The aims of the study were: (i) to examine how well the hillock forests are distinguishable from the forests of comparable site types, (ii) to ascertain the hillock forest indicator species, (iii) to elucidate the hillock forests’ community types, (iv) to assess the practical requirement for differentiation of hillock forests as a self-sufficient typological unit for forestry. The data included altogether 160 descriptions of hillock forest communities and 42 relevés of other site type stands for comparison. The historical continuity of hillock forests was estimated from old topographic maps. We established that the hillock forests constitute statistically a reliable distinct group of forests. Among the characteristic species having an indicator value significance level of up to 0.50, by the ecological strategy 43.2% belong to competitors, 37.8% are competitors and stress-tolerant ruderals, and 10.8% competitors and stress tolerators; according to the hemeroby, 62.5% of these species are apophyts and 30.0% hemeradiaphors. Most hillock forests (55.6%) are located on former slash-and-burn areas (bushlands) or reforested agricultural land (36.1%). The hillock forests can be classified into four forest types: 1)