THE DYNAMICS OF EVOLUTIONARY CHANGE 



These two postulates are complementary to each other. To- 

 gether, they express the conclusion that in sexually reproducing 

 organisms both mutation and genetic recombination are essen- 

 tial contributors to the gene pool. This can be defined as the 

 supply of hereditary variation in a population on which nat- 

 ural selection can act. The existence of a great store of varia- 

 bility in the gene pool is best shown by two types of experi- 

 mental evidence: first, from experiments with artificial selec- 

 tion and second, from experiments which have uncovered con- 

 cealed genetic variability by obtaining certain chromosomes or 

 chromosome segments in the homozygous condition. 



To understand the meaning of recent experiments with 

 artificial selection, we should first make predictions as to what 

 the limits of selection would be on the basis of two contrast- 

 ing hypotheses. One hypothesis assumes that the gene pool is 

 very small, so that selection changes populations by increasing 

 the frequency of favorable mutations just as soon as they occur, 

 and cannot be effective unless favorable mutations are constantly 

 occurring at a reasonably high rate. We may call this the hy- 

 pothesis of direct selection of favorable mutations. Either by 

 direct statement or by implication, we can recognize this hypo- 

 thesis as basic to the thinking of De Vries, Johannsen, Bateson, 

 Morgan, and most geneticists of a generation ago. 



The second hypothesis assumes that the gene pool is very 

 large, so that, if a population is placed into a new environ- 

 ment and subjected to selection in a new direction, new gene 

 combinations are sorted out from the supply of genes already 

 present in the gene pool, and selection can be effective for many 

 generations even if no new mutations occur. We can call this 

 the hypothesis of indirect action, since it assumes an indirect 

 rather than a direct connection between mutation and selection. 



If evolution usually proceeds by direct selection of new mu- 

 tations, then the limits of selection would be set by the rate of 

 favorable mutations, which is always very low. Selection would 

 change the population for only a very few generations. This 

 assumption was made by De Vries, Bateson, and particularly 

 Johannsen, and was apparently borne out by Johannsen's well- 

 known experiments with the garden bean. On the other hand, 

 if the gene pool of most populations is large and diverse, then 

 selection can be expected to change a population in the same 



43 



