Contribution of L.maackii to deer diet

To address the contribution of L. maackii to deer diet, we estimated the monthly consumption of L. maackii by deer in MUNA, summed this for 12 consecutive months, and divided this by various estimates of the total mass of food consumed by deer in MUNA.

To estimate monthly consumption of L. maackii by deer in each habitat, we estimated the number of browsed L. maackii twigs per unit area, multiplied this by the average leaf mass and average stem mass of a browsed twig, multiplied each of these products by the area of habitat in MUNA, then summed these totals (Eqn. 1

b_i= number of twigs browsed subsequent to the last census on a marked branch in quadrat i; o_i= number of twigs observed on marked branch in quadrat i; t_i= total number of twigs in quadrat i; a= quadrat area (0.25 m²); N₁= number of quadrats in Juniperus forest =100; N₂= number of quadrats in forest/field edge =100; N₃= number of quadrats in deciduous forest interior =100; m₁= average leaf or stem mass of a browsed twig in Juniperus forest (g); m₂= average leaf or stem mass of a browsed twig in forest/field edge (g); m₃= average leaf or stem mass of a browsed twig in deciduous forest interior (g); H₁ = area of Juniperus forest in MUNA (m²); H₂ = area of forest/field edge in MUNA (m²); H₃ = area of deciduous forest interior in MUNA (m²).

). Twigs browsed per unit area was estimated by summing across quadrats the product of the total number of twigs in the 0.3–2.1 m height range x the proportion of twigs on the marked branch that were browsed since the last monthly census; this sum was divided by the number of quadrats (100) x quadrat area (0.25 m²). This quotient was then multiplied by the average leaf or stem mass of a browsed L. maackii twig for that habitat (see below paragraph) to estimate the leaf mass and stem mass of browsed L. maackii per unit area for that month. We then multiplied each of those products by the area of MUNA comprised of that habitat. Those three values were summed to estimate the total browse of L. maackii in MUNA that month.

In order to estimate the leaf and stem masses of browsed portion of twigs in each of the three habitats, we used a method similar to that of Schmitz (1990), who weighed comparable portions from unbrowsed twigs. We predicted the leaf and stem mass of each browsed twig using allometric regressions relating leaf and stem masses of clipped twigs to dimensions that could be measured on browsed twigs. The dimensions measured were the length and diameter remaining at the point the twig was browsed or clipped.

Since long shoots have different morphology than regular twigs, and there is a suggestion they experience higher deer browse (D. Lieurance, pers. comm.), we counted long shoots separately in terms of scoring browse events and in parameterizing allometric regressions. In July 2015, we collected 30 unbrowsed twigs and 20 unbrowsed long shoots in each habitat, sampling from shrubs near each transect. For each of these twigs and long shoots we measured both the remaining length (L) of each twig and the diameter (d) of each twig at the point it was clipped. For each sample, we separated the leaves from woody tissue and dried each at 65 °C for 3 days before weighing. Each of the four sets of dry masses (leaf or stem, twig or long shoot) was regressed on L and d. As twigs approximate cylinders, use of d² gave a better fit than d for each of these multiple regressions. We confirmed that these regression equations (Table 2) accurately predicted the mass of L. maackii twigs and leaves, by regressing observed mass on predicted mass. For each of the four data sets the regression model was a good fit (R²>0.6, Martinod 2016).

Regression equations and statistics to relate leaf and stem mass of clipped L. maackii twigs to dimensions that could be measured on twigs browsed by deer. L is the length remaining (cm) and d is the diameter (mm).

The leaf and stem mass of each browsed L. maackii twig or long shoot was estimated from its L and d as measured the month it was first scored as browsed and the appropriate equation (Table 2). For twigs browsed between December 2015 and March 2016 we only estimated stem mass because L. maackii was leafless.

In April 2016, there were new twigs that expanded, requiring us to distinguish three types of browse: (1) browse on new twigs, (2) browse on old twigs (2015 growth) that showed no new growth, and (3) browse on old twigs that left some new growth. To quantify browse on new twigs (type 1), we counted the new browsed twigs on each marked branch and for each of these measured L and d and then collected an unbrowsed new twig with the same L and d and obtained its dry mass of leaf and stem tissue. The mean dry masses of leaf and stem tissue for each habitat [seeSupporting Information—Table S2] was used these to estimate leaf and stem mass consumed on browsed new twigs.

For browse in April 2016 on old twigs we first scored whether or not any new (2016) stems and leaves remained on the browsed twig. If there was no new growth (type 2) we assumed browse occurred before new twig expansion and therefore estimated only ‘old stem’ mass from measures of L and d and the allometric equation (Table 2).

If there was new growth on a browsed old twig (type 3), we recorded its L and d, then collected an unbrowsed old twig with new growth with the same L and d, separated, dried, and weighed the old stem tissue, new stem tissue, and new leaves. For each habitat we calculated the mean mass of old stem, new stem, and new leaves for these samples [seeSupporting Information—Table S2] and used these in estimating the mass of each tissue consumed.

In order to estimate the total mass of food consumed by deer in MUNA, we used several different estimates of daily consumption by deer from the literature (Table 3) and multiplied by the estimated number of white-tailed deer in MUNA (Eqn. 2).

Estimated daily dry mass intake of white-tailed deer from the literature (c in equation 2) and the corresponding estimate of annual deer diet comprised of L. maackii based on equations 1 and 2. Literature values for intake are based on captive deer in winter to early spring, except for the first line, which is based on wild deer in winter. The range in the diet composition corresponds to mean ±1 SE of estimates of deer density in MUNA from Barrett (unpubl.).

We estimated deer density in MUNA as 14.0 ± 2.3 (SE)/km², based on ten estimates (transects in five areas x two seasons (summer, winter)) for MUNA in 2013 made by Barrett (2014) using pellet-based distance sampling (Urbanek et al. 2012).