The ratios K+/(Ca++ + Mg++)1/2 and K+/(C++ + Mg++) in the soil and the responses to potassium fertilization.

Abstract

The present study tries to clarify the fact, largely found in literature of soils with low values of exchangeable K⁺ which do not respond to potassium fertilization, and of soils with values of exchangeable K⁺ considered satisfactory, which respond to potassium fertilizer. The results of 74 experiments were studied together several soils with their analyses and several crops, applying the relations K⁺/(C⁺⁺ + Mg)^1/2 and K⁺/(C⁺⁺ + Mg⁺⁺). The statistical method of logistic regression was used, carried out through CATMOD procedure of Statistical Analysis System. It was concluded that the 0.20 index obtained for a ratio K⁺/(C⁺⁺ + Mg)^1/2 is the limit from which no response to potassium can be expected and that 0.13 index marks the limit from which response to potassium fertilizations are progressively lower and, therefore, antieconomic. It was concluded that exchangeable K⁺ values, by themselves, do not give correct information about potassium fertilization needs, and that the potassium supplying power of soils and liming to be applied which changes the ratio K⁺/(C⁺⁺ + Mg)^1/2 are also responsible for unexpected results in experiments with potassium fertilization.