from year to year. Group VI on the other 

 hand decreases. GroUpB III and IV and also 

 V do not show any clear tendency to increase 

 or decrease. Consequently we can see that of 

 the six size groups which appear in this fish- 

 ing ground the three near the middle do not 

 present any clear tendency to increase or de- 

 crease but the smaller size groups show a 

 relative increase while the groups of larger 

 size fish show a relative decrease. 



Of the three size groups nearest the 

 middle, in group III and group IV another con- 

 spicuous annual change can be seen. This is 

 the alternate year cycle appearing in the quan- 

 tity of occurrence of these two groups. Group 

 III occurs conspicuously in even-numbered 

 years and is scarce in odd-numbered years. 

 Group IV appears in small numbers in even- 

 numbered years and is outstanding in odd- 

 numbered years. Consequently they alternate 

 every year in predominance. 



Discussion of the Results 



Basically, in analyzing the size composi- 

 tion into a number of groups with different 

 characteristics, we should take as a staindard 

 the characters which they actually possess. 

 If we take as a basis only the form of the size 

 distribution and thus separate them, it i s 

 thought that some degree of subjectivity and 

 error will be introduced. Therefore the va- 

 lidity of the size groups separated by this 

 method should be exannined by comparing for 

 each character the results of the direct analy- 

 sis of the size composition with actual phe- 

 nomena. However, at the present stage we do 

 not have data which would enable us to make 

 such an examination for each character, and 

 furthermore it is thought that of the various 

 /possible^/ characters there is the greatest 

 possibility that the groups are based on age. 

 Consequently in what follows we shall try to 



examine one or two points in terms of age 

 groups .2.1 



a. In Japan, Uno {1936a, b) and Aikawa 

 (1937) and Aikawa and Kato (1938) have studied 

 the age of albacore by means of the vertebrae. 

 The following is the comparison of their results 

 with the spacing of the modal lengths. 



aj Comparison with the results of 

 Aikawa and Kato: 



The absolute values of the differences 

 can not be said to be large, A consid- 

 erable difference in the trend can be 

 seen. According to Aikawa and Kato's 

 results, the interval between each age 

 group and the next, with the exception 

 of that between age group I and age 

 group II, is 9 cm. Consequently, the 

 five intervals between age groups I to 

 VI are all about 2 cm. larger than the 

 values shown by Aikawa and Kato. 



Aikawa and Kato did not separate 

 the size composition directly by the 

 number of rings in the vertebrae, but 

 they used the indirect method of finding 



— Differences in size of the fish depending 

 on sex do not appear to participate in the for- 

 mation of the size groups. Figure 3 shows the 

 results of an examination of the sizes of fish by 

 sex carried out on a research vessel in March 

 1953. It is unavoidably somewhat rough because 

 it was not possible to carry out an investigation 

 of age at the same time, but at any rate between 

 70 cm. and 100 cnn. no marked difference can 

 be seen between the modal lengths of males and 

 females. Since the intervals between the six 

 size groups are 10 cm. or more, it is not 

 thought that these can be based on differences 

 associated with sex. 



20 



20 



90 



LENGTH ( CM ) 



Figure 3. --Length frequencies of males and females. 



11 



115 



