Longicornia Malayana. 6 



additional sub-tribes instituted. These are the " Hippopsitse" 

 from the " Oncideritas, " and the " Tapeinitoe " from the 

 *' Saperditae." The two principal things that strike us in tliis 

 arrangement are, first, that no provision has been made for tlie 

 Dorcadion forms in any of these sub-tribes, — perhaps, beca<ise they 

 do not occur in the Amazonian fauna, — and the second, the placing 

 of the " Compsosom'Uce" in the same sub-tribe with the ^' Desmi- 

 phoritcB," a position wholly unaccountable from anything I can 

 gather in the description. Mr. Bates lias not, however, over- 

 stated the case when he says, that " it is a matter of great diffi- 

 culty, perhaps impossibility, to find constant characters for the 

 subordinate divisions ;" and therefore it must not be expected 

 that systematists will always agree in their disposition of par- 

 ticular forms in a group so confessedly difficult as the Lon- 

 gicornia. 



This difficulty arises from the great diversity of forms, and 

 the gradual modifications of even tiie most important organs in 

 nearly connected species, so that if we were to insist strictly on 

 definitions, we should add almost indefinitely to the genera, and 

 many of these again would have to be raised to the rank of sub- 

 families. It frequently happens that we can only judge of the 

 value of a character when we are able to follow it up into allied 

 forms : an apparently excellent generic diagnosis, drawn from 

 a single isolated species, may break down altogether when 

 there are half-a-dozen to be dealt with ; and in the case of 

 higher groups than genera, we shall often find beyond its typical 

 members a tendency only to the character which more than any 

 other affects that particular group, or in other words the cha- 

 racter becomes modified, or is lost altogether. In each of the 

 sub-families of the Longicornia, I have noticed that there is 

 generally one preeminent character, and other characters, that 

 elsewhere were of primary importance, then become of secondary, 

 or, perhaps, only of specific value. The absence of humeral 

 angles and, consequently, of wings in the Dorcadion group, for 

 instance, being of this special importance, we find other characters 

 so subordinate as almost to startle us at meeting the most oppo- 

 site of them in apparently nearly allied species. 



In the arrangement of my collection, I have come to the 

 conclusion, that there are at least twenty-four types among the 

 Lamiidae which can be clearly demonstrated. It is true, that in 

 no case, after we have referred the various genera to their respec- 

 tive types, can any of them be distinguished by any absolute, 

 much less by any series of characters, but by an examination of 



