An increase in atypical petal numbers during a shift to autogamy in a coastal sand verbena and potential evolutionary mechanisms

PremiseIn plants, meristic traits, such as petal and sepal numbers, are usually considered invariant within taxa, yet certain species consistently exhibit great variability in these traits. The factors contributing to "atypical" counts are not well-known, published hypotheses include relaxation of pollinator selection, inbreeding, and hybridization, among others. The sand verbenas, Abronia (Nyctaginaceae), usually have five perianth lobes ( petals), yet certain taxa exhibit marked departures from this norm. MethodsHere we integrate an analysis of images from community science data (iNaturalist) and common garden experiments to evaluate a comprehensive set of adaptive and nonadaptive explanations for the production of these atypical flowers across an evolutionary transition from xenogamy (outcrossing) to autogamy (selfing) in the coastal sand verbena Abronia umbellata. Key resultsThe shift to autogamy in this lineage correlated with a higher frequency of atypical flowers from ~7% to ~20% and a significant reduction in mean petal number per inflorescence. Autogamous success did not change with petal number, and neither hybridization or up to three generations of inbreeding consistently increased production of atypical flowers or decreased mean petal number, all in contrast to previously-published hypotheses. In contrast, intra-inflorescence, inter-plant (intra-population), inter-population, and inter-variety comparisons demonstrated a correlation of reduced floral size with reduced petal number, suggesting correlated evolution due to a well-established relation between organ number and meristem size. ConclusionsThe reduction in petal number was probably a consequence of selection for smaller flowers associated with increased selfing. While we could not completely eliminate several alternative hypotheses, including a long-term history of inbreeding or relaxed selection on petal number constancy, those are less likely to explain the observed changes, though they may have contributed to the trend. In general, we develop a framework of hypotheses for evolutionary investigations of meristic variation in floral organs.