Zygomorphic flowers have fewer potential pollinator species

Abstract

Botanists have long identified bilaterally symmetrical (zygomorphic) flowers with more specialized pollination interactions than radially symmetrical (actinomorphic) flowers. Zygomorphic flowers facilitate more precise contact with pollinators, guide pollinator behaviour and exclude less effective pollinators. However, whether zygomorphic flowers are actually visited by a smaller subset of available pollinator species has not been broadly evaluated. We compiled 53 609 floral visitation records in 159 communities and classified the plants' floral symmetry. Globally and within individual communities, plants with zygomorphic flowers are indeed visited by fewer species. At the same time, zygomorphic flowers share a somewhat larger proportion of their visitor species with other co-occurring plants and have particularly high sharing with co-occurring plants that also have zygomorphic flowers. Visitation sub-networks for zygomorphic species also show differences that may arise from reduced visitor diversity, including greater connectance, greater web asymmetry and lower coextinction robustness of both plants and visitor species-but these changes do not necessarily translate to whole plant-visitor communities. These results provide context for widely documented associations between zygomorphy and diversification and imply that species with zygomorphic flowers may face a greater risk of extinction due to pollinator loss.

Keywords: ecological specialization; floral morphology; interaction networks; pollination.

Figure 1.

The global plant-visitation dataset. (a) Global distribution of the plant-visitor networks, with points scaled to indicate numbers of plant taxa. (b) Bar plots giving the number of plant taxa for each network, coloured according to floral symmetry ((i) full dataset, (ii), all networks except for the largest, to provide better visibility). (c) Visitor species per plant, grouped by floral symmetry. (d) Sharing of visitor species with all co-occurring plant species, and co-occurring species with each type of floral symmetry. (e) Mean visitor species per plant in sub-networks for plants with differing symmetries, with grey lines linking points for zygomorphic and actinomorphic taxa from the same network, plotted against latitude of the network location. In (c) and (d), asterisks indicate significant differences in one-tailed Wilcoxon tests: *p < 0.01, **p < 10⁻⁴, ***p < 10⁻⁵.

Figure 2.

Floral symmetry and plant-visitor network structure. (a) Network descriptive statistics for sub-networks based on floral symmetry. Grey lines link points representing zygomorphic and actinomorphic sub-networks from the same source network, horizontal black bars mark the median value within each floral symmetry type, and asterisks mark significant differences in one-tailed paired Wilcoxon tests. (b) Scatterplots of relationships between the proportion of zygomorphic flowers in a network and the set of network structure metrics in (a); linear regression lines (white with grey 95% CI) are given when the correlation is significant with p ≤ 0.05.