ik ee eh ae Oe irate - BRITISH MUSEUM (NATURAL HISTORY) INSECTS OF SAMOA AND OTHER SAMOAN TERRESTRIAL PDA : se ae ME ee PART. Il. ~ LEPIDOPTERA a FASC. 1. Pp. 1-64 BUTTERFLIES OF SAMOA AND SOME NEIGHBOURING ISLAND-GROUPS By G. H. E. HOPKINS, M.A., F.E.S. -~ WITH ONE TEXT-FIGCURE AND FOUR PLATES _ LONDON PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM SOLD AT Taz Brrvise Moszum (NatvuRAn Histogy), OrRomwsett Roap, S.W.7 AND BY B, Quaritch, Lrp,; Doutav & Co., Lrp.; Tas Oxrorp University Panss; anp WuHeEwpon & Wasuny, Lrp,, LONDON; Auso By Oniver & Boyp, EpinsuraH 1927 [Price Five Shillings Issued 9 April, 1927] INSECTS OF SAMOA AND OTHER SAMOAN TERRESTRIAL ae VIE 3 VLE: ARTHROPODA PROPOSED ARRANGEMENT :— Orthoptera and Dermaptera. . Hemiptera. Lepidoptera. Coleoptera. Hymenoptera. Diptera. Other Orders of Insects. Terrestrial Arthropoda other than Insects. The work will be published at intervals in the form of numbered fascicles. Although individual fascicles may contain contributions by more than one author, each fascicle will be so arranged as to form an integral porion of one of the Parts specified above. c INSECTS OF SAMOA ta Parr III. Fasc. 1 BUTTERFLIES OF SAMOA AND SOME NEIGHBOURING ISLAND-GROUPS By G. H. E. Hopkins, M.A., F.E.S. With 1 Figure in text, and Plates I to IV. INTRODUCTION THE area dealt with in this paper includes Samoa, Tonga, the Ellice and Tokelau Islands, and Swain’s Island. The collection on which it is mainly based was made by Dr. P. A. Buxton and myself during two years (1924 and 1925) spent in Samoa while employed in research work on Filariasis for the London School of Hygiene and Tropical Medicine. During the course of our work we visited all the islands of Western Samoa, and in addition were able to do a little col- lecting in Tutuila (American Samoa) on several occasions. I have also been greatly helped by records and specimens from Dr. J. 8. Armstrong and Dr. A. J. Brass; such records, together with all others not my own, are given below with the recorder’s name appended. Most of the Tongan records were obtained during a visit I made to the group during February and March, 1925.* I was in Togatabu tf from 14th to 26th February, in Haapai on the 13th and 27th, and in Vavau on the 12th February and from the 28th February to the 13th March; unfortunately a large number of the specimens collected by me in Tonga were destroyed on the way to England. I also owe to Dr. Armstrong a number of Tongan specimens which he collected in the same localities in * This visit, and one of about a week’s duration to American Samoa, were made possible by a grant from the Fund for Promoting the Study of Evolution, presented to the University of Oxford by Prof. J. M. Baldwin, and administered by Prof. E. B. Poulton, F.R.S. T Spelt ‘“‘Tongatabu’”’ on most Atlases. It. 1 2 INSECTS OF SAMOA. March 1926, and to Dr. T. D. A. Cockerell, of the University of Colorado, the records of a small collection made by Mrs. Cockerell in the same localities in July 1924. I was not able to visit any of the higher (volcanic) islands in the Tongan group, nor can I find in the literature any records from them. The records from the Ellice Islands and neighbouring groups were obtained by Dr. Buxton during September 1924. [ am indebted to the Bishop Museum, Honolulu, for the opportunity of examining three small but very interesting collections made by Dr. H. C. Kellers in 1917-1918, Mr. E. H. Bryan in 1924, and by Mr. A. F. Judd in 1926, for the most part in American Samoa. I have no records from a greater elevation than 2,000 feet in Savaii and Upolu, or 1,000 feet in Tutuila, but Bryan, who reached altitudes of four or five thousand feet in Savai’i on several occasions, did not see any butterflies much above 3,000 feet. Rose Atoll has only four species of plants (Pisonia grandis, Boerhaavea diffusa, Portulaca sp., and the coconut), and, as might be expected, has no butterfly fauna (Mayor, p. 74). The authorities of the British Museum and Professor Poulton have allowed me to make full use of the collections under their charge, and have helped me in many ways. For the identifications of plants I am indebted to the authorities at Kew and at the British Museum. The coloured plate I owe to the skill of Miss O. F. Tassart, and the figures of early stages to my friend Dr. V. B. Wigglesworth, who drew them from specimens in spirit with the aid of photographs from life taken by Dr. Buxton. The literature on the butterflies of the area (like that on all the other groups of insects) is unfortunately extremely scattered, and much confused by synonymy and misidentifications ; in several instances identical insects are recorded under totally different names (sometimes under two different names in the same list), while in at least one case (that of Zizera labradus Godt. and Z. alsulus H.8.) two very distinct species have, in several lists, been lumped together as being synonymous. It is hoped, therefore, that this paper will serve a useful pur- pose in bringing together our knowledge of the subject and clearing up some of the synonymic muddles ; no attempt has been made to make the synonymy exhaustive, but I have tried to ensure that it is as complete as possible for the area under consideration. It is most unfortunate that few of the Samoan butterflies in collections have more accurate indications of locality than “Samoa” or (as in the case of the Bourke collection) “Apia and Pago-Pago”’; this has entirely obscured BUTTERFLIES OF SAMOA AND SOME NEIGHBOURING ISLAND-GROUPS. 