The adaptive surface in ecology

We review the adaptive surface as it applies to the kinds of continuously varying traits studied in ecology and behaviour. The adaptive surface is a plot of the mean fitness of a population against the mean phenotype, and can be used to predict directions of evolution. Under frequency-independent selection the mean phenotype of a population evolves to an equilibrium position near a peak in the adaptive surface. However, under other forms of selection, peaks in the adaptive surface often do not coincide with evolutionary equilibria. Even when selection does drive a population toward peaks, the peaks may bear little relationship to measures of population "fitness", such as population size. Nevertheless the concept of an adaptive surface is heuristically useful because shifts between alternative locally stable states are likely to be a common mechanism of evolution. Furthermore, modelling peak shifts using the frequency-independent case indicates how easily different kinds of evolutionary transitions can occur. Theoretical models of peak shifts indicate that they most easily occur in response to the fluctuating selection pressures associated with environmental change, and are very unlikely to occur in response to genetic drift. Other mechanisms of peak shifts include environmental induction and mutations of large effect. The measurement of adaptive surfaces in nature is becoming increasingly common, although typically the individual fitness function rather than the true adaptive surface is measured, and frequency dependence is ignored. We suggest that more attention needs to be paid to the distinction between the individual fitness function, the adaptive surface, and the positions of evolutionary equilibria. If this is done, the value of the general concept will increase, and mechanisms of evolutionary transitions become further clarified.