POPULATION GENETICS AND EVOLUTIONARY CHANGE 447 



they sometimes exhibit, as compared to characteristics shown by large 

 populations of their relatives, and how large populations descended from 

 a few immigrants may differ from the population from which the immi- 

 grants came (Mayr, 1954; Sheppard, 1960). 



Genetic Drift and Natural Selection 



We have noted that chance may lead to increased frequency of occur- 

 rence of a gene, without regard to usefulness. In this connection it is only 

 fair to state that differences of opinion exist as to the importance of ge- 

 netic drift in producing the variations in gene frequency actually found in 

 small populations, and especially as to the importance of drift in giving 

 rise to the first steps in evolutionary change. Some investigators conclude 

 that the differences between even small populations arise through the ac- 

 tion of natural selection rather than through the action of chance in ge- 

 netic drift. 



This point of view was set forth by Ford (1949), who described the 

 spread of a gene through an isolated population of moths. In 1928 the fre- 

 quency of the gene was 1 .2 percent; by 1939 it had risen to 9.2 percent, and 

 in 1940 to 11.1 percent, after which it fluctuated between 6.8 and 4.3 per- 

 cent. The isolated population fluctuated in numbers between 1000 and not 

 over 8000, but the fluctuations in population size did not afl'ect the fre- 

 quency of the mutant gene under study. Ford concluded, "It has been pos- 

 sible to show that the chances are less than one in a hundred that the ob- 

 served variations in the gene-ratio could be produced by random survival, 

 indicating that they must be due to the influence of selection, varying in 

 direction and intensity from year to year. This is the first time that these 

 alternatives have been studied experimentally, and the result gives no 

 support to the view that random survival plays a significant part in evolu- 

 tion in populations of 1000 individuals or more." Furthermore, he feels 

 that populations of smaller size than this are not permanent, are particu- 

 larly liable to extinction, and hence are not likely to contribute signifi- 

 cantly to evolution. 



Doubtless the conditions of effective population size vary from animal 

 to animal; an effectively small population of butterflies may not have the 

 same size range, in absolute numbers, as an effectively small population 

 of mice, for example. Extremely small populations are sometimes re- 

 corded. Thus Miller (1950) recorded a unique dwarf species of fish con- 

 fined entirely to a single hot-spring hole in Death Valley. "Since every fish 

 is in view at one time, reliable estimates of the total population are possi- 



