Mantidactylus ambohimitombi ambohimitombi Boulenger, 1919 Type material.— Mantidactylus ambohimitombi Boulenger, 1919 is based on a series of syntypes that include BMNH 1947.2.26.25–32 from ‘ Ambohimitombo Forest’. Blommers-Schl̂sser and Blanc (1991) noted that syntypes BMNH 1947.2.26.31– 32 may be referable to Mantidactylus curtus, without justification. We here designate the DNA-barcoded specimen BMNH 1947.2.26.25, a large-sized individual of 65.0 mm SVL, as lectotype. Lectotype designation is justified by the need to stabilize this and other nomina in Brygoomantis, given the uncertain identity and morphological similarity of many taxa in the subgenus. Identity.— Mantidactylus ambohimitombi Boulenger, 1919 is typically considered a valid species (BlommersSchl̂sser & Blanc 1991; Frost 2021; Glaw & Vences 1992a, 1994), although it was considered of uncertain status by Glaw and Vences (2007). It is morphologically close to M. curtus but distinguished by its distinctly larger body size. It was not assigned to a genetic lineage by Vieites et al. (2009). A 16S sequence of the lectotype surprisingly clusters with a lineage predominantly known from the Ankaratra Massif. We therefore redefine this lineage provisionally as corresponding to M. ambohimitombi. In previous studies the populations from the Ankaratra Massif have been considered as confirmed candidate species M. sp. 19 by Vieites et al. (2009), and M. sp. Ca19 by Perl et al. (2014). They were referred to as ‘ M. sp. aff. curtus “Ankaratra”’ by Schmidt et al. (2009). We emphasize that this attribution is preliminary; since only mtDNA data (no genomic information) are available from the lectotype, and no fresh samples are available from Ambohimitombo forest where the original syntype series was collected, we cannot exclude that mitochondrial introgression has taken place, potentially blurring a hypothetical differentiation between the Ankaratra and Ambohimitombo populations. Our Phylonetworks analysis provided evidence for gene flow of syntopic M. curtus into M. ambohimitombi marefo ssp. nov. from Itremo (Fig. 5), providing a first hint that reticulated evolution may have played a role in the origin of the various morphologically divergent frogs that we here subsume in the species M. ambohimitombi. A more in-depth analysis of ranges, and of gene flow among various lineages of the M. curtus clade (M. curtus, M. bourgati, M. ambohimitombi) is necessary to understand their evolutionary history and verify their taxonomy. In our phylogenomic tree, two other lineages form a monophyletic group with specimens of M. ambohimitombi from Ankaratra, and these are in subsequent accounts described as subspecies of M. ambohimitombi; see the rationale in the respective accounts below. Diagnosis.— Mantidactylus ambohimitombi is a member of the M. curtus clade; it is here defined as containing three deep genetic lineages considered as subspecies, and is sister to the morphologically very distinct M. madecassus. See Table 4 for a list of diagnostic morphological characters. The following account only diagnoses the nominal subspecies M. a. ambohimitombi from other species in Brygoomantis; see below for diagnoses and comparisons of the two other subspecies. The combination of relatively large body size of up to 51 mm at Ankaratra (up to 65 mm in the type series from Ambohimitombo forest), slightly granular skin without dorsolateral ridges in most specimens, strongly developed foot webbing with fully webbed fifth toe, small tympanum diameter with a maximum of 11% of SVL in males, distinguishes M. a. ambohimitombi from species of the other clades. Within the M. curtus clade, M. alutus, M. madecassus and M. pauliani have smaller body sizes and are distinguished by either a usually shorter snout (M. madecassus, M. pauliani), or advertisement calls emitted in regular series (in M. alutus, vs single notes) (Table 4); M. curtus usually has a smoother skin and a somewhat shorter snout; M. bourgati is morphologically very similar but appears to occur only on the Andringitra Massif. For detailed distinction from new species described herein, see the respective species accounts.A full list of molecular diagnostic sites in the 16S gene of M. a. ambohimitombi in pairwise comparisons to all other Brygoomantis species is provided as Supplementary appendix. Variation.