34 



Fishery Bulletin 101(1) 



160° W 1 



I KAUAI 

 I ^ OAHU 



^ MOLOKAI 



Malaekahana 

 (1994) 



N 



Kahana Bay 

 (1993, 1994, 1996, 1997, 1998) 



Kailua Bay 

 (1996, 1997, 1998) 



10 km 



Figure 1 



Island of Oahu, Hawaii showing locations for releases of hatchery-reared Pacific threadfin iPoly- 

 dactylus sexfilis) from 199,3 to 1998. 



enhancement program. Fishermen were interviewed at 

 boat docks and along the coasthne to collect information 

 about their catch and effort (species, lengths, number of 

 hours fished per gear type). All Pacific threadfin in the 

 catches were scanned for coded wire tags with a hand-held 

 Northwest Marine Technologies cwt sensor ("wand"). 



Because Pacific threadfin were scarce in creel samples, 

 we offered a $10 reward to Oahu fishermen for each Pacific 

 threadfin (hatchery-reared and wild) caught in the open 

 season between March 1998 and May 1999 (no retention 

 is allowed from June to September). Posters, newspaper 

 articles, and radio talk shows were used to publicize the 

 reward program. We also informed participating fishermen 

 that we were primarily interested in samples from the 

 windward side of the island; thus, catches of Pacific thread- 

 fin from other parts of the island were underrepresented in 

 our effort. To obtain the reward, fishermen provided date, 

 location, gear used, fork length, and the head of each Pa- 

 cific threadfin caught. Beginning in September 1998, fish- 

 ermen had to return whole fish to obtain the $10 reward. 

 This requirement allowed us to gather more accurate 

 length information, as well as information on weight, sex, 

 and gonadal condition. We obtained a relatively unbiased 

 random sample from the fishery because fishermen were 

 unable to determine visually whether a fish had a coded 

 wire tag; this unbiased sample enabled us to calculate the 

 contribution of hatchery-reared fish to the fishery. 



A Mann-Wliitney rank sum test (T-statistic) was used 

 to compare distance traveled among sites and years for 

 hatchery-reared fish. This test was also used to compare 

 fork lengths of wild fish by sex between the 1999 and the 

 1962-68 sampling periods. We compared number of Pacific 

 threadfin recovered in each size class in relation to the 

 number of hatchery-reared fish released in that size class 

 by using chi-square goodness of fit. Expected number of 

 fish recaptured in each size class was calculated by taking 

 the total number of recaptures for each year from each 

 release site and multiplying it by the proportion of fish 

 released in that size class at that site during that year 



A PCruskal-Wallis ANOVA on ranks with Dunn's multiple 

 comparison procedure (Sokol and Rohlf, 1981) was used to 

 identify pair-wise difference in CPUE among gear types; 

 « = 0.05. A two-way ANOVA (F) was conducted to compare 

 fish length among gear types for wild and hatchery-reared 

 Pacific threadfin during each reward year Tukey multiple 

 comparison procedures were conducted to determine pair- 

 wise differences in fish length between gear types; a = 

 0.05. Condition factor was calculated as 



C = (W/L'')x 10,000, 



where C = condition factor; 



W = fish weight in giams; and 

 L = fork length in mm (Anderson and Neumann, 

 1996). 



