140 THE Is^ATURAL niSTORT EEYIEW. 



" 7. Viewed laterally, tlie upper edge of the second abdominal 

 joint (counting the peduncle as the first joint) describes a circular 

 arc of about 60°, in both forms. Taking the chord of this arc as a 

 definite and permanent basis of measurement, in spongijlca P the 

 lower or ventral edge proceeds straight downwards, exactly at right 

 angles with this chord, for a distance equal to half or one-third the 

 length of the chord, before it curves gradually backwards, to form the 

 ventral arch. In aciculata $ on the contrary, instead of being at 

 right angles (90°) with the chord, it forms with it an angle of about 

 11.0°, so as to exhibit a most extraordinary bulge in front, and it 

 curves much farther downwards from the peduncle, and in a more 

 compressed and knife-edged form, so that the abdomen is vertically 

 at least as wide as long, and almost always much wider, whereas 

 in spongiJiGa j it is always longer than wide, generally much 

 longer. The above variation in each form is caused by the terminal 

 abdominal joints being more or less telescopically drawn out in 

 difierent specimens, so that in each form the second abdominal joint 

 sometimes occupies dorsally half the entire length of the abdomen, 

 exclusive of the peduncle, and sometimes almost two-thirds. St. 

 Eargeau has observed the same thing of the genus MegachilCy 

 (Hymenopt II., p. 338,) and I only notice it here because Osten 

 Sacken, having only a few specimens of each form on hand, supposes 

 the relative length of the second abdominal joint with regard to the 

 terminal joints to be a terminal character of each (Proc. Ent. Soc. 

 Phil. I., p. 246.)" 



" 8. In consequence of the above bulge on the anterior abdomen 

 in aciculata, (see Appendix, Pig. 1) the distance from the " ventral 

 valve" (Fig. 1, v.) to the "dorsal valve" (Pig. 1, 7) is proper* 

 tionately twice as long as in spongijica, and consequently the sheaths 

 of the ovipositor (Pig. 1, s.s.) are also proportionally twice as long, 

 though their proportional breadth in both forms is nearly the same." 



" 9. With the exception of a single specimen, my 30 $ aciculata 

 are one quarter broader and longer than my 5 P spongijica and my 

 9 P inanis, all 14 of which are remarkably uniform in size, save a 

 single >? inanis which is a little smaller than the rest." 



" These nine differences are sufiiciently remarkable, and, but for 

 "the evidence of dimorphism, would," as Mr. "Walsh truly observes, 

 " undoubtedly be viewed by every entomologist as of specific value." 



Mr. AValsh's reason for regarding these two forms as belonging 

 to the same species is as follows: — In May and June, 1863, he 

 gathered a niy&iber of galls from a black oak (Q. tincforia), some of 



