Content uploaded by Darroch Whitaker
Author content
All content in this area was uploaded by Darroch Whitaker on Apr 27, 2015
Content may be subject to copyright.
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors,
nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of
maximizing access to critical research.
NIGHT ROOST SELECTION DURING WINTER
BY RUFFED GROUSE IN THE CENTRAL
APPALACHIANS
Author(s): Darroch M. Whitaker and Dean F. Stauffer
Source: Southeastern Naturalist, 2(3):377-392.
Published By: Eagle Hill Institute
DOI: http://
dx.doi.org/10.1656/1528-7092(2003)002[0377:NRSDWB]2.0.CO;2
URL: http://www.bioone.org/doi/
full/10.1656/1528-7092%282003%29002%5B0377%3ANRSDWB
%5D2.0.CO%3B2
BioOne (www.bioone.org) is a nonprofit, online aggregation of core
research in the biological, ecological, and environmental sciences. BioOne
provides a sustainable online platform for over 170 journals and books
published by nonprofit societies, associations, museums, institutions, and
presses.
Your use of this PDF, the BioOne Web site, and all posted and associated
content indicates your acceptance of BioOne’s Terms of Use, available at
www.bioone.org/page/terms_of_use.
Usage of BioOne content is strictly limited to personal, educational, and
non-commercial use. Commercial inquiries or rights and permissions
requests should be directed to the individual publisher as copyright holder.
SOUTHEASTERN NATURALIST
2003 2(3):377–392
NIGHT ROOST SELECTION DURING WINTER
BY RUFFED GROUSE
IN THE CENTRAL APPALACHIANS
DARROCH M. WHITAKER1,*, AND DEAN F. STAUFFER1
ABSTRACT - In northern regions, ruffed grouse (Bonasa umbellus) conserve
considerable energy during winter by burrowing under snow cover to roost.
When conditions are unsuitable for snow burrowing grouse almost invariably
roost in conifers. We studied selection of winter night roosts by ruffed grouse
in western Virginia, a region where snow accumulations are rare and transient.
Grouse almost always used ground roosts when snow was present even though
snow was never deep enough for snow burrowing. When snow was absent
grouse did not show any clear preference in roost microsite type, and were
found roosting in and under deciduous and evergreen trees and shrubs, in
brush piles, and in leaf litter. We hypothesize that this ambivalence to conifers
was due in part to persistent accumulations of fallen oak leaves, which likely
afford grouse good thermal cover and concealment. Grouse were frequently
found at low elevations during daytime, but rarely roosted in bottoms. This
suggests daily elevational movements, possibly to avoid cold air settling in
low-lying areas during night.
INTRODUCTION
Roost site selection is a fundamental aspect of the behavior of
free-living birds, having important consequences for energy budgets
and predator avoidance (Bergerud and Gratson 1988, Walsberg
1983). This is particularly true for ruffed grouse (Bonasa umbellus
L.) wintering in temperate and boreal forests, where weather can be
severe, predator pressure can be high, and individuals spend 19–23
hours roosting each day (Gullion and Svoboda 1972, Ott 1990).
Through most of the species’ range, wintering ruffed grouse typically
burrow under snow cover to roost. Snow burrowing reduces radiative
heat loss, provides shelter from wind, and traps air heated by thermal
radiation next to the bird. This affords a thermoneutral environment,
which can reduce metabolic heat production by 30 % or more when
compared to open roost sites; dry powder snow likely represents opti-
mal thermal cover for the species during winter (Thompson and
Fritzell 1988a). Further, because the exact location of grouse beneath
the snow surface is not evident, captures of snow-burrowing grouse
by predators are rare (Bergerud and Gratson 1988, Marjakangas
1Appalachian Cooperative Grouse Research Project (ACGRP), Department of
Fisheries and Wildlife Sciences, Virginia Tech, Blacksburg, VA 24061.
*Corresponding author - dwhitake@vt.edu.
Southeastern Naturalist Vol. 2, No. 3378
1990). Predator avoidance may also be enhanced through decreased
foraging and travel time, as snow burrows confer energetic savings
and are typically available in close proximity to food sources (Doerr
et al. 1974, Ott 1990).
