Zootaxa,Larval Morphology in Four Species of Madagascan Frogs Of

TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited.

Larval morphology in four species of Madagascan frogs of the subgenus Brygoomantis (Mantellidae: Mantidactylus)

ANGELIKA KNOLL1, JÖRN KÖHLER2,6, FRANK GLAW3, MEIKE TESCHKE4 & MIGUEL VENCES5

1Technical University of Darmstadt, FB Biologie, Schnittspahnstr. 10, 64287 Darmstadt, Germany. E-mail: [email protected] 2Department of Natural History – Zoology, Hessisches Landesmuseum Darmstadt, Friedensplatz 1, 64283 Darmstadt, Germany. E-mail: [email protected] 3Zoologische Staatssammlung, Münchhausenstr. 21, 81247 München, Germany. E-mail: [email protected] 4Institute for Genetics, Evolutionary Genetics, University of Cologne, Zuelpicher Str. 47, 50674 Köln, Germany. E-mail: [email protected] 5Technical University of Braunschweig, Spielmannstr. 8, 38106 Braunschweig, Germany. E-mail: [email protected] 6Corresponding author

Abstract

We describe the tadpole morphology of four species of frogs classified in the endemic Madagascan subgenus Bry- goomantis of the genus Mantidactylus, based on larval specimens identified by DNA barcoding: Mantidactylus betsileanus and M. biporus, and two so far undescribed species that are here named M. sp. aff. betsileanus "very slow calls" and "Vohidrazana" referring to their bioacoustic features or collecting locality. The tadpoles of these four species are brown to yellowish benthic forms with a depressed body shape, dorsally directed eyes and relatively low fins. The oral discs are generalised and all species exhibit a wide gap in dorsal papillae. There is variation among different developmental stages with regard to the keratodont row formulae. In Mantidactylus betsileanus, younger stages (25 and 26) exhibit the formula 1:3+3/1+1:2, whereas it is 1:4+4/1+1:2 in older stages (36 and 38). In M. sp. aff. betsileanus "very slow calls", keratodont row formulae vary within one developmental stage (25–26) from 1:3+3/3 to 1:4+4/3, whereas it seems to be con- stantly 1:4+4/3 in all examined developmental stages of M. sp. aff. betsileanus "Vohidrazana". In general, the larvae of the different Brygoomantis species studied are morphologically similar to each other. Differences between species mainly concern colouration, shape of body and the number of keratodonts per millimetre, which is approximately 38 in M. betsileanus, 50 in M. biporus, and 60 in the two undescribed species.

Key words: Amphibia, Mantellidae, Mantidactylus, Brygoomantis, Mantidactylus betsileanus, M. biporus, tadpole descriptions, DNA barcoding

Introduction

The anuran family Mantellidae comprises a large number of species endemic to Madagascar and the Como- ros. Among them is the subfamily Mantellinae which was considered to contain two genera: Mantella Bou- lenger, 1882 and Mantidactylus Boulenger, 1895 (Vences & Glaw 2001). Recently, a molecular phylogenetic analysis revealed the paraphyly of the genus Mantidactylus which consequently was partitioned into seven monophyletic genera (Glaw & Vences 2006), plus a further genus described subsequently (Glaw et al. 2006). According to this new proposal for classification, the genus Mantidactylus sensu stricto contains six clades of subgeneric rank: Brygoomantis Dubois, 1992, Hylobatrachus Laurent, 1943, Chonomantis Glaw & Vences, 1994, Maitsomantis Glaw & Vences, 2006, Ochthomantis Glaw & Vences, 1994 and Mantidactylus. Species placed in the subgenus Brygoomantis by Dubois (1992) were formerly known as the Mantidacty-

Accepted by S. Carranza: 24 Aug. 2007; published: 17 Oct. 2007 49 TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited. lus ulcerosus species group (e.g. Blommers-Schlösser 1979). Species of this lineage are small to medium- sized (adult snout-vent length 17–68 mm) frogs, showing riparian to semiaquatic habits along stagnant or running water bodies, with diurnal and nocturnal activity (Glaw & Vences 1994, 2006). As far as known, egg deposition is terrestrial in a single clutch (e.g., Blommers-Schlösser 1979). The known larvae of Brygooman- tis represent a more or less generalised type of benthic feeders with small mouthparts exhibiting a wide dorsal gap in their marginal papillae (Blommers-Schlösser 1979). A previously underestimated problem connected with the systematics of Brygoomantis is the high diversity of unidentified cryptic species. Currently, the subgenus is considered to contain eleven species (Glaw & Vences 2006). However, at least some nominal species as defined by Blommers-Schlösser & Blanc (1991) display distribution patterns barely linked to the identified biogeographic regions, and differences in their advertisement calls as well as strong genetic divergence strongly suggest the presence of species complexes instead of one single taxon, although differentiation according to the morphology of adults is often highly problematic (Vences et al. 2006). A recent analysis by A. Aumüller, F. Glaw and M. Vences (in progress), based on morphological re-analyses of types, mitochondrial DNA sequences and analysis of advertisement calls indicates that the species M. ulcerosus, M. betsileanus and M. biporus as defined by Blommers-Schlösser & Blanc (1991) in fact need to be partitioned into at least 16 separate species. As a consequence, former tadpole descriptions (Arnoult & Razarihelisoa 1967; Blommers-Schlösser 1979) were possibly based on more than one species and the assignment of taxon names to described populations was not reliable from the present point of view. The goal of this contribution is thus to provide detailed descriptions of the larval morphology of four species of Mantidactylus of the subgenus Brygoomantis which were unequivocally identified according to a DNA barcoding approach (see Thomas et al. 2005). This paper is part of a series of ongoing tadpole descriptions aimed at documenting the morphological diversity of Mada- gascan anuran larvae (e.g., Raharivoloniana et al. 2003; Grosjean et al. 2006; Vejarano et al. 2006; Randrian- iaina et al. 2007).

