32 The Graspedosomatidse of North America. 



members of this family makes the study and preservation of the 

 dissected parts difficult unless they are mounted on microscopic 

 shdes. This method has the further advantage of allowing corre- 

 sponding parts to be viewed at the same angle, so that fictitious 

 differences are not multiplied. Some genitalia are, however, so 

 irregular in shape as to make it difficult to determine a species 

 from one drawing alone. Descriptions of genitalia are nearly 

 worthless without diagrams to aid in their interpretation. 



In describing the antennae we have adopted the form of a con- 

 tinued ratio for showing the comparative lengths of the joints. 

 The third joint, the longest, is taken as the standard, and given 

 the value 10, thus the ratio of each joint to an}- other can be 

 readily seen. 



We have not attempted to describe the immature stages of the 

 species, since the material was not collected with a view to such 

 study. Young individuals of this family are less commonly 

 found than in the others, and are usually so delicate as to need 

 special care in preservation. Such observations as we have been 

 able to make have been entireW in the line of those of Latzel. 



In the direction of describing forms and varieties we have done 

 equally little. This is partly due to the fact that our material 

 has been neither fresh nor very extensive, so that variations of 

 color and external characters have not been seen to advantage. 

 It is also due in part to the fact that we consider that in Diplo- 

 poda, as in Arachnida and Crustacea a different plan of male 

 genitalia is to be looked upon as of specific importance. There 

 is not in the habits and habitats of the Diplopoda apparent reason 

 for the multiplication of external differences, but the differences 

 of genitalia and secondar}^ sexual characters are not therefore less 

 important when they occur, but rather more so. Until we have 

 a recognized criterion of what a species is, the matter of specific 

 classification must remain largely a question of convenience. 

 The characters of the genitalia indubitably have important bear- 

 ing on the question of specific distinctness. The ultimate value 

 of every difference, as a means of specific definition must of neces- 

 sity be settled by ascertaining the constanc^^ of the particular 

 character in the forms to be separated. As a practical method 

 we have proceeded on the assumption that species with evident 

 and readilj^ assignable differences in the genitalia should be con- 

 sidered distinct until the differences are shown not to be constant. 



