not accurately estimate effective population sizes, rates of migration, and neutral mutation ( Lewontin 1974; Stebbins and Lewontin 1972).
Empirical results that agree with a theory may support it, but their agreement does not confirm that theory, and the predictions from competing theories may not be distinctly different. For example, the correspondence of the amino acid content of proteins to predictions arising from the redundancy of the genetic code does not--and cannot--reject the hypothesis that allelic variants of a protein significantly affect physiology and growth and play a major role in adaptation to environmental variation. Furthermore, although an excellent molecular clock would result from the fixation of neutral mutations, all approximate clock would result from a mixture of effectively neutral and adaptive mutations. And although an increase in genetic variation with population size is predicted by neutral theory, it is also predicted by models relying on natural selection to balance genetic variation.
We could reasonably begin a study by assuming that genetic variation is neutral, as this assumption helps us build null hypotheses and formulate predictions. However, if evolutionary biologists are curious about the fitness consequences of abundant genetic variation and want to test the assumption of neutrality, they must address this question directly, by examining the physiological and demographic consequences of the genetic variation ( Bennett 1987; Clarke 1975, Clark and Koehn 1992; Koehn, 1978, 1987; Powers 1987, Powers et al. 1994; Watt 1985, 1991). Accordingly, chapter 2 discusses enzyme kinetic studies, and chapter 4 summarizes studies of the biochemical and physiological consequences of molecular variation.
Although the substitution of one allele for another is all important evolutionary event ( Haldane 1957) and an event that is especially interesting to molecular systematists, substitutions are of little or no importance to population biologists, for they occur on a time scale much greater than that used by population biologists studying microevolutionary events. New mutations ( Kahler, Allard, and Miller 1984) and allelic substitutions occur so infrequently that they cannot play all important role in the daily, seasonal, and yearly dynamics that produce fitness differentials among individuals in populations. For these reasons, in this book I do not treat mutation or gene substitution but instead concentrate on the assortment and rearrangement of genes extant in the gene pool (chapters 7, 9, 10).
I begin with empirical data, as my primary motive is to consider the generality of several contrasting Population genetic models of the determination of fitness. In particular patterns in the empirical data have led me to examine the immediate consequences and the evolutionary implications of truncation or threshold selection acting on levels of heterozygosity in natural populations (chapters 6, 7, 8, 10).
Decades of studies of natural populations have firmly established that genetic variation is abundant and that selection is common. Oddly enough, we know little about the degree to which selection affects this abundant genetic variation.
New mutations and allelic substitutions occur so infrequently that their significance in the adaptation of populations to the typical range of heterogeneity ill their environments must be slight.
Questions concerning the dynamics of abundant genetic variation constitute major