Neurellipes rhoko, Sáfián, Szabolcs, 2014

Sáfián, Szabolcs, 2014, Neurellipes rhoko sp. n. from the Cross River Loop, Eastern Nigeria (Lepidoptera: Lycaenidae: Polyommatinae), Zootaxa 3860 (2), pp. 195-200 : 196-199

Neurellipes rhoko sp. n.

( FIG 1 A,C, 2 A)

Holotype: ♂, Nigeria, Cercopan Research and Conservation Area, Rhoko Forest, Iko Esai, Cross River Loop, 18.V.2012. Leg.: Szabolcs Sáfián. Coordinates: 5°39'27.20"N, 8°15'42.16"E. Type depository: African Butterfly Research Institute, Nairobi ( ABRI).

Paratype: 1♂, with the same data as the holotype (in coll. SZS).

Additional material examined: 1♂, Cameroon, Ebogo, III.2005. Leg.: ABRI (in coll. ABRI), 1♂ Nigeria, Cercopan Research and Conservation Area, Rhoko Forest, Iko Esai, Cross River Loop 10.XII.2008. Leg: Oskar Brattström (in coll: ABRI).

Description. Forewing length: 9 mm. The upperside ground colour is black, with a large orange patch in the discal area of the forewing, covering approximately half of the wing surface. The orange scaling covers approximately one-third of the forewing in spaces 5-6, its inner edge curving gently towards the base without a break. There is diffuse orange scaling in space 1, but it leaves a very narrow black margin along the costa. The hindwing has a smaller orange patch in the outer half, leaving a 1 mm broad black margin around it. The underside has the usual Neurellipes - Anthene zebra striation; it is of greyish tone, and with a tiny turquoise tornal eyespot, ringed faintly with orange. Female unknown.

Genitalia: The genitalia are small (1 mm) and very weakly sclerotised. The structure is rather simple with completely reduced uncus; the two sides of the tegumen are bridged only by a narrow belt (as seen on ventral view on Fig. 2 A). The sub-unci are slender and long, curving gently upward. The valves are rather rectangular, with two spines on their dorsal edge.

Diagnosis. Although their underside colour and pattern are almost identical, the new species is easily distinguishable from N. mahota , as the latter species is significantly larger than N. rhoko with an average forewing length of 12 mm ( Fig. 1 ). N. gola is also larger, and generally has a much whiter ground colour on the underside, the striation leaves larger white areas in the sub-marginal area of the hindwing, as well as in the median area of the forewing. The transverse bands are black in colour, rather than dark grey as appears in N. mahota , N. georgiadisi and N. rhoko ( Fig. 2 B). N. rhoko , apart from minor scattered scaling near the tornal eyespot—visible only in large magnification—has no orange colour in the hindwing submarginal area, while the orange scaling stretches down to the margin on all other species in the N. mahota -group, including N. georgiadisi , appearing at least as a well visible ring around the tornal spot. The shape of the orange patch on the hindwing is also different, extending further towards the base in N. mahota and N. gola . The new species is similar in size to N. georgiadisi but the orange patch on the hindwing of the latter species extends towards the margin, and the width of the orange patch is much narrower in space 5 on the forewing, ending with a sharp angle in the cell (also in N. mahota and N. gola ), while the margin of the orange patch is curving gently from the sub-apex towards the base, without a break on the forewing of N. rhoko . The wing shape of both wings of the holotype of N. georgiadisi ( Fig. 1 ) and another three specimens (collected in November 2012 by the author and Robert Tropek in the northern parts of Sapo National Park in Liberia) is also different; the forewing apex is more acute, the outer margin is almost straight (significantly more rounded in N. rhoko ) while the hindwing is also sharply angled at the tornus and the apex (rounded in N. rhoko ). The male genitalia also show some differences from those of N. gola and N. mahota , especially in the tooth-like projections (or spines) on the valves, and the shape and tip of the aedeagus (the tip of the aedeagus of N. mahota and N. gola are less pointy), being more similar to N. georgiadisi . The bridge of the tegumen is much broader in N. georgiadisi , also, it has a spine or tooth right on the tip of the valves, while the tip of the valves are lobed beyond the spine in N. rhoko . As the aedeagus of N. rhoko was damaged during dissection, it could not be included in the diagnosis.

