Butterflies (Lepidoptera: Papilionoidea) of Mount Kilimanjaro: Nymphalidae subfamilies Libytheinae, Danainae, Satyrinae and Charaxinae

ABSTRACT This paper, which presents an annotated checklist of the ‘lower Nymphalidae’ (Libytheinae, Danainae, Satyrinae, Charaxinae), is the fourth in a series on the butterfly fauna of Mount Kilimanjaro. Four genera of lower Nymphalidae (Danaus, Amauris, Bicyclus, Charaxes), with a total of 11 included species, are known or believed to occur within the main forest zone, from c. 1800 to 2800 m. Of these, only three species of Charaxes (Charaxes berkeleyi, Charaxes ansorgei, Charaxes xiphares) may be restricted locally to this primary forest. The lower slopes fauna, below 1800 m, is considerably richer, with a total of 11 genera and 41 species listed (8 species of which extend into the forest zone). Possible additional species, dubious earlier records, problems with African subspecies of Danaus chrysippus, a need for more work on certain Satyrinae, and classification of the genus Charaxes are discussed. An identification key to the subfamilies of Nymphalidae, and the 19 genera of Libytheinae, Danainae, Satyrinae, Charaxinae that occur in Tanzania, together with a key to the adults of all the species of these four subfamilies considered to occur or have occurred on Kilimanjaro, with 206 colour images, are included as online Supplementary Information.


Introduction
This paper, which offers a synopsis of the 'lower nymphalid' butterflies of Mt Kilimanjaro, is the fourth in a series intended to present a checklist for all butterflies currently known or likely to occur on the peak and its lower slopes (Liseki and Vane-Wright 2011, 2013, 2014. As indicated by Liseki and Vane-Wright (2011), the ultimate goal of this inventory programme is to facilitate monitoring the butterfly fauna of this exceptional mountain, including the potential to use these conspicuous day-flying insects as a focal group to document possible impacts of climate change in eastern Africa. Keys and photographs, provided as online Supplementary Information (SI), offer a practical means for identification of the adult butterflies.
CONTACT Steven D. Liseki sdliseki@yahoo.com Supplemental material for this article can be accessed here. Wahlberg et al. (2009) divided the Nymphalidae sensu lato into four major groups: the Libytheinae (long regarded as the sister group of all other nymphalidsbut not found to be so in all analysese.g. Wu et al. 2014; some other recent work suggests a relationship with Danainae -N. Wahlberg, pers. comm.), the Danainae (the next branch, or even the first - Wu et al. 2014), and two highly diverse sister taxa: the satyrine clade (to include among other higher groups Satyrinae and Charaxinae) and the nymphaline clade (to include Apaturinae, Limenitidinae, Heliconiinae and Nymphalinae, among other subfamilies). The monophyly of these two major groups, separate from the Libytheinae and Danainae, could receive additional support if the suggestion that their wing eyespots (= border ocelli of Nijhout 1991) are controlled by a network of co-opted genes that evolved just once within the Papilionoidea c. 90 million years ago (Oliver et al. 2012, Monteiro 2015 were corroborated.

Nymphalidae
Although there would be no logical justification for regarding either of the major terminal clades as more basal than the other (Krell and Cranston 2004), for convenience we propose to refer to the first three branches (Libytheinae + Danainae + satyrine clade) as the 'lower Nymphalidae'. The fourth branch, the 'higher Nymphalidae', will be dealt with in the following papers of this series.
For Tanzania as a whole, Kielland (1990) and Congdon and Collins (1998) recognized just over 180 species of lower Nymphalidae from Tanzania, as defined above, divided here among 19 genera ( Table 1). Many of these species and several of the genera do not occur in the northeastern region of the country, where Kilimanjaro is situated. Since 1998 there have been further changes and additions, but these are only noted where they affect the Kilimanjaro fauna.
Neocoenyra parallelopupillata (Karsch, 1897) is listed below (and illustrated) as a result of what has since proven to be an erroneous Kilimanjaro label added during curation. The specimen concerned was collected by SDL in the Usambaras. Normally we would have eliminated this 'record' altogether, especially if N. parallelopupillata is truly endemic to the Usambaras, as usually supposed. However, our investigations suggest the possibility that this species may be more widespread, perhaps extending south to Malawi and north to Kenya. Consequently we have retained N. parallelopupillata in our treatment, to draw attention to the need for further workbut there is no empirical evidence that it occurs on Kilimanjaro.

Conventions
In most cases male and female forewing lengths are given for each species. The means and standard deviations (SD) have been calculated from small samples of BMNH (and some OUMNH) specimens, as in Liseki and Vane-Wright (2011). The ranges are estimates based on these measurements. Wherever possible we have used material entirely from northeastern Tanzania. Where necessary, however, BMNH material from elsewhere in Tanzania, Kenya or East Africa has been used. All sizes given should be regarded as indicative only (see discussion in Liseki andVane-Wright 2011, p. 2392). For most Charaxes we have also noted the forewing length ranges given by van Someren in his 'revisional notes' (van Someren 1963(van Someren -1975. Altitudinal ranges are given in metres as rounded estimates, largely by reference to the work of Kielland (1990), but sometimes modified from other sources. In the SI figure legends and at certain places in the text, however, heights at which individual specimens were stated to have been captured are given in the units used on the data labelsnotably 'feet' (conversion to metres gives a spurious impression of accuracy, or rounding difficulties). Place names are generally standardized to modern spellings (e.g. Moshi for 'Moschi'; Engare Nairobi for 'Ngare Nairobi').

