Persistent pollinators and the evolution of complete selfing

Abstract

0.8) are common in self-compatible populations where inbreeding depression is below 0.5 (Fig. 2 , sections IV and V). Here, we turn the original question around to ask, "Why might outcrossing persist in populations despite relatively low inbreeding depression?" and consider constraints posed by pollinator mutualists. Understanding the evolution of selfing given high inbreeding depression is perhaps best explained considering outcross pollen limitation. Th e reproductive assurance hypothesis posits that unreliable pollinators or low mate availability leave plants no better fi t-ness option but to self, given the alternative of making fewer or even no seeds. A number of excellent models incorporate outcross pollen limitation (Goodwillie et al., 2005), and empirical studies support components of the reproductive assurance hypothesis (Busch and Delph, 2012). While this hypothesis is well accepted as an important driver of the evolution of selfi ng and the maintenance of mixed mating, it is one-sided in shedding light on mating system evolution because it does not explain moderate to high levels of outcrossing when inbreeding depression is low. We proff er that an explanation for persistent outcrossing given low inbreeding depression may lie in viewing the evolution of selfi ng through the lens of mutualism breakdown (Sachs and Simms, 2006) and the potential difficulties for self-compatible plants in divesting from their mutualist pollinators. Conflict could arise from the evolution of higher autonomous selfing rates because, while autonomously selfi ng plants are less reliant on their pollina-tor mutualists, there are ecological conditions where pollinators may remain strongly reliant on the plant for pollen and nectar. The resulting asymmetry would effect continued outcross pollen delivery , significantly slowing the progression of mutualism aban-donment by the plant and thus evolution to high selfi ng. Notably, the asymmetry should be greatest where pollinators are abundant. Yet, pollen limitation is often seen as ubiquitous. Despite this perception, conditions with abundant pollinators providing adequate outcross pollen may be common. Indeed, Knight et al. (2006) revealed clear publication bias for reporting pollen limitation and that, when measured over a plant's lifetime, the magnitude of pollen limitation on average is actually low, especially for monocarpic species, which constitute half of the species inhabiting the seemingly paradoxical state space of moderate to high outcrossing levels with low inbreeding (Fig. 2 , sections IV and V).