Mature Mesophytic Hardwood Forest: Community Transitions, by Layer, from East‐Central Minnesota to Southeastern Michigan

Abstract

Fragments of old—growth forest in which sugar maple (Acer saacharum) is a principal dominant, i.e., late—successional mesophytic forests, were sampled layer by layer across a floristic gradient from east—central Minnesota to southeastern Michigan, USA. This includes the northern Maple—Basswood region ("Big Woods" section) and northwestern Beech—Maple region as described by E. L. Braun (1950). All sampled areas were relatively undisturbed, at least 0.5 ha (surrounded on all sides by >30 m of forest), and occupied similar sites: deep, gently sloping, well—drained drift with high water—holding capacity. The object was (1) to determine whether the community importance of species that are common throughout the study area differs significantly from region to region, perhaps as a result of within—layer competition with species that are restricted floristically, but become important as soon as their ranges are entered (e.g., beech, Fagus grandifolia), and (2) to determine whether regional differences in tree composition correlate with regional differences in the composition of inferior layers. Regional differences in composition of mature trees tended to support Braun's earlier findings that "climax" forests on mesic sites west of the beech border are characterized by the prevalence of basswood (Tilia americana) as a codominant, but as soon as the beech border is crossed, beech is the principal codominant. Basswood ranked at least second in density in 70% of the Big Woods stands (mean density = 53.4 boles/ha), but in only 20% of the stands within the range of beech (mean density = 8.3 boles/ha), where beech ranked at least second in 73% of the stands (mean density = 42.0 boles/ha). Mean basal area of basswood was 14.0 m2/ha west of the beech border and 1.6 to the east, compared with 8.7 for beech. Basswood, in fact, attained about the same relative density (0.20) and basal area (0.36) averaged over all stands in the Big Woods as beech did within its range (i.e., 0.23 and 0.29, respectively). (Mean total density per stand in the two regions was 267 ± 38 and 184 ± 37 boles/ha; mean total basal area was 39.0 ± 5.4 and 30.8 ± 4.6 m2ha.) Regional differences in composition were not readily apparent when inferior layers (vascular plants <2 m tall) were sampled quantitatively, excluding rare species and species with restricted floristic distributions. When understories so defined were clustered according to several measures of species importance, broad overlap of stand groups resulted in each case. Overlap was often similar to that obtained when understories were clustered on the basis of species weightings randomly allocated to stands, that is, when the distribution of scores for each species is not changed, but scores are assigned to stands randomly.