Records
According to Kielland (1990, p. 91), this butterfly occurs in forested areas and forest margins at elevations up to 2000 m, in the 'Eastern and northern part' of Tanzania. Although a very distinctive species, L. laius has a very variable underside pattern. Cordeiro (1990, p. 33) recorded it from Lake Manyara National Park. Included here as a member of the Kilimanjaro lower slopes fauna on the basis of a single male from 'Slopes of Kilimanjaro' collected by Hannington, and a ?male from New Moshi, collected by Selous, v.1916 (BMNH; both specimens listed by Kawahara 2013, p. 30). Libythea butterflies sometimes migrate or disperse in huge numbers and, although not encountered by Liseki (2009), it is possible this species occurs from time to time in the lowest zone of the montane forest. More widely, L. laius occurs in eastern and southern Africa, from Sudan to Malawi, Zambia (Heath et al. 2002), Mozambique and South Africa, including eastern Kenya (Kawahara 2013, p. 30). The very closely related L. labdaca Westwood, 1851, has an essentially parapatric distribution, being found throughout much of the rest of Africa, from Ethiopia through western Kenya and western Tanzania to the Congo Basin, Angola and West Africa, including São Tomé and Principe. The claim of Ackery et al. (1995, p. 472, as L. labdaca laius) that L. laius also occurs west as far as Angola may be in errorbut, given Zambia and Angola share a border, this needs to be checked.
Subfamily Note: see Appendix 2 for an account of recent debate regarding the taxonomy of African Danaus.

Records
In most parts of the country, from sea level to high mountains, flying all year (Kielland 1990, p. 73). Godman (1885, p. 537, as Danais dorippus (Klug, 1845)) recorded this butterfly from Kilimanjaro up to c. 1500 m, but Aurivillius (1910a, p. 2, as Danaida dorippus) gave elevations up to 3200 m. Kielland (1990) did not give specific records but the continuing presence of this butterfly on the mountain was confirmed by Liseki (2009). Beyond Tanzania this subspecies occurs throughout the Afrotropical Region, much of Arabia and Asia, including Asia Minor and the Levant, and even breeds occasionally in parts of southern Europe (notably Italy and Greece) (Ackery and Vane-Wright 1984;Ackery et al. 1995, p. 268, as Danaus c. aegyptius (Schreber, 1759). In many parts of Africa, including Tanzania, D. chrysippus exhibits unimodal wing-pattern polymorphism (terminology of Vane-Wright 1975). In the Kilimanjaro area most individuals appear to be of the 'dorippus' phenotype, but forms 'transiens', 'semialbinus' and 'albinus' also occurindicating that the population is polymorphic at the loci controlling both hindwing coloration (the A-locus, with white recessive) and forewing pattern (the C-locus, with pre-apical forewing bar recessive). In OUMNH there is a single male f. 'chrysippus' from North Kilimanjaro (Kenya), collected 19 June 1905, ex Brodie Collectionwhich can be presumed to have the genotype AAcc (see summary table in Ackery and Vane-Wright 1984, p. 95)however, the nominate form appears genuinely rare on Kilimanjaro. In Asia, form 'dorippus' is found as far east as Sri Lanka. In addition to colour pattern polymorphism, both sexes (and all forms) of this species vary very greatly in size.

Records
Highland forest, forest edges and forest mosaic up to 2300 m in north and northeastern Tanzania, south to Mufindi, and inland to the Rubeho Mountains, with disjunct populations to the west in Mpanda and Kigoma; also in lowland forest at Kasoge and Sanje, and common in semi-evergreen bush near Lufusi River, in the Rubehos (Kielland 1990, p. 73). Although not specifically recorded from Kilimanjaro by Kielland (1990), and not found there by Liseki (2009), it is included here as a member of the lower slopes fauna based on Rogers' (1908, p. 494) Heath, September 1981(Heath et al. 2002. The collective species comprises in total four recognized subspecies, ranging from Nigeria and Gabon east to Ethiopia and Somalia, south to Zambia (Ackery et al. 1995, p. 270). Tirumala formosa is very distinct from all other members of the genus (Ackery and Vane-Wright 1984), including the only other African species included. Records Kielland (1990, p. 73

Records
According to Kielland (1990, p. 75), occurs in eastern, northern and southwestern areas of Tanzania, from sea level to over 2300 m. Carcasson (1963, p. 23 Rogers (1908, p. 499) evidently found it abundantly. Not encountered in the forest zone by Liseki (2009). We regard this as a member of the lower slopes faunabut it would appear to have the potential to enter the lower levels of the forest. Subspecies dominicanus occurs widely in eastern Africa, from Kenya south to Natal, whereas the nominate race occurs from western Kenya west to Guinea. A third subspecies is found in northern Uganda, southern Sudan and Ethiopia. According to Ackery et al. (1995, p. 271;based on Talbot 1940, p. 320), it is possible that a fourth, unnamed race occurs on Bioko (Fernando Po)however, although the species is confirmed for the island by Spearman et al. (2000, p. 457), and there do appear to be minor differences in phenotype (photographs made available by J.D. Weintraub), these authors make no comment on status.
Amauris ( Talbot (1940, p. 320, 321) included 11 named forms under A. tartarea tartarea Mabille, 1876. Unpublished research suggests the possibility that some of these may represent separate speciesand this may also apply to the two other taxa currently regarded as separate subspecies: A. t. damoclides and A. t. tukuyuensis Kielland, 1990.

Records
Amauris t. damoclides was said by Talbot (1940, p. 321) to come from 'Tanganyika Territory and south-east Kenya', implying its likely presence in the Northern Highlands of Tanzania. Kielland (1990, p. 75), however, indicated eastern Tanzania only, at altitudes up to 2000 m, from the Ulugurus to Rubeho and Mufindiapparently unaware of Carcasson's (1963, p. 24) records from Moshi and Himo. The BMNH has a number of specimens from Moshi and Taveta. The presence of damoclides in the Moshi area was confirmed by Cordeiro (1995, p. 195), who noted it as common in the Rau and Kahe forest reserves. Included here as a member of the lower slopes fauna. Beyond Tanzania, A. t. damoclides also occurs in Malawi and coastal areas of southern Kenya. Southwestern Tanzania is inhabited by an endemic subspecies (subsp. tukuyensis), whereas the highly polytypic nominate race occurs from southern Sudan to Guinea south through western Tanzania and DRC to Namibia (Ackery et al. 1995, p.272).

