296 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



with caution, however, because the fishei-y is 

 limited geographically to the immediate island 

 area (fig. 4). Furthermore, the catches are also 

 influenced by availability, fishing effort, and selec- 

 tivity on the part of the fishermen, which may not 

 be constant throughout the year. Brock (1954, p. 

 100-103) ha? shown, by means of sex ratios, that 

 the availability of female skipjack is not constant. 

 He suggests that spawning activity may be the 

 crucial element. The cyclical nature of fishing 

 effort has been shown previously (fig. 11), and the 

 fact that the fishermen are selective in the schools 

 they fish is common knowledge. These biases 

 appear to have an annual cycle and between-year 

 comparisons may not be affected by them to the 

 extent of within-year comparisons. 



Fluctuations In the Catch of Large Skipjack In the 

 Oahu Region 



The growth rate of Hawaiian skipjack has been 

 studied by Brock (1954, pp. 96-97) by means of 

 length frequency distributions. During the sum- 

 mer there are two distinct modes, one at about 45 

 cm. (4 pounds) and one at about 70 cm. (18 

 pounds) } The mode of large skipjack represents 

 season-fish and presumably a smaller number of 

 the previous year's season-fish. It is this mode 

 that is considered here as large skipjack. The time 

 of year when a mode, which Brock assimies to be a 

 year class, passes through the weight ( 10 pounds) 

 which separates small and large fish in this study, 

 is apparently winter or early spring. During the 

 period from May to October, it may be assumed 

 with reasonable certainty that the small skipjack 

 are a year younger than the large skipjack. 



In 1952 there appears to be little consistency in 

 the average weight for large skipjack (fig. 12), 

 which fluctuates widely from one biweekly period 

 to the next. By contrast, 1953 has an interval 

 from May 3 through October 3 (periods 10-20) 

 with a regularly increasing weight for large fish. 

 The rate of this increase, 0.25 pound per week, is 

 in agreement with Brock's curve for skipjack 

 growth, which yields a linear weight increase of 

 0.25 pound per week. The coincidence of the sea- 

 son of gi-eatest productivity (as indicated by the 



> ConverslonH of length (millimeters) to weigh (pounds) were 

 made accordilng to the formula: Log welght= — 8.2755-|-3.34913 

 loc total length. 



Figure 12. — Catch of large skipjack in the Oahu region, 

 1952 and 1953. 



number of large fish caught) with the interval of 

 uniform weight increase suggests that in 1953 a 

 single population of season skipjack was available 

 to the fishery, but the erratic fluctuation in the 

 numbers taken indicates that variations in this 

 availability were quite marked. 



Period 5 of 1953, with its unusual numbers of 

 very large fish, must consist of skipjack greater in 

 size than the 1953-season fish. It seems probable 

 that these very large skipjack are 1952-season fish, 

 which were present only briefly that year. The 

 relatively high average weights during other 

 periods of early 1953 imply that 1952-season fish 

 may have been generally present during that time. 

 In early 1952, on the other hand, few of the previ- 

 ous year's season fish were present, as judged by 

 the average weights during periods 4 to 7. 



