Renge (Astragalus sinicus L.) is not self-sterile and produces self-fertilized seeds easily by artificial self-pollination, but most of seeds set in open fields might have been cross-fertilized by insects. Natural crossing percentages were studied in individual Renge plants and also in populations surrounded by other variety in the field. In a propagation field of normal red-flower variety, white-flower plants and “ever-flowering” plants grown in pots were arranged at the rate of one plant per 3.3m² during ghe glowering period. Then seeds were harvested form the white-flower plants and the eber-flowering plants individually. Characters of each progenty were examined. Because white flower or ever-flowering is recessive to red-flower or nomal type, it is considered that normal or red flower plants in the progenies are cross-fertilized, and crossing percentage of mother plants can be estimated form the ratio of normal type and recessive type the next generation. Average crossing percentage of white-flower and ever-flowering plants were 85.71±5.95 and 87.62 ±7.24, respectively (Table 1 and 2). As the crossing percentage of wheite-flower plants and ever-flowering (red-flower) plants were not so different, it was ^suggested the visiting frequsncies of insects to white-flowers and red-flowers were not defferent. A white-flower variety was cultured in square, 21.8m×21.8m, and a red-flower variety was grown around it and during the flowering period they were open-pollinated. Then seeds were harvesed from the white-flower platns seperated by direction and distance from the red-flower variety (Fig. 1). The crossing percentage of them were examined by the character of their progenies. The results are shown in tables 3 and 4. Difference of crossing percentage was not signigicant in ghe direction, but was significant in the distance. Average crossing percentage in the distance 0-0.9m from red-flower variety was 18.2%m 10, 0% for 1.8-2.7m, 6.6% for 3.6-4.5m, and inside of them, 6.1-4.5%. From the data presented, it was suggested that in natural condition crossing percentage of an individual Renge ploant was high, the fertilizing pollen being brought from its neighbourhood. Methods to protect the new variety from crossing with other varieties in the propagation field were discussed.