Records
Highland forests and coffee plantations at 1400-2600 m, on Mts Kilimanjaro, Meru, Lolkisale, Kwaraha, Longido, the Oldeani-Ngorongoro highlands and Mbulu forests (Kielland 1990, p. 74), Arusha and Moshi (Carcasson 1963, p. 26). Talbot (1940, p. 330) listed numerous specimens collected on Kilimanjaro by Cooper, but all at or below 5000 ft (c. 1500 m), with one male from Moshi, and form 'luxurians' from Doringo Erok, northwest of the mountain, on the Kenya border. There are two males from the 'slopes', ex Rogers in OUMNH (who evidently considered it a common species up to about 1500 m - Rogers 1908, p. 493, 511). Given these data, and the fact that the species was not encountered by Liseki (2009), who worked in the protected forest from 2000 to 3000 m, we list this butterfly as part of the lower slopes faunabut it must surely have the capacity to enter at least the lower levels of the protected forest. Amauris e. meruensis is endemic to Tanzania. Amauris echeria, which includes 18 recognized subspecies, ranges widely across forested biotopes in Africa, from Bioko to Ethiopia and south to South Africa (Ackery et al. 1995, p. 274).

Records
According to Kielland (1990, p. 74), A. ellioti Butler, 1895, is represented in Tanzania by a single subspecies, A. e. junia Le Cerf, 1920, which flies in montane forests at 1100-2400 m in the northern highlands (including Kilimanjaro), and in eastern, southern and southwestern areas, including the Uluguru Mountains south to Songea, and thence west to Mt Rungwe and Tukuyu. Aurivillius (1910a, p. 2) recorded a pair of this species (as A. ansorgei Sharpe, now restricted as a race of ellioti from parts of Kenya and Uganda) from Kilimanjaro, Kibong'oto, 2000 m. However, Liseki (2009) did not encounter this species on the mountain and, apart from a male collected by Cooper at Moshi, 2500 ft, noted by Talbot (1940, p. 335), there does not appear to be any other Kilimanjaro material of this species in the BMNH. No material was found in OUMNH. Given that Talbot saw a distinction between the two subspecies (altumi with submarginal spots on hindwing, junia without), there is some uncertainty regarding the subspecific name to apply to the Kilimanjaro population of A. ellioti. Provisionally, we have followed Talbot (and de Jong and Congdon 1993;and Ackery et al. 1995) rather than Kielland. So we consider that A. e. altumi flies from Kilimanjaro northwards to Mt Kenya, and thence west to Uganda east of the Rift Valley (Talbot 1940, p. 335). Amauris ellioti has four recognized subspecies, all of which are restricted to highlands in eastern Africa (Ackery et al. 1995, p. 274). Records Kielland (1990, p. 75) considered the nominate race of A. ochlea in Tanzania to be coastal, occurring inland up to 1000 m in the Uluguru and Nguru Mountains, and on the Udzungwa escarpment. Although we have not found any Kilimanjaro material in BMNH or OUMNH, and Cordeiro (1990) did not encounter this butterfly in Lake Manyara National Park, with several specimens in BMNH from the East Usambaras, we consider this distinctive species a possible member of the lower slopes fauna. The nominate race is east African, found from Kenya south to Natal. The four other recognized subspecies occupy western Tanzania, the Horn of Africa and the Comoro Islands (Ackery et al. 1995, p. 276).
Subfamily SATYRINAE Genus Gnophodes Gnophodes betsimena diversa (Butler,  Records Kielland (1990, p. 77) stated that this subspecies occurs in the north, east and south of Tanzania, at altitudes up to 1600 m and 'occasionally to 2000 m'. Not found at 2000 m or above by Liseki (2009), but included here as a member of the lower slopes fauna on the basis of three male and three female specimens in BMNH from Moshi, New Moshi and Engare-Nairobi (where it was collected at 4000-5000 ft), and a single female from Taveta, c. 2500 ft, in OUMNH. More widely subspecies diversa extends from northern Kenya (Marsabit) south to the Cape, with the species as a whole found throughout the Afrotropics, including Madagascar (Smiles 1973;Ackery et al. 1995, p. 280).

Records
The whole of Tanzania, in forest and savanna, including semi-arid regions, from sea level to 2000 m (Kielland 1990, p. 77, as M. leda africana). Although this species varies considerably not only in coloration (most notably the underside) but also in forewing shape, it is unmistakeable in the local fauna. Not encountered on Kilimanjaro at 2000 m or above by Liseki (2009). On the basis of two specimens from Moshi, and a specimen collected '6 miles NW of Moshi . . . May 1916' in BMNH, and one male and three females from Taveta, c. 2500 ft (OUMNH), we include this butterfly as a member of the lower slopes fauna. However, it is possible that it does occur within the lowest levels of the protected forest, its crepuscular habits often making it difficult to detect. Melanitis leda helena occurs in suitable areas throughout the whole of the Afrotropical region; other subspecies of the Evening Brown occur in Asia and the Indo-Australian tropics.
Genus Bicyclus Bicyclus anynana anynana (Butler, 1879) Larsen 1996: pl. 29, fig. 419  Records Kielland (1990, p. 79) states that this butterfly is common in woodlands and forests from sea level up to 2000 m in all parts of Tanzania. Recorded by Cordeiro (1990, p. 29) from Lake Manyara National Park, where it was 'very common'. Included here as a member of the lower slopes fauna on the basis of one male labelled 'Kilimanjaro', without further data or provenance, ex Rothschild Collection (BMNH), and Condamin's (1973, p. 295, fig. 384) distribution map, which has a spot centred on southern Kilimanjaro. The nominate subspecies occurs in eastern Africa south from Kenya to the Transvaal and Natal, and the Comoro Islands. There are two further subspecies recognized, one from Uganda to northern Angola, the second on Socotra (Condamin 1973;Ackery et al. 1995, p. 288). Note: in line with general lepidopterological practice, the original orthographies campina and ocelligera are maintained here rather than altered to 'agree' with the supposed masculine gender of Bicyclus. As noted by Condamin (1973, p. 144 (Aurivillius, 1901), ranges from southern Tanzania into the DRC, Zambia, Malawi, Zimbabwe and Mozambique (Condamin 1973, p. 146;Ackery et al. 1995, p. 288). Note: the name of this species should be conserved (Larsen and Vane-Wright 2012).

