Northeastern Naturalist 55 B. Heinrich 2017 Vol. 24, Special Issue 7 Winter Strategies of Ruffed Grouse in a Mixed Northern Forest Bernd Heinrich* Abstract - I examined behavioral flexibility with respect to potentially conflicting demands of Bonasa umbellus L. (Ruffed Grouse) adapted to winter conditions in a mixed forest of western Maine. At the beginning of winter, during the first snow of 15 cm, the grouse roosted overnight 2–4 m above ground in dense stands of conifers. After 40 cm of snow depth, they dove into the snow in flight in open areas of deciduous forest and tunneled, at varying angles and distances from the entrance approach, to den at nigh t and most of the day. With new snow on top of packed snow near the end of winter, they switched tactics again and rested and spent the overnight hours in snow molds on the ground under cover of conifer branches or against tree trunks. The 3 behaviors are discussed in the context of winter diet and anti-predation strategies. Introduction Animals exhibit a variety of solutions for coping in winter. Birds in particular face the interrelated problem of securing enough food in often greatly reduced foraging time, and maintaining an elevated body temperature for long durations of the diurnal cycle. Greater rate of heat loss due to cold and small body size (Calder 1984) may be compensated for by better insulation, and/or increased shivering which increases heat production (Cooper and Swanson 1994, Dawson and Carey 1976, Walsberg 1983) abd requires greater food intake. Hypothermia and seasonal acclimation are alternate tactics (Chaplin 1976, Liknes et al. 2002, Marsh and Dawson 1989). Behavioral strategies used by birds in winter include overnighting in existing tree cavities (e.g., Poecile atricapillus (L.) [Black-capped Chickadee; Smith 1991], Sitta canadensis L. [Red-breasted Nuthatch; B. Heinrich, pers. observ.], and Sitta carolinensis Latham [White-breasted Nuthatch; B. Heinrich, pers. observ.]); excavating tree cavities in the fall specifically for overnighting (Dryobates pubescens L. [Downy Woodpecker] and Picoides villosus (L.) [Hairy Woodpecker]; B. Heinrich, pers.observ.; Kilham 1992); huddling in groups (Regulus satrapa (Lichtenstein) [Golden-crowned Kinglet]; Heinrich 2004); or burrowing in the snow (Acanthis flammea (L.) [Common Redpoll]; Heinrich 2014). Winter survival strategies of Bonasa umbellus L. (Ruffed Grouse) could potentially encompass several of the above strategies (Ber gerud and Gratson 1988). Ruffed Grouse are one of the best-known game birds living year-round in forests throughout North America (Atwater and Rusch 1991, Bump et al. 1947, Rusch et al. 2000), where they feed in the winter on tree buds (Hewitt and Messner 2000, Huepner and Tester 1988, Jakuba and Gullian 1991). Tree buds are potentially available in large amounts, although they contain high-cellulose, low-caloric bulk *Box 153, Weld, ME 04285; bheinrich153@gmail.com. Manuscript Editor: Susan Z. Herrick Winter Ecology: Insights from Biology and History 2017 Northeastern Naturalist 24(Special Issue 7):B55–B71 Northeastern Naturalist B. Heinrich 2017 56 Vol. 24, Special Issue 7 (Casaway et al. 1976). Additionally, given standard avian energy economics (Dawson and Hudson 1970, Reinertsen and Haftorn 1986), Ruffed Grouse, because of their large size (near 500–600 g), face a low rate of heat loss relative to other winter residents such as kinglets, chickadees, and small finches weighing 5–13 g. They should thus be less constrained by the energy demands imposed by low temperatures; however, like many small birds that remain in the Northern Hemisphere in the winter, Ruffed Grouse overnight in dense coniferous foliage where they gain energy economy (Thompson and Fritzell 1988a, Whitaker and Stauffer 2003). Unlike Lagapus spp. (ptarmigan), which molt to cryptic white plumage in the winter, Ruffed Grouse are camouflaged only in the summer and may thus be constrained in their foraging as well as roosting behavior in the winter. They are choice prey of many mainly visually oriented predators (Bergerud and Gratson 1988, Gormley 1996, Gutierez et al. 2003, Small et al. 1991). They would presumably be especially vulnerable in the winter to their major aerial predators such as Accipiter gentilis L. (Northern Goshawk; Eng and Gullion 1962, Tomberg and Fritzell 1988b), both on snow-covered ground as well as in the tops of then-bare deciduous trees where they often feed. Like some other birds, their choice of night roosts appears to be based on both safety (Marjakangas 1990) and energetics (Swenson and Olson 1991). I herein examine Ruffed Grouse snow-use and roosting behavior in the context of seasonal changes at a site in Maine where the birds experience snowcover varying from absence to shallow to deep, and have a choice of habitat from open to stands of dense conifers. The standard term for typical bird behavior of perching at various levels of elevation and exposure is “roosting”. However, grouse are also well-known for “snow-roosting” to avoid predation by burrowing to shelter under the snow (Gullion 1984). For differentiating these 2 behaviors, and also that of resting on the snow surface in molds, and in common usage with respect to function in other animals, I here refer to the under-snow sheltering as “denning”. However, no account had been made of the relation of when, where, and how denning relates to roosting and the use of molds, and to season or to the lengths of time birds may stay in residency in these places. Given the ~15 hours of local darkness in winter, and several hours of den-residence time added either before night or after, snow-den residences could have been as long as 18 hours. I here examined these behavioral changes in relation to seasonal and presumably environmental conditions. Methods Site description My observations were conducted in ~50-ha of woodland (44°41'18.89883"N, 70°22'20.37437"W) on York Hill, in western Maine, at an elevation of 350–400 m. The area contains a patchwork of a wide range of habitat choice and food. Portions of the forest had been logged at various times over the past 5–50 years, creating variable-aged stands of Picea–Abies (spruce–fir) and Pinus strobus L. (White pine) as well as mixed-species deciduous patches of early successional spruce–fir adjacent to a maturing hardwood forest on a sloping hillside (Fig. 1A) adjacent Northeastern Naturalist 57 B. Heinrich 2017 Vol. 24, Special Issue 7 to recently logged areas containing dense regrowth of 4–12 cm diameter Abies balsamea (L.) Mill. (Balsam Fir) (Fig.1B). The primary winter food of grouse, buds of primarily Betula spp. (birch), Populus spp. (poplar), and Ostrya virginiana (Mill.) Koch (Eastern Hop Hornbeam), were distributed throughou t the area. Figure 1. Ruffed Grouse winter habitat at York Hill, western Maine: (A) deciduous forest and (B) adjacent Balsam Fir thickets. Northeastern Naturalist B. Heinrich 2017 58 Vol. 24, Special Issue 7 I have informally observed Ruffed Grouse at this site for decades, and as expected from studies elsewhere (Bergerud and Gratson 1988), they routinely roost overnight at several meters height in the branches of dense spruce–fir thickets in spring and fall. In winter, Ruffed Grouse, often seen in groups of 2–4 individuals, feed at dusk and dawn on buds in the tops of tall birch and poplars. Their overnighting winter roosts in conifers are easily recognized by numerous scat scattered on the snow. Similarly, I had routinely encountered over the years in this area grouse snow dens and molds, and have determined grouse occupancy in a den both by scat and/or flushing into flight when approached. Environmental conditions Since I do not want to assume the environmental variables affecting behavior, I here separate the observations into the 3 main seasonal periods observed. In the first period, “February 2015” (2 February to 3 March 2015), the snow was powdery and near 40 cm in depth on top of additional packed snow below. Both day- and night-time temperatures were, as is typical in other years, not above 0 °C in the daytime, and usually near or below -20 °C at night. Almost all of the snow denning was observed specifically in this period. In the second period, “March 2015” ( 4 March to 3 April 2015), it had warmed briefly to daytime temperatures of 8 °C, but temperatures at night continued to remain below freezing. The snow softened and then crusted, but there were 4–5 cm of fresh fluffy snow on that crust when the observations were made. The third period, “January 2016”, (31 December 2015 to 26 January 2016) was immediately after that year’s first significant storm that deposited ~14–17 cm of fluffy snow that remained throughout the period. Night-time temperatures ranged from -25 to -7 °C. Grouse activity Ruffed Grouse activity and approximate timing of snow-den construction and roosting site use were determined by twice-daily surveys for roosting, snow dens, and resting molds as they occurred anew along an arbitrary 5-km route, chosen to traverse different habitat. Visibility of Ruffed Grouse activity marks on/in the snow was ~5 m to either side of the trail. Frequent light snow dustings and snow disturbance by wind, and varying times of day at which I conducted the surveying, were also used as aides for estimating den making and occupancy times. The survey route led through nearly mature hardwood forest of primarily Acer rubrum L. (Red Maple), A. sacharum Marsh (Sugar Maple), Fraxinus americana L. (White Ash), Fagus grandifolia Ehrh. (American Beech), Betula papirifera Marsh (Paper Birch), Betula alleghaniensis Britt. (Yellow Birch), and Populus balsamifera L. (Balsam Poplar) (Fig. 1A), to secondary growth of the same tree mixture with Balsam Fir and Picea rubens Sarg. (Red Spruce), and also through young growth of the latter (Fig. 1B). I located the snow dens by the entry marks left in the snow. I made scat counts after