Relations between soil factors and herbage yields of natural grassland on sandy soils in the south‐eastern Transvaal

Abstract

Forty systematically‐placed observation points were established in a 1.5‐ha area of floristically unifoim veld in which productivity varied spatially. At each point grass yields were determined by harvesting quadrats at eight‐week intervals during the growing season, and the volumetric water content of the soil was measured weekly. In addition, root mass and root‐N content, and soil variables involving C and N contents, extractable P, K, Ca and Mg, cation exchange capacity and base saturation, pH, clay content and soil depth, were determined. The data were examined using linear correlations between variables, principal components analysis and multiple regression. The major factors associated with variations in yield were, in order of importance, soil water content, the N supply (as expressed by root‐N content or root‐N mass) and the cation supply (as expressed by the topsoil variables, extractable Ca or cation exchange capacity or percentage base saturation). There was a relatively weak positive association between yields and the H⁺ concentration in the subsoil, and a weak negative association between yields and the clay content of the topsoil. A multiple regression function involving all five factors accounted for 77.3% of the variation in dry matter yield.