So that Walshii, during a period of eleven years, has been seen as 

 early as 9th March, and as late as 30th May ; and Telamonides as early as 

 29th April and as late as 9th June. In the one case there has been a 

 range of 92 days, in the other of 41. The first appearance of Walshii 

 depends altogether on the state of the weather. If a few fine days come 

 early in March, some individuals will surely be seen long before the bulk 

 of their generation ; and these early comers are invariably cut off by the 

 severe weather which follows. Walshii is in fact an April form here, 

 while Telavioiiides is scarcely ever seen in April, but is a May form, not 

 descended from Walshii of the same year, but like Walshii, coming from 

 over-wintering chrysalids of all three forms of the preceding year. After 

 I St June, Marcellus coming from the eggs of Walshii, laid last of April 

 and early in May (before which time the food plant Pawpaw, Anona 

 triloba, does not put forth the first flower bud or leaf bud) abounds, and 

 continues in successive generations till late in the season. 



During the middle of the Walshii period many females may be seen, 

 worn and broken, evidently some time out of chrysalis, and in their last 

 days. In 187 1, loth April, I recorded that I took 3 % Walshii "bat- 

 tered and broken." Whereas, a few days later, on 16th April, I took 4 

 fresh ^ . One half the generation was passing away while the other was 



