Varanus bushi, Aplin, Kenneth P., Fitch, Alison J. & King, Dennis J., 2006

Aplin, Kenneth P., Fitch, Alison J. & King, Dennis J., 2006, A new species of Varanus Merrem (Squamata: Varanidae) from the Pilbara region of Western Australia, with observations on sexual dimorphism in closely related species, Zootaxa 1313, pp. 1-38 : 24-29

Varanus bushi sp. nov.

Figures 1 , 3–6

Holotype: Western Australian Museum R108999, adult male from Marandoo, Western Australia in 22 0 37’ S 118 0 08’ E. Collected on 20 June 1991 by Greg Harold. Specimen fixed in 10% formalin then stored in 70% ethanol. Liver sample stored in ­80 0 C ultrafreeze at WAM.

Paratypes: Three additional specimens from Marandoo: WAM R54230, an adult male collected on 25 October 1976; WAM R56834, an adult female collected in April 1977; and WAM R62171 a juvenile of undetermined sex collected on 28 February 1979.

Diagnosis: A small­bodied member of the subgenus Odatria distinguished from most others by the combination of a longitudinally striped and only moderately spinose tail, unkeeled head and body scales, non­overlapping ventral primary scales, and an absence of longitudinal streaks on throat. Distinguished from V. g i l l e n i by its slightly lesser average size, more elongate dorsal scales, more densely spotted venter and more irregularly spotted dorsum, less prominent linear patterning on the head and neck, and its more numerous presacral vertebrae, pedal subdigital lamellae and ventral scales. Males are further distinguished from V. gilleni by having hemipenes with an undivided inner hemibaculum. Distinguished from V. caudolineatus by its slightly greater average and maximum size, proportionally shorter fore­ and hind­limbs, more elongate snout, higher average midbody and ventral scale counts, higher average sub­digital lamellar counts on pes, more finely scaled and less rugose proximal portion of the tail, more numerous presacral vertebrae, less conspicuously spotted head, more orderly alignment of dorsal pattern into transverse rows, and presence of transverse bands on the basal one­third of the tail. Males are further distinguished from V. caudolineatus by having a shorter hemipenis with more numerous papillose distal frills.

Etymology: We take pleasure in naming this species after naturalist and educator Brian Bush who has contributed enormously to our knowledge of the herpetofauna of Western Australia and of the Pilbara region in particular.

Distribution and sympatry: Endemic to the Pilbara region of Western Australia, northwest to the vicinity of Cooya Poonya and Tambrey, southwest to Mt Brockman and Mt Tom Price in the Hamersley Range, southeast to Mt Whaleback and northeast to Marillana and Hope Downs ( Fig. 9 ). Two instances of regional sympatry with V. caudolineatus are recorded—at West Angelas and at Hope Downs, both on the southern margin of the Pilbara Uplands. Details of the relevant specimens are provided in the Discussion.

Description of holotype: Adult male measuring SVL 145 mm, Tail 190 mm, Forelimb 32 mm, Hind limb 43 mm, Head and neck 50 mm. Both hemipenes are fully everted.

Vertebral column includes 30 presacral vertebrae. Phalangeal formula of manus 1.2.3.4.2; and of pes 3.4.3.2.1.

Head moderately depressed, depth at pineal organ 8.7 mm. Snout relatively elongate; lacking canthus rostralis. Nostril positioned 6.8 mm from tip of snout; 3.4 mm from anterior corner of eye. Ear aperture is vertically narrow and obliquely oriented.

Dorsal head scales unornamented; most are irregular polygons, sub­rounded in shape, and majority have a single scale organ. Supraocular scales are smaller than those on the rostrum, frontal and parietal regions. Granules are absent from head except in temporal region and on throat forward into genal groove. Dorsal primary scales on neck are raised but lack keels; all have scale organs. The dorsal primary scales on neck are rounded anteriorly but become more ovate posteriorly, merging smoothly with body scalation. Ventral primary scales on neck are small and almost round anteriorly, but become larger and progressively more elongate to rear; all ventral primary scales on neck are ringed laterally and posteriorly by small granules. Approximately one third of ventral neck scales bear scale organs, usually one per scale but not infrequently two or three. Gular fold is distinct, located 5 mm forward of anterior base of forelimb, and consisting of 4 rows of small scales, all lacking scale organs.

Body scalation is relatively unspecialised. Dorsal primary scales are ovate, approximately twice as long as wide and lack keels; the majority have a single scale organ. Transverse scale rows are well­organised on the dorsum but tend to subdivide on flanks. Ventral primary scales are approximately 50% longer than and twice as wide as dorsal primary scales. All are ringed laterally and posteriorly by small granules. Single scale organs are present on the majority of scales towards the flanks but only about 15% of scales located along the mid­ventral area. Inguinal fold indistinct but located 1.5 mm forward of the anterior base of hind­limb. Position is marked by a change in scale size and shape, those behind the fold and back to vent being smaller, rounded and not arranged in transverse rows, and by occurrence of incomplete transverse series that terminate before reaching flank.

