Supplying capacity, response to fertilization and residual effect of K fertilization were evaluated in a twelve-year soybean field experiment on a Haplohumox soil, in Campos Novos, State of Santa Catarina, Brazil. Treatments consisted of: four levels of K2O (0, 80, 160 and 320 kg ha-1), applied only in the first cropping, on the whole plots; three levels of K2O (0, 40 and 80 kg ha-1), applied annually in the sub-plots from the fifth year; one treatment with application of 40 kg ha-1 of K2O in the furrow. All treatments were applied in a split plot arrangement in RCB design with three replications. Soil analysis and grain yield were measured every year. There was no response on soybean yield to K fertilization during the first four years. Beginning the fifth year, all treatments had higher yields than the control. Yield response to annual fertilization increased with time and with decreasing residual K rates in the soil. On the average of 12-year period, the grain yield increased 18 kg per kg of K2O applied at the first cropping. A fertilization of 60 kg ha-1 K2O year-1 may maintain soybean yield at about 90% of maximum productivity over a 12 year period.
Glycine max (L.) Merrill; grain yield; potassium; availability and residual effect