564 
Fishery Bulletin 99(4) 
of bias to standard error was less than 0.25 (Efron and 
Tibshirani, 1993). However, the daily instantaneous mor- 
tality rate (2=0.29) from the bootstrap was higher than 
0.21 from the original data. The bias-corrected 2 was 0.13 
(SE=0.12). For the entire survey area, the bootstrapped P 0 
was 2.60/0.05 m 2 (SE=0.71; CV=0.27), similar to the origi- 
nal estimate (2.57/0.05 m 2 ). 
The bootstrap results indicated that the weighted non- 
linear regression produced a downward biased estimate 
ofP 0 (bias=-1.26=3.50-4.76) in the high-density stratum. 
The standard error (1.60) of P 0 from bootstrap was also 
higher than that computed from the original data (0.86). 
The fact that the ratio of bias to standard error (1.26:1.60) 
was greater than 0.25 may indicate that the weighted non- 