Material and methods

Tadpoles were collected in the field, euthanised by immersion in chlorobutanol solution, and immediately sorted into series based on their morphology. From each series at least one specimen was selected and a tissue sample from its tail musculature or fin taken and preserved in 99% ethanol. These specimens are here named “DNA vouchers”. After tissue collection, all specimens were preserved in 4% formalin. Specimens were deposited in the Zoologische Staatssammlung München, Germany (ZSM). Tadpoles were identified using a DNA barcoding approach based on a fragment of the mitochondrial 16S rRNA gene, which is known to be sufficiently variable among species of Malagasy frogs (Thomas et al. 2005). The ca. 550 bp fragment was amplified using primers 16Sa-L and 16Sb-H from Palumbi et al. (1991) applying standard protocols, resolved on automated sequencers, and compared to a near-complete database of sequences of adult Malagasy frog species. Identification was considered to be unequivocal when the tadpole sequence was 99–100% identical to an adult specimen from the same geographical region, and not more similar to any sequence from another species. DNA sequences were deposited in Genbank (accession numbers EF606875- EF606886; accession numbers of comparative adult specimens included in the sequence sets AY847959-AY848683 and AJ315909-AJ315913). As outlined in the introduction, taxonomy of Brygoomantis is currently under revision, and many new species are awaiting description as well as synonyms awaiting resurrection. Providing stable names for the taxa identified is therefore not possible in all cases. The most recent comprehensive account on Malagasy frogs is the audioguide of Vences et al. (2006). We here partially follow the names introduced in this publica- tion to refer to forms that have not yet been assigned Linnean nomenclatural names, and we explain the iden-

50 · Zootaxa 1616 © 2007 Magnolia Press KNOLL ET AL. TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited. tity of each form in a small taxonomic note in the descriptive accounts. Because the DNA sequences of each tadpole series (as available from Genbank) provides an unambiguous tag, it will be possible to trace the iden- tity of these tadpoles, also if in the future the taxonomy of Brygoomantis changes. Terminology in morphological descriptions and landmarks for measurements follow Altig & McDiarmid (1999) and Grosjean (2006). Developmental stages were determined according to Gosner (1960). Keratodont row formulae follow the scheme of Dubois (1995). Body length and tail length were measured using a digital caliper (effective range: 0–150 mm; precision: 0.03 mm; repeatability: 0.1 mm). Other measurements were taken using a stereo microscope with measuring device. Drawings are mainly based on pictures taken with a digital camera with close-up lenses. Pictures of the oral apparatus were taken using a stereomicroscope with a mounted microscopy camera. All specimens of each series were examined to complete information on struc- tures missing in the DNA voucher according to the tissue sampling as well as to assess morphological vari- ability. Abbreviations used in the decriptions are as follows: BL, body length (distance from the tip of the snout to the body-tail junction); TAL, tail length; TL, total length (sum of body length and tail length); BW, maximum body width (at the widest point); BH, maximum body height; ED, eye diameter; IOD, interorbital distance (measured between the centers of the pupils); IND, internarial distance (measured between the centers of the narial apertures); S–N, distance snout-nostril (measured between the center of the nostril and the tip of the snout); N–E, distance nostril-eye (measured between the center of the nostril and the center of the pupilla); S– S, distance snout- spiracle (measured between tip of the snout and center of the spiracular apertures); TMH, tail muscle height (measured vertically from the junction of the body wall with the ventral margin of the tail muscle); TMW, tail muscle width (measured horizontally from the sinistral and dextral junction of the body wall with the tail muscle); MTH, maximum tail height (including fins and caudal musculature, taken at its maximal vertical extent); ODW, oral disc width (maximum width of the oral disc); DG, dorsal papillae gap.

