Myers et al.: Population growth rate of Gadus morhua 
765 
time series of data for the Gulf of Maine (NAFO 5Y) 
is not included in the comparative analysis because 
the time series is too short (less than 10 years) (Myers 
et al., 1995b). 
Maturity and weight-at-age were estimated from 
research surveys carried out for each population. 
Typically hundreds of fish are measured from a strati- 
fied random or systematic design each year. Adult 
natural mortality (M) has been investigated by a 
variety of methods; for each of the 20 populations, 
the value of M= 0.2 was accepted by the researchers 
studying each population. This natural mortality 
corresponds to adult survival of p s = e~ M ~ 0.8 (which 
will be our working value). The uniformity (across 
populations) of acceptable values for p s suggests that 
one can safely assume that the true p s lies in the 
range 0.7 to 0.9, as shown in Figure 1. 
Bottom temperature was estimated by Brander 
( 1994) for all the populations except the Baltic popu- 
lations. We used data in Dickson et. al. (1992) to make 
estimates for the Baltic populations. We estimated 
independent temperatures for the Northwest Atlan- 
tic from deYoung et. al. (1994). These changes were 
small (less than 2°C) and did not have an important 
effect on our findings. 
Estimation 
The slope at the origin, d , and other parameters of 
the Ricker model were fitted by using maximum-like- 
lihood estimation and by assuming lognormal vari- 
ability in the residuals (Hilborn and Walters, 1992); 
the assumption of gamma variability in model re- 
siduals resulted in only minor changes in the esti- 
NORTH ATLANTIC 
ON AN AZIMUTHAL EQUAL AREA 
PROJECTION CENTERED AT 
40° N AND 35° W 
Figure 2 
Map of the North Atlantic showing the regions used to define fish populations for fisheries management. 
