(A) POLE -AND- LINE 



(B) LONGLINE 



20 

 15 

 10 

 5 

 

 15 

 10 

 5 

 



15 



10 



TOKARACSWOF JAPAN) 



FEB 1951-55 

 (N = ?) 



U.S. TRUST TERRITORY 



4°N-23''N ; I32°E-I55°E 

 APR a OCT. 1967 

 (N = 970) 



HAWAI I 



3" QUARTER 1960 



(N = 4,I80) 



MARQUESAS ISLANDS 



OCT 1957- MAY 1958 

 (N = I,2I6) 



^^rUi^lrr^ 



_i_ 



35 40 45 50 55 60 65 70 75 80 85 

 TOTAL LENGTH (CM) 



35 40 45 50 55 60 65 70 75 80 85 

 TOTAL LENGTH (CM) 



Figure 2. — Length compositions of skipjack tuna taken with pole-and-line and longline gear in various parts of the Pacific Ocean. (Source of 

 data: Panel A— Tokara from Kawasaki (1965); U.S. Trust Territory from Tohoku Regional Fisheries Laboratory (n.d.); Hawaii from Roth- 

 schild (1965); Marquesas Islands from Wilson, Nakamura, and Yoshida (1958). Panel B - Northwestern Pacific from Miyake (1968|; U.S. 

 Trust Territory from Murphy and Otsu (1954); central Pacific from Murphy and Shomura (1953, 1955); and Shomura and Murphy (1955); 

 central South Pacific from files NMFS Honolulu Laboratory.] 



Table 1.— Estimated ratio of skipjack tuna recorded to skipjack tuna 

 landed in the Japanese longline fishery, 1962-67. 



'Data from Fisheries Agency of Japan, Research Division (1969). 

 'Data from Japan, Ministry of Agriculture and Forestry. Statistics 

 and Survey Division (1963, 1964, 1966, 1967, 1968, 1969). 



Miyake (1968) pointed out the possibility that differ- 

 ences between small and large vessels in reporting 

 skipjack tuna catch data could result in geographical 

 bias of the catches, since smaller vessels (20- to 50-ton 

 class) operate only in near coastal waters, whereas 

 larger vessels (>50-ton class) tend to operate farther 

 away (vessel sizes are in gross metric tons). The data, 

 in fact, show that large vessels fished also in areas 

 fished by small vessels. To determine whether there 

 was any difference in recording by vessel size, the 

 quarterly catch rates (skipjack tuna per 1,000 hooks) of 

 small and large vessels in each 5° area fished by both 

 were compared, using the Wilcoxon matched-pair, 

 sig^ed-rank test (Siegel 1956). The tests were made on 

 the premise that the catch rates of small and large 

 vessels fishing in the same area and same quarter would 

 be comparable, so that any large differences would be 

 attributable to differences in recording. Zero catches by 

 both vessel size categories in the same area were 

 excluded, since it could not be established whether these 



