48 CASEY 



stant throughout, so that even Lacordaire, generally so happy in his 

 choice of primary characters, has, seemingly in despair, seized upon a 

 few antennal features for major group division, without testing them 

 with sufficient thoroughness; the result is that the grouping suggested 

 by that author, especially that founded upon variations in the ridge 

 above the antennal foveae and in the pores and fossae of the outer 

 antennal joints, is not satisfactory, as indeed remarked by LeConte 

 in his monograph of the family. At the same time the group char- 

 acters adopted by LeConte are also valueless to a great degree, for 

 it is not apparent that the front is narrowed by antennal insertion to 

 any greater degree in Chrysohothris than in Gyascutiis, and the nature 

 of the meso-metasternal suture, although very important, completely 

 fails as a group character in the sense intended by that author, while 

 the differences in prosternal conformation are too feeble, inconstant 

 and illusory to serve any useful purpose. It is perhaps also due in 

 part to this lack of evident organic structural diversity, that early mis- 

 takes in generic assignment were so frequently made by systematists; 

 but it is difficult to comprehend, even under this assumption, the 

 original referring of what we now know as Gyascutus planicosta and 

 ohliteratus, as well as many exotic species, to the genus Chalcophora. 

 It has been impracticable for me to survey the entire family with 

 sufficient thoroughness to definitely pronounce upon primary group 

 division, if there be any worthy of adoption, and the genera here 

 reviewed are therefore arranged in close succession in a single state- 

 ment and without tribal delimitation. The principal generic charac- 

 ters assumed relate almost wholly to antennal and elytral structure, 

 but, for more closely related genera, diverse characters of the front, 

 pronotum and tarsi have proved to be of more or less utility. There 

 are two sufficiently distinct types of antennal structure among the 

 genera having smoothly anchylosed sterna, and, if found to be of gen- 

 eral application in the family at large, there is no reason why they 

 should not form the basis of major group or tribal division. In the 

 first group we might suggest Hippomelas and allied genera, having a 

 terminal antennal process, as one tribal group, Psiloptera, Spinthoptera 

 and probably others now confounded with Psiloptera, as another, 

 Chalcophora and related forms as a third and Buprestis with its im- 

 mediate allies as still another, and, in the second division, which is 

 probably by far the larger, Dkerca, with related genera, should be 