Descriptions

Mantidactylus betsileanus (Boulenger, 1882)

Series examined. ZSM 848/2004 from Andasibe (1 specimen), ZSM 876/2004 (2 specimens) and 880/2004 (10 specimens) from Fierenana, ZSM 1080/2004 (1 specimen) from Ranomafana. Tadpoles were collected from small and very shallow, slow-flowing streams. Sampling was done in January/February 2003 and Febru- ary 2004. Taxonomic note. We here follow the definition of Mantidactylus betsileanus introduced by Blommers- Schlösser (1979). Calls of M. betsileanus were first described by Blommers-Schlösser (1979) and documented by Vences et al. (2006) on CD 2, Track 62. This species is common at Andasibe and Ranomafana (where it occurs sympatrically with M. sp. aff. betsileanus "very slow calls"), and was the only species of this complex heard at Fierenana. Description (based on a tadpole out of the series ZSM 880/2004, field number 2003.1668; Genbank accession number EF606877; Fig. 1). Tadpole in stage 38 and in good state of preservation (tip of tail slightly deformed naturally). Colouration in preservative, tan brown, posterior body dorsally and ventrally scattered with irregular brown flecks, caudal dorsal fin covered with some fawn flecks and spots, belly, caudal muscle and ventral fin almost uniformly transparently drab. In dorsal view, body shape elliptical, head and body separated by a constriction of the body wall at the plane of the spiracle. In lateral view, body depressed, BW 1.25 of BH. Nostrils rounded, not protuberant, positioned dorsolaterally, IND 0.62 of IOD, slightly closer to the tip of the snout than to the eyes, S–N 0.77 of N–E. Eyes positioned dorsally, directed dorsolaterally, moderately sized, ED 0.10 of BL. Spiracle sinistral, visible in dorsal view, attached to the body wall but its tip lifting up

LARVAL MORPHOLOGY OF MADAGASCAN FROGS Zootaxa 1616 © 2007 Magnolia Press · 51 TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited. from the body wall, closer to the end of the body than to the tip of the snout, S–S 0.66 of BL, spiracular opening oval, laterally positioned, oriented posteriorly, closer to venter than to dorsum. Intestinal spiral slightly visible in dorsal view, visible in lateral view, well visible in ventral view. Vent tube median, cloacal aperture dextral to caudal fin. Caudal musculature well developed, TMH 0.64 of BH and 0.60 of MTH, gradually tapering from base to the almost reached tip. Dorsal fin originating at the tail-body junction, slightly convex, point of maximum height located a bit further as midtail, MTH 1.07 of BH, margins almost parallel to caudal muscle. Oral disc of moderate size, generalized, emarginated, ODW 0.20 of BL and 0.33 of BW, anterioventrally positioned, with robust dark coloured beaks, nearly surrounded by marginal papillae. Oral disc margins not visible in dorsal view, but visible in lateral view. Keratodont row formula 1:4+4/1+1:2, about 38 keratodonts per mm. Keratodont row length reduces from A2 to A4. Upper labium with large dorsal gap, DG 0.84 of ODW, 8 marginal papillae and one submarginal papilla on each side, lower labium with two rows of papillae (app. 39 in marginal row, app. in 28 submarginal row). All papillae with a rounded tip, less than 1.0 mm in length each. Upper beak M-shaped, lower beak V-shaped, both with almost equal-sized serrations. Measurements (in mm). BL 12.8; TAL 25.7; TL 38.5; BW 7.6; BH 6.1; ED 1.3; IOD 4.5; IND 3.9; S–N 1.7; N–E 2.2; S–S 8.4; TMH 3.9; TMW 4.3; MTH 6.5; ODW 2.5; DG 2.1. Variation within the series. Series ZSM 880/2004 with 9 additional tadpoles in stages 25–36, some in poor state of preservation (measurements only partly included). Morphological proportions vary as follows: BW 1.09–1.33 of BH, ED 0.07–0.10 of BL, S–N 0.67–1.00 of N–E, IND 0.54–0.75 of IOD, S–S of BL 0.53– 0.63, TMH 0.46–0.61 of BH, TMH 0.42–0.53 of MTH, MTH 0.89–1.24 of BH, ODW 0.20–0.25 of BL, ODW 0.33–0.39 of BW (see Tab. 1). Keratodont row formula is not consistent. Stages 25 and 26 differ from older stages by keratodont row formulae 1:3+3/1+1:2 and 1:3+3/3. In earlier stages, the skin is more transparent and the intestinal spiral is well visible in dorsal view, lying under the apex of myotomes of the tail musculature.

FIGURE 1. Drawings of the preserved voucher tadpole of Mantidactylus betsileanus (stage 38), series ZSM 880/2004; (A) dorsal view, (B) lateral view, (C) oral disc.

TABLE 1. Variation in measurements, body proportions and keratodont row formulae of Mantidactylus betsileanus (undamaged specimens from series ZSM 880/2004). All measurements in millimeters. Each column represents measurements from a single specimen.

Stage25252626273638 BL 6.4 7.1 9.6 7.3 10.3 11.5 12.8 TAL 13.5 14.0 18.7 13.7 22.4 19.4 25.7 TL 19.9 20.2 28.3 21.0 32.7 30.9 38.5 Keratodont 1:3+3/1+1:2 1:3+3/1+1:2 1:3+3/3 1:3+3/1+1:2 1:4+4/1+1:2 1:4+4/1+1:2 1:4+4/1+1:2 row formula TAL/BL 2.1 2.0 1.9 1.9 2.2 1.7 2.0

52 · Zootaxa 1616 © 2007 Magnolia Press KNOLL ET AL. TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited.

