Nyctophilus timoriensis (Geoffroy, 1806)

Parnaby, Harry E. & Helgen, Kristofer M., 2023, Rediscovery of the Long-eared Bat Genus Nyctophilus (Chiroptera: Vespertilionidae) in Timor and a Reassessment of Nyctophilus timoriensis, Records of the Australian Museum 75 (5), pp. 629-652 : 634-643

A revised concept of Vespertilio timoriensis Geoffroy

Since the mid nineteenth century, Geoffroy’s timoriensis has been viewed as one of the largest members of Nyctophilus and its identity has been closely intertwined with N. major Gray, 1844 from south-western Western Australia with which it has often been synonymized. We present a new concept of timoriensis , based on details of the illustration accompanying Geoffroy’s description, a re-interpretation of measurements given in Geoffroy’s account and a detailed examination of the basis for the prevailing view that timoriensis is one of the largest Nyctophilus species that resembles N. major . As set out below, we conclude that timoriensis is not conspecific with N. major , differing in important morphological distinctions that have previously been overlooked.

Geoffroy’s account of timoriensis

The illustration of the head

A frontal drawing of the head (with the skull in situ) in Geoffroy’s (1806) plate 47 is the sole illustration accompanying his description of timoriensis (reproduced in Fig. 3 ). Two aspects of this drawing have been overlooked in previous assessments of the status of this species. First, there is an elevated mound on the rear of the snout behind the narial foliations. The oblique ventral angle of the illustration obstructs a clear view of the dorsal snout region and a rudimentary post-nasal elevation, such as that characterizing N. major , would not be visible from this angle. The illustration resembles the distal part of the snout prominence of those Nyctophilus species in which the post-nasal mound is well developed, such as in N. heran , N. gouldi and N. geoffroyi . The snout mound in Nyctophilus consists of two separate bodies joined in the midline by an elastic membrane of variable extent. The bilobed structure is not seen in Geoffroy’s illustration. However, it is likely that the artist did not have an accurate understanding of the shape of this structure, which might have shrunk in the specimen, and therefore was unable to accurately depict it. The small size of the illustration might be one reason these snout features have been overlooked, but the illustration contains an impressive level of detail. The enlarged snout mound is in sharp contrast to the low, rounded mound characteristic of N. major . This difference alone would suggest that timoriensis and major are not conspecific.

A second notable but overlooked feature of the illustration is a scale bar adjacent to the head of timoriensis on Geoffroy’s plate 47. Geoffroy (1806: 205) states that head length is indicated by the scale line beside the head of each species illustrated. The head length line for timoriensis is ca. 17.5 mm long, measured on a hardcopy of the journal illustration. A comparison of his frontal and lateral illustrations of the head of “ V. nigrita ” on the same plate clearly indicates that “length” was measured along the long axis of the head and was not measured in another manner, such as from ear to ear.

We suspect that the head illustrations on Geoffroy’s plate 47 are reproduced at life size, although he does not explicitly state so. Of the 19 species accounts in his paper, ear length is provided in the text only for V. auritus and his measurement of 33 mm is an exact match against his illustration, if ear length is taken from the notch as per modern measurements. If his illustration of the head of timoriensis is roughly life size, it is far too small to be a species the size of N. major but instead approximates a species the size of N. geoffroyi . Ear length measured from his illustration is ca. 22 mm, but the ears are not fully erect and could easily have attained about 25 mm. (Both of the latter ear measurements are not especially informative, as both fall within the expected range for many Nyctophilus including N. geoffroyi and N. major .)

Body measurements

Establishing the approximate body size of Geoffroy’s type specimen of timoriensis relative to other species of the genus is fundamental to an interpretation of its possible identity, and therefore requires detailed examination. The prevailing view that Geoffroy’s type was of one of the largest species of Nyctophilus originated from Tomes (1858a), yet as discussed below, only one of the measurements provided by Geoffroy (1806) implies a large-bodied species.

Three measurements are cited in Geoffroy’s brief description of timoriensis , which he gave in millimetres: body length, 70 mm; tail length, 40 mm; and wingspan, 270 mm. These were the standard measurements given by Geoffroy (1806) for each species in his account of vespertilionid bats. Geoffroy provided a fourth measurement for timoriensis , head length of ca. 17.5 mm, as noted above. Geoffroy generally used three standard body measurements (head-body, tail, and wingspan) for bats in his other taxonomic papers (e.g., Geoffroy, 1810, 1813). Geoffroy (1806) does not indicate how these measurements were taken, which could differ somewhat from modern standard measurements. His revision was published in an early phase of French taxonomic research when standard measurements for bats were evolving; e.g., Desmarest (1821) used twice as many body measurements, and the taxonomic value of forearm length, now a fundamental measure of size, was not recognized until later ( Geoffroy, 1832). Though we suspect that they would have been regarded at the time as having self-explanatory definitions, wingspan and body length could have been measured in several ways that would result in significantly disparate measurements.

