Calumma juliae, David Prötzel & Miguel Vences & Oliver Hawlitschek & Mark D. Scherz & Fanomezana M. Ratsoavina & Frank Glaw, 2018

David Prötzel, Miguel Vences, Oliver Hawlitschek, Mark D. Scherz, Fanomezana M. Ratsoavina & Frank Glaw, 2018, Endangered beauties: micro-CT cranial osteology, molecular genetics and external morphology reveal three new species of chameleons in the Calumma boettgeri complex (Squamata: Chamaeleonidae), Zoological Journal of the Linnean Society 2018 (20), pp. 1-28 : 15-20

CALUMMA JULIAE SP. NOV.

urn:lsid:zoobank.org:act:6A92C95D-6383-45DC-9FFC- 0943B08F064F

Remark: Despite intensive research at the type locality in January (by L. Randriamanana, F.G. and D.P.), August (by N. Raharinoro, K. Glaw, T. Glaw, J. Forster, F.G. and D.P.) and November 2016 (by A. Rakotoarison and M.D.S.) in the rainy and dry seasons, only female specimens of this new species were found. In total, eight adult females, two subadult females and two juveniles were encountered (but not all of them collected).

Holotype: ZSM 143/ 2016 ( FGZC 5235 ), adult female, collected in small forest fragment 5 km east of Moramanga, just south of the Route Nationale 2 (18.9520°S, 48.2707°E, at 950 m a.s.l.), Toamasina Province, eastern Madagascar on 6 January 2016 by F. Glaw, D. Prötzel and L. Randriamanana.

Paratypes: ZSM 142 / 2016 ( FGZC 5233 ) , FGZC 5232 and FGZC 5234 (two uncatalogued specimens in UADBA), all three adult females, collected from the same location as the holotype (18.9519° S, 48.2705° E, within a radius of 50 m, at 950 m a.s.l.) on 6 January 2016 by F. Glaw, D. Prötzel, and L. Randriamanana . ZSM 254 / 2016 ( FGZC 5274), adult female, and ZSM 255/ 2016 ( FGZC 5275) and FGZC 5276 (uncatalogued specimen in UADBA), both juveniles, all three collected on 30 July 2016 ; FGZC 5277 (uncatalogued specimen in UADBA), subadult, collected on 31 July 2016 at the same location as above by F. Glaw, D. Prötzel, and N. Raharinoro.

Diagnosis: Male specimens are unknown so far, hence the diagnosis refers only to females of this species. Calumma juliae sp. nov. is a member of the phenetic C. nasutum species group ( Prötzel, Ruthensteiner & Glaw, 2016), on the basis of the presence of a soft, dermal unpaired rostral appendage, absence of gular or ventral crests, and heterogeneous scalation on the lower arm, consisting mostly of tubercles of a diameter of 0.7–0.8 mm. Within the group, it is a large (TL 105.3–111.6 mm), grey–beige chameleon that is characterised by a long and distally rounded rostral appendage, a dorsal crest of 11–14 tubercles, occipital lobes that are clearly notched but not completely separated, and absence of axillary pits.

