A central assumption of life history theory is that the evolution of the component traits is determined in part by trade-offs between these traits. Whereas the existence of such trade-offs has been well demonstrated, the relative importance of these remains unclear. In this paper we use optimality theory to test the hypothesis that the trade-off between present and future fecundity induced by the costs of continued growth is a sufficient explanation for the optimal age at first reproduction, alpha, and the optimal allocation to reproduction, G, in 38 populations of perch and Arctic char. This hypothesis is rejected for both traits and we conclude that this trade-off, by itself, is an insufficient explanation for the observed values of alpha and G. Similarly, a fitness function that assumes a mortality cost to reproduction but no growth cost cannot account for the observed values of alpha. In contrast, under the assumption that fitness is maximized, the observed life histories can be accounted for by the joint action of trade-offs between growth and reproductive allocation and between mortality and reproductive allocation (Individual Juvenile Mortality model). Although the ability of the growth/mortality model to fit the data does not prove that this is the mechanism driving the evolution of the optimal age at first reproduction and allocation to reproduction, the fit does demonstrate that the hypothesis is consistent with the data and hence cannot at this time be rejected. We also examine two simpler versions of this model, one in which adult mortality is a constant proportion of juvenile mortality [Proportional Juvenile Mortality (PJM) model] and one in which the proportionality is constant within but not necessarily between species [Specific Juvenile Mortality (SSJM) model]. We find that the PJM model is unacceptable but that the SSJM model produces fits suggesting that, within the two species studied, juvenile mortality is proportional to adult mortality but the value differs between the two species.