Competition for pollination has served as a model for the integration of ecological and evolutionary perspectives in the study of species interactions. Its study has elucidated both obvious and more subtle mechanisms, and has documented a range of outcomes. However, the potential for this interaction to inform our understanding of both pure and applied aspects of pollination biology has only begun to be realized.

No abstract

Summary 1.Pollinators visiting large floral displays may probe several flowers in sequence, leading to geitonogamous (among-flower) self-pollination. To investigate the relationship between floral display size and patterns of pollinator movement, we studied foraging by several pollinator species in four replicate arrays of Mimulus ringens (Scrophulariaceae). In each array displays were trimmed to two, four, eight and 16 flowers per plant. 2. Bees preferred large displays, and probed more flowers in sequence on large than on small displays. However, the proportion of available flowers probed decreased with display, resulting in nearly equal floral visitation rates across treatments. 3. Because pollinators probed more flowers in sequence on large displays, plants with numerous flowers should experience more geitonogamous self-pollination than plants with small displays. 4. In all four treatments, pollinators frequently visited only one flower before leaving the plant. As the first flower probed on a plant cannot receive geitonogamous pollen, this potentially reduces selfing rates for those flowers, compared to flowers probed late in a long visitation sequence on a plant. Such differences among flowers in pollination history should increase variation in geitonogamous self-pollination among fruits within plants. 5. The three most abundant pollinator species differed significantly in behaviours that could influence plant mating patterns, including number of flowers probed per plant; interplant movement distances; and grooming. Variation in foraging patterns was also evident among individuals within species. These subtle differences in response should affect the pollination services provided to plants.

The influence of population density on pollinator movements and outcrossing rates was studied in experimental arrays of Mimulus ringens (square-stemmed monkeyflower), a wetland perennial species with a mixed-mating system. Each population was composed of genets with unique multilocus combinations of homozygous genotypes, facilitating determination of outcrossing rates through paternity exclusion. Replicate arrays were cloned from the same set of genets to minimize differences in floral and vegetative morphology among density treatments. Two arrays were planted at each of three spacings typical of the range of densities found in natural M. ringens populations. Both the proportion of pollinator flights between plants and the frequency of outcrossing were significantly greater at high density. These results suggest that ecological and demographic factors, such as population density, can significantly influence levels of inbreeding in species with mixedmating systems.Keywords: density, experimental arrays, geitonogamy, outcrossing rate, paternity analysis, pollinator movements. IntroductionIn several species of animal-pollinated flowering plants, outcrossing rates vary dramatically among populations (Harding & Barnes, 1977;Valdeyron et a!., 1977;Ellstrand et al., 1978;Schoen, 1982;Schemske & Lande, 1985; Ritland & Ganders, 1987a;Waller & Knight, 1989; Barrett & Husband, 1990;Dole, 1991;Godt & Hamrick, 1991). This variation is often attributed to ecological factors, such as population density, that influence pollinator movements (Aide, 1986;Waller, 1986;Brown et al., 1989;Watkins & Levin, 1990;Murawski & Hamrick, 1991;Lloyd & Schoen, 1992). Despite theoretical predictions that outcrossing rates will be positively correlated with population density (Bateman, 1956;Handel, 1983;Brown et a!. 1989), the relationship between these variables remains poorly understood.Most studies documenting intraspecific variation in plant mating systems have involved comparisons of natural populations. However, in nature, population density is often confounded with other demographic variables such as plant size, population shape and population size (Barrett & Eckert, 1990;Kunin, 1993;Karron, 1995). In addition, differences among populations in genetically determined floral traits influencing *Correspondence the mating system (Campbell & Abbott, 1976;Schoen, 1982; Marshall & Abbott, 1984; Wyatt, 1984a,b;Holtsford & Ellstrand, 1989 Barrett & Husband, 1990;Clegg & Epperson, 1988;Morgan & Barrett, 1990) may further limit the utility of descriptive studies correlating population density with outcrossing rate.The influence of density on plant mating systems is well suited to experimental study. Spacing between plants is readily manipulated in population arrays, and recent experiments have examined the effects of density on patterns of pollinator movement and seed set in self-incompatible species (Kunin, 1992(Kunin, , 1993. Artificial arrays have also been used to determine outcrossing rates of distinct floral morphs in polymorphic populations...

No abstract