A Californian population of the Yellow Monkey Flower, Mimulus guttatus Fischer ex D.C., contains a floral fragrance polymorphism such that the plants can be divided into two classes: fragrant and non-fragrant. This thesis has examined the effects of the fragrance polymorphism on the natural pollinators of M. guttatus, with a view to investigating the role and evolutionary significance of floral fragrance in apidophilous plants. Published work by many researchers has suggested that a fragrance polymorphism in an apidophilous chasmogamous plant would lead to reduced inter-phenotype pollinator movements due to the phenomenon of odour-constancy, and hence to genetic divergence and possibly sympatric speciation. Alternatively, if the fragrance phenotypes differed in pollinator attractiveness, the more attractive phenotype would have a selective advantage, and would therefore eventually become fixed. Both of these possible scenarios have been explored in the course of this project, both by investigation of bee foraging behaviour on experimental plots containing fragrant and non-fragrant Mimulus plants, and by investigation of the fitnesses of the two phenotypes in the natural population and under glasshouse conditions. It was found that the two phenotypes did not differ significantly in terms of measured fitness, and that in the natural population there was no evidence of reduced gene-flow between fragrant and non-fragrant plants. Experimental plot results showed no evidence of odour-constancy by bees, but revealed that some bumblebee species discriminated weakly between the two phenotypes. Pollinators were found to be visiting Mimulus flowers for pollen, and variation in plant pollen production had significant effects on within-plant foraging behaviour. However, bees failed to learn to associate fragrance phenotype with pollen production, in contrast to previously published research that has demonstrated that bees readily learn to associate fragrance phenotype with nectar rewards, suggesting that pollen-collecting bees may respond differently to floral cues than nectar-collecting bees. For a variable such as floral fragrance to have evolutionary significance it is essential that the character is heritable, and part of the investigation has concentrated on investigating the heritability and penetrance of the character. It was found that inheritance of fragrance is best explained by a simple additive genetic model, and that fragrance production is dominant to absence of fragrance. The possibility that this fragrance polymorphism has been previously influenced by other pollen vectors and the population's potential for switching from bee-pollination to alternative pollen vectors is discussed.