Observations on the biology of Afrotropical Hesperiidae (Lepidoptera). Part 9. Hesperiinae incertae sedis: Zingiberales feeders, genera of unknown biology and an overview of the Hesperiinae incertae sedis Cock, Matthew J. W. Congdon, T. Colin E. Collins, Steve C. Zootaxa 2016 4066 3 201 247    Discussion, Hesperiinae incertae sedis This paper completes our treatment of Hesperiinaeincertae sedis. We have treated this assembly of genera in five parts: dicotyledon feeders ( Cock & Congdon 2013), palm feeders ( Cock et al. 2013), grass and bamboo feeders (Cock & Congdon 2014),  Dracaenafeeders (Cock et al.2015) and Zingiberalesfeeders (this paper). The 37 genera treated and their food plant families are summarised in Table 2, leaving only 11 genera for which nothing is known of the food plants or life history.   TABLE 2. Summary of the genera of Hesperiinae incertaesedis and their food plant families.    Genus (species) Evans’ Genera Group Food plant families1 Stages documented2 Ovum Caterpillar Pupa    Acleros   Acleros   Fabaceae, Anacardiaceae, Malpighiaceae +++ +++ +++    Andronymus   Acleros   Fabaceae,  Sapindaceae +++ +++ +++    Hypoleucis   Acleros   Costaceae +++ +++    Meza( sensu lato)   Acleros  Fabaceae, Linaceae, Dichapetalaceae + +++    Meza( sensu stricto)   Acleros  Poaceae    Paronymus   Acleros  Dicotyledon ++    Gorgyra   Ampittia   Connaraceae Dichapetalaceae +++ +++ +++    Kedestes   Ampittia   Poaceae +++ +++    Prosopalpus   Ampittia   Poaceae +++    Astictopterus   Astictopterus   Poaceae +++ +++    Lepella   Astictopterus  Poaceae    Tsitana   Astictopterus  Poaceae + + +  ......continued on the next page   TABLE 2.(Continued)    Genus (species)  Acada Ankola Evans’ Genera Group  Ceratrichia Ceratrichia Food plant families1  Fabaceae Poaceae Stages documented2 Ovum Caterpillar +++ +++ Pupa +++ +++     Ceratrichia Osmodes   Ceratrichia Ceratrichia   Poaceae Marantaceae +++ +++ ++ +++ +++     Pardaleodes Parosmodes Rhabdomantis Teniorhinus   Ceratrichia Ceratrichia Ceratrichia Ceratrichia   Poaceae Combretaceae, Phyllanthaceae,  Myrtaceae Marantaceae Fabaceae +++ +++ +++ +++ +++    Xanthodisca Xanthodisca astrape3  Fresna Melphinyet   Ploetzia Ploetzia Gegenes Gegenes   Zingiberaceae Marantaceae Fabaceae, Sapindaceae  Euphorbiaceae +++ + + + +++ +++ + +++     Platylesches Erionota Artitropa   Gegenes Plastingia Ploetzia   Chrysobalanaceae Musaceae Asparagaceae +++ +++ + +++ +++ +++ +++ +++ +++    Caenides(  dacela)  Caenides(  dacena)   Chondrolepis   Ploetzia Ploetzia Ploetzia   Arecaceae Costaceae Poaceae ++ +++ +++ +++ +++ +++ +++    Gamia Gretna(  balenge)   Ploetzia Ploetzia   Asparagaceae Arecaceae +++ +++ +++ +++ +++    Gretna(  carmenetc.)  Gretna(  cylinda)   Leona Moltena   Ploetzia Ploetzia Ploetzia Ploetzia   Arecaceae Marantaceae Arecaceae, Zingiberaceae  Strelitziaceae ++ +++ +++ +++ +++ ++ +++ +++ +++ +++ +++     Monza Perrotia   Ploetzia Ploetzia   Poaceae Arecaceae, Poaceae +++ +++ +++ +++     Ploetzia Pteroteinon Semalea Zophopetes   Ploetzia Ploetzia Ploetzia Ploetzia   Arecaceae Arecaceae Zingiberaceae Arecaceae +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++  Total in photographs 13 32 32 1 Families in bold were documented in our work, others are from the literature. 2 + from literature or museum specimens; ++ our observations; +++ photograph. 3 Asnoted above, T.B. Larsen (pers. comm. 2015) intended to create a new genus for this species. Evans (1937)placed it in  Xanthodiscain  Ceratrichiagroup, but its affinities are not clear.   