Figure 2. Vital and population growth rate relationships with PC1 for each focal species. Low values of PC1 represent closed canopy, high litter, high soil fertility plots. PC1 is standardised to a mean of 0 and standard deviation of 1. Black lines and light grey fields represent linear regressions from models with 95% confidence intervals holding continuous explanatory variables at their means, under the ambient watering treatment. Points for emergence and survival represent emergence fractions and survival of individuals. To facilitate plotting, we have added one to seed production values and show them on a log scale. Significant interactions between PC1 and neighbour abundance or presence (for population growth) are plotted with points and regressions representing the presence (red) or absence (blue) of neighbours, holding abundance at their respective means for survival and seed production. Asterisks indicate significance of the main effect of PC1 where models did not include interactions of other factors with PC1: * p < 0.05.
Soil pH and the watering treatment had few main effects on species’ vital and population growth rates. Soil pH (PC2) had a negative or convex relationship with emergence for 3/9 species, was positively related to survival for 1/8 species, negatively related to seed production for 1/8 species, and positively related to population growth rate for 1/8 species (Table S2, p <0.05). Interestingly, survival responses to PC1 depended on the watering treatment for half of the species but without a clear pattern in directions of responses among species (Table S2 and Figure S4). This interaction was more prevalent for survival than seed production (1/8 species) or population growth rate (2/8 species, Table S2 and Figure S4). There was consistently no evidence that the watering treatment influenced species’ responses to neighbour abundance or presence, except for T. cyanopetala which had lower population growth rate in the dry compared to ambient watering treatment but only where neighbours were present, not absent (Table S2).
Survival, seed production, and population growth rate responses to biotic factors were similar among species. We found no clear main effects of neighbour abundance on survival or seed production for any species, except for P. debilis where seed production decreased as the number of neighbours increased (Fig. 3, Table S2). Mean population growth rates were lower in the presence than absence of neighbours for all species, although the main effect of neighbours on population growth rate (where there were no interactions with other factors) was only significant for 2/8 species, including P. debilis (Fig. 3, Table S2, p <0.05). The presence of the invasive parasiticC. campestris, did not affect survival, and only influenced seed production, negatively, for 1/8 species (Table S2).