1. Controlled self- and cross-pollinations were performed on field bean plants. It was found that while manipulation of the flower stimulated seed setting, both self- and cross-pollination were equally effective.

2. Estimations of pollen germination on artificial medium, to which had been added extracts of styles or ovaries from the same or a different plant, gave some slight evidence of a specific inhibition by ‘self’ extracts. No analogous effect was detected on pollen tube growth.

3. Experiments in which the style was cut at varying intervals after pollination detected no difference between self- and foreign pollen in rate of growth through the style. Pollen placed on stigmas of immature flowers began growth at once although the fertility achieved was less than when mature flowers were used.

4. The composition of seed from mixed self- and cross-pollination was examined. Early application of foreign pollen increased the proportion of crossfertilization but the results were similar in hybrids and inbreds.

5. Flower manipulation was more effective in stimulating seed setting than shaking the plants. Spraying with hormone solution was ineffective.

6. It was found that plants varied in their capacity to set seed without flower manipulation (autofertility). Hybrid plants were more autofertile than plants from inbred lines.

7. X-ray pictures of unopened flowers suggested that the pollen may be more abundant and less diffuse in hybrids than in inbreds.

8. The expression of autofertility was affected by many genetic factors.

9. Seed produced by open pollination of F1 hybrids contained a much higher proportion derived from self-pollination than did seed from inbred plants grown in the same plot. Possible explanations for this difference, and some consequences for the population genetics of the species, are discussed.