Abstract
An important site index study on the black wattle (Acacia mearnsii De Wild.) in southern Africa had been performed in the 1960s and many of the recommended site requirements in this region for this species have been derived directly or indirectly from this study. Serious deficiencies, however, were apparent in the statistical manner in which the study had been pursued, which prompted a re-evaluation of the data using a more appropriate multivariate technique (principal components regression). While this re-evaluation revealed that the overall conclusions of the original study were in essence correct (particularly that soil depth is the most important of the site characteristics measured, in estimating site index of black wattle), undue credibility had been placed in the use of numerous laboratory-derived soil chemical and physical properties.