34 



May 20, July 1, July 17, August 2, September 7, October 12, jiud Octo- 

 ber 10; full-grown larva3 were observed May 11, May 22, Juue 16, July 

 0, August 13, Sei^tember 13, and October 14. 



C 



Fig. 5.— Nematus vkntraus. a, a, a, young larvae; 6, full-growu larva; c, cocoou; d, adult; all slightly 



enlarged (original). 



In the grounds of the Agricultural Bei^iartment in Washington tbese 

 slugs were extremely abundant upon isolated willows on July ; eggs 

 were deposited July 17, which produced the second brood. A third 

 brood became nnmerous August 2 ; eggs for the fourth brood were de- 

 posited September 7, and eggs for a fifth brood hatched October 14. 

 No saw-fiies had been observed during the j^revious year nor prior to 

 July 6, thus showing that they came from some other willows, and that 

 the larva* then noticed were the offspring of females of a first or perhaps 

 second brood, which had been forced to search for new breeding-places, 

 the former food supply having become exhausted. 



The number of broods on these isolated willows could be studied with 

 great ease, since each brood of slugs consumed in turn nearly all the 

 younger leaves in the course of their larval existence. Before the winged 

 flies appeared and before their eggs had hatched, the young leaves and 

 shoots had attained quite a size, and offered enough suitable food for a 

 new brood to feed to full growth. In larger patches of willows such ob- 

 servations are more difficult, because in the course of time the several 

 broods overlap each other, so that in late summer slugs of all sizes can 

 be found at the same time, and there is no indication of the brood to 

 which they belong. 



