Fitness in Random Environments

Abstract

The effect of random environmental changes on fitness is modelled for a wide range of life histories and patterns of environmental and demographic variability with the following main conclusions: 1. Independent individual variation in fitness does not influence selection on an allele or a trait in a large mixed population. 2. Mean log fitness is the measure of long term fitness when fitness varies between nonoverlapping generations. In this case there is no coexistence between alternative strategies. 3. With long-lived overlapping generations, a mean ratio greater than 1 between the fitness of a rare type and the fitness of the common type is the criterion for the increase of the rare type. This gives a selective advantage to the rare type, which allows coexistence between strategies that have some independently varying fitness. This also applies to a long lived seed bank in the soil. 4. In very small populations, especially with single founders, the long term fitness, which is given by the probability of establishment, is determined by the probability of extinction, which increases when the variance of individual fitness increases for the same mean fitness. Mutants with higher mean fitness have a higher long term fitness however, and eventually displace low variance mutants established earlier. A long term selective advantage for low variance mutants is maintained in continuously changing environment, or in structured populations whose subpopulations periodically go extinct. 5. The variation of different components of fitness does not influence fitness if it occurs independently in different individuals in large populations. However, it does influence fitness if it is synchronised to some extent in the whole population, for example in synchronised age dependent cohorts in nonoverlapping generations. 6. The optimal tradeoff between reproduction and survival is expected to take into account only the mean survival if survival is distributed independently in a large population, e.g. risk of predation, while the variance of survival has to be taken into account if mortality is synchronised, or when there is a single or few founders. 7. The variation of resources which contribute to individual fitness, e.g. food, decreases fitness over the convex range of the fitness function and increases fitness over the concave range. 8. Behavioural, developmental, or life history strategies which reduce the variance between generations are selected for: e.g. dormancy, long living perennial habit in plants, dispersal, and phenotypic polymorphism.