FISHERY BULLETIN: VOL. 81. NO. 1 



estimates are plotted against time, the intercept 

 at time zero is an estimate of (/ for bluefin in the 

 eastern North Pacific. 



As tagged bluefin were not fully dispersed dur- 

 ing the season of tagging, monthly estimates of q 

 were calculated as the monthly mean, per 1° area 

 of latitude and longitude, for 1° areas from which 

 tagged fish were caught. For the second and 

 third seasons, when tagged fish appeared to be 

 fully dispersed, monthly estimates were calcu- 

 lated for the entire bluefin range; then, the nat- 

 ural logarithms of these values and those for the 

 first season were plotted (Fig. 17). Effort and 

 therefore q are expressed in boat-days. The re- 

 gression line fitting these points (Y= —8.7363 — 

 0. 1725 X, R 2 « 68%) was weighted by the number 

 of tagged fish released each year, since the num- 

 ber of tagged fish varied between 35 and 960/yr. 



The best estimate of q from the tag-recapture 

 data is the antilogarithm of the regression line 

 intercept, 1.66 X 10~ 4 /boat-day with a 95% con- 

 fidence interval of 0.99 X 1(T 4 to 2.63 X KT 4 / 

 boat-day corrected for geometric mean bias 

 (Beauchamp and Olson 1973). The slope of the re- 

 gression (—0.17, S 2 = 0.02) is an estimate of the 

 monthly instantaneous mortality coefficient (Z), 

 and was expanded to estimate the yearly instan- 

 taneous mortality (Z = 2.07, S 2 = 0.24) including 

 immigration and emigration. This estimate com- 

 pares favorably with Bayliff and Calkins' (1979) 

 and Bayliff s (1980) estimates (Z = 2.08, S 2 = 

 0.8) for 1962-66. They call these estimates "rates 



1 2 



8 9 10 11 12 13 

 RETURN MONTHS 



222324 



Figure 17.— Natural logarithms of adjusted return rates for 

 tagged bluefin tuna plotted against number of months between 

 tagging and recapture, for the years 1962-64, 1966, and 1968. 

 The predicted catc liability coefficient (q) from straight-line re- 

 gression and the 95% confidence interval around (§) are shown 

 at the zero-month intercept. 



of attrition," since immigration and emigration 

 are included. 



The ratio of fishing mortality to instantaneous 

 total mortality is an estimate of the exploitation 

 ratio (Ricker 1975) and was calculated as a mean 

 for the period 1962-70 because q was also calcu- 

 lated for that period. The mean annual fishing 

 effort in that period was 4,215 boat days which, 

 multiplied by (/, estimates a fishing mortality of 

 0.7/yr. Dividing this value by estimated Z(2.07/ 

 yr) yields an exploitation ratio of 0.34, and then 

 multiplying by the annual mortality or "attri- 

 tion" (0.87) yields a 30% exploitation rate. 



DISCUSSION 



The review and analysis of data concerning the 

 bluefin tuna fishery in the eastern North Pacific 

 show large fluctuations in the catch to be a major 

 part of two important phases. The decline in 

 catch near the end of the first phase (1921-50) is 

 offset by the development of a "high seas" purse 

 seine fleet and the resultant increased catch of 

 bluefin off Baja California. The current decline 

 (1963-present) is probably due to a decline in the 

 abundance of bluefin as indicated by CPUE evi- 

 dence. The effect on the resource of Mexico's 200- 

 mi regulations was not assessed at this time; how- 

 ever, the apparent decline in catch and CPUE 

 cannot be attributed to such regulation since it 

 has been enforced only recently. 



The declines in catch and CPUE in the eastern 

 North Pacific are significant and are reflected 

 by an even greater decline in catch and nomi- 

 nal CPUE in the western Pacific (Figs. 18, 19). 



50 



O 

 o 

 o 



I 

 o 



40 



30 



20 



o 10 



_L 



1948 1956 



1964 1972 1980 



YEAR 



Figure 18.— Annual Japanese landings of northern Pacific 

 bluefin tuna for the years 1951-59 (metric tons X 1.000) and 

 1962-79 (thousands of fish). 



118 



