Discussion
Tabwecala robinsoni shows features which agree with what is nowadays mostly referred to as the Ophiusini + Poaphilini clade (e.g., Homziak et al. 2019). The concept of Ophiusini, historically revised many times (Kitching 1981; Kühne and Speidel 2004; Homziak et al. 2016), has recently been split between the tribes Ophiusini and Poaphilini (Lafontaine and Schmidt 2010, 2013; Zahiri et al. 2012), the latter being revived from Euclidiini, mainly based on molecular evidence. In the analyses by Zahiri et al. (2012), confirmed also by Homziak et al. (2019), each group has been recovered as monophyletic with good support but also as sister to the other in a well-supported clade, which raises the issue whether splitting them taxonomically was strictly necessary. Further to this, there are no known autapomorphies supporting the separation of the two tribes.
Zahiri et al. (2012) mention for the Ophiusini that “Ophiusa, Thyas and Artena all have a strongly modified apex to the proboscis, with strong and enlarged spines and erectile, reversed hooks that are used in fruit-piercing or lachrymal-feeding behaviour”. Such statement, attributed to Holloway (2005), was evidently a lapsus for the Calpinae (or Calpini), the erectile reversed hooks being a well-known character exclusively occurring in Calpinae/i (Speidel et al. 1996; Kitching and Rawlins 1998; Zaspel et al. 2011). In fact, Holloway (2005: 11) noted that the three genera have only enlarged spines and that these were not erectile like those of the Pericymini, and stated that strongly sclerotised, erectile, reversed hooks occur instead in Calpini and Scoliopterygini. During the present study, it was decided to compare the structure of the proboscis of Artena (A. dotata Fabricius, 1794, examined) and Achaea (A. serva Fabricius, 1775, examined), genera attributed respectively to Ophiusini and Poaphilini (Zahiri et al. 2012). No tearing hooks were observed. In contrast, the same basic structure was detected, both taxa showing only thin, nail-like spines and other sensilla in the apical section of proboscis. The fruit-piercing behaviour does not allow the separation of the two groups either, as also noted by Zahiri et al. (2012), since it has been recorded by Bänziger (1982) both in genera of the restricted Ophiusini (e.g., Artena, Ophiusa, Thyas) and others now placed in Poaphilini (e.g., Achaea, Ophisma, Bastilla, Grammodes) (Zahiri et al. 2012; Lafontaine and Schmidt 2013). In relation to lacriphagy, Bänziger (1973) summarised all data known to date and neither recorded Ophiusini nor Poaphilini. These results are in full agreement with Bänziger’s (2021: 139–140) independent review of the issue just published.
Homziak et al. (2016), in their historical review of the classification of the Erebinae, noted the treatment of Ophiusini and Poaphilini by most recent authors but also that early authors did not separate them, and often recorded features shared or scattered across taxa that are currently separated into the two groups. They noted that the restricted concept of Ophiusini matches the section comprising Artena, Ophiusa Ochsenheimer, 1816 and Thyas Hübner, 1824 that was circumscribed by Holloway (2005) on basis of absence or strong reduction of dorsal valval coremata and absence of a powdery or waxy bloom on the pupa. Holloway had also mentioned a dorsal elliptical mark on larval A5 uniting Artena and Thyas, but its absence in other Ophiusini such as Ophiusa and Clytie Hübner, 1823 (cf. Leong 2009; Herbison-Evans and Crossley 2021; Mazzei et al. 2021) excludes this feature from the list of putative synapomorphies.
As regards the valval coremata, Artena lacks them, in Ophiusa they are generally shallow (Holloway 2005), but those of O. trapezium (Guenée, 1852) are bigger than those of a poaphiline species such as Ophisma gravata Guenée, 1852 (cf. Holloway 1979; and pers. obs.), and the ophiusine Stenopis reducta Mabille, 1880 has long tubular coremata (De Prins and De Prins 2011–2021) not dissimilar from those of some Poaphilini in the so-called Parallelia-complex like Pindara Moore, [1885] or Macaldenia Moore, [1885] (cf. Holloway and Miller 2003; Holloway 2005). Splitting the two groups on this character looks therefore untenable. More promising seemed to be the character of the bloom on pupae, widespread in the Erebinae but apparently absent in the Ophiusini, so that its loss could have served as an autapomorphy delimiting the restricted concept of this tribe. Nonetheless, good illustrations of pupae of Ophiusa disjungens (Walker, 1858) and, particularly, Thyas coronata (Fabricius, 1775) clearly show that a pupal bloom also occurs in some species of this group, the pupa of the latter being also explicitly stated to be “covered in a white waxy powder” (Herbison-Evans and Crossley 2021).
