100 PROCEEDINGS OF THE NATIONAL MUSEUM vol. m 



made on specimens of either sex by the use of quahtative external 

 characters as pronounced as those utiUzed to distinguish species in 

 Coleoptera, Orthoptera, and other insect groups. 



Cranium: In nearly all of the species examined, the head is com- 

 pletely smooth and pohshed except for the normally concealed surface 

 of the vertex, which tends to be finely rugose. The eurelines, however, 

 often develop faint sculptm-e on the lower part of the head, the sculp- 

 ture being well developed as genal striation in Centrelus and thus 

 providing a means for distinguishing this genus from the closely 

 related Eurelus. 



Clypeal foveglae: In aU the species of the family, the total range 

 of variation in this character is from 3-3 to 5-5, and this range may be 

 observed in a single species. There is no defensible reason for the 

 erection of species upon shght variation in this character. In nearly 

 every case, comparison was made between a single specimen of an 

 allegedly new species with a single specimen of another. The tacit 

 assumption that the clypeal foveolae do not vary ^\^thin a species 

 indicates nothing more than unfamiliarity with the group. 



Ocelli: In a general way, the number of ocelli tends to be charac- 

 teristic of the subfamilies although there is some overlap. Species of 

 the Eurelinae usually have 40 to 50 in each ocellarium (a new term 

 to replace the incorrect "eye" and the awkward "ocellus patch"), 

 and those of the Arinohnae have from 30 to 40, but the variation in 

 a single species is usually as great as for the entire subfamily. We 

 have detected no significant variation in size, shape, or arrangement 

 of the ocelli in species or genera. As a rule, the ocelh are difficult to 

 count with accuracy. We have found that the only way to be abso- 

 lutely sm-e of the number is to boil the cranium in strong KOH or 

 NaOH until the exoskeleton is rendered colorless and the ocelli become 

 very sharply defined. As a ruJe, the smaller species have smaller 

 ocellaria, which are correspondingly more widely separated, and it 

 may be entirely possible to take accurate measm-ements and work out 

 ratios that would be of diagnostic importance. Such refinements, it 

 seems, mxy profitably be postponed until an adequate volume of 

 material is available for analysis. 



Antennae : The antennal articles are generally similar in size, pro- 

 portion, and vestiture throughout the family, with the second nor- 

 mally the largest, followed by the sixth, third, fourth, fifth, seventh, 

 and first. In the Eurelinae the antennae are shortest in proportion 

 to the size of the animal, and usually do not reach beyond the caudal 

 edge of the collum when straightened and extended caudad. The 

 smaller species, such as those of Arinolus and Onychelus, have much 

 longer antennae; these extend back to the third segment, and the 

 individual articles are much longer in proportion to then* width. The 



