Stumpffia miery, Ndriantsoa, Riemann, Vences, Klages, Raminosoa, Rödel & Glos, 2013

Ndriantsoa, Serge H., Riemann, Jana C., Vences, Miguel, Klages, Johannes, Raminosoa, Noromalala R., Rödel, Mark-Oliver & Glos, Julian, 2013, A new Stumpffia (Amphibia: Anura: Microhylidae) from the Ranomafana region, south-eastern Madagascar, Zootaxa 3636 (4), pp. 575-589 : 579-586

Stumpffia miery sp. nov.

Figs 1–3

Remark. The new species has been listed as confirmed candidate species Stumpffia sp. 1 by Vieites et al. (2009). Not included by Glaw & Vences (2007) and Wollenberg et al. (2008); the names Stumpffia sp. 1 and S. sp. “Ranomafana” as used by these authors refer to the second (large-sized) species of Stumpffia in the Ranomafana region which in turn was named S. sp. 9 by Vieites et al. (2009).

Holotype. ZMB 77453 (field and tissue #_611), adult male, Madagascar, Ranomafana region, Ambolo forest fragment , S 21°15.752', E 047°30.412' (site name: FFst04), app. 660 m a.s.l., near a small stream, 1 March 2011, coll. J.C. Riemann & A. Telo.

Paratypes. ZMB 77454 (field and tissue #_677), S 21°17.824' E 47°35.971' (site name: MBP 02), app. 515 m a.s.l., banana plantation with coffee and fruit trees, bush and shrub vegetation; below a forest fragment in Andalangina near Ifanadiana , 10 March 2011, other data as holotype ; ZMB 77455 (field and tissue #_768), Ambolo forest fragment , S 21°15.818', E 47°30.540' (site name: FFst09), app. 640 m a.s.l., near a small stream, 4 April 2011, other data as holotype ; ZSM 121/2011 (field and tissue #_804), S 21°17.962', E 47°35.861' (site name: MGSst02), app. 450 m a.s.l., coffee and banana plantation in Andalangina , 6 April 2011, other data as holotype ; ZMB 77456 (field and tissue #_09), Ambolo forest fragment , S 21°15.815', E 47°30.553' (site name: FFst09), app. 670 m a.s.l., 9 March 2010, coll. SHN, JCR, J. Solo, A. Telo & MOR ; UADBA_A 62120 (field and tissue #_621), S 21°15.748', E 47°30.447' (site name: FFns04), app. 680 m a.s.l., forest app. 150 m to next stream, 2 March 2011, other data as holotype ; ZSM 2447 /2007 ( ZCMV 5868 ; GenBank#: FJ559309 ), Ranomafana region, Ambohitsara, S 21°21.431' E 47°48.941', 294 a.s.l., 3 March 2007, coll. MV ; UADBA_A 62121 (field and tissue #_770), Ambolo forest fragment , S 21°15.835', E 047°30.557' (site name: FFns09), app. 700 m a.s.l., forest app. 120 m to next stream, 4 April 2011, other data as holotype ; UADBA_A 62122 (field and tissue #_1393), Andalangina, S 21°17.909', E 047°36.201' (site name: FFst02), app. 490 m a.s.l., near stream in forest fragment, 2 February 2011, coll. SHN and J. Solo ; UADBA_A 62124 (field and tissue#_1738), 10 March 2011, other data as ZSM 121/2011 ; all adult, calling males.

Diagnosis. A small species of Stumpffia assigned to the genus (and distinguished from all other Malagasy anurans) based on combination of its small size (male SVL <16 mm; species of all other Malagasy microhylid genera are larger), tips of fingers and toes not broadly expanded, absence of maxillary and vomerine teeth, terrestrial habits, advertisement calls consisting of single tonal note series, and digital length reduction in external view (not shown by any other species belonging to Malagasy microhylid genera). Molecular phylogenetic data are in agreement with this genus assignment as they consistently place the new species into a clade ( Stumpffia sensu stricto) containing most nominal species and also the type species of the genus ( Figs. 4–5 ; see Vieites et al. 2009 for a tree containing a near-complete outgroup sampling of all Malagasy microhylids).

