The yield response of soybean [Glycine mau (L.) Merr.] to foliar fertilizat
ion during reproductive stages has been inconsistent. This study evaluated
soybean responses to foliar applications of a 3-8-15 (N-P-K) fertilizer at
early vegetative stages in 48 trials conducted on Iowa soils that tested mo
stly optimum or above in P and K. In 1994, the treatments were a control, s
ingle applications of 19, 28, or 38 L ha(-1) at the V5 stage, and 38 or 56
L ha(-1) split as one-half at the V5 stage and one-half 8 or 9 d later. In
1995 and 1996, only the single rate of 28 L ha(-1) and the split rate of 38
L ha(-1) were used. The lowest and highest volumes used encompassed N-P-K
rates of 0.8-2.1-3.9 to 2.4-6.3-11.7 kg ha(-1). All treatments were replica
ted four times. Some or all treatments increased yields in se,en sites and
decreased yields at two kites. Mean yield increases were 60 kg ha(-1) in 19
94 (not significant), 30 kg ha(-1) in 1995 (not significant), 60 kg ha(-1)
in 1996, and 54 kg ha(-1) across the 48 sites. Differences between treatmen
ts were small and inconsistent, and the single application of 28 L ha(-1) (
one of the lowest rates used) produced the highest mean yield increase acro
ss all responsive sites (375 kg ha(-1)). In 1994, the higher yield response
s occurred on ridge-till and no-till fields and when the P concentration of
young plants was low. No treatment caused leaf burning. Foliar fertilizati
on seldom increased P and K concentrations of leaves at the R2 growth stage
and did not affect plant maturity, grain moisture, or the weight of grains
. No simple set of measurements explained the occurrence of yield responses
. Results of factor and regression analyses suggested that responses tended
to occur in soils with high cation exchange capacity, when plant-available
P was Low, and/or when rainfall in spring and midsummer was low. This grou
p of variables explained only 14 to 23% of the responses in different years
, however. Effective use of foliar fertilization of soybean at early growth
stages in areas with predominantly high-testing soils requires further res
earch to identify the conditions under which positive responses rue more li
kely.