3 the fact that in several cases the butterflies of Western Samoa (Upolu and Savai’i) show very important differences from their relatives in Tutuila (of which island Pago-Pago is the capital). On the other hand, I have been fortunate enough to capture specimens of several species of butterflies previously unrecorded from Samoa, and to work out a portion of the life-histories (for the most part pre- viously unknown) of many of the species on my list. Many measurements of expanse are given in the following notes; these were all taken by measuring the distance from centre of thorax to tip of forewing and doubling the figure thus found; many of the measurements given by German authors are approximately half mine, and are evidently the expanse of one forewing only, but the method adopted herein appears to me to be more logical. The types of all new forms have been presented to the British Museum, paratypes to the Bishop Museum, Honolulu, and to the Hope Department, Oxford. NatuRAL ENEMIES AND Mimicry An attempt was made to ascertain the natural enemies of butterflies in Samoa. Ants swarm everywhere, and captive larvae were destroyed by them on several occasions both there and in Tonga, but I am not able to produce any evidence that they attack larvae in nature ; they do not appear to interfere with larvae of Badamia exclamationis, though the bushes on which these latter occur are always overrun with ants. Birds were never observed to capture butter- flies, but that they frequently do so is proved by the fact that specimens of several species (including Danaida melittula, Euploea eleutho bourker, EL. schmeltza, and Melanitis leda in Samoa, and Ewploea eleutho mathewi and Hypolimnas antilope in Tonga) were not uncommonly captured bearing unmistakable beak- marks on their wings (PI. III, figs, 1, 2, 3). Specimens with what looked like lizard-injuries were quite common (PI. III, figs. 4-7). Spiders undoubtedly take a toll of butterflies, and specimens of Danaida archippus, Danas melittula, Euploea schmelizt and Hypolimnas imconstans, as well as smaller species, were found in their webs. On one occasion a Pentatomid bug was caught in the act of sucking a larva of Atella exulans. I was not able to find any convincing evidence of mimicry in any of the islands visited : Hypolimnas errabunda somewhat resembles Ewploea schmeltz, and some forms of H. pallescens are not unlike Huploea e. mathew and EH. e. bourkez, 4 INSECTS OF SAMOA. . but these very doubtful cases seem to be the only possible instances of mimicry in the area. H. lutescens greatly resembles EF. schmeltzi, but the two species do not occur in the same islands. DISTRIBUTION With the single exception of the comparatively recent American immigrant, Danaida archippus, the butterflies inhabiting Samoa and the neighbouring eroups of islands are all Indo-Malayan in origin; most of them call for little note on the subject of distribution, being very wide-spread throughout the islands of Polynesia. The majority of them seem to have reached Samoa by way of Fiji, which is the Eastern limit of quite a number of species, but the cases of Kuploea schmeltzr, Hypolimnas errabunda, and Atella exulans are different. The first-named has no close relatives nearer than the Loyalty Isles, the other two both find their closest allies in Papua (and in the case of H. errabunda in the Solomons also), and have not been recorded in any form from the islands intervening between these localities and Samoa (a distance of about 1,200 miles). Many of these islands are not very well-known, and, since both H. errabunda and A. exulans are mountain species, it is possible that they have been over- looked, as they had been in Samoa, and that races of them will turn up in Fiji and perhaps elsewhere ; but this argument does not apply to E. schmeltzi, which is a coastal species and could not be overlooked in any locality where any col- lecting at all has been done. A possible explanation of the absence of this species from any locality between the Loyalties and Samoa is that it existed there at one time, but has been worsted in the struggle for existence by the larger and more vigorous H. eleutho, which feeds on the same food-plants and would, therefore, come into competition with it. Though both species occur in Samoa, they are confined to separate islands, and strong support to this suggestion would be found if this should prove to be the state of affairs in the Loyalties also. Unfortunately, the fact that the material available from the latter group is insufficiently supplied with data makes the test impossible of application at present. Neither H. errabunda nor E. schmeltzi is recorded in any form from localities east of Samoa, but Aftella gaberti, which appears to be fairly closely related to A. exulans, occurs in Tahiti, though apparently not in the intervening groups. The other point of special interest, in the distribution of the butterflies BUTTERFLIES OF SAMOA AND SOME NEIGHBOURING ISLAND-GROUPS. 