In contrast to more northerly regions, conditions suitable for snow
burrowing by grouse are rare and transient in the central and southern
Appalachian Mountains. Consequently, ruffed grouse in the Appala-
chians typically must use alternate roost sites that afford reduced ener-
getic benefits (Ott 1990, Thompson and Fritzell 1988a). Several studies
have demonstrated that under such circumstances ruffed grouse and
other tetraonids almost invariably roost in the thickest conifer cover
available (Bump et al. 1947, Ott 1990, Pietz and Tester 1982, Pekins et
al. 1991, Swenson and Olson 1991, Thompson and Fritzell 1988a,
Woehr 1974). However, in the central Appalachians conifers are often
rare and dispersed and most are pines (Pinus spp.) that have relatively
sparse foliage, so conifers may be of less value as roosting cover. Also,
during many winters Appalachian grouse endure a prolonged period of
nutritional stress and energy deficit (Hewitt and Kirkpatrick 1997,
Norman and Kirkpatrick 1984, Thomas et al. 1975). This deficit is
closely tied to poor hard mast crops during fall and can lead to overwin-
ter reductions in total carcass fat of approximately 60 % in male and 75
% in female grouse (Norman and Kirkpatrick 1984). Although this
deficit does not lead directly to grouse mortality due to exposure or
starvation, it has important population consequences by reducing repro-
ductive success during the following spring (Norman and Kirkpatrick
1984, Servello and Kirkpatrick 1987; P.K. Devers, Virginia Tech, un-
published data). Consequently, strategies Appalachian grouse employ to
reduce energy expenditure in the absence of preferred roosting cover are
of interest to wildlife managers.
Given the energetic stress Appalachian ruffed grouse often endure
and the apparent rarity of preferred roosting cover, our objective was to
describe night-roosting behavior of ruffed grouse in the central Appala-
chians during winter. Specifically, we wanted to describe the general
types of roost microsite selected, habitat structure around roosts, the
topographic placement of roosts, and the influence of environmental
conditions on roost site selection.
METHODS
Data were collected from January 1998 through March 2002 at
three study sites located in western Virginia (2,000–10,000 ha/site).
The northeastern (VA1; Augusta Co.) and southwestern (VA3;
Smyth Co.) study sites were separated by ≈ 270 km, while the third
study site (VA2; Botetourt Co.) was located between these two, ≈ 80
D.M. Whitaker and D.F. Stauffer2003 379
km from VA1. Topography on all three sites is complex, with relief
exceeding 500 m and abundant mountains, ridges, and hollows. Oak-
hickory (Quercus spp., Carya spp., respectively) and oak-pine for-
ests were the dominant cover on all study sites. A number of ever-
green species were dispersed throughout or locally common on these
sites. These include white (Pinus strobus L.), pitch (P. rigida
Miller), Virginia (P. virginiana Miller) and Table Mountain pine (P.
pungens Lambert), eastern hemlock (Tsuga canadensis L.), and east-
ern red cedar (Juniperus virginiana L.). Mountain laurel (Kalmia
latifolia L.), great rhododendron (Rhododendron maximum L.), and
Catawba rhododendron (R. catawbiense Michx.) were common in
the forest understory.
We trapped grouse from September through mid November each
year on each site. Sex and age (juvenile or adult) of captured birds were
determined from feather criteria (Kalla and Dimmick 1995). Grouse
were then equipped with necklace-style radio transmitters (10 g,
148.000–152.000 MHz; Advanced Telemetry Systems, Isanti, MN) and
released at the capture site.
Like Thompson and Fritzell (1988a), we located most roosts by
flushing radioed grouse at dawn during winter (0630–0700 h; January
7–March 23). Additional roost sites were located opportunistically,
primarily during nightlighting recapture attempts (Huempfner et al.
1975). We tried to avoid flushing an individual grouse on more than
one occasion. Once a roost site was located we recorded the number of
grouse present and categorized the roost as one of the following
microsite types: conifer tree, conifer ground (i.e., on the ground under
a conifer), evergreen shrub, deciduous tree, deciduous ground, decidu-
ous shrub, open ground, snow burrow, or “other” (with a written de-
scription). A ground roost was determined to be associated with the
nearest tree or shrub whose foliage dripline encompassed it. Air tem-
perature at dusk (1700 h) was obtained from data loggers recording
hourly temperature readings at a central location on each study site
(Hobo® XT, Onset Computer Corp., Pocasset, MA). We used this tem-
perature as it reflected conditions during the period when roost sites
were being selected. However note that overnight low temperatures
usually were considerably colder and occurred near dawn (e.g., Fig. 1).
Overnight precipitation (none, rain, or snow), snow depth, and snow
cover (none, continuous, or patchy) also were recorded. Species,
height, Diameter at Breast Height (DBH), and dominance class (domi-
nant, subdominant, or understory) were recorded for each roost tree.
When we observed a grouse in a tree roost we recorded the roost
height. However tree roosting grouse were often difficult to locate
before they flushed, in which case we simply recorded the bird as
roosting “above ground.” Finally, slope aspect (North, South, East, or
Southeastern Naturalist Vol. 2, No. 3380
West) and slope position were recorded, with slope position classes
being defined as the upper, middle, or lower one-third of the slope
(ridge, midslope, or toe/bottom, respectively).
