The genesis of size hierarchies in seedling populations of impatiens capensis meerb.

TLDR: In this article, the authors used path analysis to summarize the interactions among the continuous variables and to disentangle direct from indirect causal effects, concluding that seed weight enhanced performance primarily through its effects on emergence date and cotyledon area.

Abstract: Summary In 1978 and 1980 I planted 660 and 900 seeds of Impatiens capensis Meerb. at four densities in the greenhouse to examine how variation in plant performance is generated. I noted seed weight, seed type (cleistogamous or chasmogamous), maternal parent, and population of origin, and monitored emergence date, cotyledon area, the biomass of competitors within the same flat, and the final size reached by the seedlings after 63 to 80 d of growth. Larger seeds tended to germinate sooner, as did seeds derived from chasmogamous flowers. Seeds from northern populations took longer to germinate. Seeds from different maternal parents also germinated at different rates. Cotyledon area strongly depended on seed weight, and, to a lesser extent, germination date. Maternal parent and seed type significantly affected cotyledon area in 1980. Plant density and cotyledon area influenced final size the most, but almost every factor proved to be statistically significant. As expected, earlier emerging seedlings with larger cotyledons growing at the lowest density grew into the largest plants. I applied path analysis to summarize the interactions among the continuous variables and to disentangle direct from indirect causal effects. This technique revealed that seed weight enhanced performance primarily through its effects on emergence date and cotyledon area. Chasmogamous seedlings outperformed cleistogamous seedlings, and their advantage was expressed during most phases of growth. Together, the predictor variables accounted for over four-fifths of the total variation in final size. The inequality of plant sizes, as measured by the Gini coefficient, increased appreciably during the course of the experiment. There was no simple relation to density, however, suggesting that both intrinsic differences in growth rate, and competitive dominance and supression fuel the establishment of size hierarchies.