5 of the Central Pacific, is the great difference between the fauna of Western and that of American Samoa. I have shown this in tabular form in Table I. In this table I have marked with a query those species which are rare, or which occur only at a considerable altitude, and may therefore yet be met with in the islands from which they have not hitherto been recorded. It will be noted at once that, leaving out of consideration the cases of H. errabunda and A. exulans, both of which come in this category, there are very interesting differences between Western and American Samoa, and unexpected resem- blances in several cases between the latter and Tonga. Huploea eleutho occurs in both Tonga and American Samoa but not in Western Samoa, as also does Hypolimnas antilope ; EHuploea schmeltz, on the other hand, though abundant at low levels in Western Samoa, is absent from the eastern part of the group and from Tonga. Furthermore, specimens of Hypolimnas bolina and of Precis villida from American Samoa do not appear to be separable from Tongan ones, while the same species occur in Western Samoa as quite distinct races: of the sixteen species known from American Samoa five do not occur (at least as the same race) in Western Samoa. There appears to be at least prima facie evidence that the fauna of Western Samoa has a slightly different origin from that of the other two localities; the difference is not confined to butterflies, but was observed in other orders also. Hven among the birds similar differences are to be found: the king-hunter of Western Samoa is Todirhanvphus recurvirosiris, which in both Tonga and American Samoa is replaced by races of Halcyon sacra ; the former is confined to Western Samoa, but other species of the genus are recorded as occurring in the Paumotu and Tahiti groups. It seems possible that, while the faunas of Tonga and American Samoa have been derived from Fiji, that of Western Samoa has come more directly from New Guinea. The intervening islands may give evidence for or against this theory when their insects are better known. Concerning local distribution in a particular island there is little to note. As a rule the insects of the coast-belt are not found at higher elevations, except where artificial clearings and roads have opened up the country to them. The insects of any particular elevation are, however, usually identical in all parts of the same island, as might be expected on account of the comparatively small size of the latter. INSECTS OF SAMOA. TABLE I DISTRIBUTION OF BUTTERFLIES IN SAMOA AND TONGA B. exclamationis | Species. Western Samoa. American Samoa, Tonga. D. archippus x x x D. melissa ssp. melittula ssp. tutuilae ssp. angustata E. eleutho — ssp. bourket ssp. mathewt E. schmeltza x — ot A. andromacha ssp. polynesiaca 2 ssp. polynesiaca M. leda . ssp. solandra ssp. solandra ssp. solandra D. bisaltide ae = ssp. tonganus H. errabunda x 2 q H. antilope aan ssp. lutescens ssp. lutescens H. bolina ssp. inconstans ssp. pallescens ssp. pallescens P.-villida Ssp. samoensis ssp. villida ssp. villida I. sinha . ssp. bowdenia ssp. bowdenia ssp. bowdenia A. exulans x g g P. godeffroyt xX xX — C. jacquinotiu . ssp. manaia ssp. manaia ssp. manara B. java x 2 ssp. schmeltze T. hecabe = — ssp. aprica D. epijarbas ssp. doris 2 ssp. armstrongt J. argentina x xX — J. carissima a — x J. morphordes . — — 2 C. cnejus ssp. samoa ssp. samoa ssp. Samoa C. lithargyrea . ssp. pepe ssp. pepe xX N. vitiensis SSp. samoensis ssp. samoensis ssp. samoensis Z. alsulus x — x Z. labradus xX x x T. fraserv x 2 2 Xx BUTTERFLIES OF SAMOA AND SOME NEIGHBOURING ISLAND-GROUPS. 7 SpEciIFIC DESCRIPTIONS 1. Danaida archippus (F.). Danaida archippus ; Fruhstorfer, 1910, p. 193. Danais archippus ; Butler, 1874, p. 275. Rebel, 1910, p. 416. Schmeltz, p. 177. Danais pleaippus L. ; Fraser, p. 147. Anosia plexippus ; Butler, 1883, p. 408. Danaida plecippus ; Walker, p. 187. Collenette, 1925. Anosia menppe ; Waterhouse, 1904, p. 491. Swezey, 1921, p. 601. Danais erippus ; Semper, 1905, p. 247. Walker, who gives an excellent summary of the known history of this species in the Pacific, states (p. 187) that it was first noted in Tonga in 1863, in Tutuila in 1867, and in Western Samoa not until 1869. There appear to be no previous records from the Ellice Islands, but on Nui Island in September 1924 Buxton saw a specimen feeding on Iron-tree (Pemphis acidula), while he also noticed the butterfly as well as its food-plant on Vaitupu Island; both islands are in the Ellice group. Collenette’s notes on the species tend to show that it is becoming much less common in some parts of the Pacific, probably on account of the decrease in abundance of its food-plant, Asclepias curassavica L. This may also be true in Tonga, where I found both the butterfly and its food- plant by no means common, though, so far as my investigations went, generally distributed ; Schmeltz (l.c. p. 70), records it from Niuafou also. In Samoa, however, the insect is very abundant, occurring commonly at sea-level through- out the group, and extending up into the hills wherever clearing has allowed the Asclepras to become common. | The larvae sometimes entirely defoliate the plant, and are reduced to eating the stalks and seed-capsules ; thus they may be of some slight economic use as a check on this obnoxious weed. The adult frequents the flowers of Ageratum coryzoides L. (ComposiraE), Lantana camara L., Stachytarpheta indica Vahl. (VERBENACEAE), and its own food-plant. The early stages, which may be found throughout the year, are too well- known to require description. 