^24 THE ORIGIN OF SPECIES 



Besides the various types of geographical isolation, such as 

 by water, deserts, mountains and mere distance, it is probable 

 that other types may have been important in speciation. Many 

 insects have races with different food plants, which makes 

 ecological or habitat isolation. Parasites with different hosts 

 are similarly separated, and it is possible that habitat preferences, 

 reinforced by a homing instinct, are important in fishes, birds 

 and mammals, although many ornithologists would deny this. 

 Seasonal isolation is possible, where different mutants become 

 sexually active at different times of the year, and in species 

 with copulatory organs a single mutation could produce structural 

 isolation. In the higher animals there may be psychological or 

 behavioural isolation — refusal to mate with an individual who 

 does not show the characteristic recognition marks or precoitional 

 behaviour of the species. All these may lead to subspeciation 

 and then to a genetic isolation or incompatibihty ; at this point 

 or before the one species has become two. 



DRIFT 



The second discovery of the mathematicians is that both the 

 rate and the type of evolution depend on the size of the breeding 

 population or deme. Evolution will be most rapid when this is 

 moderate. When it is small (perhaps less than about looo) 

 selection becomes less important, and may be entirely overcome 

 by drift, a random spread of genes depending on the chance of 

 their mutation and survival. Not everyone is agreed on the 

 importance of drift, but if it does occur it would account for the 

 apparently useless quality of many specific characters. 



SUMMARY 



We have then a modern theory of natural selection, which 

 depends on mutation, isolation and selection ; to these some 

 writers would add drift. It sometimes escapes notice that this 

 leaves the most important part of evolution, the origin of muta- 

 tion, without explanation, and here we are Kttle better off than 

 was Darwin. We know, as he did not, that the evolutionarily 

 important part of variation is produced by changes in the genes, 

 but we can go little further. It is a reasonable guess, as we have 

 seen above (p. 700), that a mutation involves a change in the 



