A procedural model for vegetation analysis is presented. Suggestions are made that analysis methods can test theory as well as examine vegetation-environment correlations. Gauch and Whittaker's propositions regarding species behavioural properties expected for an individualistic continuum are tested on a eucalypt forest data set. The data set is carefully stratified to control environmental heterogeneity. The shape and distribution of species response curves are then examined along a temperature gradient using 750 sites. The conclusions are:
(1) Bell shaped response curves to environmental gradients are not universal (2) Positive-skewed curves are characteristic of major canopy species in eucalypt forest in southern New South Wales (3) Species richness increases with temperature along the gradient (4) Tests of other propositions regarding species modes and ranges are confounded by the change in species richness along the gradients (5) More rigorous statistical analysis and analyses on other vegetation types are needed. * 1 thank my colleagues A. O. Nicholls and C. R. Margules for collaboration in collecting data and analysis, and P. R. Minchin, D. P. Faith, C. R. Margules and E. Van der Maarel for comments on the manuscript. Thanks are also due to J. L. Stein for assistance with the analysis and numerous Australian ecologists who have made their data available to me.