The vascular flora of 62 Danish beech forests of eastern Jutland ranging in size from 1-445 ha, was investigated for species-area relations. Species richness reflecting total diversity, forest diversity, and of different habitat groups, were corrected for non-linearity by means of a log-log power function transformation and regressed to forest area. The transformed forest species diversity showed a negative relationship with forest area. It is highly questionable whether the often stated positive correlation between forest area and species number is valid when the total species numbers are corrected for the area-error. The often stated theorem that the larger the forest, the smaller the disturbance, is not valid in Danish broad-leaved forests. On the contrary, if the number of ruderals is taken as an indication of human disturbance, the correlation of area with the proportion of ruderals to all species and to forest species, both failed to detect correlation with area. This is probably due to the less intensive management regimes in small forests compared with larger forests. A substantially larger number of forest species is found in ancient forest than in middle-aged forest. This shows the great importance of forest continuity for the species diversity. Lists of indicator species for ancient and old broad-leaved forest are presented. The corrected species diversity value is recommended as a good nature quality indicator, to be used in comparative studies and for assessing biological quality for management and conservation actions.