Previous descriptions. In comparison to the data provided by Blommers-Schlösser (1979) our series differs only slightly. Our tadpoles in stages 25 and 26 have maximum body lengths of 7.1 and 9.6 mm, respectively, and therefore are approximately 0.6 mm larger than tadpoles of the same stages measured by Blommers-Schlösser (1979). This also refers to the maximum recorded body length in stages 26–30, which was given as 9.0 mm by Blommers-Schlösser (1979) and is 10.3 mm in a stage 27 larva from our sample. The average ratio of tail to body length is 2.05 (range 2.0–2.1) in our sample and 2.2 according to Blommers- Schlösser (1979). See discussion.

Mantidactylus sp. aff. betsileanus "very slow calls"

Series examined. ZSM 841/2004 (4 specimens) and 843/2004 (2 specimens) from An’Ala, ZSM 1114/2004 (1 specimen) and 1193/2004 (1 specimen) from Ranomafana, Maharira. Tadpoles from An'Ala were collected in ephemeral ponds of about three to five meters diameter, moderate depth and with a thick layer of dead leaves covering the ground. Sampling was carried out in January and February 2003 and February 2004. Taxonomic note. This is a sibling species of M. betsileanus, differing morphologically by a different shape of femoral glands in adult males, and by a conspicuously slower pulse repetition rate in advertisement calls. Calls of this species are documented by Vences et al. (2006) on CD 2 as Track 64 (as M. sp. aff. betsileanus "very slow"). Description (based on a tadpole from series ZSM 841/2004, field number 2003.1604; Genbank accession number EF606879; Fig. 2). Tadpole in stage 25 and in fairly good state of preservation. Colouration in preservative is transparent tan with a yellowish shade. Dorsum uniformly coloured transparent yellowish, above each nare an orange coloured spot/structure shining through the skin, also visible in ventral view. Dorsal fin, belly and ventral caudal fin almost lacking darker flecks. In dorsal view, body shape elliptical. In lateral view, body depressed, BW 1.43 of BH. Body shape in lateral view flattening toward snout. Nostrils directed dorsolaterally, round, not protuberant, positioned closer to the tip of the snout than to the eyes, S–N 0.86 of N–E. Eyes positioned dorsally, oriented dorsolaterally, moderately sized, ED 0.07 of BL, not visible in ventral view. Snout rounded in lateral view, nearly rounded in dorsal view. Spiracle sinistral, closely attached to the body wall, slightly visible in dorsal view. Spiracular opening oval, upper margin situated approximately at level of lower margin of eye, at mid-distance between end of body and tip of the snout, S–S 0.51 of BL. Intestinal spiral well visible in lateral and ventral view, also visible in dorsal view. Short, flattened vent tube, almost median, cloacal opening dextral to caudal fin, attached directly to the ventral fin. Caudal musculature well- developed, upper origin well visible in dorsal view, TMH 0.61 of BH and 0.57 of MTH, gradually tapering from base to the rounded tip. Dorsal fin originating on base of caudal musculature, increasing in height at about half of length of tail, MTH 1.07 of BH, maintaining equal height throughout two-third of the tail, slightly diminishing to finely rounded tip. Dorsal and ventral fin slightly convex in outline, dorsal fin at its maximum height slightly higher than ventral fin, margins nearly parallel with the margins of the tail muscle. Oral disc small, generalized, ODW 0.28 of BL and 0.53 of BW, transversely elliptical, directed anterioventrally, laterally emarginated, not visible in dorsal view, but margins visible in lateral view. All but its anterior margin free from the body wall. Large dorsal gap, DG 0.57 of ODW, upper labium with 8 marginal papillae and one submarginal papilla on each side. Lower labium with two continuous rows of papillae (app. 50 in marginal row). All papillae with a rounded tip, less than 1.0 mm in length each. Keratodont row formula 1:4+3/3, about 60 keratodonts per mm. Keratodont row length reduces from A2 to A4, with A5 consisting of 4 keratodonts only on the left side of the oral disc. Large medial gap in second anterior keratodont row. Beak well developed and dark brown, both jaw sheaths with serrations at their cutting edges. Upper beak M-shaped, lower beak V-shaped. Measurements (in mm): BL 7.5; TAL 14.4; TL 21.9; BW 4.0; BH 2.8; ED 0.5; IOD 2.0; IND 1.2; S–N

LARVAL MORPHOLOGY OF MADAGASCAN FROGS Zootaxa 1616 © 2007 Magnolia Press · 53 TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited.

