18G4.] 581 



willow on which it grows, but always entire. When young and immature, 

 the galls are spherical and are enveloped in a dense mass of foliage, which 

 gradually falls off towards the autumn, and by November the twigs on 

 which they grow, if small, are already killed for an inch or two down- 

 wards. Occasionally at the extreme tip of the gall the leaves open out a 

 little, as in S. sti-obih'scus n. sp., but without projecting from the tip as in 

 that species. Easily distinguished from that gall by the portion of 

 each leaf which lies "to the weather," towards the base of the gall, not 

 terminating in a rectangular point, but describing a circular arc. The 

 leaves are also more densely pubescent, especially the portion that lies 

 " to the weather." Appears early in the summer and is full-sized by 

 the middle of July, at which time that which is reddish brown in the 

 dry gall is greenish white. The pubescence on the leaves retains its 

 glaucous-white color to the last, except where they are badly weather- 

 beaten. On the same bush throughout the summer may be seen the 

 old, dry, last year's galls, and the young growing galls of the current 

 year. I have already referred to the Orchc/hnuni eggs often found 

 under the scales of this gall. (Froc. Ent. Sor. Phil. III. p. 232.) In 

 one gall examined this autumn I counted no fewer than 71 of these 

 eggs. In September I detected a species of A'^jj/u'rfmm, which accord- 

 ing to Mr. Uhler is undescribed, ovipositing in the pith at the tip of a 

 broken stem of Golden-rod (Solidago). Probably Locusturise Latr- 

 (=(Tryllida3 Leach) do not so generally oviposit in the earth as authors 

 have hitherto led us to believe. 



Larva. — Five specimens examined Nov. 15 and many subsequently 

 did not differ from the larva of S. hrassicoides, the breast-bone being 

 similar and varying in the same manner. Length .08 — .20 inch, breadth 

 .04 — .07 inch. Out of nearly 20 galls opened at this date all contained 

 the cocoon, though many cocoons contained another cocoon in which 

 lay a Proctotrupide imago about .10 inch long. The cocoon differs 

 from that of S. hrassicoides in being 21 — 3 times as long as the larva 

 and truncate at tip, the tip end forming a kind of diaphragm not far 

 from the tip of the lanceolate cell formed by the interior leaves. The 

 diameter of the cocoon does not greatly exceed that of the larva, which 

 is always found lying closely in its basal end, the rest of it being hollow 

 and empty. Specimens of the larva examined Feb. 20 were of a pale 

 orange-color, and others examined March 20 of a deep orange-color. 



