Jamides vasilia sp. n., from montane West New Britain Province, Papua New Guinea, is described and illustrated. The new species is strongly divergent from other known Jamides Hübner, 1819 in possessing a high antenna-forewing length ratio, long androconia on the hindwing upperside and a strongly convex forewing inner margin in the male. It is compared by external structures, male genitalia and mtDNA sequence data to putative related species in the cyta group of Jamides. Notes on various Jamides taxa from the Bismarck Archipelago are also provided, with J. pseudosias (Rothschild, 1915) and J. reverdini (Fruhstorfer, 1915) recorded from New Britain for the first time.

Taxonomy, Lepidoptera , Lycaenidae , Polyommatinae , new species, Bismarck Archipelago, androconia

Jamides Hübner, 1819, butterflies, commonly known as Caeruleans, belong to the subfamily Polyommatinae. The genus is distributed throughout much of the Oriental, Australian and Pacific region tropics (Rawlins et al. 2014) and comprises approximately 60 described species, with a concentration in South East Asia and New Guinea. (Hirowatari 1992) recognised 57 species and additional species have been described by Tennent and Rawlins (2012), and Takanami (1994). From approximately west to east, Kunte et al. (2015), Corbet and Pendlebury (1993), Treadaway and Schröder (2012), Rawlins et al. (2014), Parsons (1998), Tennent (2002) and Tennent (2006) identified six species for India, 16 for the Malay Peninsula, 16 for the Philippines, 21 for Maluku, 15 for Papua New Guinea, eight for the Solomon Islands and nine for the remainder of the South West Pacific, respectively.

Hirowatari (1992) proposed two groups and eight subgroups for Jamides. He noted that, together with members of the genus Nacaduba Moore, 1881, there is exceptional diversity of the male genitalia valva and their morphology is useful in placing species and making phylogenetic inferences about the group.

The new species introduced herein is exceptionally distinct from all other Jamides species. A number of other phenotypically distinct butterflies have recently been recorded from West New Britain Province (Müller 2013, 2014; Müller and Wills 2013), the type locality of the new species.

Examination of type and other relevant material was carried out in various institutions (as listed below). Adult specimens were photographed using a Nikon D300s Digital SLR Camera with a Nikon AF-S VR Micro-Nikkor 105mm f/2.8G IFED Macro lens and Nikon R1C1 Close-up Kit Flashes Speedlights. RAW images were edited using Adobe Photoshop CS6. Editing included alignment, auto contrasting and removal of background. A standardised procedure was followed with photography and image editing to ensure consistency of image output. Genitalia were extracted following maceration of abdomens in 10% KOH at room temperature for 36 hours. Genitalia were photographed in glycerol using the fore-mentioned camera body adapted to a Meiji Techno EMZ-5TR-P-FOI Trinocular Stereozoom Microscope, with OPTEK FL95E Fibreoptic Illuminator and twin arm optical fibre. Individual images were taken with the remote acquisition software DIYPhotoBits Camera Control 5.2. Sliced genitalia photographs were stacked and concatenated using the software Helicon Focus 6.0 and edited in Adobe Photoshop CS6. The plates were assembled in Adobe InDesign CS6 and the phylogeny tree in CorelDRAW X6. Genitalia were stored in small glycerol-filled vials pinned beneath the specimen.

Descriptions of external facies follow that of the numerical vein system of Corbet and Pendlebury (1993). Nomenclature of genitalic descriptions follows the same reference.

Tissue material (two legs) was collected from representatives of all of the eight Jamides subgroups of Hirowatari (1992), except that of the monotypic celebica Eliot, 1969 group, originally ascribed to Epimastidia Druce, 1891. Genomic DNA was extracted using the Qiagen DNEasy extraction kit, following the guided protocol by the manufacturer. Nucleotide sequence alignment was done by eye using Bioedit. For COI, a 654 bp fragment was amplified using Folmer et al. (1994) LCO (5'-GGTCAACAAATCATAAAGATATTGG-3') and HCO (5'-TAAACTTCAGGGTGACCAAAAAATCA-3').