A further character which has been used to unite the modern Ophiusini is the similarity in pattern, though considering the range of patterns expressed by genera of the Ophiusini and Poaphilini the overall picture is so intermingled that it is debatable that a phylogenetic signal may be traced here. Tabwecala itself, which in leg features corresponds to Poaphilini (see description above and comments below), shows the greatest similarity in external appearance to species of Artena (Ophiusini), especially members of the A. rubida group (cf. Zilli and Hogenes 2004; Zilli and Lourens 2011). Solely upon basis of the habitus it had provisionally been marked by Gaden S. Robinson with “Closest ally Artena rubida Walk.”.
Emphasis has also been laid on the widespread acceptance of Euphorbiaceae as hostplants by members of the Poaphilini (Holloway and Miller 2003; Homziak et al. 2016). The host range of Poaphilini is wide in terms of plant families and many overlaps occur with the Ophiusini. Following a recent analysis run on the hostplant data of Robinson et al. (2001), Holloway (2019 and pers. comm.) notes that of the 15 families exploited by the Ophiusini in the Indo-Australian Region, eight are shared with the Poaphilini, while the latter feed on a range of at least 35 plant families (Phyllanthaceae being then still subsumed into the Euphorbiaceae), though the Euphorbiaceae take the lead. This feature seems therefore quite well embedded into the poaphiline lineage, especially in the Parallelia-complex studied by Holloway and Miller (2003). However, (A) the existence of several poaphilines feeding on other plant families or at least not known to feed on Euphorbiaceae, (B) the diffuse polyphagy by some species which also exploit euphorbs (Robinson et al. 2001, 2021; Holloway and Miller 2003) and (C) the varying rate of their acceptance, indicate that such habit has to be better considered as a ‘syndrome’ rather than a synapomorphy uniting the whole Poaphilini. For example, out of 35 hostplant associations recently recorded by Staude et al. (2020) for Achaea in South Africa, 34.3% were Fabaceae, 14.3% Anacardiaceae, 8.6% each Combretaceae and Rhamnaceae, 5.7% each Myrtaceae, Ochnaceae and Zygophyllaceae, and only 2.9% each (1 record) Euphorbiaceae, Malvaceae, Phyllanthaceae, Plumbaginaceae, Sapindaceae and Sapotaceae. Euphorbs are also known to enter the larval diet of many genera of Erebinae outside of the Ophiusini + Poaphilini clade, e.g. Asota Hübner, [1819], Cosmophila Boisduval, 1833, Homodes Guenée, 1852, Hypocala Guenée, 1852, Tamba Hübner, [1823], Ugia Walker, 1858 (cf. Robinson et al. 2021), thus also this feature appears to be a symplesiomorphy. There is also an indication that Ophiusa sp. may feed on Hevea brasiliensis Müll. Arg. (Euphorbiaceae) in Malaysia (Robinson et al. 2001, 2021; Holloway 2019 and pers. comm.), though such records are drawn from a time when the use of this generic concept among applied entomologists may have been different from the current one (e.g., Rao 1965), even comprising poaphiline taxa such as members of Achaea. For instance, A. janata (Linnaeus, 1758) (= melicerta Drury, [1773]), highly polyphagous and especially feeding on Euphorbiaceae, has long been recorded in the past under Ophiusa (e.g., Ridley 1904; Lefroy 1908; cf. Waterhouse 1993).
Apparently good characters to differentiate between the two groups remain those outlined by Berio (1959, 1965) for his ‘phyla’ of Anua (= Ophiusini) and Achaea (= Poaphilini), both having from one to three spines on the male forefemur (Homziak et al. 2016 record four in Poaphilini but this also seems a lapsus), but differing in the position of the androteca, on the profemur and mesotibia, respectively. For this reason, the systematic position of Tabwecala, with its typical mesotibial androteca, is here considered to be in the Poaphilini. However, even if these characters have a diagnostic role, they are unsuitable in the delimitation of either tribe, being symplesiomorphies shared with other erebine groups. A mesotibial androteca is also typical of the Catocalini and Pericymini (Berio 1959: 289; Kühne 2005: 34), and it cannot therefore be used to circumscribe the Poaphilini, while the proleg type is also found in Audea Walker, [1858] and Crypsotidia Rothschild, 1901 (Audeini) (Berio 1959; Kühne 2005). Ulotrichopus Wallengren, 1860 (Catocalini) may have both types (Kühne 2005, and pers. obs.), in some species (U. variegata Hampson, 1902, examined) even showing an unusually oriented bunch of modified setae which overlap from the trochanter onto the femur as in some Ophiusini, e.g. Ophiusa tirhaca (Cramer, [1777]) (examined), O. dilecta Walker, 1865 and Stenopis tumiditermina (Hampson, 1910) (cf. Berio 1965).
