Abstract
Multivariate techniques were utilized to examine even-aged jack pine stands on upland sandy sites at the southern edge of the boreal forest near Elk Lake, Ontario. Cluster analysis of 180 stands led to the recognition of 10 vegetation types, each showing a unique combination of floristics, physiognomy, and environmental components. Classification of common species led to the recognition of five ecological groupings, which show varying degrees of association with the vegetation types. Nonmetric multidimensional scaling of the stands suggested a vegetational continuum in response to overall moisture availability. A corresponding ordination of common species indicated the development of interspecific associations related to soil moisture conditions. It is suggested that the vegetational composition of upland jack pine forests is determined by both probabilistic and deterministic effects, and this is discussed in the context of vegetation structure and dynamics.