EXPERIMENTAL AND THEORETICAL APPROACHES 263 



lab. and studied should yield a steady value lying within the field range but 

 much more definable. This did happen in Daphnia. A coarse field estimate 

 may be used to give an idea of a range, and a species can then be selected 

 and studied in the laboratory to clarify the picture. Laboratory studies are 

 also valuable if a species or system can, under such conditions, be examined 

 outside its natural range. Furthermore, some phenomena may be demon- 

 strated not to occur in the lab. under conditions which permit the assumption 

 that they also do not occur in the field. 



R. J. H. Beverton: I think Slobodkin's approach and my own are 

 complementary. His system is a more complete one: mine takes a natural 

 system and attempts to break it into functions. 



D. Chitty: In experimental work on populations of small mammals a 

 great deal of confusion has arisen from failure to appreciate the point made 

 by Slobodkin. It may be possible to refute hypotheses in the laboratory; 

 but it is a serious error to assume that what happens in the lab. is necessarily 

 relevant to the field. Just as field data should suggest experiments, so the 

 experiments should provide lines of thought which should be rigorously 

 tested in the field. 



Even in the laboratory it is extremely hard to determine causes of death; 

 in the field it is often impossible to find the corpses. A common error is to 

 suppose that because population size is sometimes related to food resources 

 that surplus animals therefore die of starvation. There is no more justification 

 for such an inference than there would be for saying that animals die of cold 

 or damp simply because population size might be related to temperature or 

 relative humidity. 



E. D. Le Cren: It is always hard to tell why an animal dies. But is the 

 precise cause important? Suppose that in a dense population there is a 10 per 

 cent survival of young, and in a sparser one, 50 per cent survival. Then the 

 rate of survival will be proportional to density, tending towards the main- 

 tenance of a constant number of individuals. A starved fish larva may survive 

 in a perfect artificial environment when it would fall easy prey to a predator 

 in nature. The slow growth which Beverton showed in a crowded popula- 

 tion results in longer exposure to a predator, but the two are alike conse- 

 quences of limited resources whatever the nature of the actual mortality 

 controlling numbers. 



Field experiments may not be feasible working with large populations in 

 the whole North Sea, But under some conditions, as in small streams, or 

 lakes, they are possible and may provide guidance as to the soundness of the 

 extrapolation from the laboratory to the full-scale field problem. 



K. R. Ashby: It is certainly hard to define causes of death. People 



