POLLEN‐OVULE RATIOS: A CONSERVATIVE INDICATOR OF BREEDING SYSTEMS IN FLOWERING PLANTS

The evolutionary shift from xenogamy (outcrossing) to autogamy (selfing) has been mediated through decreased flower size and alterations in floral morphology (Ornduff, 1969) which reduce the energetic cost per flower (Cruden, unpubl.) and facilitate self-pollination, respectively. Recently, a number of workers have shown that the flowers of self-incompatible and other xenogamous taxa produce more pollen grains than closely related self-compatible and/or autogamous taxa (Arroyo, 1973; Baker, 1967; Cruden, 1973; Gibbs et al., 1975; Lloyd, 1965; Vries, 1974). In contrast to other floral characteristics the number of ovules per ovary in the taxa studied did not decrease. In these taxa the evolutionary shift from xenogamy to autogamy has been accompanied by a decrease in the ratio of pollen grains to ovules (PIO). It is generally accepted that differences in breeding systems tend to be correlated with successional stages. Colonizers of disturbed habitats and plants of early successional stages tend to be self-compatible and/or autogamous (Baker, 1955, 1959, 1965; Stebbins, 1958) whereas plants of later successional stages, including tropical forest trees, tend to be xenogamous (Baker, 1959; Bawa, 1974; Bawa and Opler, 1975; Stebbins, 1950, 1958). Since, as I will show, PIO's are an integral part of a plant's breeding system, they also should be correlated with habitat or successional stage. My overall objective is to show that PIO's reflect the likelihood of sufficient pollen grains reaching each stigma to result in maximum seed set. The more efficient the transfer of pollen, the lower the PIO should be. It follows logically that cleistogamous flowers should have the lowest PIO's, and that autogamous flowers will have lower PIO's than xenogamous flowers, i.e., that PIO's are correlated with breeding systems. My data show that this is the case and suggest that PI O's are integrated with other facets of a plant's breeding system. Second, I show that there is a strong correlation between the breeding system, hence PIO's, and successional stage. In addition, I briefly circumscribe what constitutes a sufficient number of pollen grains to assure maximum seed set and discuss evidence that supports the notion that some plants tend to minimize pollen production. Finally, I discuss deviations from the general pattern to show that PIO's are a better predictor of a plant's breeding system than other morphological characteristics.