Population strategies in severe environments: Alpine plants in the northwestern Caucasus

This study attempts to show a differentiation into strategy types among plant species of four alpine communities in the northwestern Caucasus, Russia. Four groups of population characteristics (traits) were examined: above-ground biomass, seed size, seed yield and occurrence of a persistent seed bank. In addition, values for relative growth rate and a morphology index were estimated. Agglomerative Cluster Analysis and Principal Component Analysis were applied to a data set with 42 species from these communities and six species characteristics. In the Geranium gymnocaulon meadow, the most productive community, large gaps occur created by wild boars and bears. Here, the response of 13 species to disturbance was also analysed. Strategy types were approached both according to Grime: 'competitive', 'ruderal', 'stress-tolerating' and according to Ramensky, Rabotnov and Romanovsky: 'violent' dominant (V),'explerent' explorative(E) and 'patient' endurant (P). Differentiation into strategy types was more obvious in productive alpine grasslands than in lichen heaths and snowbed communities. The grassland dominants Festuca val-ia, Geranium gymnocaulon and Hedysarum caucasicum showed a more or less 'violent' strategy. Species approaching an E-strategy, e.g. Matrical ia caucasica, Sibbaldia procumbens and Gnaphalium supinum, were also found in the grasslands but prevailed in the snowbed communities. Species with a 'patient' strategy prevailed in the lichen heaths. We conclude that plant strategy types are well-differentiated, this is also seen in plant communities occurring under severe high-mountain conditions. We also suggest that the approaches by Crime and Ramensky/Rabotnov/Romanovsky are not so similar as has been assumed earlier and that the R/R/R types are better differentiated and ecologically better interpretable. We emphasize the gradient character of strategy 'types' and would characterize plant species by the degree of adaptation to conditions: (I) from small seed size and large persistent seed bank to large seed size and absence of a seed bank; (2) from high to low biomass and seed production, with relative growth rate as a less important partially correlated trait.