SKIPJACK IN HAWAII FISHERY 



287 



weight per fish, a certain amount of error will re- 

 sult from mixed catches of small and large fish; 

 this error should disti-ihute itself more or less ran- 

 domly, however, so that neither size group is 

 consistently favored. 



ERROR IN ESTIMATION OF FISHING EFFORT 



There is no way to determine from the catch 

 records the actual effort, i.e., the fisherman-days 

 whether productive or not, put forth on a skipjack 

 boat. In this study only positive fishing results 

 (catch' reports) are available, and the productive 

 fislierman-trip is of necessity used in lieu of the 

 fisherman-day. Sources of error in the productive 

 fisherman-trip have been discussed in the section, 

 Clioice of the Unit of Fishing Eifort, and on the 

 basis of the performance of two skipjack boats for 

 which logbooks are available, it appears to be a 

 reasonable substitute. 



OTHER SOURCES OF ERROR 



The weight of the catch of skipjack taken in tlie 

 Hawaiian live-bait fishery is affected by complex 

 factors which present sources of error that are 

 difficult to estimate. Among these factors are 

 variations in bait supply, response of skipjack to 

 chum, behavior and niuuber of birds in the flocks 

 which serve to locate schools, the size and behavior 

 of the skipjack schools, selection by the fishermen, 

 and probably several others. 



Yamashita (1958, p. 270) has discus-sed the 

 problem of ascertaining the influence of variations 

 in bait supply on the skipjack catch in terms of 

 annual production and suggests that in certain 

 years, when skipjack have been plentiful, the 

 availability of bait may be a limiting factor in the 

 fishery. Royce and Otsu (1955) have investigated 

 many aspects of Ijeliavior of skipjack schools and 

 birds; Yuen (1959) has studied the response of 

 skipjack to live bait. 



In the present study no attempt has been made 

 to evaluate the sources of error introduced by the 

 factors considered alx)ve. Information available 

 is not adequate to discern which of these may be 

 important at any particular time. It seems reason- 

 able that most of these factors act relatively in- 

 dependent of one another so that over a period of 

 time their combined effects should not introduce 

 bias. However, it is just as plausible that at cer- 

 tain times several of these elements may act in 



unison resulting in considerable deviation from 

 the normal state. The investigation of the role of 

 these factors in the fishery awaits a more sophisti- 

 cated study than is attempted here or is possible 

 with the present sources of information. 



CONCLUSIONS ON SOURCES OF ERROR 



None of the sources of eri-or appears to be so ex- 

 tensive as to destroy the usefulness of the catch 

 report as the basis for a study of distribution and 

 abundance. Some of the sources of error tend to 

 reduce the bias introduced by othei'S. With re- 

 spect to time, geogi-aphy, and size, the categories 

 employed in this study have deliberately been 

 made broad. Were the study concerned with only 

 a few vessels, very short time periods, or several 

 size groups, the probability of error would be in- 

 creased, but as only the most general of categories 

 are used, the influence of error on the results 

 should be slight. 



RESULTS AND DISCUSSION 



There are small discrepancies between the official 

 total catches for 1952 and 1953 as listed by Yama- 

 shita (1958, table 2) and the totals obtained in the 

 present study (table 3). These differences amount 

 to 1 percent and are probably the result of catch 

 reports, which were turned in too late to be in- 

 cluded in official summaries and to records lost or 

 misplaced during the interval of storage. The 

 pi-oportion of unusable data in 1953 was greater 

 than in 1952 (table 4). largely because of the poor 



Table 3. — Comparative data from 2 studies of the Ha- 

 waiian skipjack catch for lHoZ and 1953 



1 Source: Yamashita (195S, table 2). 



' Figures adjusted to correspond with calendar year. 



Table 4. — I'sahility of 1952 and 195S catch report data 



