Lygodactylus verticillatus, Mocquard, 1895

Puente, Marta, Glaw, Frank, Vieites, David R. & Vences, Miguel, 2009, Competitividad del Sector Forestal de México Tendencias y Perspectivas, Zootaxa 2103, pp. 1-76 : 44-59

Lygodactylus verticillatus group

Contains: Lygodactylus arnoulti , L. blancae , L. decaryi , L. heterurus , L. klemmeri , and L. verticillatus .

Although well-defined by one character (the whorled tail), this is in general a morphologically rather heterogeneous group (previously known as occidental lineage) in which we currently accept six species (including L. praecox as new synonym of L. klemmeri , and one subspecies of uncertain status, L. heterurus trilineigularis ). Molecular data ( Puente et al. 2005b) are in accordance with a possible monophyly of this group and support its affinities to Lygodactylus tolampyae .

Species in this group are defined mainly by the distinct whorls on the original tail, i.e., narrow rings of enlarged scales around the tail (which however may be indistinct or lacking in L. decaryi ). This character is not observed in any other Malagasy Lygodactylus , although several species have rings of scales of slightly different colour around the tail which remind whorls ( Fig. 21 ).

Further characters of all species in this group are as follows: dorsal scales granular; mental scale semidivided with usually three postmentals and five postpostmentals; first finger present, bearing a claw; males with more than nine preanal pores, usually 9–11, but 15 in L. decaryi . Two species ( L. heterurus and L.

klemmeri) with a distinct throat pattern of longitudinal black lines. The hemipenis is bilobed, with relatively long lobes, and covered with fields and serrated ridges of pointed papillae.

All species of the L. verticillatus group typically occur in arid parts of Madagascar, some ( L. verticillatus , L. decaryi ) in the extremely dry South, two species ( L. arnoulti and L. blancae ) in the central highlands, and one species ( L. heterurus , with the subspecies L. h. trilineigularis) in the West and reaching humid coastal areas in the North East.

The diagnosis of various species in this group is rather difficult and based on characters of rather high variability only (such as body size and coloration); however, considering also the distribution areas, a reliable diagnosis is possible in most cases.

Lygodactylus arnoulti Pasteur, 1964

( Fig. 22 )

Lygodactylus arnoulti Pasteur, 1964 .— Name-bearing type: holotype, male of 21 mm SVL, MNHN 1966.1001 (preliminary collection number of G. Pasteur, BGP 204 ), collected by M. de Saint Ours. — Type locality: “montagne de l’Ibity, Centre ”, according to the original description, “montagne de l’Ibity vers 1900 m sous pierres”, according to the MNHN catalogue.— Other types: no paratypes mentioned in the original description.— Etymology: dedicated to Jacques Arnoult.

Diagnosis. Lygodactylus arnoulti , a species endemic to the central massifs of Ibity and Itremo, is characterized as a member of the L. verticillatus group as defined above by its clearly whorled tail. It differs from other species of the group as follows: from L. decaryi , L. heterurus , L. klemmmeri , and L. verticillatus by the characteristic dorsal pattern with grey ground colour and ocellae or short dark transversal markings (vs. absence a clear dorsal pattern, sometimes with a tendency of longitudinal striation in L. verticillatus , light dorsolateral bands in L. klemmeri , or pairs of large light markings in L. heterurus ); from L. heterurus and L. klemmeri by the absence of longitudinal dark lines on the throat (vs. presence); from L. verticillatus by the absence of contact between the posterior projection of the mental scale and the first infralabials (vs. usually presence); from L. decaryi and L. verticillatus by a larger body size (mean SVL 30.2 mm vs. 25.1 and 23.1 mm); from L. decaryi by a lower number of preanal pores in males (9–11 vs. 15). Distinction is most difficult from L. blancae ; however, the dorsal pattern of that species is also less distinctly ocellated or marked with transversal markings than in L. arnoulti , and L. arnoulti usually has few or no dark spots on the throat whereas L. blancae has usually many such spots, sometimes forming indistinct longitudinal lines.

Description. (1) A moderately large Lygodactylus , adult SVL 21.5–36.9 mm (30.2± 3.7 mm, n=50); (2) TAL 14.8–35.9 mm (27.2± 5.9 mm, n=12); (3) granular dorsal scales; (4, 5) first finger present, bearing a claw (usually strong claws and relatively large fingers); (6) three pairs of lamellae under fourth toe (n=55); (7, 8) mental scale divided into three parts by sutures, a faint contact between posterior projection of mental and first infralabial can be present; (9, 10) three usually bisymmetrical postmental scales (n=55); (11) 5–7 postpostmental scales (5.40±0.59, n=55); (12) 4–6 infralabial scales (4.89±0.56, n=55); (13) 4–7 supralabial scales (5.76±0.51, n=55); (14) 1–3 internasal scales (1.47±0.54, n=55); (15) males with 9–11 preanal pores (9.71±0.84, n=22), and with enlarged and unpimented preano-femoral scales; (16) tail whorled, with 12–17 distinct whorls; (17, 18) usually with 3–6 dorsolateral tubercles, each composed of 1–7 scales; (19, 20) grey dorsal colour, usually with a pattern of indistinct ocellae or of regularly arranged transversal dark lines or markings; (21) ventrally light, often yellow (especiallly in males), usually with dark spots on the throat; (22) 182 dorsal scales along the body (n=1); (23) 75 dorsal scales around the body (n=1) ( Figs. 21a–b , 22 ).

