1864.] 608 



be sometimes the case in >S^. batatas, assumed the form of a slight, 

 elongate-oval enlargement of the twig; but from these I bred nothing. 

 < )n the whole, further investigation will be required to determine, whe- 

 ther tbe imago produced from these galls on S. cordata and S. discolor 

 is identical with Cec. s. batatas. It is rather singular, that in these 

 »S. discolor galls I found no larvae at all of the Decatoma (?) and Calli- 

 mome, which so greatly outnumber the larvae of C. s. batatas in the S. 

 humilis galls. 



It is worth remarking, that the only two Cecidomyidous galls which 

 appear to grow on more than one species of our Rock Island Willows — 

 viz. aS'. batatas and C. siliqua — occur on as many as three species of 

 them, and that these three species should in both eases be the samf", 

 three, viz. S. humilis, S. cordata and S. discolor. The chances are very 

 greatly against such an event happening, without some good and suffi- 

 cient cause for it. Mr. Bebb informs me that there is a close alliance 

 between S. humilis and S. discolor; but that neither S. cordata (=S. 

 rigida) nor S. lucida is allied to the first two. While on this subject 

 I may say that Mr. Bebb has re-examined the doubtful species of Wil- 

 low referred to above (p. 546), and has concluded that it is certainly 

 neither 8. nigra nor S. alba, and that it is not improbably S. fragilLs 

 Lin., a species which has been introduced into N. A. from Europe. 



Larva. — July 24 and 30 the larvje were orange-colored, .09 inch long and with 

 a slender, black, clove-shaped breast-bone, and some of them with curdy, bowel- 

 like markings above and below. The surface of the cell was then opaque and 

 rather rough and scaly. Nov. 11 and subsequently the surface of the cells was 

 smooth, the cocoon apparently adhering to their sides but being scarcely sepa- 

 rable except piece-meal. Larvae from 12 to 20 in number examined at various 

 times in November were .08 — .10 inch long and .04— .05 inch wide, not shining 

 but rather opaque as in most of the preceding, except C. s. nodulus, of a sangui- 

 neous color with dominant, bowel-like, yellowish markings, which are situated 

 mostly between the sutures, the sanguineous color generally forming a wide dor- 

 sal vitta, widely interrupted between the sutures on each joint. Breast-bone 

 coal-black, clove-shaped, the head of the clove towards the mouth, 1-4 — 1-6 as 

 long as the whole breast-bone and composed of two.short, robust, acute branches, 

 which are divergent at base but afterwards run parallel to each other, with an 

 appearance often of another lateral branch on each side. The stem of the clove 

 is straight, uniform in breadth except that it is minutely clavate at base, and 

 4 — 5 times as long as broad ; and the whole breast-bone is half as long again as 

 one of the middle abdominal joints. Feb. 26 and 28 the insect was still in the 

 larva state, and was then deep orange-color almost sanguineous : and a speci- 

 men occurred in that state even as late as April 23. 



