Pilodeudorix (Pilodeudorix) cf. catori ( Bethune-Baker, 1903 )

Takano, Hitoshi, 2024, Descriptions of three new species of Lycaenidae (Lepidoptera: Papilionoidea) from Parc National de Nouabalé-Ndoki, Republic of Congo with additional notes on little-known species, Zootaxa 5496 (2), pp. 214-230 : 224-225

Pilodeudorix (Pilodeudorix) cf. catori ( Bethune-Baker, 1903)

Three female Pilodeudorix Druce, 1891 specimens with all-brown uppersides ( Fig. 35 ) pertaining to the P. catori group ( Libert 2004b) were caught at PNNN, one at Mbeli camp and two attracted to light traps at Ndoki Formation ( Fig. 39 ). Based on the similar underside colouration to the male holotype, it was presumed that these were the unknown females of P. hamidou Libert, 2004 , which at the time of description had not been identified with any certainty; Libert (2004b) had postulated that it was probably close to P. catori ( Bethune-Baker, 1903) based on three all-brown females he had studied from Cameroon. Surprisingly, the results of barcoding analyses revealed that the PNNN insects diverged from both P. catori ( Fig. 34 ) and P. hamidou (a publicly available sequence ALYML432- 13) with PWDs of 4.0% suggesting that it is either an unknown brown form of another species in this group or an undescribed species. Libert (2004b) indicated that females of another member of the species-group, P. anetia ( Hulstaert, 1924) , were probably dimorphic in northeast DRC having studied all-brown and all-blue forms from the same locality. With its wide distribution spanning Central Africa, the PNNN insects could represent a form of this taxon but sadly, the figured female specimens of P. anetia from DRC and Gabon ( Figs 36–37 ) were not sequenced and thus any firm conclusions, especially in the absence of male specimens, cannot be drawn. It is however worth noting that the appearance of the females from PNNN are constant yet distinct from P. catori and P. anetia ; the upperside is noticeably greyer especially on the hindwing, there are densely packed dark brown scales along the hindwing veins giving it definition, and the lobe and tornal region of the hindwing is a contrasting orangey-brown. The underside is most similar in appearance to the female of P. anetia from DRC, but with a warmer orangey-brown tone towards the outer margin of the forewing. Further investigations of members of this group are required and it is figured here as an insect of interest to look out for in the future.

Distribution patterns of butterflies in the Sangha River Interval

PNNN is located between two main tributaries of the Congo River, the Ubangi to the east and Sangha to the west, and at its heart is the watershed region for the Nouabalé River which flows into the former and the Ndoki River into the latter. The park also falls within a region known as the Sangha River Interval (SRI), a 400 km-wide area that saw recurrent phases of forest disturbance in the Quaternary ( Bostoen et al. 2015). During the late Holocene, major perturbation at the forest core led to fragmentation and a rapid expansion of pioneer-type vegetation that resulted in the formation of forest and fluvial refuges ( Fig. 41 ). The presence of savanna species deep within the humid forests suggests that the SRI may have acted as a corridor for woodland/savanna species linking the Sudanian savannas of the north to the Bateke Plateau in the south ( Maley et al. 2017). This peculiarity provides one potential explanation for the initial divergence and modern-day distribution of forest-dwelling sibling taxa in this region of which there are numerous examples. Furthermore, in the case of taxa found in sympatry such as N. ngoko and N. smithi sp. nov. (or parapatry), the most parsimonious inference is that they too speciated in allopatry and the current sympatry is the result of secondary contact (e.g., McBride et al. 2009).

However, despite the present-day continuous expanse of the equatorial rainforests in Central Africa, the SRI seems to be the eastern-most (e.g., Neurellipes mahota ( Grose-Smith, 1887) ; see Libert 2010) or western-most (e.g., Triclema lutzi Holland, 1920, the PNNN specimens being a new country record for this species and a westerly extension to the localities listed in Libert 2010) extent of many species’ distribution which may suggest that other factors have and are circumscribing the distribution of these species. At a finer scale, the direct comparison of the species assemblages of two localities situated within the SRI, PNOK and PNNN, is of particular interest. Although the majority of species are found at both sites, there are numerous examples of species recorded at PNOK, both on Dowsett’s (1997) list and from samples collected on a short trip to the park by Violette Dérozier (ANHRT) but absent from PNNN. Species such as Ornipholidotos dowsetti Libert, 2005 , Hewitsonia bitjeana Bethune-Baker, 1915 and Oxylides bella Aurivillius, [1899] are present at the former but (so far) not observed from the latter and conversely, vice versa for the three new species described above. Despite long series of two species of Oxylides being sampled at PNNN, O. bella was never collected and based on the distribution of this ( Libert 2004a) as well as the presence-absence of the other aforementioned species, it may be possible to speculate that the biogeographic barrier is much narrower than the SRI and that the Sangha River itself has acted as a barrier to gene flow ( Fig. 41 ). These observations may be a sampling artefact: this is after all a region that has in general been poorly-collected ( Larsen & Congdon 2011) and basing such reasoning on the distribution of Lycaenidae , a group which in the past has attracted less attention ( Larsen 1991) and can be difficult to collect, is perhaps misconceived. Moreover, the dataset is based on a relatively short sampling period—93 fieldwork days with three individuals sampling for 5–6 hours a day—and it is by no means complete; although sampling took place throughout the different seasons, it does not account for annual fluctuations in species abundance. There are however numerous examples of even relatively narrow tropical rivers acting as suture zones or barriers to gene flow for animals that have the capacity to cross such bodies of water (e.g., Hawkins et al. 2006; Allen et al. 2021) and unlike larger Nymphalids that are known to disperse across rivers (e.g., Rosser et al. 2021), some small forest-understory butterflies show species differentiation across them ( Hall & Harvey 2002). With specific reference to this region of Central Africa, there is a growing body of work where similar distributional patterns have been observed in other groups of Lepidoptera (e.g., László & Volynkin 2023; Taberer 2024; Takano & László 2024) as well as Orthoptera (e.g., Massa 2023) and although conjecture, all the while noting the caveats as outlined above, the SRI and potentially the Sangha River may be acting as a barrier for these forest-dwelling Lycaenids.

Figures

FIGURE 41. A map of rainforest refugia in Central Africa ca. 2500–2000 cal. yr BP (adapted from Maley et al. 2017). The present-day extent of the rainforest is outlined in pale green, the forest refuges in dark green, fluvial refuges in dark blue and the Sangha River Interval (SRI) between the dotted lines. The national parks (1: Nouabalé-Ndoki, 2: Odzala-Kokoua) are indicated in yellow, and the known distribution of Oxylides bella Aurivillius is overlayed in red dots.

FIGURES 34–37. Pilodeudorix (Pilodeudorix) species. Adult ♀♀, upperside (above) and underside (below). 34. P. catori (Bethune-Baker), Ivory Coast, Mont Sângbé NP (ANHRTUK 00289930).35. P. cf. catori, ROC, PNNN (ANHRTUK 00289890). 36. P. anetia (Hulstaert) Gabon, Monts de Cristal (ANHRTUK 00189583). 37. Id., DRC, Biakato (ANHRTUK 00321794).

FIGURES 38–40.Parc National de Nouabalé-Ndoki, Republic of Congo.38.Mbeli camp, typelocality of Falcuna nouabaleensis sp. nov. and Anthene sangha sp. nov. 39. Ndoki formation, type locality of Neurellipes smithi sp. nov. 40. F. nouabaleensis sp. nov., paratype ♂, taken in natura at Mondika camp. (All photographs V. Dérozier).