A pair of 400 m2 (11.3 m radius) habitat plots was sampled for each
roost site (Noon 1981). The first plot was centered on the roost, while
the second was placed 60 m away in a random direction, but within the
same forest stand. At each plot we recorded tree basal area (estimated
using an optical prism) and counts of all trees within the plot classified
by species and size class (8–20, 20–40, 40–60, and > 60 cm DBH). We
established strip transects along the east-west and north-south axes of
each plot (22.6 m each). To estimate small stem density, we counted all
stems < 8 cm DBH and > 1.3 m tall within 1 m of each transect. Care was
taken to not double count stems in the area where transects overlapped
at the plot center (area sampled = 86.4 m2). We recorded the presence or
absence of deciduous and coniferous tree canopy cover at 10 evenly
spaced points along each transect. This yielded 20 observations/plot,
which we multiplied by five to obtain estimates of percent overhead
cover. Finally, we recorded the number of stems at ground level within 3
m of the roost site (alive or dead, all diameters).
Typically each radio-equipped grouse on a study site was located
twice weekly. Sets of 3–8 azimuths were collected in < 20 minutes
Figure 1. Nighttime air temperatures (°C; 1700 h–0700 h) along ridgetops and
bottoms on the VA1 study site (Augusta Co., VA) from January 26–February 1,
2002, the first week during which such data were collected. In total, thermal
inversions developed on 15 of 36 nights sampled.
D.M. Whitaker and D.F. Stauffer2003 381
by a roving technician using a using a handheld Yagi antenna from a
network of fixed telemetry stations (White and Garrott 1990).
Grouse locations were then estimated for each set of azimuths using
Lenth’s maximum likelihood estimator and specifying a bearing er-
ror of 7û for confidence ellipse calculations (Lenth 1981). These ra-
diotelemetry locations were used to estimate the daytime distribution
of grouse relative to slope position. For each study site we created a
combined dataset comprising locations from the 25 grouse having
the most extensive tracking histories (81–303 locations/grouse, mean
= 145 locations; 16 juvenile females, 19 adult females, 21 juvenile
males, 19 adult males). From this set we retained locations collected
between 0900–1600 h from January 1–March 20 each year, having
maximum likelihood error ellipses < 1.5 ha, and a Geometric Mean
Distance (GMD) between receiving stations and the location esti-
mate not exceeding 500 m (n = 882). Previously we conducted error
assessments at all three study sites using beacon transmitters having
known locations (White and Garrott 1990), which indicated that
mean linear error should be ≈ 45 m for locations taken from 500 m
and having error ellipses of 1.5 ha. However, as this error estimate
is for the least reliable locations used here, overall mean location
error would be considerably less than 45 m. From this dataset we
randomly selected 150 locations per site and overlaid them on
United States Geological Survey Digital Elevation Models (DEMs)
of the study areas. Slope position for each daytime location was
assessed using the same classification scheme used for roost sites in
the field.
To describe the effect of slope position on nighttime air temperature,
we placed temperature loggers on the ridge, middle, and toe of various
slopes on the Augusta Co. study site during winter 2002 (Hobo® Temp,
Onset Computer Corp., Pocasset, MA). Loggers were left in place for 3–
4 nights and then moved to a new slope. We sampled 10 slopes, with
slope relief ranging from ≈ 25–60 m. The position of individual data
loggers was randomized on each slope and following data collection all
three loggers were placed in the same location for 24 h to verify that
they were recording temperature comparably.
We used DEMs to make a digital slope aspect map for each
study site, and from this calculated the expected number of roosts
by slope aspect class. For each site we created a Minimum Convex
Polygon (MCP; White and Garrott 1990) containing all grouse
radiotracking locations, and then multiplied the proportion of the
MCP having a particular slope aspect by the number of roosts ob-
served on that site. We summed values for each aspect class across
the three study sites to obtain an estimate of the expected number
of roosts by slope aspect.
Southeastern Naturalist Vol. 2, No. 3382
Statistical analyses of categorical data were performed using Chi-
squared tests (Sokal and Rohlf 1995). Paired t-tests were used to com-
pare habitat measurements between roost site and random habitat plots
(Sokal and Rohlf 1995). A significance level of a = 0.05 was used for all
statistical tests. Analyses were carried out using Minitab statistical
software (Version 11.21, Minitab Inc., State College, PA).
