This study evaluated soybean (Glycine max (L.) Merr.) response to foliar fertilization at early vegetative stages in Iowa soils that tested mostly high in P and K. Several studies evaluated fertilization effects on grain yield; N, P, K, composition of vegetative tissue; and photosynthesis. A 3-8-15 (N-P-K) mixture was evaluated at 21 trials in 1994. Treatments were a control, single applications of 19, 28, or 38 L ha-1 at the V5 stage, and 38 or 56 L ha-1 split one half at the V5 stage and one-half 8 to 9 days later. Three fertilizers were evaluated at 27 trials in 1995 and 1996. Treatments were a control, 28 L ha-1 of 3-8-15, a split application (in two) of 38 L ha-1 of 3-8-15, a single application of 28 and 56 L ha-1 of 10-4-8, and a single application of 42 L ha-1 of 8-0-7. Soil and plant samples were collected before the first spray. Trifoliolate leaves and whole-plants were collected at the R2 stage and apparent photosynthesis was measured in eight sites. Some or all treatments increased yield at 19% of the sites and decreased yield at 8% of the sites. Differences between treatments were inconsistent and the 28 L ha-1 of 3-8-15 usually produced the highest yield. The mean response to this treatment was 270 kg ha-1 at responsive sites and 70 kg ha-1 over all sites. Only the higher rate of 10-4-8 caused moderate leaf damage but damage was not obviously related to yield decreases. Fertilization seldom increased plant nutrient content, photosynthesis, plant maturity, or weight of individual grains. In 1994, the higher yield responses occurred in no-till and ridge-till fields and when plant P concentration was low. Factor and regression analyses over sites showed that positive responses occurred mostly in soils with high cation exchange capacity and when plant growth, plant-available N or P, and (or) rainfall in spring or summer was low. These variables explained 14 to 27% of the responses in different years. Further research is needed to identify the conditions under which responses are more likely.