Volunteers deriving from unharvested seeds of a crop can lead to persistent feral populations and participate in genetic exchanges across the agro-ecosystem, both between crop varieties and between crops and their wild relatives. A first step to understand the importance of volunteers is to characterize their capacity to reproduce autonomously for several generations. For that purpose, we constructed and evaluated a maximum-likelihood method to estimate the genetic age of a population deriving from one of the most common field crop type: an F1-hybrid variety. The method estimates the number of reproduction cycles that occurred since the cultivation of that variety. It makes use of genotypic data at a number of linked microsatellite loci pairs, thus exploiting the recombination of parental haplotypes, which is expected to occur as the population is reproducing. Estimates with moderate bias and variance were found for a broad range of parameter values in simulations, and the method revealed robust to some deviations from the assumptions of the underlying model. We propose a specific procedure to test the hypothesis of persistence, that is has a given volunteer population experienced more than one cycle of reproduction since the F1-hybrid state? The method was applied to both an experimental and a natural sunflower volunteer population and revealed promising, considering these ideal case studies. Possible further developments toward more complex natural systems are discussed.

Cultivated plants have been molded by human-induced selection, including manipulations of the mating system in the twentieth century. How these manipulations have affected realized parameters of the mating system in freely evolving cultivated populations is of interest for optimizing the management of breeding populations, predicting the fate of escaped populations and providing material for experimental evolution studies. To produce modern varieties of sunflower (Helianthus annuus L.), self-incompatibility has been broken, recurrent generations of selfing have been performed and male sterility has been introduced. Populations deriving from hybrid-F1 varieties are gynodioecious because of the segregation of a nuclear restorer of male fertility. Using both phenotypic and genotypic data at 11 microsatellite loci, we analyzed the consanguinity status of plants of the first three generations of such a population and estimated parameters related to the mating system. We showed that the resource reallocation to seed in male-sterile individuals was not significant, that inbreeding depression on seed production averaged 15-20% and that cultivated sunflower had acquired a mixed-mating system, with B50% of selfing among the hermaphrodites. According to theoretical models, the female advantage and the inbreeding depression at the seed production stage were too low to allow the persistence of male sterility. We discuss our methods of parameter estimation and the potential of such study system in evolutionary biology. Keywords: sunflower; mating system; gynodioecy; population genetics; crop evolution INTRODUCTION Cultivated plants are the product of the adaptation of wild species to cultivation and use by humans. Domestication resulted in dramatic changes in morphological (for example, plant architecture), biochemical and life-history traits (seed dormancy, phenology, mating systems; Gross and Olsen, 2010) and made crops outstanding study systems for investigating the genetic bases of adaptation (Ross-Ibarra et al., 2007). Moreover, a great deal of genetic mechanisms involved for instance in the resistance to pathogens or in different aspects of plant reproduction (self-incompatibility, restoration of male fertility) have been first unraveled in crop plants (Hancock, 2005). Reciprocally, when human control is at least partially relaxed, populations of cultivated plants are confronted with a new set of evolutionary forces and can adapt quite quickly to new environmental conditions. This is exemplified by the recurrent evolution of weeds from escaped cultivated plants (Ellstrand et al., 2010), or by dynamic management experiments where artificial populations of crops have been placed in contrasted environments (for example, differentiation of heading time after 10 generations among populations of dynamic management of wheat; Goldringer et al., 2006).Mating system has been affected by domestication and the subsequent improvement of cultivated plants (Glémin and Bataillon, 2009). Recently, it has been the target of intentional m...