Wingspan

The simplest interpretation of wingspan is a straight-line distance between wing tips of the extended wing. Alternatively, wingspan might have been measured along the bones of the leading edge of the wing, thus avoiding underestimates in specimens for which the wings could not be fully extended. Perhaps Geoffroy used both methods, depending on whether the specimen had fully extended wings, a procedure progressively adopted by Tomes throughout 1858. In his revision of Nyctophilus, Tomes (1858a , read 12 January) gave “expanse of the wings” for some species, while for others he cites “expanse of the wings, following the phalanges”. In a subsequent paper, Tomes (1858b: 125, read March 9th) explains that to overcome this problem with wingspan measurement he used a thread placed along the bones of the wing to the shoulder, adding the distance between shoulders. We are uncertain what method was adopted by Geoffroy (1806) in his description of timoriensis but for at least one species in his account ( Geoffroy, 1806) he seems to have measured wingspan as a straight-line span between wing tips, as revealed by his life-size illustration of one of his syntypes of Vespertilio emarginatus (= Myotis emarginatus ). Although he does not cite a wingspan measurement for that species in the text, a comparison with the measurements of the type specimens examined by Tomes (1858c) suggests a straight-line measurement. The linear wingspan that we measured from a hard copy of Geoffroy’s figure for emarginatus is ca. 258 mm and wingspan measured along the bones of the leading edge of the wing is ca. 306 mm. Geoffroy does not indicate which of his specimens of emarginatus was illustrated, but our linear wingspan measurement of 258 approximates the value 254 mm given by Tomes (1858c) for the specimen from Abbeville (the lectotype) and a wingspan of 267 mm for the Charlemont specimen.

Geoffroy’s wingspan measurement of 270 mm would seem to be too small for a species the size of N. major . It falls at the upper range for wingspan given by Churchill (2008) of 208–275 mm (n = 22, mean = 245) for the small-bodied species N. geoffroyi from northern Australia, and at the lower end of the intermediate-sized N. daedalus (275–323 mm, mean = 300, n = 61). Few wingspan data are available for N. major major from south-western Western Australia other than from Bullen & McKenzie (2002), who cite a mean of 349.5 mm (n = 8) but do not provide maximum or minimum measurements. Bullen & McKenzie (2001) provide a mean wingspan of 321.9 mm (standard deviation = 10.8 mm, n = 11) for N. major tor , from which we estimate the sample range to fall within 290–354 mm (based on 3 standard deviations from sample mean) but that form has a significantly smaller body size than N. major major . An important consideration is that wingspan taken along the bones of the leading edge of both wings will exceed the straight-line span between wingtips. Consequently, had Geoffroy measured the timoriensis wingspan along the leading wing edge, the straight-line span would be appreciably less than 270 mm, further indicating a relatively small-bodied specimen.

Head length and tail length

Head length measured on a study skin might seem a fairly imprecise measurement, although perhaps less so if the skull was in situ, as indicated in Geoffroy’s illustration. Nevertheless, the head length measurement given by Geoffroy suggests that his specimen might not have been as large as N. major . Geoffroy’s measurement of 17.5 mm (“8 lines”; 1 line = 2.1 mm) is smaller than the “10 lines” given by Tomes for his two specimens of N. major , yet falls within the range of species regarded by Tomes to be of small and intermediate body size (given as 7.5–9 lines), i.e. N. geoffroyi and N. gouldi . The range for GL of N. major major (18.8–20.7 mm, n = 20, Parnaby, 2009) also implies that Geoffroy’s timoriensis might belong to a relatively smaller-bodied species. Tail length (40 mm) is a relatively uninformative discriminator between species of Nyctophilus , e.g., Geoffroy’s measurement of 40 mm falls within the range of both the small-bodied N. geoffroyi and the large-bodied N. corbeni (see Churchill, 2008). (This applies regardless of whether tail length is measured from the root of the tail, or from the vent as for Churchill, 2008).

