Transformation processes in countries like Kyrgyzstan often. lead to intensification or extensification of grazing on mountain pastures. In order to reveal the impact of livestock grazing on vegetation patterns, we examined traits of dominant and frequent species in four previously classified plant communities. In particular we analysed trait-environment relationships using multivariate RLQ analysis and univariate fourth-corner statistics in order to identify most relevant environmental gradients. Functional Response Groups (FRG's) of plant taxa were derived from clustered RLQ ordination space, and were subsequently analysed for their proportionality in previously identified plant communities. Plant height, growth form, lateral spread, flowering time and life cycle showed a high degree of correlation with grazing. RLQ analysis revealed a high differentiation of plant trait values along a grazing gradient on the first RLQ axis. FRG's matched three of four vegetation types by constancy and fidelity values (phi) well. Thus, trait-based analyses were successfully used to identify functionally similar groups of species with regard to different regimes of grazing pressure.