Montane Tadpoles in Madagascar: Molecular Identification And
Total Page:16
File Type:pdf, Size:1020Kb
Copeia, 2005(1), pp. 174–183 Montane Tadpoles in Madagascar: Molecular Identification and Description of the Larval Stages of Mantidactylus elegans, Mantidactylus madecassus, and Boophis laurenti from the Andringitra Massif MEIKE THOMAS,LILIANE RAHARIVOLOLONIAINA,FRANK GLAW,MIGUEL VENCES, AND DAVID R. VIEITES The larval stages of three species of frogs from montane habitats in the Andrin- gitra Massif, southern central Madagascar at 2100–2500 m above sea level, were identified through mitochondrial DNA sequences and are described herein. The tadpoles of Boophis laurenti agree with the previously known tadpoles of the closely related Boophis microtympanum, whereas the tadpoles of Mantidactylus madecassus are similar to those of other species of the subgenus Brygoomantis that occur at lower altitudes. The tadpoles of Mantidactylus elegans are very large (up to 106 mm total length) and show mouthparts largely agreeing with those of species of the subgenus Guibemantis, with a relatively high number of upper labial tooth rows (one contin- uous and six interrupted). These tadpoles are uniformly blackish on the dorsum, indicating a possible general trend of high frequency of dark color and melanism in montane amphibians. Molecular identification provides a fast and very efficient tool to identify larval stages of amphibians, especially in cases of specialized tad- poles from remote areas in which rearing is difficult. HE central high plateau of Madagascar en- the specific or subspecific level. The status of T compasses several mountain massifs which Boophis laurenti, which is morphologically close locally reach altitudes higher than 2800 m to Boophis microtympanum, remains to be clari- above sea level. The three highest massifs (Tsar- fied. atanana, Ankaratra, and Andringitra) harbor a We recently started a project to describe the highly specialized and endemic montane fauna, larval stages of Malagasy frogs, based on their which is poorly known in terms of ecology and identification through mitochondrial DNA se- biology (Raxworthy and Nussbaum, 1996b). Vir- quences (e.g., Hebert et al., 2003; Blaxter and tually nothing is known about the amphibians Floyd, 2003; Tautz et al., 2003). This method is of Tsaratanana in northern Madagascar (Blom- considerably faster and potentially more reli- mers-Schlo¨sser and Blanc, 1991; Raxworthy and able than identification through rearing of tad- Nussbaum, 1996b), whereas recent inventories poles and determination of juveniles. of the Ankaratra mountains in central Madagas- Knowledge on the morphology and habitat of car have resulted in a reasonable state of knowl- tadpoles is highly relevant for the understand- edge (Glaw and Vences, 1994; Vences and Glaw, ing of the ecological requirements and natural 1999; Vences et al., 2002a). The third high-alti- history of frog species. Recent global trends of tude massif, Andringitra, has been intensively amphibian declines (Kiesecker et al., 2001) surveyed (Raxworthy and Nussbaum, 1996a; Ra- seem to especially affect montane species (e.g., selimanana, 1999; Rasolonandrasana and Good- Young et al., 2001). Hence, the understanding man, 2000), but detailed ecological data of the of the natural history of montane Malagasy am- montane herpetofauna specialized to altitudes phibians bears relevance for conservation biol- above 2000 m is so far lacking. ogy, because these species have a very limited According to present knowledge, three am- distribution, and several may qualify for a phibian species (Anodonthyla montana, Boophis threatened category in terms of IUCN catego- laurenti, and Mantidactylus madecassus) and one ries (IUCN, 2001). reptile (Lygodactylus intermedius) are endemic to In this paper, we provide the first descriptions high-elevations of Andringitra massif (Vences of the tadpoles of three frogs from the Andrin- and Glaw, 1999; Vences et al., 2002a). The local gitra Massif belonging to the family Mantellidae: populations of several further amphibians (e.g., B. laurenti, Mantidactylus elegans, and Mantidac- Scaphiophryne madagascariensis, Mantidactylus tylus madecassus. We discuss our findings in the brevipalmatus, and Mantidactylus curtus) appear context of the evolution of larval morphology to be either genetically or morphologically dis- among mantellids and review the utility of ge- tinct and may also represent separate entities at netic identification of larval stages. ᭧ 2005 by the American Society of Ichthyologists and Herpetologists THOMAS ET AL.