scraping away the snow surrounding a den and separating the scatclump. At over-nighting roosts in trees, the scat were clearly visible scattered over the snow surface. Northeastern Naturalist 59 B. Heinrich 2017 Vol. 24, Special Issue 7 I assayed scat to make inferences about den- and mold-occupancy durations. Ruffed Grouse produce 2 kinds of scat. The first is derived from coarse material that is passed directly into the large intestine, whereas the other is from fine pasty liquid material that is retained much longer in the cecum (Casaway et al. 1976). The intestinal scat of Ruffed Grouse was ~2.5 cm long and relatively solid sausageshaped (weighing on average 1.8 g each). The cecal scat was amorphous brown semi-liquid that had usually frozen solid by the time it was assayed in this study in the field. Given the high fiber diet of tree buds in the winter, Ruffed Grouse produce the intestinal scat in bulk. An absence of den sites along the survey route in late afternoon versus their presence the next morning, as well as the number and kind of scat left in them in the morning, were used as evidence for overnighting dens. The cecal scat was found only at overnight roosts and dens, and it was deposited shortly before or directly during the birds’ den departure. At night tree roosts, the cecal scat, when not seen on the ground, was located instead stuck onto the tree branches on or under where the bird had roosted. Ruffed Grouse routinely made snow dens within a meter or two of the snowshoe tracks of the survey course, and they were therefore apparently not disturbed by my tracks. To test if the sometimes-observed clumping of den-site locations (within about 10 m of one another) resulted from grouse using the den-entrance marks on the snow as a signal for their own denning there, or if alternately they were derived from birds who had stayed together after their just previous feeding, I made marks on/in the snow that mimicked grouse snow-den entry marks. One hundred of the dummy den sites were made available on one day of fresh snow at approximately equi-distant intervals of at least 30 m over the 5-km den survey trail. These pseudoden entrances that closely mimicked grouse snow-den entry marks were created by tossing a bantam rooster carcass on a string into snow, and immediately yanking it back out. They were accessed twice daily on the regular route survey for 4 successive days. Predators As determined by tracks and sightings, potential predators in the area included Canis latrans Say (Eastern Coyote), Lynx rufus (Schreber) (Bobcat), Mustela frenata (Lichtenstein) (Long-tailed Weasel), Neovison vison (Schreber) (Mink), and Martes pennanti (Erxleben) (Fisher), all of which left tracks in the area. An Accipiter gentillis L. (Northern Goshawk) and an Accipiter cooperii (Bonaparte) (Cooper’s Hawk) were both seen once near the area during the current study as well as before it. No avian predators were currently nesting nearby, although a pair of Bubo virginianus (Gmelin) (Great-horned Owl) and Buteo jamaicensis (Gmelin) (Red-tailed Hawk) had nested there earlier. Results February 2015 The only Ruffed Grouse I encountered during ~150 hr in the woods in the daytime were those I flushed from their snow dens on approach, although on the 2 Northeastern Naturalist B. Heinrich 2017 60 Vol. 24, Special Issue 7 occasions that I was out for an hour at dusk I saw a pair feeding on buds near the tops of Yellow Birch. During that month, no grouse tracks were found on the snow except those occasionally associated within ~1–2 m of a den. In the 111 dens I examined during this time, 16 were day-dens (the grouse were in and then flushed from them). Seventy-two were used overnight, as determined from their absence along the survey route in the evening and presence in the morning. In 23 dens, the difference between day versus night residency was undetermined. All grouse dens noted during this time of low temperatures, no wind, and deep fluffy snow were in relatively open hardwood forest characterized by areas of cleared snow surface with access by flight. In 109 of the 111 denning occasions, there were no tracks leading into the den site. Den exit holes had imprints of 1 or 2 wing-beats and usually 1 or 2 foot-steps, although occasionally grouse foottracks extended for up to 2 m from the exit holes. The den cavity, which contained grouse intestinal fecal pellets, were usually located within 0.5 m of the entrance hole, but sometimes the tunnel to the den extended 1 m or more, with one situated 2.7 m in from the entrance (Fig. 2). The tops of the tunnels were 2–5 cm beneath the snow surface, and the longer tunnels commonly had what appeared to be a deliberately made small peek-hole (Fig. 3). As opposed to passive cave-ins, these holes to the snow surface had snow extruded from them and were vertical and round. Fig. 2. Diagrammatic configurations of a sample of Ruffed Grouse dens during 40-cm snow conditions, showing