Scales on the inner surface of the fore­ and hind­limbs are small and rounded while those on the outer surfaces are larger, more elongate and weakly keeled. Plantar surfaces of manus and pes with primary scales on small mounds of fine granules; all primary scales bear scale organs and some are pigmented. Subdigital surfaces with transverse lamellae made up of two or three rounded primary scales surrounded by granules. The apical lamella on each digit is single and more intensely pigmented. Longest digit (IV) of manus has 20 lamellae; longest digit (IV) of pes has 23. Claws on manus and pes are moderately large and darkly pigmented; all are laterally compressed and bear sharp, recurved tips.

Tail is slightly wider than deep at base, becoming more rounded in cross­section towards tip; it lacks a dorsal keel or other obvious specialisation. Dorsal primary scales near tail base are elongate but otherwise unspecialised. Moving distally, the primary scales first become more elongate, then develop a distinct midline keel, and finally develop smaller lateral keels. Ventral primary scales also change from elongate and lacking keels near the tail base, to even longer and with a strong midline keel distally. The overall effect is that the tail feels relatively smooth near the base but becomes more distinctly rasp­like distally. Circumferential scale counts decrease from 52 near tail base to 32 at one third of total length, then to 22 at two­thirds of total length. Longitudinal scale counts at the same positions are 8, 7 and 6 rows per cm.

The post­cloacal scale cluster is relatively poorly developed and consists of two rows of modified but unpigmented scales. The outer row consists of six nodular scales of which the lateral four scales are distinctly spinose. The inner row consists of four smaller, bluntly pointed scales.

The hemipenis measures 6.3 mm without the projecting hemibacula and 6.8 including these structures. The nude basal portion measures 3.3 mm. Further details were provided in an earlier section.

Ground colour of dorsum in preservative is a uniform pale­brown from rostrum to base of tail, becoming greyer on the flanks. Patterning on the head includes an irregular mottling of dark brown on the dorsum and sides, tending towards longitudinal streaking on the occiput, and a subdued temporal stripe on each side, running from the posterior corner of the eye to above the ear. Irregular dark brown mottling is also present on the dorsum of the body, each ‘spot’ usually consisting of one or a few dark primary scales. The pattern is irregular on the anterior body but tends towards transverse alignment on the lower half of the back and above the hind limbs. The last three bars of this series are quite distinct and alternate with rows of small dark spots. The under surface of the throat, neck and body bears distinct pale grey spotting, each ‘spot’ usually comprised of two or three pigmented primary scales. These are most abundant on the throat and neck, and along the sides of the body, and less so in the mid­ventral region. The insides of the limbs are similarly patterned. The proximal 35 mm of the tail bears a series of eight more or less complete transverse bars. The next 38 mm bears a series of broken bars and irregular spots. The remainder of the tail supports a linear pattern consisting of five, more or less continuous dark stripes, each of which is one scale wide. The ventral surface of the tail is unpatterned throughout its length.

Variation among referred specimens: Mensural and meristic data for the complete sample of V. b u s h i sp. nov. is presented in Table 1 ; locality details for referred specimens are given in Appendix II. The holotype is the largest male specimen of V. b u s h i sp. nov.; SVL of the largest female ( WAM R135340) is 140 mm.

Dorsal patterning in most individuals is similar to that of the holotype, with fine spotting on the anterior body, conspicuous banding only on the rear of the body and on the proximal segment of the tail, and longitudinal striping of the distal tail. In some individuals (e.g. WAM R125521 from the ‘northern Pilbara’) the entire head and body are more intensely marked with variable sized spots; these are randomly distributed on the head and neck but are aligned into more or less regular transverse rows posterior to the level of attachment of the forelimbs. All specimens of V. b u s h i sp. nov. have ventral spotting on the throat, neck and body. In comparison with the holotype, most individuals show more intense and evenly distributed spotting on the body. Some individuals also show a strong differentiation between abundant fine spotting on the throat and neck, with each spot consisting of a single pigmented primary scale, and less abundant but larger blotches on the body.

Dorsal ground colour in life is pale grey­brown, typically with a reddish tinge on the upper surface, from the crown of the head to between the hind limbs ( Fig. 1 ). The ground colour of the tail is always paler than the body and generally has a cream wash on the distal one third of the upper surface.

Morphological comparisons: Varanus bushi sp. nov. is most similar overall to V. gilleni , a phenetic resemblance that is consistent with the phylogenetic conclusions of the molecular analyses. These species share a number of morphological attributes including a relatively elongate body with proportionally shorter limbs, high mid­body and ventral scale counts, high subdigital lamellar counts on the pes, a relatively non­spinose basal portion of the tail, and a tendency to include transverse bands in the dorsal patterning. As indicated in the diagnosis, V. bushi sp. nov. differs from V. gilleni in its lower average and maximum body size, its more elongate dorsal scales, its more densely spotted venter, its less conspicuous linear patterning on the head and neck, its more numerous presacral vertebrae, its simpler heimbacular morphology, and some minor meristic differences.

