84 



"red spider" attack. It is almost certain tlmt this same red spider 

 had killed the sawfly eg'gs or at least prevented their development. 



DEVELOPMENT. 



The first of the larval ofl'spring- of the female found March 17 

 attained mature growth and deserted the plant on which it had fed 

 May 3. In the absence of more accurate information we may assume 

 that this female issued about March 14, and that eggs were deposited 

 during the next three days. It may l)e assumed also that the egg 

 stage was at this season about eighteen days, which would leave an 

 active larval period of al)out four and a half Aveeks. 



From eggs that were deposited l)y a lly which was confined alone 

 with a violet plant April 19, the first lar\a hatched Ma}' 1, giving as 

 the duration of the egg state for this period, which was colder than 

 normal, twelve da3'S. 



From May 1 until about the 10th or later the other larvic of this 

 brood, about a dozen in number, deserted the plant and crawled about 

 on the glass cylinder confining them. 



May 25 the first fiy of this brood was found to have issued, having 

 passed the entire cycle from egg to imago in about ten W'Ceks. 



The first larval molts of the first spring generation were not observed. 

 A number of mature larvffi were isolated and ol)servations made from 

 the time of the penultimate molt to the issuance of the adult. 



No. 1 molted April 21 and afterwards, and transformed to pupa 

 May 25. 



No. 2 molted April 22, also afterwards, and transformed to pupa 

 May 2(3. 



Nos. 3 and 4 molted May 10, and had assumed the contracted form 

 by May 21. May 25 both transformed to pupa, and to imago May 31. 



No. 5 molted May 6, began May \) to bore into the pith of a stem 

 supplied for the purpose of pupation, and in the course of an hour 

 had obtained entrance and closed the aperture with the comminuted 

 bits of pith produced by its boring. Ma}^ 11 a portion of pith was 

 removed that further transformation might be observed. Pupation 

 ensued May 25 and the adult appeared June 1. The last fly observed 

 issued June 11. 



These periods, if so we may term them for present convenience, are 

 of course varial)le according to atmospheric conditions — heat or cold, 

 humidity or dryness — and are probably also subject to individual vari- 

 ation in the larval stage. Approximately, we may say that the pe- 

 nultimate larval molt in the first spring brood takes place about five 

 weeks from the time of pupation, the final larval molt occurring about 

 eighteen or twenty days later than this. The duration of the final 

 contracted larval stage could not definitely be determined, as it is 

 gradual and hence it is difficult, if not impossible, to determine where 



