Danaida chrysippus

Smith, David A. S., Lushai, Gugs & Allen, John A., 2005, A classification of Danaus butterflies (Lepidoptera: Nymphalidae) based upon data from morphology and DNA, Zoological Journal of the Linnean Society 144 (2), pp. 191-212 : 203-204

OF D. CHRYSIPPUS

Schreber (1759) described (as Papilio aegyptius ) a butterfly from ‘Aegypto’ – Egypt – closely similar to ssp. chrysippus . A more recent treatment of D. chrysippus in the Indo-Australian region ( Morishita, 1985) supported Talbot’s arrangement of subspecies (distribution map in Lushai et al., 2005a). Ever since Linnaeus (1758) and Schreber, respectively, described them, the Asian chrysippus and African aegyptius races have been treated, either as species (e.g. Moore, 1883) or, more recently, as vicariant subspecies of D. chrysippus s.l. (e.g. Talbot, 1943).

The prevalent view of D. chrysippus in Africa ( Owen & Chanter, 1968; Rothschild et al., 1975; Smith, 1975, 1980) has followed Talbot (1943) in recognizing aegyptius as the only African subspecies, with three main ‘forms’, aegyptius s.s., alcippus (Cramer) and dorippus (Klug) , and several ‘minor forms’, including liboria (Hulstaert). The crucial distinction between polytypic and polymorphic variation has been widely ignored (Smith et al., 1997). Other ‘minor’ African forms, merely listed by Talbot (1943), include (extra-limital) alcippoides (Moore) , transiens (Suffert), klugii (Butler), albinus (Lanz) and semialbinus (Strand), all of which are now known to be F 1 or backcross phenotypes from crosses among the major colour forms, aegyptius ¥ alcippus ( alcippoides ), aegyptius ¥ dorippus (transiens and klugii) or dorippus ¥ alcippus (albinus and semialbinus) ( Owen & Chanter, 1968; Clarke et al., 1973; Smith, 1975, 1998; Smith et al., 1998).

The failure to distinguish polytypism and polymorphism ( Talbot, 1943) has caused major confusion, since names such as aegyptius , alcippus and dorippus have been applied to both allopatric subspecies and sympatric morphs. The exceptional size of the hybrid zone in East Africa, where the ranges of semi-isolated subspecies and species overlap (Smith et al., 1997, 1998; Lushai et al., 2003a, 2005b), has frustrated the emergence of a biologically meaningful nomenclature.

Talbot (1943) describes the coloration of aegyptius as darker than that of chrysippus , but this is only generally true for butterflies from southern Africa. Whereas in southern Africa the ground colour is almost invariably nut-brown (genotype B-), the form that prevails from Kenya and Uganda northwards is tawny orange (genotype bb), as in the nominotypical Asian form. There is, however, a broad band of polymorphism in central and east Africa where both forms coexist and hybridize ( Smith et al., 1993, 1997, 1998; Lushai et al., 2003a); furthermore, examination of museum material ( Smith et al., 1998) shows that both orange and brown forms occur as rare outliers far beyond their heartlands. chrysippus (Asia) and orange aegyptius (North Africa) are probably synonymous as specimens from India (N = 4), Oman (N = 8) and Kenya (N = 4) were identical for both 12S (344 bp) and COI (676 bp) sequences (haplotype ST 1 in Lushai et al., 2005b). We have therefore relegated the name aegyptius , as applied to D. chrysippus from North Africa, to the status of a junior heterotypic synonym of chrysippus ( Lushai et al., 2005a) .

Although Talbot (1943) noted that two aegyptius - like races occur in Africa, he did not distinguish them biogeographically. The two races are differentiated by three phenotypic characters ( Talbot, 1943; Smith & Owen, 1997): (1) coloration (brown, or orange); (2) subapical band of five white spots on the forewing (spaces 4–6, 8–9), either broad and fused or narrow with some spots separated; (3) presence or absence of a submarginal white spot in the brown area of space 2 (Cu 1b) of the forewing. All these characters are under genetic control (Smith & Owen, 1997). The narrow-banded, orange form (genotype bbll), that we now describe as chrysippus , ranges from Kenya and Uganda northwards, whereas the brown, broad-banded form, with the spot in space 2 (genotype BBLL), has a southern distribution in Africa. We have named the latter form orientis ( Lushai et al., 2005a): the argument for thus naming it is convoluted.

Pennington (1994: pl. 41, fig. 1aiv) lists and illustrates f. liboria Hulstaert, 1931 as one of seven forms of D. chrysippus aegyptius that have been recorded in South Africa, but fails to indicate that it is the predominant one. However, Hulstaert’s type locality for liboria, ‘ Inde Continent, Afrique orientale’, is unhelpful to say the least! Talbot (1943) adds further confusion by applying the name liboria, not only to a ‘Malagassic race’, occurring on most islands in the Indian Ocean, but also to a form of ‘ aegyptius’ that occurs on the African continent, distribution unstated! However, Aurivillius (1909) had earlier described orientis (as a variety of Danaida chrysippus ), giving the type localities as ‘Comoren’ ( Comoro Islands), ‘Madagaskar’ ( Malagasy Republic) and ‘Aldabra’ ( Seychelles).

Having ourselves collected and/or examined large samples of D. chrysippus from the Indian Ocean islands and southern Africa, and also inspected Aurivillius’ type specimen of orientis in the Natur Historiska Riksmuseet, Stockholm (NHRS), we believe that liboria and orientis are heterotypic synonyms, the latter name having priority ( Lushai et al., 2005a). Talbot himself recognized the synonymy but failed to give the due priority to orientis. Therefore, the subspecific epithet orientis should henceforth apply to all D. chrysippus populations from islands of the Indian Ocean and southern Africa. However, in contrast to southern African populations, which are invariably brown, orientis from Indian Ocean islands (N = 760 examined) are polymorphic for orange (bb) or brown (B-).

In summary, our phylogenetic analysis identifies two forms of chrysippus s.s. in Africa: chrysippus (central and northern Africa) and orientis (central and southern Africa). The two forms are characterized, respectively, by a single adenine-guanine transition at site 330 in the mitochondrial COI locus ( Lushai et al., 2005b), a character that is constant in the specimens examined (N = 22, 25, respectively, Table 1), and six base pair substitutions (1.5%) in the nuclear EF1-a gene ( Lushai et al., 2005b). There is thus no doubt that the two forms are distinct clades that are separated by distribution, colour/pattern genes, nDNA and mtDNA characters. Although there is good evidence for assortative mating at the B locus in Tanzania and Ghana ( Gordon, 1984; Smith, 1984), as the two forms interbreed readily in Uganda ( Smith et al., 1993), we classify them as subspecies.