Body length

The different interpretations of “body length” need to be considered, given that we do not know how this measurement was taken and that modern concepts of the term might not have applied in the early 19th century. We assume Geoffroy’s “body length” included the head, and thus equates to “head-body length”. The latter interpretation was applied by Desmarest (1821), who gave Geoffroy’s body length of 70 mm for timoriensis as “length of body and head”. Body length was presumably taken from nose tip to base of tail on a stuffed specimen. An alternative interpretation could be that it was taken from the ear tips, because in long-eared bats such as Plecotus and Nyctophilus , the ears project forward well past the nose tip. However, for the one species for which Geoffroy (1806) provided ear length in the text in addition to “body” length ( Plecotus auritus , a species with exceptionally long ears), it is apparent that he measured body length from snout tip and not to the projecting ear tip.

Head-body measurements for small mammals are generally reported as a linear measurement and have been for more than a century, but it is possible that Geoffroy measured head-body length along the body contours of the dorsal surface along the midline, as was often done with skin mounts of large mammals, at least in the latter half of the nineteenth century. Our dorsal measurement on the apparent type of timoriensis taken along the spine (67 mm, Anja Divljan pers. comm. 2019) is a close match to the 70 mm given in the original description.

A body length of 70 mm is the only measurement give in Geoffroy’s description that matches a large species such as N. major , but only if this represents a linear measurement. If so, 70 mm would seem to exclude all but the largest species of Nyctophilus . Geoffroy’s head-body measurement would be some 5 mm longer if he had measured from the rear of the body, rather than the current practice of measuring from the vent. Head-body measurements provided by Churchill (2008) show that small and intermediate-sized species of Nyctophilus do not exceed about 50 mm. Head-body measurements are available only for three adult female and two male N. major major , as the species is poorly represented in collections. These field measurements of snout-vent length were taken from specimens now in the AM and range from 56–62 mm. However, Churchill (2008) provides snout-vent measurements of 50–75 mm (n = 33) for N. corbeni which is of comparable size to N. major . We compared body measurements given in fourteen nineteenth century accounts of timoriensis published in the decades after Geoffroy’s description (see below) in the hope that subsequent authors might have re-measured Geoffroy’s type but all appear to be re-iterations of his account. There is no indication that any of those authors had examined Geoffroy’s material, contrary to Parnaby (2009), who mistakenly believed that Temminck (1840) had done so. Some accounts are short ( Griffith, 1827; I. Geoffroy, 1832) while others seem to provide identical measurements when accounting for possible error from conversion to mm from the variety of European definitions of the inch of that time ( Desmarest, 1819, 1821; Lesson, 1827; Fischer, 1829, 1830; Temminck, 1840; Giebel, 1855, 1859; Wagner, 1840, 1855; Fitzinger, 1872).

Reconciling Geoffroy’s measurements

To summarize, the only clear indication of large body size, Geoffroy’s measurement of 70 mm for head-body length, seems to clash with other measurements given in his account. The wingspan of 270 mm would appear to be too small for such a large body length. Perhaps the wings were not fully extended on the type specimen, but the smaller body size implied by wingspan is supported by the head length measurement provided by Geoffroy, and the small size of the head illustration, which we suspect was reproduced at approximately life size. As noted above, we suspect that the head-body length reported by Geoffroy (1806) seems disproportionally large because it is measured as an arc length along the dorsal contours of the prepared specimen, and not as a linear measurement as usually reported today.

The account of N. timoriensis by Tomes (1858a)

The entrenched view that timoriensis is a large-bodied species similar to N. major that arose from Tomes (1858a) has remained unchallenged. Significantly, Tomes (1858a) had examined “the” original specimen of timoriensis in Paris, but his statement that it was “absolutely identical” to his Nyctophilus specimens from south-western Western Australia is not tenable in light of modern understanding of morphological variation in Nyctophilus . As previously mentioned, Tomes did not recognize differences in external morphology, other than size, between N. geoffroyi and N. major (for which he used the name timoriensis ), which are distinctive species. His account is bereft of a description or measurements of the Paris type specimen and we suspect that his assessment of timoriensis was based only on external features of the apparent type skin. In particular, Tomes seemed unaware of the diagnostic value of the relative size and structure of the snout mound, which he does not cite as a character for differentiating N. geoffroyi from N. major , two species that exhibit opposite extremes of development of that character. Most of the material available to Tomes consisted of dry skins, in which snout morphology might have been difficult to assess. A full appreciation of the diagnostic value of snout morphology in the genus was first recognized by Thomas (1915a), who assembled a large collection of fluid-preserved Nyctophilus on loan from the Australian Museum, Sydney.

