Species size and light penetration
Light penetration was significantly lower in plots with greater plot-level mean height (R2 = 0.08, P = 0.025; Fig. 5) but was unassociated with large species abundance. Mean intraplot variation in light penetration was significantly positively correlated with mean light penetration (Spearman correlation, rho= 0.49, P < 0.001; Fig. S2).
Small species abundance and richness in plots were significantly higher in plots with high light penetration, but not for both definitions of small species. Small species under the ‘1st quartile’ definition had significantly higher flowering abundance in plots with higher light penetration (R2 = 0.08, P = 0.027; Fig. 6A). Median flowering abundance and richness were significantly higher in plots with increased light penetration (Abundance: R2=0.1, P = 0.012; Richness: P = 0.014, R2 = 0.1; Figs. 6B,D). The richness of small species (1st quartile) did not significantly change with light penetration (Fig. 6C, Table S1). Within-plot variation in light penetration was not significantly related to small species abundance or richness using either definition (Table S1).
The composition of small plant species defined using the median significantly changed with mean plot-level light penetration (dbRDA1,P = 0.031, pseudo-F 1,47 = 2.06; Fig. 7), but was unaffected by mean intraplot variation. Mean plot-level light penetration explained 4.2% of the total variation in plot composition for ‘median’ small species. Eighteen species had higher abundances associated with higher light penetration (as demonstrated by positive loadings along the light penetration axis), while eight species tended to have lower abundances associated with higher light penetration (as demonstrated by negative loadings along the light penetration axis; Fig. 7). Conversely, ‘1st quartile’ small species composition was unaffected by mean light penetration or mean intraplot variation, and neither were retained as axes.
We performed analyses using both total abundance and flowering abundance; however, because the analyses of flowering abundance were more informative, we have focused on these. Half of the analyses using the total number of individuals demonstrated significant relationships between light penetration and small species abundance and richness, compared to two-thirds of analyses using only the number of flowering individuals. We provide further information on analyses using the total number of individuals in the supplementary material (Fig. S3-S7).