At Swiss long-term forest ecosystem research sites, ground vegetation was assessed during the period 1994-2003/2008 following two approaches: (1) visual assessment of the cover of species occurring in sixteen 1 m(2) quadrats, distributed over a 43 × 43 m area, and (2) phytosociological relevés in concentric circular plots of 30, 200, and 500 m(2). We first compared the two approaches with respect to diversity assessment. The number of species recorded in the 16 quadrats was in general higher than in the 30 m(2) plot and it represented 42% to 108% of the number of species recorded in the 500 m(2) plot. In a second step, we tested whether any temporal trends were apparent. In a few cases, a decrease or increase in Landolt's mean indicator values for light, nitrogen availability, soil pH, soil moisture, or temperature was found to be significant. However, these changes were usually restricted to one approach or one area. The only clear trend was detected in an unmanaged former coppice beech stand, for which all survey approaches indicated canopy closure. At another site, vegetation reacted to the local opening of the canopy following windthrow. In a third step, we compared the leaf area index (LAI), measured with an LAI-2000 instrument (Licor, Inc.) over each quadrat, with the indicator value of the vegetation for light (L). Within a site, there was no clear relationship between LAI and L values per quadrat. In contrast, across all sites, the relationship between LAI and L, averaged per site for all available years, was highly significant.