8 INSECTS OF SAMOA. 2 (a). Danaida (Tirumala) melissa melittula (H.S.). Danais melittula ; Herrich-Schaeffer, 1869, p. 70 (Upolu Island). Schmeltz, pp. 175-177. Butler, 1847, p. 275. Rebel, 1910, p. 415, Pl. XVIII, fig. 4. Tirumala melittula ; Waterhouse, 1904, p. 492. Moore, p. 233. | Swezey, 1921, p. 602. Danaida melissa melittula ; Fruhstorfer, 1910, p. 203. Poulton, pp. 604-606. Danais melissa ; Butler, 1870, p. 360. “Miniature edition of Tirumala hamata”’ ; Fraser, p. 147. As has been pointed out by several authors, the Samoan record of D. obscurata Butler is certainly erroneous. It is due to the fact that the type specimens from the Solomons were unfortunately labelled Upolu as well. The form is quite distinct from any found in Samoa. D.m. melittula (PI. II, fig. 3) is found in all the islands of Western Samoa, where it is always common, and frequently very abundant, from sea-level to about 800 feet, and, as a strageler, up to 2,000 feet. In American Samoa it is replaced by the closely-allied D. melissa tutuilae. It is found in greatest numbers on the flowers of Ageratum coryzoides L., but is also fond of those of Stachytar- pheta indica ; Rechinger states that it prefers the flowers of Asclepias curassavica, but I am not able to confirm this. It flies in mist or light rain, but is not on the wing until after 7 a.m., and does not become fully active even in sunshine until after 8 a.m.; it remains on the wing until dusk. Among captured specimens males outnumber females by at least five to one. There does not appear to be any difference in markings or in size between specimens from the various islands of Western Samoa; in a series of 49 males from Upolu and 43 from Savai’l, the maximum, mean and minimum expanse are 82, 72, and 62 mm. for Upolu, and 78, 72, and 62 mm. for Savai1; ignoring one abnormally large specimen from Malololelei, Upolu, the figures for the two islands are identical. Females are very slightly smaller, the corresponding figures for a series of 21 specimens from the two islands being 78, 71, and 62 mm. The race appears to be confined to Western Samoa, but the species has a wide BUTTERFLIES OF SAMOA AND SOME NEIGHBOURING ISLAND-GROUPS. 9 distribution through the Indo-Malayan region, and occurs in many races in the islands of the Pacific. The present race appears to be the smallest known form of the species. The eggs, which are laid singly on the underside of a leaf of Tylophora samoensis A. Gr. (ASCLEPIADACEAE), are creamy white and barrel-shaped, with about 14 longitudinal ribs, each of which is joined to its neighbour by about 19 narrow transverse-bars; the height is about 1°35 mm. and the diameter 0°76 mm. The head of the larva (PI. IV, fig. 1) is black, except two broad transverse bands and the labrum, which are white. The body above the spiracles is white, with one broad and two or three narrow black transverse bands on each seg- ment; below the spiracles the colour is orange-brown. Spiracles, legs and prolegs are black. There is a pair of black fleshy filaments on the mesothorax, and another pair on the eighth abdominal segment. Eges and larvae were found in June and September. The pupa (PI. IV, fig. 2) is much more compressed in shape than that of D. archippus, and of a beautiful translucent jade-green, totally different from the waxy green of that of archippus ; the girdle and a few minute dots on the dorsal surface are golden, and the cremaster and two small spots on the ventral surface black. 2 (b). Danaida (Tirumala) melissa tutuilae, ssp. n. Danars melittula ; Schmeltz, p. 175. Rebel, 1910, p. 415. Differs from melittula in both sexes owing to its much larger size and more extensive blue markings. The latter feature is best seen in the state of develop- ment of the hook-shaped marking formed by the coalescence of the blue streak along the inner margin of the forewing with the blue spot external to it, and of this latter with the spot internal to it and anterior to the streak. The extent of this coalescence varies in both the forms, and the hook may be ‘‘ Complete ”’ (Pl. II, fig. 1), “Incomplete” (Pl. IT, fig. 2), or “Absent”; I call it incom- plete when the streak is attached to the external spot, but the latter is not joined in turn to the internal one, or when the two spots are jomed to one another but not to the streak. In a series of one hundred and twelve specimens of D. m. melittula from Upolu and Savai’i, the hook is complete in seventeen males and nine females, incomplete in seventeen males and five females, and 10 INSECTS OF SAMOA. absent in fifty-eight males and six females ; in a series of forty-eight specimens of D. melissa tutuilae, all from Tutuila, it is complete in eleven males and seven females, incomplete in twelve males and four females, and absent in thirteen males and one female. That is to say, the hook, absent in 63 per cent. of males of D. m. melittula, is developed to some extent in 64 per cent. of male D. melissa tutuilae ; the difference in the females is not so great, and the numbers are rather small. The difference in size between the two races is much more striking than that of markings; the maximum, mean, and minimum expanse of thirty-five males of D. melissa tutwilae being 88, 80, and 71 mm. as against 78, 