Is there silicon in flowers and what does it tell us?

Abstract

The emergence of flowers marked an important development in plant evolution. Flowers in many species evolved to attract animal pollinators to increase fertilisation chances. In leaves, silicon (Si) discourages herbivores, for example by wearing down mouthparts. Flowers are essentially modified leaves and hence may also have the capacity to accumulate Si. If Si in flowers discourages animal visitors as it does in leaves, Si accumulation may be disadvantageous for pollination. Whether flowers accumulate Si, and what the implications may be, was not known for many species. We analysed leaves and flowers of different taxa, separated into their different anatomical parts. Flowers mostly have low Si concentrations in all parts (mean ± SE of BSi in mg g^-1 was 0.22 ± 0.04 in petals, 0.59 ± 0.24 in sepals, 0.14 ± 0.03 in stamens, 0.15 ± 0.04 in styles and stigmas and 0.37 ± 0.19 in ovaries for a subset of 56 species). In most cases, less Si was accumulated in flowers than in leaves (mean ± SE of BSi in mg g^-1 was 1.51 ± 0.55 in whole flowers vs. 2.97 ± 0.57 in leaves in 104 species) though intriguing exceptions are found, with some species accumulating more Si in flowers than leaves. The large variation in concentration among flowers across the taxa examined, with a particularly high concentration in grass inflorescences, tantalisingly suggests differences in the use of Si for flowers across plant groups. We conclude that the study of the functions of Si for flowers warrants more attention, with pollination strategy a potential contributing factor.

Keywords: angiosperms; biogenic silica; evolution; pollination.

FIGURE 1

(a) [BSi] for different flower parts and leaves for 56 species. (b) [BSi] for whole flowers and leaves for 104 species. Note log scale of y‐axes. The values for flower [BSi] are overall higher in (b) because the larger dataset included more high Si accumulating species, and typically petals and sepals, which had (non‐significantly) higher [BSi], contributed a larger proportion of whole flower biomass than other plant parts shown in (a).

FIGURE 2

Comparisons between leaf and whole flower [BSi] by taxonomic group. The axes are log₁₀ scale, with the regression line compared with a dotted 1:1 line.

FIGURE 3

(a) Comparisons between whole flower (white) and leaf (grey) [BSi] by order, with plant groups indicated also (C = Chloranthales). The plant orders are arranged such that they align from L‐R with the phylogenetic tree in Figure S1 from top to bottom. (b) Comparisons between whole flower (white) and leaf (grey) [BSi] by plant group, with mean flower: leaf BSi ratios given. (c) Phylogenetically independent contrasts between leaf and whole flower [BSi] for 102 contrasts calculated at dichotomies across 104 species for logged data. The slope was forced through the origin as the sign value for each pair is arbitrary.