Distributed under Creative Commons Cc-by 4.0

The toad species Sclerophrys capensis Tschudi, 1838 was erected for a single specimen from South Africa which has never been properly studied and allocated to a known species. A morphometrical and morphological analysis of this specimen and its comparison with 75 toad specimens referred to five South African toad species allowed to allocate this specimen to the species currently known as Amietophrynus rangeri. In consequence, the nomen Sclerophrys must replace Amietophrynus as the valid nomen of the genus, and capensis as the valid nomen of the species. This work stresses the usefulness of natural history collections for solving taxonomic and nomenclatural problems.


INTRODUCTION
In zootaxonomy, early nomina (scientific names) often have a complex story. Many original type specimens have been lost and thus the status of the nomina cannot be assessed. In other cases, type specimens are present but have never received the necessary attention and been studied properly. This is the case of the amphibian nomen Sclerophrys capensis Tschudi, 1838, a toad from Cape of Good Hope described by Tschudi (1838) as a member of his family Bombinatores. This nomen is based on a single specimen (MNHN RA 0.742), whose holotype (holophoront) by original monotypy (monophory) has since then been kept in the collections of the Paris Muséum national d'Histoire naturelle (MNHN) where it has permanently been available for study by taxonomists.
The specimen was collected in the Cape Province of South Africa by Pierre Antoine Delalande and Jules Verreaux (then 12 years old). They made three field trips to the east and northwest of this province and collected 13,400 specimens, including 322 specimens of Reptiles and Amphibians (Delalande, 1822). The first travel to the east started on 11 November 1818 and was stopped by the battle of Grahamstown where 10,000 Xhosa people confronted English troops on 22 April 1819. A second trip started on 5 July 1819, led north to the Olifants River and enabled collection in the marshes of the Berg River. The third expedition started on 2 November 1819 and lasted 8 months. Delalande and his men went by boat to the Algoa Bay and then inland to the Keiskamma River. On 1 December 1820, Delalande returned back to France where he arrived at the end of the year (Delalande, 1822;Barnard, 1956;Varley, 1956).
Among the specimens collected by Delalande and Verreaux was a subadult toad which was described by Tschudi (1838) as a new genus and species Sclerophrys capensis. The nomen Sclerophrys was mentioned in the taxonomic review of volume 8 of the Erpétologie Générale (Duméril & Bibron, 1841, 300), but not in any synonymy list of amphibian species of this book. All Bufo specimens collected by Delalande were allocated by these authors to Bufo pantherinus Smith, 1828. However, apart from the holotype of Sclerophrys capensis, all the bufonid specimens collected by Delalande presently in the MNHN collection belong in the species now known as Vandijkophrynus angusticeps (Smith, 1848), but until 2006 known as Bufo angusticeps. The holotype of Sclerophrys capensis then was listed in the type catalogue of the Paris Museum of Guibé (1950, 15-16), who stated that by its characters it should be considered a member of the genus Bufo Garsault, 1764.
The problem posed by the existence of a generic nomen of bufonids unallocated to a group of recent species was not a real one as long as the genus Bufo was not divided, but it becomes a real one if the genus has to be split, at least in subgenera. Among 33 synonyms of the generic nomen Bufo listed by Duellman & Trueb (1985, 533), Sclerophrys was the only one having an African type species, and it remained so until 2006: it was therefore bound to be the valid one for an African genus or subgenus of bufonid if such taxa had to be erected. We do not think that ignoring available nomina whenever they pose nomenclatural problems is good taxonomy: keeping such nomina as nomina dubia is providing the conditions for such nomina to become a problem whenever any taxonomist decides to unearth them. We consider that solving the Sclerophrys capensis nomenclatural problem should have been a preliminary action to take before coining new nomina for African bufonid genera that might possibly include this species. The same applies to the description of new species of this group or the resurrection of some specific nomina from synonymy, as has been the case in the recent decades. The aim of the present paper is to allocate unambiguously this specimen to a bufonid species present within the region visited by Delalande and Verreaux. For this, after unsuccessful attempts at obtaining nucleic acids from this specimen, we used both a morphometrical and a morphological approach.
Based on the geographic origin of the specimen, there are two possibilities, which both would result in the invalidation of a nomen coined by Frost et al. (2006): Amietophrynus if the nomen is finally applied to Bufo pantherinus, Bufo pardalis or Bufo rangeri, and Vandijkophrynus if it is applied to Bufo angusticeps or Bufo gariepensis. The nomina created in Frost et al.'s (2006) work are very recent and do not in the least qualify for conservation under the nomen protectum rule of the Code.

Methods
A total of 75 specimens belonging to five species of South African bufonids, Amietophrynus pantherinus (2 males, 2 females, 4 young), Amietophrynus pardalis (3 females, 4 young), Amietophrynus rangeri (6 males, 5 females, 2 young) Vandijkophrynus angusticeps (7 males, 23 females, 4 young) and Vandijkophrynus gariepensis (2 males, 9 females, 2 young), that occur in the Cape region, were studied and compared with the holotype of Sclerophrys capensis. On every specimen, 35 measurements (Dubois & Ohler, 1999) were taken by a single observer (AMO), either with a slide caliper or with an ocular micrometer (measurements smaller than 5 mm). In order to correct for size, every measurement was transposed into its logarithm and divided by the mean of the 35 logarithm-transposed measurements of the specimen (Mosimann, 1970). The ln-transformed variables were analysed using Discriminant Analysis. The holotype of Sclerophrys capensis was included without group membership for subsequent allocation to one of the groups.
The following morphological characters were used for morphological description of the specimens studied and for allocation of the holophoront of Sclerophrys capensis to a group (Poynton & Lambiris, 1998;Du Preez & Carruthers, 2009): spot on snout (distinct; indistinct; absent); spots on eyelids (paired spot; continuous band on head); skin on throat (granular; smooth); distal subarticular tubercle of finger III (a unique tubercle; paired tubercles); extension of web on outer side of toe III (number of phalanges free of web; see Savage & Heyer, 1967); fringes on digits (absent; narrow; broad); fringes on toes (absent; narrow; broad); separation of parotoid and eye (touching; separated); shape of parotoid (elongate; oval; broadened oval); parotoid height (flat; prominent); middorsal line (present; absent).

