SATURNIA PAVONIA. 329 



Worcester, a cocoon constructed with three valvular emergence 

 apertures, two of which are quite perfect and beautifully formed, the 

 third being deficient in the usual converging filamentary portion ; 

 this was constructed by a larva from which a fine $ specimen was 

 afterwards bred. Reaumur refers (Mem., i., pp. 634 — 635) to a 

 larva that spun a cocoon without a normal opening, the resulting 

 imago being quite unable to force its way out on emergence, and 

 Rye records a similar one " rounded at both ends like an ordinary 

 ' eggar ' cocoon.'' Sheppard exhibited a similar one at the meeting of 

 the Entomological Society of London, August 4th, 1851, and Poulton 

 mentions another (Trans. Ent. Soc. Loud., 1892, p. 311) spun without any 

 apparent valvular opening, whilst Bell notes (Ent. Rec, vii., p. 71) two 

 cocoons without the usual outlet, almost spherical in shape, 

 and of small size. Newport describes (Zoo/., xii., p. 4175) a large 

 flattened cocoon, the joint production of two larvae, divided by a 

 septum into two chambers, in one of which was a dead 

 larva and, in the other, a pupa from which the imago had 

 failed to escape ; Harwood mentions (Ent. Wk. Int., vii., 

 p. 28) a cocoon that contained two S pupae, and Frohawk describes 

 and figures (Ent., xvii., p. 73) a similar one, one of the imagines 

 from which emerged, but failed to get out of the cocoon ; Fowler 

 notes two cases in which two larvae, taken at Ringwood, spun 

 common cocoons, each the work ot two larvae, whilst Milton mentions 

 (Ent. Rec, hi., p. 12) a cocoon containing two pupae; SeyfHer states 

 that such cocoons have frequently come under his notice in Wurtem- 

 berg. Such cocoons appear to result only from an overcrowding in 

 confinement. At the meeting of the Entomological Society of London, 

 held December 4th, 1901, Pickett exhibited several abnormal cocoons, 

 one with two openings, one with no opening, and a third containing 

 three pupae, whilst at the meeting held on May 4th, 1868, Trimen 

 exhibited a cocoon in which the imago, having failed to get out 

 of the anterior end, forced its abdominal segments through the 

 posterior end. 



Variation in colour of cocoons. — Cocoons of this species vary 

 in tint from pale yellowish or white to almost black. Quail and others note 

 light cocoons as occurring among those spun by Wicken larvae, the silk 

 almost white; Culpin a long white one with two exits, whilst other pale- 

 coloured cocoons are on record, all, however, made by larvae reared in 

 confinement. Poulton (Colours of Animals, 1890, pp. 142-6) suggested 

 that this difference in colour was due to an ability on the part of 

 the larva to spin a cocoon in response to the conditions of its 

 surroundings. This view was challenged by Bateson and others, 

 and the former (Trans. Ent. Soc. Lond., 1892, pp. 45 et sea.) detailed 

 at length observations which he had made in 1891 on certain 

 experiments that he had carried out, that tended to contradict 

 Poulton's conclusions. He notes : 



(1) Eleven cocoons found spun in hedges in a state of nature, all of full colour. 

 (2) Fifteen larvae, shut up in and among various dark substances, spun cocoons which 

 were all light in colour, some being quite white, no dark cocoon being spun by a 

 larva thus confined. (3) Fifteen lame fed in a large vessel on food surrounded with 

 crumpled white paper, produced 7 dark cocoons (3 attached to the paper and 4 

 among leaves), 4 light brown in colour, 4 white (attached to paper and leaves). 



As a general fact Bateson found it difficult to obtain any 

 considerable number of dark cocoons from S. pavonia larvae in 



