Variability in the dynamics of northern peripheral versus southern populations of two clonal plant species, Helianthus divaricatus and Rhus aromatica

Abstract

1. The clonal perennial herb Helianthus divaricatus and the clonal shrub Rhus aromatica reach the northern limit of their distribution in southern Quebec (Canada), where both are rare, but they are more common southwards in the adjacent province of Ontario. We tested whether peripheral populations of these species maintain themselves by having highly variable individual growth and survival rates. Their annual rates of population growth, λ, would then vary more, from year to year, than those of more southern populations in neighbouring Ontario, where the species are common. 2. Demographic analyses compared, over two transition intervals, size-based transition matrices for two northern (Gatineau Park, southern Quebec) and two southern populations (Pinery Park, southern Ontario) of each species. Comparisons used log-linear analyses, bias-corrected λ-values with 95% confidence intervals, and elasticity values. 3. For both species, the northern populations showed a larger variation between transition intervals and local populations in some of their vital rates and in λ-values. For H. divaricatus, elasticity values combined with log-linear analyses indicated that differences in the growth and survival of medium-sized ramets contributed most to the differences in λ-values between northern and southern populations. For R. aromatica, they indicated that differences in the fate of both medium and large ramets contributed to variation in λ. 4. Most of our data indicated a greater demographic variability in the northern populations of H. divaricatus and R. aromatica. At the northern periphery, these species grow in very few of the numerous habitat patches that appear to be suitable for them, whereas the two species together present a near continuous cover in all suitable habitats further south. High demographic variability therefore appears to contribute to determining the distribution of plant species, by increasing the extinction probabilities of local populations at the distribution limit.