A range of tundra plant communities were experimentally perturbed in 1984 by 1) application of N P K-fertilizer to three stable and two frost-heaved sites and 2) by removal of the dominant deciduous shrubs from the three stable tundra sites. Effects on vertically projected plant area (PPA) on perturbed plots and on unperturbed controls were measured by point intercepts in 1985, 1986 and 1988. There were no significant changes of the PPA of vascular plants and bryophytes on the control and removal plots during 1984-1988, and few changes among lichens. Fertilization increased the PPA of vascular plants strongly, reduced the PPA of lichens and increased or reduced the PPA of bryophytes. The PPA of only two vascular species declined, whereas that of 21 species increased. Species diversity did not change consistently on control or removal sites. It increased considerably, however, among vascular plants after fertilization of stable tundra and declined on fertilized frost-heaved tundra. Among cryptogams (bryophytes plus lichens), species diversity always declined after fertilization. The lack of changes in remaining vascular species on stable tundra after removal of the shrubs suggests that competition from them was not a major limiting factor to the growth of co-occurring species. The increased rather than decreased diversity of vascular plants after fertilization of stable tundra, and the lack of negative effects on abundance of almost all species during the growth flush, demonstrate low competitive exclusion once nutrient limitation is released: indeed some positive plant interactions were identified. The decline of diversity among cryptogams on fertilized sites and among vascular plants on fertilized frost-heaved tundra indicate a higher sensitivity to competitive exclusion than among the vascular plants of stable tundra. It is suggested that the differences relate to community structure and differences in the clonality among the vascular components of the communities.