Ecological context of breeding system variation: sex, size and pollination in a (predominantly) gynodioecious shrub

Abstract

Background and aims: Species that exhibit among-population variation in breeding system are particularly suitable to study the importance of the ecological context for the stability and evolution of gender polymorphism. Geographical variation in breeding system and sex ratio of Daphne laureola (Thymelaeaceae) was examined and their association with environmental conditions, plant and floral display sizes, and pollination environment in a broad geographic scale was analysed.

Methods: The proportion of female and hermaphrodite individuals in 38 populations within the Iberian Peninsula was scored. Average local temperature and precipitation from these sites were obtained from interpolation models based on 30 years of data. Pollination success was estimated as stigmatic pollen loads, pollen tubes per ovule and the proportion of unfertilized flowers per individual in a sub-set of hermaphroditic and gynodioecious populations.

Key results: Daphne laureola is predominantly gynodioecious, but hermaphroditic populations were found in northeastern and southwestern regions, characterized by higher temperatures and lower annual precipitation. In the gynodioecious populations, female plants were larger and bore more flowers than hermaphrodites. However, due to their lower pollination success, females did not consistently produce more seeds than hermaphrodites, which tends to negate a seed production advantage in D. laureola females. In the northeastern hermaphroditic populations, plants were smaller and produced 9-13 times fewer flowers than in the other Iberian regions, and thus presumably had a lower level of geitonogamous self-fertilization. However, in a few southern populations hermaphroditism was not associated with small plant size and low flower production.

Conclusions: The findings highlight that different mechanisms, including abiotic conditions and pollinator service, may account for breeding system variation within a species' distribution range and also suggest that geitonogamy may affect plant breeding system evolution.

Fig. 1.

The locations of the 38 Daphne laureola populations surveyed for sex ratio at the Iberian Peninsula. Population #32 (Roblehondo) is the only one represented here that is located within the Natural Park of Sierras de Cazorla, Segura y Las Villas, where previous studies of altitudinal variation in sex ratio referred to in the text were conducted. White areas of the pie chart graphs indicate the percentage of females in every study population. Fully hermaphroditic populations were not found in the NW region. For reference, the grey shading shows the species distribution range in the Iberian Peninsula.

Fig. 2.

Relationship between the mean annual local temperature and sex ratio of the studied Iberian Daphne laureola populations. Northern and southern locations are as indicated.

Fig. 3.

Regional differences in average plant size (A) and total flower production (B) of Daphne laureola based on data of seven NE populations and three NW populations, and 40 individuals per site. Bars represent the least-square means (± s.e.) after accounting for sex (fixed effect) and population (random effect) variation.

Fig. 4.

Differences between sexes in three measures of quantitative pollination success of Daphne laureola. (A) Mean stigmatic pollen load, (B) mean number of pollen tubes per ovule and (C) the percentage of flowers without pollen tubes. Bars represent the least-square means (± s.e.) after accounting for region (fixed effect) and population (random effect) variation, and are based on ten gynodioecious and eight hermaphrodite populations sampled in 2003. A priori contrasts were used to test specifically for differences between female and hermaphrodite plants in gynodioecious populations, and also between hermaphrodite plants growing in gynodioecious and hermaphroditic populations, ***P < 0·0001 and **P = 0·002.

Fig. 5.

Relationship between the mean annual local temperature and population pollination failure estimated as the average proportion of unfertilized flowers per plant. Female and hermaphrodite scores in gynodioecious populations are as indicated.