Pseudohemisus longimanus Angel, 1930

Name-bearing type: Lectotype by present designation, MNHN 1929.237, from “Ambongo” according to the original description.

The original syntype series included a second item, MNHN 1929.236, with probably non-conspecific tadpoles (see below) .

The type locality of this nomen, Ambongo, is located in north-western Madagascar, not far from the karstic massif of the Tsingy de Bemaraha, according to Blommers-Schl̂sser & Blanc (1991). The nomen was originally based on two syntypes. One of these, MNHN 1929.237, is a specimen in a moderately poor state of preservation that is probably an adult and is clearly assignable to the genus Scaphiophryne . The second number, MNHN 1929.236, refers to a series of about eight tadpoles in poor state of preservation, conserved in the same jar as MNHN 1929.237. Most of these tadpoles, upon superficial examination, show distinct blackish jaw sheaths and some are large; consequently, by these characters and general appearance clearly do not correspond to tadpoles of the subgenus Pseudohemisus in Scaphiophryne (Grosjean et al. 2007) . For this reason, we hereby designate MNHN 1929.237 as lectotype of Pseudohemisus longimanus Angel, 1930 . This stabilizes this nomen, in the event that future work on the S. calcarata complex reveals the existence of further species in addition to S. calcarata and S. obscura, for which this nomen may then be available.

Morphology. For measurements of the lectotype, see Table 1. The tarsal tubercle and enlarged inner metatarsal tubercle are clearly visible (Fig. 9). Subarticular tubercles are difficult to recognize, but are not fully flattened (as remarked in the original description). On the left side of the body, a dark stripe runs from the eye backwards for a few millimeters, reminiscent of the dorsolateral fold typical for the S. calcarata complex. Osteological examination revealed close resemblance to the other Pseudohemisus specimens examined here (Fig. 10, Supplementary Figs S2–S 3).

Assignment. This taxon was considered to be a synonym of S. calcarata by Guibé (1978) and Blommers-Schl̂sser & Blanc (1991). However, it is important to mention that due to the somewhat elongated head with a distinct neck, indistinct subarticular tubercles, and very large inner metatarsal tubercle, the lectotype specimen is also reminiscent of S. brevis, and some S. brevis also show a thin dorsolateral line or dorsolateral ridge in the anterior part of the body (as recognizable in the longimanus lectotype). According to Raselimanana (2008), S. brevis does occur in the Tsingy de Bemaraha area (along with populations of the S. calcarata complex). Based on its locality, it therefore cannot be ruled out that longimanus is a junior synonym of S. brevis . However, considering the probable presence of a dorsolateral fold, we here follow Guibé (1978) and Blommers-Schl̂sser & Blanc (1991) in assigning the nomen to the S. calcarata complex. Micro-CT scans of S. brevis, which were not produced for this study, might help to clarify this further.

No recently collected material and no DNA sequences of the S. calcarata complex are available from the Tsingy de Bemaraha area, which is located geographically between the known distribution areas of the western and the north-western lineages of S. obscura . While the data are insufficient to assign the nomen to either of these deep conspecific lineages, based on geography we assign Pseudohemisus longimanus Angel, 1930 as junior synonym to Scaphiophryne obscura .