﻿10 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 102 



have 30-33 +i and the females 30+i-34 rings (fig. 14). The range is 

 also great in A. longicaudatus from Wyoming (U.S.N.M. No. 58766), 

 where the males have 41-44 and the females 39+i-43 rings (fig. 31). 

 These are relatively extensive samples, the former containing 54 males 

 and 42 females, the latter 33 males and 98 females. In smaller 

 samples the range of variation is not so great. The extremely high and 

 low numbers of rings in one sample are usually represented by rather 

 few specimens. 



No doubt the average number differs in various samples of the 

 same species. But we cannot reasonably compare samples with 1, 2, 

 or 10 specimens with samples containing a hundred or more specimens. 

 Small samples give very uncertain figures, and museum w^ork, in this 

 respect, is much hampered by the fact that the great majority of 

 samples contain a very small number of specimens. Statistical 

 calculations, however desirable, are thus rendered very difficult. 



There seems to be a general rule that species with a large number of 

 rings have a greater range of variation than species with few rings, 

 both in Lepidurus and m Apus. There are some exceptions to this 

 rule, but not many, m the material I have seen. The fewest rings in 

 Apus occur in forms that have a clear parthenogenetic tendency and 

 that show a very small range of variation. About these more will be 

 said later on (see p. 12). Similar forms belonging to a group with 

 relatively few body-rings are found among species of Lepidurus. 



Body-rings in the species of Lepidurus range from 25 to 34, those of 

 Apus from 30 to 44. In the former genus, the high numbers of 30-34 

 are represented by two species only, L. hilohatus and a new species 

 described in this paper, both of which differ from the other species of 

 the genus in the number of leg-bearing abdominal rings, and, most 

 evidently, in the number of legs. Apart from this grouping, and from 

 what has been mentioned about parthenogenetic forms, I cannot find 

 clear correlations between the number of rings and other characters. 

 I doubt the validity of Rosenberg's (1947) method of distinguishing taxo- 

 nomic units — and not lesser units at that, but species — by associating 

 a certain number of rings with a certain size of specimen, because 

 I think that further research is necessary to establish the conditions 

 of the variation in size, which in some lots is considerable, in others 

 rather small. 



It is usual for the males to have a higher total number of rings than 

 the females of the same species, though the rule has many exceptions. 



It is, of course, necessary to know whether the number of rings is 

 really fixed in adult specimens — non-adult stages are not considered 

 here — or whether it increases with an increase in size of the specimen. 

 I have attacked this problem, which has never before been investi- 

 gated, in two ways: first by comparing the number of rings in small 



