In order to test for differences in forest composition due to deer browsing among vascular groundcover species, two 10 X 10 meter deer exclosures were constructed in 1995 using a randomized selection of grid points within that specific management stand. The exclosures were built using chicken wiring of two grades tightly wrapped around 12 treated wooden posts approximately 10 feet high. The two grades of wiring were meant to exclude the affects of smaller mammals as well as deer, with the smaller grade wiring placed closer to the base of the exclosure. Within each silvicultural management unit, the two exclosures were replicated with three open 10 X 10 meter plots established on the same grid system used for establishing the exclosures. The open plots were located in close proximity to the fenced plots, so that the likelihood of any observed differences between the plots were due to the differences in their location within the silvicultural treatment would be minimized. A total of 25 plots were established, five plots in five different management units.

Within each forest management unit, a weather tower station was constructed in 1994, designed to hold machinery specialized for the collection of miroclimatic data. Data were gathered on the precipitaion, wind speed, minimum and maximum temperatures, snowfall, irradiation levels, and PAR. These data will be used to describe the local climatic patterns influencing deer yarding habits, populations and consequently, effects upon the environment.

At each of the 25 plots, a forest vegetation description was completed. The dominant overstory, understory, and groundcover species were described, as well as the slope, aspect, and position in the landscape relative to lowland forests. A spherical densiometer was used at the center of each plot to obtain an estimate of the amount of light reaching the groundcover. As it is postulated that white-tailed deer increasingly browse in forests with a greater percentage of light hitting the forest floor, the placement of five plots within these five different silvicultural management units is appropriate.

For three years; 1997 - 1999, and three times during the growing season; early spring, summer, and fall, a complete species list was created for each of the five 10 X 10 m plots within each treatment to generate a species richness index, as quantified by a count of species, for paired comparisons. Each 10 X 10 m plot was divided into four quadrants to ease the counting of species. Percent Cover of each species was determined for each plot using a randomly selected 5 X 5 m quadrant of each plot for similar comparisons. As these plots are intended to assess the impacts of deer browsing on northern hardwood forest biodiversity and regeneration, species richness and percent cover will be compared from the open/unfenced plots to deer exclosures between paired treatments, as well as across different silvicultural treatments. The mean species richness and percent cover calculated will establish and reflect a range of baselines for ecological diversity in each treatment as a function of the presence or absence of deer browsing. A single value of species richness will also be determined as sum of the number of unique species sampled in each plot across all seasons and years for ease in comparisons of a site's ecological diversity over a growing season.

To statistically analyze for significant differences (p < 0.05) in the composition of the northern hardwoods forest species due to deer herbivory, measurements of species richness will be compared using a crossed two-factor repeated measures ANOVA. A paired t-test will be used to compare differences in species richness between silvicultural treatments meant to replicate each other (second growth plots and even-aged management, managing for old-growth characteristics and unmanaged old growth, and unmanaged old growth and uneven-aged to management). Percent cover will be similarly compared across different silvicultural treatments using a Kruskal-Wallis nonparametric analysis of variance and the Mann-Whitney U-test statistic.

In addition, the species richness and percent cover of groups of species and specific species will be examined in order to detect differences in percent cover within each plot and across each treatment, season, and year. The species richness of the heavily browsed lilies (Family: Liliaceae) will be examined across each season and year sampled for differences due to silvicultural treatments and the presence or absence of deer browsing. Similarly, the percent cover of sugar maple (Acer saccharaum) will be tested for similar differences (silvicultural treatment, season, and year). As sugar maple is so ubiquitous in these forests, comparisons of the affects of deer browsing upon the percent cover of sugar maple over three years can be used to illustrate the current status of regeneration of timber species.

Similarly, two common ferns, the lady fern (Athyrium filix-fimina) and wood fern (Dryopteris spinulosa) will be compared across silvicultural treatment, seasons and years, to determine detrimental affects of deer browsing. Again, a crossed two-factor ANOVA and Tukey's Test for Differences will be used to detect significant differences between the species richness indices due to the browsing pressures across the five silvicultural treatments, seasons sampled and year. Percent cover will be analyzed over the same seasons and years with a Kruskal-Wallis analysis of variance, and a Mann-Whitney U test to rank the percent cover.

Addressing the morphological indications of deer browsing, certain understory species will be examined for significant differences across treatments and seasons. Differences in the mean height, leaf area, number of leaves, flower/fruiting frequency, and seasonal shoot growth of a particular species can be compared between particular species found within the exclosures and open, unfenced plots located near the exclosure. A paired t-test will be used to examine community structure of the exclosures and paired open plots between the paired treatments (second growth plots and even-aged management, managing for old-growth characteristics and unmanaged old growth, and unmanaged old-growth and uneven-aged to management). By using a paired plot which is close in proximity to the exclosure, the differences in morphological features then result from browsing pressures, rather than other local environmental conditions. While micro-climatic differences might cause small-scale differences in these measurements of leaf area and particularly mean height, deer browsing more dramatically alters the net differences in these morphological measurements, and can be detected through these statistical analyses. By pairing the plots, the greatest number of outside factors can be controlled. These paired t-tests allow for the detection of statistically significant differences in morphological measurements resulting from the effects of browsing. This study aims to follow up on the work of Marquis (1981) in the Hardwood forests of Pennsylvania, and give further validity to the potential of deer browsing to negatively affect northern hardwoods regeneration of certain silvicultural treatments.