FISHERY BULLETIN: VOL. 74, NO. 3 



and, therefore, is invalid, but rather that the 

 apparent change in growth was caused by seasonal 

 changes in availability of the stock due to some 

 seasonal size-age migration phenomenon or possi- 

 bly by changes in the selectivity of the fishery. 

 Preliminary examination of Japanese longline 

 statistics suggested that the striped marlin stock 

 available to the Hawaiian fishery shifts its center 

 of density northward in the months with the 

 warmest water temperature and becomes less 

 available to the local fishery (Heeny S. H. Yuen, 

 Southwest Fisheries Center, pers. commun.). The 

 decreased growth most commonly seen in the third 

 quarter might then be due to smaller fish being 

 associated with the periphery of the stock. 



There are of course other possible e.xplanations. 

 For example, one reviewer suggested that this 

 period of slow or negative growth represented an 

 asymptote followed by the initiation of a new 

 growth phase. Fitting a two-cycle Gompertz 

 growth curve to the pooled data, this reviewer 

 found both sexes tending toward an asymptote at 

 age 2.46 yr followed by another growth phase 

 where females tended toward an asymptote at 

 320 cm, but no solution was found for males. Such 

 changes in growth phase are common at sexual 

 maturity and at other times when body form 

 changes. Change in growth form is commonly 

 accompanied by a corresponding change in the 

 length-weight relationship, and Skillman and 

 Yong (1974) found no indication of a change in the 

 length-weight relationship over the range 

 142.2-310.1 cm. Also, the age of fish in the 

 Hawaiian fishery, having a calculated mean length 

 corresponding to length at first maturity found by 

 Eldridge and Wares (1974) and Kume and Joseph 

 (1969) for the eastern tropical Pacific, was 4.2 yr. 

 This age is nearly double the hypothesized age of 

 first asymptotic growth. Thus, while it is possible 

 to fit a segmented growth curve to the data, 

 biological evidence given above does not support 

 such a procedure. 



Another possible explanation involves the 

 separation of age-groups. The aberrant growth 

 occurred most frequently in the third quarter, and 

 since the sample size was smallest in this quarter, 

 the precision of the estimates is probably less than 

 for the other quarters. However, the aberrant 

 growth did not always occur in this quarter, and its 

 repeated occurrence among cohorts suggested 

 that it was real and not an artifact of the estima- 

 tion procedure per se. With any probabilistic 



means of separating age-groups from a mixed 

 distribution, there is always the danger that 

 age-groups from diff"erent cohorts of the same 

 spawning stock will be so similar in size that they 

 cannot be separated, especially with increasing 

 age and varying growth rates of the cohorts. We 

 acknowledge that this may be a problem, but if it 

 is, it would seem from Figure 1 to be more impor- 

 tant for the growth period following the period of 

 aberrant growth. This problem would be increased 

 if there were more than one spawning stock 

 involved, and this seems to be the case for some 

 quarters. In spite of the small sample sizes in the 

 third quarter, there seemed to be little doubt that 

 the 167-cm age-group was real, since its mean 

 length is quite removed from that of the next 

 age-group at about 194 cm and since the age-group 

 was found for the pooled data analysis and for 8 

 out of 9 yr for females and for all years for males in 

 the analysis of yearly data. Because the two 

 spawning stocks would continue to have quite 

 different lengths for the next couple of quarters 

 and no comparable age-groups were separated in 

 these quarters, it was reasonable to assume that 

 this other secondary stock was not represented in 

 the catches in the subsequent fourth, first, and 

 probably second quarters. But what about the 

 following third quarters? If similar growth curves 

 are assumed for this other stock, then the 200.0-cm 

 female age-group in the third quarter of 1966 and 

 the 204.6-cm male age-group in the third quarter 

 of 1968 could also belong to the secondary spawn- 

 ing stock. If this secondary stock was present in 

 other years but in numbers too small to be separ- 

 ated out, it would tend to bias downward the 

 estimates of the similar-sized age-group of the 

 primary spawning stock. With the accuracy of the 

 present set of data, it is impossible to comment on 

 the likelihood or importance of this possibility. 



The occurrence of these age-groups at approx- 

 imately 167 cm in the third quarter presents an 

 additional problem. Where do they come from? If 

 the male and female growth curves are used to 

 back calculate the probable time of spawning for 

 the age-group at approximately 167 cm in the 

 pooled data analysis, January is estimated as the 

 time of peak spawning. We hypothesize that these 

 fish could come from a stock spawning in the 

 equatorial region, probably north of the equator, 

 during months corresponding to the southern 

 summer. It is hard to visualize a hypothesis that 

 would account for a stock spawned 6 mo out of 



562 



