Key points are not available for this paper at this time.
(1) Conventional ordination of repeated enumerations of the same set of stands over a period of time is liable to result in confusion between spatial and temporal trends. (2) An alternative technique for isolating temporal trends is to eliminate the effect of initial differences between stands. Between-species correlations are calculated from the species x time matrix for each stand separately. The correlation matrices from individual stands are summed, and the summed matrix is subjected to principal components analysis. (3) The method is illustrated with data from montane grasslands from which sheep have been excluded, in Snowdonia, North Wales.
Swaine et al. (Sat,) studied this question.