FISHERY BULLETIN: VOL. 87, NO. 3, 1989 



Max 



Average 

 Min 



10 



20 30 UO 



Age in days 



Figure 5. — The average daily growth rate (mm/day) estimated from 

 otolith analyses (■) with ±1 SD. The minimum (min), average, and 

 maximum (max) growth rate (mm/day) calculated from Pacific herring 

 surviving to the termination of the experiment (day 62) are indicated. 



and the actual age of the larvae; this difference 

 corresponds well with the end of yolk-sac stage 

 of the same larvae (Wespestad and Moksness fn. 

 1); it is three days later than that found by 

 McGurk (1984b) for Pacific herring larvae at the 

 same temperature (8.0°C). 



The results are the same as earlier investiga- 

 tions on Norwegian spring spawning herring in a 

 4,400 m"^ outdoor basin (Gj0saeter and 0iestad 

 1981) in which herring form one otolith incre- 

 ment per day. However, the results of one incre- 

 ment per day contradicts findings from labora- 

 tory experiments in which otolith increments 

 were not formed daily. Geffen (1982) on Atlantic 

 herring and McGurk (1984a) on Pacific herring 

 found the growth rate of the larvae and the num- 

 ber of increments in the otoliths to be correlated, 

 while our results show a poor correlation be- 

 tween growth and daily increments. The ex- 

 pected standard deviation was ± 1 day, based on 

 the observed range in hatching time; however, 

 the estimated variation was much greater than 

 this. A possible explanation for this discrepancy 

 might be that the minimum growth rate (0.31 

 mm/d) observed in this study was gi'eater than 

 growth rates observed in laboratory studies. 



Geffen (1982) reported that gi'owth coefficient 

 tended to increase with the increase in size of 

 aquarium used — from the 120 L laboratory tanks 

 up to 4,000 m^ mesocosms. The difference in data 

 between experiments using small and experi- 

 ments using large rearing tanks results from an 

 inability of larvae to form daily increments at 

 low growth rates (Moksness et al. 1987). In our 

 work, the average growth rate corresponded to 



the gi'owth rate observed in nature. Checkley 

 (1982)' reported otolith increments and fish 

 length for juvenile herring captured in Bristol 

 Bay in autumn 1981. From these data, we esti- 

 mated the average daily growth rate over the 

 first summer of life to be 0.74 mm/d, which is 

 similar to the average growth rate observed in 

 this experiment. Therefore, it appears that con- 

 ditions in the basin were similar to the average 

 conditions larvae experience in its natural 

 habitat. 



Otolith increment size was well correlated 

 with the measured growth in standard length of 

 the larvae. The preferred growth model (see 

 Figure 4a), gave a good fit to the observed val- 

 ues, however small the sample size and espe- 

 cially for the smallest larvae, but the growth 

 model did not rule out the applicability of other 

 models. An exponential model might fit better 

 with more available data. When fitting the data 

 on dry weight of the larvae to the radius of the 

 otolith, a very good fit was observed for the 

 exponential equation. 



The resulting daily length increment from the 

 fourth order polynomial growth model approxi- 

 mated the calculated daily length increment 

 based on observed length-at-age reported by 

 Wespestad and Moksness (fn. 1). By estimating 

 the relationship between the standard length of 

 the larvae and the radius of the otolith, the daily 

 length-increment of the fish could be described. 



^Checkley, D. M. 1982. The ageing of juvenile Pacific 

 herring by otolith analysis. Final Report, NOAA contract 

 82-ABA-lOOl. Northwest and Alaska Fisheries Center. 



512 



