Natural disturbances have been traditionally defined in terms of major catastrophic events originating in the physical environment and, hence, have been regarded as exogenous agents of vegetation change. Problems with this view are: (1) there is a gradient from minor to major events rather than a uniquely definable set of major catastrophes for each kind of disturbance, and (2) some disturbances are initiated or promoted by the biotic component of the system. Floras are rich in disturbance-adapted species. Disturbances have probably exerted selective pressure in the evolution of species strategies. Heathland cyclic successions and gap-phase dynamics in forests have been viewed as endogenous patterns in vegetation. When death in older individuals imposes a rhythm on community reproduction, dynamics may indeed be the result of endogenous factors. However, documented cases of senescence in perennial plants are few and many cyclic successions and cases of gap-phase dynamics are initiated by physical factors. forest dynamics range from those that are the result of individual tree senescence and fall, through those that are the result of blowdown of small groups of healthy trees, to those that are the result of large windstorms which level hectares of forest. The effect of wind ranges from simple pruning of dead plant parts to widespread damage of living trees. Wind speed is probably inversely proportional to occurrence frequency. Disturbances vary continuously. There is a gradient from those community dynamics that are initiated by endogenous factors to those initiated by exogenous factors. Evolution has mediated between species and environment; disturbances are often caused by physical factors but the occurrence and outplay of disturbances may be a function of the state of the community as well. Natural disturbances in North American vegetation are: fire, windstorm, ice storm, ice push on shores, cryogenic soil movement, temperature fluctuation, precipitation variability, alluvial processes, coastal processes, dune movement, saltwater inundation, landslides, lava flows, karst processes, and biotic disturbances. Disturbances vary regionally and within one landscape as a function of topography and other site variables and are characterized by their frequency, predictability, and magnitude. The landscape level is important in assessing disturbance regime. Disturbances and cyclic successions belong to the same class of events-that of recurrent dynamics in vegetation structure-irrespective of cause. Dynamics may result from periodic, abrupt, and catastrophic environmental factors or they may result from an interaction of the changing susceptability of the community and some regular environmental factor. In any case, the dynamics result in heterogeneous landscapes; the species adapted to this heterogeneity are numerous, suggesting their long time importance. The importance of disturbance regime as part of the environmental context of vegetation means that allogenic and autogenic models of vegetation are difficult to apply. Species composition can be seen to be a function of disturbance regime, as well as other environmental variables. Competitive replacement in succession occurs, then, only as disturbances cease to operate and can be viewed as allogenic adjustment to a new disturbance-free environment. Competitive divergence, separation of role, and competition avoidance may, in fact, underlie successional patterns traditionally viewed as the competitive replacement of inferior species by superiorly adapted climax species. The importance of ongoing dynamics is also difficult to reconcile with the concept of climax, founded as it is on the idea of autogenesis within a stable physical environment. Climax is only arbitrarily distinguished from succession. Climax as an organizing paradigm in plant ecology has obscured the full temporal-spatial dimensions important in understanding the vegetated landscape and the evolution of species which contribute to the landscape patterns. Whittaker's coenocline concept is accepted with modifications: (1) natural disturbance gradients and Whittaker's complex gradient are intimately related, (2) temporal variation in the community should be viewed as an added axis of community pattern, and (3) ongoing dynamics have important effects on specificity of species to site relations and the predictability of vegetation patterns. Recent work has suggested an r-K continuum in species strategy. In general, colonizing ability is seen as a trade-off against specialization. Frequent disruption of the community and the creation of open sites seems to result in mixes of species that are fleeting in time and do not repeat in space. Species in such mixes are often tolerant of wide environmental extremes but are compressed into early successional time if disturbance ceases. The composition of such communities is not predictable from site characteristics. Even communities with low disturbance frequency lack complete environmental determinism, and historical events are important in understanding present composition. Communities vary in level of environmental determinism and species differ in niche breadth and degree of site specificity. Management implications of vegetation dynamics are discussed.