Dylan Childs

A central concept in life history theory is that organisms are fundamentally constrained by trade-offs among fitness components. Demographic data provide a valuable means to explore these relationships—reanalyses of published matrix population models (MPMs) have identified consistent life history variation that appears to fall onto to a fast-slow continuum and a reproductive strategy axis. These relationships are assumed to reflect fundamental trade-offs. However, demographic constraints may also impact on patterns of covariance between life history metrics in natural populations. We simulated sets of size-structured integral projection models (IPMs) under the simple constraint that the long-term growth rate is close to one, which is to be expected when populations are subject to density dependence.

The IPMs were then discretised to mimic the stage structures found in published MPMs. We find that a PCA of life history metrics derived from the simulated models produced patterns of life history trait covariance that are very similar to patterns seen in empirical data. Our results indicate that comparative analysis of MPM-derived life history traits may not identify meaningful life history trade-offs and that a significant component of the covariance among MPM-derived life history matrics is consistent with non-adaptive constraints arising from density dependence.