72, and 62 mm. for D. m. melittula, and of eleven females 86, 79, and 74 mm. as against 78, 71, and 62 mm. for the same sex of the latter. As will be seen from the above figures, among captured specimens, males are much commoner than females, though to a less extent than in D. m. melittula, outnumbering them by only about three to one instead of by about five to one as in the latter. The difference, although I do not think it probable, may possibly be seasonal, since most of the specimens from Tutuila were captured in August. If not seasonal, it 1s very interesting in view of the fact that the proportions of the sexes in Hypolomnas bolina show an exactly opposite difference in the two parts of Samoa, females bemg very much commoner in Western Samoa than they are in Tutuila. D. melissa tutuilae is very common at low elevations in Tutuila, and pro- bably throughout American Samoa. In habits the race exactly resembles D. m. melattula, but even in flight is readily separated by its superior size. I have no knowledge of the early stages. 2 (c). Danaida (Tirumala) melissa angustata (Moore). Twumala angustata ; Butler, 1883, p. 408. Danada melissa angustata ; Fruhstorfer, 1910, p. 203. Poulton, pp. 604-609. At the time of my visit in February and March, 1925, this appeared to be a very rare insect in Tonga, and only two specimens were seen, both at Nukualofa, Togatabu; Armstrong, in March 1926, did not meet with it, nor did Mrs. Cockerell in July 1924. It was evidently commoner during Mathew’s visit in July 1884, since he captured four specimens at Nukualofa during the single day on which he was there; the difference may possibly be seasonal. There do not appear to be any records from the other islands of the Tongan BUTTERFLIES OF SAMOA AND SOME NEIGHBOURING ISLAND-GROUPS. 11 group, and the form does not occur elsewhere. Schmeltz’s records of D. melittula and D. neptunia from Togatabu (1876, pp. 175 and 177) presumably refer to this race. Poulton suggests (p. 607) that the somewhat reduced pattern may give the form a superficial resemblance to the much commoner FE. eleutho mathewi when on the wing; the markings do not seem dissimilar in flight, but D. melissa angustata appears quite blue, while ZL. eleutho mathew looks very white. Possibly the female (which in this group of Danaines is usually of a considerably paler blue than the male) would show more resemblance to the Huploea. 3 (a). Euploea eleutho bourkei (Poulton). Huploea eleutho ; Schmeltz, p. 180. Moore, p. 272. Nipara helcita ; Moore, p. 258. Euploea eleutho escholtz ; Swezey, 1921, p. 602. Euploea helcita bourker ; Poulton, p. 585. Talbot (p. 26) has recently cleared up the confusion between EL. “ eleutho,” never found east of 150° E. longitude, and F. “ helcita,”’ which does not occur to the west of 160° E.: he has shown that the error is to be traced back to Boisduval, the following of whose mistake by various later authors accounts for the records of H. eleutho from Samoa, Tonga, the Ellice Islands, etc., in all of which it has certainly never occurred in its typical form. Talbot also pointed out that the two “ species’ were probably only geographical races of one, of which the prior name is eleutho. The main difference between the two “species ” is the presence in eleutho of a sexual brand, absent in “ helcita”’ : that Talbot’s view is correct is shown by the specimens of FH. eleutho bourker (described below) which had well-developed brands. The existence of specimens of bourkei, taken on the same day and within a few yards of each other, some with brands and some without, shows (if further proof were needed) how utterly fallacious is the use of the male brands as generic or sub-generic characters in Huploea. The forms of eleutho with brands and those without, which were formerly placed in different genera, are now shown to occur, not only in the same species, but in the same geographical race of that species! Felder’s EL. eleutho escholtzi is a Fijian race, and the Samoan record of it is erroneous. LZ. aglaina, a form with much reduced white markings, was described by Fruhstorfer (1908, p. 276, Pl. 864) as occurring in Tutuila, but, as pointed out by Poulton, this is almost certainly 12 INSECTS OF SAMOA. erroneous. If the label be correct, it must be a very rare aberration, for there are no specimens showing the least approach to it in any of the collections examined, nor in the long series collected by mein Tutuila ; on the other hand, judging from Fruhstorfer’s figure, which does not agree with his description, it almost exactly resembles specimens from the Cook Islands, which are its probable habitat. As mentioned by Poulton (p. 586), there is in the British Museum a “ Male walkerv, with the hindwing pattern of escholtzi,”’ labelled Navigators’ Islands ; this has the grey suffusion of the underside typical of bourke: and mathewr, and is probably correctly labelled, but there are no specimens showing any approach to this pattern in the very long series of other Samoan examples examined by me. Variation is not extensive, but the spots on the hindwing are slightly more developed in some specimens than in others. Females always have the ground- colour paler than males, but do not seem to exhibit better developed spots. The most interesting point about the race, however, is the occurrence of males with a