1.2; N–E 1.4; S–S 3.8; TMH 1.7; TMW 1.8; MTH 3.0; ODW 2.1; DG 1.2. Variation within the series. The series ZSM 841/2004 and ZSM 843/2004 contain two undamaged tadpoles at stage 25 and one in stage 26. Proportions of stage 25 tadpoles vary as follows: BW 1.39–1.48 of BH, ED 0.07 of BL, S–N 0.93–1.00 of N–E, IND 0.70 of IOD, S–S 0.48–0.57 of BL, TMH 0.54–0.64 of BH, TMH 0.54–0.62 of MTH, MTH 1.00–1.04 of BH, ODW 0.27–0.28 of BL, ODW 0.51–0.54 of BW. Stage 26 differs by the following ratios: S–N 0.79 of N–E, IND 0.75 of IOD, S–S 0.63 of BL, TMH 0.52 of BH, TMH 0.45 of MTH, MTH 1.15 of BH, ODW 0.23 of BL, ODW 0.42 of BW (see Tab. 2). Keratodont row formula is not consistent, with larvae at stage 25 exhibiting keratodont row formulae 1:4+3/3, 1:4+4/3 and 1:3+3/3. The stage 26 larva has the keratodont row formula 1:3+4/3. General morphology and colouration is homogenous within the sample. Previous descriptions. See discussion.

FIGURE 2. Drawings of the preserved voucher tadpole of Mantidactylus sp. aff. betsileanus "very slow calls" (stage 25), series ZSM 841/2004; (A) dorsal view, (B) lateral view, (C) oral disc.

TABLE 2. Variation in measurements, body proportions and keratodont row formulae of Mantidactylus sp. aff. betsileanus "very slow calls" (undamaged specimens from series ZSM 841/2004, 843/2004). All measurements in millimeters. Each column represents measurements from a single specimen.

ZSM 841/2004 841/2004 841/2004 843/2004 Stage 25 25 25 26 BL 7.5 7.4 7.1 8.6 TAL 14.4 14.3 13.5 16.2 TL 21.9 21.7 20.6 24.8 Keratodont row formula 1:4+3/3 1:3+3/3 1:4+4/3 1:3+4/3 TAL/BL 1.9 1.9 1.9 1.9

Mantidactylus sp. aff. betsileanus "Vohidrazana"

Series examined. ZSM 1042/2004 (3 specimens), 1043/2004 (3 specimens) and 1063/2004 (1 specimen), all from the forest of An’Ala. Two of the tadpole series were collected in a quiet running brook of about two meters in diameter and 30 to 70 cm depth. The stream bed was sandy. The third series was sampled from a shallow system of connected puddles containing dead leaves which covered the ground. Taxonomic note. This is a further sibling species of M. betsileanus of which so far no adult morphological or bioacoustic data have become available. The known DNA sequence of this species is from the locality Vohidrazana, and the tadpoles from the nearby site An'Ala indicate that this species may be widespread at mid to low elevations of central eastern Madagascar.

54 · Zootaxa 1616 © 2007 Magnolia Press KNOLL ET AL. TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited.

Description (based on tadpole ZSM 1063/2004, field number 2003.2015; Genbank accession number EF606886; Fig. 3). Tadpole in stage 37 and in a fairly good state of preservation. Colouration in preservative is transparent tan with a yellowish shade. Dorsum uniformly transparent yellowish, above each external nare an orange coloured spot visible through the transparent skin, not visible in ventral view. Dorsal fin, belly and ventral caudal fin lacking darker flecks. In dorsal view, body shape elliptical. In lateral view, body very depressed, BW 1.41 of BH. Body shape in lateral view flattening toward snout. Nostrils directed dorsolaterally, round, not protuberant, positioned closer to the tip of the snout than to the eyes, S–N 0.83 of N–E. Eyes positioned dorsally, oriented dorsolaterally, moderately sized, ED 0.08 of BL, not visible in ventral view. Snout rounded in lateral view, nearly rounded in dorsal view. Spiracle sinistral, closely attached to the body wall, visible in dorsal view. Spiracular opening oval, upper margin situated approximately at level of lower margin of eye, slightly closer to the end of the body than to the tip of the snout, S–S 0.58 of BL. Intestinal spiral well visible in lateral and ventral view, also visible in dorsal view. Short, flattened vent tube, almost median, cloacal opening dextral to caudal fin, attached directly to the ventral fin. Caudal musculature well developed, upper origin well visible in dorsal view, TMH 0.63 of BH and 0.59 of MTH. Dorsal fin originating on base of caudal musculature, increasing in height at proximal third of tail, MTH 1.07 of BH, maintaining equal height throughout two-third of the tail, slightly diminishing to finely rounded tip. Dorsal fin slightly convex in outline, ventral fin almost straight, equally sized, margins nearly parallel with the margins of the tail muscle. Oral disc small, generalized, ODW 0.22 of BL and 0.40 of BW, transversely elliptical, directed anterioventrally, laterally emarginated, not visible in dorsal view, but margins visible in lateral view. All but its anterior margin free from the body wall. Large dorsal gap, DG 0.74 of ODW, upper labium with 8 marginal papillae and one submarginal papilla on each side. Lower labium with two continuous rows of papillae (app. 46 in marginal row). All papillae with a rounded tip, cores pigmented, less than 1.0 mm in length each. Kerat- odont row formula 1:4+4/3, about 60 keratodonts per mm. Keratodont row length reduces from A2 to A5, large medial gap in second anterior keratodont row. Beak well-developed and dark brown, both jaw sheaths with serrations at their cutting edges. Upper beak M-shaped, lower beak V-shaped. Measurements (in mm). BL 10.3; TAL 19.8; TL 30.4; BW 5.8; BH 4.1; ED 0.9; IOD 3.1; IND 2.0; S–N 1.5; N–E 1.8; S–S 6.1; TMH 2.6; TMW 2.4; MTH 4.4; ODW 2.3; DG 1.7. Variation within the series. The series ZSM 1043/2004 and ZSM 1042/2004 contain three additional undamaged tadpoles at stages 26, 37 and 40, respectively. TL ranges from 26.3 mm (stage 26) to 32.9 mm (stage 40), and BL from 8.6 mm (stage 26) to 10.8 mm (stage 40). The proportions vary as follows: BW 1.19– 1.42 of BH, ED 0.08–0.10 of BL, S–N 0.75–0.84 of N–E, IND 0.60–0.79 of IOD, S–S 0.56–0.63 of BL, TMH 0.65–0.87 of BH, TMH 0.60–0.75 of MTH, MTH 0.92–1.16 of BH, ODW 0.23–0.24 of BL, ODW 0.40–0.48 of BW (see Tab. 3). Keratodont row formula is homogenous within the sample, as are general morphology and colouration. Previous descriptions. See discussion.