Material examined. MNHN 1966.1001 (holotype, M. de Saint Ours, Mont Ibity, 1900 m, "sous pierres") , MNHN 1966.1002 (C.P. Blanc, Mont Ibity 1700 m) , MNHN 1990.1869 – 1871 (C.P. Blanc, 6 Jan. 1973, Ambatomenaloha ) , MNHN 1990.1872 – 1874 (C.P. Blanc, 15–16 Jan. 1973, Analabe ) , MNHN 1990.3398 – 3429 (C.P. Blanc, Dec. 1965, Mont Ibity ) , MNHN 1990.3429 – 3440 (C.P. Blanc, 14 Jan. 1973, South of Mont Ibity , Ampandrianombilapa) , ZSM 394–395 /2000 (M. Vences, 26 Mar. 2000, Mont Ibity ) , ZSM 490 /2001 (D. R. Vieites & M. Vences, 11 Mar. 2001, Itremo ) , ZSM 18 /2004 (F. Glaw, M. Puente, M. Thomas & R. Randrianiaina, 20 Jan. 2004, Col des Tapias ) .

Distribution. According to the specimens examined in this study, Lygodactylus arnoulti is known from the following localities: (1) Mont Ibity (type locality), (2) Col des Tapias (next to Mount Ibity) and Itremo (Ambatomenaloha and probably other sites). We are uncertain about the coordinates of the locality Analabe but assume that it also is located in the Ibity or Itremo area. No further localities are mentioned in the literature.

Habitat. According to Pasteur (1964) this species lives in semi-arid savannah areas at an altitude of about 1700 m. During numerous occasions (M. Vences in March 2000; M. Vences & D. R. Vieites in March 2001 and February 2008; F. Glaw, M. Puente, M. Thomas & D. R. Vieites in January 2003; F. Glaw, M. Puente, M. Thomas & R. Randrianiaina in January 2004) we observed the species active during the day on rocks and hidden under stones on the Mount Ibity, the Col des Tapias, and Itremo. While the sympatric L. blanci has only been found in the summital regions of Mount Ibity, L. arnoulti reaches lower elevations such as on the Col des Tapias. In 2001 we observed numerous eggs glued under a large flat stone lying on a rock on Mount Ibity; we assume this may have been a common egg-laying site of L. arnoulti which would thus be an egggluer.

Lygodactylus blancae Pasteur, 1995

( Fig. 23 )

Lygodactylus blancae Pasteur, 1995 .— Name-bearing type: holotype, MNHN 1990.39 , female, 32 mm SVL (according to the original description).— Type locality: “Ampefy, à 1 km au NW du lac Itasy" (NW de l’Ankaratra , vers 1400 m), according to the original description.— Other types: paratypes, MNHN 1990.40 – 42 and MHNG 2541.29 (adult males), MNHN 1990.43 (juvenile male), MNHN 1990.44 – 45 (recently born), MNHN 1990.46 – 47 (embryos), MNHN 1990.38 (adult male) “sur arbres près du lac Itasy”, according to the original description.— Etymology: dedicated to Françoise Blanc.

Diagnosis. Lygodactylus blancae , a species apparently endemic to a small area in the central highlands, is characterized as a member of the L. verticillatus group as defined above by its clearly whorled tail. It differs from other species of the group as follows: from L. heterurus and L. klemmeri by the absence of distinct and well-defined longitudinal dark lines on the throat (although a large number of black spots can be present on the throat which sometimes form irregular longitudinal lines) and a yellow venter in life, especially in males (vs. presence of well defined throat lines and absence of yellow pigment of venter in life); from L. decaryi and L. verticillatus by a larger body size (mean SVL 30.3 mm vs. 25.1 and 23.1 mm); from L. decaryi by a lower number of preanal pores in males (9–11 vs. 15); from L. verticillatus by a pigmented throat with distinct black spots that sometimes form longitudinal lines (vs. no distinct dark spots on throat). Distinction is most difficult from L. arnoulti ; however, this species has a clear dorsal colour and pattern (usually grey with a pattern of indistinct ocellae or of regularly arranged transversal dark lines or markings) whereas the pattern is more diffuse and variable in L. blancae , and L. arnoulti usually has only few or no black spots on the throat whereas in L. blancae the throat usually is densely spotted, the spots sometimes forming diffuse longitudinal lines.

Description. (1) a moderately sized species of Lygodactylus , adult SVL 24.3–35.0 mm (30.3± 3.7 mm, n=12); (2) TAL 20.2–31.9 mm (27.0± 6.1 mm, n=3); (3) granular dorsal scales; (4, 5) first finger present, bearing a claw; (6) three pairs of lamellae under fourth toe (n=14); (7, 8) mental scale divided into three parts by sutures, the posterior projection in indistinct contact with first infralabial; (9, 10) 2–4 bisymmetrical postmental scales (3.07±0.47, n=14); (11) 4–7 postpostmental scales (5.21±0.69, n=14); (12) 4–6 infralabial scales (5.43±0.64, n=14); (13) 4–7 supralabial scales (6.00±0.87, n=14); (14) 1–3 internasal scales (1.47±0.54, n=14); (15) males with 9–11 preanal pores (10.20±0.83, n=5) and with enlarged but unpigmented preano-femoral scales; (16) tail whorled, with 9–18 distinct whorls; (17, 18) dorsolateral tubercles absent; (19, 20) dorsally greyish to brownish, without a clear pattern, but often with striations around the eye; (21) ventrally light (usually yellow in life), usually with dense dark spots which sometimes form diffuse longitudinal lines; spots are also present on the throat and a characteristic dark line on the mental scale; (22) 184–194 dorsal scales along the body (188±4.64, n=4); (23) 58–62 dorsal scales around the body (60±2.3, n=4) ( Figs. 21c–d , 23 ).