den locations versus entry and exit points into the snow, tunneling, and (at top left) magnification of a den and (at top right) side view showing locations of the cecal and intestinal scat arrangement of an over -nighting den. Northeastern Naturalist 61 B. Heinrich 2017 Vol. 24, Special Issue 7 Figure 3. Surface observations at sites of Ruffed Grouse snow-dens, February 2015. E = entry to den, P = “peep hole”, X = exit, W = wing marks on leaving, F = furrow. (A) 3 February—bird tunneled left on entered snow, overnighted by peep hole, and then exited the next day, leaving wing marks as it took flight. Scale: backpack length = 48 cm. (B) 25 February—birds travelled under snow to right, spending night by peep hole. (C) 27 February— after spending night by peep hole, bird exited and walked across snow surface leaving furrow, before taking flight. Northeastern Naturalist B. Heinrich 2017 62 Vol. 24, Special Issue 7 Overnighting dens contained 45–85 (mean of 61) intestinal scat. Scat counts of the 16 birds flushed from their dens in the daytime varied from 2 to 36, with those dens from which I flushed the grouse early in the day containing the fewest, whereas those with over 30 intestinal scat were flushed out near noon. Other, presumably new, dens near the end of the day also had low scat counts. Day-dens lacked cecal scat. Dens that were made in the evening and occupied until shortly after dawn contained 2–5 cecal scat along with the pile of inte stinal scat. In night-dens, the cecal scat was visible at a glance on the snow surface in the den, and sometimes at the den exit. But the intestinal scat was piled up, and often partially buried within the den in a tight multi-layered clump of ~5 cm diameter, suggesting the bird had apparently not relocated from the same spot for hours. The usual thin snow layer between the cecal scat on top and intestinal scat under it, and a lack of disruption of shape of the almost fluid paste-like cecal scat on issuing and its placement, indicates that the latter was not voided by the grouse until upon, or just before, leaving the den. I estimated den residence time using an inferred scat clock proceeding at an intestinal scat production rate of about 4.7 /hr.; this rate was derived by dividing the mean of 61 scat found in night-dens by the 13 hrs of nighttime residency. Ruffed Grouse (n = 16) were flushed from their dens at various times from morning to late afternoon, on days of sunshine and no snowstorm activity nor unusual temperature lows. Either they made/entered new snow dens in the morning after their dawn feeding and then remained denned for various lengths of time, or they had not left their night dens by morning. The scat number left in those snow dens of grouse I had flushed in the daytime increased approximately linearly until at least 13:00 hr, when it reached 36 (Fig. 4). Applying my calculated scat production rate described above, the grouse should have produced 37.6 scat if they denned after feeding near dawn from 05:00 hr until 13:00 hr., and ~28 scat if they had denned from 07:00 hr to 13:00 hr. This estimate suggests that the grouse could have snow-denned for up to at least 6 hours of the day until I had flushed them. Low scat numbers late in the day presumably indicate new dennings that I had interrupted. March 2015 On 4 March, the first day with above-freezing temperatures that year, the Ruffed Grouses’ behavior changed abruptly. From that day and through the next week, temperatures rose to +8 °C in the day, although night-time temperatures continued to be near -7 °C. The top layers of snow crusted at night with occasional dustings added. I found no more snow dens in the open areas of the hardwood forest on my circuits at that time. However, there then appeared molds pressed into the softening snow surface. The 18 snow molds (several cm deep) observed were all under the cover of live conifer branches or adjacent to thick tree stems. All were associated with tracks on the snow leading into and out of them. Scat (only the intestinal was present) numbers per mold ranged from 3 and up to 36 (mean = 7.3). By 18 March, after 4–6 cm of fresh snow covered a crust, the birds remained in dense conifer thickets where they left numerous track-ways that, in the length of Northeastern Naturalist 63 B. Heinrich 2017 Vol. 24, Special Issue 7 several km of them that I followed, were all under sunlight-excluded cover. Upon reaching thicket edges, the tracks returned back into the cover. The birds had fed on seeds shed by several birches that towered over or along the thicket, and on the evergreen fronds of Dryopterus marginalis (L.) (Marginal Fern) at its edge. Of the 21 observed individual overnighting locations, none were in either the deciduous forest nor in mature conifers. All were molds several meters into and under the cover of the dense young Balsam Firs and spruces growing in the thickets. January 2016 Evidence of near identical behavior to that