(^) Before using this relationship, we must examine its main assumption, that Y/X is
proportional to B. Note immediately that this is a dangerous equation statistically,
since even for two series of random variates, Yi and Xi, Yi/Xi will be negatively
correlated with Xi. That is, Y/X can be artifactually correlated with X. Fortunately, in
many cases the relationship is close enough not to be attributed to this artifact alone.
One way is to examine Y/X in a fishery that is somehow associated with a total count
of the stock. The Bristol Bay sockeye salmon (Oncorhynchus nerka) fishery fits the
ff
(ff)
#1