RESULTS
From 1998–2002 we located 90 roost sites used by 72 radio-equipped
and 12 unmarked grouse on the three study sites (46 At VA1, 26 at VA2,
and 18 at VA3). A broad range of roost microsites was used with no one
type dominating the sample (Table 1). The most commonly used
microsite was evergreen shrubs, but this only accounted for about one-
third of all roosts. Roosts were almost evenly divided between evergreen
and non-evergreen microsites (Table 1). Evergreen shrub roosts were
fairly evenly divided between rhododendron (n = 16) and mountain laurel
(n = 15), while white pine was used for 14 of 18 conifer tree roosts.
Table 1. Winter roost microsites used by ruffed grouse at three study sites in western
Virginia (n = 90, 1998–2002).
Roost type Count %
Evergreen Shrub (ground) 13 14
Evergreen Shrub (above ground) 18 20
Conifer Tree (ground) 5 6
Conifer Tree (above ground) 13 14
Evergreen total 49 54
Deciduous Shrub (ground) 5 6
Deciduous Shrub (above ground) 0 0
Deciduous Tree (ground) 19 21
Deciduous Tree (above ground) 10 11
Deciduous total 34 38
Snow Burrow* 0 0
Open Ground 7 8
* Grouse were regularly observed ground roosting in snow (see Table 4), but snow depth
was never sufficient to allow birds to completely bury themselves.
Table 2. Habitat measures from paired 400-m2 sampling plots centered on roosts and
control sites (i.e., 60 m away in a random direction) in western Virginia, 1998–2002.
Differences were tested using a paired t-test (n = 44).
Feature Roost ± S.E. Control ± S.E. tP
% Deciduous cover 84.66 ± 2.18 84.55 ± 2.92 0.03 0.970
% Conifer cover 23.48 ± 4.50 16.93 ± 3.60 1.54 0.130
Conifers (8-20 cm)/ha 85.25 ± 19.75 53.00 ± 20.75 1.28 0.210
All conifers/ha 114.75 ± 25.00 77.75 ± 26.00 1.17 0.250
Basal area (m2/ha) 25.41 ± 2.15 28.18 ± 3.27 0.78 0.440
Stems < 8 cm/ha 7453 ± 793 5436 ± 656 2.95 0.005
Stems within 3 m 51.63 ± 5.79 33.62 ± 4.74 2.77 0.009
D.M. Whitaker and D.F. Stauffer2003 383
Deciduous roosts (n = 34) were associated with nine tree and three shrub
species, with no one species being used for more than four roosts. Seven
roosts were situated away from any immediate woody vegetation. Grouse
typically deposit a large number of fecal pellets during the night, and it
was evident that individual roost sites were not reused.
Habitat was sampled at 44 pairs of roost site and control plots.
Proportions of roosts for which habitat was sampled were similar to the
overall distribution of roosts across study sites (VA1 = 26, VA2 = 11,
VA3= 7) and between ground and above ground roosts (ground = 27,
above ground = 17). Most habitat variables did not differ between roosts
and controls (Table 2). There was a trend towards higher numbers of
conifers on roost plots than control plots, however this was not statisti-
cally significant. Density of small stems and number of stems within 3
m were greater around roost sites than on control plots (Table 2).
However only eight of the roosts sampled were located in sapling stands
(trees < 12.5 cm DBH), while 29 were in pole stands (12.5–28 cm
DBH), and 7 were in sawtimber stands (> 28 cm DBH).
Forty five percent of all roosts were classified as above ground
(Table 1). Of these above ground roosts, 16 were in understory shrubs or
trees, nine in subdominant (midstory) trees, and six in dominant
(canopy) trees. Roost trees ranged in DBH from 2.5–39.0 cm. Of 81
observations where roost height could be determined accurately, 49
were on the ground, 26 were < 5 m above ground, four were 5-10 m
above ground, and two were > 10m above ground. The highest roosting
grouse we located was in the forest canopy 17 m above ground. Because
of the ubiquity and persistence of oak leaf litter on the floor of oak forest
types in the region, virtually all grounds roosts were associated with this
substrate. Ground roosting grouse generally formed a nest-like depres-
sion in this leaf litter and, as evidenced by accumulations of droppings,
rarely relocated during the night.
Table 3. Slope position of daytime and nighttime (i.e., roosting) grouse locations at three
study sites in western Virginia, 1998–2002. Divisions between midslope and toe/bottom,
and between ridge and midslope were one-third and two-thirds of the way up slopes,
respectively.
