Hyperolius viridiflavus (Duméril and Bibron)

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 : 1990-1993

Hyperolius viridiflavus (Duméril and Bibron)

(figure 10)

This name encompasses a confusing complex of morphologically and ethologically similar forms of reed frog that especially vary in coloration. Some populations are spatially isolated, while in others clinal variation is evident. Experiments revealed that at least some populations are not reproductively isolated from each other. Numerous names are available for the different forms scattered over almost the entire tropical and southern subtropical Africa, the validity of many of which seems unwarranted. Apparently, a viridiflavus subgroup from western, eastern and south-eastern Africa can be identified within this taxonomic chaos. According to molecular data, it seems to contain 10 species, at least (for a summary see Schiøtz, 1971, 1975, 1999; Largen, 1998; Wieczorek et al., 2001). The nominotypical form is known from forest and bush land in western Kenya, and ranges into adjacent Uganda and north into Ethiopia ( Schiøtz, 1975, 1999). According to Wieczorek et al. (2001), it is conspecific with a phenotype from Uganda originally named Rappia bayoni Boulenger.

Diagnosis. (1) SVL males 29.2¡ 1.16 mm (26.8–31.2 mm, n ~11) (302 non-collected males had mean SVL 29.5¡ 1.98 mm, range 22.3–38.7 mm), SVL of preserved females unknown (21 non-collected females had mean SVL 31.5¡ 2.15 mm, range 27.2–35.5 mm); (2) TIBL/SVL 0.51¡0.03 (0.45–0.55, n ~11), HW/SVL 0.3¡0.01 (0.29–0.33, n ~11); (3) dorsal surface with scattered pointed warts, tuberculate below eye and in tympanic area; (4) snout shape dorsally semi-circular to rounded and laterally rounded, nares visible from above; (5) E-N/EYE 0.71¡0.06 (0.60–0.81, n ~11), canthus rostralis straight or slightly convex from tip of snout to nostril and concave from nostril to eye; (6) tympanum distinct or covered by thick skin, TYMP/EYE 0.30¡0.02 (0.28–0.34, n ~6); (7) FOOT/TIBL 0.86¡0.06 (0.77–0.97, n ~11); (8) foot webbing formula: 1(0–1), 2i(1) 2e(0–K), 3i(1–1K) 3e(0–K), 4i(1–1K) 4e(1), 5(0); hand webbing formula: 1(1), 2i/e(1), 3i(1K–2) 3e(1–1K), 4(1); (9) PhJ is dorsally tan or brown with a dark brown lateral band from behind the eye to the groin and occasionally two dark brown mid-dorsal lines and a dark brown canthal line; dorsal pointed warts always dark brown; dorsolaterally sometimes and ventrolaterally always bordered by a cream area, labial area and ventral side whitish, venter occasionally with tiny red spots, inner extremities more or less dark red, and sole and palm, webbing as well as parts of the upper sides of toes and fingers always dark red. PhF is dorsally light to dark green and yellow to orange with one of these colours dominating, dorsal pointed warts always brown or dark green, usually surrounded by yellowish spots on a greenish ground and sometimes with additional dark green spots, ventrolateral area yellowish, most intense in labial area, ventrally as PhJ; the gular flap of males is bright yellow; the iris is brownish (information provided is based on collected material 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 AY323920 (~NMK A/3866/5); (11) LTRF 1/2 or 1/3.

Species that are similar to H. viridiflavus include other species within the viridiflavus complex (cf. Wieczorek et al., 2001). These can be distinguished from H. viridiflavus sensu stricto by their unique colour patterns as summarized by Schiøtz (1999).

Life history. In the Kakamega Forest, the nocturnal arboreal H. viridiflavus was common between March and October. Males and females were found in disturbed primary forest at ponds. Females were also observed migrating through the forest. Males emitted vocalizations from 0.5 to w 2 m above ground. Breeding took place throughout the entire observation period; hence, H. viridiflavus clearly is a prolonged breeder. Eggs were deposited in small assemblages (of ca 30 eggs) below the water surface and attached to vegetation. They are half-pigmented. Larvae can be characterized as lentic and omnivorous.

