Fig 1.
(a) The layout of the 51 woodland patches used for the field experiment (n = 51, green) and other surrounding patches (n = 81, unfilled). Map of study site was created by authors using QGIS v3.22 [47]. Basemap was acquired from ‘GADM’ open-access data (https://gadm.org/). (b) Bluebells were pollinated by bumblebees (Bombus spp.). (c) Seed capsules of bluebell, H. non-scripta.
Fig 2.
The species-area relationship of a) total number of plant species and b) number of spring flowering species. A linear model was shown in a solid line, with 95% confidence intervals (CIs). Numbers of species (points) were log10 transformed.
Fig 3.
The effect of bagged flowers (sample size n = 6 pots, with three bluebell plants in each) versus open-pollination (sample size = 130 pots, i.e. the plants placed in the woodland patches) on the number of seeds and seed capsules and the number of flowers that failed to set any seeds.
Numbers were log10 transformed. Significant levels: 0 ‘***’, 0.001 ‘**’, 0.01 ‘*’, 0.05.
Fig 4.
The effect of patch proximity (i.e., the opposite of isolation) on the number of seeds (a) and seed capsules (b) of bluebells. Dots represent data points and lines are model outputs, with 95% CIs. Significant levels: 0 ‘***’, 0.001 ‘**’, 0.01 ‘*’, 0.05.
Table 1.
Summary of fitted GLMs on the effect of patch area, proximity, wild bluebells, and the interaction of patch area and proximity on the number, size, quality of seeds, and number of seed capsules and undeveloped flowers.
Significant p values were highlighted in bold. Significant levels: 0 ‘***’, 0.001 ‘**’, 0.01 ‘*’, 0.05.