Records
The most widespread and common Tanzanian Bicyclus, recorded from every part of the country and in most habitats, from sea level to 2200 m (Kielland 1990, p. 81). This butterfly was not encountered by Liseki (2009), and is therefore listed as a member of the lower slopes faunabut it may well also occur in the lower zones of the protected forest. Two specimens ex Rothschild Collection (BMNH) are labelled '6 miles NW of Moshi, 8.v.1916, Buchanan'. There are numerous specimens in BMNH from Lake Manyara. The OUMNH has six males and seven females from the slopes of Kilimanjaro and Taveta. Beyond Tanzania this butterfly is found throughout almost all of Africa south of the Sahara (Aurivillius 1911b, p. 93;Condamin 1973, p. 236), with the population in Ethiopia regarded as a separate subspecies (Ackery et al. 1995, p. 292 Note: Kielland (1990, p. 83, as Henotesia perspicua) considered this to be a polytypic species, with 'a distinct race in Cameroun', named in a later publication (Kielland 1994). Only treated as monotypic by Ackery et al. (1995, as Henotesia perspicua), and listed as without representation in Cameroon, because Kielland's 1994 paper came too late for inclusion. Like many Mycaelsina, this species shows seasonal variation, notably with respect to expression of the border ocellifor which Riley (1925, as Henotesia perspicua) still offers a useful summary in relation to two closely related species, Heteropsis simonsii (Butler, 1877) and Heteropsis teratia (Karsch, 1894), both of which occur elsewhere in Tanzania.

Records
Described by Kielland (1990, p. 84) as 'very common in woodland and savanna from sea level to 2150 m . . . throughout the country in suitable habitats'. In contrast, noted as 'rare' during the dry season only at Lake Manyara National Park (Moehlman and Liseki 2003). This butterfly is included here as a member of the lower slopes fauna on the basis of 10 males in OUMNH from Taveta collected at c. 2500 ft by Rogers (see also Butler 1901, p. 23) and, in BMNH, several specimens labelled Kilimanjaro (mostly collected by F. J. Jackson), together with a single male from southeast Kilimanjaro obtained by Cooper at Marangu, 4000-5000 ft, during January 1937. Liseki (2009)

Records
According to Kielland (1990, p. 88), this butterfly occurs in arid thorn-bush, at 1400-1900 m, in northern Tanzania (with an isolated record from Mbeya). Included here as a member of the lower slopes fauna on the basis of ten male and three female specimens in OUMNH collected by Rogers in 1905 and1906 at Taveta, c. 2500 ft (see also Butler 1901, p. 23), and, in BMNH, three males from West Kilimanjaro collected by Cooper at 4500-5000 ft., two males from Taveta (ex Rogers), and a further example from Taveta (sex uncertain). Not encountered by Liseki (2009) at 2000 m or above. More widely, according to Ackery et al. (1995, p. 315), this monotypic species ranges from Somalia southwards to Uganda and the Rwanda/DRC border.

Records
Open areas at 1200-2200 m in the Northern Highlands and Singida area; possibly also at Iringa, and in the Mpanda and Kigoma districts (Kielland 1990, p. 89). Not encountered by Liseki (2009), but included here as a member of the lower slopes fauna on the basis of a single male collected by Cooper on West Kilimanjaro at 4500-5000 ft (BMNH). In addition, the BMNH has a male from Mt Meru, June-July 1938, also ex Cooper. However, there does not appear to be any Kilimanjaro or Taveta material in OUMNH. Beyond Tanzania this species occurs in Somalia, Kenya, Uganda, DRC (east) and Malawi, at altitudes up to 3000 m (Larsen 1996, p. 279). Larsen (1996, p. 280) also recorded it from Zambia, but this was not substantiated by Heath et al. (2002).
The status of this species on Kilimanjaro requires further investigation.
[Neocoenyra parallelopupillata (Karsch, 1897)] Kielland, 1990: 276 (1 fig.); d'Abrera 1997: 245 (3 figs). SI: Figure 11a- Although we now realize that, without doubt, this is a subsequent labelling error (confirmed by examination of the original collecting envelope, still attached, with the data ' Magamba, 26.v.2001Magamba, 26.v. , 2000.40 hrs'), it led us to consider the possibility that N. parallelopupillata might also occur on Kilimanjaro. This species has been reported in the past from Malawi and Kenya (Gifford 1965, p. 91; the record for 'Kenya' being credited to Elliot Pinhey). Although these records have generally been discounted (e.g. Carcasson 1981, p. 182;Ackery et al. 1995, p. 316), it is not inconceivable N. parallelopupillata has a wider range than previously thought. The BMNH has a male of what appears to be this very dark species from 'Itumba District', collected by G. Wood Note: as indicated by Kielland (1990, p. 85), who only knew of one reliable record for Tanzania based on a single male collected near Amani by T.H.E. Jackson, without dissection this butterfly is very difficult to separate from other species. Larsen (1996, p. 274), however, offered exophenotypic characters for separation, including a difference in antennal segment number between Y. asterope and Ypthima yatta Kielland, 1982, a Kenyan and Ethiopian species considered otherwise to be 'similar'. These differences were first noted by Kielland (1982, p. 108), including Y. asterope reported as having 32 antennal 'joints' [segments], but only 29 in Y. yatta.

Putative records
According to notes apparently made by the late Jan Kielland, a pair of Ypthima from Lake Manyara in BMNH (collected by Cooper), formerly identified as Y. asterope, are Ypthima antennata van Son, 1955 (SI: Figure 12a-d). Moehlman and Liseki (2003) list Y. antennata as rare during the wet season at Lake Manyara National Park. Ypthima asterope is included here as a possible member of the lower slopes fauna based on four males and two females in OUMNH, long standing over the name Y. asterope, and which do not appear to be Y. antennata, collected by Rogers at Taveta, c. 2500 ft, on various dates in 1905. Supposedly found mainly in arid bush, according to Ackery et al. (1995, p. 305), Y. asterope occurs from India through parts of the Arabian peninsula and much of Africa south of the Sahara, with a separate subspecies (or replacement species : Larsen 1996, p. 275) recognized in the southwest (western Cape, Namibia and Angola). However, Kielland (1982, p. 107) noted that he had not seen reliable specimens or records from DCR to southern Nigeria and Ghana, Zambia, Burundi, Rwanda or Malawi. Heath et al. (2002) did not include it among the 11 species they recorded for Zambia. Gifford (1965, p. 92) did include asterope for Malawi, but his material (presumed to be in RSM Edinburgh), identified long before Kielland's 1982 revision, needs to be checked. If possible, a long series of Ypthima should be sought from various areas on the lower slopes, and subject to critical examination. Based on Kielland (1982, p. 114), male forewing length c. 15-17 mm, females 15-21 mm.
Note: reliable identification of this species is difficult based on external characters alone. Larsen (1996who did not illustrate this species) discussed its separation from Y. asterope and Y. yatta (see also Kielland 1982), considering it to be more similar to the latter in lacking pale marginsbut Larsen did not illustrate Y. yatta either.

