Differential selection on floral traits of Ipomopsis aggregata growing in contrasting environments

Abstract

Interactions for pollination between co-flowering plant species have been hypothesized to shape the evolution of their floral traits, but this hypothesis has rarely been tested. I tested the prediction that the presence of a co-flowering plant species influences the strength and/or direction of pollinator-mediated selection on floral traits. I measured phenotypic selection via female fitness on four floral traits of Ipomopsis aggregata in five populations. Three contained only conspecifics (I only) and two also contained the co-flowering species Penstemon barbatus (P + I). Directional selection via fruits/plant on corolla length and width differed in both strength and direction between P + I and I only populations. On average, selection on corolla length and width (1) was stronger in P + I than I only populations and (2) was consistently negative in P + I populations, but consistently positive in I only populations. However, these differences in selection on I. aggregata between P + I and I only populations were not caused by interactions for pollination with P. barbatus. Although plants in P + I populations received approximately 31% less conspecific pollen/flower than plants in I only populations, this difference in pollination did not translate into differences in reproductive success, which indicates that P. barbatus and I. aggregata do not strongly compete for pollination. In addition, selection via fruits/plant and conspecific pollen deposited/flower was not congruent. For example, selection on corolla length via pollen/flower was uniformly positive and did not differ between P + I and I only populations. These data suggest that the presence of P. barbatus does influence selection on floral traits of I. aggregata, but not by influencing pollination. Instead, differences in selection between P + I and I only populations appear to be the result of post-pollination modification of selection by a factor correlated with the presence of P. barbatus.