Hemipenial structure. Based on ZSM 498/2001, adult male from Lac Itasy. Hemipenis subcylindrical, its total length ca. 4 mm, with a large pedicel measuring ca. 2.5 mm in length. The apex is divided into two lobes, each measuring ca. 1.5–2 mm. The truncus and the lobes are covered with serrated ridges and fields of pointed papillae. Sulcus spermaticus formed by one channel on the pedicel, dividing into two subchannels of which each runs to a terminal position on either lobe. No distinct sulcal lips, and no papillae along sulcus ( Fig. 23 ).

Material examined. MNHN 1990.38 (paratype, C.P. Blanc, 5 Nov. 1966, Lac Itasy ) , MNHN 1990.39 (holotype, C.P. Blanc, 13 Feb. 1973, Ampefy ) , MNHN 1990.40 – 45 (paratypes, C.P. Blanc, 13 Feb. 1973, Ampefy ) , ZFMK 48242–48243 (F. W. Henkel & R. Seipp, Apr. 1988, 70 km W Antananarivo) , ZSM 497–500 /2001 (D. R. Vieites, 22 Feb. 2001, Lac Itasy, hotel close to Ampefy ) .

Distribution. According to the specimens examined in this study, Lygodactylus blancae is only known from the type locality, Ampefy close to Lake Itasy. It is likely that all specimens have been found around one hotel near Ampefy.

Habitat. According to Pasteur (1995) this species was found living in small trees near the lake. According to our own observations (D. R. Vieites in 2001), specimens were active during the day on a single large tree in front of a hotel named Kavitaha, near Ampefy. The whole area around the Lac Itasy is largely deforested and there are very few rocks in the grassland. If the species indeed is endemic to this area, it clearly must be an extremely localized and rare species, but it is possible that it is also found on other mountains of central Madagascar where more trees and/or rocks provide appropriate habitat.

Lygodactylus decaryi Angel, 1930

( Fig. 24 )

Lygodactylus decaryi Angel, 1930 .— Name-bearing type: male holotype, MNHN 1930.271 , SVL 26.8 mm.— Type locality: “capturé dans le Massif de l’Angavo , province de Fort Dauphin, en julliet 1926; altitude 400 mètres; région sub-désertique. Se trouvait sous des écorces.”, according to the original description.— Other types: probably none; no information given in original description.— Etymology: dedicated to the collector, Raymond Decary: “Matériaux de la Mission R. Decary, en 1926.”

Diagnosis. Lygodactylus decaryi is a very poorly known species from dry areas in the South of Madagascar that we here assign to the L. verticillatus group although the state of its tail (whorled or not) is uncertain pending further study (see below). It differs from other species of the group as follows: from all species in the group by the number of 15 preanal pores in males (vs. 9–11, rarely 6, in the other species); from L. heterurus and L. klemmeri by the absence (vs. presence) of distinct and well-defined longitudinal dark lines on the throat; from L. arnoulti and L. blanci by the smaller body size (mean SVL 25.1 vs. 30.2 and 30.3 mm); from L. blanci by the absence of a dense pattern of black spots on the throat (vs. presence); from L. arnoulti by an undefined dorsal pattern (vs. well-defined pattern of grey with indistinct ocellae or regularly arranged transversal dark lines or markings).

From external characters, L. decaryi appears to be close to L. verticillatus , and a more detailed revision of these two species is necessary to understand their differentiation. If the tail of L. decaryi would be confirmed to be not whorled, it may also result to be related to the L. mirabilis group (but it has granular dorsal scales) or to L. tolampyae .

Description. (1) A relatively small species of Lygodactylus , adult SVL 22.2–26.8 mm (25.1± 2.5 mm, n=3); (2) no TAL data because all specimens examined with broken tail; (3) granular dorsal scales; (4, 5) first finger present and bearing a claw; (6) three pairs of lamellae under fourth toe (n=3); (7, 8) mental scale divided into three parts by sutures, its posterior projection sometimes with indistinct contact to first infralabial; (9, 10) three bisymmetrical postmental scales (n=3); (11) five postpostmental scales (n=3); (12) 4–5 infralabial scales (4.67±0.57, n=3); (13) 5–6 supralabial scales (5.33±0.57, n=3); (14) 1–2 internasal scales (1.33±0.57, n=3); (15) the single known male with 15 preanal pores, without enlarged preano-femoral scales; (16) tail with whorls in the female specimens, unverified in the male holotype which has a broken tail and which was reported to lack whorls ( Angel 1930); (17, 18) dorsolateral tubercles absent; (19, 20) brownish dorsal colour, without a well-defined dorsal pattern; (21) ventrally light, without throat spots, except in the specimen MNHN 1950.262 (female) in which is possible to see a few faint spots on the throat; (22) 190 dorsal scales along the body (n=1); (23) 66 dorsal scales around the body (n=1) ( Fig. 24 ).

Material examined. MNHN 1930.271 (holotype, Angavo massif) , MNHN 1950.262 ( Behara , Bevia forest) , MNHN 1930.270 ( Andrahamana ) .

Distribution. According to the specimens examined in this study, Lygodactylus decaryi is known from the following localities: (1) Angavo Massif (type locality), (2) Andrahamana, and (3) Behara. No further localities mentioned in the literature.

Habitat. According to Angel (1930) the holotype was found in a subdesertic region, at an elevation of 400 m above see level, in arid savannah. The specimen was collected under tree bark.