Records
Found in Tanzania in montane and semi-montane grassland at altitudes up to 2200 m (Kielland 1990, p. 86). Specific localities are given as Meto Hills, Mt Longido and West Kilimanjarowith Kielland (loc. cit.) suggesting that the Kenya/Tanzania populations 'may very well belong to a distinct race'. The type locality of Y. simplicia is generally treated simply as Ethiopia, but it is in fact either Atbara town (Sudan), or somewhere along the Atbara Riverwhich, although it rises in northwestern Ethiopia, flows for most of its length through the Sudansee Appendix 4. Ypthima simplicia was not encountered on Kilimanjaro over the range 2000-3000 m by Liseki (2009), nor have we been able to locate material of this species from the mountain in the BMNH or OUMNH. Its inclusion here, as a member of the lower slopes fauna, is based on Kielland (1982, p. 115;1990, p. 86). Beyond Tanzania, Y. simplicia is thought to occur in Sudan, Ethiopia, Somalia and Kenya (Ngong Hills) (Ackery et al. 1995).
Note: van Someren (1974, p. 428) gave forewing length for this population (both sexes) as 30-32 mm, and suggested that it is slightly larger than Charaxes p. bebra Rothschild and Jordan, 1900, from Uganda and DRCbut our figures hardly bear this out.

Records
Coastal areas inland to Usambara, Nguu and Uluguru mountains, Udzungwa Rift and Magombera Forest, at altitudes up to 1600 m, with a record by Cordeiro (pers. comm. to Kielland) from Rau Forest (Kielland 1990, p. 108). Although not mentioned by Cordeiro (1995), he confirms that he did see this distinctive species in the forest, although it 'was very hard to capture' (Cordeiro pers. comm. 29 September 2014). There does not appear to be any Kilimanjaro area material in OUMNH or BMNH, and it is therefore not certain that this population does belong to C. p. orienswhich subspecies extends outside Tanzania to coastal areas of Kenya. Collectively, the five named subspecies of C. pleione (Godart, 1824) occur widely across a central belt of Africa, from Sierra Leone east to Kenya, and south to Angola, DRC, Rwanda and Tanzania. The female illustrated (SI: Fig. 13c,d), from western Kenya, represents the similar subspecies Charaxes p. bebra Rothschild, 1900which ranges from west Kenya to Uganda and northeastern DRC (Ackery et al. 1995, p. 454-455;Larsen 1996, p. 302).
Note: this species can be regarded as an example of class 2 polymorphism, in which both sexes occur in multiple but essentially identical colour formsexcept insofar as males have only one well-developed pair of hindwing tails, whereas females always have two (Vane-Wright 1975). Intermediate colour forms also occur, with individual variation (including very darkly marked males originally named as subspecies obscuratus Suffert, 1904); it seems likely this is more a case of seasonal polyphenism rather than genetic polymorphism.
Records Kielland (1990, p. 112) indicates that this butterfly can be found in most parts of Tanzania, at altitudes up to c. 2100 m, usually in thorn-bush but also entering montane forests where its primary habitat is adjacent. Cordeiro (1990, p. 35) noted it as common in Lake Manyara National Park. Recorded [as the synonym Charaxes neanthes (Hewitson)] by Butler (1901, p. 24) (2009), so included here as a member of the lower slopes.
Charaxes zoolina occurs widely is eastern Africa, from Somalia to South Africa. Three additional subspecies have previously been recognizedof which one occurs in Madagascar, and another in Angola. The third, C. z. mafugensis Jackson, 1956, is found in montane forests in Rwanda and southwest Uganda (Henning 1989, p. 376;Ackery et al. 1995, p. 463), and Rumanyika Orogundu Game Reserve in the bordering Karagwe District of western Tanzania (Kielland 1992, p. 51). However, molecular evidence now suggests all four could be regarded as specifically distinct ).
Note: The name of this butterfly is sometimes incorrectly attributed to Poulton, 1926who originally introduced the name aubyni as a female (infrasubspecific) form of C. etheocles (Cramer, 1777). Charaxes aubyni was first made available as a species group name by van Someren and Jackson (1952, p. 272), when they recognized it as a distinct, separate species (this may be of significance in assessing the primary type material). The discal bands of the females vary from pale cream to ochreous; as noted by Henning (1989, p. 351), an extreme of the latter type was named 'female f. ochrefascia' by van Someren and Jackson, and it is arguable that the females should be considered dimorphic.
Charaxes (Eriboea) baumanni tenuis van Someren, 1971 Henning 1989: 251 (2 figs). SI: Figure 13e- (Kielland 1990, loc. cit.). Henning (1989, p. 252) notes Moshi and Arusha (probably based on van Someren 1971, p. 219  Note: The female differs from other subspecies on the forewing upperside, on which the postdiscal spotting is reduced, and on the hindwing upperside, which has the discal band broader. In females of marci, the bands on the upperside of both wings are the same colour, whereas other subspecies have the hindwing markings paler than those of the forewings (Congdon and Collins 1998, p. 52).

