Variation in plant and pollinator abundance in space and, through variation in the timing of reproduction, also abundance in time affects genetic diversity of pollen loads. Genetic diversity through multiple-donor pollination is expected to enhance offspring diversity, number and quality, through post-pollination selection against inbreeding, genetic incompatibility or selfish genetic elements. Here I propose a multi-level approach to investigate post-pollination selection against inbreeding in S. latifolia. Small populations are frequently genetically less diverse than large, core populations, resulting in higher homozygosity that can lower individual performance (inbreeding depression) and hamper population persistence. Higher likelihood of pollination among neighbouring individuals will increase the risk of inbreeding, if neighbouring individuals tend to be related. Such an increased risk of inbreeding in natural plant populations thus results from limited seed and pollen dispersal. As known for many species, including humans, inbreeding often results in fewer or less performant progeny and despite being favoured under some conditions, usually generates net fitness costs. Because of the mostly detrimental effects of inbreeding on fitness, mechanisms should arise to avoid fertilizations by related pollen (or, alternatively, to select against inbred offspring), as has been found in some animal species, where unrelated males can be preferred as mates or at fertilization. In plants the opportunity for post-pollination selection against inbreeding may be facilitated by gene expression in pollen. As known for Arabidopsis and other species, pollen tubes express many genes. This may allow recognition between related alleles during pollen-pistil interactions. Several examples suggest that genetically–based pollen-pistil interactions and thus possibly post-pollination, pre-zygotic selection may occur in plants, including female gametophyte control over pollen tube reception in the embryo sac, selection against self pollen, and more generally genetic self-incompatibility systems. Thus, when several pollen tubes from different donors grow through the stigma to reach the ovary, there may be scope for selection against pollen stemming from donor plants closely related to the recipient plant, or after fertilization against inbred embryos. Negative effects of inbreeding on individual performance have been documented in several plant species. However, often inbreeding is studied under controlled laboratory conditions, while the costs of inbreeding may be higher under natural environmental conditions. For a realistic assessment of the conditions that should favour post-pollination selection, it is therefore important to evaluate the impact of inbreeding depression on plant fitness under natural conditions, in particular in the presence of natural enemies, and throughout the life cycle (i.e. including for instance also the analysis of inbreeding depression for male fertility under natural pollination). Finally, it is essential to relate costs and benefits of inbreeding avoidance to the actual risk of inbreeding in natural populations and to validate experimental results to the pollen load diversity that characterizes natural populations. Thus, ideally a set of interrelated questions should be addressed in one system: • how strong is the risk of receiving pollen from related plants in natural populations? • do plants possess mechanisms of post-pollination selection against inbreeding (against related pollen or inbred embryos) despite receipt of related pollen?, • does post-pollination selection against inbreeding benefit offspring fitness? • do inbred males have reduced fertility (lower pollination and fertilization success) when exposed to natural pollinators? • which ecological factors (population size, flowering synchrony) affect the genetic diversity within pollen loads, and thereby the opportunity for post-pollination selection, offspring fitness, and selection on pollen performance?, and finally, • can results of experimental crosses be validated with effects observed for naturally-occurring variation in pollen load diversity and under conditions of natural pollination? By simultaneously answering these questions in one system, we should be able to understand the conditions favouring inbreeding avoidance in this dioecious species. This study should also allow us to evaluate whether costs and benefits differ for male and female plants, for which theoretical work predicts non-overlapping optima for inbreeding avoidance.