Effects of range position, inter-annual variation and density on demographic transition rates of Hornungia petraea populations.

The central-marginal model assumes unfavourable and more variable environmental conditions at the periphery of a species' distribution range to negatively affect demographic transition rates, finally resulting in reduced population sizes and densities. Previous studies on density-dependence as a crucial factor regulating plant population growth have mainly focussed on fecundity and survival. Our objective is to analyse density-dependence in combination with the effect of inter-annual variation and range position on all life stages of an annual plant species, Hornungia petraea, including germination and seed incorporation into the seed bank. As previous studies on H. petraea had revealed a pattern opposite to existing theory with lower population densities at the distribution centre in Italy than at the periphery in Germany, we hypothesised that (1) demographic transition rates are lower, (2) the inter-annual variation in demographic transition rates is higher and (3) the intensity of density-dependence is weaker in Italy than in Germany. To analyse demographic transition rates, we used an autoregressive covariance strategy for repeated measures including density and inter-annual variation. All the three hypotheses were confirmed, but the impact of range position, density-dependence and inter-annual variation differed among the transition steps. All transition rates except fecundity were higher in the German populations than in the Italian populations. Germination rate and incorporation rate into the seed bank were strongly density-dependent. Central populations showed a larger inter-annual variation in fecundity and winter survival rate. Winter survival rate was the only transition step with a stronger density-dependence in peripheral populations. In most cases, these differences between distribution centre and periphery would not have emerged without taking density-dependence and inter-annual variation into account. We conclude that including range position, inter-annual variation and density-dependence in one single statistical model is an important tool for the interpretation of demographic patterns regarding the central-marginal model.