88 HYMENOPTERA. 



thicker than a good sized pin, to the largest which are of 

 several times their bulk. Likewise in puncturation a wide 

 range of variation is found, which will be discussed later 

 under the species. In this series four or five forms could be 

 selected which would undoubtedly stand as distinct species 

 in the absence of the rest, but in the complete series their 

 identity is utterly lost among the intermediate forms. Out 

 of this group several species have been described and have 

 stood as such for years, but they are unquestionably onlj'- 

 forms or varieties. In two or three other species of wide 

 range this same condition exists, causing great confusion. 

 As might be expected the most variable species are those 

 occurring in the greatest numbers, and over the greatest ter- 

 ritory P. cressoni and P. modesties being striking examples. 

 These two species include a relatively large part, (about 

 one-fouth) of the collections as a rule ; and the fact that they 

 show such variability makes it more probable that other 

 species will show a similar tendency when collected in large 

 numbers, and when their entire range is worked over. 



At present very little can be said about the distribution 

 and ecology of the genus. So little material has been ob- 

 tained, and this from such small and widely separated locali- 

 ties, that only an infinitesimal part of what will be necessary 

 for adequate treatment, has been collected. The shaded 

 areas on the accompanying map are intended to show ap- 

 proximately the ground covered up to the present time, and 

 to emphasize the fact that we know comparatively nothing 

 of the genus. The shaded areas cover many times the re- 

 gions actually worked, which would be better represented 

 by mere dots. Before any really comprehensive work can 

 be done, then, large collections must be built up, including 

 series from intermediate localities. The importance of col- 

 lecting in large series cannot be too much emphasized, for 

 it is only in this way that any reliable and usable knowledge 

 of our bees can be obtained. At present only two of our 

 species of Prosopis are known reasonably well, and these are 

 P. cressoni and P. inodestus. Of the former I have about one 

 thousand specimens, and of the latter about two hundred and 