Further doubt regarding Tomes’ emphatic judgement that timoriensis and major were “identical” or at least of similar body size arises from a comparison of Geoffroy’s measurements with those of Tomes (Tomes himself did not make that comparison.) When compared to Tomes’s measurements of his four species, head-body length of 70 mm (= 2 inch 9 lines) is the only one of Geoffroy’s four measurements that unequivocally fits the largest Nyctophilus recognized by Tomes, i.e. the southwestern Western Australian material that Tomes called N. timoriensis . As noted above, Geoffroy’s remaining three measurements (tail length, wingspan, and head length), when compared to measurements provided by Tomes, fit the species considered to be intermediate in size by Tomes, i.e. N. gouldi , and N. unicolor from Tasmania (currently a synonym of N. geoffroyi ). Tomes appeared to place some credence in wingspan as a character, as it was the only measurement directly cited by him when discussing size differences between his species. Had Tomes compared the wingspan of 270 mm given by Geoffroy against wingspan measurements for his own material, the intermediate size suggested by that measurement would have been apparent. Wingspans given by Tomes are: geoffroyi 9–10 inches (228–254 mm), gouldi and unicolor , ca. 10–11 inches (254–279 mm), and timoriensis from Western Australia, 12.75–13.5 inches (324–343 mm). Tomes measured wingspan for timoriensis as “expanse of the wings, following the phalanges” which is not equivalent to measurements of the remaining three species, measured as “expanse of the wings”, implying a direct tip to tip measurement. His measurements of ca. 323–343 mm are much larger than Geoffroy’s 270 mm, even when accounting for the fact that wingspan measured along the bones of the leading edge of the wing will be greater than a direct span between wing tips.

In conclusion, we suggest that the concept of timoriensis as a large-bodied species has a far weaker foundation than previously thought and it seems more likely to be of intermediate size in the genus. The concept of timoriensis as a large bodied species rests largely with the outdated assessment by Tomes (1858a) and on the body measurement of 70 mm given in Geoffroy’s account. Although Tomes examined Geoffroy’s original specimen, he based his understanding of timoriensis on specimens from south-western Western Australia from Gould’s collection, one of which was later designated the type (lectotype) of N. major by Thomas (1914). In effect, Tomes published the first diagnosis of what was later to become known as major , but under the name timoriensis . A further source of confusion arose because Tomes did not mention the name major anywhere in his paper. This omission was noted by Peters (1861), who proposed, in a footnote, that major should therefore be placed in the synonymy of timoriensis . Perhaps Tomes did not consider major to be a published name. The fact that N. major remained undiagnosed throughout the nineteenth century has also contributed to the erroneous conflation of timoriensis with major . The first diagnosis of N. major was provided by Thomas (1915a). The written account of N. major and the accompanying illustration of an animal from southwestern Western Australia was published by Gray (1875) but that illustration, accompanied with the name Nyctophilus major , was published separately and widely circulated privately in the 1840s (the publication date of major has been determined to be 1844 by Mahoney & Walton, 1988). Gray (1875) did not provide measurements or a description of N. major . His brief account consisted solely of a statement that he could not determine what species of Nyctophilus should be applied to his previously published plate.

The suspected holotype of timoriensis

Jansen (2017) noted that for birds collected by the Baudin expedition, none of the specimens have original field tags attached, and no original tags are known to have survived. Jansen indicates that original specimen data was communicated by the naturalists Peron and Lesueur to MNHN taxidermists and transcribed to pedestal bases. We suspect that the same applies to the Baudin mammal material, and we note that the identity of the type specimen of timoriensis is uncertain. The earliest registers of bird and mammal specimens in the MNHN that assign specimen numbers began in the early 1840s ( Jackson et al., 2021), and it is possible that the identity of Geoffroy’s original material might have become confused before the 1840s.

The specimen currently labelled the holotype of “ Nyctophilus timoriensis ” is CG1990- 36 in the MNHN. Although forearm length is not given in the original description, this specimen is a medium-sized Nyctophilus with forearm length of 43 mm. It is a puppet skin (see Fig. 4a,b ) from which the skull has been extracted at an unknown date and is now apparently lost ( Figs 4c–e ). Three other numbers are associated with the skin. The first published attribution of type status to this specimen is the catalogue of bat type specimens by Rode (1941), stated that the skull was lost and who might have assigned the number 217 to the skin. We have not found any other reference in the literature to the skull. The skin has 884 written in old ink on the right wing, to the left of which is some partially obliterated writing ( Fig. 4f ). There were no tags associated with the skin when examined in Sydney by HP in 1990. Later in 1990, Michel Tranier inventoried the MNHN collection and apparently registered the specimen as CG 1990-36 and concluded that there were no other specimens in the collection that could be Geoffroy’s type material (see Parnaby, 2009). It appears that Tranier also added numbers to an old board label that could have originated from the 19th century when the species was referred to Plecotus ( Fig. 4g ). A tag now attached to the specimen, presumably by Tranier, has an additional number 160a, evidently an old pedestal number ( Fig. 4h ). That number is listed in the Nouveau Catalogue des Galeries (New Catalogue of Galleries for skin mounts) ( Jackson et al., 2021), which commenced around the 1840s. We do not know the source of the pedestal number 160a. Perhaps it originated from the initial taxidermy procedures immediately following the Baudin expedition, as described for bird specimens by Jansen (2017).

