Modelling the proximate basis of salmonid life-history variation, with application to Atlantic salmon,

The great diversity of life-history patterns in the salmonids has stimulated many theoretical studies. However, virtually all studies are based on ultimate considerations, in which predictions are made by comparing the expected reproductive success of dierent developmental or life-history pathways and choosing the one (or ones) with the highest fitness. Such models are post hoc because they attribute fitness to individuals at the completion of the particular phase of the life cycle and do not attempt to characterize the mechanisms that animals use to achieve the life-history pattern. We describe a model, based on proximate considerations, for salmonid life histories, focused on Atlantic salmon Salmo salar L. The model involves identification of the times at which developmental conversions are initiated or inhibited and the connection between physiological states and the thresholds for such conversions. Developmental paths are based on the comparison of the current physiological status of the fish (and its change of state) with a genetic threshold. The state of the fish and rate of change of state are determined by environmental opportunity, but the threshold is genetic. This approach therefore immediately generates a genotype–environment interaction. We use expected reproductive success to determine the fitness of individuals with dierent genetically determined thresholds. Instead of finding an optimal life history, our theory generates fitness surfaces for dierent life histories, so that variation is inherent in this approach. We describe and explain the structure of the model and present evidence on which this structure is based, thus providing a framework within which one can understand how ecology relates to the physiological mechanisms leading to the developmental changes of smolt metamorphosis and maturation