Individual sequence properties were assessed using MEGA, version 4.1 (Tamura et al. 2006). Bayesian analyses of the dataset were carried out using MrBayes, version 3.0b4 (Ronquist and Huelsenbeck 2003). Three independent Bayesian runs at temperature settings in the range 0.2–0.4 were performed on the data using metropolis coupled Markov chain Monte Carlo simulations, from one to 5 million generations each, and tree sampling every 100 generations. Bayesian topology and branch posterior probabilities were computed by majority rule consensus after deleting the first 1000 000 generations (10 000 trees) as ‘burn-in’, after confirming that likelihood values had stabilized prior to the 100 000th generation.

Sequences were uploaded to GenBank (http://www.ncbi.nlm.nih.gov/genbank/) and are listed for each individual in the phylogenetic tree presented in Fig. 59. Additional sequences of Jamides taxa (also listed in the tree) were downloaded from Genbank and manually aligned with the dataset. Where possible, the species identity of the latter sequences was checked through the BOLD online database image gallery (http://www.boldsystems.org/).

Jamides vasilia exhibits significant disparity from other members of Jamides. The male wing shape, with rounded tornus and convex inner margin, is unique, as is the black, tapering terminal border. On the forewing underside of the male the inner margin is broadly white-grey, relative to the remainder of the wing, a feature apparently unique in the genus. Also in the male, the hindwing upperside bears conspicuous androconia (Fig. 53), covering a purple-brown patch. Such prominent androconia are either absent or poorly developed in related species (Figs 54–58). The underside bands of both sexes are very curved and displaced at veins, unlike other Jamides taxa. J. vasilia is exceptionally bright, the male with the upperside luminescence possibly unmatched in Indo-Pacific Lycaenidae.

The male genitalia of J. vasilia (Fig. 46) are equally distinct and are unlike any of its congeners (Figs 47–52). The exaggerated teeth-like projections of the valva (Figs 46h, 46i) are unique. The dorsal projection of the valva is characteristic of the genus Jamides and in Indo-Pacific Polyommatinae appears to be shared only with the genus Callictita Bethune-Baker, 1908.

Figures 31–45. 

Jamides adults (left side upperside and right side underside, where halved) and label data. 31 Jamides areas holotype ♂ upperside 32 J. areas holotype ♂ underside 33 J. areas holotype ♂ label data 34 J. areas allotype ♀ upperside 35 J. areas allotype ♀ underside 36 J. areas allotype ♀ label data 37 J. areas ♀ (halved) (Guadalcanal) 38 J. nitens holotype ♂ upperside 39 J. nitens holotype ♂ underside 40 J. nitens holotype ♂ label data 41 J. nitens ♂ (halved) (Telefomin) 42 J. nitens ♂ (halved) (Mianmin Range) 43 J. nitens ♀ (halved) (Telefomin) 44 J. reverdini ♂ (halved) (New Britain) 45 J. reverdini ♀ (halved) (New Britain). Scale bar = 10 mm.

Figures 46–48. 

Jamides male genitalia. (a genitalia in dorsal view with aedeagus removed, b genitalia in ventral view with aedeagus removed, c genitalia in lateral view with aedeagus removed, d aedeagus in dorsal view, e aedeagus in lateral view, f valva in lateral external view, g valva in lateral interior view, h valva in dorsal view, i valva in ventral view. 46 Jamides vasilia holotype ♂ 47 J. cyta cyta (New Britain) 48 J. cyta amphissina (Sepik, mainland NG). Scale bar = approx. 1 mm.

Figures 49–52. 

Jamides male genitalia. (a genitalia in dorsal view with aedeagus removed, b genitalia in ventral view with aedeagus removed, c genitalia in lateral view with aedeagus removed, d aedeagus in dorsal view, e aedeagus in lateral view. 49 J. pseudosias (Mianmin Range) 50 J. areas (New Georgia, Solomons) 51 J. nitens (Mianmin Range) 52 J. reverdini (New Britain). Scale bar = approx. 1 mm.

Figures 53–58. 