We are unable to definitely establish that CG 1990-36 is Geoffroy’s original specimen, but we have no reason for rejecting it, other than the incongruent head-body length of 70 mm. However, detecting a mis-matched skin of a similar-sized Nyctophilus species is hampered by the poor condition of the skin. Our measurements of the putative type wingspan and tail length are consistent with Geoffroy’s (1806) account: our wingspan measurement of 264 matches his 270 mm given that part of the wing tip is missing ( Fig. 4 ) and our measurement of tail length, taken from the base of the body, equates to his 40 mm. Thomas (1914) reported a forearm length of 42 mm for the alleged type of timoriensis , taken by Trouessart at MNHN. This is a close match with our measurement of 43 mm taken on the putative type CG 1990- 36. Crucially, we have not been able to determine the nature of the post-nasal elevation, which is hidden in the shrivelled skin folds. Finally, we note that Geoffroy (1806) did not state the number of timoriensis specimens that he examined, but he does not mention more than one and it has generally been assumed that his description was based on a single type specimen. Parnaby (2009) mistakenly believed that there were two specimens; his confusion arose from a statement by Temminck (1840) to that effect, but Temminck was instead referring to material of Vespertilio peronii I. Geoffroy, 1832 . On balance, we accept that MNHN CG 1990-36 is possibly Geoffroy’s original specimen of timoriensis .

Comparisons with Maubisse specimens

Nyctophilus timoriensis sensu stricto

The Maubisse specimens share several features with our revised concept of Geoffroy’s N. timoriensis : both are of medium body size, with a conspicuously developed post-nasal mound, similar to species such as N. geoffroyi , N. heran , and N. gouldi . The largest Maubisse specimen, though not fully mature, is of medium body size for the genus as reflected by a forearm length of 40 mm compared with 43 mm for the suspected N. timoriensis holotype. Field head-body measurements for male AM M.38841 and female AM M.37639 from Maubisse approximate that of the suspected type specimen (56 and 50 mm vs. ca. 53 mm).

Nyctophilus heran and N. geoffroyi

Our comparison with these two species focuses on the Maubisse male AM M.38841, given that sexual size dimorphism occurs within Nyctophilus (males averaging smaller) and that the unique specimen of N. heran and the holotypes of geoffroyi and pallescens included in our analyses are all males. Measurements of an extensive series of N. geoffroyi from throughout Australia indicates that Tasmanian animals average larger than those from mainland Australian. Tasmanian specimens are excluded from further consideration here, given that mainland Australian “ N. geoffroyi ” is a composite of at least two species ( Eldridge et al., 2020) and the taxonomic status of Tasmanian populations has not been assessed. Our analyses treat N. geoffroyi as one entity because we did not identify any obvious geographic or morphological groupings in the morphometric data.

The Maubisse specimens undoubtedly most resemble N. geoffroyi and N. heran in overall morphology rather than any other species of the genus (other than N. timoriensis ) based on the following combination of external, cranial and bacular characters:

1 Overall body size is medium within the genus.

2 The post-nasal prominence is well developed, consisting of two elevated mounds tapering to their distal tips and joined in the midline by an elastic membrane, forming the “Y” shape characteristic of N. geoffroyi (see Fig. 5 ).

3 The external ears are large relative to body size, and joined in the midline for a substantial proportion of the length of the ear and general colour of body fur is a mouse-brown dorsally, with much paler ventral fur.

3 The auditory bullae are large relative to skull size ( Fig. 6 ).

4 Baculum shape is consistent with 12 specimens of N. geoffroyi examined from mainland Australia and resembles that of the holotype of N. geoffroyi pallescens illustrated by Hill & Harrison (1987). A slight groove is evident in the distal tip of M.38841 ( Fig. 7 ). It is not clear if this represents incomplete ossification in this subadult animal, but a similar indentation is apparent in the holotype baculum of N. geoffroyi pallescens . Baculum length of AM M.38841 falls within the range for 12 specimens N. geoffroyi , and height and breadth are smaller in this subadult animal ( Table 2 ). Baculum morphology of N. heran remains imperfectly defined ( Kitchener et al., 1991).

