of the body Treight, and the large fish had large gonads irtiile the small fish had 

 small ones. This fact is shown by figi^.re 3, which is a graph of the logarithms of 

 the gonad and body weights with the males and females plotted separately. As can 

 be seen from the graph, no difference could be perceived between the gonad weights 

 of male and fenale fish. VJhen I called at the Nank5 Fishing Company at Saipan on 

 my way to Palao, I '/»as told that, "The small fish have small eggs and the large 

 fish have largo ones." I wondered iriiat they were talking about, but now my doubts 

 have been cleared up by the realization that by "eggs" they meant the ovaries, and 

 I have found that they weire quite right. 



There are some wide variations in the weights of the gonads as shown in figure 

 3. It is thought that the smaller variation in the case of fish of medium weight 

 is due to the fact that there were few specimens in this size range, and the parti- 

 cularly great variation in the small specimens is thought to be the result of using 

 logarithmso At any rate there is a wide variation. One wonders vrtiether it may be 

 due to inherent variations in the gonads themselves, or whether it may not result 

 from lumping together in one graph specimens '.vhich may have been at different 

 stages of maturity either because they were taken at different seasons or ftrom dif- 

 ferent schools. Figure U was drawn on the basis of data from only those specimens 

 iriiich were taken at the beginning of the study (the middle of October) and at its 

 end (the first part of December) . As the figure shows, there was a tendency for the 

 gonad weights of the specimens taken in December to be slightly larger than those of 

 the fish collected in October, It is thought that in such a study the difference 

 in the season at which the specimens were taken is one reason for the large varia- 

 tion. It cannot, however, be said that this increase in weight is due to the devel- 

 opment of the gonads within a period of two months. Since nothing is known of the 

 ecology of the skipjack urtiich provided the material for study, it may well be that 

 this merely indicates a difference in the sexual development of the schools. While 

 the author was at Palao the schools which were fished in the adjacent waters were 

 of three types, those consisting only of small skipjack, those consisting only of 

 large skipjack, and those characterized by the admixture of yellowfin tuna of the 

 same size as the skipjack, 7*iich might be of any size from large fish to small. An 

 attempt was made to find out whether the variation js due to the difference in the 

 schools from which the specimens were collected by plotting separately the cor- 

 relation between the gonad weights and body weights of fish from three schools, but 

 this experiment revealed no difference in the gonad weights among the three schools, 

 and showed that the variation noted above is not due to the fish of the three 

 schools having different sized gonads. 



Next only females were selected which, according to the criteria described 

 above, were in the low gz*ades 1 and 2 or the high grades 3 and U of sexual maturity, 

 and the correlation between their body weights and gonad weights was examined. The 

 material used in the experiment was jrather scanty, but the results showed no cor- 

 relation between the progress of the maturation of the gonad and its weight. It is 

 naturally to be expected that the gonads will increase in weight as they develop, 

 but if this increase follows the general laws of growth it can be assumed that it 

 will be sudden and marked at first and will thereafter become slower and slower. 

 Therefore it is thought that the gonads collected for this study had already reached 

 a certain degree of maturity and consequently their increase in weight was at a rate 

 too small to be shown on the graph. Thus, of the large variation in the relation of 

 gonad weight to body weight it can only be said that there was a seasonal difference 

 in the material studied, and the reason for it cannot, unfortunately, be clearly 

 established. It may also be thought that since there are considerable differences 

 between the gonad weights of the individual fish, the correlation shown in figure i^ 

 may have been arrived at by chance because of the small number of schools and in- 

 dividual specimens which were used. 