—Variation in measurements is given in Table 5. See Fig. 13 for colouration in life and its variation. A light vertebral stripe occurs in few individuals. There is pronounced sexual size dimorphism (at Ankaratra, largest confirmed male SVL 37.7 vs female SVL 50.6 mm). As discussed above, the type series is comprised of particularly large-sized animals reaching 65.0 mm in SVL. Femoral glands are only distinct in some male specimens, possibly due to seasonal effects; in two specimens, ZSM 190/2021 (ACZCV 334) and ZSM 195/2021 (ACZCV 342), FGL and FGW are 4.7 mm x 3.8 mm and 4.0 mm x 3.0 mm, respectively. Natural history. — A common species on theAnkaratra Massif, especially above the tree line along streams and swamp in montane savanna and heatland, but also in rainforest (reported by Vences et al. 2002 under the name M. curtus). Males were found calling during the day underwater. Calls.— The advertisement call of M. a. ambohimitombi, recorded on 16 February 2006, 14:50 h, at Ankaratra, consists of a short, pulsed note (Fig. 14), emitted in series at slow succession and somewhat irregular intervals. The calls were emitted from several shy animals at the border of a mountain stream, partly underwater, and identification of the calling specimen was therefore impossible. Notes exhibit slight amplitude modulation, with maximum call energy occurring either at first third of the note’s length or at the centre of the note, and the terminal pulse of the note always being of lowest energy. The initial pulse is sometimes separated from the second pulse by a slightly longer inter-pulse interval. Numerical parameters of 20 analysed calls are as follows: call duration (= note duration) 136– 218 ms (180.2 ± 18.7 ms); 9–14 pulses per note (10.9 ± 1.5); pulse duration 5–8 ms (6.3 ± 1.0 ms); pulse repetition rate within notes 56.1–69.3 pulses/s (62.3 ± 5.1); dominant frequency 925–1012 Hz (959 ± 30 Hz); prevalent bandwidth 500–3100 Hz; call repetition rate (= note repetition rate) in regular series ca 11–20 calls/ min. Tadpoles. —A tadpole of M. a. ambohimitombi was described under the name ‘ M. sp. aff. curtus “Ankaratra”’ by Schmidt et al. (2009). Distribution.— Apparently endemic to a small area of the central highlands of Madagascar (Fig. 7). The nominal form is known from Ankaratra. Mitochondrial sequences assignable to this lineage have also been recorded from Analafohy and Antoetra; however, confirmation is needed, especially at Antoetra, as to whether this is evidence of true co-occurrence, or is a result of introgression with M. curtus, which is common in this locality. The type locality Ambohimitombo forest is close to Antoetra. Elevation range: 1150–2380 m a.s.l. Etymology.—Formulated from the type locality, ‘Ambohimitombo Forest’.

Published as part of Scherz, Mark D., Crottini, Angelica, Hutter, Carl R., Hildenbrand, Andrea, Andreone, Franco, Fulgence, Thio Rosin, Köhler, Gunther, Ndriantsoa, Serge Herilala, Ohler, Annemarie, Preick, Michaela, Rakotoarison, Andolalao, Rancilhac, Loïs, Raselimanana, Achille P., Riemann, Jana C., Rödel, Mark-Oliver, Rosa, Gonçalo M., Streicher, Jeffrey W., Vieites, David R., Köhler, Jörn, Hofreiter, Michael, Glaw, Frank & Vences, Miguel, 2022, An inordinate fondness for inconspicuous brown frogs: integration of phylogenomics, archival DNA analysis, morphology, and bioacoustics yields 24 new taxa in the subgenus Brygoomantis (genus Mantidactylus) from Madagascar, pp. 113-311 in Megataxa 7 (2) on pages 178-182, DOI: 10.11646/megataxa.7.2.1, http://zenodo.org/record/7441023