Soyo Maru encountered a large school of saiiry and material from this catch 

 was supplied by investigator Gorozo Okamoto so that the author was able to 

 measure 60 of these fish on November 23, These results are assem.bled and 

 shovm in figures 10 and 11, and the following empirical formulas for 

 saury of \U - 39 cm body length have been obtained. 



For fish of August-September and November Jf^ 0.00 285/ ......(15) 



(the units here are gr for l/if and cm for / ) 

 For fish of August -September K= 0,13/ + 1.. 6 , (16) 



With respect to the K of the August fish (3.3 - 5.5), the K for the 

 November fish is scattered between 3.5 and 5 and is roughly the same, but 

 it is hard to recognize any regular correlation between K and / . The large 

 saury with an / of 27 cm or over, in which K has a tendency to diminish 

 regressively, are thought to be spent fish. 



August-September>2--0,16/-0.8 , 

 " c/^0,1 1-0,6 . 



" Z-1.1/ 



" s.O.U 1-1,7., 



in late November in the southern area (near Nojimaaaki) have a lesser 

 body depth in relation to body length, and the breadth and circumference 

 are also smaller so that it may be said that the fish are leaner, however, 

 the ratio of body length to total length is roughly the same in both 

 groups. Further, as shovvn in Figure 12 



Both Aug. -Sept. and Nov. fish (leaving out two deviations) 



K = 0,23Ac/+3.2 ..... (21) 



Both Aug. -Sept. and Nov. fish (leaving out two deviations) 



^= 0o32s-f 2.1 ..(22) 



Aug. -Sept. fish K = 27 5 d./8+ 2 ,0 ,.(23) 



November fish K= 275^/3+2. 2 .... (23 ' ) 



If in formula (23) we let /o'=-.l, then/-*- — • 275. ; and putting into formula 

 (23') the /0 = 1.05 derived from actual measurements, v;e get/-^ =.262.* 



For the correlation between / and cL-hJl and /J~'-'/( in the August- 

 September fish, a! = 0.0027 /f 0,073 and /?= 0,002/ + 0.025 show a manner of 

 increase with / which is scattered in the vicinity of a straight line. In 

 the November fish, however, they appear to increase somewhat with / but 

 the increase is not clearly marked. Also the correlation between s/A. and 

 / is not clear, and it appears to be scattered between 2,2 and 2.8 regard- 

 less of the / . The correlation between s// and / is also obscure, how- 

 ever, a tendency can be detected for it to increase somewhat with / , The 

 relationship between K and -s// can be shown as 



* The actual measurement of P was done according to the same method used by 

 Kimura ^3) on sardines. Thanks are due to r*1r, Kinosuke Kimura for much 

 valuable advice and the loan of measuring instruments for taking this 

 measurement. 



22 



