Following introductory comments expressing doubts about the validity of genetic load and Haldane's "cost of natural selection," the role of selection (expressed as the average number of adult daughters per female) on gene frequencies in populations has been partitioned into population and time arenas. The population arena (a geometric plane) deals with the fitnesses of different genotypes under the many situations encountered by individual members of the population in a single generation; average fitnesses of carriers of various genotypes are obtained by calculating across these many situations. The population arena includes the point signifying that, on the average, each mother leaves one daughter as her replacement within the population. It is the plane within which evolutionarily significant norms of reaction exist. The time arena is also a (geometric) plane, one that is composed of the edge-on limit (average fitness) of each successive population arena. It does not include the effects of individual situations on relative fitnesses within each population arena; it encompasses only the temporal sequence of average relative fitnesses. Amino acid substitutions in proteins and base-pair substitutions in DNA are events of concern in the time arena; within the population arena, however, gene action (not merely gene structure) is a matter of considerable concern. Thus, the discussions of the 1950s and 1960s regarding genetic variation which were reasonable within the population arena seem less so within the time arena where structural, rather than functional, variation is stressed. The function-structure dichotomy is entangled with the neutralist-selectionist controversy.