A) Sexual dimorphism. The data on which the present study is based were obtained from the records of the Hokkaido Fisheries Experimental Station concerning four-year old dog salmons, consisting of 151 males and 116 females, captured in North Kuril Islands from June 11 to July 30, 1932. The results are given in Tables 1-4, and may be summarized as follows: -(1) Males are larger than females in size. (2) The ratio, body length/body height, in female is a little greater than that in male, and two ratios, body length/head length and head length/ snout length, are in female considerably greater than those in male. This is probably due to the well known change in shape of males in the spawning season. (3) Correlations between body length and body height, body length and head length, head length and snout length, and head length and height of caudal peduncle respectively are comparatively high in both sexes and these are without exception higher in male than in female. Correlation coefficient between head length and snout length in male is the greatest of all. The partial orrelation of head and snout length in male, holding the date of capture constant, is much higher than the ordinary correlation between these characters, since there occurs the prolongation of snout in male during this period.
B) Races of dog salmon. According to S. KAWAKAMI the dog salmon in Hokkaidô is to be grouped into two distinct races, Pacific and of Japan Sea, which latter may be subdivided into two races. For the distinction between the last two, he has given means and modes of branchiostegals, developed dorsal fin rays, anal fin rays, gill rakers, and pyloric eoeea. The results of the x2-test applied on his data, po ?? nted to the conclusion that according to all characters, except pyloric coeca, the two groups are really differentiated samples, and they may be safely separated into two races.
