51ÖØ¿ÚÁÔÆæ

Gene dispersal through pollen is affected by both abiotic and biotic environmental factors. In this study, I mainly investigated the effect of environmental factors: nutrient availability, simulated herbivory and floral morphology (flower colour), on the siring success of competing pollen donors and fitness of the resulting offspring. In a greenhouse experiments, maternal plants were treated with the above environmental factors and pollinated with two pollen donors which, are of different genotype and pollen vigour (measured in terms of pollen tube growth rate in vitro).



In both the nutrient manipulation and simulated herbivory treatments, the genotype of the donor and pollen tube growth rate had an effect on the outcome of pollen competition. In Cucumis sativus, when nutrient availability was low, females preferred donors with slow pollen tube growth rate. When nutrient availability was high, however, pollen tube growth rate was less important. Analysis of nutrient content of the pistil (as boron/calcium ratio) also showed that when plants had lower nutrient availability, the boron/ calcium ratio in the pistil decreased compared to pistils from plants in the nutrient rich soils. Both results suggest that the nutrient availability in the soil could affect the rate at which pollen tubes grow in the pistil and thus the female plant could be selected to adjust preference according to environmental circumstances. Simulated herbivory on the maternal plants of Nicotiana alata had no overall effect on the proportion of seeds produced by the two donors (red- and white-flowered morphs). However, Pollen donors that produce pollen with high growth rates in vitro had a higher siring ability, especially when the maternal plants were subject to simulated herbivory. The outcross pollen (red-flowered) produced higher proportion of seeds however, it produced small-sized offspring.



In plants therefore there are ranges of effects that induce variation in gene dispersal and there is no single expected outcome. This web of influences is too complex to sort out by experimental design. Risk assessments of gene flow from cultivated or genetically modified (GM) crops to their close relatives must thus rely on a qualified estimate based on knowledge about the factors influencing the gene flow. Studying the influence of female preference and male siring ability on gene flow, which unfortunately is given little attention, can augment this. Studies on the mechanisms that affect gene dispersal under different environmental conditions could thus contribute to the accuracy of predictions of ecological effects of gene flow from GM crops to their wild/weedy relatives.