376 
ELYMNIAS. By H. Fruhstorfer. 
differs greatly, the $ only slightly from the general type, hence, among the hypermnestra forms the $ is the 
most conservative, whereas in other species, e. g. agondas, the resist, the $$ incline to variation. Larva 
similar to that of undularis Drury, but head more yellowish brown. Pupa green with yellow longitudinal 
stripes, which are bordered by red strokes and spots, head and thorax with three short tubercles. Ceylon, 
hyper- nowhere uncommon. - — hypermnestra L. (= leucocyma Godt. [ ?] (87 a) comes much nearer to the North Indian 
mnestra. undularis than to tinctoria Moore or violetta Fruhst. like the jraterna Btlr. <$<$. Some Java BS are i n fact iden¬ 
tical with undularis and most of them differ only in the rather darker, violet rather than blue submarginal 
markings of the forewings. The $ of hypermnestra is as a rule rather smaller than undularis and the more 
or less darkened forewings together with the diminution of the white markings on the hindwing stamp it 
as a melanic Island race. The apically broader, oblique white bar on the forewings has the periphery more 
richly glossed and darker blue than in undularis $, yet I do not possess a single £ in which this gloss is so bril¬ 
liant as in Cramers otherwise excellent figure, which is perhaps somewhat fancifully equipped in this respect. 
East Javanese are characterized on the upper surface by brighter red-brown and paler interspaces in the 
forewing and especially strongly white spotting on the hindwing; on the underside by the richer white decora¬ 
tion of the apical and submarginal tracts of the hindwings. BS of the east-java dry season form approach 
the balinese race (baliensis Fruhst.) in their dwarfed size, the light violet gloss of the forewings and the pale 
reddish external border of the hindwings. The entirely didl under surface is divided into a somewhat dar- 
perpusUla. ker basal, and a faded brown-grey distal half. Such BS I have designated as B fa. perpusilla Fruhst. — In 
Java hypermnestra extends from the sea coast up to about 3000 feet, and is one of those butterflies which the 
native collectors first supply in large numbers. Wherever we find, in or near the villages, gardens sha¬ 
ded by groves of bananes or other fruit trees, there at all seasons we meet with the sluggish hypermnestra, 
which always sit in the thickets, dull and tired, and, even when roused, only fly a few meters forward, when 
they at once assume the attitude of rest, close their wings and turn to the observer their finely striate 
under surface. The $$ are again and again mistaken for Danais intensa Moore, which is common in the 
same localities, but whereas the Danaidae are always searching for blossoms, the Elymnias never go upon 
baliensis. flowers. West and East Java; Basean, Ivangean (Snellen). — baliensis Fruhst. According to the statements 
of Elwes and Niceville, it appears that two species of Elymnias of the hypermnestra group occur on Bali. 
This view is founded upon a great error, for Bali only yields hypermnestra baliensis Fruhst., the sister form 
of the Javanese protogenia Cr. (recte hypermnestra hypermnestra L.). baliensis comes very near the fa. per¬ 
pusilla Fruhst., but is larger, has more strongly dentate wings and a darker type of colouring. (Satellite Is¬ 
land character). The violet subapical marks of the forewings seem to be considerably broader than in the 
Java BB> brighter violet, and the other lower spots are also more prominent. Distal border of the hindwings 
cacao-coloured, under surface darker, more black in place of red-brown. Bali, October. Elwes and Niceville 
state that baliensis does not differ from the javan protogenia, an observation on which one cannot place much 
dependence, since both authors (Journ. As. Soc. Beng. 1886) were also unable to detect any difference between 
undularis Drury and tinctoris Moore, but compare Moores figures of the these unusually pronounced races in 
“Lepid. Indica”. 
caudata. E. caudata Btlr. (87 a, b), a bramch of the undularis seriesj-which has departed furthest from the type 
of the species, with entirely different coloration larger size and longer tails on the hindwing. The undularis 
colouring is almost entirely retained especially in the with its extended white instead of blue spots on the 
forewing gives a totally different impression. Larvae on the Areca, Phoenix and Coco palms, also on Caryota 
urens L. and Calamus pseudotenuis (which also belong to the Palmaceae). Larva with longer processes on 
the head, shorter and thicker than in the undularis larvae. Throughout South India, in Bamboo forests. 
cottonis. E. cottonis has a very circumscribed habitat, only Burma and the Andamans. Two geographical 
races: cottonis Heu\ (87 b) probably a mimic of Euploea simulatrix, not uncommon at Port Blair in the South 
Andamans from April to August. Dry form (87 b) small, with almost orange distal margins to all wings. Wet 
obnubila. form with dark red terminal margin to all wings. — obnubila Marsh. To Bingham belongs the credit of ha¬ 
ving not only discoveder obnubila, but also of having assigned to its the correct systematic position, for Bing¬ 
hams “race” which he applies to obnubila, is always identical with the conception of subspecies. It is almost 
certain that cottonis represents undularis on the Andamans, while occuring together with it in the Mergui Archi¬ 
pelago. The under surface is very similar to that of undularis, but the existence of obnubila side by side 
with undularis tinctoria Moore in the Mergui Archipelago supports the specific rank of the former, obnubila 
belongs to the few recent species of butterflies which are common to Burma and the Andamans, and their si¬ 
multaneous presence must be traced back to a land connection in the Tertiary Period, from which also proba¬ 
bly dates the broadly black framed Hestia common to Burma and the Andamans. Central and South Tenas- 
serim, Mergui Archipelago, Salanga Island. 
nigrescens. E. nigrescens Btlr. the commonest, most widely distributed, and at the same time most inconspicuous 
species of the Elymnias. Common as it is, the larva is still unknown, though Martin (Iris 1895, p. 251) assumes 