FIGURE 3. Drawings of the preserved voucher tadpole of Mantidactylus sp. aff. betsileanus "Vohidrazana" (stage 37), series ZSM 1063/2004 (left leg cut for tissue sample); (A) dorsal view, (B) lateral view, (C) oral disc.

LARVAL MORPHOLOGY OF MADAGASCAN FROGS Zootaxa 1616 © 2007 Magnolia Press · 55 TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited.

TABLE 3. Variation in measurements, body proportions and keratodont row formulae of Mantidactylus sp. aff. betsileanus "Vohidrazana" (undamaged specimens from series ZSM 1042/2004, 1043/2004, 1063/2004). All measurements in millimeters. Each column represents measurements from a single specimen.

ZSM 1063/2004 1042/2004 1042/2004 1043/2004 Stage37374026 BL 10.6 10.6 10.8 8.6 TAL 19.8 22.3 22.1 17.7 TL 30.4 32.9 32.9 26.3 Keratodont row formula 1:4+4/3 1:4+4/3 1:4+4/3 1:4+4/3 TAL/BL. 1.9 2.1 2.0 2.1

Mantidactylus biporus (Boulenger, 1889)

Specimen examined. ZSM 879/2004 (1 specimen) from Fierenana. The tadpole was collected in a puddle beside a footpath on a muddy ground. Collection was done in February 2003. Taxonomic note. Based on the original description and on our own examination of the types (M. Vences, unpublished), the defining character of M. biporus are two separated rudimentary clusters of femoral glands in females. Furthermore, the type specimens of M. biporus exhibit larger body sizes than many other populations assigned to M. biporus. Based on these observations, the population from Fierenana (with double glands in females and relatively large body sizes) qualifies as representing this species more closely than any of the populations listed by Vences et al. (2006) as M. sp. aff. biporus. Description (based on a single tadpole, ZSM 879/2004, field number 2003.1667; Genbank accession number EF606883; Fig. 4). Tadpole in stage 37 (no other conspecific larvae were found) and fairly good state of preservation (bit from the dorsal tail fin has been excised for DNA analysis). The colouration of the preserved tadpole is pale brown to tan with a yellowish shade. Dorsum scattered with slightly darker irregular flecking, as dorsal fin, especially on the body-tail junction; belly and ventral caudal fin transparent yellowish, almost lacking darker flecks. In dorsal view, body shape elliptical. In lateral view, body depressed, BW 1.72 of BH. Body shape in lateral view flattening toward snout. Nostrils directed dorsolaterally, round, not protuberant, positioned closer to the eyes than to the tip of the snout, S–N 1.25 of N–E. Eyes positioned dorsally, oriented dorsolaterally, moderately sized, not visible in ventral view, ED 0.09 of BL. Snout rounded in lateral view, nearly rounded in dorsal view. Spiracle sinistral, closely attached to the body wall, not visible in dorsal view. Spiracular opening oval, upper margin situated approximately at level of lower margin of eye, closer to the end of the body than to the tip of the snout, S–S 0.61 of BL. Intestinal spiral slightly visible in lateral view and well visible in ventral view. Short, flattened vent tube, almost median, cloacal opening dextral to caudal fin, attached to the ventral fin. Caudal musculature well developed, TMH 0.74 of BH and 0.62 of MTH, gradually tapering from base to not clearly visible pointed tip. Dorsal fin originating on base of caudal musculature, as far as noticeable due to the excision, increasing in height at the end of the proximal third of tail, MTH 1.19 of BH, maintaining equal height throughout two-third of the tail, slightly diminishing to finely rounded tip. Dorsal and ventral fin slightly convex in outline, equally sized, margins nearly parallel with the margins of the tail muscle. Oral disc small, generalized, ODW 0.21 of BL and 0.33 of BW, transversely elliptical, directed anterioventrally, laterally emarginated, not visible in dorsal view, but margins visible in lateral view. All but its anterior margin free from the body wall. Large dorsal gap, DG 0.69 of ODW, upper labium with 8 marginal papillae and 4 submarginal papillae on each side. Lower labium with two rows of papillae (app. 34 in marginal row, app. 26 in submarginal row). All papillae with a rounded tip, cores of papillae are pigmented, less than 1.0 mm in length each. Keratodont row formula 1:3+3/1+1:2, about 50 keratodonts per mm. Kerat-

