed for the 17 specimens he had available. Recently, Fabian et al. (2023), in describing several fossil acrobatids, carried out a comparison of dental morphology of extant and extinct acrobatids. From their appendix of material examined we were able to determine that they examined dentaries from 12 specimens of A. pygmaeus and 24 A. frontalis (Supplementary Table S1 ). Fabian et al. (2023) effectively had done a “blind” analysis of dental variation in that they did not identify the vouchers to species level prior to their examination and subsequently stated that: “we did not detect significant variation in the dental features we examined, including in m1 morphology (see Results). We conservatively refer to all of these specimens as A. pygmaeus ”.

Distribution. Broadly distributed across forested and wooded regions of south-eastern Australia, including: the south-eastern corner of South Australia; Victoria with the exception of the inland draining catchments of the Murray Basin; the Great Dividing Range of New South Wales, north to the Border and McPherson Ranges in south-eastern Queensland; and the coastal drainages of the Great Dividing Range, north at least to the Wallamba River, Tuncurry—AMS M25806 ( Fig. 7 ). Acrobates pygmaeus is regionally sympatric with A. frontalis throughout its entire range, with numerous instances of syntopy at the level of museum locational data (Supplementary Table S1 ).

Remarks. The lectotype and paralectotype of Didelphis pygmaea Shaw 1794 are leached of all pigment and stained a uniform brown colour. The lectotype is in poor condition, with much of the fur stripped from the venter and from the tail. The manus and pes are clenched and dehydrated, and no details of the plantar surface are discernible. However, the terminal pads of pedal digits 4 and 5, while folded and somewhat shrivelled, seem to be relatively unexpanded and thus consistent with Acrobates pygmaeus rather than A. frontalis .

The holotype of Acrobates pulchellus Rothschild, 1892 (NHMUK 1939.2988) is a well-preserved puppet skin and a cleaned skull which lacks the rear part of the skull Supplementary Fig. S2 B ). The purplish-brown colour of the dorsal fur, noted as a diagnostic feature by Rothschild (1892), falls within the observed variation of both species. The tail is short but the density of ventral fur at the tip leaves us in no doubt that it is damaged, possibly to a significant degree. The pes is consistent in morphology with A. pygmaeus , featuring relatively small, ovate terminal pads on each of digits IV-V and a prominent accessory hallucal pad on which the striae are clearly visible. The extensively grey-based ventral pelage of NHMUK 1939.2988, though not strictly diagnostic at species level, nonetheless is also consistent with the typical condition in A. pygmaeus .

Our assessment of the holotype of A. pulchellus supports and extends the previous findings of Tate (1938). In all morphological respects, including the morphology of the skull and teeth examined by Tate (1938), the purportedly New Guinean holotype of Acrobates pulchellus resembles A. pygmaeus of south-eastern Australia, rather than northern populations of A. frontalis . Like Tate, we favour the view that the locality details accompanying NHMUK 1939.2988 are spurious and that it was derived from somewhere in south-eastern Australia, a conclusion that could be tested further by DNA sequencing of the holotype.

Because of the extensive regional sympatry of the two species of Acrobates in south-eastern Australia, retrospectively identifying which species was used in published studies of Acrobates could be problematic. For instance, Ward (1990) reported four mammae in Acrobates but the specific identity of the material he examined is not known. A possible exception would be for studies carried out on the Southern Tablelands of New South Wales at elevations above 600 m.

Figures

FIGURE 2. Phylogenetic relationships among Acrobates based on: A) mitochondrial ND2 gene; B); a 96 bp segment of the mitochondrial 12S rRNA gene (Bayesian tree); and C) a 148 bp segment of the BRCA1 gene (ML tree). ML bootstrap proportions from 100 pseudoreplicates (left) and Bayesian posterior probabilities (right) indicated at nodes.

FIGURE 3. Comparison of the morphology of ventral tail tip A, B) and pads of pedal digits C, D) in Acrobates. The extent of naked tail tip and relative degree of fur density at tip differs between A) Acrobates pygmaeus, and B) A. frontalis. Pedal digit morphology in C) A. pygmaeus (ANWC M24072)—terminal pads of pedal digits 4 and 5 rounded, subequal in length and width, and with narrow distal groove (black arrows); base of hallux with two distinct raised pads (blue arrows); and D) A. frontalis (ANWC M30739)—terminal pads of pedal digits 4 and 5 broadly heart-shaped, much wider than long and with broad distal groove (black arrows); base of hallux with a single raised pad (blue arrow). Note the syndactylous toes 2 and 3 in C, D).

FIGURE 4. Ventral and dorsal pelage colour in Acrobates. Acrobates frontalis—A) AMS M1370 Belltrees, Scone, NSW; B) AMS M3001 Belltrees, Scone, NSW; C) AMS M860 Hornsby, NSW; D) AMS M8326, Maraylya, NSW; E) AMS M8326, Maraylya, NSW dorsal view; Acrobates pygmaeus—F) AMS M3048 Newnes Forest, via Newnes Junction NSW; G) AMS M5060 Arandale, Burragate, NSW; H) AMS M3295, Cox’s River, Lithgow, NSW; I) AMS M3723, Portland, NSW; J) AMS M3723, Portland, NSW dorsal. Images for AMS M3723 and M8326 by Ron Lovatt © Australian Museum DigiVol. Scale bars are 1 cm.

FIGURE 5. Ear canal disc morphology in acrobatid possums—external (left) and internal (right) views for each pair of images via 3D reconstructions from Micro-CT scans of A–E) A. pygmaeus, A) ANWC M3745 Bondo NSW, B) ANWC M3912 Michelago ACT, C) ANWC M10812 Coolangubra State Forest, Bombala NSW, D) ANWC M24072 Nadgee Nature Reserve NSW, E) ANWC M24706 Mongarlowe River, 3 m S Mongarlowe NSW; and F–I) A. frontalis, F) ANWC M183 Mt Brown, Richmond Range State Forest NSW, G) ANWC M29876 Julatten Qld, H) ANWC M30739 Dalmeny Qld, I) Coonabarabran; and J) Distoechurus ANWC M12247 Mt Moiba, near Opanabu, Milne Bay Province, PNG. K) Lateral view of an Acrobates skull showing position of the ear canal disc.In external view, the disc occupies between 60 and 70% of the width of the tympanic canal in A. pygmaeus, versus 70 and 90% in A. frontalis. In the internal view, the disc lacks sculpturing in A. pygmaeus versus sculpturing that emphasises the raised margin in A. frontalis. In A. frontalis in external view, the interior of the disc is depressed so that the margin appears raised. The disc morphology in Distoechurus pennatus most closely resembles that of A. pygmaeus.

FIGURE 6. Images of cranium and mandible of Acrobates via 3D reconstructions from Micro-CT scans. A–E) A. frontalis ANWC CM29876 Julatten Qld, F–J) A. pygmaeus ANWC CM3745 Bondo NSW, K–O) Distoechurus pennatus AMS M14752 Mt Karimui, South Simbu Province, PNG, in dorsal, ventral, lateral cranial views and buccal and occlusal mandible views. The crania, mandibles and teeth of the two Acrobates species are very similar (also see Supplementary Fig. S2 for illustration of variation within each species). Acrobates differs from Distoechurus cranially in having a proportionally larger and more inflated neurocranium, a proportionally shorter and deeper rostrum, a narrower mandibular angular process. See the text for a more extensive description.

FIGURE 7. Map showing distribution records of Acrobates based on identified vouchers (Supplementary Table S1). A) A. frontalis—blue squares; B) A. pygmaeus—red circles.