Remarks. Lygodactylus decaryi differs from all the other Malagasy Lygodactylus by the number of 15 preanal pores, but this is based on the holotype only. The two additional specimens MNHN 1950.262 (Behara, Bevia forest), MNHN 1930.270 (Andrahamana) are two females that are assigned to this species in the MNHN catalogue. Whether this identification is correct, and how L. decaryi differs from L. verticillatus which is widespread in the South of Madagascar, can only be determined once that new material becomes available and a detailed revision of this group is undertaken. The two additional specimens differ from the male holotype by having 5 versus 6 supralabials and one vs. two internasal scales.

According to the original description ( Angel 1930), this species lacks whorls on the tail. We could not verify this state in the male because the tail of the holotype was missing when examined in 2001. Two female specimens from Behara and Andrahamana assigned to this species had broken tails but we observed the presence of four whorls on the tail base (although less distinct than in L. verticillatus ). More material from the type locality of this species is necessary to ascertain its identity.

Lygodactylus heterurus Boettger, 1913

( Fig. 25 )

Lygodactylus heterurus Boettger, 1913 .— Name-bearing type: lectotype SMF 8953, designated by Mertens (1967), collected by A. Voeltzkow, Jan. 1897, according to Mertens (1967). However, in Boettger (1913: 323) it is stated (apparently by Voeltzkow himself) that Voeltzkow had collected on Nosy Be only in the years 1890 and 1895; Krüger (2001) considered SMF 8953 as holotype.— Type locality: Nosy Be.— Other types: two paralectotypes (the original description was based on one male and two females).— Etymology: no information given in original description.

Lygodactylus heterurus trilineigularis Rösler, 1998 .— Name-bearing type: male holotype, MTKD D39054.— Type locality: “Ampahana (Locus typicus), nordöstliches Madagaskar ”, according to the original description.— Other types: paratype, MTKD D39055, female.— Etymology: derived from latin tri - (three), linea - (lines), gula (throat), referring to the presence of three lines on the throat.

Diagnosis. Lygodactylus heterurus is characterized as a member of the L. verticillatus group as defined above by its clearly whorled tail. It is distinguished from L. arnoulti , L. decaryi and L. verticillatus by having 3–5 distinct longitudinal black stripes on the throat (vs. no black lines on throat). L. blancae and L. klemmeri also show black lines on throat, but these are only present in some specimens and then rather diffuse in L. blancae (vs. always present and very distinct in L. heterurus ), and they are slightly diagonal, directed towards the center of the throat in L. klemmeri (vs. not diagonal in L. heterurus ). Another species with black lines on the throat, here not assigned to the L. verticillatus group, is L. ornatus ; in this species, the throat lines are transversal, and L. ornatus in addition differs by its non-whorled tail. L. heterurus further differs from L. arnoulti and L. blancae by its white ventral colour in life (vs. mostly yellowish to deep yellow, especially in males); from L. decaryi by a lower number of preanal pores in males (9–11 vs. 15); from L. verticillatus by the grey dorsal colour with pairs of light grey markings (vs. more brownish colour without pairs of light markings); and from L. klemmeri by a dorsal pattern of pairs of light markings (vs. absence of such markings).

Description. Besides the type specimens, no morphological data have become available from the type locality Nosy Be. We therefore here provide first a brief summary of data from the female specimen MRSN R1911, collected at Nosy Be: (1) SVL 24.6 mm; (2) no data on tail length; (3) granular dorsal scales; (4, 5) first finger present, and bearing a claw; (6) three pairs of lamellae under fourth toe; (7, 8) mental scale divided into three parts by sutures, its posterior projection not in distinct contact with first infralabial; (9, 10) three bisymmetrical postmental scales; (11) five postpostmental scales; (12) five infralabial scales; (13) six supralabial scales; (14) three internasal scales; (15) no males examined for preanal pores; (16) tail distinctly whorled; (17, 18) dorsolateral tubercles absent.

The following is a comprehensive description, based on seven specimens, excluding the one from Nosy Be: (1) a small species of Lygodactylus , adult SVL 22.9–27.5 mm (24.8± 1.5 mm, n=7); (2) TAL 22.9–26.0 mm (24.2± 1.4 mm, n=4); (3) granular dorsal scales; (4, 5) first finger present and bearing a claw; (6) three pairs of lamellae under fourth toe (n=7); (7, 8) mental scale divided into three parts by sutures, its posterior projection can be in indistinct contact to first infralabial scale; (9, 10) three bisymmetrical postmental scales (n=7); (11) 5–6 postpostmental scales (5.14±0.38, n=7); (12) five infralabial scales (n=7); (13) 6–7 supralabial scales (6.14±0.37, n=7); (14) 1–2 internasal scales (1.71±0.48, n=7); (15) males with 9–11 preanal pores (9.75±0.95, n=4), and with enlarged preano-femoral scales which especially in specimens from the North East are distinctly pigmented; (16) tail with 14–15 distinct whorls; (17, 18) dorsolateral tubercles absent; (19, 20) grey dorsal ground colour, without a very distinct pattern, but often with a series of about five pairs of light grey markings; (21) ventrally white or light grey, with 3–5 longitudinal lines on throat; (22) 160–190 dorsal scales along the body (179.20±11.43, n=5); (23) 78–99 dorsal scales around the body (73.60±32.50, n=5); (24) 90 ventral scales (n=1) ( Figs. 21e–f , 25 ).

Material examined. MRSN R1911 (F. Andreone, 10 Feb. 1999, Nosy Be). ZSM 560–561 /2000 (M. Vences, 20 Mar. 2000, Sambava), ZSM 915 /2003 (F. Glaw, M. Puente & R. Randrianiaina, 20 Feb. 2004, Montagne des Français ), ZSM 285–287 /2004 (F. Glaw, M. Puente & R. Randrianiaina, 26 Feb. 2004, Ankarana), ZSM 308 /2004 (F. Glaw, M. Puente & R. Randrianiaina, 28 Feb. 2004, Montagne des Français) .

