NATURAL SELECTION 



647 



favorable case, while under artificial selec- 

 tion such combinations are established in 

 relatively few years. If a dominant gene 

 increases the fitness of its carriers by one 

 part in 1000, its frequency in a population 

 mating at random will increase from 0.001 

 per cent to 99 per cent in 23,400 genera- 

 tions. If this difference of fitness affected 

 viability, we should need a population of at 

 least 16 milhon to detect it even with the 

 degree of certainty afforded by a deviation 

 equal to twice its standard error. Simpson 

 (1944, p. 82), following Wright, says that, 

 with a selective advantage as low as 0.0001, 

 selection w'ould still be a major factor in 

 the fate of a gene in a breeding population 

 of 5000 or more mdividuals (p. 654). Such 

 survival differences would obviously be 

 hard to detect within a population; but af- 

 ter frequencies have become established in 

 different populations, it may then be possi- 

 ble to measure the survival values that the 

 evolutionary process has brought about. 

 Under artificial selection, divergencies may 

 be detected much more rapidly. 



Some investigators (Errington, 1943, p. 

 903; Pearl, 1930a, p. 178) seem to think 

 that minor agents contributing to the total 

 mortality can have no appreciable effect 

 upon the direction of evolution, while the 

 major agents causing the greatest mortality 

 may select without genetic discrimination. 

 Dewar (quoted by Pearl, 1930a) says: 

 "From the facts that the greatest destruc- 

 tion is to eggs and young birds and that 

 the forces which destroy adult birds for the 

 most part act indiscriminately as opposed 

 to selectively, the inference must be drawn 

 that, speaking generally, the individuals 

 which survive longest in the struggle for ex- 

 istence are the lucky ones, rather than the 

 most fit." Great mortafity, without involv- 

 ing genetic differential survival, must be 

 taken into account (SaHsbury, 1936). A 

 single giant puffball produces about 

 7,000,000,000,000 spores, which are distrib- 

 uted by the wind into many lethal habi- 

 tats. Large numbers of seeds of plants are 

 distributed at random by the wind, and 

 those that happen to alight on a favorable 

 spot may sprout and grow to reproductive 

 age, while all others die regardless of the 

 favorable gene arrangements in their cells. 

 Eggs of some invertebrates and fishes are 

 distributed almost as completely at random. 



Haldane's mathematical analysis (1936) 



indicates that, even though individual sur- 

 vival may in large part be haphazard, the 

 influence of selection may still guide adap- 

 tive evolution as rapidly as is indicated in 

 nature. The horns of titanotheres, the first 

 mutations in the direction of mimetic re- 

 semblance, and slight changes in ammonite 

 sutures, have been used as arguments 

 against the selection theory, because of the 

 probably small initial advantage of such 

 characters. But as has been stated, small 

 selective coefficients can influence adaptive 

 evolution. 



Beside the misinterpretations of the 

 principle of preadaptation, which go to the 

 length of assuming a small role for selec- 

 tion in the evolution of adaptation, the 

 concept has sometimes been grossly misap- 

 pHed. The fact that complex structures are 

 sometimes inhibited in development by 

 simple genetic factors has been interpreted 

 by some to signify that these complex 

 structures can arise in phylogeny through 

 simple gene mutation. Villee (1942, p. 

 168) states: "In the course of phylogeny, 

 one major and a few minor changes in 

 genotype can suddenly produce dorsal ap- 

 pendages and then change them into wings 

 or halteres, thus demonstrating that macro- 

 evolution is possible without the accumu- 

 lation of micromutations under the pressure 

 of selection." Villee thus agrees with the 

 equally startling conclusions of Gold- 

 schmidt (1940), in his discussions of mac- 

 roevolution (see Wright, 1941a, for a criti- 

 cal review). 



Even though simple genetic mutations 

 may change wings into halteres or elim- 

 inate them altogether as visible structures, 

 this is no indication that the wings are the 

 product of a few genes. A single mutation 

 might influence a developmental threshold 

 that makes possible the growth of a charac- 

 ter that has been somatically, but not genet- 

 ically, lost. If, as has been shown in many 

 Hymenoptera, a haploid egg usually pro- 

 duces a male, while a diploid egg usually 

 produces a female, it would be a rather 

 rash conclusion (see Whiting, 1945; White, 

 1945) to assume that sex evolved in one 

 jump by a mere doubling or halving of the 

 chromosome number (see discussion by 

 Goldschmidt, 1945). Such structures as 

 wings, eyes, appendages, and sexual or- 

 gans of Drosophila are modified by a large 

 number of genes located in various chro- 