The Maubisse male M.38841 is compared against the following differences between N. heran and N. geoffroyi cited by Kitchener et al. (1991):

1 Nyctophilus heran has smaller and less inflated bullae relative to skull length. The bullae of N. heran are smaller and less inflated than N. geoffroyi . In absolute size, BUL of N. heran falls within the size range for N. geoffroyi . However, the larger skull size of N. heran results in relatively smaller and less inflated bullae, as illustrated in a plot of BUL vs GL ( Fig. 8a ) and BTB vs MAS ( Fig. 9 ). The bullae of M.38841 appear to be relatively larger than those of the holotype of N. heran as evident in the lateral skull view (compare Fig. 6 with figure 2 of Kitchener et al., 1991), in which M.38841 is far more typical of N. geoffroyi . However, BUL of M.38841 is at the low end of the size range for N. geoffroyi ( Table 2 ) and a bivariate plot of BUL vs GL ( Fig. 8a ) indicates that relative to GL, BUL is smaller relative to most specimens of N. geoffroyi . Kitchener et al. contrast the smaller ratio BUL / GL of the holotype of N. heran (0.233) compared to the smallest ratio of 0.247 in their sample of six male N. geoffroyi . The ratio of 0.232 for M.38841 is similar to that of N. heran but this ratio ranged from 0.225 –0.314 in our sample of 53 adult male skulls of N. geoffroyi from throughout mainland Australia. However, the ratio exceeded 0.237 in all but one of the 53 males that we measured, and the trend for a relatively larger BUL in N. geoffroyi is clear.

2 Nyctophilus heran has a more sharply angled anterior edge of the mesopterygoid fossa. The anterior edge of the mesopterygoid fossa is gently curved toward the base of the post-palatal spine in M.38841 ( Fig. 6 ), similar to N. geoffroyi , but in contrast to the more linear margin in the holotype of N. heran . We have examined photographs of the holotype skulls of pallescens and pacificus , both of which resemble that of M.38841. However, this feature is not invariant, and occasional mainland Australian N. geoffroyi specimens had angled edges.

3 Nyctophilus heran has a more pronounced post-palatal spine. The post-palatal spine of M.38841 is relatively shorter, similar to that of the holotypes of pallescens and pacificus and other N. geoffroyi specimens examined, compared to N. heran .

4 Hypocones on M 1 and M 2 more developed in N. heran . The hypocones of M.38841 are present but are relatively undeveloped. Kitchener et al. (1991) state that the hypocones are more developed than in N. geoffroyi and we assume that hypocone development in the latter species is variable, given that they are absent in the N. geoffroyi that we examined. We are unable to evaluate this further because we cannot clearly discern hypocone morphology from the illustrations of N. heran given by Kitchener et al. (1991), although it appears that they are more developed than those of M.38841. We note that the latter authors did not include this character in their diagnosis of the species. Cusp terminology used by those authors is possibly the same as in figure 2 of Kitchener & Caputi (1985).

5 Nyctophilus heran has a relatively longer third commissure on M 3. The third commissure is relatively much shorter in M.38841 compared with N. heran . Kitchener et al. (1991) suggest that the greater development of the third commissure has resulted in a greater M 3 width than that of N. geoffroyi , however M 3 length and breadth of M.38841 approximates that given for N. heran , and the likely level of measurement error suggests that M 3 is effectively the same size as the holotype of N. heran , both of which fall at the upper end of the range for the six N. geoffroyi measured by Kitchener et al. (1991).

6 Nyctophilus heran has a less rounded distal end on the glans penis.

7 Dorsal crest on the glans penis is absent in N. heran . The external morphology of the glans penis of M.38841 resembles that of the holotype of N. heran (see fig. 4 of Kitchener et al., 1991), rather than that of N. geoffroyi , in having a broadly rounded distal tip, and no dorsal crest.

8 Larger absolute size of N. heran . The holotype of N. heran is clearly larger in overall size than mainland Australian N. geoffroyi of equivalent sex, as noted by Kitchener et al. (1991). This is evident, for example, in bivariate plots of ZYG vs GL ( Fig. 8b ) and GL vs FA ( Fig. 8c ), in which N. heran falls well outside N. geoffroyi but close to N. daedalus . We have added the latter species to these plots as a yardstick to the magnitude of interspecific differences that can occur for Nyctophilus species. The four examples of larger skull and dental dimensions cited by Kitchener et al. (1991) are GL, ZYG, BRH and CM 3, all of which are corroborated by our much larger sample sizes ( Table 2 ).

