The effects of contrasting environments on the reproductive ecology and genetics of Campanula rotundifolia L. (Campanulaceae)

The purpose of this work was to test the hypothesis that the alpine environment imposes strict limitations on plant reproduction due to low levels of insect diversity and activity, and that outcrossing in alpine plant populations is therefore unreliable. If this is true, alpine populations, relative to lower elevation populations, are expected to exhibit several genetic and reproductive characteristics including, (1) self-compatibility and/or autogamy, (2) decreased inbreeding depression, and (3) limited within population genetic variability and increased among population differentiation. My research focused on Campanula rotundifolia L., a widespread herbaceous perennial that occupies a broad elevational range in the Rocky Mountains, and addressed the degree to which alpine populations exhibit these characteristics in comparison to conspecific populations in the foothills. The pollination ecology of these populations was also investigated. In Chapter Two the population genetics of high and low elevation populations along the Colorado Front Range is described. Enzyme electrophoresis was used to assess levels of genetic variability and population structure. Measures of genetic variability did not differ between high and low elevation populations and were consistent with Hardy-Weinberg expectations. Low F{dollar}rmsb{lcub}st{rcub}{dollar} values indicated low levels of differentiation among all populations, whereas Rogers' index of genetic similarity suggested that alpine populations may be more differentiated from each other than were foothill populations. These results suggest either high levels of gene flow among populations or recent establishment from a common ancestral population. In Chapters Three and Four I discuss the mating systems and magnitude of inbreeding depression in alpine and foothill populations. Controlled pollinations indicated that plants in all populations were self-compatible, and I found no pattern of self-compatibility or pollinator limitation corresponding to elevation. However, in contrast to expectations, autogamous fruit production was highest in the foothill populations. In all low elevation populations, selfed seedlings had significantly lower dry weight, petiole length and leaf number than outcrossed seedlings. In contrast, there were no differences between selfed and outcrossed seedlings from the alpine populations for any of these fitness variables. Geographic variation in the magnitude of inbreeding depression, such as described here, has not been widely studied and suggests that, in contrast to many models, the ecology of populations may play a larger role in mating system evolution than inbreeding depression. Pollinator visitation rates were significantly lower, as was pollinator diversity, in the alpine than in the foothill populations as described in Chapter Five. However, "corrected visitation" rates that included information on pollinator diversity and abundance, pollen deposition, and the duration of stigma receptivity indicated that pollen transfer in foothill populations was nearly an order of magnitude less than estimates based on visitation rates alone. Therefore, the probability of fertilization was similar in alpine and foothill populations, and, more importantly, these data show that pollination in the foothills might be significantly overestimated by visitation rates alone. Highly efficient pollination (probably in conjunction with multiple aspects of reproductive ecology) maintained outcrossing in alpine populations of C. rotundifolia, which therefore did not exhibit the reproductive and genetic characteristics which are often associated with alpine plants.