RALSTON and POLOVINA: COMMERCIAL DKKP-SKA HANDLINK FISHKRY 



Table 4.— Comparisons of correlations of CPUE and fishing 

 effort if) for two different measures of/. The total aggregate 

 incorporates all 13 species. 



'MLKM = Maui-Lanai-Kahoolawe-Molokai 

 KNK= Kauai, Niihau, and Kaula Rock 

 Significant at P = 0.05 level, one-tailed test, df 



18. 



is apparent that in 13 of the 16 possible compari- 

 sons, fisherman-days showed a stronger negative 

 correlation with CPUE than did catch-records. 

 Based on these results we conclude that 

 fisherman-days predicts the behavior of CPUE 

 more precisely than catch-records. Use of this 

 measure also eliminates repeated counting of 

 effort statistics when more than one species in a 

 group is caught on a particular day and has 

 greater intuitive appeal as well. For these 

 reasons we conclude that fisherman-days is the 

 best measure of fishing effort available at 

 present. It is worth noting that these two 

 different measures of effort are approximately 

 linear in their relationship to one another, imply- 

 ing that the superiority of fisherman-days over 

 catch-records as a measure of effort is probably 

 due to a smaller residual variance of instanta- 

 neous fishing mortality (F) on the former statistic 

 than on the latter. 



Stock Production Analyses 



In this section the Schaefer model is applied to 

 the deep-sea handline data in which the catch is 

 aggregated at three different levels. At the first 

 level a single-species Schaefer model is fitted to 

 each species separately. Next, the TBSM is fitted 

 to each of the three species groups delimited by 

 the cluster analysis. In the final section the total 

 aggregated catch of all 13 species taken together 

 is analyzed with the TBSM. Fisherman-days 

 was used as the measure of fishing effort 

 throughout, but equilibrium approximation 



(Gulland 1972) was not attempted because no 

 information was available concerning the 

 longevity of these species and a previous appli- 

 cation of this method to the data had shown little 

 improvement in the results (Ralston footnote 

 4). 



When each species is treated independently 

 there are 52 separate analyses (4 banks with 13 

 species each). In only two of these regressions of 

 CPUE on /is the null hypothesis /?>0, where j3 is 

 the slope of the regression, rejected in favor of the 

 alternative hypothesis /3<0. Both involved the 

 MLKM bank where opakapaka (t = -2.91, df = 

 18) and uku (t = -1.82, df = 18) demonstrated 

 significant inverse regressions in which respec- 

 tively, 32% and 16% of the total variation in CPUE 

 were explained. The significance of these two re- 

 gressions can easily be attributed to the Type I 

 error and consequently nothing can be concluded 

 from these results concerning the productivity 

 of these fishes. 



The fit of the TBSM to the data is much 

 improved when the three species groups are con- 

 sidered. The model was applied to the HDFG 

 data 12 times; once for each species group and 

 bank combination. Significant results (P = 0.05, 

 one-tailed test) were obtained in 5 of the 12 appli- 

 cations of the model (Table 5). The three analyses 

 from the MLKM bank were significant in every 

 case and those for Group III were significant in 

 three out of the four regressions tested. The ob- 

 servation that the results from the remaining 

 banks and species groups are not significant is 

 not so disturbing because 56% of all bottom fish 

 landings are harvested from the MLKM bank 

 (Table 2). An estimate of the maximum sustain- 

 able yield (MSY) and optimum effort was then 

 computed for each of the five significant com- 

 binations, as well as a standardized measure of 

 productivity, calculated as the sustainable yield 

 of bottom fish per nautical mile of 100-fathom 

 isobath. Assuming logistic growth of the stocks 

 the catchability coefficient was estimated using 

 the computer program PRODFIT (Fox 1975). 

 The t value in the table refers to the test of the 

 null hypothesis that the slope of a regression is 

 zero or positive. 



Pope (1979) has proposed an interactive model 

 to describe multispecies fisheries in which total 

 yield is depicted as the sum of the yields of 

 individual species with additional terms to 

 account for community interactions. In the 

 simple two-species case the equation describing 

 surplus production (Y) is: 



441 



