FISHERY BULLETIN: VOL. 87, NO. 4, 1989 



oocytes ripen and are successively spawned was 

 indicated by the wide range in GSI values calcu- 

 lated for both species during the spawning 

 season. 



Sex Ratio 



The sex ratio of male to female uku was 1.05:1 

 (51.2% males, A'^ = 559 individuals combined). A 

 chi-square goodness of fit test suggested that 

 this ratio was not significantly different from the 

 expected 1:1 ratio (x" = 0.302, P > 0.05). By 

 examining sex ratio at 50 mm FL intervals, inde- 

 pendence was determined and the percentage of 

 females was shown to decrease between the 600 

 and 750 mm FL categories from 50.0 to 40.9%, 

 then increase 65.5% at 800 mm FL before reach- 

 ing 100% beyond 900 mm FL. Based on two-way 

 contingency table analysis, the percentage of 

 females caught increased significantly towards 

 the end of the spawning season (September- 

 October; Table 3). Males were caught in higher 

 percentages in the spawning months from May 

 to July. 



DISCUSSION 



The spawning season for both uku and onaga 

 extends throughout the summer months in 

 Hawaii, as evidenced by the advanced condition 

 of the ovaries and the peaks in ova diameters and 

 GSI values during this period. This pattern has 

 been reported in Hawaii for other snapper spe- 

 cies, including E. carhuncuhis (Everson 1984) 

 and P. filamentosus (Kikkawa 1984). Spawning 

 activity also occurs during the austral summer 

 for populations of A. viresce>is from New Cale- 

 donia (Fourmanoir and Laboute 1976) and East 

 Africa (Talbot 1960; Nzioka 1979). Likewise, 

 peak summer spawning with intermittent activ- 

 ity throughout the rest of the year seemed to be 

 the pattern for E. coruscans in Vanuatu 

 (Brouard and Grandperrin 1985). The seasonal 

 peak in spawning activity may be most closely 

 tied to increases in water temperature and day 

 length, as suggested by Walsh (1987) who ob- 

 served that spawning activity for Hawaiian reef 

 fishes declined rapidly in September to October 

 as maximum water temperatures were reached. 



Table 3. — Tests of the hypothesis that sex ratio of uku and onaga did not vary 

 significantly from the sample populations during the year. Data are pooled by 2 mo 

 intervals for samples collected in 1984-87 (df = 5). 



P<0.05. 



The sex ratio of male to female onaga differed 

 significantly (x^ = 8.99, P < 0.05) from 1:1 in 

 favor of males (61.4%, N = 347 individuals com- 

 bined). The overall ratio of males to females was 

 1.59:1. There was a significant preponderance of 

 males within the 50 mm FL intervals from 600 to 

 750 mm. However, females predominated above 

 850 mm FL, reaching 100% of the individuals in 

 the 950 mm FL category. The two-way contin- 

 gency table analysis suggested that sex and 

 month of capture were not independent (Table 

 3). 



This pattern would ensure optimum tempera- 

 ture conditions for developing larvae. A similar 

 post-summer spawning decline associated with 

 changing local environmental conditions was 

 noted by Grimes and Huntsman (1980) for 

 Rhomboplites aurorubens from North and South 

 Carolina and by Everson (1984) for E. car- 

 bwnculus. The extension of onaga's spawning 

 season into November, with a peak in October, 

 may reflect that the genus Etelis is restricted to 

 much greater depths than are the other reef- 

 associated lutjanid species. Seasonal changes in 



884 



