Abstract
Measurements of canopy area (A) and visual scores for clove bud differentiation (S) were used to describe the variability in clove tree yields using regression analysis. Of the fitted variates, A + S accounted for more variability than A or S individually. For some datasets, unstable variation in yield, A and S necessitated the use of logarithmic transformations before the sum was fitted in the regression analysis; the relationship between yield and A and S was then of the form of the product between A and S. The use and some of the limitations of this function as a substitute for yield data or as a means of estimating tree yields are discussed.
