The problem of interpreting the effects of cultural treatments on the chemical composition of leaves is studied by successively applying a number of alternative methods to the same example involving repeated sampling over a number of years. These methods are the analysis of variance applied to each element separately, the use of ratios of pairs of elements, the use of a system of trilinear co-ordinates to express the N-P-K balance and the K-Ca-Mg balance, and the simultaneous study of all elements by means of Principal Component Analysis. The relative merits of the different methods are assessed by their ability to produce consistent results from year to year. As the number of elements to be examined simultaneously is increased, so there appears a progressive increase in consistency until the method of interpretation is one involving all elements. It is concluded that a reasoned application of the Principal Component Analysis approach, which places no restriction on the number of elements that can be examined, can lead to a greater understanding of the effects of treatments on leaf composition as a whole than any examination of individual elements or simple ratios.