Introduction

The longevity of the neutralist–selectionist controversy in population genetics reflects the difficulty of measuring the extent of natural selection in both field and laboratory studies. The advent of powerful molecular tools for genetic analysis has advanced but not resolved this issue. Although there is extensive indirect evidence suggesting a major role for selection in molecular evolution, direct measures of these selection pressures are difficult. Indeed, most selective advantages are probably too small to be measurable in practical terms. However, much may be learned about the likely nature of very weak selection pressures, and of their effects, by analysing less typical large selection pressures that are within range of current studies.

Our extensive and ever-increasing knowledge of human molecular genetic diversity makes ours a species particularly amenable to observational studies of natural selection. This knowledge combined with epidemiological and clinical surveillance offers an alternative approach to studies of experimental animals with more convenient generation times and manipulable environments. Thus, the classical example of heterozygote advantage operating in natural populations remains the resistance of human heterozygotes for haemoglobin S to severe Plasmodium falciparum malaria (Allison, 1964). Other major infectious diseases of humans have been less studied, with the exception of investigations of the major histocompatibility complex where associations have been documented for diseases caused by several major pathogens (Todd, West & McDonald, 1990; Hill et al., 1991; Brahmajothi et al., 1991; Thursz et al., 1995).