sexual brand, a feature never recorded before in any race of “ helcita.” Of twenty-six males captured by Buxton and myself in Tutuila, seven have a well-developed brand (PI. III, figs 9 and 10), and two more show a trace of one, so that more than a third of the specimens have the brand developed to some extent. As might be expected in what is evidently a vestigial and obsolescent structure, the brand varies very much in size, but I have never seen a trace of it in the hundreds of male “ helcita” from Tonga and other groups in the neighbourhood of Samoa that I have examined. The maximum, mean and minimum expanse of twenty-six males are 82, 74, and 66 mm., and of fourteen females 76, 70, and 65 mm. Common in Tutuila Island, American Samoa, at Pago-Pago, Leone, and other localities at or near sea-level, frequenting the flowers of Ageratwm coryzoides and Stachytarpheta ; there are also three specimens in the collection from Tau Island, American Samoa. It is very easily distinguished from EL. schmeltz, even on the wing, by its much darker colour and larger white spots. This form does not occur in Western Samoa ; the types were stated by Poulton to be from “ Apia and Tutuila ” (EK. Bourke) and “‘ Apiaand Pago-Pago ” (G. F. Mathew), but Mathew, on referring to his note-book, found that all his Samoan specimens of the species were from Pago-Pago, and it is quite certain that Bourke’s speci- mens also came from there, since he lumped the two localities together. As represented by its several races, this is one of the most widely-distributed butterflies in the Pacific Islands, often occurring on atolls which support no BUTTERFLIES OF SAMOA AND SOME NEIGHBOURING ISLAND-GROUPS. 13 other butterfly-life except H. bolina and P. villida. Forms without a brand are found from Tahiti in the east to the New Hebrides and New Caledonia in the west, and (as stated above) are replaced still farther west by the form with a brand (typical eleutho). The larva (PI. IV, fig. 5), in addition to the two pairs of fleshy filaments borne by the caterpillars of EH. schmeltzi and D. melittula, has a pair of these processes on the metathorax and second abdominal segment respectively. In colour it is brown-olive above, with a pair of small diamond-shaped white spots on each segment; the head is black, marked with two white bands; the black spiracles are borne on a broad cream-coloured spiracular line; the legs are black, the first pair much reduced; the filaments plum-coloured. The speci- men figured is not quite full-grown. The larva feeds on Ficus tinctoria Forst., and doubtless also on other species of the same genus. 3 (b). Euploea eleutho mathewi (Poulton). Nupara eleutho ; Butler, 1883, p. 408. Euploea helcita mathewi ; Poulton, p. 586. This race, the Tongan representative of H. eleutho, was found commonly in all the localities visited in that group. It has a somewhat bolder flight than E. schmeltzi, and in the evening, about 5.30 to 6 p.m., may be met with flying low, apparently in preparation for roosting for the night. At these times the butterfly circles round close to the ground, often settling on bushes and trees, or on dead fallen leaves. Mathew records it (Poulton, p. 607) as roosting in flocks, as do many species of the genus. Apparently it is not attracted to Stachytarpheta flowers, but is fond of those of Ageratum ; I suspect that the former only attract butterflies in the absence of more attractive flowers. Like other species of Huploea it has a habit of settling on the dead twigs and branches of Tournefortia (Buxton, 1926, Hopkins, 1926), but, as in the other cases, this seems to apply only to the males, females not being found at this tree. Males are in general more often captured than females, but I am not able to give exact figures. Variation is not extensive, but there is an interesting difference between specimens from Togatabu and those from Vavau: in mathewi the sub-terminal spots of the hindwing are usually fused in pairs, and in most specimens from Togatabu this is found to be the case; in Vavau specimens, on the other hand, though this form is common, there is a definite tendency for the spots in question to be separate. It would appear that we have here the beginning of the 14 INSECTS OF SAMOA. — separation of a new race, which, however, has not yet proceeded far enough to render the forms separable. The maximum, mean, and minimum expanse of seventy-one males from both localities are 84, 77, and 72 mm., and of eleven females 82, 77, and 70 mm. This race, unlike EH. e. bourker, has the female more heavily spotted than the male. The chief food-plant is a large tree, the Tongan sacred fig, the local name for which is Ovava, but I also saw a female ovipositing on Ficus tinctoria, so that probably any species of Ficus is acceptable. The black areas of the wings evidently contain a pigment, not present in the white areas, which is distasteful to some other creatures. On the way to England a batch of Tongan butterflies, which included many specimens of the present species, was attacked and partially destroyed by Dipterous larvae ; in the case of specimens of L. eleutho mathewr, the white markings were eaten away while the black parts were left untouched (PI. III, fig. 8). 