Hemipenial structure. Based on ZSM 560/2000, adult male from Sambava. Hemipenis subcylindrical, total length ca. 3.3 mm, with a relatively large pedicel measuring ca. 1.5 mm in length. The apex is divided in two lobes, each measuring ca. 1.8 mm. The truncus and the lobes are covered with fields and serrated ridges of pointed papillae which are more densely arranged on the lobes. Sulcus spermaticus formed by one channel on the pedicel, dividing into two channels of which each runs to a terminal position on either of the lobes. There are no distinct sulcal lips and no papillae around the sulcus ( Fig. 25 ).

Distribution. According to the specimens examined in this study, Lygodactylus heterurus is known from the following localities: (1) Nosy Be (type locality), (2) Ankarana, (3) Montagne des Français, (4) Sambava. Further localities mentioned in the literature are (5) Ampahana (type locality of subspecies trilineigularis), (6) 4 km S Antsiranana, and (7) Sahafary. Additional specimens clearly to be assigned to this species were recently collected by P. Bora, F. Glaw and J. Köhler at Tsingy de Bemaraha National Park in western Madagascar, in sympatry with L. klemmeri . These specimens, which were not examined morphologically in detail for the present study provide a considerable extension of the known distribution area southwards.

Habitat. According to Rösler (1998), this species lives in trees in dry forest in north-western Madagascar as well as in humid areas of the North East. At Montagne des Français we observed specimens on trees, at 0.5–2 m from the ground, active during the day (F. Glaw, M. Puente & R. Randrianiaina, personal observation in February 2004). At Sambava, specimens (probably to be assigned to subspecies trilineigularis) were observed on trees along the beach, active during the day (M. Vences, personal observation in 2001) .

Remark. According to Rösler (1998), the main difference between the subspecies trilineigularis and the nominate subspecies are the number of black lines on the throat (three indistinct vs. five distinct). Other differences are extended dark pigment in preanal region, fewer supralabials and infralabials, fewer longitudinal vertebral scales, and fewer scales between eyes, but these are based on the comparison of only two vs. three type specimens. We did not examine the type specimens of neither heterurus nor trilineigularis, and found some specimens from other localities difficult to assign to either of these forms: for example, specimens from Sambava, near the type locality of trilineigularis, agreed with that subspecies in showing rather indistinct gular stripes, although more than three were recognizable ( Fig. 3g), and a largely dark pigmented preanal region (see figure on p. 385 of Glaw & Vences 2007). As already emphasized by Rösler (1998) the distribution of the two subspecies and the relationships between populations of this species need further analysis. Nevertheless, it seems clear that all of these species are morphologically more similar to each other than they are to any other species of Lygodactylus , and the presence of some variability thus does not compromise the diagnosis of L. heterurus .

Lygodactylus klemmeri Pasteur, 1964

( Fig. 26 )

Lygodactylus klemmeri Pasteur, 1964 .— Name-bearing type: male holotype, MNHN 50.259 (currently missing), SVL 28 mm (according to the original description).— Type locality: “forêt de l’Antsingy (Nord- Ouest)”, according to the original description.— Other types: probably none; no information given in original description.— Etymology: dedicated to Konrad Klemmer, herpetologist and previous curator of the Senckenberg Museum, Frankfurt.

Lygodactylus praecox Pasteur, 1995 .— Name-bearing type: holotype MNHN 1990.48, hatchling, according to the original description.— Type locality: “Antsingy (Ouest central)”, according to the original description.— Other types: paratype MNHN 1990.49, hatchling.— Etymology: “ praecox fait allusion au fait que les nouveau-nés de cette espèce peuvent sortir de l’oeuf à un stade inhabituellement précoce, alors que la tête est encore (jusqu’aux épaules) aussi longue que le tronc entre bras et cuisses, et très sensiblement plus large” (according to the original description).

Diagnosis. Lygodactylus klemmeri is characterized as a member of the L. verticillatus group as defined above by its clearly whorled tail. It is distinguished from L. arnoulti , L. decaryi and L. verticillatus by having distinct diagonal black stripes on the throat (vs. no black lines on throat). L. blancae and L. heterurus also show black lines on throat, but these are only present in some specimens and then rather diffuse in L. blancae (vs. always present and very distinct in L. klemmeri ), and they are longitudinal, following the body axis, in L. heterurus (vs. slightly diagonal, directed towards the center of the throat in L. klemmeri ). Furthermore, in L. klemmeri , the ground colour of the throat is often at least slightly yellowish in life, whereas no ventral yellow colour is known from L. heterurus . Another species with black lines on the throat, here not assigned to the L. verticillatus group, is L. ornatus ; in this species, the throat lines are transversal, and L. ornatus further differs by its non-whorled tail. L. klemmeri further differs from L. decaryi by a lower number of preanal pores in males (9 vs. 15).