Jamides male hindwing apex, showing development (or absence) of androconia. 53 Jamides vasilia paratype ♂ 54 J. pseudosias (Mianmin Range) 55 J. cyta cyta (New Britain) 56 J. areas (New Georgia, Solomons) 57 J. nitens (Mianmin Range) 58 J. reverdini (New Britain). Scale bar = approx. 1 mm.

That J. vasilia was phylogenetically recovered, as a result of Bayesian Inference, within a clade of Jamides also comprising J. cyta and J. nitens is perhaps unexpected, considering that J. vasilia is phenotypically very different from either of these taxa. Indeed, there is significant genetic divergence (in COI barcode) between J. vasilia and J. cyta or J. nitens. However, a potential similarity between J. vasilia and J. cyta is that the latter has been recorded feeding on Syzigium in Australia (Braby 2000) and, based on adult female behaviour (see Remarks section), is possibility the food plant of J. vasilia. J. nitens (Figs 38–43) is one of the, otherwise, brightest Jamides taxa. Parsons (1998) noted that J. nitens exhibits a deeper blue than related species in Papua New Guinea. Although the male upperside rich sky blue appears to be constant in all specimens, both sexes exhibit some degree of variability of the amount of white suffusion bordering the striae on the underside (hence three males of J. nitens are figured, see Figs 38–42).

J. vasilia is a montane species, having been recorded from 850 to 1050m. An additional female was observed at 700m in the Whiteman Range. Of the related Jamides taxa, J. nitens is also a montane insect, Parsons (1998) recording it from 1200–1800m, although the author has recorded this species from the Mianmin Range from as low as 550m. J. cyta is typically a lowland species and is rarely recorded above about 500m, although it has been seen very occasionally up to 900m in the Whiteman Range (pers. obs.). Parsons (1998) records J. cyta from sea level to 700m. J. reverdini and J. pseudosias appear to be predominantly montane in New Britain, the former being seen by the author at over 2,000m in the Bainings Mts, East New Britain Province. However, in mainland New Guinea, both species tend to occur in the lowlands. In the Solomons, J. areas (Figs 31–37) is generally a lowland species and the species was found to be fairly common at Uepi Island, New Georgia Group, at sea level during 1986 (pers. obs.). It is unknown to what extent, if any, vertical elevation separation has played in the role of Jamides diversification in the New Guinea region.

Various phylogenetic analyses produced a range of trees for the Jamides dataset. Indeed, in several, including Neighbour-Joining, Maximum Likelihood and Maximum Parsimony, the resultant tree suggested that J. vasilia is sister to all assessed Jamides species groups and that it may represent a separate genus. Given the variable results from the different methods, COI alone is clearly not an adequate tool and additional sequencing of nuclear gene fragments would undoubtedly assist in better resolving the phylogenetic position of the new taxon.

Figure 59. 

Fifty percent majority rule consensus phylogram for selected Jamides, from a Bayesian analysis of 650 base pairs of COI gene fragment. Numbers at the nodes are the posterior probabilities of those nodes. The scale bar represents 2% genetic distance. Each terminal indicates the taxon name, locality and Genbank sequence accession number. Note the following abbreviations in parentheses for Province (Prov.), Papua New Guinea (PNG) and Australia (Aus).

Figure 60. 

The Bismarck Archipelago, Papua New Guinea, showing all main islands (excluding the Admiralty Group), provinces (upper case), main centres and localities referred to in the text. Shading represents approximate expanse of land above 1000 m elevation.

Figure 61. 

The Whiteman Range, West New Britain; type locality of Jamides vasilia.

A number of friends and colleagues assisted the author, who is indebted to all of them. John Tennent photographed relevant specimens in institutional collections and facilitated several of the author’s trips to the UK. Geoff Martin and Blanca Huertas (NHM) allowed the study of material in their care. Ullasa Kodandaramaiah, Kerala (India), sequenced certain material. Yusuke Takanami, Japan, provided photos of lectotypes held in MNHU and SMT. Jean Weiner, AM, improved the translation of Boisduval’s description of Catochrysops cyta. Zsolt Bálint (Hungarian Natural History Museum, Hungary) and Oskar Brattström (University of Cambridge, UK) reviewed the manuscript.