Will similar forests develop on similar sites?

Abstract

Abies grandis, Taxus brevifolia, Thuja plicata, or any combination of these may dominate old-growth mesic forests of the Bitterroot Canyons, western Montana. Similar sites need not develop similar, relatively stable forests. This is shown by (i) anomalous distributional patterns of tree species, (ii) broad overlap of tree species abundance in environmental space (shown by ordination and discriminant analysis of stands in environmental space), and (iii) weak or undetectable correspondence of species × stand and site factor × stand matrices (multiple regressions of compositional dissimilarity against environmental differences; also, canonical correlation and Mantel tests). Since a one-to-one mapping from site factors to species composition in old-growth vegetation is a fundamental tenet for applications of the climax concept, caution is warranted where the concept is to be applied within a narrow range of site factors or to insular communities.