OTHER THEORIES OF EVOLUTION 321 



(a) During periods when numbers are at their lowest the 

 expectation of mutations will be correspondingly low. 



(b) Even if the chances of survival are increased during one 

 period of minimum numbers, we have still to explain the 

 phenomenon of progressive modification. If we assume that 

 a mutant has survived one period, we have still to assume that 

 a further mutation carrying the modification a step further will 

 occur in the descendants of that mutant at the next minimum. 



(c) Elton himself (I.e. p. 79) points to the objection that 

 ' at the next reduction of numbers the mutation will apparently 

 be reduced to about its original proportion in the population 

 and will never be able to spread beyond a certain point.' 



Elton (I.e. pp. 79-82) has considered two of these objections, 

 (a) and (e), and attempted to meet them ; but we are not 

 satisfied that the reasoning he adopts in this attempt is 

 sound. 



It is not to be expected that many exact observations 

 on the intensity of variation during a numerical increase of 

 population would be available. Some information on this 

 subject will be found on p. 213. So far we do not believe a 

 very strong case has been made out for Elton's theory. Never- 

 theless there is a further possibility to be considered. Robson 

 (I.e. p. 221) suggested that a non-advantageous mutation might 

 spread if its appearance happened to coincide with the occu- 

 pation of a new habitat. We know, as a matter of fact 

 (Chapter II, p. 53), that species are by no means rigidly con- 

 fined to strictly defined habitats, and that individuals are 

 often found straying into situations or adopting habits not 

 characteristic of the bulk of the species. With this tendency 

 we may consider the very definite evidence accumulated by 

 Elton (I.e.) for the frequency of migration, though in point of 

 fact in such a suitable case for studying this phenomenon as 

 the Migratory Locust no special increase of variation has been 

 noted with the swarming phase (Gause, 1927). 



In a new and relatively untenanted habitat a single 

 mutant of a given type would of course not be immune from 

 the normal risks of death, but it would be at least freed from 

 the chances of competition peculiar to a crowded habitat. 

 However, there still remains the objection (similar to (c) in the 

 criticism of Elton's hypothesis) that in order to explain sus- 

 tained change in a given direction we would have to assume 