Morphometric comparisons with N. heran and mainland Australian N. geoffroyi

Skull and external measurements of M.38841 fall within the size range of mainland Australian adult male N. geoffroyi for most dimensions (though smaller than the range for INT). However, given the specimen is subadult, it likely has not attained fully adult size, compromizing morphometric comparisons overall. In contrast, this specimen is at the upper size limit of specimens measured in this study, for sinPAL, BTB and M 3 B, for which it approaches the size of heran ( Table 2 ). AM M.38841 falls within the range of variation of N. geoffroyi as illustrated in bivariate plots of BUL, ZYG, and BRH vs GL, GL vs FA ( Fig. 8a–d ), and EAR vs FA ( Fig. 10 ). In contrast, the holotype of N. heran falls outside the range of N. geoffroyi in these plots and exceeds the upper range of mainland N. geoffroyi for most characters other than BUL and EAR, which fall within the range ( Table 2 ).

A comparison of M.38841 with N. geoffroyi and N. heran was explored further in a principal components analysis based on 9 skull and dental dimensions of 75 mainland Australian adult male N. geoffroyi . Separate analyses using a correlation matrix and a variance-covariance matrix yielded similar trends, with the holotype of N. heran a clear outliner in both. The PCA explained 59.6 and 12.1% of variance on the first and second PC axes respectively, compared to 71.1 and 7.6% in the variance-covariance analysis and we only present results of the latter. The first three PC axes account for a substantial percentage of the measurement variance ( Table 3 ) and character coefficients suggest the first PC axis is dominated by overall size, while PC 2 contrasts BRH, with M 3 –M 3 and BUL. A plot of PC scores on the first two PC axes, and on PC 1 vs. PC 3 ( Fig. 11 ) indicate that scores for AM M.38841 fall within the range of N. geoffroyi , while those of the holotype of N. heran are an outlier on the first two axes but not on PC 3. A minimum spanning tree fitted to each specimen in the PCA plots (not shown) revealed that the holotype of N. heran is a clear outlier on a plot of PC 1 vs. PC2, and PC 1 vs. PC 3, while that of the Maubisse male falls within the range of variation of N. geoffroyi .

We further compared skull and dental measurements of N. heran and M.38841 with the same sample of mainland Australian N. geoffroyi in dendrograms from UPGMA cluster analyses using euclidean distance as a measure of similarity. The holotype of N. heran formed an outgroup to both M.38841 and all mainland N. geoffroyi in all 10,000 boot-strap replications, in which there was little or no support for subgroupings within N. geoffroyi and M.38841 was an outgroup to mainland N. geoffroyi in 37% of replications (not shown). This suggests that no meaningful substructure was detected within N. geoffroyi and M.38841 with this character set using this technique.

The subadult female specimen M.37639 was at the most advanced growth stage of the three Maubisse specimens and falls at the upper end of the size range of 70 adult female mainland Australian N. geoffroyi as shown in a plot of Ear Length vs. FA ( Fig. 10 ). Its measurement of C 1 –C 1 of 4.26 mm falls within the range of 3.8–4.8 mm of 70 adult female N. geoffroyi from mainland Australia.

Summary of species comparisons

Seven potentially diagnostic criteria are available to compare the Maubisse male with N. geoffroyi and the original description of the holotype of N. heran . The character states shared by the three entities are summarized in Table 4 . The Maubisse male has a unique combination of characters shared with both N. heran (glans penis morphology and BTB) and N. geoffroyi (skull and dental morphology). This suggests that the Maubisse animals could be a separate taxon. The Maubisse male has only three of the seven criteria in common with the holotype of N. heran , but four with N. geoffroyi .

Ontogenetic changes in size and shape potentially influence at least three of the five criteria this specimen shares with N. geoffroyi . The relatively enlarged bullae, which are very characteristic of N. geoffroyi , possibly result from differential skull growth. If the cranial cavity and auditory regions attain near adult dimensions earlier than rostral dimensions, the Maubisse male might be a subadult specimen of N. heran irrespective of its relatively large bullae. Further support for this interpretation stems from the large absolute size of BTB, for which the Maubisse male matches that of N. heran and exceeds the largest of 45 male N. geoffroyi ( Table 2 ). However, this is contradicted by BUL of the Maubisse male, which is at the lower end of the range of N. geoffroyi and is substantially smaller than N. heran .Although the Maubisse male could have skull proportions not shared with either N. heran or N. geoffroyi , these differences potentially result from differential growth rates of bullae relative to the cranial vault. This could be resolved when adult material becomes available from Timor, and additional material enables an assessment of intraspecific variation in N. heran . Although the overall size of the Maubisse male falls within the range of N. geoffroyi for most individual dimensions and also in the PCA based on skull and dental dimensions, the animal is at a fairly early stage of development based on the extent of fusion of the wing epiphyses and it might not have attained full adult size. The relative length of the post-palatal spine is similarly problematic, and further growth in the length of this structure cannot be discounted. However, we suspect that the anteriorly more rounded shape of the mesopterygoid fossa of the Maubisse male reflects the adult state.

