where L, = standard length (mm) at age t (d), L,, = 

 initial length (^-intercept), and A^^ and a are fitted 

 parameters (Table 2). 



This sigmoid curve suggests relatively slow 

 growth to an age of about 50 d and a length of about 

 25 mm SL, followed by rapidly accelerating growth 

 through the juvenile stage, an inflection point at 

 113.2 mm, and an asymptotic length near 307.8 mm. 

 Since sablefish achieve lengths to 100 cm (Hart 

 1973), these results should not be extrapolated 

 beyond the ages in the present study. Also, the 

 predicted fit of zero age individuals (Lq) is 1.22 mm 

 SL (Table 2; Fig. 3). This value does not accurately 

 reflect the length of sablefish at hatching. Egg size in 

 sablefish ranges from 1.8 to 2.2 mm and newly 

 hatched larvae are 5 to 6 mm (Mason et al. 1983). If 

 daily increments are first laid down at first feeding 

 as in some other species (Laroche et al. 1982), then 

 this intercept is clearly an underestimate. Mean egg 

 size suggests a length at first feeding of about 8 mm 

 (Shirota 1970). The smallest larva taken in the pres- 

 ent study was 9.8 mm SL (Fig. 3). This part of the 

 curve may be related to the inclusion of the older, 

 slower growing neustonic specimens. Another factor 

 may be effects of shrinkage; small specimens were 

 preserved in ethanol, older juveniles frozen. The 

 magnitude of shrinkage for A. fimbria is unknown, 

 but capture and preservation of other fish larvae 

 causes shrinkage which decreases with increasing 

 age or size (Theilacker 1980). Thus increases in ac- 

 tual length for small individuals may have been 

 relatively greater, changing the fitted equation and 

 possibly increasing the length at time zero (Fig. 3). 



Heyamoto (1962) estimated growth for young 

 sablefish, suggesting that specimens 12.3 to 16.4 cm 

 FL (11.1 to 14.8 cm SL) were 6 mo old. His data, 

 however, were based upon estimating the age at 

 collection by difference between capture and an 

 assumed spawning season. In our study, 6-mo-old 

 specimens were > 24 cm SL. The specimens cap- 

 tured by Heyamoto (1962) were taken by trawl in 

 320 to 412 m, much deeper than the epipelagic 

 juveniles in our study. Beamish et al. (1983) used 

 daily increments as part of a study to validate an- 

 nulus formation in sablefish. In nine specimens 23 to 

 27 cm FL (208 to 245 mm SL), they observed from 

 270 to 350 (mean 313) increments but suggested 

 that the fish were 1 yr old due to the inability to 

 count all increments. Based upon our growth curve 

 (Fig. 3), their ages would be overestimates. 



Recent observations of laboratory growth are in 

 substantial agreement with growth described by our 

 curve. Shenker and Olla^ found average growth 

 rates as high as 2.3 mm/d for juvenile sablefish fed ad 



Table 2.— Fitted parameters of the Laird- 

 Gompertz growth model for larval and juvenile 

 Anoplopoma fimbria in the present study. The 

 curve is fitted to all larvae and juveniles (W = 

 105) based upon counts of otolith increments. 



Parameter 



Estimate 



Asymptotic 

 standard error 



1.2203 

 0.1084 

 0.0196 



0.4675 

 0.0146 

 0.0015 



libitum. These fish were near the lengths where our 

 curve predicts fastest growth (2 mnVd, Fig. 3). High 

 growth rates were also observed for fish smaller 

 than 25 mm, where our data suggest relatively slow 

 growth. Grover and Olla'* noted starvation of field- 

 collected sablefish larvae based upon morphological 

 criteria; thus food probably limits sablefish growth in 

 the field. This species apparently has a great scope 

 for growth given high laboratory rations or patches 

 of high prey density in the field. 



The distribution of dates of first increment forma- 

 tion were estimated by back calculating from the 

 ages of all specimens in our study. Since larvae and 

 juveniles were from different years and sampling 

 gears, it is possible that differences would be ob- 

 served in this distribution. Since the plankton gear 

 selects for smaller larvae due to avoidance by later 

 stages, the results could be biased if the entire 

 spawning season were not sampled. The median 

 dates for the 1982 larvae (8 April) and the 1981 

 juveniles (18 March), however, were similar. Thus all 

 105 samples were combined and the distribution of 

 the dates of first increment formation plotted (Fig. 

 4). The distribution has a mode in early April. If the 

 first increment is formed in association with first 

 feeding, as in most other species studied (Brothers et 

 al. 1976; Taubert and Coble 1977; Laroche et al. 

 1982), then the spawning dates would precede the 

 distribution in Figure 4. Ware (1975) provided an 

 egg size-incubation time relationship for fishes; 

 sablefish, with a 2 mm egg, would have an incubation 

 time of 13 d. If a similar time is spent in yolk absorp- 

 tion before first feeding, peak spawning would occur 

 in early March. This generally agrees with most 

 other reports of the spawning season for A. fimbria. 



^Shenker, J., and B. L. 011a. Laboratory growth and feeding of 

 ]uwem\e ssb\eT\s\ Ano^lcrpoma fimhris. Unpubl. manuscr. 



■•Grover, J., and B. L. OUa. Field evidence for starvation of larval 

 sablefish, Anoplopoma fimbria. Manuscr. in prep. Northwest 

 and Alaska Fisheries Center, Newport Field Office, National 

 Marine Fisheries Service, NOAA, c/o Marine Science Center, 

 Marine Science Drive, Newport, OR 97365 (direct correspondence 

 to B. L. 011a). 



479 



