Spatial Dynamics of a Primary Succession: Nucleation

Abstract

Changes in spatial pattern during the course of succession have attracted considerable interest in recent years. Such changes are of interest not only for their own sake, but because they may throw light on mechanisms of successional change. Greig-Smith (1964) summarized the currently accepted hypothesis of successional variations in pattern. He suggested that the variations are determined by the establishment and growth of clumps (often clones) of persistent species. As these clumps enlarge, and eventually coalesce, the scale of pattern will increase, not necessarily smoothly, to a maximum at climax; on the other hand, the intensity of pattern will decline as the range of scale of contagion increases with age. Work by Whitford (1949), Greig-Smith (1952), Margalef (1958), Kershaw (1958), Pielou (1966) and Brereton (1971) supports this view. Williams et al. (1969) have examined the question of whether succession is a gradual or stagewise process, i.e. does the rate of change of composition of the vegetation fluctuate? In a secondary succession in a Queensland rain forest they found a change from predominantly temporal to predominantly spatial variation after 6 years. If successional stages exist they are likely to influence variations in spatial pattern: patterns may differ from stage to stage but the pattern at times of transition may also differ from that within stages. Shafi & Yarranton (1973) have demonstrated stages of differing heterogeneity during a post-fire succession in the boreal forest. Their discovery that a stage may be succeeded by another of greater homogeneity implies that the scale of pattern may decrease during at least part of a succession. Anderson (1967) has also noted a decrease in the scale of pattern early in a succession. Morrison & Yarranton (1973) showed that heterogeneity is greatest during a sand dune succession at the transition between stages. Webb, Tracey & Williams (1972) raise the question of whether patches of vegetation occurring in a succession represent different stages of maturity; both Morrison & Yarranton and Shafi & Yarranton (1973) suggest that transitions from stage to stage take place heterogeneously on the ground so that this is indeed the case. If this is admitted as at least a possibility, it establishes the connection between the mechanism of succession and variations in spatial pattern. On the basis of this reasoning it was decided that detailed examination of the pattern of vegetation on the ground before, during and after a stage-to-stage transition in a succession might not only clarify the sequence of spatial pattern but also help to elucidate the mechanism of succession itself. The sand dune succession at Grand Bend, Ontario, previously described by Morrison (1973), was selected for such a study.