110 Mr. E. Saunders on the terminal ventral 



It is, I think, quite clear that the abdomen of an 

 aculeate hymenopterous insect should consist of nine 

 segments. Packard, in the ' Proceedings of the Boston 

 Society of Natural History.' vol. x., points out that of 

 the fourteen segments of the larva (counting the head as 

 one), the first represents the head in the perfect insect, 

 the next four the thorax {i.e., including the 5th larval 

 segment, which he shows is taken up into the metathorax 

 during the larval changes), and the remaining nine the 

 abdomen ; Packard also points out that the male geni- 

 talia are visible in the larva as three pairs of tubercles 

 on the sternal portion of the 9th ring, which 9th ring 

 no doubt represents what we call the genital armature 

 in the imago. Eight segments now remain to be 

 accounted for, and all these, I find, present in such bees 

 as I have examined. 



The form of the 8th ventral plate in the male varies 

 very greatly, and appears to afford excellent characters 

 in some genera for sectional purposes. It is this 8th 

 ventral plate which projects somewhat in the form of a 

 spoon at the apex of the abdomen in Andrena, and 

 which has been considered as the 7th by many authors ; 

 the 7th, however, in this genus lies under the 6th, and 

 is rarely, if ever, visible ; the same form of the segment 

 occurs in Panurgus, Cilissa, Dasypoda, and Macropis. 

 In Paniirgm all the ventral segments are visible, as the 

 apex of the 6th is largely and widely emarginate, allow- 

 ing the 7th to be seen through the emargination. In 

 Megachile, where only four ventral segments are visible, 

 the 8th is a mere tongue-like plate lying close to the 

 under side of the genital armature. In our other British 

 genera the 8th is almost always hidden, but in Prosopis 

 it is hidden in nearly all the species, and very con- 

 spicuously exhibited in one (liyalinata). On the Conti- 

 nent there are two or three others which belong to the 

 same section as hyalinata, but the section is a very small 

 one compared to the great number of species known. 



In Prosopis the forms of the 7th and 8th ventral plates 

 are most curious and interesting in all the species, and I 

 have figured these segments of each to show how charac- 

 teristic they are. As will be seen, they are perfectl}' dis- 

 tinct in shape, and differ to such an extent in some cases 

 as to make one doubt whether all the species really belong 

 to one genus. The great similarity, however, of the general 

 structural characters throughout the species makes me 



