Historically, there have been two contrastng views of succession: Gleasonian and Clementsian. The Clementsian model argues for a predictable, orderly process which culminates in a stable climatic climax. The Gleasonian model is more complex, invoking interactions between the physical environment, population-level interactions between species and disturbance regimes in determining the composition and spatial distribution of species.
The obvious extension of the Gleasonian view is the view that continuous change in vegetation is the norm (similar to Cowles’ view of succession) and that multiple steady states exist in ecosystem dynamics. This challenges the popular concept of reference (or 'original') conditions in most ecosystems.
The development of some ecosystem attributes, such as pedogenesis and nutrient cycles, are critical factors which influence community development but take considerable time. Coupled with the stochastic nature of disturbances events and other long-term (e.g., climatic) changes, it is doubtful whether a true 'climax' exists in any ecosystem, except as a purely theoretical construct.
At the turn of the 20th century, Henry Chandler Cowles was one of the prime movers in the emerging study of "dynamic ecology", through his study of the Indiana Dunes, sand dunes at the southern end of Lake Michigan. Cowles found that he could relate the vegetation at any point in the dunes to several variables: the distance from the lake shore, the estimated age of the dune, and the type of soil that had developed. In 1899 his classic paper "The ecological relations of the vegetation of the sand dunes of Lake Michigan" appeared in the Botanical Gazette.
In modern times, less stress has been placed on the idea of a single, natural "climax forest" or other climax vegetation, and more study has gone into the roles of contingency in the actual development of several possibilities at any one site.