3 (c). Euploea eleutho distincta (Butler). Nipara distincta ; Butler, 1874, p. 278. Euploea eleutho ; Butler, 1878, p. 296. Moore, p. 272. Euploea helcita distincta ; Fruhstorfer, 1910, p. 235. Euploea helcita walkert ; Poulton, p. 582. We obtained this form only in the Ellice group, where Buxton captured a series of thirteen males and nine females on the islands of Nui, Nanomaga, Niutao and Nukulailai; he did not see it on Funafuti, and Rainbow does not record it from there, but it is possible that Whitmee’s specimens, referred to below, are from this island, since the name “ Ellice Is.,” now generally applied to the whole group, was formerly used for Funafuti only. Of the specimens captured by Buxton, all are typical HE. eleutho distincta except the two males from Nukulailai, both of which have the white pattern much reduced ; in one cf them the reduction affects both wings equally, but in the other the hind- wing pattern is almost obsolete, the inner row of spots showing more reduction than the outer, while on the forewing the spots, though obscured by dark suffusion, are larger than in the first specimen. Unfortunately these are the only two specimens captured on this island. Of two males and three females in the British Museum, labelled “ Ellice Is., Whitmee,’ both males and two females are typical H. e. distincta; the third female is almost identical on the upper- BUTTERFLIES OF SAMOA AND SOME NEIGHBOURING ISLAND-GROUPS. 15 side with Tutuila specimens, but on the underside there is no trace of the greyish suffusion so characteristic of the latter, so that I see no reason to suppose that it is incorrectly labelled. The form, however, would appear to be uncommon. A large male in the British Museum labelled ‘‘ Atafu, Union Is., J. J. Lister ”’ (apparently the only record of a Huploea from this group), is like the better- marked specimens from the Ellice group in pattern, and HE. e. distincta also appears to be the common form in Wallis Island (Poulton, p. 584, Pl. XLII, figs. 8-10). Maximum, mean, and minimum expanse of specimens from the Ellice Is. are: male 81, 75°5, and 72 mm. ; female 77, 72, and 68 mm. The larva was found on a species of Ficus, which grows either as an independent plant, or as an epiphyte in a tuft of fern on a coconut palm. Unfortunately it was not preserved. Buxton notes that the flight of the adult in the Ellice Is. is not at all like that of £. schmeltzt, beng much more soaring, and that the insect makes long flights from tree to tree ; this is also true of the races of #. eleutho found in Tutuila and in Tonga. The butterfly was fond of sitting on the broad leaves of various trees, and was not common except in Nui. 4. Euploea schmeltzi schmeltzi (H.8.). Euploea schmeltzi ; Herrich-Schaeffer, 1869, p. 70, PI. I, fig. 8. Butler, 1874, p. 277. Fruhstorfer, 1910, p. 241. Rebel, 1910, p. 416, Pl. XVIII, figs. 2 and 3. Euploea schmeitzw ; Schmeltz, 1876, p. 181. Pagenstecher, p. 302. Deragena schmeltzw ; Fraser, p. 147. Moore, p. 272. Waterhouse, 1904, p. 492. Deragena schmeltz ; Swezey, 1921, p. 602. Euploea schmeltzu schmeltz ; Poulton, p. 596. This species occurs very commonly in Western Samoa with D. m. melittula, and frequents the same kinds of flowers, but in addition it is found in flocks of many hundreds on Tournefortia argentea L. (BORAGINACEAE), a common strand tree (Buxton, 1926, Hopkins, 1926). All the specimens captured on this tree (several dozen in number) turned out to be males; Armstrong on one occasion saw ‘“ About 150 on one dead branch below the tree, all males.” Trees of the 16 INSECTS OF SAMOA. © same genus are very attractive to Huploea mathewi in Tonga, and to several species of Huploea in the New Hebrides (Buxton) and the Solomons (Woodford, 1890, p. 94). In all these cases it is noticeable that only males are attracted, though both sexes frequent flowers, and that it is not the flowers but the dead and withering twigs and branches that are attractive. I once saw many hundreds of Huploea schmeltzi on the fruit-clusters of 7. argentea in Savaii, but even in this case it was the dead and withered clusters that were preferred. The tree does not seem to be attractive to any other butterflies, except moderately so to males of D. m. melattula, and it is difficult to imagine in what the attraction lies; no exudation of any sort was observed. Since the attraction is almost confined to males of the genus Huploea, it seems not impossible that the scent may resemble that of a virgin female, and that the volume of it is sufficiently great to make up for slight differences, and hence make it attractive, not to one species of Huploea only, but to many. It should be noted that this habit is quite distinct from that of roosting in flocks, which is so common in the genera Huploea and Danaida. In Tonga, in addition to seeing the males of L. mathewi on Tournefortia, | was able to find both sexes going to roost in the evening about half a mile inland; they did not appear to show a preference for any particular tree, and there were no specimens of Tournefortia in the neighbourhood, this tree being entirely confined to the beach. The case seems to some extent parallel with the attraction of isoeugenol and methyleugenol for certain TRYPETIDAE, and of naphtha and kerosene for Ceratitis captitata ; this also has not yet found a satisfactory explanation, but has been attributed to some form of sex-stimulus, since in this instance also it is the males