58 S. J. HOLT 



samples can be determined are well known; methods of estimating mortality 

 coefficients in similar circumstances were reviewed by Beverton & Holt 

 (1956) and there is not, I think anything fundamentally new to add to that 

 review, though there has since that date been some refinement of certain 

 methods [see particularly documents of the International Commission for 

 the Northwest Atlantic Fisheries by Paloheimo and by Taylor, and also 

 Paloheimo (1958) and Bishop (1959)]. 



When age determination is difficult or impossible, attention is turned to 

 the possible use of length frequency distributions of catches. These distribu- 

 tions, or 'catch curves', are, when averaged for all seasons and over several 

 years, typically peaked curves The ascending left-hand Hmb of such a curve, 

 represents recruitment into the exploited stock by growth of fish to a size 

 at which they are retained by the gear, or by movement of young fish of 

 catchable size into the fishing area. The descending right-hand hmb may be 

 interpreted in several ways, depending principally on the fishable life-span 

 of the fish, and the type of gear used to catch them. Some gears, particularly 

 gill nets and possibly hooks, do not retain large (or small) fish so well as they 

 retain middle-sized fish (these terms are of course relative, and must be 

 understood in relation to the sizes of mesh or hook in use). A first step is 

 therefore to determine the relative probabihties of retention as a function of 

 fish length, giving a 'gear selection curve', ordinates of which may be 

 divided into the corresponding ordinates of the 'catch curve' to estimate the 

 length composition of the fish population accessible to the fishing gear. A 

 general method of determining the selection curve in such cases from 

 comparative fishing experiments is now available (see Holt, i957^; Olsen, 



1959)- 



The descent of the right-hand limb of the 'accessible population curve' 

 may be due, if several year classes or broods offish are present in the popula- 

 tion, to the decline in numbers offish present with increasing age. If there is 

 only one or a very few broods present the descent will merely represent the 

 variability of size among fish of the same age, but there is usually a good 

 chance of recognizing this situation from the seasonal changes in size-distribu- 

 tion, from which indeed the growth rate might be estimated. 



When the slope of the right-hand hmb of the accessible population curve 

 is an expression primarily of decrease in number offish with age, the decrease 

 may be ascribed to mortahty or to a net emigration from the fished area, or 

 both. For the assessment of that fishery, however, the processes of death or 

 migration are equivalent; they both result in fish becoming inaccessible to 

 the fishery and may be regarded as 'mortahty'. Only when emigrating fish 

 may thus become accessible to a fishery elsewhere is it essential to distinguish 

 these processes for assessment purposes. 



