Bioacoustics of Mantidactylus betsileanus

Diagnosis .—A member of the M. betsileanus clade as revealed by the phylogenomic analysis, and sister to the poorly known M. incognitus sp. nov. described below. See Table 4 for a list of diagnostic morphological characters. The combination of a relatively small body size in males (SVL 22–29 mm ) and distinctly larger size in females (SVL 30–37 mm ), slightly tubercular dorsal skin with distinct continuous dorsolateral ridges, reduced webbing (one phalanx of fifth toe free of web), absence of white spots on flanks, presence of a white marking on snout tip, and advertisement call consisting of a single, long note composed of more than 100 pulses distinguishes M. betsileanus from species of all other clades. Within the M. betsileanus clade, the species differs from all species except possibly M. incognitus sp. nov. (for which calls are unknown) by a higher number of pulses in its advertisement calls ( Table 4 ); it also differs from M. noralottae by smaller size of males, and from the sympatric M. katae sp. nov. by smaller size of femoral glands (see account of that species below). For a detailed distinction from its sister species M. incognitus sp. nov. , and from all other new species described herein, see the respective species accounts.A full list of molecular diagnostic sites in the 16S gene of M. betsileanus in pairwise comparisons to all other Brygoomantis species is provided as Supplementary appendix.

Calls .—The advertisement call of M. betsileanus , recorded on 29 January 1994 at Mandraka, 23–24°C air temperature ( Vences et al. 2006 : CD2, track 62, cut 1), consisted of a long, regularly pulsed note ( Fig. 35 ), emitted at more or less regular intervals, but always in slow succession. Notes exhibited some amplitude modulation, with call energy increasing rapidly to the maximum at the beginning of the note, followed by continuous decrease of energy towards the note’s end. Pulse repetition rate was higher in the centre of the note and slightly lower at its beginning and end. Call energy was distributed across a wide frequency range. Numerical parameters of four analysed calls were as follows: call duration (= note duration) 2907–3103 ms (3046.5 ± 93.3 ms); 165–178 pulses per note (173.0 ± 7.7); pulse duration 4–7 ms (5.5 ± 1.0 ms); pulse repetition rate within notes 46.3–64.5 pulses/ s (57.0 ± 5.1); dominant frequency 1399–1421 Hz (1410 ± 12 Hz), with a second peak of almost identical energy at around 2650 Hz; prevalent bandwidth 1100–6000 Hz; call repetition rate (= note repetition rate) ca 3.7 calls/min. Calls recorded on 12 February 2008 at a forest fragment northeast of Lake Alaotra, at an estimated 20–25°C air temperature, perfectly agreed with the call described above from Mandraka. Numerical parameters of five analysed calls were as follows: call duration (= note duration) 2464–3650 ms (3235.0 ± 438.3 ms); 181– 199 pulses per note (192.3 ± 9.9); pulse duration 4–7 ms (5.3 ± 0.9 ms); pulse repetition rate within notes 39.5–71.4 pulses/s (56.8 ± 12.1); dominant frequency 1356–1464 Hz (1404 ± 42 Hz), with a second peak of almost identical energy at around 2800–2900 Hz; prevalent bandwidth 1100–6000 Hz; call repetition rate (= note repetition rate) ca 3.1 calls/min. Calls recorded on 25 March 2006 , 18:00 h, at the crossing of the Moramanga-Anosibe An’Ala and Besariaka roads, estimated 20–25°C air temperature, also agreed with the calls described from Mandraka. Although difficult to analyse in detail due to many overlapping calls (including those of syntopic M. inaudax ), following numerical parameters could be measured (10 calls originating from at least three individuals analysed): call duration (= note duration) 1560–2279 ms (2010.9 ± 278.6 ms); 107–173 pulses per note (136.0 ± 33.7); pulse duration 2–5 ms (3.6 ± 0.8 ms); pulse repetition rate within notes 63.5–81.6 pulses/s (74.0 ± 6.5); dominant frequency 1261–1574 Hz (1414 ± 148 Hz), with a second peak of almost identical energy at around 2700–3150 Hz; prevalent bandwidth 1000–4200 Hz Calls recorded on 25 March 2006 , at night, at the orchid garden in Andasibe, estimated 20–25°C air temperature, agreed with the calls described above as well. Numerical parameters of six analysed calls, originating from two individuals are as follows: call duration (= note duration) 1887–2868 ms (2439.0 ± 351.9 ms); 173–201 pulses per note (184.3 ± 14.7); pulse duration 4–7 ms (5.5 ± 0.9 ms); pulse repetition rate within notes 56.0–105.3 pulses/s (70.1 ± 17.1); dominant frequency 1238–1340 Hz (1289 ± 38 Hz), with a second peak of almost identical energy at around 2700–2900 Hz; prevalent bandwidth 1000–4600 Hz; call repetition rate (= note repetition rate) ca 4–5 calls/min. A divergent call has been recorded from specimen MRSN A6343 (FAZC 13875; accession number of 16S sequence HM364713 ), at Betampona. The calls of this individual were recorded at 19:00 on 14 November 2007 , at 20°C air temperature. They consisted of a very long, regularly pulsed note, with very short pulses. Numerical parameters of five analysed calls were as follows: call duration (= note duration) 3490–7507 ms (6148.8 ± 1552.6 ms); 78–164 pulses per note (134.8 ± 33.6); pulse duration 2 ms (2.0 ± 0.0 ms); pulse repetition rate within notes 19.2–24.6 pulses/s (21.9 ± 1.7); dominant frequency 1670–1826 Hz (1719 ± 61 Hz); prevalent bandwidth 1000–4800 Hz. Especially the pulse repetition rate in these calls differed strongly from all other available recordings and we have therefore not included it in the characterization of this diagnostic feature in Table 4 . More data on the advertisement calls and genetics of M. betsileanus at Betampona are necessary to understand the identity of this population and exclude the possibility of mitochondrial introgression into another Brygoomantis species at this site. Tadpoles.— The tadpole of M. betsileanus was described by Blommers-Schl̂sser (1979), Knoll et al. (2007) , and Scheld et al. (2013) . The effect of diet on their development was reported by Soamiarimampionona et al. (2015) .

Natural history .—A common species whose typical calls are often heard from small running water bodies in rainforest but also in forest fragments and degraded areas or plantations nearby. Calls are emitted during day and night, especially during the day often from concealed positions directly at the edge of water. Very common in swamp areas, also ricefields next to rainforest, as long as the water is shallow and is at least very slightly flowing. Usually sitting in shallow water or along the stream bank, hiding in the leaf litter nearby water bodies or sometimes found sitting on low vegetation about 0.3 m hight. In Ranomafana and surrounds found at an elevational range between 550–1132 m a.s.l. M. betsileanus was successfully bred in captivity in the Mitsinjo amphibian husbandry research and captive breeding facility at Andasibe ( Edmonds et al. 2012 ). The size and structure of a population from near Andasibe was studied by Edmonds et al. (2019) .