Asymmetrical disassortative pollination mediated by long-/short-tongued pollinators in a distylous Limonium myrianthum (Plumbaginaceae) with a short corolla tubular small flower

Abstract

In heterostylous plants, short-tongued pollinators are often ineffective/inefficient owing to the limitations imposed by a long corolla tube. However, it is unclear how disassortative pollen transfer is achieved in small flowers. We investigated the pollination pattern and floral morph variation by analyzing heterostylous syndrome, pollinator groups, and pollen deposition after a single visitation in two Limonium myrianthum populations with short-corolla-tubular small flowers. The predominant pollinators in the Hutubi population were pollen-seeking short-tongued syrphids, which can only transfer pollen between high-level sexual organs. In the Xishan population, nectar-seeking short-tongued insects were efficient pollinators with symmetrical disassortative pollen transfer between high- and low-level sexual organs, whereas long-tongued pollinators had a low efficiency between high-level sexual organs due to the low contact probability with the stigma of long-styled flowers (L-morph), which no longer offered the same advantage observed in tubular flowers. Asymmetrical disassortative pollination may cause the female fitness of short-styled (S-morph) individuals in the Hutubi and L-morph individuals in the Xishan population to suffer greater selection pressure and exhibit a higher degree of floral morph variation. Limonium myrianthum exhibits an unusual pollination pattern in which the small flowers with short corolla tubes make it possible for short-tongued insects to become effective pollinators. However, factors such as the position of stigma-anther within the flower, pollinator species and their preference further caused asymmetrical disassortative pollen transfer. Therefore, more factors should be considered when evaluating the effectiveness of short- and long-tongued insects in pollination service.

Keywords: Limonium myrianthum; floral morph variation; heterostyly; long‐ and short‐tongued pollinators; pollination patterns; small flower with short corolla tube.

FIGURE 1

Two types of plants and floral morphs with different pollen–stigma morphologies in two populations. (a1–a3) The plants and floral morphs with “cob” stigma morphology (icon: ○); (b1–b3) the plants and floral morphs with “papillate” stigma morphology (icon: ). The deeper the icon color, the larger their stigma–anther separation.

FIGURE 2

Floral morphs, morphology of stigmas and pollen grains, and pollinators of Limonium myrianthum. (a) Long‐styled morph flower; (b) short‐styled morph flower; (c) the short‐tongued pollinator Syrphidae in Hutubi population; (d) the short‐tongued pollinator Muscidae in Xishan population; (e) the long‐tongued pollinator Bombylius sp. in Xishan population; (a1–a5) the long‐styled morph flowers and their variants, and their “cob” stigma and coarse reticulate pollen morphology; (b1–b5) the short‐styled morph flowers and their variants, and their “papillate” stigma and finely reticulate pollen morphology.

FIGURE 3

Visiting frequency of pollinators of Limonium myrianthum in two populations (bars indicate mean ± SE). (a) Hutubi; (b) Xishan. Bars with asterisks indicate a significant difference between pollinator groups (***p < .001) according to linear mixed‐effects models (LMMs).

FIGURE 4

Number of legitimate pollen grains on stigmas after a single visitation of long‐tongued/short‐tongued pollinators in two populations of Limonium myrianthum (bars indicate mean ± SE). (a) Hutubi; (b) Xishan. Bars with asterisks indicate a significant difference between morphs (***p < .001; **p < .01; ns, p > .05) according to generalized linear mixed‐effects models (GLMMs).

FIGURE 5

The fruit set under different pollination treatments (bars indicate mean ± SE). Same letters above the bars indicate that there is no statistically significant difference. Different letters indicate a statistically significant difference (p < .01) according to generalized linear mixed‐effects models (GLMMs).