Within the genus, Stumpffia miery sp. nov. is characterized by the following combination of characters: (1) small adult size, male SVL 13.2–14.6 mm (n= 10); (2) non-expanded finger discs; (3) outer palmar tubercle round; (4) canthus rostralis rounded in cross-section; (5) large distinct tympanum, its diameter 40% of the eye diameter (holotype; range 40–57%); (6) dorsal colour beige, typically with dark teddy-bear-shaped markings in life; (7) four fingers recognizable in external view, but first finger highly reduced and second and fourth fingers considerably reduced in length; (8) all five toes recognizable, with length reduction of first toe; (9) length of tonal notes in advertisement calls 40–88 ms.

Stumpffia miery differs from Stumpffia be, S. grandis , S. hara , S. megsoni , S. roseifemoralis , and S. staffordi by a smaller size (male SVL 13–15 vs. 17–28 mm), and by more reduced fingers I and IV; furthermore from Stumpffia be, S. hara , S. megsoni , and S. staffordi by non-expanded finger disks (vs. expanded), from S. grandis by less distinct marbling of dark colour on throat and chest (vs. distinct black marbling on white-bluish ground colour), and from S. roseifemoralis by absence of reddish colour on limbs (vs. presence). It differs from S. tridactyla by the presence of four fingers of which one is vestigial (vs. three of which two are vestigial) and of five toes (vs. three). Stumpffia tetradactyla resembles the new species by its teddy bear shaped dorsal pattern, but S. miery differs in the length reduction of fingers (first, second and fourth finger vs. first finger only) and presence of five toes (vs. four). It differs from S. pygmaea by a larger size (13–15 mm vs. 10.9–12 mm) and a slightly larger tympanum (40–57% vs. 33–42% of eye diameter). Stumpffia miery differs from S. psologlossa by a length reduction in first, second and fourth finger (vs. not reduced), and by differences in call characteristics (a tonal note vs. a distinctly pulsed note). It differs from S. gimmeli by a length reduction in first, second and fourth fingers (vs. no digital reduction). It differs from S. madagascariensis by a larger size (13–15 mm vs. 9.5–11.5 mm). Stumpffia helenae , a species possibly in need of generic reallocation, resembles S. miery in body size, but can be distinguished by its expanded disks of fingers and toes (vs. non-expanded in S. miery ).

Furthermore, S. miery shows genetic divergences higher than 8 % with respect to all other nominal species of Stumpffia for the 16 S rRNA gene fragment analyzed.

Description of the holotype [measurements in mm]. Adult male specimen in moderate to good state of preservation; right foot missing ( Fig. 3 ). Snout-vent length 13.2 mm; body slender; head slightly longer (4.3) than wide (3.6), not wider than the body; snout rounded to slightly pointed in dorsal view, protruding in profile; large oval nostrils directed laterally, slightly protuberant, nearer to the tip of the snout (0.8) than to the eye (0.9); eye large (1.5), pupil round; internarial distance (1.4) smaller than interorbital distance (2.8); canthus rostralis distinct and rounded; loreal region straight to slightly convex; tympanum distinct in life, less distinct in preservation, round, its diameter (0.6) is 40% of the eye diameter; the supratympanic fold indistinct; tongue long and narrow, tip broadly rounded, converging towards base, tongue base bulbous; maxillary teeth absent; vomerine teeth absent; choanae small and round. Arms slender; lower arm length 2.4; hand length to tip of finger III: 2.2; only finger III with two small, flat, round, subarticular tubercles; two large flat palmar tubercles, inner tubercle ovoid, outer tubercle round; fingers without webbing; relative length of fingers I<II=IV<III; finger I highly reduced in length and fingers II and IV considerably reduced; finger discs not enlarged, broadly rounded; prepollex absent; nuptial pads absent; tibiotarsal articulation reaches the tympanum; hindlimbs slender; femur length (4.6) slightly shorter than tibia length (5.6); length of tibiofibulare including toe IV: 8.1; lateral metatarsalia strongly connected; inner metatarsal tubercle indistinct, flat, ovoid; outer metatarsal tubercle absent; toe tips slightly broader than rest of toe without forming discs; toe tips with circummarginal grooves; distinct small, round subarticular tubercles; relative length of toes I<II<V<III<IV; no pedal webbing. Toe I reduced in length.