Ova. A priori, it seems likely that differences between ova are likely to be more conservative, than any observed amongst caterpillars or pupae, which are likely to show adaptations to avoid predation and parasitism and related to their food plants and leaf shelters (e.g. Greeney et al. 2012). However, this is not necessarily true as mortality of ova has been shown to be important in pest Lepidopterasuch as the polyphagous noctuid  Helicoverpa armigera(Hübner), where egg predators play an important role in Kenya(Van den Berg et al.1999). MJWC noted that significant numbers of  Artitropa erinnyis(Trimen)ova disappeared from a  Dracaena steudneriplant in his garden in Nairobi (Cock et al. 2015). Hence, adaptations to maximise adhesion of ova to different typesof plant surfaces, or minimise the ability of predators such as ants to remove ova are certainly possible. We have images for only 14 genera ( Figure 41), but they do seem to indicate some groups within the incertae sedis. The genera  Parosmodes,  Semaleaand  Platyleschesare united by have having wall-like ribs; further, the ova of  Parosmodesand  Semaleaare somewhat flattened compared to the relatively round ova of  Platylesches, and the dorsal surface of the  Parosmodesova is rugose. Hence, there are significant differences between these three genera as well as a striking similarity. There is a group of genera with rather similar ova: dome shaped, with fine ribs, including  Erionota,  Zophopetes,  Pteroteinon,  Leona,  Artitropaand the dicotyledon feeding genera  Aclerosand  Andronymus. The similarity of the Asian  Erionotaand the African  Zophopetesand similar genera is striking.  Gretna balengediffers in that the ribs are very fine and the egg shape is bun-shaped (an oblate spheroid). The two grass feeding genera,  Astictopterusand  Ceratrichiaare united by smooth dome-shaped ova, without apparent ribs. The bunshaped ovum of  Gorgyraspp. differs from others, although this could be linked to its placement between leaves.   Caterpillars. The caterpillars of Hesperiinae incertae sedisdo not divide as easily into groups ( Figure 42). In general, the dicotyledon feeders have stouter caterpillars, and the grass feeders tend to have more slender caterpillars. The species currently placed in  Gretnacan be grouped together because of the well-developed setae and white waxy powder on the head and body. Apart from this no clear groupings can be recognised, although  Semaleaand  Caenides dacenahave relatively narrow heads,  Parosmodesand Melphinyethave distinct scale-like setae on the head (the similarities of the leaf shelters of these two genera have already been noted in Cock & Congdon (2013)),  Acleros,  Andronymusand tube shelter building  Platylescheshave strongly contrasting pale markings on the head, and so on.  Pupae. Representative pupae of the Hesperiinae incertae sedisare shown in Figure 43. None are of the Baorini type, which is perhaps slightly surprising since other tribes of Hesperiinaedo include genera with pupae of this type( Cock & Congdon 2012). Possible groupings based on the pupae are not that obvious.  Gorgyraand  Pteroteinonspp. have a bifurcate frontal projection,  Leonahas a blunt frontal projection,  Chondrolepishas a short down-turned frontal projection,  Semaleaand to a lesser degree  Xanthodiscahave a small frontal protuberance and the other genera have none; this does not seem to be a helpful character. We have noted above that the three species of  Gretnawhich are doubtfully congeneric have quite similar caterpillars, but the pupae are very different, implying they belong in three different genera, none of which show much affinity to any other genera. The frontal projection of the pupa of  G. balengeis unique, and the tufted pupa of  G. cylindais also very distinctive, but this may reflect that it is the only Afrotropical species that we have documented to form its pupa on the food plant leaf without the construction of any shelter. Aspreviously noted the pupae and pupal shelters of  Parosmodesand Melphinyetshow affinities ( Cock & Congdon 2013), and the pupae of at least some  Platyleschesspp. are generally similar. The pupae of  Aclerosand  Andronymusshow affinities in shape and colouring as well as pupal shelter construction. Above, we have noted that the pupae of  Caenides(both the palm-feeding  C. dacelaand the Costaceae-feeding  C. dacena),  Semalea,  Xanthodiscus, and  Hypoleucisare united by the broad C-shaped rim to the T1 spiracle. The pupae of  Gorgyraand  Osmodesdo not show any obvious affinities to any other genera. The pupae of  Zophopetes,  Moltena,  Monzaand  Perrotiaare similar in shape and colour, while those of  Gamia,  Artitropa, and  Erionotaare similar but paler.  