704 



Fishery Bulletin 103(4) 



/ 4 (cm) 



E 29- 



B 



• PB 

 O LB 



26 



~l 

 27 



/ 7 (cm) 



Figure 3 



Bootstrap estimates of reparameterized von Ber- 

 talanffy growth function mean lengths at age for 

 Notolabrus fucieola, by site. (A) / 4 versus /- (B) l- 

 versus / 10 mean length-at-ages at 7 and 10 years. 

 Contrasting crosses show the location of parameter 

 estimates based on the original data set (+, PB = Point 

 Bailey, x, LB = Lord's Bluff). 



25 



20 - 



15 



10 



05 



00 

 -0 05 

 -0.10 - 

 -0 15 

 -020 



-0 25 



"I I I I T~ 



2 5 3 3.5 4 4 5 



g 20 (cm/yr) 



B 



05 1 



U 



Figure 4 



Bootstrap estimates of GROTAG parameters for Noto- 

 labrus fucieola, by site: (A) g 20 versus g 30 , mean 

 annual growth at initial length 20 and 30 cm and 

 (B) u versus w, magnitude and timing of seasonal 

 growth. Contrasting crosses show the location of 

 parameter estimates based on the original data set 

 (4-, PB = Point Bailey, x, LB„ =Lord's Bluff). 



at or=0.05 when tested individually (Table 8A), in agree- 

 ment with the results of the randomization tests. 



Sex comparisons Bootstrapped parameter estimates 

 from sex-specific data sets were approximately sym- 

 metrical about the original estimates (Table 7). The 

 largest divergence from 0.5 was evident in estimates of 

 s for females at Lord's Bluff and males at Point Bailey. 

 Bootstrap estimates of u for Lord's Bluff males occasion- 

 ally extended into spurious negative values, lowering 



confidence estimates of the extent of seasonal growth 

 in this data set (Table 7). 



Based on simple overlap of CIs, no single parameter 

 differed significantly between sexes at either site (Ta- 

 ble 7). Plots of the bootstrap estimates of the growth 

 parameters g., and g., 8 showed minimal overlap between 

 males and females, and separation was most evident 

 along the g 20 axis (Fig. 5A). Plots of bootstrapped es- 

 timates of the seasonal growth parameters u and w 

 (Fig. 5B), and the measurement error parameters m and s 



