The purpose of this study is to see if it is possible to determine the course of forest succession over several hundred years by the dissection and close scrutiny of live and dead plant material. The study area was a one-tenth-acre (0.04-ha) square plot in a forest that had never been cut, located near Ashuelot in southwestern New Hampshire. Within it stem-location and growth-rate data were collected from three different classes of stems: 1) living trees, 2), dead unburied stems and stem fragments, and 3), woody remnants buried in the forest floor. The vegetational history of the plot was reconstructed from before 1665 to 1967. During this time, autogenic succession did not contribute significantly to compositional changes, whereas disturbance was an important mediator of such changes. Furthermore, changes in forest structure were often manifestations of species behavior: each tree species had a distinctive stem-distribution pattern in relation to microtopography, mode of germination, and growth. The approach seemed to provide a useful means to obtain evidence about forest change through time that may help prediction and illuminate ecological theory. Limits and modifications of the procedure are discussed.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.Wiley is collaborating with JSTOR to digitize, preserve and extend access to Ecology Abstract. Hypothetical vegetation models were made to simulate numerical changes in species populations along a single environmental gradient. A single ordination procedure was evaluated by its ability to detect the ecological information in the hypothetical models. The procedure was reasonably successful when the data were drawn from a short length of the gradient but became progressively less so as longer lengths of the environmental gradient were included in the data. This parallels an increase in the number of stands from which each species is absent in the total data set. Zero values appear to mask ecological information, and an intuitive method of assigning "degree of absence" values to the data is described. After this adjustment, ordination patterns were easier to interpret because ecological information was concentrated in fewer axes.

Drx, R. L., and J. M. A. SWAN. 1971. The roles of disturbance and succession in upland forest at Candle Lake, Saskatchewan. Can. J. Bot. 49: 657-676. Eighty-nine upland forest stands were selected to cover the ranges of tree species composition, stand ages, understory composition, and site in the area. The role of each tree species as a pioneer, transient, or self-maintaining component of the forest was determined from the number of stems, their vigor, and distribution in tree, sapling, and seedling strata anlong all stands and from growth increment cores of trees and saplings in 39 stands. Conclusions are drawn regarding the species likely to dominate different sites following severe fire and the kinds of vegetational change likely to occur on them between disturbances. Changing patterns in the non-arboreal vascular flora, moss, and lichen cover were related to changes in tree species composition with site and time. The forest and its environment are linked in an irregular "pulse" strategy of alternating disturbance and regrowth that repeatedly rejuvenates the growing stock.

Data from a survey of lowland, mainly peatland, vegetation were subjected to environmental ordination based on measurements of water level and water conductivity, and to vegetational ordination derived from principal component analysis (P.C.A.). Analyzed were the total set of the data ("all types"), half sets ("nonwoody" and "woody types") and quarter sets (stands of "marshes", "meadows", "shrub fens", and "other woody types"); the number of distinct physiognomic groups in a set of data, and presumably the amount of contained heterogeneity, decreased at each segmentation.The effectiveness of the ordination models was tested by correlating measured distances in two-dimensional ordination models with 2W/(A + B) indices of vegetational similarity for randomly selected pairs of types or stands. As the physiognomic complexity decreased, the effectiveness of the P.C.A. vegetational ordination increased whereas that of the environmental ordination decreased. The environmental ordination seemed most appropriate to the data encompassing high complexity (total data set), while the P.C.A. vegetational ordination seemed most appropriate to data with low complexity (quarter sets of the data).