Description. Detailed data given in Puente et al. (2005a). (1) A relatively small species, adult SVL 23.9–25.2 mm (24.5± 0.5 mm, n=4); (2) TAL 22 mm (n=1); (3) granular dorsal scales; (4, 5) first finger present and bearing a claw; (6) three pairs of subdigital lamellae under the fourth toe (n=4); (7, 8) mental scale divided into three parts by sutures, its posterior projection can be in faint contact with first infralabial scale; (9, 10) three bisymmetrical postmental scales (n=4); (11) 5–6 postpostmental scales (5.50±0.95, n=4); (12) 5–7 infralabial scales (5.75±0.95, n=4); (13) 6–7 supralabial scales (6.75±0.50, n=4); (14) one internasal scale (n=4); (15) males with 9 preanal pores (n=1) and with non-pigmented enlarged preano-femoral scales; (16) tail with 11 distinct whorls; (17, 18) dorsolateral tubercles absent; (19, 20) the main dorsal colour is brownish and grey-olive, often with poorly defined broad light dorsolateral bands; the head can have a shade of greenish ( Fig. 21g ); (21) ventral side light, sometimes yellowish to yellow on throat, with 6–8 distinct black lines on throat that start on infralabials and run diagonal backwards and to the center of the throat where the anteriormost lines get into contact and can fuse. The posteriormost lines can run less distinctly diagonal (see Fig. 3).

Synonymy of Lygodactylus praecox . Pasteur (1995) described Lygodactylus praecox based on two recently hatched specimens from the Tsingy de Bemaraha (=Antsingy). The author was aware that sympatrically with this presumed new species, a further species occurred at Bemaraha, L. klemmeri , described by himself in 1964: " L. praecox est-il limité à l’Antsingy, comme semble l’être le sympatride L. klemmeri Pasteur 1964 ." However, he described the species as belonging to verticillatus group, related to L. heterurus , not to L. klemmeri ( Pasteur 1995) . The major difference to L. heterurus was the colouration of the throat: “4 paires de raies convergentes partent des labiales et aboutissent postérieurement, après fusion des deux plus antérieures, à 7 stries quasi parallèles dont les latérales parviennent jusqu’au cou” according to the original description ( Pasteur 1995). In fact, this description almost perfectly conforms to the throat pattern of L. klemmeri , as easily visible in the photo provided by Pasteur (1995) and the drawing herein. At the time of description of L. praecox , the holotype of L. klemmeri was probably already lost in the Paris museum (see Puente et al. 2005a), and Pasteur was therefore unable to carry out a direct comparison between the two species. In the original description of L. klemmeri , its throat pattern had been only briefly mentioned as short convergent stripes (see Puente et al. 2005a), and therefore the similarity may not have been immediately obvious. In fact, also the description of L. praecox is very brief, and provides a comparison only with L. heterurus (mainly emphasizing the differences in throat stripe pattern). Because of the brief original description and missing holotype of L. klemmeri , its identity cannot be assessed with full reliability but Puente et al. (2005a) provided a rationale why indeed the specimens with diagonal throat pattern are to be assigned to this species.

In the following, we provide a standardized description of L. praecox based on the holotype and the paratype specimens: (1) SVL 9.6–14.5 mm; (2) TAL not measured; (3) granular dorsal scales; (4, 5) first finger present and bearing a claw; (6) three pairs of lamellae under fourth toe (n=2); (7, 8) mental scale divided by sutures into three parts, its posterior projection not in distinct contact with first infralabial scale; (9, 10) three bisymmetrical postmental scales; (11) 5–6 postpostmental scales; (12) 5–6 infralabial scales; (13) 5–6 supralabial scales; (14) 1–2 internasal scales; (15) number of femoral pores unknown (specimens not adult); (16) tail without recognizable whorls; (17, 18) dorsolateral tubercles absent; (19) dorsally brownish; (20) without a distinct dorsal pattern, but tends to be striated; (21) ventrally light with four pairs of black stripes starting at the margins and running diagonally to the center of the throat where the anterior pair fuses; (22) 185–188 dorsal scales along the body; (23) 62–69 dorsal scales around the body ( Fig. 26 ).

Our re-analysis of the types of L. praecox revealed a single major character differentiating them from L. klemmeri , namely the lack of whorls on the tail. However, we assume that this is a juvenile character state, and based on (a) the throat patterning that is unique among Malagasy Lygodactylus and identical between L. klemmeri and L. praecox (and clearly differing from that of sympatric L. heterurus from the Bemaraha area), and (b) the fact that both species were described (and are only known) from the Bemaraha area, we consider Lygodactylus praecox Pasteur, 1995 as junior synonym of Lygodactylus klemmeri Pasteur, 1964 .

Material examined. MNHN 1990.48 – 49 (holotype and paratype of L. praecox , C. & F. Blanc, 15 Nov. 1964, Antsingy , Ouest central) . UADBA 17819–17820 (A. Raselimanana, 24 Nov. 2001, Bekopaka , Bemaraha National Park) , UADBA 17821 (A. Raselimanana, 30 Nov. 2001, Bekopaka , Bemaraha National Park) , UADBA 17822 (A. Raselimanana, 3 Dec. 2001, Bekopaka , Bemaraha National Park) .

Distribution. According to the specimens examined in this study, Lygodactylus klemmeri is known from the Tsingy de Bemaraha area, from various sites: (1) Antsingy forest (type locality), (2) Bekopaka. No further localities are mentioned in the literature. Two further specimens of L. klemmeri have recently been collected in the Tsingy de Bemaraha by P. Bora, F. Glaw and J. Köhler, in sympatry with L. heterurus . These specimens were not examined in detail for the present study, but one of them (a juvenile) provides further evidence that the typical throat pattern of the species is already visible in the juvenile stage ( Fig. 3h–i).

Habitat: According to Pasteur (1964) this species is arboreal. The specimens UADBA 17819, 17820 and 17822 were found active during the day on tree trunks in dry semi-deciduous forest, within a limestone area ( Puente et al. 2005a).

