﻿NORTH AMERICAN NOTOSTRACA — LINDER 59 



the impression that (Hfferent populations might be characterized by- 

 different nmnbers, but this is by no means statistically certain, and 

 I am convinced that this problem can be solved only by studying 

 large lots. Also, we must not forget that the number of legless rings 

 is quite a dubious character: it is the result of the interaction 

 of the number of leg-bearing rmgs and the total number of rings. 



Size and form of the telson are worth attention because this has 

 proved a reasonably good character in Lepiduru.'^. As did Packard, 

 I have found some variation in this structure. However, the various 

 forms present a reasonably continuous series, and I have not been 

 able to discover any correlation with other characters. As the speci- 

 mens in one lot are usually similar to each other as far as the telson 

 is concerned, perhaps it ought to be determined whether the differ- 

 ences can be of some use in distinguishing lower taxonomical units. 



a bcde fghi 



'l II II II II I J II II II 



^^^ Sp„^^ Sp..- 



•? 1 12 1 1 92 1 1 5 114j 



9 d* 2 16 1 1 20]^ 



2? 3 14 17] 5^ 



3.t ' ' -H 



5 



3 



? 



*• ? 1 



Figure 29. — Patterns (a to i) of dorsal central spines on the telson in representative specimens 

 of Apus longicaudatus LeConte from: (1) Near Aurora, Wyo. (U.S.N.M. No. S'i766); (2) 

 Moores Lake, Lubbock, Tex. (U.S.N.xM. No. 617S0); (3) GalSpagos Islands (U.S.N.M. 

 No. 84240); and (4) Kansas (Stockholm Mus., types of A. lucasanus Packard). M, dorsal 

 hind margin of telson; Sp, well-developed spine; Spi, rudimentary spine; Spu, short spine. 



Finally, in considering the pattern of the central dorsal spines on 

 the telson, we must remember that Pearse found a considerable 

 variation here, and that Rosenberg (1947) found the pattern variable 

 in his material. Aly studies have shown that the variation presents 

 a regular series with many intermediate patterns between the most 

 common ones, so that in a sample with manj'- specimens almost all 

 patterns may be represented. 



We see from figure 29 that some patterns are more common than 

 others. These patterns were given the value of specific characters 

 by Packard (1883); he assigned patterns a and b to Ajpus newberryi, 

 pattern b to A. aequulis, pattern c to A. longicaudatus, and pattern 

 f to A. lucasanus. Even in his small collection, however, significant 

 variations must have been known to him. He gives two different 

 patterns for A. newberryi, while the paratypes of A. lucasanus that 