Records
The type locality is 'West Mt Kilimanjaro, 7000 ft.' Collected in March 1993 by S.C. Collins, the type series (three males, six females, with one additional male) is in the African Butterfly Research Institute, Nairobi. Endemic to Kilimanjaro, and known only from the type locality (Congdon and Collins 1998, p. 52), it was not encountered by Liseki (2009). The two other recognized subspecies of C. berkeleyi van Someren and Jackson, 1957, are restricted to highland forests in Kenya (Ackery et al. 1995, p. 435).
Note: the females are very different to the black males, and are also polymorphicthree named female forms are recognized in this subspecies: 'ethalion', 'rosae' and 'swynnertoni' (Henning 1989, p. 305-306). However, with individual variation in pattern and size, to some extent these intergrade.
Records Kielland (1990, p. 104) recorded subsp. kenyensis from Ukerewe Island (Lake Victoria), the Pare Mountains, the Usambaras, and Pemba, at altitudes up to 1500 m, describing it as 'uncommon and local'. Henning (1989) gives no specific records for Kilimanjaro, and there is no Kilimanjaro material of this taxon in OUMNH. However, van Someren (1975, p. 425) lists 'Taveta-Kilimanjaro area' as part of its rangeand on this basis, together with a male from New Moshi and another from Taveta in BMNH (where they have been placed, perhaps incorrectly, as the more southerly Charaxes j. argynnides Westwood), we include C. j. kenyensis as an element of the lower slopes fauna. There must be significant doubt, however, regarding the subspecies assignment (including the issue of size, as noted above). Beyond Tanzania, subsp. kenyensis is restricted to Kenya, where it occurs up to the Somalian border, with transitional forms in northwestern Kenya and eastern Uganda. Divided into eight races, C. jahlusa (Trimen, 1862) is an east African insect, found from Somalia to South Africa (Ackery et al. 1995, p. 446).
Note: see Appendix 5 regarding the name and status of this taxon. In females, the pale hindwing band varies from off white through cream to pale yellow.
Records Kielland (1990, p. 93) gave the distribution of C. tiberius tiberius in Tanzania as lowland forest, up to 1350 m, in eastern and northeastern parts of the country, from south of Ifakara to the Usambaras, with an isolated sighting for Lake Duluti, Arusha, by Arthur Rydon. Cordeiro (1995), who concluded that this butterfly must be very rare in the Kilimanjaro area, cited Smiles' (1985) record for Moshi. This appears to be based on a male in BMNH collected by Cooper, 2500 ft, January-February 1938; there is also a Cooper female from West Kilimanjaro, Engare-Nairobi, at 4500-5500 ft, collected during the same period. This has smaller forewing postdiscal white spots than the female from Amani illustrated (SI: Figure 19c). In females from Tanzania the pale hindwing 'window' varies from white to pale yellow, but is apparently never so buff as in C. tiberius meruensis (van Someren, 1936) from Kenya. There is no tiberius material from the Kilimanjaro area in OUMNH. Included here as a member of the lower slopes fauna.
Subsp. tiberius extends north into the coastal areas of Kenya. The only other race of C. (E.) tiberius is found in the vicinity of Mt Kenya (Ackery et al. 1995, p. 468).
Note: In museum specimens, the blue membrane colour of the males often fades to yellowish or straw colour.

Records
Lowland forests, up to 600 m, in coastal areas of Tanzania northwards from Ifakara to the Usambaras, and inland as far as Udzungwa, Mbulu Forest and Arushawhere it may sometimes occur at 2000 m or even higher (Kielland 1990, p. 93). Cordeiro (1990, p. 35) recorded it from Lake Manyara National Park. The BMNH has specimens from Taveta, New Moshi, Moshi and Rau, all localities at approximately 750 m. In OUMNH there are four males from Taveta collected by Rogers in 1905. Not observed in the forest by Liseki (2009). On the evidence above included here as a member of the lower slopes faunawith perhaps the capacity to enter the lowest zone of the forest. More widely C. (E.) wakefieldi occurs from coastal areas of Kenya south to South Africa, including populations on Pemba and Zanzibar (Henning 1989, p. 401;Ackery et al. 1995, p. 467).
There is almost no sexual dimorphism in colour pattern in this species, but the leaflike undersides vary considerably.

Records
Mountains of the Northern Highlands, including Oldeani, Ngorongoro, Mt Meru and Mt Longido, at elevations of 1700-2600 m (Kielland 1990, p. 94). Encountered on Kilimanjaro by Liseki (2009, p. 105) at 2000 m during January and March. Four male and 1 female specimens in BMNH were collected at altitudes from 4000-5500 ft by Cooper. In addition there are three males from Moshi district, one collected at Rau Forest on 8 September 1943, one from '6 miles NW of Moshi', obtained 13 May 1916 by Buchanan, and the third from Mwika, 4000 ft, ex Adams Collection. In OUMNH there is a single male from Kilimanjaro collected by Rogers, May 1905. Beyond Tanzania this subspecies occurs in southeastern Kenya (Mt Mbolo, Taita Hills, Chyulu Hills, Ol'Doinyo Orok : Henning 1989;Larsen 1996). Over a dozen subspecies of C. acuminatus Thurau, 1903, are recognized, distributed in highland forests from Kenya and Uganda to Zambia and Zimbabwe (Ackery et al. 1995).
Note. See Appendix 6 regarding the type material of this taxon.

Records
A common butterfly throughout most of Africa south of the Sahara, although supposedly relatively uncommon in West Africa (Ackery et al. 1995;Larsen 2005). Found in all suitable habitats in Tanzania, up to 2600 m, including Pemba (Kielland 1990, p. 98).
Although not encountered during this study, C. brutus was recorded from Kilimanjaro by Rogers (1928, p. 153, as brutus natalensis Staudinger, 1885), and is included here as a member of the lower slopes fauna. Rogers (in Butler 1901, p. 23) noted it as 'fairly common' at Taveta (four males in OUMNH). The BMNH collection has two males from West Kilimanjaro (Ngaserai, and Engare-Nairobi) collected by Cooper at altitudes between 3000 and 5000 ft, a female from the 'slopes', and two pairs from Arusha collected by A.H.B. Rydon. Beyond northeastern Tanzania, C. b. alcyone occurs only in eastern and coastal Kenya. C. b. natalensis occurs from the Cape northwards to Angola and southern and western Tanzania.

