446 



BULLETIN OF THE BUREAU OP FISHERIES 



The specimens from the two lakes are of approximately the same average size, and 

 the figures given are comparable, therefore. It appears that the Ontario form has 

 many more gill rakers, much shorter paired fins, and a somewhat shorter head. Other 

 characters, as number of scale rows around the body, number of fin rays, and mandible 

 length, are approximately as in the typical form. The Ontario representative tends 

 to have, on the average, longer gill rakers, a shorter dorsal, a broader caudal, and 

 blunter head, but these characters are so variable within each race that no further 

 account of their variability is given than is found in the detailed comparison of 10 

 specimens from each lake in Tables 49 and 55. 



The form appears sufficiently well marked to merit a name, and I propose to 

 call it orientalis. Specimen No. 54064 of Table 55, taken on July 19, 1921, off Wilson, 

 N. Y., in 65 fathoms is designated as the typp. It is catalogued as No. 88352 in the 

 United States National Museum. 



The color in life is not different from that of the typical form. Alcoholics show, 

 on the average, more pigment on the head and body and on the abdominal fins, 

 especially on the ventrals and the anal. 



At least the males of the species develop pearl organs in the breeding season. 

 Specimens collected in the latter part of July, 1921, off Wilson, N. Y., had traces of 

 pearls, and specimens taken in early September off Oswego, N. Y., showed well- 

 developed pearls. For the most part these have been lost by friction in the preserved 

 specimens, so that no detailed description is possible, but in general they are dis- 

 tributed over the head and body and have the same general shape, size, and location 

 on the scale as in other members of the genus whose breeding dress has been described. 



VARIATIONS 



Racial variations. — No differences are observable between the groups of speci- 

 mens collected from the various parts of the lake, but it is not improbable that, if 

 sufficient numbers were gathered together, local races might be differentiated. 



Size variations. — Most of the collected specimens have been gilled in 23^-inch 

 nets and are therefore longer than 200 millimeters, so that there are no groups of 

 specimens of different sizes available to determine how the body parts change in 

 size with growth. 



Four specimens as small as 148 to 177 millimeters long have been seen, and all 

 had maturing gonads. 



COMPARISONS 11 



Kiyi can be confused only with Tioyi or possibly nigripinnis jrrognathus. Kiyi 

 and Jioyi attain about the same maximum size and resemble one another rather 

 closely. The shape of the body, as seen from the side, is decidedly less elliptical in Tciyi, 

 as the predorsal contour rises rather abruptly from the occiput and the head is rather 

 more elongated. Absolute differences in characters that can be expressed numerically 

 are wanting, but there are several characters that show average differences, and by 

 the use of these most specimens can be identified properly. 



» Figures in this section are given for all specimens collected except those of proportions for arledi, which are based on speci • 

 mens 225 millimeters and more in length. 