Number of locations
Site Toe/Bottom Midslope Ridge χ2
2 P
VA1 Night 4 18 24 30.30 < 0.001
Day 63 63 24
VA2 Night 0 19 7 9.15 0.010
Day 33 69 48
VA3 Night 4 6 7 0.21 0.900
Day 43 51 56
Total Night 8 43 38 20.85 < 0.001
Day 139 183 128
Southeastern Naturalist Vol. 2, No. 3384
Local topography influenced roost site selection by grouse. Daytime
radiotelemetry locations of grouse were fairly evenly distributed across
slope position categories. In contrast, on all three study sites night-
roosting grouse were more common on ridges and midslopes than in
bottoms, with significant differences between the distribution of diurnal
and nocturnal observations being detected on two sites (Table 3). Night-
time air temperature also was affected by slope position, with thermal
inversions developing as cold air pooled at lower elevations on 15 of 36
nights sampled (Fig. 1). The observed distribution of roosts across slope
aspects did not differ from proportional availability (χ2
3 = 1.24, P =
0.745; Fig. 2).
Figure 2. Expected and observed distributions of grouse roost sites by slope
aspect during winter in western Virginia, 1998–2002.
Table 4. Influence of snow cover on incidence of ground roosting by grouse in western
Virginia, 1998–2002.
Ground Above ground χ2
1P
No snow on ground 29 30 6.88 0.009
Snow on ground 20 5
Table 5. Influence of overnight precipitation on incidence of ground roosting in western
Virginia, 1998–2002.
Ground Above ground χ2
2P
No precipitation 37 31 6.65 0.036
Rain during night 3 6
Snowfall during night 9 1
D.M. Whitaker and D.F. Stauffer2003 385
Weather-related variables influenced roost site selection by grouse.
Although snow depths were never sufficient to allow snow burrowing (≈
20 cm; Thompson and Fritzell 1988a), ground roosting increased when
any snow cover was present (Table 4). Precipitation during the night
also influenced the incidence of ground roosting, with the smallest
proportion of ground roosts observed during nights having rainfall, and
the greatest proportion observed during nights having snowfall (Table
5). Grouse shifted from predominantly using above ground roosts at
temperatures above freezing to using ground roosts when temperatures
fell below 0 °C (χ2
1 = 7.35, P = 0.007; Fig. 3). However this relationship
may have resulted from the association of snow cover with cold
weather, and no difference was detected when observations were re-
stricted to nights having no snow cover (χ2
1 = 1.40, P = 0.237). Air
temperature did not influence the proportion of grouse roosting in ever-
green vegetation (Fig. 4).
At five roosts a second grouse was flushed within a few meters of the
radioed bird, while at four roosts we observed three grouse in close
proximity to one another (nroosts = 90). Thus, 21% of the 103 grouse we
observed in roosts were in groups. Grouse populations on our study sites
have slightly male-biased sex ratios (Reynolds et al., 2000), but of seven
radioed (i.e., known-sex) grouse in these groups, six were females
(small sample size precludes statistical testing).
Figure 3. Proportion of ground and above ground night roosts selected by grouse
relative to air temperature (°C) at dusk (1700 h) during winter in western
Virginia, 1998–2002.
Southeastern Naturalist Vol. 2, No. 3386
DISCUSSION
Our observations suggest the roosting behavior of ruffed grouse in
western Virginia differed from that observed elsewhere. Roost site
selection is predictable in other regions, with grouse almost invariably
selecting snow burrows when snow cover is suitable, and dense conifers
otherwise (Bump et al. 1947, Ott 1990, Pietz and Tester 1982, Pekins et
al. 1991, Swenson and Olson 1991, Thompson and Fritzell 1988a,
Woehr 1974). Although conditions were never suitable for proper snow
burrowing during our study, grouse in western Virginia did exhibit the
expected affinity for ground roosting in snow. In contrast to findings
from other regions, in the absence of snow we observed grouse using an
array of alternate roost microsites, with no one type being clearly
preferred. Particularly unexpected was that roosts were approximately
equally divided between evergreen and deciduous or open roost
microsites, and were often situated in areas having little or no conifer
cover (Tables 1 and 2). Indeed on several occasions grouse were found
roosting on the forest floor away from any immediate vegetative cover.
However this is not to say that no selection was occurring, as a number
of patterns were observed.
Roost Microsite Selection
There are several potential explanations for the intermediate use
of evergreen roosts we observed. An obvious reason would be that
Figure 4. Vegetation types selected for nocturnal roosting relative to air tem-
perature (°C) at dusk (1700 h) during winter in western Virginia, 1998–2002.