only that are attracted. Variation is considerable in both sexes, both in size and markings, but all forms may be taken together at the same time and place. The white markings on the upperside are usually much better developed in the female than in the male; the reduction of pattern, especially in the latter sex, may go so far as to leave no indication of the white markings except the short row of sub-apical spots, while on the other hand all the markings may be well developed. In some specimens the forewing pattern is well developed, and that of the hindwing almost obsolete. My largest specimen is 69 mm. in expanse, and my smallest 60 mm. ; both are males. On the 7th April, 1924, Buxton observed a very interesting flight of this species in the district of Aleipata, Upolu Island. At about 9 a.m. he found the BUTTERFLIES OF SAMOA AND SOME NEIGHBOURING ISLAND-GROUPS. 17 butterflies in numbers flying out from the mainland towards the small out- lying islands of Namua and Nuutele (Text-fig. 1). As there was no perceptible wind, and the directions of the flights to the two islands were nearly at right angles, he concludes that the islands must have been visible to the butterflies from the shore, a distance of more than a mile in the case of Nuutele. There does not appear to be any other possible explanation of the facts.* The form is found commonly through- WP OL Uv out Western Samoa, but we did not CCranvarary meet with it in Tutuila; Schmeltz (p. Z-[ Gann 181), however, records a specimen from pee there; Rechinger (Rebel, 1910, p. 417) \ : ewes Lutuila) vas; one or the localitiés |, -.0 <2 2 3 ames Ginus where he observed it; and Poulton (p- TEXT-FIGURE 1.—Coast-line in the Aleipata. 606) states that Mathew found “ Huploea district of Upolo Island, Samoa, show- [helcita bourke: and schmeltz.]” at Pago- me ee flight of Huploea schmeltze : rom mainland to outlying islands. Pago. The names in brackets were added by Poulton, who informs me that the original record merely gives the genus and no specific names ; there can be no doubt that the species referred to was Ly. eleutho bourket. Rechinger’s record is equally unsatisfactory ; all the specimens (twelve in number) brought home by him were from localities in Upolu and Savai'l, nor were any of his specimens of other species of butterflies taken in Tutuila (Rebel gives full particulars of the number captured in each locality and the total number obtained), so that it is obvious that he did not collect in Tutuila at all, but merely saw a species of Huploea (which must have been E. eleutho bourket) and assumed that it was LE. s. schmeltz.. This is further borne out by the fact that he did not capture either H. e. bourke: or the male of Hypolimnas bolina, although these are the commonest butterflies in Tutuila. There are no specimens from Tutuila in the British Museum, and I think we are justified in ignoring Schmeltz’s old record, and stating that E. e. schmeltzi is entirely confined to Western Samoa. Outside Samoa its nearest relative is H. schmeltzi whitmer (Butl.), which occurs in the Loyalty Islands. These closely related forms are widely separated geographically, * See p. 46 for a similar habit of Catophoga jacquinotit manaia. II. 1 2 18 INSECTS OF SAMOA. and no connecting forms have yet been described; the intervening islands are, however, so little known that it is quite possible that such forms exist on them, but have not yet been met with. The egg is barrel-shaped, 1°6 mm. in height and 1-0 mm. in diameter, pale yellow, with some twenty-four longitudinal and about the same number of transverse ribs dividing the surface up into roughly rectangular areas, which are more irregular towards the apex. ‘The shell is particularly soft, and the egg is easily crushed. The ova are laid singly on the upper or underside of a leaf of the food-plant, Ficus tinctoria Forst., a common bush which grows either as an epiphyte in the crowns of Pandanus and other trees, or as an independent plant. The butterfly is not strictly confined to this species, but sometimes selects other members of the genus Ficus. - The larva spins a pad of silk on the leaf to improve its foothold. When young it is green, with a pale yellow sub-spiracular line; head, legs, prolegs and two pairs of fleshy filaments, one on the mesothorax, the other on the eighth abdominal segment, are black. The full-grown larva (PI. IV, fig. 3) is about 30mm. long, and varies considerably in colour; the form found invariably in a wild state has the body, legs, and prolegs glaucous-green ; the head is pale brown, with two pale green lines; there is a broad pale yellow sub-spiracular line, and the spiracles are black ; the fleshy filaments, which are large in proportion to the size of the larva, are purplish-brown in colour. When larvae are reared in captivity, either from eggs or from very young caterpillars, the commonest form when full-grown is very dark, almost black, with a yellow sub-spiracular line and black filaments; a modification of this form has the head black, the body above the yellow sub-spiracular line deep black with one rather broad and three broken and rather narrow white transverse bands on each segment, and spiracles, filaments, legs and prolegs black. Almost all larvae in captivity are of one of these dark forms, unless captured when already half-grown.