GRIMES ET AL.: REPRODUCTIVE BIOLOGY OF TILEFISH 



Table 8.— Sex ratio at age and log-likelihood tests that sex 

 ratio at age was not different from 1:1. All G scores were 

 calculated using Yates correction for small sample sizes. 

 Critical chi-square values are x^qos (^ <^^ = 3.84 and x^o.oi 

 (1 df) = 6.64. NT = not testable. 



Table 9.— The proportion of male tilefish in the 1978 



n 



population >50 cm FL, P = X L, M, where L and M = 



the proportion of all fish and males, respectively, in the /'^ 

 5 cm FL interval, and n = the number of size intervals. L, 

 was determined from the 1978 commercial longline catch 

 of tilefish. 



fractional spawners from about March through 

 November, although most of the reproduction 

 evidently occurs from May to September. Some 

 females with free ova in the ovarian lumen (running 

 ripe) v^ere present in March through August and in 

 October and November. From May through August, 

 89-98% of the females were ripe or running ripe 

 (Fig. 5). Running ripe or ripe males were not as 



frequently observed as were females in a similar 

 reproductive state. In fact, only very large males 

 (75-80 cm FL) were observed with a large creamy 

 white swollen testicular mass. Ripe males were 

 found in January, March, May through August, 

 October, and November, but the highest proportions 

 (23-46%) were present in May through August (Fig. 

 5). 



GSI data for females indicated a similar seasonal 

 spawning pattern (Fig. 6). Highest GSI values con- 

 sistently occurred from May through August, when 

 ovaries accounted for 3.5-8.3% of gutted body 

 weight. 



Analysis of ovum-diameter data also suggested 

 that spawning occurred mostly from May to Sep- 

 tember, and indicated that spawning was fractional, 

 i.e., ova were released in batches. Highest monthly 

 mean developing ovum diameters (0.30-0.42 mm) 

 occurred in May through August (Fig. 7). During 

 these months mean ovum diameter was usually 

 >0.35 mm (Fig. 7). During other months, mean 

 ovum diameter was always <0.30 mm. The size- 

 frequency distribution of ova from running ripe 

 females was polymodal (Fig. 8), suggesting multi- 

 ple spawnings by individual females during the 

 reproductive season. 



Tilefish may not utilize the liver and soma to store 

 energy as fat for mobilization to the gonads in prep- 

 aration for spawning as many species are thought 

 to do (Hoar 1957). HSI for both males and females 

 showed a distinct pattern of seasonal variation, but 

 highest values occurred during summer (spawning 

 season) and lowest in winter (Fig. 9). Somatic con- 

 dition factor (eviscerated weight, g/FL^) showed no 

 discernable seasonal pattern in males or females. 



Fecundity 



Because tilefish are fractional spawners, fecun- 

 dity (ovarian egg count) was estimated from females 

 collected early in the spawning season (i.e., May to 

 early June) to minimize the chance of using a par- 

 tially spawned ovarj' and underestimating egg num- 

 ber. Estimates of fecundity ranged from approx- 

 imately 195,000 for a 53 cm FL (2.1 kg) female to 

 10 miUion for a 91 cm FL (13 kg) female, with a 

 mean egg count of 2.28 million (r? = 49, SD = 1.02). 

 The 91 cm FL female with 10 million eggs was ex- 

 ceptional; all other estimates were <4.1 million, even 

 for other large females 80 and 86 cm FL. Therefore, 

 we judged that the two largest fish, 91 and 86 cm 

 FL, were outliers and developed predictive 

 equations for egg count without using outlier data. 



753 