Skin on dorsum finely tuberculate; ventral skin smooth; throat with longitudinal fold of loose skin. After 10 months in preservative, dorsum light brown with distinct dark brown hour glass (teddy bear) pattern, a small line extending almost to tip of snout, pairs of semicircular convexities of this dorsal pattern extending to lower parts of eyelids, the region just dorsal of forelimbs insertion, and anterior to pelvic region; posterior and inner parts of this pattern a bit lighter than edges; teddy bear pattern bordered by thin white lines; darker median round patch just anterior to vent; dark canthal stripe; loreal region and upper lip dark, the latter with a few small white spots; posterior upper part of tympanum bordered by black semicircular patch, lower part pointing to forearm bases; distinct black lateral band runs from above the forearm to almost the groin area on right flank, on left flank it is broken into two blotches; black inguinal bands; black bands on flanks and hips bordered by light lines; upper arm light brown, elbow with dark spot; lower arms and hands with dark spots or cross bars; vent framed by black blotch; thighs brownish with irregular black dots; shanks with three dark cross bars, the median ones being most distinct; foot with dark spots; lower mandible dark with white spots; throat densely covered with dark brown points, few small round areas almost white; pectoral region and belly with irregular pattern of narrow brownish blotches, otherwise white; ventral parts of thighs slightly mottled brown; ventral parts of arms, hands, thighs and feet darker brown. Colours in life slightly more contrasting, back with reddish brown basic colour, iris red, upper arms red; ventral parts greyish instead of whitish ( Fig. 1 a, b, g).

Variation. The paratypes, all being adult males, equal the holotype in general measurements (see Table 1 ) and body shape. The most conspicuous variation in measurements was that the tibiotarsal articulation reaches the tympanum in some specimens (ZMB 77453, 77454; UADBA_ A 62120 , 62121) and eye in others (ZMB 77455; ZSM 121/2011; UADBA_ A 62122 , 62124). Further variation was found in tympanum diameter relative to eye diameter (40–57%). Larger variation was observed in colour pattern ( Fig. 1 ). Whereas the teddy bear like black dorsal pattern was always present, the basic dorsal colouration varied from light brown ( Fig 1 a, b, c) to darker reddish brown ( Fig. 1 d, g, h); in some specimens additionally interspersed with shorter dark lines and spots, a red dorsolateral line ( Fig. 1 c) or many tiny bluish white points ( Fig. 1 d, g, h). The teddy bear figure may be bordered by lighter lines ( Fig. 1 a, b, c, h) or lack them ( Fig. 1 d). The flanks mostly carry a continuous or broken broad black band, on a more or less mottled background. In UADBA_A 62121 the lateral spots are less conspicuous, the flank is completely mottled in dark brown and white ( Fig. 1 d). Ventral colour may deviate from the holotype to whitish venter, throat and lower parts of thighs, densely mottled dark brown to grey (UADBA_ A 62120 , Fig. 1 e). The skin texture in live frogs varied from almost smooth ( Fig. 1 a) to slightly granular ( Fig. 1 g) to very granular ( Fig. 1 d). Females of the new species are unknown.

Genetics and phylogeny. In a fragment of the 16S rRNA gene that is often used for assessments of molecular diversity in anurans (e.g., Vences et al. 2005) the new species had divergences higher than 8% with respect to all other nominal species of Stumpffia , which is similar to or higher than the divergences between most nominal species of the genus. It should nonetheless be mentioned that this value might be slightly inflated given that the compared fragment is shorter than the typically used one and contains a relative higher proportion of hypervariable (loop) nucleotide sites. Three out of the four specimens of S. miery sequenced (see Fig. 4 ) had identical 16S rRNA sequences, including two specimens from Ranomafana and the one from Ambohitsara (non-identical branch lengths in Fig. 4 being caused by slightly different sequence lengths). Only the holotype (ZMB 77453) differed by two substitutions (0.8% pairwise distance) from the other specimens. According to a more comprehensive data set by Klages et al. (submitted) the new species also displays a strong 16S divergence from all undescribed candidate species assigned to Stumpffia .

