442 INTRODUCTION TO EVOLUTION 



had become fixed. When fixation occurs the drift becomes irreversible, but 

 at any point before fixation is reached reversal of drift may occur. 



A model to illustrate the operation of "drift" was constructed by 

 Dubinin and Romaschoflf (described in Dobzhansky, 1941, p. 162). In this 

 model the gene pool was represented by 100 marbles in a bowl. Each mar- 

 ble bore a different number. In order to simulate the loss of 25 percent of 

 the mutant genes described above (p. 440) the investigators discarded 

 from the bowl twenty-five marbles, taken at random. In order to simulate 

 the doubling in frequency of 25 percent of the mutant genes (p. 440) they 

 withdrew twenty-five more marbles at random and then replaced them, 

 accompanying each marble by a second one bearing the same number. In 

 this way the total number of marbles remained 100, but 25 percent of the 

 numbers designating individual marbles were lost, and 25 percent of the 

 designating numbers were doubled in frequency. This procedure repre- 

 sented the action of chance in the production of one generation of off- 

 spring. The process was then repeated time after time. As the "genera- 

 tions" passed it was found that fewer and fewer different numbers 

 remained in the bowl, until finally all 100 marbles came to bear the same 

 number. This culmination was reached in from 108 to 465 "generations," 

 in different experiments. 



In order to demonstrate the influence of size of population upon drift 

 Dubinin and Romaschoff repeated the experiment with a gene pool con- 

 sisting of but 10 marbles. In this case "homozygosis" (all marbles having 

 one number) was attained much more rapidly than it was in the larger 

 gene pool, only fourteen to fifty-one "generations" being required. This 

 observation emphasizes the point that drift is primarily a phenomenon 

 characteristic of small breeding populations. 



Instructive as is the model just described, the present author felt that a 

 model mimicking more closely the actual conditions of bisexual reproduc- 

 tion might have enhanced value. He also wished to avoid the artificial reg- 

 ularity imposed by discarding 25 percent of a gene pool, and doubling 

 another 25 percent, at each generation. Accordingly, he devised a simple 

 model in which chance was free to operate in two phenomena at each 

 generation: (1) in arranging of matings and (2) in production of off'spring 

 from these matings (Moody, 1947). In the model, individuals were repre- 

 sented by pairs of beads tied together; two red beads stood for a homo- 

 zygous dominant individual (MM); two blue beads for a homozygous re- 

 cessive (mm); a pair consisting of one red and one blue bead represented 

 a heterozygous individual (Mm). 



The model began with a small population conforming to the Hardy- 



