240 DISCOVERY REPORTS 



the existence of this process." I agree with Hansen, and I beUeve that Groups a and d 

 are natural groups but that Groups h and c are not. 



The best criterion for separating the species of Eiiphamia is the copulatory organ of 

 the male. Those of Group a and those of Group d each show some characters which no 

 others have. 



The members of each group are shown in the table below. 



GROUP a 



Hansen has figured the copulatory organs of every species of Group a, and I have 

 seen those of E. (imericana, E. recurva, E. brevis and E. diomedeae in specimens from the 

 Discovery collections. In every one of the seven species the terminal process has near 

 the end a slender thin-walled finger-like secondary process such as is not found in any 

 other species of the genus. There is nothing so distinctive in the proximal process, but 

 there is a general similarity between those of all : they have heavy bases and are more or 

 less strongly curved ; the distal ends are expanded and flattened, and the beginning of 

 the expansion on the inner and hinder side is in each abrupt and marked by an angle 

 or protuberance. The lateral process has no secondary tooth or spine. 



It is interesting to consider some of the other characters by which these species are 

 distinguished — the structures on the antennular peduncle. The lobe on the first seg- 

 ment is either pectinate or bifid (except that it is acute in the male of E. recurva). The 

 second segment has on the distal end of the dorsal surface a pair of tubercles or spines 

 one at each corner, a tubercle on the outer corner only, or no tubercles. There is no 

 constant correlation between these two sets of characters, those of the first and those of 

 the second segment, as the following table shows. 



From the fact that these seven species have the peculiar and distinctive characters of 

 two pairs of lateral denticles on the carapace and of a slender thin-walled secondary 

 process near the distal end of the proximal process of the male copulatory organ, I con- 

 clude that they form a natural group within the genus. Because in this natural group 

 the structures on the antennular peduncle — the lobe on the first segment, the spines 

 or tubercles on the second — show considerable variation I think that for the purpose of 



