OPTIMAL RESOURCE-ALLOCATION IN A RANDOMLY VARYING ENVIRONMENT

1. We construct a simple strategic population model to investigate opt imal allocation of resources (in excess of those required for maintena nce) to growth and/or reproduction. 2. Analysis of the model for a con stant environment demonstrates that determinate growth (where growth c eases at reproductive maturity) is always the optimal strategy. 3. We conduct numerical competition experiments to investigate optimal alloc ation in randomly varying environments, under three different noise mo dels. 4. indeterminate growth (simultaneous growth and reproduction ov er some of the individual's lifetime) is optimal in varying environmen ts where the variability is intense and on a time-scale comparable wit h that of an individual's lifetime. 5. The long-run growth rate and th e correlation between phenotype biomass and environment are maximized by sucessful competitors in the numerical contests. The presence of a competitor is shown to be an essential component defining the 'environ ment'. Optimization of various fitness measures in a single phenotype situation does not reveal the optimum for the competitive situation.