simultaneous oceanographic observations taken every year in August, and 

 the approximate positions of the fishing grounds have been entered on the 

 charto From this it can be seen that the grounds lie within several 

 hundred miles of the coast at places where the temperature difference 

 between the upper and lower strata is comparatively greats and in the 

 vicinity of projections of the warm current system at; the 100-meter 

 level (tending somewhat to center on the west sidejc 



0)-^s f (OpS) - If we consult the statistics to see the correlation 

 between skipjack catch and salinity, we findj as shown in Figure 5, that 

 the fish can be taken within a range of ^i* of 33=35 /6o , however, the 

 maximum catch frequency for both 1936 and 1937 appears where J" is 34.o6- 

 34o8 /6o c In this regard, however, there are few data and accuracy is not 

 high. There is a need for an investigation by areas and by seasons 

 based on a larger volume of accurate datao 



An interesting point to note here is the correlation, shown in Figure 

 6, between the temporal changes in the landings of skipjack jV at Makurazaki 

 in Satsunan for each ten-day period of the years from 1933 to 1937 and the 

 changes in chlorinity S' taken by hygrometer at the Yakushima lighthouse „ 

 In this case the curve ofyiAand the curve of ^ agree rather well in their 

 general tendency during the period from the middle of March and the latter 

 part of September, showing a true correlation, and the variations in 

 maximal points of /^ and S from year to year come out in rather good agree- 

 raentc Particularly well matched are the curves for the period of the 

 decline of yV and S from thei.r maximuTn value Se 



The origin of this phenomenon is presumed to be the spreading of 

 water of low chlorinity from the East China Sea system in the surface 

 layer of the Kuroshio system in June, July, and August. Through the 

 autumn and winter the S again increases and the temperature also reverts 

 to its spring level, but the fact thatvVdoes not increase in proportion 

 is probably due to the fact that it is not yet time for the schools, which 

 were broken up during the summer, to reassemble and they are probably 

 engaged in a migration related to their spa'/vning season^ 



{L) J^ z j i^s^) ' li we attempt to examine the relation between the catch 

 and transparency, we find, as shown in Figure 7, that the most fish are 

 taken at transparencies of H • 28m and that the maximum catch frequency 

 appears in the vicinity of 20m„ The relationship of transparency to natural 

 food, bait" taking, and catch will probably become gradually clearer in 

 the future as more data are made available,, 