Records
The Green-veined Charaxes is a common butterfly throughout most of its vast range (Larsen 1996, p. 283). According to Kielland (1990, p. 98), found in all suitable habitats in Tanzania, up to 2600 m, including Pemba. Although van Someren (1974) and Henning (1989) gave no specific records for Tanzania, and this butterfly was not encountered by Liseki (2009), the BMNH has one male from the east side of Mt Meru collected at 5000 ft by B. Cooper, June-August 1937, two more Cooper males (West Kilimanjaro, 4500-5000 ft, December 1937 to February 1938; Engare-Nairobi, 4000-5000 ft, February-March 1937), and two females from Moshi area collected in 1942 and 1943. Cordeiro (1990, p. 34) recorded it from Lake Manyara National Park. We thus include this taxon as a member of the lower slopes fauna. Beyond Tanzania, the nominate subspecies occurs throughout almost the whole of tropical Africa; the only other recognized races are insular forms on São Tome and Socotra.
Records Kielland (1990, p. 98) states that C. c. flavifasciatus is found in northern and eastern parts of Tanzania, in woodlands, forests and coastal shrubland, at altitudes up to 2000 m, with its range said to include the 'Kilimanjaro area' (van Someren 1971, p. 188). The only certain material we have located comprises three males from Taveta collected by Rogers at c. 2500 ft, April andMay 1905 (OUMNH). Included here as a member of the lower slopes fauna. More widely this east African subspecies extends northwards into Kenya, and south as far as Natal and Transvaal. The nominate subspecies extends widely throughout central and western Africa, and to the north in Ethiopia (which may represent a separate taxon); local endemic insular races occur on Grand Comore and Pemba (Ackery et al. 1995, p. 437).
Note: the separation of C. c. nairobicus van Son, 1953, from subspecies kennethi appears tenuous. Females vary in width of the hindwing lilac band or 'window'.
Note: van Someren (1963, p. 238) and Kielland (1990) only tentatively associated material from the Northern Highlands of Tanzania with this race. Henning (1989, p. 127) indicates that true teita is endemic to eastern Kenya (Taita Hills, and the Chawia, Wandanyi, Wesu and Mbololo forests). For convenience we include the Kilimanjaro population as subsp. teita, but its status needs to be confirmed.
More widely, C. druceanus Butler, 1869, occurs as a series of some 15 named subspecies, including three others found in Tanzania, collectively covering a huge part of Africa, from Nigeria to Angola, Kenya and South Africa (Ackery et al. 1995).

Records
'Mostly arid habitats . . . Mlamba Forest in North Pare Mountains, Kilimanjaro, Mto wa Mbu (below Ngorongoro), Usambara region, Mwanza' (Kielland 1990, p. 104). Cordeiro (1990, p. 34) recorded it from Lake Manyara National Park, where it appeared to be scarce. Not encountered by Liseki (2009), and there does not appear to be any Kilimanjaro area material in OUMNH or BMNH. Although not specifically recorded by van Someren (1971) or Henning (1989) as a Kilimanjaro species, it is included here, on the basis of Kielland's general statement, as a member of the lower slopes fauna. Outside Tanzania this savanna and woodland subspecies occurs in Rwanda (east), Kenya (including kulalae van Someren 1975;treated by Henning 1989; as distinct, but synonymized by Larsen 1996, p. 287), Uganda (north), Sudan (south) and Ethiopia (south). The nominate subspecies is found in parts of Ethiopia, Somalia and Arabia, with a third subspecies in Oman (Ackery et al. 1995, p. 445), and a fourth in Yemen, Charaxes h. yemeni Turlin, 1998.  van Someren (1970, p. 232-233) gave male forewing length as 35-38 mm, female 39-42 mm.

Records
Judging by the NHM this is a rare species, at least in collectionsbut there is no suggestion of this in van Someren (1963). For C. lasti lasti, Kielland (1990, p. 105) noted 'lowland forest to 900 m . . . From the Usambaras to Pugu Hills, Morogoro and Turiani. A record from Moshi is probably this race. ' Cordeiro (1995, p. 195) indicates that the Moshi record refers to material that he collected at Rau Goundwater Forest Reserve, where it was apparently first collected by the Rev. Baker. We failed to find Kilimanjaro area material in OUMNH, but the BMNH has single females from Moshi and New Moshi, and on this basis we include subsp. lasti in the lower slopes fauna. Beyond Tanzania lasti lasti also occurs in coastal regions of southern Kenya; there are two other subspecies, both endemic to Tanzania (Kielland 1984(Kielland , 1990Ackery et al. 1995, p. 449).
Note: Populations from Mbulu and Mt Kwaraha have the median bands relatively narrow and the basal area darker; specimens from Mt Meru are paler than those from Mt Kilimanjaro (Kielland 1990, p. 108). The type material of C. p. maua is supposedly in the Iain Grahame Collection (Suffolk, UK).
Note: there are considerable differences of opinion concerning the taxonomic status of this butterfly and its relatives (see e.g. Larsen 1996, p. 286 (Henning 1989, p. 85), or a further species in this complex (Larsen 1996, p. 286). Kielland (1990, p. 104, as C. jasius saturnus) describes this butterfly as occurring in savannah and open woodland, 200-2200 m, from the whole of Tanzania where suitable habitats occur. van Someren (1963) does not give any specific records for the Kilimanjaro area. Based on three males in OUMNH from Taveta, c. 2500 ft, collected by Rogers during 1905, we include C. saturnus as a member of the lower slopes fauna. More widely, this butterfly extends north throughout Kenya, and south and west to the Cape and Namibia. If admitted as separate but belonging to C. saturnus, C. s. brunnescens Poulton, 1926, occurs from northern Angola to the Central African Republic, while C. s. pagenstecheri Poulton, 1926, flies in Ethiopia and Somalia (Henning 1989). However, as already noted, the respective status of all these taxa and other members of the C. jasius species complex remain uncertain.  van Someren (1974, p. 478) gave male forewing length as 35-45 mm, female 45-51.
Note: the underside of both sexes is dead-leaf-like, and very variable.
Records Kielland (1990, p. 110) records C. varanes vologeses from woodlands and open areas, thickets, and occasionally forests and montane forests, up to 2300 mapparently throughout Tanzania (mentioning the possibility of a separate subspecies on Pemba). There does not appear to be any Kilimanjaro area material of this common species in

Records
Occurs in forests and thickets at 300-1700 m through much of inland eastern Tanzania (i.e. other than the coastal area), westwards to Mt Meru and Mt Kilimanjaro, and south to Tukuyu (Kielland 1990, p. 110 (Ackery et al. 1995, p. 461).

