NY.MPIIALINAE : BASILAIICIIIA ARCHllTUS. 275 



the base, they collect the debris into <a little bunch, which they fasten with silk 

 to the midril). When the hibernaculum is llnished the seam is perfectly smooth and 

 the whole inside is lined with sillc. The larva, after corapletinii; its work, composes 

 itself for the winter, Avith the tail towards tlie orifice. Here it remains till the catkins 

 are in bloom the next sprinj;, when it retreats from its honse and commences feeding. 

 Not the least wonderful part of the phenomenon is, tliat it is only the autumnal brood 

 of larvae that form pitcher-like houses to live in during the inclement season of the 

 year, the summer brood having no occasion to shelter themselves from the cold. 



I am indebted to Miss Guild and to Messrs. Lintner, Sanborn and 

 Trouvelot for many specimens of this larva, and to IMessrs. Emery, Mark, 

 Shute, Smith and Trouvelot for chrysalids. 



Life history. In New England, this is at least a double brooded 

 insect, tliough it is more than likely that it is triple brooded in the southern 

 stations and probable that a third supplementary brood will be found 

 everywhere. It starts the year in its larval hibernaculum as a caterpillar 

 ordinarily in its second stage, but not at all infrequently in its first, and 

 rarelv in its third stage. To escape from its narrow winter quarters it 

 must retreat backwards, and when it does this it is to find only the bursting 

 catkins, and as yet no leaves of its food plant; upon these it breaks its 

 lono- fast, and even sometimes attains nearly half its grow'th, but in May 

 becomes full fed upon the tender leaves. The butterflies appear early in 

 June, usually by the 5th or 6th, in the latitude of Boston, the earliest 

 date recorded being May 31, Avhen Mr. Lintner has taken specimens near 

 Albany, N. Y. I have also found the larva nearly full grown at New 

 Britain, Conn., on May 15. But they rarely become abundant before 

 the middle of June, and continue to emerge from the chrysalis until the 

 end of the month. Pairing takes place at once (Mr. Lintner records a 

 case June 7 ) , and eggs are laid in about a fortnight after the first 

 appearance of butterflies, the earliest known to me being indicated by 

 recently hatched caterpillars in the first stage as well as eggs found 

 June 26, when the butterflies had been flying at that season about three 

 weeks. In the north they ordinarily appear a week or ten days later than 

 about Boston, the earliest dates known to me being June 9, at Hallowell, 

 Me. (Miss Wadsworth), and June 11, at Wing Road, N. H. (Scudder), 

 and I have also found badly battered females at Franconia, N. H., on 

 June 27, which must have been out a fortnight. It is therefore a little 

 earlier than B. arthemis, having apparently a little the start of it in the 

 earlier provender provided by natin-e for the use of the hibernating cater- 

 pillars. Eggs are certainly laid in June, for I have found caterpillars in 

 their first stage during the last week of June at Plymouth, N. H. 



There is no little difference, apparently, in the rapidity with which the 

 transformations are made, the eggs alone varying from four to eight days 

 (ordinarily about six and one-half), but although the caterpillar is sluggish 

 and by no means a gluttonous beast (as many are), the transformations 



