Pyxicephalus

Scott, Elizabeth, Visser, John D., Yetman, Caroline A., Oliver, Lauren & Broadley, Donald G., 2013, Revalidation of Pyxicephalus angusticeps Parry, 1982 (Anura: Natatanura: Pyxicephalidae), a bullfrog endemic to the lowlands of eastern Africa, Zootaxa 3599 (3), pp. 201-228 : 222-223

Sympatry of Pyxicephalus species

According to Channing et al. (1994) and Channing (2001), P. adspersus and P. e d u l i s both occur on the Mozambique coastal plain, and local inhabitants reportedly eat the smaller species early in the rainy season, and the larger species after very heavy rainfall in February. Channing (2001: 351) notes that local tribesmen in Mozambique even prepare them for consumption differently. Channing (2001) also mentions that the local people believe the small animals from early in the season grow into the large animals, which are only found later in the season after heavy rains. This is unlikely, because skeletochronological research on P. adspersus indicates that it takes at least 3–4 years for a male to reach breeding size (Yetman et al. 2012).

These field observations and folklore are consistent with the occurrence of two phenologically distinct species of Pyxicephalus on the Mozambique plain. However, we suggest that these two species, one larger and the other smaller, are in fact P. edulis and P. angusticeps respectively; not P. adspersus and P. e d u l i s, as interpreted by Channing et al. (1994). It is important to recognise that P. e d u l i s is not a small species, and male specimens may attain considerable size. Parry (1982) reports that males of P. e d u l i s average 100 mm. Our unpublished data indicate that typical P. edulis males from Mozambique can reach 138 mm SVL (recorded for MNHNP 2010.0153, Fig. 7 ), but are not as large as the largest P. adspersus (maximum 191 mm SVL recorded for AMNH-A 23621). Misidentifications of P. edulis as P. adspersus that appear to be based on size are common in museum collections. Furthermore, no characters reliably distinguish the metamorphs of P. adspersus from P. e d u l i s until a fairly large size class is reached (hence our list of ‘indeterminate’ metamorphs in Appendix 1). A large proportion of museum material of Pyxicephalus comprises small metamorphs, which emerge in large numbers and are easily collected, but often poorly preserved. This material has generally been variously assigned to P. adspersus or P. edulis . The erroneous sympatric distribution reported for P. adspersus and P. edulis appear to be due largely to size-based museum determinations, together with random determinations of juveniles, compounded by general confusion arising from failure of some traditional characters to adequately diagnose these two taxa (see revised key). Parry (1982) misidentified some large individuals of P. e du li s from Botswana as P. adspersus (asterisks in Appendix 1). Specimens from Botswana have previously been difficult to identify. Parry (1982) and Poynton & Broadley (1985) mention ‘introgression’ of P. adspersus and P. e d u l i s characteristics in material from Botswana. However, Parry (1982) did not misidentify any adult material from Mozambique as P. adspersus (see Appendix 1). Our revised distribution data for re-examined adult museum specimens from Mozambique corroborates the studies of Parry (1982) and Poynton & Broadley (1985), neither of which listed any specimens of P. adspersus from the Mozambique plain, despite examining all material available in several major collections. Based on this evidence, we conclude that P. adspersus does not occur on the Mozambique plain.

We agree with Pickersgill (2007) that the apparent sympatry of P. adspersus and P. edulis mentioned by Channing et al. (1994), on the basis of an unspecified recording from Naboomspruit (Limpopo Province, South Africa) wherein both species were allegedly present, remains to be verified. The only available published sonograph from the locality (see Fig. 13 B) indicates only the long call. This also applies to Channing’s (2001: 350) statement that both species occur together at Nylsvlei (near Naboomspruit) and at Beira in Mozambique. The recording(s) and/or sonographs which form the basis of these statements of sympatry need to be re-examined. In the absence of adult voucher specimens of both P. adspersus and P. edulis from these localities, and given the apparent confusion of P. e d u l i s and P. angusticeps documented above, it is more parsimonious to assume that P. adspersus and P. e d u l i s are allopatric.

We suggest that P. angusticeps and P. e d u l i s may be sympatric on the Mozambique plain, although these two species have not yet been observed in the same microhabitat. If syntopy were the case, other isolating mechanisms must be in force separating P. edulis and P. angusticeps , e.g. differences in breeding phenology, diet and/or male vocalizations. Channing et al. (1994) documented differences in breeding phenology between the two species on the Mozambique plain (discussed earlier), but clear differences in calls between these two taxa have not rigorously been established yet, requiring further investigation. Larger general size in concert with a wider gape in P. e d u l i s could theoretically be responsible for ecological niche differentiation in diet between adults of this species and P. angusticeps . Lynch (1975) mentioned that wide gape of the wide-headed Pyxicephalus of Africa could be correlated with the specialized habit of predating on other frogs. Lynch (1975) restated the views of Hutchinson (1959) that greater head width and concomitant gape size in frogs seems to be a means of affecting niche separation without having to become larger. Pyxicephalus angusticeps is not as squat as its congeners, with relatively thin, long fingers and a slight increase in toe webbing, compared to P. e d u l i s. The narrower head, more gracile body form, slight increase in webbing and longer digits of P. angusticeps , as compared with P. edulis , suggest a more aquatic lifestyle. Pyxicephalus angusticeps may require more swamp-like habitats or flooded grasslands, based on the habitats (e.g. flooded rice paddies) in which it has predominantly been collected in Mozambique.

Figures

FIGURE 7. Pyxicephalus edulis Peters, 1854: MNHNP 2010.0153, adult 3 from Nhica do Rovuma, Mozambique. 7 A. Dorsal view. 7 B. Ventral view. 7 C. Lateral view, left. 7 D. Frontal view. 7 E. Hand. 7 F. Foot. Scale bar = 1 cm.

FIGURE 13. Spectrograms of Pyxicephalus Tschudi, 1838 produced via a DFT filter of 256 points. 13 A. Pyxicephalus adspersus Tschudi, 1838 from Bloemfontein, South Africa (recorded by Louis du Preez). 13 B. Pyxicephalus adspersus Tschudi, 1838 from Naboomspruit, South Africa (from collection of Alan Channing), with numerous calls of Kassina senegalensis Duméril & Bibron, 1841 visible above [published in Channing et al. 1994]. 13 C. Pyxicephalus edulis Peters, 1854 from ‘ Transvaal’ (from collection of Alan Channing). 13 D. Pyxicephalus edulis Peters, 1854 from Tschaneni, Swaziland (recorded by Vincent Carruthers). 13 E. Pyxicephalus angusticeps Parry, 1982 New Status from Beira, Mozambique (recorded by Alan Channing) [published in Channing et al. 1994].