584 



Fishery Bulletin 92(3). 1994 



mean SL, variance, and corrected 

 mean SL for all samples collected 

 in the upper 100 m. The corrected 

 mean SL will be hereafter referred 

 to as mean SL. 



A nonparametric ANOVA by 

 ranks showed that mean SL dif- 

 fered significantly between the 

 shallow samples from each sta- 

 tion (Kruskal-Wallis test statis- 

 tic=746.5, P<0.001 ). A Tukey type 

 nonparametric multiple compari- 

 son (Zar, 1984) indicated signifi- 

 cant differences (P<0.05) between 

 the innermost station pair (RES 1 

 and RES 2) and each of the sta- 

 tions outside RES 2 (RES 2.5, 3, 

 4, GAK 1 ). Among the outside sta- 

 tions, the only significant differ- 

 ence was found between RES 3 

 andGAKl(P=0.003). 



When samples from all depths 

 were pooled and mean SL com- 

 pared between stations, results 

 were very similar. An ANOVA 

 showed a highly significant differ- 

 ence in mean SL between the stations (F=80.00, 

 P<0.001). A Tukey HSD multiple comparison again 

 indicated that significant differences (P<0.05) exist 

 between both of the two innermost stations and any 

 one of the stations outside RES 2. 



Larvae at stations RES 1 and RES 2 were signifi- 

 cantly larger and older than those at stations out- 

 side RES 2. The observed size differences translate 

 into an age difference of 8.5 days between the aver- 

 age at the two inner stations (RES 1 and RES 2) and 

 that at the outer stations (RES 2.5, 3, 4, and GAK 1 ). 

 Age was calculated by using growth equations ob- 

 tained in this study. The average age of larvae col- 

 lected at stations RES 1 and RES 2 was estimated at 

 15.1 days. The average age of larvae at the other four 

 stations was estimated to be 6.6 days, relative to 2 

 May. Thus, the results of size and age comparisons 

 suggest that the stations can be divided into two dis- 

 tinct groups on the basis of larval size. 



Growth rates 



Growth rates were determined for larvae collected 

 1-4 May 1989 at station RES 2 in the inner basin 

 and at station RES 4 in the outer basin. At station 

 RES 2, 62 larvae collected at 7 m on 4 May 1989 

 were measured and dissected to remove otoliths, of 

 which 54 could be aged. The increment count ranged 

 from 6 to 40 increments for larvae between 5.1 mm 



and 11.1 mm SL. A linear regression model relating 

 mean SL and increment count yielded a growth rate 

 of 0.18 ± 0.028 mm/day (95% CI)(r 2 =0.75, Fig. 5), 

 assuming each increment represents growth of one 

 day. From a sample collected at RES 4, at 18 m on 2 

 May 1989, 38 larvae ranging in length from 5.3 mm 

 to 10.1 mm were aged. The growth rate at this sta- 

 tion was estimated to be 0.19 ± 0.016 mm/day 

 (r 2 =0.79, Fig. 5). 



We compared the regression lines of standard 

 length on increment count from RES 2 and RES 4 

 (Fig. 5) using the ^-statistic according to Zar ( 1984) 

 and found no significant difference between the 

 slopes (r=1.048; 0.20<P<0.50). This indicated that the 

 growth rate was not different between the two sta- 

 tions. A common slope for both data sets was computed 

 by using a weighted regression coefficient (Zar, 1984). 

 The resulting combined growth rate for all walleye pol- 

 lock larvae in Resurrection Bay was 0.18 mm/day. 



In addition to length at age (increment count), we 

 examined the relationships between otolith size and 

 increment count and between otolith size and stan- 

 dard length. The regressions of otolith diameter on 

 increment count resulted in a much tighter fit for 

 both stations (r 2 =0.85 for RES 2 and r 2 =0.91 for RES 

 4; Fig. 6). Regressions of length on otolith size indi- 

 cated a close relationship between body length and 

 otolith diameter for the limited size range studied 

 here (r 2 =0.83 and r 2 =0.86). 



