Hyperolius kivuensis Ahl, 1931

Lötters, S., Schick, S., Scheelke, K., Teege, P., Kosuch, J., Rotich, D. & Veith, M., 2004, Bio-sketches and partitioning of sympatric reed frogs, genus Hyperolius (Amphibia; Hyperoliidae), in two humid tropical African forest regions, Journal of Natural History 38 (15), pp. 1969-1997 : 1987-1989

Hyperolius kivuensis Ahl

(figure 8)

This hyperoliid is known from south-western Ethiopia to western Kenya and Uganda, and from there south to Tanzania, Angola, the Democratic Republic of Congo and Zambia (cf. Schiøtz, 1975, 1999; Channing, 2001). It usually occurs in open bush land, savannah and rarely forest. Junior synonymy of H. bituberculatus Ahl from Rwanda (also considered to be a subspecies of H. kivuensis by several authors; e.g. Laurent, 1950, 1972) and H. simus Ahl from Burundi seem to be warranted because of variation among populations (cf. Schiøtz, 1975, 1999 by implication). The syntypes of H. simus clearly coincide with H. kivuensis , and consequently we support considering them synonymous. In contrast, Schiøtz (1975) and Largen (1998) discussed whether H. bituberculatus may be recognized as a valid taxon (e.g. as a subspecies of H. kivuensis ), because it differs from typical H. kivuensis in having a pigmented ventrum (versus unpigmented in H. kivuensis ). Interestingly, reading the descriptions by Ahl (1931), both H. kivuensis and H. bituberculatus are said to have pigmented ventral sides (i.e. dark grey). Unfortunately, the type material of H. bituberculatus is lost; that of H. kivuensis was available to us and has the ventrum well pigmented. We therefore suggest H. tuberculatus to be treated as a junior synonym of H. kivuensis . Hyperolius ipianae Ahl and H. raveni Ahl from Rwanda are in provisional junior synonymy, according to Laurent (1943, 1950). We examined the holotype of H. ipianae which seems unrelated to H. kivuensis (it may be related to H. quinquevittatus Bocage ). The type material of H. raveni is lost; however, according to the description by Ahl (1931: reddish with three white median lines and white lateral spots), we conclude it is not conspecific with H. kivuensis .

There are some species that can be similar to H. kivuensis but from which it appears to be reproductively isolated, i.e. H. balfouri (Werner) , H. cinnamomeoventris , H. concolor (Hallowell) , H. quinquevittatus Bocage and H. tuberilinguis Smith (cf. Schiøtz 1975, 1999; Poynton and Broadley, 1987; Largen, 1998).

Diagnosis. SVL males 28.9¡ 3.08 mm (25.1–32.3 mm, n ~4) (16 non-collected males had mean SVL 28.7¡ 1.61 mm, range 26.5–33.1 mm), SVL of one female 30.2 mm (two non-collected females had SVL 31.3 and 31.7 mm); (2) TIBL / SVL 0.46¡0.02 (0.43–0.48, n ~5), HW/ SVL 0.28¡0.01 (0.27–0.29, n ~5); (3) dorsal surface smooth to finely coarse, tuberculate below eye; (4) snout shape dorsally subovoid to subelliptical and laterally rounded, nares visible from above; (5) E-N/EYE 0.79¡0.07 (0.73–0.87, n ~5), canthus rostralis convex from tip of snout to nostril and straight (one male) or concave (one female) from nostril to eye; (6) tympanum distinct or covered by thick skin, TYMP /EYE 0.31 (one female) to 0.42 (one male); (7) FOOT / TIBL 0.90¡0.04 (0.84–0.93, n ~5); (8) foot webbing formula: 1(1), 2i(1) 2e(0–1), 3i(1) 3e(0–K), 4i(1K) 4e(1), 5(K); there may be rudimentary hand webbing as noted in some non-collected males; (9) PhJ and PhF are dorsally brown, olive or green (occasionally with brownish mottling), always with a black dorsolateral line from tip of snout to groin, usually also with a whitish line below the black one; ventral sides whitish, occasionally with greyish mottling; hidden parts of limbs usually red; the gular flap of males is whitish; the iris is golden to dark brown (information provided is based on collected and numerous non-collected specimens of both sexes); (10) for sequence of 560 bp fragment of mitochondrial DNA of the 16S ribosomal gene see GenBank under AY323919 (~ NMK A/3867/4); (11) LTRF 1/2 or 1/3.

Species that can be confused with H. kivuensis (depending on intra-specific variant) include H. balfouri ( Democratic Republic of Congo, Ethiopia, Uganda), H. cinnamomeoventris , H. concolor (West Africa), H. quinquevittatus ( Angola, Tanzania, Democratic Republic of Congo, Zambia, Malawi) and H. tuberilinguis (south-eastern Africa), which are all sympatrics except H. concolor and H. tuberilinguis (at Kakamega Forest, H. cf. cinnamomeoventris only) ( Laurent, 1950; Schiøtz, 1975, 1999; Poynton and Broadley, 1987). The first mentioned may become larger (up to 34.0 mm), exhibits a dark line on the snout ‘above’ the canthus rostralis (while H. kivuensis has a canthal line) and males have small dorsal excrescences which are lacking in H. kivuensis . Hyperolius quinquevittatus is smaller than H. kivuensis (maximum male size 24.0 mm) and H. tuberilinguis lacks a dark lateral line. The variable H. concolor is sometimes difficult to distinguish from H. kivuensis on the basis of external features alone, but the two are regarded as distinct allopatric taxa (cf. Schiøtz, 1975, 1999). For differentiation from the similar H. cf. cinnamomeoventris see below.

Life history. Hyperolius kivuensis is nocturnal and arboreal. In the Kakamega Forest, it was found between March and October in disturbed primary forest and at its edge. Males were always abundant and called from vegetation (up to w 2 m above ground) in swampy areas or at ponds, while females were occasionally also found migrating inside the forest. At one pond breeding occurred throughout the above-mentioned period, while at another smaller pond females appeared between May and June only. From our observations, H. kivuensis is a prolonged breeder. One single clutch was attached to vegetation above the water surface (up to 50 cm above it), having ca 200–243 cream eggs. Tadpoles are lentic and omnivorous .

Hyperolius kivuensis males produce a ‘crack’ composed of several pulsed notes (figure 2F; table 1), as described by Schiøtz (1975) for H. k. kivuensis and H. k. bituberculatus .