Calumma juliae sp. nov. differs from C. fallax , C. gallus, C. nasutum, C. peyrierasi , C. vatosoa, and C. vohibola of the C. nasutum group by the presence of occipital lobes; from C. gehringi, C. guibei, and C. lefona sp. nov. in the completely closed brain case (vs. frontoparietal fenestra); additionally, from female C. gehringi and C. guibei in body size of 53.3–59.4 mm SVL (vs. 47.5–52.3 mm); from female C. gehringi in the shorter rostral appendage of 2.3–2.7 mm (vs. 3.2–4.4 mm); from C. guibei (both sexes) in the notch between the occipital lobes of 0.2–0.8 mm (vs. completely separated with notch of 1.2–1.9 mm; see Brygoo, 1971); from C. lefona sp. nov. (one male) in the shorter (2.3–2.7 mm) and rounded rostral appendage (vs. 5.6 mm, pointed), the absence of a temporal and parietal crest (vs. presence) and the number of dorsal cones of 11–14 (vs. 23); from C. boettgeri (both sexes) by the higher number of large (0.7–0.8 mm diameter) juxtaposed tubercle scales on the extremities (17–19 in line vs. 7–14, diameter of 0.2–0.5 mm and isolated from each other); from female C. uetzi sp. nov. in the larger body size of 53.3–59.4 mm SVL (vs. 42.0 mm SVL in females), the absence of a temporal and parietal crest (vs. presence of both) and the higher number of infralabial scales of 14–15 (vs. 11–12); from its most similar taxon C. linotum by the clearly notched occipital lobes with a depth of 0.2–0.8 mm (vs. not or slightly notched with depth 0–0.2 mm), presence of a dorsal crest in females consisting of 11–14 conical scales in C. juliae sp. nov. (vs. zero in C. linotum and six in C. cf. linotum; one specimen from Andampy), higher number of infralabial scales in females of 14–15 (vs. 12–13), absence of temporal and parietal crest in females (vs. both present), and in generally larger body size in females of 53.3–59.4 mm SVL (vs. 42.7–54.5 mm). In skull morphology, the squamosal and parietal do not meet in female C. juliae sp. nov. (vs. broad in contact in male and female C. linotum ( Prötzel et al., 2015); frontal of triangular shape and narrower, e.g. 16.8–18.2% of SkL at border to prefrontal (vs. 22.1–24.4%), 30.7– 31.1% of SkL at border to postorbitofrontal (vs. 34.4– 35.8%) and 25.0–26.3% of SkL bordering the parietal (vs. 32.5–33.6%); also, parietal is narrower at border to frontal with 28.5–29.5% of SkL (vs. 34.1–35.1%).

Description of the holotype ( Fig. 3C ): Adult female, in a good state of preservation; SVL 55.8 mm, tail length 55.4 mm; for other measurements, see Table 1; distinct rostral crest, laterally compressed dermal rostral appendage that projects forward of the snout tip over a length of 2.7 mm with a diameter of 2.2 mm, rounded distally; 14 infralabial and 13 supralabial scales; supralabials dorsally serrated; no supra-orbital crest; distinct lateral crest running horizontally; no temporal crest; occipital lobes clearly developed and separated but still slightly connected by a notch of 0.5 mm; casque raised; dorsal crest present, starting 4.3 mm from the base of the notch between the occipital lobes, consisting of a row of 12 separated, small conical scales spaced at regular intervals of ~ 2 mm, continuing on the tail with smaller and narrower spaced cones; no traces of gular or ventral crest. Body laterally compressed, with fine homogeneous scalation with the exception of the extremities and head region; limbs with rounded tubercle scales of maximal diameter 0.7 mm; heterogeneous scalation on the head, with largest scale on temporal region having a diameter of 1.3 mm; no axillary or inguinal pits.

Skull osteology of the holotype ( Fig. 4C , Supporting information, Video S4): Skull length 13.2 mm; snout– casque length 16.1 mm; broad paired nasals meeting each other; anterior tip of frontal exceeding less than half of the naris and separated from premaxilla; prefrontal fontanelle and naris fused; frontal smooth, and parietal with only a few tubercles; frontal with width of 2.4 mm (18.2% of SkL) at border to prefrontal extending to 4.1 mm (31.1% of SkL) at border to postorbitofrontal; broad parietal tapering strongly from a width of 3.9 mm (29.5% of SkL) at the border to frontal to 0.8 mm (6.1% of SkL) at midpoint, then extending mostly straight beyond the posterodorsal extension of the squamosals, and finally tapering again at its tip; posterodorsally directed parietal does not meet the squamosal; squamosal thin without tubercles. For further measurements, see Table 2.

Coloration of the holotype in preservative ( Fig. 3C ): The body of the holotype in preservative is of dark grey colour without any distinct pattern; inner side of extremities and tail beige; extremities and temporal and postocular region of the head speckled with greyish–blue tubercle scales.

