PATTERNS IN SPECIES ASSOCIATIONS IN PLANT-COMMUNITIES - THE IMPORTANCE OF SCALE

Present discussions on competitive interactions and the occurrence of predictable patterns in species composition including assembly rules - are likely to benefit from appropriate analyses of the spatial struct ure in plant communities. We suggest such an analysis when we specific ally want to detect scale regions where fine-scale local processes may affect the spatial pattern of species composition. We combine indirec t ordination in the form of Detrended Correspondence Analysis (DCA) an d geostatistics in the form of variography. The species abundance data in the sampled quadrats are summarized as positions on the axes in th e ordination. Each axis is used as a regionalized variable in the vari ography to obtain the spatial dependence of the quadrats. The spatial pattern found will suggest the relevant scale region in which to perfo rm an analysis of species associations. A significant spatial dependen ce (the 'range' in geostatistical jargon) will define the size of a sa mpling plot that will minimize both the problem of being too small and thus having the risk of oversampling of e.g. clonal individuals and o f being too large which will risk including individuals that do not in teract. We also suggest that plots are spaced at least a 'range' apart to insure spatial and statistical independence. Comparisons of specie s compositions in such plots will reveal any positive or negative asso ciations between species on a scale where these should reflect species -species interactions. To illustrate the method it is applied to three different data sets from two different plant communities.