The essence of the contradiction between traditional ecological complexity-stability hypothesis and recent theoretical results is clarified. The distinction between resilience and resistance is stressed. The possibilities of field verification of May's model are discussed. No satisfactory method for estimation of connectance and mean interaction strength in plant communities has been found. Relation between these parameters and stability in real communities remains an open question. The relation between connectance and stability (resilience) in purely competitive model communities is more complicated than May's rule predicts. The certain value of connectance having been achieved, stability increases with increasing connectance. We assessed the positive relation between species diversity and resistance, and negative relation between species diversity and resilience in plant communities during old-field succession in xeric habitat. But there is no causal relationship between species diversity and both kinds of stability. Resistance and resilience of the plant communities studied were determined primarily by life history strategies of constituent species. The results are interpreted in terms of Grimes' life history strategies and imply the validity of Gleasonian, population-centered explanation of community phenomena.