Variation: All adult female C. juliae sp. nov. that have been found so far show a consistent morphology and consequently agree well with the holotype. ZSM 254/2016 and FGZC 5234 have a slightly shorter rostral appendage than the holotype, at 2.4 and 2.3 mm (vs. 2.7 mm); the notch between the occipital lobes is deeper in FGZC 5232 at 0.8 mm and less deep in ZSM 142/ 2016 at 0.2 mm (vs. 0.5 mm); the dorsal crest of ZSM 255/ 2016 consists of only nine tubercles (vs. 12); this might be referred to its juvenile developmental stage. In osteology, no significant variations were found.

Coloration in life ( Fig. 7 ): Females are indistinctly grey–beige coloured; a netlike dark brown pattern on the skin between the scales and two dorsoventrally compressed blue blotches may occur on the body; tubercle scales on extremities and body may be bright green; rostral appendage not highlighted and of same colour as the body; a dark lateral stripe may stretch from the rostral appendage across the eyes to the occipital lobes. If stressed, only the head coloration changes to dark brown or green pattern at the dorsal head region, eyelids with radially aligned blue/violet spots, and rostral crest and appendage of turquoise/ green colour.

Available names: Apart from C. boettgeri and C. linotum, there is no other valid species or synonym in the C. nasutum group with slightly notched occipital lobes.

Etymology: D.P. dedicates the first new species he discovered himself to Julia Forster, in recognition of her generous support and understanding of our research on Madagascan chameleons and her help in collecting specimens of this species.

Distribution: Calumma juliae sp. nov. has so far been recorded only from a fragment of degraded primary rainforest, covering an area of just 15 ha, east of Moramanga (Toamasina Province, eastern Madagascar; Figs 6 , 8A ); for geographical coordinates, see the descriptions in the subsections ‘Holotype’ and ‘Paratypes’ above. The forest fragment spreads over a hill that rises to a maximum elevation of 1010 m a.s.l.; specimens were found only along the trail at 950 m a.s.l., but we expect them to occur on the hill as well.

A further specimen was observed just north of the Route Nationale 2 (18.9513°S, 48.2721°E, 950 m a.s.l.) in secondary bushes.

Natural history and ecology: Calumma juliae sp. nov. is an arboreal, diurnal species found in bushes and trees in a small forest fragment of degraded primary rainforest. Roosting sites at night were thin branches or, rarely, leaves that were not exposed, but hidden inside tree/bush cover 0.3–2 m above the ground. In contrast to the syntopically occurring population of C. cf. nasutum, the present species preferred a horizontal sleeping position (vs. head pointing downwards). When disturbed, some specimens dropped immediately and stayed curled and motionless on the ground. In appropriate habitat, specimens occurred a few metres from one another. In January 2016, only adult females were collected; in August 2016, additionally subadult and juvenile specimens; in November 2016, only one adult female was found. None of the collected females showed signs of being gravid. The size of the juveniles suggests that they hatched during the rainy season from approximately January to March. Under the occipital lobes of two specimens (ZSM 143/ 2016 and 142/2016) we found three and two mites, respectively; the lobes might function as mite pockets comparable to axillary pits in other chameleon species to limit and locate the damage caused by these ectoparasites (Arnold, 1986).

Recommended IUCN status: As no attempts have yet been made to estimate the population size or status of C. juliae sp. nov. directly, we suggest that it should be assessed under the IUCN Red List criterion B ( IUCN, 2012). The distribution area of the single known forest in which the species occurs has an area of ~ 0.15 km 2, which we interpret as the area of occupancy (AOO) of the species as defined by the IUCN (2012), but which at present is also equivalent to the EOO of the species (criterion B1). This constitutes a single threat-defined location (criterion B, subcriterion a). The forest in which it occurs is under heavy active anthropogenic pressure and is, in our opinion, in imminent danger of disappearance [criterion B, subcriterion b(iii)]. As the AOO of the species is considerably <10 km 2, consists of a single threat-defined location, and is experiencing on-going decline in its quality and extent, the species qualifies as Critically Endangered under IUCN criterion B1ab(iii).

