558 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.46. 



As the work progressed it became more evident that many of the 

 characters used diagnostically were very superficial and subject to 

 considerable variation. In many instances very closely related in- 

 dividuals had to be placed in widely distinct genera, and, again, 

 several specimens placed in as many different genera by whomever 

 had worked on the collection before me were found to belong without 

 a doubt to the same species. Such instances as these naturally 

 forced me to the conclusion that other characters must be found 

 which are more constant. 



I suppose that every careful student of this and related groups has 

 been impressed with the same fact and several have attempted a 

 solution of the difficulty, but most have merely followed along the 

 same ruts and expressed the hope that somebody would take the 

 time to work out the true relationships. Having these very extensive 

 collections, which included many known and new species from North, 

 South, and Central America, as well as a considerable number of deter- 

 mined species from Europe, I have been able to work out the rela- 

 tionships a great deal more accurately than one could with only a 

 small collection representing a less extensive distribution. In view 

 of this fact, I have ventured to rearrange the grouping of genera and 

 species to a considerable extent, in the hope that it will prove to be 

 more usable both by the expert and the beginner. 



As was stated above, there is great homogeneity in the family and 

 it is, therefore, difficult to find characters which are both constant 

 in subgroups and at the same time usable generally. The trend of 

 the lateral keels of the pronotum, for instance, has been used exten- 

 sively to separate two large groups of genera, but this trend is variable 

 in a great many of the species, and certainly in many genera at least, 

 and, moreover, is very difficult to appreciate. At one angle they ap- 

 pear to have one trend and at another they have another trend. 

 James Edwards, Mr. Distant, and the late Mr. Kirkaldy, I beUeve, 

 were right in disregarding this as a good character. Again the shape 

 of the outhne of the vertex and frons has been used very extensively, 

 but this, also, is difficult both to describe and to appreciate. The 

 last-named characters, while they are in the main unsuited to generic 

 separation, may in many cases be good secondary specific characters. 

 The venation of the forewings (elytra) is so variable, even in the two 

 mngs of the same insect, that it can not be used with any certainty 

 at all. 



The object in view, of course, is to find characters which are constant 

 within certain reasonable limits, and, also, may be appreciated by 

 other students besides the one who first uses them. The form and 

 relative lengths of the two antennal segments afford, within certain 

 limits, a usable character. To distinguish genera, however, on a small 

 difference in length of similarly shaped antennse is wholly without 



