Abstract
BACKGROUND AND AIMS: Pollen:ovule ratios are often lower in species and populations with higher selfing rates. This may be due either to higher pollination efficiency through selfing, or to lower male competition when less allo-pollen is available. Changes in pollination can also impact pollen traits, such as the number of apertures. Viola arvensis has experienced a rapid recent increase in selfing rates, and a rapid floral trait evolution towards the selfing syndrome. This study tests the hypothesis that V. arvensis is also undergoing a rapid evolution in its pollen:ovule ratio and pollen heteromorphism. METHODS: Using the resurrection ecology methodology, we compared four ancestral populations (from ca. 30 years ago) to their descendants (from 2021). We counted ovules and pollen and measured the different pollen aperture morphs in 50 individuals per population. We also developed a model to better understand the links between the number of apertures and the mating system. KEY RESULTS: We found no temporal change in pollen or ovule production. However, populations with the lowest pollen:ovule ratios were also those with the highest ancestral selfing rates, suggesting that the pollen:ovule ratio could have evolved on a similar timescale to population differentiation. Our model predicts a positive correlation between number of apertures and selfing rates, if pollination parameters remain constant. However, this positive correlation was not found in our results, neither across populations nor through time. CONCLUSIONS: Unlike floral morphology, pollen and ovule production did not evolve rapidly with increased selfing rates, suggesting a delayed change of the pollen:ovule ratio compared to other traits of the selfing syndrome. The absence of correlation between pollen heteromorphism and selfing rate can be explained by multiple (non-mutually exclusive) factors: a decrease in allo-pollen deposition correlated with the evolution of the selfing syndrome, pollinator declines or the absence of selection in this trait.