332 BRITISH LEPIDOPTERA. 



anus. The pupa of 6". pyri is, in shape, almost, if not quite, 

 cylindrical, tapers only very slightly to the shoulders (base of wings), 

 and then abruptly to head, forming a blunt front. It tapers more 

 evenly towards anus, which is rather blunt, and does not curve up 

 ventrally as the pupa of S. pavonia does, though the segments are 

 somewhat compressed ventrally. The pupa of S. py?'i is not unlike 

 that of Mimas tiliae in shape, only rather wider in regard to its length, 

 and, of course, differing at head and anus. In general shape and 

 appearance, omitting the antenna-cases, it tends rather towards the 

 Amorphic! (Smerinthid) than the Saturniid shape, as typified by S. 

 pavonia (Bacot). The pupa of S. pyri is much nearer the typical 

 Saturnian pupa than is that of S. pavonia, which is really a rather 

 exceptional form, both in its curvature and antero-posterior flattening. 

 The Smerinthid (practically =Citheronian) pupa is the basal form of 

 Saturnian pupa (Chapman). 



Pupal habits. — Like Lachneis lanestris, Dimorpha versicolora 

 and other species that appear to have had a northern origin*, 

 Saturnia pavonia frequently goes over two, three or more winters 

 in the pupal stage*, and this habit appears to be almost as frequent 

 among broods of more southern as among those of more northern 

 origin. Chapman records breeding imagines in May at Glasgow 

 after passing two years in the pupal stage, and Barrett notes a 

 whole brood, reared from Yorkshire ova, that pupated in August, 

 1866, went over the whole of 1867 in the pupal stage, and produced 

 imagines in the spring of 1868. Mead observes that from 1890 pupae he 

 bred imagines January 22nd (c?)> February 2nd ( 2 ), February 7th 

 (2 $ s), and February 9th ( $ ), 1892 ; whilst Nash bred a ? March 31st, 

 1890, and a $ April 5th, 1890, at Hardwick, from pupae of the 

 second year. Fenn reared, in 1890, from Bournemouth ova of 1889, 

 a large number of imagines between March 8th and May 19th, but 

 a few pupae went over a second winter, and the first of these emerged 

 June 1 st, 1 89 1 ; from the same locality Mrs. Cowl has repeatedly 

 bred imagines after the pupae have gone over two winters. Beadle 

 had several emerge {Ent. Rec, xi., pp. 280, 306) after being three 

 years in the pupal stage, and King {loc. cit., i., p. 109), in the spring 

 of 1890, bred several that had been in the pupal stage three years. 

 [See also anted, p. 315.]. Hoffmann notes that, in the Hartz, the 

 pupal stage sometimes lasts 2, 3 or 4 years, such pupae generally 

 producing $ s, whilst Hewett insists that the bulk of the imagines 

 reared in the York district from two- and three-year old cocoons 



* We suspect strongly that this custom lias much to do with the hereditary 

 persistence of a habit engendered, in the case of these species, when they had to 

 meet a much longei winter (and shorter summer) than now. We may suppose 

 that this happened in their present breeding-grounds, or that the habit was 

 engendered in a more northern latitude, and that the species have since spread, 

 maintaining the habit as being even now occasionally of service to them. It 

 occurs in a much more southern latitude than ours, for Warburg states that six 

 larvae from southern France, fullfed in June, 1885, gave imagines in February, 1886 

 and 1887, and Turner reared a brood (parents captured wild at Digne in the 

 Basscs-Alpes in April, 1898), that gave some examples that emerged after being two 

 years in the pupal stage, viz., in April [899, 21 S s; in April 1900, 2 ^ s and 7 ? s. 

 Standfuss observes (Handbuch, etc., p. 105) that 50 per cent, of pupae from Dantzig 

 did not give their imagines the first spring (i.e., after one hybernation), whilst only 

 30 per cent, of 'pupae from Naples and" Capri lay over. The maximum period of 

 delay observed by Standfuss has been four winters. 



