Hemlock Stands from Wisconsin to Nova Scotia: Transitions in Understory Composition Along a Floristic Gradient

The understories of 71 small stands strongly dominated by hemlock (Tsuga canadensis [L.] Carr.) were sampled in a transect from northern Wisconsin to Nova Scotia. Stands selected for sampling were at least 100 yr old and had homogeneous overstories composed of at least 75% hemlock. Stand area varied from 0.04 to 2.4 ha, with most stands between 0.2 and 0.8 ha. All stands were local areas of strong hemlock dominance within a much larger mixed forest. The hemlock understories, as a group, are not a distinctive community type with respect to composition (vascular plants). There are no species relatively abundant in the hemlock stands that are not also abundant in other types of mesophytic forest. Presence and coverage of species vary greatly from stand to stand, even on similar soils and topography, suggesting that community organization is haphazard. However, a few species are almost ubiquitous (Tsuga canadensis, Acer rubrum, Maianthemum canadense, Dryopteris austriaca), and others consistently attain high coverage when present (e.g., Taxus canadensis, Oxalis montana). Also, the mean association (Cole's index) between pairs of common species (50—69% presence) significantly exceeds the mean association in stands not strongly dominated by a single overstory species (i.e., mixed hardwood forests with varying proportions of sugar maple and basswood in the overstory). Half of the 66 understory species that are at least moderately common in non—hemlock forest throughout the study area are rare or absent under hemlock in some parts of the transect, yet common in other parts. Regional differences in composition, however, are not discrete. Regional differences in understory composition are poorly correlated with regional differences in the potential hemlock forest flora (i.e., plants observed growing in hemlock forests). Community transitions also correspond poorly to certain climatic differences between regions, such as the length of the freeze—free period.