D.M. Whitaker and D.F. Stauffer2003 387
evergreen vegetation was rare and thus often unavailable as a roost-
ing substrate. However this is seldom the case in western Virginia,
where evergreen trees or shrubs are present or common through most
forest stands (Table 2; Fearer 1999). Another possibility is that
weather conditions were typically mild enough that grouse did not
incur a large thermodynamic cost by roosting in more open
microsites. However the lower critical temperature of ruffed grouse
is approximately 1.5 °C (Thompson and Fritzell 1988b), and we
found no evidence that grouse preferred evergreen roosts at lower
temperatures (Fig. 4). Other possible explanations which we find
more plausible include that the evergreen species common on our
study sites afford inferior roosting conditions, that non-evergreen
microsites in the region offer relatively high quality roosting habitat,
and that non-thermodynamic factors (e.g., predator avoidance) are
playing an important role in roost site selection. These explanations
are not mutually exclusive.
The sparse evergreen foliage of pines, mountain laurel, and rhodo-
dendron may afford thermal benefits intermediate between deciduous
trees and the thicker conifers used by grouse in other regions (i.e.,
spruces [Picea spp.], firs [Abies spp.] and cedars [Juniperus spp. and
Thuja spp.]; see Ott 1990, Thompson and Fritzell 1988a). Ott (1990)
reported that ruffed grouse thermostatic energy demands were similar
for hemlock and Norway spruce. However, hemlock may have received
little use on our sites because of its association with low-lying areas in
the region (see “Topographic Effects” below).
Ground roosts in oak-hickory forests may be superior to ground
roosts in other forest types. After leaf drop, oak leaves are much
more resistant to decay and saturation than those of other deciduous
tree species, particularly aspen (Populus spp.) and birch (Betula spp.)
typical of more northern forests in the core of the ruffed grouse’s
range. Consequently, oak leaves remain relatively dry and unmatted
(lofted) on the forest floor through the year, often accumulating to
depths exceeding 30 cm. We typically found that ground-roosting
grouse had created a nest-like depression in oak leaf litter; in one
case a grouse had completely buried itself in oak leaves. Such “leaf-
roosting” or “leaf-burrowing” would presumably afford thermal ben-
efits by reducing heat loss through convection and radiation. Other
observations also suggest that fallen leaves afford favorable roost
microsites. Ground roosting was frequent on nights when tempera-
tures fell below freezing and least common on nights having rainfall,
when leaves were wet (Table 5, Fig. 3). Because snow cover was
always shallow, ground roosts in snow typically extended down into
the leaf litter. Snow cover is associated with low temperatures, and it
Southeastern Naturalist Vol. 2, No. 3388
may be that the combination of these factors contributes to a general
superiority of ground roosts during cold weather. Finally, grouse in
the region are red-phased, so are inconspicuous to predators when
buried flush with the similarly colored leaf litter. Given the behav-
ioral parallel to snow burrowing, the development of such leaf-roost-
ing behavior seems reasonable.
Ruffed grouse were found to roost in stands having relatively high
stem densities (Table 2; see also Thompson and Fritzell 1988a). This
affinity for areas of high stem density is interesting, though not unex-
pected. Numerous studies have reported stem density as being posi-
tively correlated with ruffed grouse habitat quality due to reduced pre-
dation risk, and this feature is widely viewed as the cornerstone of
ruffed grouse habitat needs. Though many roosts were found in mature
stands having relatively low stem densities, such roosts were generally
located in microsites having locally high densities of small stems, such
as evergreen shrub thickets and areas around windthrown trees. Stem
density is likely important even when grouse roost in trees and shrubs,
as most such roosts were situated within a few meters of the ground and
grouse often walk to and from roost sites.
Topographic Effects
One of the most striking aspects of roost site selection we ob-
served was the different distribution of day and nighttime grouse
locations across slope positions (Table 3), which suggest daily
movements downslope in morning and upslope in evening. As birds
are simply moving out of local hollows, such movements typically
would involve relatively small elevation changes and short travel
distances. A study of diurnal habitat selection by grouse on the VA3
site found that low-lying mesic sites were preferred, likely due to
the higher abundance of foods such as forbs and soft mast (Fearer
1999). However preference for higher elevation roost sites remains
unexplained. We considered the possibility that this might reflect
selection for some microsite type restricted to higher elevations.
However, given that grouse did not show any strong preference in
roost microsite type (Table 1), this seems unlikely. Indeed all roost
microsite types used were commonly available at all elevations on
our study sites.
