58 



BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. 



that the two are almost completely separated from each other, since indi- 

 viduals with 29 rays are very rare. It is also clear that the group with 

 fewer rays varies less from the normal number than does the other. It is 

 worth while therefore to examine the species separately (omitting the ob- 

 viously young) to bring out the difference in variability. As cumingii 

 and polybrachius are so closely allied, they may be considered together, 

 especially as there is no essential difference between them when tabulated 

 separately. We will omit multiradiatus altogether as the number of 

 available specimens is too few to make a reliable tabulation possible. 



The diversity in the number 

 of rays in kubiniji is remark- 

 ably slight and is clearly shown 

 in Diagram 2, from which it 

 will be seen that practically 80 

 per cent have 23-25 rays and 

 that nearly 69 per cent have 

 an odd number. In canopus 

 on the other hand (Diagram 

 3) only 48 per cent, have 

 23-25 rays, and only 37 per 

 cent have an odd number. 

 Although the small number 

 of specimens available for com- 

 parison undoubtedly accounts 



in part for these peculiarities 

 of canopus, it can hardly be 



doubted that this species shows 



a much greater tendency to 



variability in the number of 



rays than does kubiniji. In 



helianthus (Diagram 4) the 



but 64 per cent have 34-36 



In cumingii and 



50 2\ 2X 11 V\ 25 26 27 2S 11 



Diagram 2. 



To show the relative abundance per thousand, 

 of //. kubiniji with 21-28 rays. Based on 

 110 individuals. 



number of rays varies from 30 to 40 



rays, while only 46 per cent have an odd number. 



polybrachius (Diagram 5) the number of rays ranges from 29-40 x and 



only about 36 per cent have 34-36, while almost exactly half have an 



odd number. The great variability of these two short-rayed species 



is especially notable in view of the fact that microbrachius, which is 



also short-rayed, agrees strikingly with helianthus, 63 per cent of the 



1 The two specimens of polybrachius with 42 and 43 rays respectively are 

 omitted from the diagram. 