Chondrolepis,  Pteroteinonand  Leonaare also quite similar to this group, but have a frontal projection and the cremaster of  Chondrolepisis less developed and may not be functional, although this may be a reflection of a tightly rolled pupal chamber obviating the need for pupal attachment. The dicotyledon-feeding  Mezaspp. may also relate to this group, but their exposed pupation site is perhaps linked to the unusual dorsal markings. The simple pupa of  Acada, which could easily be mistaken for that of a member of the Noctuoidea, might also belong with this group. The pupae of  Pardaleodesand  Ankolaare of similar shape and both are very flimsy (Cock & Congdon 2014), but their shape suggests they may have affinities with this group. There may also be affinities with  Kedestes. In summary, based on the pupae alone, we have nine possible groups: 1  Parosmodes, Melphinyetand perhaps  Platylesches 2  Aclerosand  Andronymus 3  Caenides dacela, Semalea,  Xanthodiscus, and  Hypoleucis4  Zophopetes,  Moltena,  Monza,  Perrotia,  Gamia,  Artitropa, and  Erionotawith possible subgroups or related groups (a)  Pteroteinonand  Leona, (b)  Chondrolepis, (c) dicotyledon-feeding  Meza,(d)  Acada, (d)  Pardaleodesand  Ankolaand (e)  Kedestes 5  Gorgyra 6  Osmodes 7 Gretna carmen8 Gretna balenge9 Gretna cylinda These are not incompatible with the tentative groupings suggested for the limited observations of ova. The weak groupings suggested for caterpillars are also mostly compatible except that the similarities of the caterpillars of the three species of  Gretnasuggest that they are more closely related than do the very dissimilar pupae. We look forward to seeing if and how these groups may align with future development of the classification of Hesperiinae incertae sedisbased on the T.B. Larsen’s work on male genitalia and adult characters (mostly unpublished at present) and future expansion of the recent molecular and adult character analysis ( Warren et al. 2009). In doing so, we recognise that any classification based on the early stages (and hitherto we are only considering a morphological classification rather than a phylogeny) will not necessarily align with classifications or phylogenies based on the adults, but analyses based on both combined are likely to yield a more robust phylogeny ( Meier & Lim 2009). For example, Penz et al. (2013)showed that a phylogeny based on early stages of Brassolini (Nymphalidae)was less well resolved than one based on adult characters, but that the phylogeny based on the combined dataset was similar, but not identical to that based on adult characters only. On the other hand, caterpillars and pupae living in an exposed way (as Brassolinido) are likely to be subjected to stronger and perhaps more convergent natural selection, compared to those of Hesperiidae, which mostly live in shelters. We have already commented on the similarity of the early stages of diverse Hesperiinaethat feed partially exposed, particularly on grasses and other monocotyledons. The caterpillars mostly have green bodies with longitudinal lines, and the pupae have similar colouring and a pointed frontal spike, which we have referred to as Baorini typepupae (Cock & Congdon 2014). At the time we hesitated to suggest whether this was a conservative character or the result of convergent evolution to address a common problem; today we are more inclined to the later explanation. We have noted another example of convergent evolution that will obscure relationships between genera in this series of papers: the development of ‘counterfeit predator eyes’ on the heads of caterpillars and pupae which live in shelters and should gain some advantage by frightening small vertebrate predators ( Janzen et al. 2010). Thus natural selection will obscure phylogeny where selection acts to converge on particular models for life style reasons.