A similar pattern of daily movement was observed in Idaho by
Hungerford (1951), who reported that during August and September
ruffed grouse broods moved upslope each evening to avoid thermal
inversions. Thermal inversions develop in convoluted terrain when
heavier cold air drains into low-lying areas during calm, clear, dry
nights (Geiger 1950) and are common during winter in western Virginia
D.M. Whitaker and D.F. Stauffer2003 389
(e.g., Fig. 1). The maximum temperature inversion we observed was
8.8 °C. On ridgetops, 57 % of hourly nighttime temperature readings
were < 1.5 °C (the lower critical temperature for ruffed grouse; Thomp-
son and Fritzell 1988b), while in bottoms 67 % were < 1.5 °C. Conse-
quently, it seems that in mountainous areas ruffed grouse may make
regular short-distance movements upslope during evening to avoid ener-
getically costly microclimates developing in low-lying areas. At times
wind chill may be greater at higher elevations due to increased expo-
sure. However, as mixing by wind prevents the formation of thermal
inversions (Geiger 1950), wind chill should be negligible on nights
when inversions develop. Though we did not measure it directly, it is
likely that grouse also avoid wind chill through selection of sheltered
microsites (e.g., Thompson and Fritzell 1988a).
Bump et al. (1947) reported that, though all slope aspects were used
regularly, grouse in New York were more often found on west-facing
slopes. Though we also located the greatest proportion of birds on west-
facing slopes, the distribution we observed did not differ from the
availability of slope aspects across study sites (Fig. 2).
Snow Roosting
In more northern forests at the core of the species’ range, grouse
numbers and production may be positively correlated with snow
roosting conditions during the preceding winter (Gullion 1970,
Kubisiak et al. 1980). In southern portions of the species’ range
ruffed grouse also snow burrow whenever accumulations of snow are
sufficient (≈ 20 cm; Ott 1990, Thompson and Fritzell 1988a). Given
that individual grouse in the South may never experience conditions
sufficient for proper snow burrowing, it is noteworthy that they ex-
hibit a strong predisposition for this behavior. Even when snow cover
was limited grouse buried themselves with their backs flush with the
snow surface (see also Ott 1990). On one occasion we were able to
backtrack from a grouse’s ground roost to a tree roost it had aban-
doned soon after a heavy snowfall had begun near midnight. These
observations suggest snow roosting is a strongly and broadly devel-
oped behavior of ruffed grouse.
Management Implications
Natural resource managers in the Appalachian region often devote
considerable energy and resources to improving habitat for ruffed
grouse. These efforts often include planting or favoring conifers, par-
ticularly white pine, both for their timber value and perceived value as
roosting cover for grouse and other upland game birds. While we often
found grouse roosting in understory white pines, our findings do not
Southeastern Naturalist Vol. 2, No. 3390
suggest a strong selection for roosting in pine forest types in the
central Appalachians, and we question the benefits of this practice as a
means to improve roosting cover. If managers feel suitable roosting
cover is limited, we suggest management for conifers having denser
foliage (e.g., eastern red cedar; Thompson and Fritzell 1988a). Also,
stands having high stem densities afford good escape and roosting
cover for grouse, and clearcuts or heavy selective cuts are often used to
create such habitat. Managers should consider orienting these habitat
management units such that they connect foraging habitat on lower
slopes with roosting habitat on upper slopes, thereby reducing expo-
sure of grouse to predators.
ACKNOWLEDGMENTS
The authors are grateful for the assistance of several ACGRP cooperators,
who worked unusual hours in adverse weather to collect the data presented here.
The manuscript was greatly improved through critical reviews and input from
Patrick Devers, Carola Haas, Roy Kirkpatrick, Gary Norman, Jeff Walters, and
two anonymous reviewers.
LITERATURE CITED
Bergerud, A.T., and M.W. Gratson. 1988. Survival and breeding strategies of
grouse. Pp. 473–577, In A.T. Bergerud, and M.W. Gratson (Eds.). Adaptive
Strategies and Population Ecology of Northern Grouse. University of Min-
nesota Press, Minneapolis, MN. 809 pp.
Bump, G.R., R.W. Darrow, F.C. Edminster, and W.F. Crissey. 1947. The
Ruffed Grouse: Life History, Propagation, and Management. New York
State Conservation Department, Buffalo, NY. 915 pp.
Doerr, P.D., L.B. Keith, D.H. Rusch, and C.A. Fischer. 1974. Characteristics of
winter feeding aggregations of ruffed grouse in Alberta. Journal of Wildlife
Management 38:601–615.
Fearer, T.M. 1999. Relationship of ruffed grouse home range size and move-
ment to landscape characteristics in southwestern Virginia. M.S. Thesis,
Virginia Tech, Blacksburg, VA. 95 pp.
Geiger, R. 1950. The Climate Near the Ground. Harvard University Press,
Cambridge, MA. 482 pp.
Gullion, G.W. 1970. Factors influencing ruffed grouse populations. Transac-
tions of the North American Wildlife and Natural Resources Conference
35:93–105.
Gullion, G.W., and F.J. Svoboda. 1972. Aspen, the basic habitat resource for
ruffed grouse. Pp. 113–119, In J.H. Ohman, F.H. Kaufert, M.R. Allen
(Eds.). Proceedings of the Aspen Symposium. U.S. Forest Service General
Technical Report NC-1. 154 pp.
