armatures in the Hi/mcnojjierotcs gcjius Colldes. 20 



settled by the future monographist of the genus, and to 

 him I leave it. Still I do occasionally mention points 

 which seem to me deserving to be considered when that 

 question comes up for definite settlement. 



Let us now consider in a general way (a) the concealed 

 abdominal plates in CoUetes $ $, and {h) the armature 

 proper. 



The concealed abdomincd plates. — Of these some are 

 dorsal, some ventral. They all lie, normally, enclosed 

 within the apex of the abdomen and immediately adjacent 

 to the " armature." To this, I believe, they both act as a 

 protective sheath, and also in some way co-operate with it 

 in its special functions (clasping, etc.). For these purposes 

 they are always extraordinarily modified from what we 

 must suppose to have been their original form, viz. that of 

 semi-annuli. And it may be noticed here, that at least on 

 the ventral side this sort of modification commences in a 

 rudimentary way even in the preceding (unconcealed) 

 segments. The sixth, especially, when mounted in balsam 

 and rendered semi-transparent, is seen to be far from 

 uniform in its structure and chitinization, and is often 

 tuberculated and foveated in a manner which probably 

 bears some relation to the forms of the organs underlying 

 it. (Of PI. IX, figs. 49, 50, 52, 53.) 



From between the armature and the seventh dorsal 

 plate may be extracted (though seldom unbroken) the 

 object which I figure in PI. IX, 40. This, following Dr. 

 E. Zander, I take to consist of a pair of dorsal plates — the 

 eighth and ninth of the abdomen reckoning from the 

 constriction, i. e. exclusive of the propodeum. I have not 

 found any specific characters in them ; and they are so 

 extremely fragile and difficult to study, that I confine 

 myself to figuring a single successful extraction of them. 



The concealed ventral plates are the seventh and eighth 

 of the abdomen (exclusive of the ventral plate, if such exists, 

 which answers to the propodeum — I have never succeeded 

 in finding any trace of one !). 



The eighth plate seems very similar in all species. (See 

 PI. IX, 44.) It has none of the paradoxical variations of 

 form which make this segment so characteristic in some 

 species of Prosopis, Andrena, etc. 



The seventh ventral plate, however, is a most interesting 

 and surprising object. Its base is a ribbon-like strip or 

 arch of chitin, joined at each end to the corresponding 



