376 BRlllSH LEPlbOPTERA. 



some time. There is abundant evidence that the late moths are 

 killed off by cold without having deposited their eggs, and that 

 they are exterminated every winter following their appearance in 

 our country \ but there is no evidence whatever in support of 

 Robson's statement (Young Nat., hi., p. 267) that "the species hybern- 

 ates as an imago and reappears in June." As tending to prove its 

 migrating habits, Boswell states that the species is sometimes 

 abundant (e.g., 1875) m the Orkneys, but that no Convolvulus grows 

 there, for, though Neill gives C. arvensis as an Orkney .plant, no 

 one has found it there since. As bearing on the question of the 

 immigration of our British examples we have two very definite sets of 

 facts, illustrated excellently by the specimens captured in 1885 and 

 1901 respectively, both years in which they appeared in this country in 

 great abundance. In 1885, the species appeared suddenly in August 

 and continued on into September, and were in no way preceded by the 

 finding of larvae or pupae in this country, nor by any early (June or July) 

 imagines ; these August and September imagines were, therefore, 

 possibly direct immigrants from more southern climes and not 

 bred in this country. This view is supported by the facts that, 

 on August 14th, 1885, Druitt caught a $ at Christchurch that 

 laid him 4 eggs, a most unusual proceeding for late, and 

 undoubted British-bred, examples (Ent., xviii., p. 259). In 

 1 90 1, on the other hand, imagines were taken on June 2nd at 

 Portland (Ent., xxxiv., p. 314), 2 s on June 30th, July 2nd, 10th, 

 and a $ on July nth at Elstow {Ent., xxxiv., p. 230), one also 

 on July 1 6th at Wandsworth (Ent:, xxxiv., p. 230), and, by early 

 August, larvae were found, e.g., at Portland on August 2nd (Ent., 

 xxxiv., p. 314), at Maldon on August 14th, on the sandhills of the 

 Lancashire coast on August 19th (loc. cit., p. 254), at Dovercourt 

 from August 1 6th to September 10th, and at Dover, on August 

 19th (loc. cit., p. 281), at Dawlish, on August 28th (loc. cit., p. 313), 

 at the end of August at Hayling Island and at Bournemouth (loc. 

 cit., p. 314), also in Northumberland {loc. cit., p. 358), whilst pupae 

 were found on the Sussex coast (loc. cit., p. 295), near Haddenham, 

 on September 19th (loc. cit., p. 313), at Portland, on September 22nd, 

 and at Bournemouth (loc. cit., p. 314); larvae and pupae were 

 obtained between September 6th-2oth at Ringwood, from which 

 two imagines were bred on October 3rd and 14th (loc. cit., p. 321). 

 But imagines, principally $ s and not in good condition, were taken 

 throughout early and mid- August, followed by fine recently-emerged 

 ones during September and October, but not in any very great 

 numbers. It is well-known that this insect is long-lived, and that its 

 eggs are slowly matured during the whole imaginal life, and 

 that, as we have already stated, without food the imagines 

 soon die, and that their eggs remain undeveloped *. There 

 appears to be no doubt, therefore, that the early August, 

 1901, examples were the remnants of the probable immigrants of 

 * When Poulton exhibited larvae of this species, reared from eggs laid August 

 29th by a ? , taken in South Devon (see antea, pp. 339 et seq.), at the meeting 

 of the Ent. Soc. of London, October 5th, 1887, McLachlau is reported to have 

 remarked that " ? s of this species, taken on former occasions when the insect 

 had been unusually abundant" had been found upon dissection to have the 

 ovaries aborted." We have found no evidence of this abortion although there 

 is plenty to show that the eggs are ill-developed, and that the food + temperature 

 obtainable in these islands in September and October are not altogether sufficient 

 to enable the ? s to develop them satisfactorily. (See also antea, p. 338, footnote 2.) 