Hewitt, D.G., and R.L. Kirkpatrick 1997. Ruffed grouse consumption and
detoxification of evergreen leaves. Journal of Wildlife Management
61:129–139.
D.M. Whitaker and D.F. Stauffer2003 391
Hungerford, K.E. 1951. Ruffed grouse populations and cover use in Northern
Idaho. Transactions of the North American Wildlife Conference 16:216–
224.
Huempfner, R.A., S.J. Maxson, G.J. Erikson, and R.J. Schuster. 1975. Recap-
turing radio-tagged ruffed grouse by nightlighting and snow-burrow netting.
Journal of Wildlife Management 39:821–823.
Kalla, P.I., and R.W. Dimmick. 1995. Reliability of established sexing and
aging methods in ruffed grouse. Proceedings of the Annual Conference of
Southeastern Association of Fish and Wildlife Management Agencies
49:580–593.
Kubisiak, J.F., J.C. Moulton, and K.R. McCaffery. 1980. Ruffed grouse density
and habitat relationships in Wisconsin. Wisconsin Department of Natural
Resources Technical Bulletin 118.
Lenth, R. V. 1981. On finding the source of a signal. Technometrics 23:149–
154.
Marjakangas, A. 1990. A suggested antipredator function for snow-roosting
behaviour in the black grouse Tetrao tetrix. Ornis Scandinavica 21:77–78.
Noon, B.R. 1981. Techniques for sampling avian habitats. Pp. 42–52, In D.E.
Capen (Ed.). The Use of Multivariate Statistics in Studies of Wildlife Habi-
tats. USDA Forest Service, General Technical Report RM-87.
Norman, G.W., and R.L. Kirkpatrick. 1984. Food, nutrition and condition of
ruffed grouse in southwestern Virginia. Journal of Wildlife Management
48:183–187.
Ott, R.A. 1990. Winter roosting energetics, roost site selection, and time-
activity budgets of ruffed grouse in central New York. M.S. Thesis, College
of Environmental Science and Forestry, State University of New York,
Syracuse, NY. 95 pp.
Pietz, P.J., and J.R. Tester. 1982. Habitat selection by sympatric spruce grouse
and ruffed grouse in north-central Minnesota. Journal of Wildlife Manage-
ment 46:391–403.
Pekins, P.J., F.G. Lindzey, and J.A. Gessaman. 1991. Physical characteristics of
blue grouse winter use-trees and roost sites. Great Basin Naturalist 51:244–
248.
Reynolds, M. C., D. F. Stauffer, R. L. Kirkpatrick, and G. W. Norman. 2000. A
Summary of Findings from Phase I of the Research Project. Appalachian
Cooperative Grouse Research Project, Virginia Department of Game and
Inland Fisheries, Verona, VA. 12 pp.
Servello, F.A., and R.L. Kirkpatrick 1987. Regional variation in the nutritional
ecology of ruffed grouse. Journal of Wildlife Management 51:749–770.
Sokal, R.R., and F.J. Rohlf. 1995. Biometry: the Principles and Practice of
Statistics in Biological Research. 3rd Ed. W.H. Freeman, New York, NY.
887 pp.
Swenson, J.E., and B. Olson. 1991. Hazel grouse night roost site preferences
when snow-roosting is not possible in winter. Ornis Scandinavica 22:284–
286.
Thomas, V.G., H.G. Lumsden, and D.H. Price. 1975. Aspects of the winter
metabolism of ruffed grouse (Bonasa umbellus) with special reference to
energy reserves. Canadian Journal of Zoology 53:434–440.
Southeastern Naturalist Vol. 2, No. 3392
Thompson, F.R., and E.K. Fritzell. 1988a. Ruffed grouse winter roost site
preference and influence on energy demands. Journal of Wildlife Manage-
ment 52:454–460.
Thompson, F.R., and E.K. Fritzell. 1988b. Ruffed metabolic rate and tempera-
ture cycles. Journal of Wildlife Management 52:450–453.
Walsberg, G.E. 1983. Avian ecological energetics. Pp. 161–220, In D.S. Farner,
J.R. King, and K.C. Parkes (Eds.). Avian Biology, Volume 7. Academic
Press, New York, NY.
White, G. C., and R. A. Garrott. 1990. Analysis of Wildlife Radio-tracking
Data. Academic Press, New York, NY. 383 pp.
Woehr, J.R. 1974. Winter and spring shelter and food selection by ruffed grouse
in central New York. M.S. Thesis, College of Environmental Science and
Forestry, State University of New York, Syracuse, NY. 128 pp.
