our crude estimate of the total mortality co- 

 efficient in 1961 was 1.3 whereas that reported 

 by Kamimura (1966) is 0.4. We cannot explain 

 this discrepancy except to note that the compu- 

 tations reported by Kamimura were undoubtedly 

 based on a longer series of data than was used 

 by us. We must also emphasize that an in- 

 crease in the per-head rate of death of 0.1 fish 

 per year with a constant instantaneous coeffi- 

 cient of natural mortality and a constant catch- 

 ability coefficient implies that effort is in- 

 creasing as a function of the reciprocal of the 

 catchability coefficient. Without knowing the 

 trend in effort, however, it is difficult to tell 

 the extent to which a trend in CPUE is related 

 to changes in effort or changes in the catch- 

 ability coefficient. 



With constant growth coefficients, constant 

 recruitment, age, etc., an increase in the per- 

 head rate of death implies that the average age 

 and therefore, also, the average size of alba- 

 core taken by the longline fishery should de- 

 crease. Such a reduction in average age is 

 shown by Suda (1959). This reduction can also 

 be seen, qualitatively, in a later study by Suda 

 1963b: p. 1250). On the other hand, we have 

 averaged the length-frequency samples given 

 by Suda (1963a) and find no conspicuous de- 

 crease in average size. If growth rate is con- 

 stant, and we have no information to the con- 

 trary, then according to the longline-fish aver- 

 age lengths depicted in figure 5, average age is 

 not decreasing. We have also examined these 

 data by 10° longitudinal strips and also find no 

 decreasing trend in average size within a lon- 

 gitudinal strip, which reduces the possibility 

 that the absence of a declining trend in size 

 could be related to the manner in which the 

 size distributions are weighted. Another con- 

 sideration is that the longline gear may be 

 highly size selective. If this were true, then a 

 decline in average size in the population would 

 not change the average size of the fish in the 

 catch, but it would reduce the catchability co- 

 efficient so that the decline in CPUE would be 

 greater than the decline in abundance of the 

 population. A final consideration concerns the 

 underlying possibility of the curvilinear decline 

 in actual abundance that was discussed earlier. 

 An exceedingly rapid--fishery induced--initial 

 decline in apparent abundance might produce a 

 sharp initial decline in average size which may 

 appear to be a random fluctuation. Thus, a 

 real trend may be masked--especially for such 

 a short series of data--by a fortuitous com- 



bination of time-dependent parametric changes 

 and random fluctuations. Thus, while the de- 

 cline in apparent abundance certainly suggests 

 a decline in actual abundance, an assessment 

 of the magnitude of this decline awaits further 

 studies on size-distribution and fishery inten- 

 sity. 



Spatial Statistics 



The next question that we considered in our 

 study was the relation of the declining apparent 

 abundance of albacore to the temporal-spatial 

 distribution of the albacore on the longline 

 grounds. In order to investigate the temporal- 

 spatial distribution of the albacore we com- 

 puted the spatial statistics. The spatial statis- 

 tics revealed several interesting characteris- 

 tics of the apparent distribution of albacore in 

 the North Pacific. 



First, as indicated previously, the spatial 

 statistics for catch and CPUE were nearly 

 identical. This agreement implies that the 

 spatial distribution of catch is coincident with 

 the spatial distribution of apparent abundance. 

 The statistics for effort, however, are not co- 

 incident with those for catch and CPUE. The 

 fact that catch and CPUE are very nearly co- 

 incident and that effort is not centered on either 

 indicates that the fishery is operating efficient- 

 ly in the sense that the catch is taken from the 

 aiea where apparent abundance is most highly 

 concentrated and that the fishing effort con- 

 sidered in this study is not totally directed 

 toward catching albacore. The fact that the 

 spatial distribution of the albacore's apparent 

 abundance is nearly constant for any month and 

 that of effort is not constant, but moves away 

 from the center of albacore apparent abundance, 

 reflects the increasing tendency tor Japanese 

 longline effort to move away from the north- 

 west Pacific Ocean. The tendency of longline 

 effort to move from the northwest Pacific toward 

 the southeast Pacific during the study period is 

 well known (e.g., Rothschild, 1966). The ten- 

 dency for effort to move away from the major 

 apparent concentration of albacore in the north- 

 west Pacific would naturally introduce a de- 

 clining trend in albacore apparent abundance 

 that merely reflects the change in the distribu- 

 tion of effort in the North Pacific Ocean rather 

 than any actual change in albacore abundance. 

 This situation cannot be studied in detail owing 

 to lack of data on the other species. We be- 

 lieve, however, that the problem of effort mov- 

 ing from the albacore grounds as well as the 



29 