In a fragment of 341 base pairs of the nuclear RAG1 gene, the haplotype of the new species was not shared by any of the other species analysed which included all small-sized Stumpffia which could potentially be confused with S. miery . The minimum observed distance was to S. tetradactyla with 3.9% uncorrected p-distance (13 substitutions); all other Stumpffia sequences differed by 4.4–7.1% (15–24 substitutions). This high differentiation among Stumpffia species in a conserved nuclear gene, without haplotype sharing among species, is also confirmed in a more comprehensive analysis with many more sequences per species and including a large number of undescribed candidate species (Klages et al., submitted).

We are aware that the gene fragments analysed here are insufficient for a conclusive assessment of the phylogeny of Stumpffia and to test the monophyly of the genus. In general, a combination of more extensive mitochondrial and nuclear data sets than used to date will be necessary to produce a reliable hypothesis of phylogenetic relationships of these and other cophyline frogs. Despite these restrictions we here briefly report the trees resulting from phylogenetic analysis which serve to further visualize the considerable genetic differentiation among all species, including S. miery . Our analysis included all nominal species of Stumpffia to assess whether the new species might show close relationships to any of them.

The combined alignment of the 16S rDNA fragments, after exclusion of hypervariable stretches, gapped sites and stretches with a high number of missing data, contained 895 nucleotides. A total of 347 characters were variable and of these, 225 were parsimony-informative. The RAG1 dataset analysed contained 341 nucleotides. 60 characters were variable, and of these, 30 were parsimony-informative. The included sequences contained only few heterozygous sites and for simplicity these were excluded from further analysis. Overall, nodes in the 16S tree are better supported than the RAG1 tree. Both trees ( Figs. 4–5 ) recover a clade of Stumpffia species distributed in the northern and northwestern regions of Madagascar which also had been found by Köhler et al. (2010), although this clade is not significantly supported in the RAG1 tree. Especially in the 16S tree, several other clades receive strong support as well (especially the relationships among the species of the north/northwestern clade), but the position of S. miery is unsupported and can be considered as unresolved in both trees. Along with the relatively long branches characterizing S. miery in both trees this suggests that it is not particularly closely related to any other nominal species of the genus.

Vocalization. The call of the holotype (ZMB 77453) was recorded on 1 March 2011 during the afternoon (15:00h, 20.5°C, ca. 30 cm distance to microphone) at the type locality ( Fig. 6 ). The one analyzed advertisement call of Stumpffia miery comprises a single tonal chirping note that ranges in duration between 51–88 ms (73 ± 12, n= 10) and is emitted after relatively regular inter-note intervals that have a duration of 2679–4247 ms (3102 ± 456, n= 10), resulting in a call repetition rate of 0.3 per second. Frequency range is 7700–8300 Hz, the dominant frequency range of ten notes is 7751–8225 Hz (8057 ± 136.9). A possible low-intensity harmonic is visible in the sonagram around 4000 Hz which could be interpreted as fundamental frequency.

The paratype ZSM 2447/2007 was recorded on 3 March 2007 shortly after dusk in degraded rainforest near Ambohitsara, at an estimated temperature of 25°C. In this individual the note duration ranges between 40–53 ms (47 ± 5, n= 10) emitted in relatively regular, but shorter intervals than the holotype, between 1850–2610 ms (2105 ± 243, n= 10), resulting in an approximate call repetition rate of 0.5/s. The frequency ranges between 7500–8100 Hz, the dominant frequency of ten notes ranges between 7708–8010 Hz (7877 ± 109). A second harmonic, less intense than the dominant frequency, was observed at 11490–11970 Hz (11757 ± 169, n= 10).