56 · Zootaxa 1616 © 2007 Magnolia Press KNOLL ET AL. TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited. odont row length reduces from A2 to A4, large medial gap in second anterior keratodont row. First posterior row with small medial gap and app. 1.5 times larger keratodonts. Beak well developed and black, both jaw sheaths with serrations at their cutting edges. Upper beak M-shaped, lower beak V-shaped. Measurements (in mm). BL 15.2; TAL 32.1; TL 47.3; BW 9.8; BH 5.7; ED 1.3; IOD 4.5; IND 2.6; S–N 2.0; N–E 1.6; S–S 9.3; TMH 4.2; TMW 4.3; MTH 6.8; ODW 3.2; DG, 2.2. Previous descriptions. None.

FIGURE 4. Drawings of the preserved voucher tadpole of Mantidactylus biporus (stage 37), series ZSM 879/2004 (part of dorsal fin cut for tissue sample); (A) dorsal view, (B) lateral view, (C) oral disc.

Discussion

The tadpoles of Brygoomantis described herein are all benthic forms (see Altig & McDiarmid 1999) with a depressed body shape, dorsally directed eyes and relatively low fins. The oral discs are not very specialized and all exhibit a wide gap in dorsal papillae, as formerly described for larvae of this species group by Blom- mers-Schlösser (1979). In external morphology, all larvae described within the subgenus are quite similar. This is furthermore confirmed when considering the results of Blommers-Schlösser (1979) for M. alutus, M. betsileanus, M. curtus and M. ulcerosus. Differences between species are slight and mainly concern overall body shape, size at a certain developmental stage and keratodont row formulae. However, the latter character has been proven to vary among individuals of one species as a consequence of environmental conditions in European species (see Vences et al. 2002) and such variation is partly also evident in some of our samples of Brygoomantis. Furthermore, keratodont row formulae may change during ontogenesis from younger to older stages (e.g. Grosjean 2006), as documented here for the two undescribed species. The tadpoles of species in the M. ulcerosus group described by Blommers-Schlösser (1979) were based on specimens originating from various different localities and due to the cryptic species diversity recently discov- ered by application of DNA barcoding, Blommers-Schlösser's (1979) allocation of species names to the populations described has to be put into question. For example, Blommers-Schlösser (1979) described tadpoles of M. betsileanus based on 45 specimens collected at Ambodiriana, Andasibe, Mandraka valley and Tampoketsa d’Ankazobe. At that time, all were treated as a single species due to the lack of significant morphological differences. From the current point of view, chances are high that several species were mixed in these descriptions from the late 1970s. Thus, comparisons of our results with those of Blommers-Schlösser (1979) suffer from great uncertainty regarding the taxon they are actually referring to. Our tadpole specimens of Mantidactylus betsileanus are slightly larger than those of M. betsileanus described by Blommers-Schlösser (1979). The average ratio of tail to body length is 1.94 (range 1.7–2.2) instead of 2.3 given by the latter author. The morphology of larvae of the two undescribed species, M. sp. aff. betsileanus "very slow calls" and "Vohidrazana", generally falls within the variation described for M. betsilea-

LARVAL MORPHOLOGY OF MADAGASCAN FROGS Zootaxa 1616 © 2007 Magnolia Press · 57 TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited. nus by Blommers-Schlösser (1979). The only disagreement seems to concern the average ratio of tail to body length, which is 1.9 in stages 25–26 of M. sp. aff. betsileanus "very slow calls" and 2.0 in stages 26–40 of M. sp. aff. betsileanus "Vohidrazana", instead of 2.2/2.3 (Blommers-Schlösser 1979). A distinctive larval feature of the two undescribed species (especially M. sp. aff. betsileanus "very slow calls") seems to be the presence of a pair of orange coloured spots posteriomedian to the nostrils. They obvi- ously constitute an internal morphological structure of unknown origin and function, visible through the skin in dorsal view. M. betsileanus and M. sp. aff. betsileanus "very slow calls" furthermore also appear to differ in the number of keratodonts per millimetre, which is approximately 38 vs. 60. This indicates that these very closely related syntopic species may be more strongly differentiated in larval morphology than in adult morphology, and possibly also occupy different trophic niches in their larval stage. Further studies on larval morphology in the subgenus Brygoomantis may likely reveal well differentiated ecological traits among the sympatric species identified, correlating with a stronger niche segregation than previously anticipated from field observations of adults.

Acknowledgements

We are grateful to Marta Puente, Roger-Daniel Randrianiaina, Liliane Raharivololoniaina and David R. Vie- ites for help during fieldwork. This work was carried out in the framework of collaboration agreements between the Institute for Biodiversity and Ecosystem Dynamics of the University of Amsterdam, the Zoolo- gische Staatssammlung München, the Département de Biologie Animale of the University of Antananarivo, and the Association Nationale pour la Gestion des Aires Protégées ANGAP. Ralf Hendrix and an anonymous reviewer provided valuable comments. Stefan Scheu (TU Darmstadt) partly supervised the work of the senior author. We thank the Madagascan authorities for research and export permits, and the Volkswagen Foundation and the Deutsche Forschungsgemeinschaft (project Ve 247/2-1) for financial support.

