Sarah M. Swope and William H. Satterthwaite. 2012. Variable effects of a generalist parasitoid on a biocontrol seed predator and its target weed. Ecological Applications 22:20–34.

Appendix A. Quantifying Centaurea solstitialis recruitment from the seed bank.

Our goal was to describe how much the seed bank contributed to recruitment of C. solstitialis in each of the two study sites described in the text. To do this we quantified seed rain in heavily invaded areas and allowed the seedlings to emerge the next growing season but removed their flowering stems to eliminate seed rain the following year. We assume that any seedlings that recruited the following year must have come from the seed bank.

In the late summer of 2007, we established six plots (0.5 × 0.5m) in each of the four invasions used in this study (two invasions at the coastal site and two invasions at the Central Valley site). Plots had high C. solstitialis density and were located 4–5m upslope of the plots used in the demography study described in detail in the manuscript. We estimated seed rain into each of the plots by counting the total number of inflorescences in the plot and multiplying by the mean number of viable seeds per inflorescence for that population in that year. (See manuscript for methods for estimating viable seeds per inflorescence.) Seedlings emerge throughout the fall and winter (2007–08) and produce a single flowering stem in the early summer. We cut all flowering stems to ground level in our plots prior to bud and inflorescence production (early summer 2008) to prevent that year’s cohort from producing any seed. Although C. solstitialis typically produces only one flowering stem per plant, it is capable of producing additional ones in response to the loss of the main stem so we cut any new flowering stems that grew in throughout that summer. Additionally, to minimize seed rain from plants adjacent to our plots, we also cut all flowering stems from plants within a 0.5m wide buffer zone around each plot. We therefore assume that any seedlings that recruited in our plots the following fall and winter (2008–09) came from the seed bank because removing the flowering stems eliminated current seed rain and removing neighbors minimized immigration.

Seedlings recruited from the seed bank in all four of our populations (but not every plot) at very low numbers. At two coastal populations, a total of nine seedlings emerged in the 12 plots from which we eliminated seed rain; at the two Central Valley populations, a total of seven seedlings emerged in the 12 plots from the seed bank (Table A1). By way of comparison, we have also provided data on the number of seedlings that emerged in the unmanipulated demography plots at the same locations at that same time (fall and winter 2008–09). We randomly selected six demography plots for this comparison. Seedlings in these plots were from the current year’s seed rain as well as the seed bank; the large differences in recruitment between the experimental plots and the demography plots indicate that recruitment is driven largely by seed rain and that the seed bank contributes relatively little to it.

TABLE A1. Results from seed bank recruitment experiment. The number of seeds and seedlings is the sum (range) for all plots, and the percentage germination. Plots in the two far right columns were randomly selected from the demography study (described in detail in the manuscript) and seedlings recruited in these plots come from both current seed rain (late summer 2008) and the seed bank.