Var a nu s bushi sp. nov. is more similar to V. caudolineatus in body patterning and this probably explains the former confusion between these species. Both species share a spotted rather than banded pattern on the dorsum, and both are typically spotted on the throat and to some degree on the venter. However, Varanus bushi sp. nov. is usually less heavily pigmented than typical V. caudolineatus , with smaller and less intense spots. The lower back and basal portion of the tail are distinctly banded in V. b u s h i sp. nov. but more irregularly patterned in V. caudolineatus . Body proportions also distinguish the two species, V. b u s h i sp. nov. having a more elongate body form with more numerous presacral vertebrae, higher ventral scale counts, more elongate dorsal primary scales, and relatively shorter fore­ and hind­limbs. Although tail length is quite variable in both taxa, the majority of V. caudolineatus have proportionally shorter tails than V. b u s h i sp. nov. The head is also shorter and stockier in V. caudolineatus , produced mainly by a shortening of the rostrum. Midbody scale counts are typically higher in V. bushi sp. nov. and the hemipenes differ in several important details from those of V. caudolineatus .

The evolutionary polarity of most of the morphological characteristics that distinguish each of the three members of the V. caudolineatus species group is uncertain. Sprackland (1991) suggests that more heavily rugose tails and smaller body size are probably apomorphic within Odatria . On both counts, V. caudolineatus would rate as more derived than either of the other taxa. Varanus caudolineatus appears to be more primitive in hemipeneal morphology than either V. bushi sp. nov. or V. g i l l e n i; of the latter two species, V. gilleni possesses a more derived hemibacular morphology.

Taxonomic remarks: Although the type specimens of Varanus caudolineatus Boulenger, 1885 and Varanus gilleni Lucas & Frost, 1895 were not examined as part of this study, the type localities of each of these taxa [Champion Bay, W.A. (= Geraldton) and Charlotte Waters, N.T.; see Fig. 9 ] is remote from the known geographic range of V. bushi sp. nov. and well within the ranges of these other taxa as currently understood. We are therefore confident that the various names are correctly associated with the biological entities as defined herein.

Ecological notes: Relatively little information is available on the ecology of Pilbara varanids and the new taxon is no exception. Most recent specimens of V. bushi sp. nov. have been taken from fallen or standing hollow trees in mulga or eucalypt woodland associations.

The sex ratio among the total sample of V. b u s h i sp. nov. is 34 males to 13 females. This situation is typical for opportunistically collected samples of varanids ( King and Rhodes 1982; Greer 1989) and probably reflects different activity patterns between the sexes. The lack of immature specimens of V. b u s h i sp. nov. and the small numbers of very young individuals in the large samples of V. caudolineatus and V. g i l l e n i is consistent with previous comment that hatchling varanids are especially cryptic and poorly represented in wild caught samples ( Horn and Visser 1991).

Identifiable stomach contents were observed in two of four V. b u s h i sp. nov. examined by D.J. King. WAM R73142 (male, SVL 96 mm) contained remains of a spider and a skink tail; WAM R54230 (male, SVL 135 mm) contained remains of a mole cricket ( Gryllotalpidae ).

Figures

FIGURE 1. An adult Varanus bushi sp. nov. from near Roebourne, photographed in life by Robert Browne­Cooper.

FIGURE 3. Details of scalation on dorsal midbody of representative specimens of three Varanus species. (a) WAM R 135340 V. b u s h i sp. nov. Paratype Ψ; (b) WAM R 75792 V. gilleni Ψ; c) WAM R 135601 V. caudolineatus ♂.

FIGURE 4. Dorsal views of representative specimens of three Varanus species. (a) WAM R 125521 V. b u s h i sp. nov. Holotype ♂; (b) WAM R 135340 V. bushi sp. nov. Paratype Ψ; (c) WAM R 75792 V. gilleni Ψ; (d) WAM R 135601 V. caudolineatus ♂; (e) WAM R 138950 V. caudolineatus ♂.

FIGURE 5. Ventral views of representative specimens of three Varanus species. Same individuals as Figure 3.

FIGURE 6. Dorsal views of head and neck of representative specimens of three Varanus species. Same individuals as Figure 3.

FIGURE 9. Distribution of the three Varanus species based on examined materials and type localities. The approximate wider distribution of V. g i l l e n i is indicated by the dashed line. Symbols: Stars — V. caudolineatus; Triangles — V. gilleni; Circles — V. b u s h i sp. nov.; Circles with cross — s V. caudolineatus and V. b u s h i sp. nov. in local sympatry; Circles with dot – type localities.