Complementarity of the nature reserve network in southernmost Finland was examined using a simple heuristic algorithm and occurrence data of 75 characteristic herb-rich forest vascular plant species in 126 protected and 120 non-protected sites. Three different minimum sets were selected to represent 1, 5, or 10 occurrences of each species. In each minimum set there were significantly more protected than non-protected sites. Thus, although efficiency is not maximal in the existing reserve network, the network does provide a representative basis for the conservation of herb-rich forest plants. However, some deficiencies were also noticed, particularly new reserves in floristically diverse herb-rich forests along watercourses would supplement the existing reserve network. On the other hand, the growing concern over the shortcomings of reserve selection studies using presence / absence data is echoed here, because many of the protected forests not included in the minimum sets harbour important nature conservation values. With regard to the five biological reserve selection criteria (e. g. occurrences of threatened species) considered here, the performance of the existing reserve network is rather good. However, examination of five reserve design criteria revealed some deficiencies, particularly vulnerability of many reserves to potential edge effects.