Records
The type locality is 'Lower slopes of West Kilimanjaro at Maua Estate' ( van Someren 1969, p. 84). Apparently confined to montane forest on the west side of the mountain. Very few specimens are known of this rare endemic (Kielland 1990, p. 112); it was not encountered by Liseki (2009), and we know of no specimens other than the original type material (only the female of which was deposited in BMNHsee Appendix 7). In addition to kilimensis, Kielland (1990) records six other subspecies of C. xiphares (Stoll, 1781) from Tanzania. In total, Ackery et al. (1995, p. 461,462) list 22 subspecies of this east African highland forest butterfly (to which at least two more must now be added), distributed from Kenya and Uganda south to South Africa,

Discussion
Ypthima and Neocoenyra The current taxonomy of these two groups, especially Ypthima, does not appear fully satisfactory. Comprehensive work including DNA sequencing could lead to significant changes in the number of species recognized. As noted under Y. asterope above, although Ypthima species are unlikely to move up into the protected forest area of Kilimanjaro, a major effort to collect a good sample of these ringlets on the lower slopes will be necessary if any Ypthima present are to be identified with confidence.

Classification of the genus Charaxes
As noted in the introductory section on Nymphalidae, Aduse-Poku et al. (2009) divided the genus Charaxes into five subgenera, including Euxanthe, formerly treated as a distinct genus (even a monobasic tribe : Rydon 1971;Henning 1989), and Polyura, often considered in the past to be a separate, exclusively Indo-Australian generic group. Robert Smiles, however, had already suggested that 'Charaxes in its present usage is paraphyletic . . . I regard [the Indo-Australian] Polyura as monophyletic, having its sister group within Charaxes' (Smiles 1982, p. 116). Since 2009 the new, five subgenera system for Charaxes has generally been well received (although certainly not by alle.g. Toussaint et al. 2015), and we decided to follow it here. In this scheme (Aduse-Poku et al. 2009, fig . 3), the interrelationships of the five subgroups can be written parenthetically (nichetes group (Polyura (Euxanthe (Eriboea + Charaxes)))). Eriboea includes the 'green charaxes', recently separated as the nominal genus Viridixes Bouyer andVingerhoedt, 2008 (Bouyer et al. 2008, p. 2). (No green charaxes have been recorded from Kilimanjaro, although a race of C. (Eriboea) dilutus Rothschild occurs nearby, in the Usambaras: Liseki 2009.) Notably in the Aduse-Poku scheme, the sister group of Polyura sensu stricto (Indo-Australia) is the C. pleione species group (Africa), with these two together sister to the zoolina species group (Africa + Madagascar)all to be included in subgenus Polyura sensu lato. Four of the subgenera recognized by Aduse-Poku et al.
(all but the as yet unnamed subgenus required to accommodate C. nichetes Grose-Smith, 1883) have representatives on the slopes of Mt Kilimanjaro. At the species level, the C. jasius complex is clearly in need of further investigation (Larsen 1996;Aduse-Poku et al. 2009).

Subspecies of Danaus chrysippus
Braby et al. (2015) argue, in effect, that despite all the complexities of the population and ecological genetics of D. chrysippus in Africa, it is probably better to regard its different phenotypes as (largely) genetically controlled forms, and that formal subdivision of D. chrysippus chrysippus into further subspecies throughout Africa, Arabia, the Indian Ocean and its vast range over much of the Asian mainland, is unhelpfulin effect, the system is too complex to be 'reduced' to a system of (widely overlapping) 'subspecies'. This accords with the earlier view of Larsen (1996, p. 256) that 'variation in the species [D. chrysippus] is not well expressed in conventional subspecific terms', and Kielland's (1990, p. 73) tacit treatment in which he did not apply any subspecies division to the whole of D. chrysippus other than to state 'there is a distinct race in Australia' (clearly referring to what is now treated as a separate species, D. petilia). Thus, despite the great geographical variation in colour pattern form frequencies throughout Africa, including populations at or near fixation (e.g. form 'alcippus' in parts of West Africa, and form 'dorippus' in parts of the Arabian peninsula), we follow Braby et al. (2015) in placing all these populations under Danaus chrysippus chrysippus.
concerning Mycalesina. David, together with Andrew Brower and Niklas Wahlberg, also responded to our queries about nymphalid higher classification, and helped reduce ambiguity in some of our terminology. Two anonymous referees also kindly made a number of very helpful suggestions for improvement and/or clarification.
a. interposita from West Kilimanjaro, Talbot (1940, p.327) also lists A. a. hanningtoni from Kilimanjaro (as well as 'Teita')! Given the current lack of certainty regarding the location of Hills of Terta, and so the true type-locality of hanningtoni, we retain Talbot's name here, in line with current usage. [With respect to James Hannington's time in this part of Africa, it is evident that he travelled the approximately 200 km return trek from Taita to Taveta, Moshi and Kilimanjaro, commencing 12 February 1885 and ending some time (well) before 23 July in the same year (when he left Rabai on a subsequent mission)see Michael (1886). His location during April is unfortunately not clear, but plausibly was in the vicinity of Kilimanjaro, rather than the Taita Hills. Even so, some form of orthographical error by Butler or other curators at the BMNH rendering Teita as 'Terta' may be the ultimate explanation.] corrected to 1969 (the 1969 date is given correctly in two online databases: Beccaloni et al. 2015;Savela 2015). With respect to the type material, there is a significant difference between the 1969 and 1972 accounts. The former states that both the holotype male and female paratype ('allotype') were deposited in the BMNH. In the latter, however, only the female is said to be so deposited. It seems likely (cf. account of Charaxes pollux maua) the holotype was retained by Iain Grahame (Suffolk, UK).