The advertisement call of H. viridiflavus is a xylophone-like ‘peep’, composed of a single pulsed note (figure 2H; table 1). Our data match the description and figures of calls related to H. viridiflavus from elsewhere (e.g. Schiøtz, 1975, 1999).

Differentiation of sympatric species in the field

Adults of the five species can be best distinguished by characters (1), (3), (4), (6), (8) and (9). Hyperolius acuticeps is the smallest, H. viridiflavus the largest species. The other species lie inbetween, in part overlapping and in H. kivuensis with overlap also with H. viridiflavus . All species except H. acuticeps possess warty skin with tuberculate areas below eye and/or tympanic area. Hyperolius cf. cinnamomeoventris , H. kivuensis and H. lateralis are more or less coarse while H. viridiflavus has pointed warts. Tympani in all species can be either distinct or covered by skin except H. acuticeps (always covered by skin). Inter-specific foot webbing variation is limited, mainly with slight modifications on toes III and V. Hyperolius acuticeps shows least webbing in general, and H. viridiflavus has relatively little webbing on toe III but has toe V fully webbed. Hyperolius viridiflavus is unique in having hand webbing (except for some H. kivuensis , which may develop rudimentary hand webbing, too). Each of the five taxa displays a species-specific colour pattern (figures 6–10). Those of H. acuticeps and H. viridiflavus (PhJ and PhF) are unique. However, due to variation some H. cf. cinnamomeoventris PhJ are difficult to distinguish from some H. lateralis PhJ and others are difficult to distinguish from some H. kivuensis (while H. lateralis and H. kivuensis cannot be confused). Hyperolius kivuensis is larger than H. cf. cinnamomeoventris and its snout shape is more pointed (cf. figures 7 and 8). The gular flap is yellow in H cf. cinnamomeoventris and whitish in H. kivuensis . In H. lateralis , the ventral sides are always translucent pink to reddish (versus cream in H. cf. cinnamomeoventris ). The gular flap is dark yellowish golden in H. lateralis and yellow in H. cf. cinnamomeoventris . Iris coloration is lighter in H. lateralis (light brownish grey versus bronze to dark brown).

All five species possess clearly distinguishable advertisement calls as shown in figure 2D–H.

Tadpoles of the five species are very similar to each other and occur synchronically in the same ponds. Their life history is poorly understood and it remains unknown how they specialize.

Partitioning of species

Our information on temporal partitioning of the five species studied is limited to a period from March to October (with an interruption in September) which excludes the dry season. All species except H. acuticeps and H. lateralis were found throughout the entire time. The two latter species were present at reproduction sites and actively breeding from April/May to June only. The other species found throughout were actively breeding at water bodies or were found in the forest or, in the case of H. cf. cinnamomeoventris , also in semi-cleared areas. Through capture/ re-capture experiments with marked specimens (unpublished), we have evidence that males of H. cf. cinnamomeoventris , H. kivuensis and H. viridiflavus remain at breeding sites for long periods (probably throughout the year) while females immediately disappear after oviposition. Circadian activity was the same in all five species with calling and reproductive behaviour at night time (in all species calling started immediately after sunset).

As shown in figure 11, there is almost no spatial partitioning among these arboreal species that were observed at three different types of water body (swampy area, permanent pond and non-permanent pond) and in forest. However, H. viridiflavus was almost exclusively found in permanent water. Only in October, few males were observed at one non-permanent pond. In contrast, H. acuticeps males were rare at permanent water.

Advertisement calls are most different in temporal parameters and in H. acuticeps also in frequency. Differences in size with H. acuticeps and H. viridiflavus as lower and upper end extremes may suggest that food resources vary throughout the five species.

Our knowledge of tadpoles of Hyperolius species at the Kakamega Forest is limited. As far as we know, they belong to the same eco-morphological guild.