OSTEOLOGY

Using micro-CT scans, we compared the skull morphology of the currently seven species of the C. boettgeri complex. Osteological measurements proved to be a useful tool to delimit these rather cryptic species. As a further important taxonomic character, C. gehringi ( Fig. 10A–C ), C. guibei, and C. lefona sp. nov. have a distinct frontoparietal fenestra (FF; also pineal or parietal foramen of other authors; see Eakin, 1973) that differs in size between the species. This fenestra also occurs in other species of the C. nasutum group, e.g. C. fallax ( Rieppel & Crumly, 1997) and C. cf. nasutum ( Table 4). Compared with the elevational distribution of the species, a highly significant correlation [P = 3.8207 × 10−6, r (Pearson) = 0.67] was found with presence/width of the fenestra and elevation ( Fig. 10D ). None of the species that live <1000 m a.s.l. has an FF, and all those> 1500 m a.s.l. have one. Additionally, FF also occur in C. peyrierasi and C. tsaratananense (Prötzel et al., 2018; A. van’t Padje et al., unpublished observations); both are montane species that occur> 1900 m a.s.l. (Brygoo, 1971, 1978), the latter of which is certainly not closely related to the C. nasutum group ( Tolley et al., 2013). All other investigated Calumma species have a closed skull roof (Prötzel et al., 2018; van’t Padje et al., unpubl.).

Figures

Figure 3. Holotypes of the three new Calumma species as preserved specimens.A, C. uetzi sp. nov. (ZSM 1688/2012). B, C. lefona sp. nov. (ZSM 2849/2010).C, C. juliae sp. nov. (ZSM 143/2016).Scale bar:20 mm.

Figure 4. Micro-computed tomography scans of the skulls of the three holotypes in dorsal and lateral view. A, male Calumma uetzi sp. nov. (ZSM 1688/2012). B, male C. lefona sp. nov. (ZSM 2849/2010). C, female C. juliae sp. nov. (ZSM 143/2016). D, male C.boettgeri (ZSM 440/2000).E, male C.linotum holotype.F, female C.linotum (ZSM 551/2001).G, male C.gehringi holotype (ZSM 2851/2010).H, male C.guibei (ZSM 2855/2010).Diagnostic characters are encircled in red. Abbreviations are given in the Material and Methods.Scale bars:2.0 mm.

Figure 7. Calumma juliae sp. nov. coloration in life during day.A, female holotype ZSM 143/2016 relaxed.B, juvenile ZSM254/2016 relaxed. C, D, portrait of female ZSM 254/2016 with stress pattern (C) and slightly displaying (D).

Figure 8. Distribution and habitat of Calumma juliae sp. nov. A, fragment of degraded primary forest of 15 ha next to Route Nationale 2; blue dots indicate records of C. juliae sp. nov. B, small fragment of primary forest (on the right) between rice fields and eucalyptus forest (on the left, behind).C, habitat of C. juliae sp. nov. next to rice fields; view from RouteNationale 2 inAugust2016.D, C. juliae sp. nov. in sleeping position photographed at night.Satellite imagery from Google Earth (17 December 2016).

Figure 10. DiceCT scan integrated into micro-CT scan of the skull of Calumma gehringi (ZSM 2840/2010), showing the position ofthe brain belowthe frontoparietal fenestra.A,lateral view on midsagittal section of three-dimensionalmodel.B, dorsalview on three-dimensional model.C, lateralview on midsagittal section(diceCT only).D, correlationof elevationand presence/width ofthe frontoparietalfenestra inthe Calumma nasutumgroup,based on datainTable4. Scale bars:2.0mm.

Figure6. DistributionmapofsevenspeciesoftheCalummaboettgericomplex inMadagascar.The typelocalityisunknown for Calumma linotum. Contour lines indicate steps of 200 m elevation. The record from Nosy Komba is based on Hyde Roberts & Daly (2014). Coordinates for Calumma boettgeri, Maromiandra (Nagy et al., 2012) in Prötzel et al. (2015) are corrected to 13.9965°S, 48.2177°E.