Two field experiments were carried out. In the first experiment, 13 variables were studied for 3 inbred groups as influenced by genetic differences, source of pollen, soil moisture, fertilizers , and size of bulbs. Path-coefficient analysis was made of var ious insect pollinators influencing seed yield in different inbreds, and of components of seed production in all inbreds together. In the second experiment, 4 variables were studied for one male-sterile inbred and one male-fertile inbred as influenced by soil moisture and fertilizers. Genetic differences among inbreds were the most important source of variation in seed yields. This points out that a major portion of the onion seed problem could be solved through plant breeding. The inbreds differed very significantly in t heir attractiveness to pollinating insects, probably as a result of varying genetic vigor which influenced nectar formation. Most of the inbreds had negative correlation or no correlation between the number of honey bee nectar collectors and seed yield. The separation of correlation with honey bee nectar and pollen collectors into components of direct and indirect effects indicates that the number of pollinating insects was excessive for all inbreds in the experiment. Negative direct effects on seed yield and negative indirect effects through percent fertilized flowers and seeds per fertilized flower disclose some important aspects as possible causes for low onion seed yields: 1. Depletion of assimilated food needed for the development of fertilized ovules through an overdraw on nectar by nectar collectors. 2. Water deficit affecting fertilized developing ovules due to increased transpiration resulting from overhandling of flowers by honey bees, especially by pollen collectors, which through nervous search for pollen on male- sterile lines might cause enough physical damage to increase transpiration even at low visitation levels. At high visitation levels, pollen collectors might also harm male-fertile lines. 3 . Removal of pollen from the stigmas of male-sterile flowers by high populations of pollen collecting honey bees. Compatibility of the pollen parents with the male-sterile inbreds was influenced by soil moisture levels. At high moisture, M611C increased seed yields while B12115-2C decreased them. At low moisture the reverse was true. In the first experiment there was no significant yield response to varying soil moisture levels due to contrasting effects of high soil moisture on compatibility with pollen parents. In the second experiment, the pollen parent B2215C was used and a significant favorable response to high soil moisture was obtained. Salinity aspects and the effects of indiscriminate use of fertilizers were tested through the method of fertilizer placement. While fertilizers reduced seed yield , high soil moisture tended to reduce the negative effects of nitrogen and to produce favorable response to phosphorus. Seed yield followed the trend in plant survival as affected by the induced salinity x moisture relationships. The size of bulbs had a great influence on seed yields, mainly through an increased number of umbels per plant. It seems that the number of umbels per plant had more influence on the number of flowers per row than did the number of flowers per umbel. Path-coefficient analysis of components of onion seed production indicates that the percentage of fertilized flowers is the most important criterion for selection to improve seed yield. It would be ideal to combine both high percentage of fertilized flowers and large number of flowers per umbel, but if compromise has to be made, selection would best be considered on the basis of greater percentage of fertilized flowers in the umbel. Plants which have a large percentage of fertilized flowers can be easily detected at the pre-seed harvest stage.