Woodbury and Ralston. Growth rates and birthdate distributions of Sebastes spp. off central California 



525 



during preparation. Otoliths were read 

 only once due to the destructive nature 

 of the preparation method. 



A photomicrograph of a shortbelly 

 rockfish otolith (Fig. 1) shows increments 

 that we interpreted to be produced daily. 

 A dark ring followed by narrow, closely 

 spaced rings was used as the starting 

 point for enumeration of increments. The 

 time of first increment formation is spe- 

 cies-specific. Cases are known where fish 

 larvae form the first increment at hatch, 

 yolksac absorption, first feeding, or at 

 other times (Jones 1986). Penney and 

 Evans (1985), in their study of larval red- 

 fish (Sebastes spp.) off Newfoundland, 

 noted a "heavy ring composed of a wide, 

 translucent band followed by a promi- 

 nent, high-contrast dark band" which 

 they used as the first increment for age 

 determination. Since all pre-extrusion 

 larvae lacked the dark band and all 

 planktonic larvae possessed it (some as 

 young as 1 day), they assumed that it 

 formed at the time of extrusion. Struc- 

 turally, their extrusion check was quite 

 similar to the marks seen in this study 

 (Fig. 1). They, like us, also observed a 

 number of thin, weakly expressed "pre-extrusion 

 rings" laid down prior to the first increment (see also 

 Campana et al. 1987). There are instances in which 

 small (~4. 8-5. Omm) field-caught planktonic larvae of 

 shortbelly rockfish lack the conspicuous first increment 

 (T. Laidig, SWFC Tiburon Lab., pers. commun.). Faced 

 with this uncertainty, we conclude that the first incre- 

 ment used in our counts forms some time between 

 extrusion and first feeding. The distance of this first 

 ring from the center of the focus varied among species 

 (i.e., 12-17/tfn), but was consistent within a species. 

 The first few increments laid down after this early 

 check ring typically were about 0.5/^m in width and 

 they increased in size thereafter. 



Several lines of evidence indicate that the periodicity 

 of otolith increments we counted was daily. For one, 

 all back-calculated birthdate distributions fell within 

 known annual parturition (spawning) seasons (Wyllie 

 Echeverria 1987, MacGregor 1986). In addition, results 

 from a separate study that sequentially sampled lar- 

 val and juvenile shortbelly rockfish in 1989 showed that 

 one increment was deposited each day from May to 

 June (Laidig et al. In press). Lastly, Yoklavich and 

 Boehlert (1987) validated the existence of daily growth 

 increments in juvenile S. melanops using autoradiog- 

 raphy to detect injected 45 Ca, as well as by tetracy- 

 cline injection. 





10 pm 



Figure 1 



Photomicrograph of a shortbelly rockfish otolith showing the first increment 

 (arrow) and subsequent development of increment microstructure. 



Under the assumption that one increment was pro- 

 duced each day, growth rate was measured by regress- 

 ing SL (mm) against estimated age (days). Analysis of 

 covariance was used to test for differences in the rela- 

 tionship of SL and age within and among years. When 

 differences were found, pairwise comparisons of class 

 means were made using the least significant difference 

 (LSD) test (Snedecor and Cochran 1967). This test 

 maintains the comparisonwise error rate at a = 0.05, 

 but, depending on the number of means compared, in- 

 creases the experimentwise probability of committing 

 a Type I error. In addition, indices of annual growth 

 performance were obtained by predicting standard 

 length at a selected standard age using regression 

 statistics from each individual year. These were evalu- 

 ated using principal component analysis (Green 1978) 

 to reveal interdependency among the data. In this 

 analysis, species were treated as variables, years com- 

 prised cases, and components were computed from the 

 correlation matrix. 



Back-calculated birthdate distributions of the fish 

 that survived until the time of sampling were obtained 

 by: (1) linearly regressing age on SL using the sub- 

 sample of aged fish, (2) using the regression equation 

 obtained to estimate the age of fish sampled during 

 May-June cruises only, (3) back-calculating to the 

 calendar date of first increment formation by differ- 



