596 Mr. G. J. Arrow ori 



meet in a sliarp angle at their extremities, and, in tlie males 

 at least, completely hide tjje abdomen as seen from above. 

 The visible part of the mesoiiotum is smooth and sliining, 

 Avith or without a cluster of minute punctures on the 

 scutellum, which is very short awd broad ; the prosternal 

 process is rather long and erect, and in the male the penulti- 

 mate joint of the front tarsus is produced beneath the last 

 joint into a slender rod almost as long as the rest of the 



j«i"f- 



In the second division, represented by H. ttnicolor, Oliv,, 

 the elytra are abbreviated and separately rounded behind (or 

 at least their posterior edges do not form a continuous line), 

 and the end of the abdomen is almost always visible from 

 above. The exposed part of the mesonotum is entirely 

 opaque and finely sculptured, the scutellum forming a well- 

 marked lobe, the prosternal process is triangular and very 

 short, and the male has only a short prolongation of the 

 penultimate joint of the front tarsus or none at all. 



The investigation of the genitalia of male D^-nastinai 

 fiequently reveals striking differences in species which have 

 a close external resemblance and is an indispensable criterion 

 in cases of difficulty. The investigation in the present case 

 has produced rather surprising results. The species of the 

 first division are entirely unlike the remaining species in the 

 form of the oedeagus, while showing only very slight 

 differences between themselves. The paramera are long and 

 slender and the left one (as seen from behind) has an over- 

 lapi)ing internal flange. 



In the second division the oedeagus is symmetrical and 

 the paramera shorter and very markedly different in shape 

 in the three species, although externally these are no more 

 dissimilar than those of the first division. 



Before making a careful study of this organ I regarded 

 some of the forms of the very variable H. unicolor as specifi- 

 cally distinct, but the uniformity of their genitalia as con- 

 trasted with the dissimilarity found in the three species just 

 mentioned obliged wc to abandon that idea. On the other 

 hand, to treat very slight diftereuces in the form of the 

 oedeagus as specifically insignificant would im[)ly the non- 

 validity of such (externally) very distinct forms as H. patella 

 and H. latissimum. Thus this structure not only provides 

 confirmation of the naturalness of the primary subdivision of 

 the genus based upon external characters, but is the only 

 really trustworthy criterion of species. 



I have therefore been impelled to accept any tangible 

 and ai)parently constant degree of diftorcnee in this respect 