—MONTANE TADPOLES FROM MADAGASCAR 175 MATERIALS AND METHODS solved on an ABI 3730 capillary sequencer. DNA sequences are deposited in GenBank (accession Tadpoles were collected with dip nets, eutha- numbers: M. elegans AY659959; tadpole of M. ele- nized using chlorobutanol, and assigned to gans AY659960; M. madecassus AY659961; tad- morphotype categories in the field. Subsequent- pole of M. madecassus AY659962; B. laurenti ly a piece of tail or fin was taken as a tissue AY659963; tadpole of B. laurenti AY659964). sample for DNA extraction from one specimen To detect possible mistakes in assigning tad- of each tadpole series, and all tadpoles were poles to series in the field, we compared the preserved in 4% buffered formalin. Institution- DNA voucher once again with all other speci- al abbreviations follow Leviton et al. (1985). mens of the series using a stereo microscope. Fieldwork was carried out at the beginning of Drawings and descriptions in this paper are February 2003 in the Andringitra National Park. based on the DNA vouchers, whereas other rep- The first locality, the sampling site of the tad- resentative specimens of the same series were pole series ZSM 608/2003, ZSM 609/2003, and examined to supplement structures missing be- ZSM 611/2003, was at a slowly running stream cause of tissue sampling. To assess morpholog- at an altitude of 2488 m above sea level located ical variability, measurements were taken from in a depression named ‘‘Cuvette Boby’’ all specimens of each series. All tadpoles were (22Њ11Ј41Љ S/46Њ53Ј23ЉE). The second site was staged according to Gosner (1960). Terminol- at a stream on the Andohariana plateau ogy and measurements follow Altig and Mc- (22Њ10Ј49ЉS/46Њ54Ј01ЉE; 2114 m above sea lev- Diarmid (1999) with some modifications. The el), sampling site of the tadpole series ZSM following measurements were taken to the near- 610/2003. The water of both sites was cold and est 0.1 mm with dial calipers: body length (dis- clear; the streams widths were 1.5–2.0 m; and tance from the tip of the snout to the body ter- average stream depth was 0.5 m. At some places, minus, which is the junction of the posterior the depth dropped down to 2 m. The bottoms body wall with the tail axis); tail length (the dis- of the brooks were covered by gravel; aquatic tance from the body terminus to the absolute plants and algae were rare. tip of the tail); total length (the sum of body For tadpole identification, we amplified a length and tail length); body width (measured fragment of up to 550 bp of the mitochondrial at the widest point of the ‘‘head’’ right behind 16S rRNA gene of each sample. We compared the eyes, not in the intestinal part); maximum these sequences of tadpoles with homologous eye diameter; interorbital distance (measured sequences of morphologically and bioacustically between the centers of the pupils); internarial well-identified adult specimens from the same distance (measured between the centers of the population. According to a large unpublished nares); distance between tip of snout and naris dataset of this gene fragment available to us, in- (up to the centre of the naris); distance be- cluding sequences retrievable from Genbank as tween naris and eye (from the center of naris of May 2004, pairwise sequence divergences are to the anterior edge of the eye); distance be- high among species of mantellid frogs (2–15%), tween tip of snout and spiraculum (up to the especially if these belong to distinct genera and center of the spiracular aperture); tail muscle species group as in the case study reported height (first, measured vertically from the junc- here. However, there are very low sequence di- tion of the body wall with the ventral margin of vergences among conspecific individuals be- the tail muscle; second, measured at midtail); longing to the same population; usually these tail height (including fins and caudal muscula- sequences are 100% identical. Hence, identifi- ture, taken at its maximal vertical extent); and cation of larvae is unequivocal if full sequence dorsal-fin origin (defined relatively to the tail identity between larvae and adults is encoun- body junction). The formula of labial tooth tered. rows follows Dubois (1995). The mouthparts in- DNA was amplified by PCR in 50 l volumes, clude upper tooth rows (UTR) and lower tooth applying 34 cycles of 30 sec at 95 C, 60 sec at rows (LTR). Values given throughout are num- 55 C, 60 min at 72 C with 20–100 ng DNA-tem- ber of individuals (N) and means of the mor- Ϯ plate, 1,5 l of 50mM MgCl2,5 l dNTPs, and phometric measurements standard deviations 1 U of Taq-Polymerase, using 5 pmol of primers (SD) as well as maximal and minimal values. 16SA-L and 16SB-H (Vences et al., 2000) as orig- inally described by S. R. Palumbi, A. Martin, S. RESULTS Romano, W. O. McMillan, L. Stice, and G. Gra- bowski (The Simple Fool’s Guide to PCR, Vers. In our comparisons of sequences of tadpoles 2.0, 1991, unpubl.). Annealing temperature was with those of adult specimens, we found three at 55 C. Double-stranded sequences were re- series of larvae showing