80 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 122 



cated that this variation was geographically, sexually, and ontogeneti- 

 cally correlated. Different samples from a relatively limited geo- 

 graphical area (i.e.. Phoenix Islands) may have quite different relative 

 cirrus lengths. Within most populations males tend to have rela- 

 tively longer cirri than females. Relative cirrus lengths of specimens 

 from India do not overlap with cirrus lengths of specimens from 

 Samoa; however, plotting lengths for other populations of the species 

 show that the cirri lengths of some populations overlap both the 

 Indian and Samoan populations. No clinal change (east to west, 

 north to south) in relative cirrus lengths was noted. In conclusion, 

 it appears that the character of main supraorbital cirrus length is 

 affected by a complex of factors and can be expected to vary con- 

 siderably from one locality to the next. 



Variation in average number of segmented dorsal or anal fin rays 

 occurs among the various populations of E. striatus. Such variation 

 prompted Schultz and Chapman (1960) to recognize subspecies within 

 this species. In an attempt to determine if the variation followed any 

 geographic pattern, I arranged frequency distributions of the counts 

 on a latitudinal north to south basis (regardless of longitude) within 

 the three major areas where the species occurs (tables 33, 34) : Pacific 

 Ocean, South China Sea, Indian Ocean. Inasmuch as the method of 

 counting anal rays generally employed in the present study (method A) 

 might obscure significant counts of the anal fin, I have also given 

 frequency distributions of the total segmented anal elements (method 

 B). 



In the Pacific Ocean the highest average fin ray counts were ob- 

 tained for populations south of the equator. A trend of increasing 

 averages for anal rays, both methods, from north to south is par- 

 ticularly noticeable if averages are taken of the counts grouped by 10 

 degree latitude intervals; dorsal rays appear to increase in both 

 directions from the equator: 



The validity of grouping the data in the above manner is not 

 known, but perhaps it is suspect because the fin ray counts grouped 

 north to south by 10 degree longitude intervals present no clear-cut 

 trend within a particular interval. The difficulty is perhaps due to 

 the fact that there is insufficient material from within the intervals 