observed in March 2015 was found in the same area after the first snowstorm blanketed the ground with 14–17 cm of powdery snow. Temperatures at night had dipped to -26 °C, and the snow remained powdery throughout the period. In my 6 days of approximately equal search in all habitats, I located 24 overnighting (tree) roosts (Fig. 5A). All were in low (estimated 2–4 m) young Balsam Fir trees. This survey also revealed 23 molds 5–15 cm into the soft snow that had been day residencies under overhanging fir branches (Fig. 5B). Tracks originating from grouse landing-marks in the snow in open snow at the thicket edges led to where the birds had made their snow molds in the thickets (most under less than 1 m below hanging snow-laden boughs, and/or usually next to a tree trunk). These molds contained 3–14 (mean of 8.8) intestinal, but no cecal, scat. In contrast, all the 24 overnighting roosts were located (by the scat on the snow beneath) at the same time in the same area in young fir trees. They were associated Figure 4. The number of scat left in snow-dens after the Ruffed Grouse had been flushed from them, indicating that grouse likely denned until mid-day, to then exit and resume denning shortly afterwards. All scat were of the intestinal variety. Northeastern Naturalist B. Heinrich 2017 64 Vol. 24, Special Issue 7 with 34–72 (mean of 61.6) intestinal and 2–4 (mean of 2.7) cecal scat. I found no evidence to indicate that any grouse had remained to overnight in the deciduous woods where they had denned during the deep snow the year befor e. Approximately 2 weeks later (13 to 26 January 2016), the roosting behavior of the grouse in the same area changed once more, even though the snow- and Figure 5. Ruffed Grouse behavior sign in early January 2016, when depth was 16 cm or less. (A) About 50 intestinal and 3 cecal scat scattered under overnight roosting place of a grouse at 3 m in live branches of a Balsam Fir tree in a thicket. (B) Three resting day-molds (M) of 3 grouse near each other at the edge of a barrier (here a brush-pile), each containing only 8–13 intestinal scat. Northeastern Naturalist 65 B. Heinrich 2017 Vol. 24, Special Issue 7 temperature conditions had remained virtually identical. Two night roosts in the Balsam Fir trees, 10 overnighting molds, and 1 snow den were located. Mean scat numbers were similar to before, intestinal at 45–85 (mean of 61.2), and cecal at 1–5 Figure 6. Grouse molds in ~16-cm snow conditions in January 2016. (A) A day mold next to a tree. Note walking tracks in and out of the ~10-cm-deep mold. (B) An over-nighting mold with cecal pellets (CP) on left. Northeastern Naturalist B. Heinrich 2017 66 Vol. 24, Special Issue 7 (mean of 2.8). Thirty-four day-molds were located with 0–34 (mean of 16.1) intestinal and no cecal scat. Both the day and night molds (Fig. 6A) were again placed adjacent to a backing such as brush or a thick tree trunk where an open flight path remained on the other side(s) of the resting place. The day and night molds (Fig. 6B) were shallow (3–6 cm deep). Three grouse were flushed at 09:30, 13:30, and 14:10 hrs from these molds holding 13, 34, and 2 scat, respecti vely. Clumping In the past I had observed Ruffed Grouse in the area in small groups (2 to 5), while foraging on the snow for birch seeds, perching in trees, and while feeding on tree buds at dusk. The snow den and mold locations also occasionally indicated clumping. During the 2015 period, there were 18 apparent groupings within the total of 111 dens, with 14 pairs of 2 dens within about 6 m of one another, and 4 sets of 3–5 dens within 10 m. In 2016, as before, some grouse molds were clumped (Fig. 5B); of 48 over-nighting and day-molds, 25 were in groups of 2–4 within 2–10 m of one another. However, during the 4 days I had made 100 psuedo-den entrances avialable, none of the 159 natural molds/dens I found were located within such close near-neighbor distances to the artificial den entrances. Predators None of the other 253 dens/molds observed in this study on the transects were associated with obvious sign indicative of a disturbance by the currently available potential predators. However, 2 incidences of grouse predation were inferred (from evidence of tracks, leavings of feathers, and grouse gizzards) in the greater area surrounding the study sites, during January 2016. A Vulpes vulpes L. (Red Fox) had killed a grouse at a den containing 25 intestinal scat, and an undetermined predator had killed another grouse. Snow conditions were not suitable for further inferences. The gizzards of both kills were packed with birch buds and grit . Discussion Birds overwintering in northern forests confront the problem of energy balance from cold, availability of suitable food and shelter, and reduced time for foraging (Aschoff 1981, Walsberg 1986). The solution to one problem likely affects others (Calder 1984). None of the small birds in New England overnight under snow, although they are