230 POPULATIONS OF THE SEA 



Freshwater en\'ironments are even more generally discontinuous 

 than land environments. Whether a piece of freshwater is large or 

 small — a lake, a pond, or a pool of water in a gutter — it is always 

 separated from similar environments by land or by dissimilar 

 aquatic environments such as a stream or river with conditions 

 to which the fauna of the lake or pool will not be adapted. The 

 result is that by far the greater number of terrestrial and fresh- 

 water species are distributed in local populations isolated more or 

 less completely from other populations of the species. Both the 

 size of the local populations and the degree of their isolation are 

 greatly variable, but it is only some of the most highly locomotory 

 species, some birds, some insects, and the larger and especially 

 the carnivorous mammals, that may not be distributed in these 

 local populations. In these species the population may be reduced 

 to single breeding pairs or families. These local and more or less 

 isolated populations have been called "demes." 



There must undoubtedly be in general some migration, active 

 or passive, between the demes; otherwise new habitats suitable 

 for the species would never be colonized. But recent work has 

 shown that the distances over which animals wander are less than 

 might at first sight be thought. Small mammals such as field mice 

 or lemmings do not wander more than 50-100 yards from their 

 homes, and among the less active invertebrates snails have been 

 found to move not more than 5-10 yards in a year. The isolation 

 of the demes is therefore real, though it is not often complete. 



From the point of view of the study of evolution, the interest 

 of these facts lies in the realization that it is in the demes that 

 differentiation is initiated. Provided the isolation is sufficient, two 

 demes will evolve independently so long as they remain isolated 

 from each other. For estimates of the sufficient grade of isolation 

 we have the work of Sewell WVight (1943), who shows that a 

 mutational difference of selective advantage \/n* will evolve 

 independently in two populations if the migration between them 

 is not more than 1/n per generation. Complete isolation is not 



* An advantage of \/n means that n + I indi\iduals bearing the mutation 

 survive to the next generation for e^■ery w individuals without it. 



