agrius cqnvolvuli. 37? 



June-July. Hyde observes (loc. cit., p. 314) that the imagines were 

 plentiful in early August at Portland, but only in fair condition, 

 whilst fewer specimens, but in much better condition, were 

 captured in September ; the fact that he obtained a larva on 

 August 2nd, and a pupa on September 22nd made him 

 believe the late September-October imagines were freshly emerged ; 

 Fowler also notes (loc. cit., p. 321) that imagines were 

 common from August 16th to the end of the month, chiefly $ s, 

 and that they occurred up to September 13th; then there was a 

 break, and no more were seen till the first week of October, when 

 four perfectly fresh $ s were netted. Several records note that 

 imagines were found, towards the end of the season, on the ground 

 more or less moribund. Careful comparison of the detailed " times of 

 appearance" (posted, pp. 379 et seq.) for the various years in which the 

 species has occurred in the British Islands will give students further 

 material in this direction. Mason notes ( Young Nat., x., p. 233) 

 that bats attack the moths when on flight at flowers ; he says 

 that, at Clevedon, he saw them swoop down on the moths, but did 

 not observe that they caught any although they spoiled his sport. 

 Time of appearance. — In the subtropical home of this species 

 the insect is possibly almost continuously-brooded, passing, how- 

 ever, its hybernating period (so far as it has any) in the pupal 

 state. Harker notes that, in Grand Bassam, the pupal period is 

 exceedingly short, rather less than a fortnight (Ent., xxix., p. 66), 

 and that the insect goes through its metamorphoses in a very short 

 time, and this appears to be true also in India, North Australia, 

 and the Pacific Islands. In Madeira, northern Africa and localities 

 with a somewhat similar limited winter, the species appears to be 

 more or less sedentary and double- or triple-brooded, appearing in 

 April-May, July-August and October, whilst parallel habits possibly 

 prevail in countries with a similar climate south of the equator. It 

 is possible that the species never passes the winter successfully in 

 the greater part of the Palaearctic region, and certainly almost all 

 the specimens taken in Europe and northern Asia are either immi- 

 grants or the direct descendants of immigrants from subtropical 

 climes. Even at Hyeres, although the imagines are abundant in 

 August and early September at flowers of honeysuckle, convolvulus 

 and petunias, the imagines have not, Powell asserts (in litt.), been 

 seen at large in June, although an imago emerged in confinement 

 on May 27th, 1903, from a pupa which had been exposed all 

 winter, to all weathers and temperatures, on an outdoor balcony. 

 He adds that if this represents a first brood, it must be a very 

 small one. Similarly in Britain, even in confinement, there are 

 no records of an imago emerging from a pupa that has spent the 

 winter as a pupa, except that already noted (anted, p. 343, footnote). 

 The summer and autumnal larvas and pupae that have been obtained 

 in Britain (anted, pp. 343-345) have either gone through their changes 

 (naturally or by forcing) rapidly to the imaginal state, or died, and 

 a similar condition occurs throughout all temperate regions, the 

 warmer the district and the less marked its winter, the greater being 

 the percentage of successful emergences, due largely to the greater 

 ease with which the larvae can obtain food. In some seasons, e.g., 

 1857, 1858, 1859, 1868, and 1887, immigrant imagines appear in June 