potentially more challenged by the cold by their small size relative to grouse. But because of their winter diet of tree buds and large body size, Ruffed Grouse face a different mix of problems compared to most other overwintering woodland birds. For grouse, the energy cost of making a snow den is presumably negligible, but the benefit of evading predators in snow that it provides could be high. Food is a major variable of winter-adaptation, and Ruffed Grouse, in contrast to the overwintering insectivorous and granivorous birds, switch to a diet of tree buds after the first snowfall (Doerr et al. 1974). This diet, despite high bulk and low nutritional content (relative to insects and seeds), has the advantage of great Northeastern Naturalist 67 B. Heinrich 2017 Vol. 24, Special Issue 7 abundance and predictability at specific locations. Thus, while most other birds in the same environment may forage nearly continuously for most of the day, grouse relying on tree buds need to feed for only brief periods. In Alberta in winter, they feed daily on average for just 16 min prior to sunrise, and only another 24 min more to again fill their crops near an hour after sunset (Doerr et al . 1974). Ruffed Grouse feed on the ground into autumn before the trees set buds. Cryptic earth-coloration is a defense in summer, fall, and spring. However, normally through winter when they feed in the tops of bare deciduous trees, these large birds are visible from several hundred meters. They form small groups presumably not only because they are drawn together by the productive feeding spots where they can quickly fill their crops in the evening and in the pre-dawn (Doerr et al. 1974). Predator avoidance is another possible reason, achieved in groups by taking advantage of the “many eyes” effect for predator vigilance. Since buds are continuously available, the crepuscular foraging in tree tops is likely an adaptation reducing predation. Similarly, during winter, the grouses’ bare-ground–based camouflage on snow could make them more vulnerable off the trees than on them. In this study, Ruffed Grouse did feed on the snow surface in late winter, but only under the very dense cover of young conifers, at locations where birch seed had been spread after a wind. Goshawk account for most Ruffed Grouse predation, exceeding that by human hunters (Bergerud and Gratson 1988, Rusch et al. 2000). Cycles of goshawk abundance respond to forest grouse populations (Tomberg et al. 2005). Great-horned Owls prey on grouse as well (Devers et al. 2007) and could hunt them at low light intensities. Regardless of which feeding time is safer, a short feeding time as such would reduce exposure, and the grouses’ shift to crepuscular activity and their tendency to group are both consistent with anti-predator behavior . On the winter solstice (21 December), at the study site in Maine, there are, between official sunrise and sunset, 8 hr of day, followed by16 hr of night. With less than an hour allocated to feeding given the availability of tree buds, Ruffed Grouse in mid-winter have potentially 23 hours of “free” time available, and being a large conspicuous bird that is targeted prey may have spurred the evolution of anti-predator behavior to occupy that time. In this study, as in others (Bergerud and Gratson 1988, Whitaker and Stauffer 2003), Ruffed Grouse commonly roosted at sites that have been shown by studies to reduce heat loss by both convection and radiation (Huepener and Tester 1988), and as determined by using grouse taxidermy mounts, the heat loss while under snow is 14% less than it is while perched in a conifer (a cedar) (Thompson and Fritzell 1988b). Dense habitat also reduces convection, and grouse generally prefer dense habitat (Endrulat et al. 2005). As shown in the present study, where conifers were available in and around the habitat where the grouse lived, the grouse remained in such habitat even after the first snowfall of ~15 cm had occurred, and then switched to using open molds in the shallow snow. On the other hand, in mid-winter when deep snow became available, they then chose open areas for denning, rather than the immediately adjacent dense stands of conifers, Northeastern Naturalist B. Heinrich 2017 68 Vol. 24, Special Issue 7 even though the snow offered only a relatively slight thermal advantage to the insulation of conifers (Thompson and Fritzell 1988b). Then, within a day after that snow became unsuitable for denning because of a crust, they switched back to overnighting in dense conifers. These behaviors are consistent with respect to both energy economizing and predator defense. The snow dens were predominantly in open deciduous woodland, possibly because of unobstructed access and exit possibilities into and out of the snow. Since the birds had dived into the snow from flight (whereas they made molds on the snow by walking around under dense conifers), they needed open space not only to enter, but also to