INTKODUCTION. 



The priinarv division in this classiticatiou has reference to tlie 

 number of visible abdominal tergites ; in the Allociastra tlie males 

 are considered to have «, and the females 7 ; while in the two other 

 groups both sexes have 7. Apart from the obvious inconvenience 

 of utilisiug for such a purpose a distinction which is confined to 

 one sex, it has been pointed out above that there are some im- 

 portant exceptions which were unknown to Leconte and Horn. 

 In various genera falling within their Otiobrhynchijd.e the 

 females have eight visible tergites and not seven (see p. 14); 

 while on the other hand I have observed several cases of males 

 with seven tergites only in genera which would fall into their 

 restricted CukculionidyE, namel}^ Bcujom, Metkdma and an uu- 

 described genus near Acallojnstas; and in this group also there 

 are two African genera the females of which have eight tergites 

 instead of seven, namely, Omoj^jhonis and Nerthops ; again in 

 examining half-a-dozen species of Cossonin.e, which are supposed 

 to have seven tergites iu both sexes, I have found an undescribed 

 African species which has eight in the male, and an Indian 

 species, Brachyiemnus cq^hahtes, Est., which has no more than 

 six; finally, Bedel has already stated tl-at the males of Ajnon 

 have eight, and not seven tergites. That the distinctions noted 

 by Leconte and Horn are of considerable taxonomic importance 

 cannot be doubted, bub it is clear that their real significance 

 cannot be projjerly appreciated until a much larger number of 

 Curculionid genera have been examined for the purpose. 



The secondary character upon which the three main groups 

 have been founded is also subject to serious exceptions. Although 

 the Haplogastra are stated to have no inferior carina on the 

 elytra, in many of the larger Old World lihyncMtes and most 

 Attelahus this " carina is distinct, and sometnnes as strongly 

 developed as iu any other CurouliojS'ID-E. On the other hand, 

 in many wingless Allogastra, such as Bi/rsops and Eremnus, 

 the carina is entirely wanting, while in others, such as Ejnsomus, 

 Stropliosonms, Amphitmetus, etc., it is often quite rudin)entary. 



In these circumstances it seems hardly probable that a com- 

 bination of the two foregoing characters would lead to a reliable 

 indication of inter-relationstnps, which is the object of all classi- 

 fication, and this is exemplified by a comparison of the two 

 divisions Allogastra and Hexerogastra. Here we find such 

 groups as the ApionidJ'. and Calakdrid.e widely separated from 

 X\\e CuRCULioNiDyE, with which they are closely related, a,nd 

 intimately associated with the Anthribid.e, a family with which 

 they have obviously nuich less afiinity. Moreover, the proposal 

 to subdivide Lacordaire's family Curculionid.e into eleven 

 families is open to the criticism that the resulting groups are 

 clearly quite disproi)ortionate ; in other words, the subdivision 

 has been carried either too far, or not far enough. 



As already indicated, it appears more satisfactory to retain 

 Lacordaire's'conception of the family, and for the purpose of this, 

 work the main lines of his classification will be followed. 



