438 BRITISH LEPIDOPTERA. 



hairs are visible to the naked eye, and are much more regular than 

 any I have seen, and are also striated, with beads or clear spots. In 

 G. americana the scales forming the dorsal tufts both on the two 

 hinder thoracic segments and on the 8th abdominal are very 

 different from those of the European species ; they are dark and 

 opaque but are long, narrow, flattened, very gradually increasing in 

 width to the end, which has a single notch, and from the single 

 notch an impressed line or stria extends along the middle for some 

 distance. These flattened hairs seem common to the family of Lasio- 

 campidae, and should bo looked for in the European species of this group. 

 In the larva of Hvteropacha rileyana there are no dorsal scales, but some 

 of those in the lateral tufts have flattened ends, which are very long 

 and slender, lanceolate oval, with the tip much attenuated. I have 

 been unable to discover these singular scales and flattened hairs in 

 Clisiocampa americana, or C. neustria of Europe, or in any other 

 family of lepidoptera, except in the hairy Noctuina or Noctuo-bom- 

 byces, or Bombycoidea, where the hairs with flattened ends probably 

 occur in the more hairy and pencilled species. In the larva of the 

 common American Acwnycta hastulifera, many of the barbed hairs 

 forming the black pencils are flattened at the end and black, but not 

 striated. These specialised and highly differentiated seta?, so like the 

 scales of adult lepidoptera, appear to be of use in rendering the pencils 

 and tufts more conspicuous and stiff. The shortest and broadest, 

 striated, scale-like setre occur on the low, broad, stout, dorsal median 

 tubercles of Gastropacha, and, perhaps, add a repellent nature to these 

 shiny dark metallic tufts. At all events the occurrence of such scales 

 is an interesting example of the acceleration of development of the 

 sets in these larval forms, and it is not improbable that in the ancestors 

 of the Lasiocampidae they were characters acquired during the later 

 stages of their larval lifetime." Bacot has observed that the larva of 

 Dendrolimus pint has also very specialised scales of a somewhat similar 

 character. 



We have already (ante, vol. i., pp. 120 et seq.) given a brief review 

 of the chief general characters presented by the Lachneid larva, and 

 have shown that the tubercles are in great measure specialised, and that 

 there is also much specialisation exhibited in the formation of a thick 

 hairy coat developed from the skin and not from the tubercular hairs, 

 whilst the tubercles proper, or their warts, become atrophied in the older 

 larval stadia. In position, i and ii form ordinary trapezoidal tubercles, 

 i usually very strongly developed compared with ii (in P. populi, ii is 

 larger), wart-like, and bearing several hairs, whilst the latter is weak 

 and bears only a few hairs (ii is atrophied in Saturniids) ; iii also is 

 poorly developed with only one or two hairs ; iv and v are both sub- 

 spiracular, iv large, v ill- developed, and generally coalesced with iv 

 even in the earlier stages, whilst vi is fairly well-developed towards the 

 base of the leg. A number of secondary hairs on the anterior portion 

 of each segment tends greatly to obscure the true tubercular structures. 

 The more generalised Lachneid larva? — e.g., Eustaudingeria vandalicia 

 — have i and ii large and elongate transversely, an Anthrocerid feature ; 

 P. populi larva has, as we have noted, ii larger than i, a somewhat 

 peculiar feature in this group, whilst in the first instar Pachygastria 

 trifolii, Lasiocampa querciis, Eutricha qtteriifolia, &c, have both i and ii 

 large, many-haired warts, iii also a large many-haired wart, and iv + v 



