Lygodactylus fritzi, Vences & Multzsch & Gippner & Miralles & Crottini & Gehring & Rakotoarison & Ratsoavina & Glaw & Scherz, 2022

Vences, Miguel, Multzsch, Malte, Gippner, Sven, Miralles, Aurélien, Crottini, Angelica, Gehring, Philip-Sebastian, Rakotoarison, Andolalao, Ratsoavina, Fanomezana M., Glaw, Frank & Scherz, Mark D., 2022, Integrative revision of the Lygodactylus madagascariensis group reveals an unexpected diversity of little brown geckos in Madagascar’s rainforest, Zootaxa 5179 (1), pp. 1-61 : 46-48

Lygodactylus fritzi sp. nov.

Lygodactylus sp. 11: Gippner et al. (2021).

Justification. This new species from coastal areas in Madagascar’s Northern Central East has previously been called L. sp. 11 in Gippner et al. (2021). It corresponds to a lineage forming part of subclade A5 of Domerguella , and occurs in a general area where L. miops is also present. It is characterized by the presence of distinct lateral spinelike scales at the base of the tail in males, as are found in several representatives of subclade A5 but not in other Domerguella . The smallest genetic distance is 9.0% to specimens assigned to L. miops . However, it differs from this species by phylogenetic position and lack of haplotype sharing in nuclear-encoded genes. While species status of these two lineages is validated by the molecular evidence, the assignment of the holotype of L. miops , and of the types of the other earlier names septemtuberculatus and spinulifer , to either of them requires further justification (see also the account of L. miops ).

The following arguments support our decision to assign the holotype of miops , septemtuberculatus and spinulifer to what we will in this section call the “widespread lowland lineage” rather than to L. sp. 11:

(1) The specimens of the widespread lowland lineage (including the three historical types) differ from L. sp. 11 by a 3–6 vs. 2 dorsolateral tubercles between limbs; dorsal scale count 207–242 vs. 247; ventral scale count 102–113 vs. 98; lateral tubercles at tail base recognizable (also in females) vs. barely recognizable. While each of these meristic differences by itself is rather weak, taken together they characterize L. sp. 11 as morphologically distinct, and the three type specimens of the earlier names as better fitting the widespread eastern lineage.

(2) L. sp. 11 is only known from two low-elevation localities near sea level (0–20 m a.s.l.), and seems to be relatively localized; for instance, it has not been found at other nearby localities at slightly higher elevation such as Sahafina or Betampona, despite a substantial number of specimens sequenced from Betampona, which all belonged to the widespread lowland lineage. It is therefore less likely that historical specimens of L. sp. 11 were collected historically, and that its range extends into the type localities of the three historical nomina, especially up to Moramanga (the type locality of septemtuberculatus and spinulifer ).

(3) The type locality of L. miops , “Senbendrana” according to the original description ( Günther 1891), cannot be located reliably at present. Senbendrana has been reported with the addition “near Tamatave” (= Toamasina) as a collecting locality of spiders ( Pocock 1895); or as corresponding to Sahembendrana or Sahambendrana. This latter synonymy is supported by the fact that the type of L. miops was provided by “Majastre” (see Puente et al. 2009), probably corresponding to A. Majastre, a collector who provided specimens of many animals and plants from this area. Some of Majastre’s collections are labelled “Sahambendrana”, e.g., the type of the orchid Eulophia grandibracteata (see Schultz 2013). The locality apparently was often misspelled; for example, we assume that “Sen Bendrana” ( Michaelsen 1891), “Senbendra” (Sharp & Ogilvie-Grant 1898), or “Schambendrama” (Bott 1963) refer to the same site as well. Blommers-Schlösser & Blanc (1991) located Sahembendrana close to Akkoraka (at higher elevations in eastern Madagascar), but it is likely that Majastre’s collecting site was situated closer to Toamasina. We could not locate current or historical maps mentioning a site with this or a similar name, but Ramananjara (2009) documents the sale of a property in 1931, in the “Canton d’Antetezambaro; sur la rivière Sahambendrana”, and more specifically “au sud d’Ambodisatrana”, which likely refers to a coastal village about 30 km north of Toamasina that can be located in historical maps from 1934 (Service Géographique de Madagascar, map “Fénérive”, 1/500,000). On the other hand, Rosa et al. (2012) report a campsite from Betampona Reserve (about 35 km north-east of Toamasina) locally known as Sahambendrana, at coordinates S17.8984, E49.2154, 458 m a.s.l. Other sources refer to a Sahambendrana river on the northern versant of Betampona ( Randriatavy 2003; Randrianarimanana 2009). Whether any of these sites corresponds to Majastre’s collecting locality cannot be decided without further evidence, but these references demonstrate that the toponym has been and is in use for sites to the north and northeast in the vicinity of Toamasina.