leave in perhaps variable directions during a forced exit at night. During escape from a potential predator coming from a random direction near a snow den, it would be important for the bird to avoid getting entangled in or hitting branches of brush or a nearby tree. Snow-denning in the daytime makes energetic sense in storms and at very low temperatures. But neither storms nor abnormally low temperatures were observed in this study, yet the grouse snow-denned on sunny days in the open areas and none were encountered either directly or indirectly by tracks. On the other hand, when snow denning was not possible, they remained day-active in the dense thickets, as evidenced by their extensive tracks there. Although the high visibility of a grouse on snow would change to near invisibility when it is under it, invisibility in snow is not necessarily invincibility to predator attacks, if a predator learns to associate den entrances with the presence of grouse. However, several features should work in favor of the grouses’ safety. In the absence of frequent snowfalls, the one-time use of dens creates numerous empty dens and dilutes their utility by a predator looking for grouse. Since dens are made by a quick dive, given suitable snow it is easy for grouse to make them on almost any occasion, and a potential predator faces a shell game scenario, especially when several empty den sites are clumped near a used one. Nearly half (43%) of the grouse dens assayed were within about 10 m of another den or resting mold, either because the association of 2 or more dens near each other is coincidental to the birds staying together, or because they seek existing dens to rest/overnight near them. If the same clumping effect had applied to the 4-day presence of 100 pseudo-dens in the same area as the 159 molds/dens observed during that time, then each had on average a 43% chance of being associated with a den. But none were, suggesting it was very improbable that they were choosing new den sites using visual cues of the presence of addtional dens. However, the observed den clumping, regardless of mechanism, could potentially be an adaptive behavior. Whether it is remains to be explored. The structure of the dens, which included odd angles from the entry point and the occasional long tunneling, suggests a design to foil a potential predator's ability to predict the grouses' location from the den entry marks on the snow. The location of the bird within the den was never deep relative to the snow surface, as would be predicted for enhancing retention of body heat. Instead, the always shallow submergence in the snow would facilitate quick escape, and the apparent peekNortheastern Naturalist 69 B. Heinrich 2017 Vol. 24, Special Issue 7 holes observed in the tops of the longer tunnels may have been used by the grouse in orienting themselves for potential escape directions. Finally, nothing could make the birds more invisible than being under the snow, with respect to anywhere above it. A molt into white feathers, such as that of ptarmigan in the Arctic, could serve the same function as hiding under the snow. However, although the duration of snow is as long or longer in the Arctic, the snow cover is thin there due to low precipitation and the conditions conducive for tunneling may be less. Furthermore, there are fewer, or no, forest thickets in the Arctic to hide in when snow conditions are unsuitable for snow-denning. I conclude that for Ruffed Grouse in a variable north temperate climate, the winter behavior of snow-denning strongly favors an anti-predation strategy. However, the observations leave open the extent to which and even how much these behaviors may contribute to winter survival in grouse with no camouflaging molt in winter. Acknowledgments I thank Charles H. Sewall and Scott R. Smedley for help with graphics, and an anonymous reviewer for thoughtfully helpful comments and suggestions . Literature Cited Aschoff, J. 1981. Thermal conductance in mammals and birds on body size and circadian phase. Comparative Biochemical Physiology 69A:611–619. Atwater, S., and D.H. Rusch. 1991. Ruffed Grouse. Stackpole Books, Harrisburg, PA. 370 pp. 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 Minnesota Press, Minneapolis, MN. 809 pp. Bump, G., R.W. Darrow, F.C. Edminster, and W.F. Crissey. 1947. The Ruffed Grouse: Life History, Propagation, Management. New York State Conservation Department, Albany, NY. 915 pp. Calder, W.A. 1984. Size, Function, and Life History. Harvard University Press, Cambridge, MA. 431 pp. Casaway, W.C., R.G. White, and D.F. Holleman. 1976. Digestion of dry matter and absorption of water in the intestine and cecum of Rock Ptarmigan. The Condor 78:77–84. Chaplin, S.B. 1976. The physiology of hypothermia in the Black-capped Chickadee, Parus atricapillus. Journal of Comparative Physiology 112:335–344. Cooper, S.J., and D.L. Swanson. 