Replacement Patterns of Beech and Sugar Maple in Warren Woods, Michigan

Factors responsible for patterns of canopy tree replacement were studied for 22 yr in Warren Woods, Michigan, USA, an old-growth forest codominated by American beech (Fagus grandifolia) and sugar maple (Acer saccharum). Our goal was to distinguish among four hypotheses: autogenic succession, allogenic succession, autogenic coexistence, and allogenic coexistence. We could discern neither successional change toward increasing dominance by sugar maple or beech nor beech self-replacement by root sprouts. In the forest as a whole, from 1933 to 1980, sugar maple remained dominant in small understory size classes, and beech remained dominant in larger understory size classes and in the canopy. We could identify no plausible species-specific canopy influences or consistent responses of understory individuals that could be the basis for autogenic succession or autogenic coexistence. No significant differences existed below canopy beech or below canopy maple in: (1) beech root influence as measured by Epifagus root parasites; (2) light intensity as measured by extension growth of 0.2-0.8 m tall sugar maple; (3) leaf litter as measured by densities of beech and sugar maple leaves just after autumn leaf fall; or (4) inhibition as measured by death of sapling beech and sugar maple. Comparisons under single canopy trees and under monospecific patches showed that understory beech were larger than maple, and this pattern was accentuated below monospecific canopy patches both of maple and beech. This suggested autogenic succession, but coring data were not consistent with this hypothesis, because the largest subcanopy stems of beech and maple were established less often after conspecific or after heterospecific canopy individuals existed above them than before or at the same time. We could not falsify an allogenic coexistence hypothesis that beech or maple advantage changes as light levels fluctuate with frequencies and sizes of treefall gaps. Maples that had recently reached canopy height had been suppressed for an average of only 20 yr, but most had undergone multiple cycles of suppression and release associated with multiple treefall gap events. In contrast beech that had recently reached canopy height had been suppressed an average of 121 yr despite having a similar number of suppression-release cycles as maple. Differences between paired individuals, matched for light microenvironments and height, confirmed our hypothesis that the strong apical dominance of maple led to an advantage in fast upward growth in vertical light gradients of gaps and the long lateral branches of beech led to an advantage in an understory with light flecks and in horizontal light gradients from nearby gaps. Beech's better performance in the understory and maple's better performance in gaps led us to predict that beech would decrease in dominance in the canopy if treefall rate increased. From 1949-1974 treefall gaps averaged 0.16 trees per hectare per year and consisted of mostly single treefalls; we projected that in these gaps beech to maple ratios in the canopy would become 1 to 1. From 1975 to 1994 treefall gaps averaged 1.64 trees per hectare per year and consisted of mostly multiple treefalls; we projected that in these gaps the ratio of beech to maple winners will become 1 to 2.5. We conclude that species-specific differences in response to light level, with allogenic spatial and temporal fluctuation in frequency and area of treefall gaps, are sufficient to explain patterns of beech and sugar maple replacement in the canopy in the old-growth forest at Warren Woods.