Lygodactylus verticillatus Mocquard, 1895

( Fig. 27 )

Lygodactylus verticillatus Mocquard, 1895 .— Name-bearing types: two syntypes (not examined by us), “deux spécimens d’une longueur totale de 40 mm ”, according to the original description. Syntype numbers are MNHN 1895.158–159 according to Krüger (2001).— Type locality: Madagascar according to the original description.— Other types: no data from the original description.— Etymology: named (by implication from the original description) after the whorled tail of the species.

Diagnosis. Lygodactylus verticillatus is a widespread species from the South of Madagascar, characterized as a member of the L. verticillatus group as defined above by its clearly whorled tail. It differs from other species of the group as follows: from L. heterurus and L. klemmeri by the absence (vs. presence) of distinct longitudinal black lines on the throat; from L. arnoulti and L. blanci by smaller body size (mean SVL 23.1 mm vs. 30.2 and 30.3 mm); from L. arnoulti further by a different dorsal colouration (no clearly defined pattern with a tendency to a longitudinal striation, vs. grey with indistinct ocellae or distinct black transversal lines or markings). Lygodactylus verticillatus is, by external characters, probably most similar to L. decaryi , but that species is characterized by having 15 preanal pores (vs. 9–11 in L. verticillatus ).

Description. (1, 2) A small Lygodactylus , adult SVL 15.2–34.2 mm (23.1± 2.5 mm, n=181); (2) TAL 12.2–31.3 mm (20.8 ± 3.4 mm, n=38); (3) granular dorsal scales; (4, 5) first finger present and bearing a claw; (6) three pairs of lamellae under fourth toe (n=189); (7, 8) mental scale divided in three parts by sutures, its posterior projection in broad contact to first infralabial scale (but less distinct than in L. tolampyae ); (9, 10) 2–3 bisymmetrical postmental scales (2.99±0.10, n=189); (11) 4–7 postpostmental scales, most frequently five (5.09±0.39, n=189); (12) 4–7 infralabial scales, most frequently five (5.06±0.50, n=189); (13) 5–7 supralabial scales (5.71±0.84, n=189); (14) 1–2 internasal scales, but two specimens (ZFMK 47287, MNHN 1990.666) with three (1.26±0.47, n=189); (15) males with 6–10, mostly nine preanal pores (8.86±0.68, n=98), usually without enlarged preano-femoral scales (if present, then always unpigmented); (16) tail with 8–16 strongly expressed whorls; (17, 18) usually without, but sometimes with 3–7 dorsolateral tubercles, each composed of 1–5 scales; (19, 20) dorsally brownish, without a clear pattern, often with a tendency of being striated, and mostly with a dark mark in the neck that includes a larger scale, often of light color and forming a tubercle (see also Mocquard 1895); (21) ventrally light coloured, usually with some throat spots; (22) 128–141 dorsal scales along the body (n=2); (23) 58–81 dorsal scales around the body (n=2); (24) 67 ventral scales (n=1) ( Figs. 21f , 27 ).

Material examined. BMNH 1930.7 .1.118 (topotype according to BMNH catalogue, E. d. White, hotel Dubound , Toliara) , MNHN 1990.422 (22 Jun. around Zampongotra, 30 km from Beloha ) , MNHN 1990.660 (15 Jan. 1966, Ampanihy ) , MNHN 1990.661 (28 Mar. 1968, Ampanihy ) , MNHN 1990.665 – 668 (15 Jan. 1966, Ampanihy ) , MNHN 1990.804 (C.A. Domergue, 17 Oct. 1962, La Sakoa ) , MNHN 1990.1738 – 1751 (C.A. Domergue, 2–4 May 1965, Ihotry ) , MNHN 1990.1752 – 1800 , 1901–1903 (C.A. Domergue, 20 Jun. 1965, Ihotry ) , MNHN 1990.1849 – 1852 (C.A. Domergue, 5–12 Feb. 1969, Manja hotel) , MNHN 1990.2103 – 2109 (C.A. Domergue, Jul. 1966, Ihotry ) , MNHN 1990.2257 (C.A. Domergue, 8–9 Apr. 1967, Ihotry ) , MNHN 1990.2349 – 2369 (C.A. Domergue, 1 Dec. 1967, Ihotry ) , MNHN 1990.2534 – 2539 (C.A. Domergue, 11 Sept. 1968, Ihotry ) , MNHN 1990.2864 – 2873 (C.A. Domergue, 18 Jun. 1966, Ihotry ) , MNHN 1990.3234 (C.A. Domergue, 15 Jan. 1966, Ampanihy ) , MNHN 1990.3282 ( Lamboromakandro ) , MNHN 1990.3283 (C.P. Blanc, 13 Aug. 1964, Manombo ) , MNHN 1990.3284 (14 Jul. 1960, Morombe district ) , MNHN 1990.3285 (Apr. 1961, Morombe district ) , MNHN 1990.3286 (Aug. 1962, Ankagoabo-Sud ) , MNHN 1990.3295 – 3301 (C.A. Domergue, 11–12 Dec. 1972, Ihotry ) , MNHN 1990.3314 – 3321 , 3287–3294 (C.A. Domergue, 15 Jan. 1966, Ampanihy ) , ZFMK 21813 (H. Meier, Jan. 1978, Toliara) , ZFMK 40512–40514 (C.A. Domergue, 4 May 1968, Ihotry ) , ZMA 19596 (D. R. Vieites & co-workers, 2003, Ifaty ) , ZMA 19601, (D. R. Vieites & co-workers, 2003, Ifaty ) , ZSM 1145 /2003 (D. R. Vieites et al., 5–6 Feb. 2003, Ifaty ) .