Behavioural notes. The type-specimens (holotype and paratype) of N. rhoko were caught on the same day at about 10.30 in a forest edge in Rhoko Forest, which is a wet evergreen forest typical to the lowlands of the Cross River Loop and Western Cameroon. The specimens were sitting low down on the vegetation, similarly to those of small Neurellipes (e.g. N. fulvimacula = chryseostictus). When disturbed, they flew quite weakly just a couple of metres before they settled again. N. mahota , which was also caught in the same forest, and the West African N. gola have more powerful flight and they usually settle higher on the vegetation (normally at least a metre from the ground), whenever they descend to the forest interior, and when disturbed, they shoot higher with their strong flight, before they settle again (often above eye level). Similarly to other smaller species in the genus, N. rhoko is probably a canopy species in dense wet forest, which only occasionally descends to ground level. It is, therefore, difficult to detect, except at the edge of paths and clearings on hot sunny days, when many Neurellipes rest on the semi-shaded low vegetation.

Etymology. The species was named after the Rhoko River and conservation area in the Cross River Loop, Eastern Nigeria, where the Nigerian NGO CERCOPAN is running a conservation project on endangered primates and their natural habitats in collaboration with Iko Easi community. The Rhoko forest probably hosts more than 800 species of butterflies, based on the extensive exploratory work of Robert David Warren and Oskar Brattström (pers. comm.), and serves as an important buffer area to the Cross River National Park, with which it is contiguous.

Discussion and notes on the biogeography. During examination of the series of over 80 male N. mahota specimens in the butterfly collection of the African Butterfly Research Institute, Nairobi, two further male N. rhoko specimens were found from Ebogo, Cameroon and from the type locality: Rhoko , Nigeria with virtually identical size and pattern. The stability of the morphology of both species confirms that N. rhoko is not a smaller variant of the significantly larger N. mahota , with which it occurs. The differences in the genitalic structure of the males of N. mahota and N. rhoko also corroborate the specific status of the latter. This finding also indicates that N. rhoko is actually a Central African species, penetrating West Africa just to the Cross River, which is the westernmost limit of the range of many other Central African butterfly species ( Larsen 2005). Although the discovery of N. georgiadisi , another species in the N. mahota -group from Liberia was quite surprising, the finding of its eastern vicariant was not particularly strange, as quite a few species-groups show similar distribution pattern. The vicariants of a few other Central African species actually occur under similar ecological conditions in the Liberian „hyper-wet” rainforests, being separated by a few thousand kilometres gap of unsuitable habitats. One of the best examples are the closely related N. mahota and N. gola , where N. mahota has the distribution area centred in Central Africa, occurring in West Africa only in Northwestern Cameroon and in the Cross River Loop in Eastern Nigeria. The distribution of its vicariant, N. gola , is restricted to the Liberian sub-region (Western Ivory Coast, Liberia and Eastern-Central Sierra Leone) ( Fig. 2 ), which is known to host a significant number of endemic species e.g. Pseudopontia gola , Aphnaeus nimbaensis , Euriphene lomaensis , E. leonis , E. taigola , Euphaedra aubergeri ( Larsen 2005, Larsen et al. 2009, Sáfián et al. 2013), all with allopatric biogeographic vicariants in Central or Eastern Africa. This indicates that the Liberian sub-region was once ecologically similar to the Central African wet rainforests and remained suitable for the wet forest dwelling species, even during the historically drier periods, when savannah covered most of West Africa. These wet forest-dwelling species must have been isolated from their main distribution area for a long period of time, which led to significant diversification from the populations occurring in the Central African core area(s) of their distributions. This strongly supports the existence of species pairs with allopatric distribution in the Guineo-Congolian forest zone ( Larsen 2005) such as N. georgiadisi and N. rhoko .

Figures

FIGURE 1. Holotype of N. rhoko (upperside A, underside C), holotype of N. georgiadisi (upperside B, underside D), both types deposited in the ABRI collection. Male of N. mahota from Cameroon (upperside E, underside G) (in coll: ABRI), and male of N. gola from Sierra Leone (upperside F, underside H) (in coll: SZS).

FIGURE 2. Male genitalia of N. rhoko (A), holotype, N. georgiadisi (B), holotype, N. gola (C) Gola Rainforest National Park, Sierra Leone and N. mahota (D) re-drawn from Stempffer’s original illustration (1944).