Description of the holotype of Sclerophrys capensis Tschudi, 1838
MNHN 0.742, young male; in rather bad state of conservation (Fig. 1). Specimen of medium size (SVL 41.2 mm), body rather robust.
Dorsal and lateral parts of head and body: snout smooth; between the eyes, back and flank with glandular warts. Large region, including upper eyelids and zone between the eyes and between the parotoids and anterior part of back of hardened aspect. Laterodorsal folds absent; supratympanic fold absent. Parotoid glands present, oval, elongate, perforate, more than two times longer (PL 10.4 mm) than wide (PW 4.1 mm), longer as distance between them (PD 8.6 mm), touching posterior border of eye. Cephalic ridges absent. (15) Co-ossified skin absent. Dorsal parts of limbs: forelimbs smooth; thigh, leg and tarsus with flattened glandular warts. Throat, chest, belly and thigh with glandular warts; foot smooth. Presence of macroglands: parotoid glands present.
Color of dorsal and lateral parts of head and body: whitish; ''hardened region'' dark brown. Dorsal parts of limbs: whitish. Ventral parts of head, body and limbs: whitish.
Male sexual characters. Presence of small testis. Nuptial spines and vocal sac openings absent.

Comparisons
In Discriminant Analysis, the percentage of variance accounted for was 70.5% on the first axis, 17.9% on the second axis (Table 1 and Fig. 2) and 7.8% on the third axis. The first axis had high loadings of the characters describing the body size, head shape, length of hand and leg, distance between parotoids and webbing, and proved useful for discrimination between the species (Chi 2 = 314.3; df = 124; p = 0.000) ( Table 2). The second axis had high loadings of the characters describing distance between eyes and some characters measuring head shape, and also showed significant discrimination between the species (Chi 2 = 169.6; df = 90; p = 0.000). The third axis had high loadings concerning eye nostril distance, outer metatarsal tubercle length and webbing, but is not significantly discriminant between species (Chi 2 = 85.9; df = 58; p = 0.010). The analysis enabled separating angusticeps, gariepensis and pantherinus but there is some ambiguity in the distinction between pardalis and rangeri. The holotype of Sclerophrys capensis, included without group membership into the analysis, was clearly allocated to the group including Amietophrynus rangeri (Table 3 and Fig. 2).
The study of the morphological characters (Table 4) supports this allocation of the specimen of Sclerophrys capensis. This toad can be distinguished from Vandijkophrynus angusticeps by its granular skin on the throat (smooth with fine furrows in V. angusticeps), and by a single subarticular tubercle distally on finger III (this tubercle is double in V. angusticeps). The other bufonid species from the Cape region are morphologically very similar. Vandijkophrynus gariepensis shares a granular skin on throat and single A. pantherinus can be distinguished by its very distinct dorsal pattern consisting of large paired neatly outlined spots. These spots are paired but smaller and not neatly outlined in A. pardalis. In A. rangeri, the pair of spots of the upper eyelids forms an uninterrupted band. Such a band can be present in the other species but in them this character state is much rarer. A middorsal line is absent in V. gariepensis and A. rangeri but present in A. pantherinus and A. pardalis. Sclerophrys capensis is clearly an Amietophrynus as it has granular skin on the throat and single subarticular tubercles on fingers and toes. Its allocation to a species is tenuous on morphological evidence because the most discriminating characters are coloration and colour pattern and these are faded in this old specimen. The dark spot in the head region of the Sclerophrys specimen does not show any lighter band in its middle so it might indicate that there was no separation of the lid spots and no middorsal line. Less than 2 phalanges are free of webbing on the outer side of toe III. These character states are concordant with Sclerophrys capensis being an Amietophrynus rangeri.

Taxonomic conclusion
The holotype of Sclerophrys capensis is allocated to Bufo rangeri by morphometric analysis. It shares character states with this species and other species presently allocated to Amietophrynus. Thus Sclerophrys is a subjective senior synonym of Amietophrynus and is the valid nomen of the genus. This nomen has been used after 1899 and its junior synonym Amietophrynus is a very recent nomen, published in 2006, and thus does not comply with the requirements of the Code for reversal of reference. Bufo regularis rangeri was described as a subspecies by Hewitt (1935) and has been considered since Poynton (1964) as a valid taxon at the species level. The validity of the species was never discussed, and this nomen appears in all recent lists of amphibians of South Africa. Searching on the 'Web of Science,' we found 29 references (including only four in which the nomen appears in the title of scientific works). However, the specific nomen capensis has been used four times in the literature after 1899 (see above). Therefore the conditions are not met to consider rangeri as a nomen protectum, and this nomen should be replaced by capensis. Thus the valid nomen of the toad currently known as Amietophrynus rangeri (Hewitt, 1935) is Sclerophrys capensis Tschudi, 1838.
The following nomenclatural changes are a consequence of the synonymy of Sclerophrys capensis with Amietophrynus rangeri:
This work stresses the importance of natural history collections for resolving taxonomic and nomenclatural problems. As shown here this can be done using morphological information only and do not always require the recourse to molecular data: compare Cappellini et al. (2014) and Dubois, Nemésio & Bour (2014).