The species of vascular plants present in 0·5 ha sample plots were recorded in 19 areas of forest planted on former agricultural land, 21 areas of old or ‘virgin’ forest and 102 areas of ‘ordinary forest’ in east central Sweden. In the case of the planted forest areas, the main correlation with floristic diversity was with the amount of daylight reaching the forest floor. In the ‘old forest’ and ‘ordinary forest’ sites, however, the main correlations were with the pH value of the surface soil, the amount which this varied within the sample plot, and the proportion of broadleaved trees in the tree canopy; these factors were inter-correlated with each other to some extent. The ‘old forest’ sites, which had the most acidic surface soil and least diverse flora, did not appear to be serving any particularly useful function in terms of conservation of the flora, at the present time. It is concluded that, if a diverse forest flora is required for any particular reason, the complete closure of the canopy should be avoided. A mixture containing some broadleaved trees is likely to be preferable to a pure crop of pine or spruce. There appears to be, however, considerable scope for the development of a diverse flora within timber-producing forests consisting mainly of spruce and pine.