Swiss forest releves were analysed in order to answer the two questions: a) which are the predominant gradients causing differences in forest vegetation all over the country and b) how does a calculated gradient system of forests fit Ellenberg's ecogram of a well defined region. Two large species data sets (6488 and 1600 releves each) were analysed with Correspondence and Canonical Correspondence Analyses (CA, CCA). Averaged Landolt's indicator values for moisture, nutrients, reaction, temperature and light served as primary environmental variables, averaged indicator values for humus and continentality, soil-, temperature- and precipitation variables served as secondary. Moisture, soil warmth, soil reaction and light turned out to be the most important factors when referring to the whole country. Soil reaction in combination with soil warmth and moisture were found to influence the submontane forests of the northern parts of the Alps predominantly. With respect to Ellenberg's ecogram, our results confirm the usefulness of the old system. In contrast, recent propositions on paralleling acidity and nutrient richness on the first ecogram axis, prove to be misleading when comparing calculated and expert-derived gradients. We consider our quantitatively derived ecogram a valuable alternative to hitherto available expert-solutions in forest classification projects.