Laidig et al.: Growth dynamics in early life history of Sebastes jordani 



615 



100 





(0 



C 

 (0 



80 



60 



40 



20 



SI, 



SL 



-15.7 + 0.49*Date 

 -22 2 + 0.54*Date 



10 



15 20 25 30 

 MAY 



10 14 

 JUNE 



Figure 5 



Regressions of standard length on calendar date of capture 

 for all juvenile shortbelly rockfish taken during the May-June 

 survey off central California, 1989 (solid line; r'- 0.265) and 

 for the subset of juveniles that were extruded during the 

 primary January-February spawn (open circles and dashed 

 line; r 2 0.484). Diamonds represent the fish (n 25) spawned 

 during March and April. 



-a 

 K 



: 



1000- 



S 500 



o 

 o 



60 80 100 



Aee (d) 



120 140 160 



Figure 6 



Predicted otolith radius as a function of age (solid line) from 

 the Gompertz growth model (multiple observations from each 

 specimen). Filled circles are means of the observed data at 

 10-day intervals, bracketed by ± 2 standard errors. Note that 

 standard errors for younger ages do not appear due to the 

 resolution of the figure. Dashed lines enclose the range of data. 



For the juveniles from the first cohort, the slope of 

 the regression of SL on date of capture was 0.54 mm/ 

 day (SE 0.0158; Fig. 5). This compares with a slope of 

 0.53 mm/increment (SE 0.0140) obtained from a re- 

 gression of length against the number of increments 

 counted for the 55 aged fish (Fig. 3). Moreover, the 

 slope of a regression of the predicted ages of the fish 

 from the first cohort on their date of capture was 0.96 

 increments/day (SE 0.028). This rate is not significantly 

 different (P>0.05) from the expected 1:1 correspon- 

 dence between increments and days (t - 1.43, df 1235). 



Most of the fish from the second cohort (Fig. 4) were 

 captured late in the May-June cruise, probably because 

 selectivity of the net did not allow their capture at the 

 beginning of the cruise when they were smaller. Be- 

 cause fish from the second cohort were smaller and 

 were caught later in the cruise, excluding them im- 

 proved the agreement between the observed rate of 

 change in length over time and the growth rate 

 estimated from daily otolith increments of aged fish 

 (Fig. 3). We feel that the exclusion was warranted 

 because of the clear separation between the cohorts 

 (Fig. 4). In any case, relatively few fish (25 out of 1262) 

 were excluded, and the effect on our estimated rate 

 of change in length of the cohort was slight (~10%; 

 Fig. 5). 



Growth of the otolith 



Otoliths of shortbelly rockfish grow at a generally in- 

 creasing rate from birth to ~90 days in age (Fig. 6). 



Thereafter, there is substantially less curvature in the 

 data. Results of fitting the Gompertz growth function 

 to the data yielded estimates of OR = 16.00^m (SE 

 0.0374), k = 5.4223 [dimensionless] (SE 0.0235), and 

 g = 0.01298/day (SE 0.000106), with an r 2 of 0.991. 

 The curve (solid line in figure) fit the data well except 

 for a minor lack of fit at about 70 days of age. 



Standard length vs. otolith radius 

 and back-calculation of length 



The segmented linear model (after Bacon and Watts 

 1971) provided a good fit when SL was regressed on 

 OR (Fig. 7, Table 2). There was no discernable pattern 

 to the residuals, and the r 2 value was quite high 

 (0.998). In addition, the model's estimate of body length 

 at the mean otolith radius at the extrusion check (17 

 |im) was 4.9mm SL; this approximated independent 

 estimates of length-at-extrusion (5.4 mm [Moser et al. 

 1977], 4.0mm [MacGregor 1986], and 4.7mm SL [this 

 study, Figs. 1, 2]). The intersections of the four linear 

 segments relating SL to OR were at 38.7, 73.1, and 

 431.3/im OR, and the corresponding lengths were 7.7, 

 8.4, and 29.9mm SL, respectively. 



Larvae were collected with two types (sizes) of nets. 

 A possible source of bias in our results could arise from 

 different selectivities of the nets (Somerton and Koba- 

 yashi 1989). For example, differential sampling by the 

 nets may be responsible for the segmenting seen in 

 Figure 7, although we do not believe this is the case. 

 Even though, on average, the nets collected different 

 sized fish, both types of nets captured fish in the range 



