Longevity and Mortality Rates of Fish in Nature 167 



greater size; they have a lower K than the females, but a 

 higher mortality rate. In^loAce {Pleuronectes platessa) of the /f 

 North Sea, and perhaps in other species, the sexual difference 

 in mortality rates is not simple; thus in both sexes the rnortalr 

 ity rate tends to vary with the age of fish, but whereas in i 

 "rnales it increases with increasing age — at least from the fifth 

 To about the fourteenth year — in the females the evidence is 

 that it may even decrease. A species of mosquito fish, 

 Gambusia holbrooki, gives evidence that males are more 

 susceptible than females to adverse conditions of temperature, 

 oxygen, ion concentration and presence of cyanide. The 

 females also overwinter more successfully and are less severely 

 affected by catastrophies due to unknown causes (Geiser, 

 1924). It would be interesting to know whether similar sexual 

 differences have been noticed in the many tests which have 

 been made with several fish species of the toxicity of various 

 substances, especially those found in polluted water; we have 

 not, however, found information of this kind in the published 

 reports of such experiments. 



Analysis of growth rates within populations of the same or 

 closely related species living in different areas suggests that 

 two factors account for most of the variation found : food and 

 temperature. The asymptotic size is greatly modified by the 

 supply of food available, but this does not affect the para- 

 meter K. Differences in environmental temperature, however, / / 

 aftect both K and L^; thus w^th an increase in water tempera- 

 ture K increases roughly proportionally with the logarithm 

 of temperature and L^ decreases, but to a lesser extent (see 

 Taylor, 1958; Holt, 1959a). 



This temperature relation at least partly explains the 

 statement often repeated in fisheries literature that in warmer 

 waters the fish tend to be smaller than in cooler waters but, 

 equally, that they grow faster in the former (see e.g. d'Ancona, 

 1937; Gunter, 1950); the size distribution that is actually 

 observed at any time depends, however, on the mortality 

 rate as well as the growth pattern. There are rather few data 