There are several other characters uniting the Ophiusini and Poaphilini. They all have membranous papillae anales without sclerotised strips following up in line with the apophyses posteriores, this being a feature of Catocalini, Audeini and Catephiini (Mitter and Silverfine 1988; Goater et al. 2003; Holloway 2005; Kühne 2005; Homziak et al. 2016), and many species of genera of the two tribes sport a dorsal peak or process on the uncus (= superuncus sensu auct. nec Klots, 1956), e.g. Artena, Ophiusa, Clytie (Ophiusini) and Achaea, Ophisma and Bastilla Swinhoe, 1918 (Poaphilini). Similarly, many members of the two groups show the coecum of phallus dorsally upcurved into a ‘foot-shape’, and a basically similar structure of the valva (cf. Berio 1965; Holloway and Miller 2003; Holloway 2005). A broader taxon sampling has therefore to be recommended in future molecular analyses to better refine the relationships within the Ophiusini + Poaphilini clade, the two groups possibly not deserving to be split taxonomically in consideration of the absence of clear autapomorphies supporting each individual group and the molecular evidence corroborating their status of adelphotaxa. In this respect, molecular analysis of taxa showing unusual combinations of characters such as Tabwecala would be a priority. Besides the features the new genus has in common with several members of both branches of the Ophiusini + Poaphilini clade, such as the corematous valvae, dorsal process of uncus and upcurved coecum, it also shows traits that are found only in one of the two ‘tribes’, such as a mesotibial androteca (Poaphilini). The subbasal rather than basal dorsal process of the uncus is instead a feature more typical of certain Ophiusini such as Clytie, with which Tabwecala also shares a remarkable similarity in the main part of the structure.
Most notable are the exclusive features of the new genus. The absence of the dorsal sclerotization of the tuba analis (scaphium) occurs sporadically in the Erebidae. Following the union with the then families of Arctiidae and Lymantriidae based on molecular evidence (cf. Mitchell et al. 2006; Zahiri et al. 2011), the importance of the scaphium as an autapomorphy characteristic of the Erebidae (cf. Fibiger and Lafontaine 2005; Lafontaine and Fibiger 2006) has been devalued. It may also be present or not within the same infrafamilial grouping. Examples are in the Toxocampinae, species of the Autophila group showing a scaphium but not so those of the Lygephila group (cf. Ronkay et al. 2014; Pekarsky et al. 2019), or the OmopteriniZale Hübner, 1818 and Matigramma Grote, 1872, scaphium present and absent, respectively (cf. Blanchard and Franclemont 1982; Franclemont 1986). However, the absence is yet unrecorded in the modern concept of Ophiusini and the Ophiusini + Poaphilini clade, which typically sport a scaphium. In species with a dorsal process of the uncus it even contributes to the characteristic ‘triple beak topping’ of the male genitalia common to many members of the clade. Possibly related to this is the development in Tabwecala of a hood-like invagination in the superior part of the tuba analis to form a large scaphial pocket that wraps the tip of uncus to likely prevent damages to the rectum. This configuration is more similar to that of the distantly related Hypeninae than to the small rugulose invagination distal to the scaphium, essentially serving the same function, that occurs in Calpinae and other Erebinae (cf. Lödl 2000; Goater et al. 2003).
Legs of Tabwecala offer a plethora of unusual features, starting from the distribution of spines on tibiae. The combination of “♂:000; ♀:010” (where 0/1 = absence/presence in pro-, meso- and metatibia, respectively), that corresponds to Berio’s (1959, 1965, 1992) grade VII of spinosity, has not been recorded by Berio in any ophiusine or poaphiline. Berio did not record in all Erebinae he studied any grade VII species with male profemural spine(s) like instead Tabwecala has. The remarkably short, teardrop-shaped male metatibiae bearing a dorsal scent pencil oriented posteriorly are unique, possibly in the whole Erebidae. Noteworthy modifications in male metatibiae to support a scent function are known in other erebids, e.g. Ascalapha Hübner, [1809] (Müller 1877; Barth 1951), but their configuration is totally different from that of Tabwecala.
In the female, the sterigma surrounding the ostium bursae that projects ventrally and posteriorly into a free lamella and the simultaneous presence of a post-ostial midventral sclerotization from intersegmental membrane A7–A8 to sternum A8, so that technically the new species shows two “lamellae postvaginales”, is also unusual but reviewing details of this area in the specialised literature is difficult because of the practice of not removing sternum A7 (lodix) during preparation of the female genitalia.
Regarding the remarkable variation in wing venation shown by Tabwecala, little can be said other than confirming that topology of veins should be always carefully checked on more specimens. Venational patterns were very much used in the past for systematic purposes, but their value has greatly been downscaled (Kristensen 2003), also in the light of remarkable intraspecific variability. This was long known to be common in the more basal lepidopteran lineages but is has also been shown to occur in the most derived families (Alberti 1954; Sotavalta 1964; Albrechts and Kaila 1997; Nath and Sevi 2009), venational patterns being subject during morphogenesis to many regulatory mechanisms that can modulate expression (cf. Shimmi et al. 2014). Nonetheless, topology of wing venation still remains much employed in systematics (e.g., Souza Moraes et al. 2021).