2d2 DISCUSSION 



by other fish. Young plaice, especially in their planktonic phase, must be 

 eaten by a variety of predators, but on present information it is not possible 

 to say whether the mortahty from this cause is important or whether it has 

 a controlling influence. For this latter purpose the mortahty need not, of 

 course, be directly density-dependent; intraspecific competition among the 

 plaice larvae for food, retarding growth, lowering their vitahty and prolong- 

 ing their pelagic phase, and so exposing them to predation for a longer 

 time, could also act as a controlling mechanism. It is very difficult to assess 

 the likely cause at present. 



A. C. Simpson: Sette, working on mackerel in America, has shown that 

 a large initial hatch is followed by a steady decline, which makes predation 

 seem probable (Sette, 1943). But more data are needed. 



I. A. McLaren: Pseudopkuronectes (winter flounder) populations have 

 been studied at Yale, in parallel to Sette's work. Early mortality has been 

 attributed especially to predation by Sarsia, and the loss may reach up to 

 25 per cent per day, in ten-day-old larvae. 



S. J. Holt: Both Beverton and Slobodkin have described experiments 

 with artificial populations conducted to improve understanding of the 

 dynamics of wild populations. Their experiments are of course of different 

 types — in the former case specific parameters of individual activity are 

 related to other characteristics of the individual: in the latter the experi- 

 mental population is studied as a whole. Nevertheless there is a common 

 problem of determining which aspects of the experimental results are Hkely 

 to be applicable to wild populations, and which are not. This problem will 

 become more acute as biologists wish to measure more, and more complex, 

 characteristics of populations, and resort to the use of convenient artificial 

 populations for the purpose. Now by analogy with the subject of anatomy, 

 we see that one result of comparative studies is to conclude that a rat and a 

 man are in many respects more hke one another than either is like a turtle, 

 so that since we cannot conduct certain pharmacological tests on men we 

 first test rats. Our problem in population dynamics may be to identify those 

 characteristics of populations which we may extrapolate to other populations, 

 and those which we may not. This particular need may help guide our 

 attempts to establish a comparative method in our field. 



L. B. Slobodkin: My own experiments are always designed in relation 

 to the field situation. In nature, estimates of the efliciency of ecological 

 processes are only possible with a fairly wide margin of error, owing to the 

 many assumptions necessary, but values from the field fall within a range of 

 5-15 per cent. This can be interpreted in many ways: it might be held that 

 all estimates of ecological eflficiency from nature are essentially the same 

 figure, blurred by the sampling technique. If so, any species taken into the 