Habitat. Most individuals were recorded in two years (2010, 2011) during the rainy season between 1 March and 4 April. In detail, ZMB 77456 was observed in 2010 during daytime in a small valley of the Ambolo forest fragment, app. 7 km east of Ranomafana. The frog was sitting on leaves on the forest floor while calling. Another calling male was observed within an accumulation of leaf litter and small branches. The habitat consisted of partly degraded forest with very dense shrub storey and partly open canopy. The calling sites were at the flanks and at the bottom of the valley, the latter close to an almost dried up small creek. Within a radius of app. 25 m another 3– 4 males were heard calling. A year later we collected ZMB 77455 at the very same site during the night. The holotype (ZMB 77453) was collected in the same forest fragment close to another small creek calling hidden in the leaf litter on a rocky hill flank during daytime ( Fig. 7 a). Two other specimens were collected during the night in the Ambolo forest fragment.

Four other males were found from a second locality, Andalangina, app. 10 km from the type locality. Two males were collected in the vicinity of a creek running through a mixed coffee and banana plantation ( Fig. 7 d). One male was captured in another nearby banana plantation also comprising coffee and other fruit trees, and dense shrub and bush vegetation ( Fig. 7 c). The fourth specimen was found in a degraded forest fragment close to a rocky almost dried up creek nearby the forest edge. Calling males were heard both during the afternoon (15:00h) and night hours (20:15–00:00h), always being well hidden in the leaf litter. Specimen ZSM 2447/2007 from Ambohitsara, at about 33 km from the type locality and 40 km east of Ranomafana, was found on the ground in a relatively open and degraded part of the rainforest, but no further habitat details were recorded.

Distribution. So far the new species is only known from three sites, close to the Ranomafana National Park.

Etymology. The species name is used as a noun in apposition, and is derived from the Malagasy word “ miery ”, which refers to the act of hiding, and alludes in this particular case to the secretive habits of the new species.

Figures

FIGURE 1. Stumpffia miery sp. nov. from the Ranomafana region, southeastern Madagascar, in life; a, b, g: male holotype (ZMB 77453); c, e: male paratype (ZSM 2447 / 2007); d, f: male paratype (UADBA _ A 62120); h: male specimen from type locality which was not collected; i: male paratype (ZMB 77454).

FIGURE 2. Hand (a) and foot (b) of male holotype (ZMB 77453) of Stumpffia miery sp. nov., ventral view.

FIGURE 3. Dorsal (left) and ventral (right) view of preserved male holotype (ZMB 77453) of Stumpffia miery sp. nov.

FIGURE 6. Sonagram and oscillograms of parts of a call (note series) of Stumpffia miery sp. nov., recorded from holotype specimen (ZMB 77453) on 1 March 2011 (15: 00 h, 20.5 ° C) at the type locality. The upper sonagram and oscillogram show a single note at a 100 ms scale whereas the lower oscillogram shows a 4 - note section of a call at a 10 s scale.

FIGURE 7. Habitats of Stumpffia miery sp. nov.; a & b type locality Ambolo forest fragment; small creek where the holotype ZMB 77453 was found (a) and terrestrial forest part (b) where the paratype UADBA _ A 62121 was collected; c & d banana plantations in Andalangina where the paratypes ZMB 77454 (c), ZSM 121 / 2011 and UADBA _ A 62123 (both d) were found.

FIGURE 4. 50 % - majority rule consensus tree from a Bayesian Inference analysis of 895 bp of the 16 S rRNA gene, including all nominal Stumpffia species. Scaphiophryne calcarata was used as outgroup. Two asterisks mark posterior probabilities of 0.99 - 1.0.

FIGURE 5. 50 % - majority rule consensus tree from a Bayesian Inference analysis of 341 bp of the nuclear RAG 1 gene. Scaphiophryne calcarata was used as outgroup. Asterisks mark posterior probabilities of 0.95 - 0.98 (one asterisk) and 0.99 - 1.0 (two asterisks).