The available evidence, however, points to the original collecting site being not directly at sea level. According to Günther (1891), the same collection that included the L. miops holotype also contained “ Rhacophorus luteus ”, which almost certainly corresponds to a treefrog species of the Boophis luteus group, which is not known from coastal sites in Madagascar (but known to be present in Betampona), and for instance has not been collected at Vohibola or Ankanin’ny Nofy ( Gehring et al. 2010) where L. sp. 11 occurs. Furthermore, Pellegrin (1933) reported fish specimens of the genus Sicyopterus from a “rivière Sahembendrana (région de Tamatave)”. These specimens were identified by Sparks & Nelson (2004) as S. franouxi , a species that according to these authors inhabits clear, swift-flowing waters and is frequently captured quite far inland, again in agreement that this site is within the range of the widespread eastern lineage but not a coastal locality within the range of L. sp. 11. The fact that Sahambendrana / Sahembendrana has on various occasions been used to refer to a river (e.g., Pellegrin 1993; Randriatavy 2003; Randrianarimanana 2009; Ramananjara 2009) allows for the possibility of an upstream collecting site of the L. miops holotype, at some distance from the coast, and thus at a moderate elevation as indicated by the accompanying fish and amphibian fauna; and probably in the area close to Betampona where our collections have only yielded individuals of the widespread lowland Domerguella lineage.

In summary, the available evidence thus suggests that none of the earlier available names miops , septemtuberculatus , or spinulifer is likely to apply to L. sp. 11, which we therefore formally name as species new to science, L. fritzi sp. nov.

Holotype. ZSM 651 /2009 ( ZCMV 8902 ), female, collected by P.-S. Gehring, F. Ratsoavina, and E. Rajeriarison at Ankanin’ny Nofy, east coast of Madagascar, geographical coordinates - S18.6058, E49.2138, roughly at sea level, on 8 April 2009.

Diagnosis. Lygodactylus fritzi sp. nov. is a species of the Lygodactylus subgenus Domerguella based on molecular phylogenetic relationships, and it can also be assigned to the subgenus Domerguella by an undivided mental scale with two postmentals, absence of a claw on the first finger, and 7 preanal pores in males. Within Domerguella , the new species is one of several species of subclade A5 known from the Northern Central East of Madagascar.

It differs from the nominal species of Domerguella occurring in northern Madagascar and belonging to subclades A1–A4 as follows: from L. expectatus by non-enlarged dorsolateral scales (longitudinal count of dorsal scales>250 vs. <170); from L. rarus by lack of regular crossbands on tail (vs. presence) and different body shape with less elongated limbs (relative hindlimb length 0.47 vs.>0.55); from L. petteri , L. tantsaha , L. salvi , L. roellae , and L. hapei by a higher longitudinal count of dorsal scales (253 vs. <241); and from L. madagascariensis as well as most of the previously mentioned species of clades A1‒A4 by a rudimentary spine-like tubercle at the base of the tail in the only known female (vs. absence).

From other species of subclade A5, the new species differs as follows: from L. miops by fewer dorsolateral tubercles between limbs (2 vs. 3–6), higher longitudinal dorsal scale count (247 vs. 207–242), lower longitudinal ventral scale count (98 vs. 102–113), and weakly expressed lateral tubercles at tail base vs. clearly recognizable in males and females; from L. guibei by higher longitudinal dorsal scale count (247 vs. 170–220), and weakly expressed lateral tubercles at tail base vs. clearly recognizable, usually large in males and females; from L. winki by fewer dorsolateral tubercles between limbs (2 vs. 5–8), higher longitudinal dorsal scale count (247 vs. 187–222), and weakly expressed lateral tubercles at tail base vs. clearly recognizable usually large in males and females; from L. ulli possibly by more weakly expressed lateral tubercles at tail base and a lower longitudinal count of ventral scales (98 vs. 110). From all these species, it differs by phylogenetic position, at least 9% 16S distance, and absence of haplotype sharing in both nuclear-encoded genes studied.