1994. Seasonal acclimation of thermoregulation in the Black-capped Chickadee. Condor 96:638–646. Dawson, W.R., and C. Carey. 1976. Seasonal acclimation to temperature in Cardouline Finches: I. Insulation and metabolic adjustments. Journal of Comparative Physiology 112:317–333. Dawson, W.R., and J.W. Hudson. 1970. Birds. Pp. 223–310, In G.C. Whittow (Ed.). Comparative Physiology of Thermoregulation. Vol. 1. Academic Press, New York. NY. Devers, P.K., D.F Stauffer, G.W. Norman, D.E. Steffern, D.M. Whitaker, J.D. Sole, T.J. Allen, S.L. Bittner, D.A. Buehler, J.W. Edwards, D.E. Figert, S.T. Friedhoff, W.W. Giuliano, C.A. Harper, WlK. Igo, R.L. Kirkapatrick, M.H. Seamster, H.A. Spiker Jr., D.A. Swanson, and B.C. Tefft. 2007. Ruffed Grouse population ecology in the Appalachian Region. Wildlife Monographs 168:1–36. Northeastern Naturalist B. Heinrich 2017 70 Vol. 24, Special Issue 7 Doerr, P.D., J.G. Keith, D.H. Rusch and C.A. Fisher. 1974. Characteristics of winter-feeding aggregations of Ruffed Grouse in Alberta. Journal of Wildlife Management 38:601–615. Endrulat, E.G., S.R. McWilliams, and B.C. Tefft. 2005. Habitat selection and home range of Ruffed Grouse in Rhode Island. Northeastern Naturalist 12:41 1–424. Eng, R.L., and G.W. Gullion. 1962. The predation of goshawks upon Ruffed Grouse at the Cloquet Forest Research Center, Minnesota. Wilson Bulletin 74:227–242. Gormley, A. 1996. Causes of mortality and factors affecting survival of Ruffed Grouse in Northern Michigan. M.Sc. Thesis. Michigan State University, East Lansing, MI. Gullion, G.W. 1984. The Ruffed Grouse. Willow Creek Press, Minocqua, WI. 144 pp. Gutieterez, R.J., G.S. Zimmerman and G.W. Gullion. 2003. Daily survival rates of Ruffed Grouse, Bonasa umbellus, in northern Minnesota. Wildlife Biology 9:351–356. Heinrich, B. 2004. Overnighting of Golden-crowned Kinglets in winter. Wilson Bulletin 115:123–124. Heinrich, B. 2014. Redpoll snow bathing: Observations and hypothesis. Northeastern Naturalist 21(4):N45–N52. Hewitt. D.G., and T.A. Messner. 2000. Ruffed Grouse (Bonasa umbellatus) foraging in aspen stands during winter in northern Utah. Western North American Naturalist 60(2):211–215. Huepener, R.A., and J.R. Tester. 1988. Winter arboreal feeding behavior of Ruffed Grouse in east-central Minnesota. Pp. 122–157, In A.T. Bergerud and M.W. Gratson (Eds.). Adaptive Strategies and Population Ecology of Northern Grouse. University of Minnesota Press, Minneapolis, MN. 809 pp. Jakubas,W.J., and G.W. Gullion. 1991. Use of quaking aspen flower buds by Ruffed Grouse: Its relationship to grouse densities and bud chemical composition. Condor 93:473–485. Kilham, L. 1992. Woodpeckers of Eastern North America. Dover, New York, NY. 240 pp. Liknes, E.T., S.M. Scott, and D.L. Swanson. 2002. Seasonal acclimation in the American Goldfinch revisited: To what extent do metabolic rates vary seasonally? Condor 104:548–557. Marjakangas, A. 1990. A suggested antipredator function of snow-roosting behavior in the Black Grouse, Tetrao tetrix. Ornis Scandinavica 21:77–78. Marsh, R.L., and W.R. Dawson. 1989. Avian adjustments to cold. Pp. 205–253, In C.H. Wang (Ed.). Advances in Comparative and Environmental Physiology 4: Animal Adaptation to Cold. Springer-Verlag, New York, NY. Reinertsen, R.E., and S. Haftorn. 1986. Different metabolic strategies of northern birds for nocturnal survival. Journal of Comparative Physiology B 156:655 –663. Rusch, D.H., S. Destefano, M.C. Reynolds, and D. Lauten. 2000. Ruffed Grouse (Bonasa umbellatus), No. 515, In A. Poole and F. Gill (Eds.). The Birds of North America. Academy of Natural Sciences, Philadelphia, PA., and Cornell Lab of Ornithology, Ithaca, NY. Small, R.J., J.C. Holzwart, and D.H. Rusch. 1991. Predation and hunting mortality of Ruffed Grouse in central Wisconsin. Journal of Wildlife Management 55:512–520. Smith, S.M. 1991. The Black-capped Chickadee: Behavioral Ecology and Natural History. Cornell University Press, Ithaca, NY. 362 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. Thompson, F.R., III, and E.K. Fritzell. 1988a. Grouse metabolic rate and temperature cycles. Journal of Wildlife Management 52:450–453. Thompson, F.R., III, and E.K. Fritzell. 1988b. Ruffed Grouse winter roost site preferences and influence on energy demands, Journal of Wildlife Management 53:15–21. Northeastern Naturalist 71 B. Heinrich 2017 Vol. 24, Special Issue 7 Tomberg, R., E. Korpimaki, S. Jungell, and V. Reif. 2005. Delayed numerical response of goshawks to population fluctuations of forest grouse. Oikos 1 11:408–415. Walsberg, G.E. 1983. Avian ecological energetics. Pp. 161–220, In D.S. Farner, T.R .King, and K.C. Parkes (Eds.). Avian Biology. Vol. 7. Academic Press, New York, NY. 542 pp. Walsberg, G.E. 1986. Thermal consequences of roost-site selection: The relative importance of three modes of heat conservation. Auk 103:1–7. Whitaker, D.M., and D.F. Stauffer. 2003. Night roost selection during winter by Ruffed Grouse in the central Appalachians. Southeastern Naturalist 2:377–392.