Figures

Figure 3. Illustration of the type specimen ofVespertilio timoriensis È. Geoffroy, 1806, reproduced from Plate 47 of Geoffroy’s original account, showing tragus and snout features characteristic of the genus Nyctophilus. A post-nasal snout mound is evident within the red ellipse of the inset. Note that the black horizontal line is part of the original illustration and represents head length but is not a scale bar for the illustration. Scale bar (white) = 5 mm.

Figure 4. MNHN CG1990-36, study skin with skull extracted, alleged holotype of Vespertilio timoriensis È. Geoffroy, 1806. (a), skin dorsal view, scale bar = 2 cm (photo: Ken Aplin); (b), skin ventral view, scale bar = 2 cm (photo: Ken Aplin); (c), dorsal view of snout (photo: Ken Aplin); (d), oblique view of snout showing nose-leaf (photo: Ken Aplin); (e), lateral view of snout (photo:Anja Divljan); (f), old writing and numerals on ventral surface of left plagiopatagium (photo:Anja Divljan); (g), front and reverse view of associated, presumed 19th century tag (photo: Anja Divljan); (h), front and reverse view of contemporary skin tag (photo: Anja Divljan).

Figure 5. Snout morphology of (a) subadult female Nyctophilus from Maubisse (AM M.37639, ET3), and (b) an adult male N. geoffroyi from Napier Downs Station, Kimberleys, Western Australia (AM M.22122) (imaging: Sue Lindsay). Scale bar represents ca. 1 mm.

Figure 6. Micro CT scans of skull and dentary of AM M.38841. Scale bar represents ca. 2 mm. (images: Fred Ford).

Figure 7. Micro CT scan of baculum of AM M.38841. (a) lateral view; (b) dorsal view. Distal end of bone faces left. Scale bar represents ca. 2 mm (images: Fred Ford).

Figure 8. Bivariate plots of adult male Nyctophilus: (a) BUL vs GL (n = 71 for N. geoffroyi); (b) ZYG vs GL (n = 79 for N. geoffroyi); (c) GL vs FA (mm) for male Nyctophilus geoffroyi (n = 50), and (d) BRH v GL (n = 75). Symbols are: N. geoffroyi from throughout mainland Australia (○); holotype of Nyctophilus geoffroyi pallescens (●); holotype of Nyctophilus geoffroyi pacificus (▲); holotype of N. heran (■); subadult male M.38841 from Maubisse, Timor-Leste (♦); N. daedalus are small dots bounded by polygon, for a comparison with a similar-sized species.

Figure 9. Plot of BTB vs MAS, for adult males: N. geoffroyi (n = 50) from throughout mainland Australia (○); holotype of N. heran (■); subadult male M.38841 from Maubisse, Timor-Leste (♦); N. daedalus (n = 15) are small dots bounded by polygon, for a comparison with a similar-sized species.

Figure 10. Plotof Ear vs FA (mm) of Nyctophilus:(a) males:adult male N.geoffroyi (n= 122) from throughout mainlandAustralia (○); holotype of N. heran (■); subadult male M.38841 from Maubisse, Timor-Leste (♦); N. daedalus (n = 18, small dots) are bounded by polygon, for acomparison with a species of similarsize to N.heran. (b) Adult female N. geoffroyi from throughout mainland Australia (n = 99); subadult female M.37639 from Maubisse, Timor-Leste (▲). Locality codes for N.geoffroyi are: centralAustralia (c); NSW (n), Queensland (q); Victoria (v), northern Western Australia and northern Northern Territory (z). Thirty specimens from Danggali, SA (d), are bounded by a polygon, indicating extent of within locality variation relative to total variation.

Figure 11. Plot of PC scores for the first three axes of a PCA based on 9 skull and dental measurements of 75 adult male specimens of N. geoffroyi; AM M.38841 from Maubisse (ET), holotype specimens of pacificus (P), pallescens (A) and the holotype of N. heran (H). Mainland Australian N. geoffroyi symbols designate specimen localities: central Australia (c); NSW (n), Queensland (q); Victoria (v) and northern Western Australia and northern NT (z).