Distribution. According to the specimens examined in this study, Lygodactylus verticillatus is known from the following localities: (1) Ampanihy, (2) Ankagoabo-Sud, (3) Ifaty, (4) Ihotry, (5) Lamboromakandro, (6) Tsimanampetsotsa, (7) Toliara, (8) Zampongotra. Several of these localities (Lamboromakandro and Ankagoabo-Sud) could not be located on the map. They probably are wrong transcriptions in the MNHN catalogue of the locality names; according to a personal communication of C. P. Blanc in 2009, Lamboromakandro could refer to a small village in dry forest, about 30 km north-east of Sakaraha, or be an alteration of Lambomakandra which refers to a forest near Sakaraha (22°42’S / 44°42’ E), crossed by the road Ihosy-Tuliara ( Carleton & Schmidt, 1990). Ankagoabo-Sud is probably is a wrong transcription of Ankazoabo-Sud, a large village in the north of Sakaraha. One further locality, Monombo, is vouchered by a specimen from the MNHN collection. The best known locality with this name is Manombo Special Reserve on the southern east coast, but we strongly doubt that L. verticillatus occurs there. Two further localities (La Sakoa and Manja hotel) most probably refer to sites close to or in Toliara or another major destination in the South of Madagascar. Further localities mentioned in the literature are Andrahomana, Amboasary, and the shores of the Onilahy river ( Angel 1942). The species is also known from Europa island in the Mozambique channel.

Habitat. According to Pasteur (1977), this species occurs in dry forest, on houses, fences and trunks. We observed specimens of L. verticillatus on isolated trees and shrubs in semi-arid areas, found on trees and fences in towns (M. Puente, M. Thomas & D. R. Vieites, personal observation in February 2003). The species reproduces over the whole year with a weak seasonality, peaking in the rainy season; for more information, see Vences et al. (2004). Pasteur (1977) and Pasteur & Lumaret (1976), based on large series of specimens collected from Ihotry, describe yearly fluctuations in the population density of L. verticillatus relative to L. tuberosus , and changes in scale count values in this population that occurred in a short time span of a few years only.

Figures

FIGURE 21. Photographs of living specimens assigned to species in the Lygodactylus verticillatus group. (a) and (b) L. arnoulti, specimens from Mt. Ibity; (c) and (d), L. blancae from Lac Itasy; (e) L. heterurus, Ankarana; (f) L. heterurus trilineigularis, Sambava; (g) L. klemmeri, Tsingy de Bemaraha; (h) L. verticillatus, south-western Madagascar. All photographs by F. Glaw & M. Vences.

FIGURE 22. Scalation characters of Lygodactylus arnoulti. (a) head in lateral view (scale bar = 2 mm); (b) lamellae of fourth toe in ventral view (scale bar = 2 mm); (c) head in dorsal view (scale bar = 2 mm); (d) throat in ventral view (scale bar = 2 mm); (e) preanal pores in ventral view (scale bar = 2 mm); all based on male specimen MNHN 1990.3420. All drawings by M. Puente.

FIGURE 23. Scalation and hemipenial characters of Lygodactylus blancae. (a) head in lateral view (scale bar = 2 mm); (b) head in dorsal view (scale bar = 2 mm); (c) throat in ventral view (scale bar = 2 mm); (d) preanal pores in ventral view (scale bar = 2 mm); (e) lamellae of fourth toe in ventral view (scale bar = 1 mm); all based on male specimen MNHN 1990.38; (f) and (g), sulcal and asulcal view of the hemipenis, based on male specimen ZSM 498/2001 (scale bar = 1mm). All drawings by M. Puente.

FIGURE 24. Scalation characters of Lygodactylus decaryi. (a) head in lateral view (scale bar = 2 mm); (b) lamellae of fourth toe in ventral view (scale bar = 2 mm); (c) head in dorsal view (scale bar = 2 mm); (d) throat in ventral view (scale bar = 2 mm); (e) preanal pores in ventral view (scale bar = 2 mm). Drawings based on holotype male specimen MNHN 1930.271 except (b) that is based on female specimen MNHN 1950.262. All drawings by M. Puente.

FIGURE 25. Scalation and hemipenial characters of Lygodactylus heterurus trilineigularis. (a) head in lateral view (scale bar = 2 mm); (b) head in dorsal view (scale bar = 2 mm); (c) throat in ventral view (scale bar= 2 mm); (d) preanal pores in ventral view (scale bar = 1 mm); (e) lamellae of fourth toe in ventral view (scale bar = 2 mm); (f) and (g), sulcal and asulcal view of the hemipenis (scale bar = 1 mm). All based on male specimen ZSM 560/2000. All drawings by M. Puente.

FIGURE 26. Scalation characters of Lygodactylus klemmeri, based on the juvenile type specimens of L. praecox that we here consider as a junior synonym of klemmeri. (a) head in lateral view (scale bar = 1 mm); (b) lamellae of fourth toe in ventral view (scale bar = 0.1 mm); (c) head in dorsal view (scale bar = 2 mm); (d) throat in ventral view (scale bar = 3 mm). Drawings (a) and (d) based on MNHN 1990.48; (b) and (c) on MNHN 1990.49. All drawings by M. Puente.

FIGURE 27. Scalation characters of Lygodactylus verticillatus. (a) head in lateral view (scale bar = 1 mm); (b) lamellae of fourth toe in ventral view (scale bar = 1 mm); (c) throat in ventral view (scale bar = 1 mm); (d) head in dorsal view (scale bar = 1 mm); all based on female specimen MNHN 1982.1246; (e) preanal pores in ventral view (scale bar = 2 mm), based on male specimen MNHN 1982.1245. All drawings by M. Puente.