For a distinction from one other species newly named and described in the following, see the respective diagnosis below.

Etymology. We are pleased to dedicate this species to Uwe Fritz, director of the Museum of Zoology, Dresden (part of the Senckenberg Natural History Collections), in recognition of his substantial contributions to the taxonomy of chelonians and squamates, and his tireless efforts to spearhead the fight for continued funding of basic taxonomic research. The name is a patronym (i.e., a noun in the genitive case).

Description of the holotype. Female in a good state of preservation, tail partly detached. SVL 26.4 mm, TAL 26.5 mm; for other measurements see Table 1 . Body broader than head. The distance from the tip of the snout to the anterior border of the eye (3.1 mm), is less than the interorbital distance anteriorly (3.5 mm), and greater than the distance between the eye and ear opening. Snout covered with granular scales larger than those on the rest of the dorsum. Nostril surrounded by four scales: rostral, first supralabial, and two supranasals. Mental scale undivided; no contact between posterior projection of mental scale and first infralabial; two asymmetrical postmental scales with four postpostmental scales; seven infralabial scales; eight supralabial scales; two internasal scales; granular dorsal scales; dorsum with small, homogeneous, granular, and unkeeled scales of similar size to those on trunk, no distinct size difference to scales on limbs; 247 dorsal scales longitudinally along the body; 98 ventral scales between mental and cloaca; venter with large homogeneous smooth scales; first finger present but very small, not bearing a claw; three pairs of subdigital lamellae on the fourth toe; two not very distinct dorsolateral tubercles, each consisting of one scale; tail without whorls; small lateral spines at the base of the tail.

Coloration of the holotype is only described from the specimen that was preserved in ethanol for 12 years. The dorsum is fawn to brownish with scattered darker brown spots, the flanks are darker than the dorsum. A distinctive dark stripe is displayed on the shoulder. The parietal exhibits two darker areas. The tail is fawn with equispaced darker brown stripes on it. Venter and the snout are uniformly whitish with small irregular brown spots.

Variation. Morphometric and meristic data are only known from a single voucher specimen. However, color patterns could be assessed from photographs of three additional individuals ( Fig. 22 ), some of which were dorsally gray-beige with an irregular contrasted pattern of larger light and smaller dark spots ( Fig. 22A ), others more uniform with a more or less symmetrical pattern of dark spots ( Fig. 22E ), or dark brown with weakly contrasted light brown dorsolateral bands with a somewhat reddish tone ( Fig. 22B–C ). Ventrally irregularly dark spotted ( Fig. 22D ).

Distribution. L. fritzi is only known from (1) the type locality Ankanin’ny Nofy and (2) Vohibola, two coastal lowland localities (0–20 m a.s.l.) in the Northern Central East of Madagascar. Littoral forest harbours a high species richness especially of plants, with several genera endemic to this habitat ( de Gouvenain & Silander 2003; Bollen & Donati 2005). Fisher & Girman (2000) identified littoral forests as one of four major areas for ant endemism in Madagascar. Previous studies in south-eastern littoral forests found no vertebrate species strictly endemic to that forest type ( Ganzhorn et al. 2000; Goodman & Ramanamanjato 2007) but recent taxonomic revisions have revealed numerous amphibians and reptiles restricted to small areas of forest directly adjacent to the Madagascar’s coast, such as for example species of the miniaturized frog genus Mini ( Scherz et al. 2019) or Pandanus-dwelling frogs of Guibemantis ( Lehtinen et al. 2011) , or the chameleon Calumma vohibola ( Gehring et al. 2011) , Lygodactylus fritzi adds to this growing list of species specialized to these highly threatened coastal forests.

Natural history. One adult specimen was photographed millimeters from the posterior end of a bug ( Fig. 22 ), and was presumably consuming honeydew excreted by the insect, as is known from gecko species ( Fölling et al. 2001).

Figures

FIGURE 22. Photos of Lygodactylus fritzi sp. nov. specimens from Vohibola in life. All specimens correspond to genotyped individuals included in Fig. 1 that were not collected.