References

Altig, R. & McDiarmid, R.W. (1999) Body plan: development and morphology. In: McDiarmid, R.W. & Altig, R. (Eds.). Tadpoles: the biology of anuran larvae. University of Chicago Press, Chicago, pp. 24–51. Arnoult, J. & Razarihelisoa, M. (1967) Le genre Mantidactylus, adultes et formes lavaires de M. betsilianus (Blgr.), M. curtus (Blgr.) et M. alutus (Prc.). Bulletin du Musée National d'Histoire naturelle Paris, 39, 471–487. Blommers-Schlösser, R.M.A. (1979) Biosystematics of the Malagasy frogs, I. Mantellinae (Ranidae). Beaufortia, 29, 1– 77. Blommers-Schlösser, R.M.A. & Blanc, C.P. (1991) Amphibiens (première partie). Faune de Madagascar, 75 (1), 1–379. Dubois, A. (1992) Notes sur la classification des Ranidae (Amphibiens, Anoures). Bulletin mensuel de la Société Linéenne de Lyon, 61, 305–352. Dubois, A. (1995) Keratodont formulae in anuran tadpoles: proposals for a standarization. Journal of Zoological System- atics and Evolutionary Research, 33, I–XV. Glaw, F. & Vences, M. (1994) A fieldguide to the amphibians and reptiles of Madagascar. 2nd ed., including mammals and freshwater fish. Vences and Glaw Verlag, Köln, Germany. Glaw, F. & Vences, M. (2006) Phylogeny and genus-level classification of mantellid frogs (Amphibia, Anura). Organ- isms Diversity & Evolution, 6, 236–253. Glaw, F., Hoegg, S. & Vences, M. (2006) Discovery of a new basal relict lineage of Madagascan frogs and its implica- tions for mantellid evolution. Zootaxa, 1334, 27–43. Gosner, K. L. (1960) A simplified table for staging anuran embryos and larvae with notes on identification. Herpetolog- ica, 16, 183–190. Grosjean, S. (2006) The choice of external morphological characters and developmental stages for tadpole-based anuran taxonomy: a case study in Rana (Sylvirana) nigrovittata (Blyth, 1855) (Amphibia, Anura, Ranidae). Contributions to Zoology, 74, 61–76. Grosjean, S., Thomas, M., Glaw, F. & Vences, M. (2006) The tadpole of the Malagasy treefrog Boophis rufioculis:

58 · Zootaxa 1616 © 2007 Magnolia Press KNOLL ET AL. TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited.

molecular identification and description (Amphibia, Anura, Mantellidae). Spixiana, 29, 73–76. Palumbi, S.R., Martin, A., Romano, S., McMillan, W.O., Stice, L. & Grabowski, G. (1991) The Simple Fool's Guide to PCR, Version 2.0., Privately published, Univ. Hawaii. Raharivololoniaina, L., Vieites D.R., Glaw, F. & Vences, M. (2003) Larval stages, habitat and distribution of the hyper- oliid frog Heterixalus rutenbergi (Boettger, 1881). Alytes, 21, 59– 65. Randrianiaina, R.D., Glaw, F., Thomas, M., Glos, J., Raminosoa, N. & Vences, M. (2007) Description of the tadpoles of two species of Gephyromantis, with a discussion of the phylogenetic origin of direct development in mantellid frogs. Zootaxa, 1401, 53–61. Thomas, M., Raharivololoniaina, L., Glaw, F., Vences, M. & Vieites, D.R. (2005) Montane tadpoles in Madagascar: molecular identification and description of the larval stages of Mantidactylus elegans, Mantidactylus madecassus, and Boophis laurenti from the Andringitra Massif. Copeia, 2005, 174–183. Vejarano, S., Thomas, M. & Vences, M. (2006) Comparative larval morphology in Madagascan frogs of the genus Guibemantis (Amphibia: Mantellidae). Zootaxa, 1329, 39–57. Vences, M. & Glaw, F. (2001) When molecules claim for taxonomic change: new proposals on the classification of Old World treefrogs. Spixiana, 24, 85–92. Vences, M., Puente, M., Nieto, S. & Vieites, D.R. (2002) Phenotypic plasticity of anuran larvae: environmental variables influence body shape and oral morphology in Rana temporaria tadpoles. Journal of Zoology, London, 257, 155–162. Vences, M., Glaw, F. & Marquez, R. (2006) The calls of the frogs of Madagascar. 3 Audio CD's and booklet, 44 pp. Alosa, Barcelona

LARVAL MORPHOLOGY OF MADAGASCAN FROGS Zootaxa 1616 © 2007 Magnolia Press · 59 TERM OF USE This pdf is provided by Magnolia Press for private/research use. Commercial sale or deposition in a public library or website site is prohibited.