# Bed-site selection by European roe deer (Capreolus capreolus) in southern Norway during winter

> Where does a roe deer choose to lie down on a cold winter night? By following tracks through fresh snow in southern Norway, researchers found 294 beds and learned that the deer either rest close to where they feed in the open or walk farther to shelter under dense conifers, travelling farther when it is colder and when they rest longer. They also favoured small hilltops and lay just a few metres apart in groups, likely to keep watch for predators.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/1995-mysterud-ostbye-bed-site-selection-european/

## Citation and access

- **Authors:** Atle Mysterud and Eivind Østbye
- **Year published:** 1995
- **Journal or source:** Canadian Journal of Zoology
- **Full citation:** Mysterud, A. and Østbye, E. 1995. Bed-site selection by European roe deer (Capreolus capreolus) in southern Norway during winter. Canadian Journal of Zoology 73(5): 924-932.
- **DOI:** 10.1139/z95-108
- **Access status:** Publisher link only
- **Where to find it:** https://doi.org/10.1139/z95-108

## Tags

- **Species:** Mammals > Roe deer
- **Region:** Europe > Norway
- **Tracking:** Beds; Feeding signs; Snow tracks; Tracking; Tracks; Tracks and signs identification; Trailing
- **Behaviour and ecology:** Behaviour; Feeding; Habitat selection; Predation risk; Predator avoidance
- **Methods and technology:** Analysis of variance (ANOVA); Mann–Whitney U test; Wilcoxon matched pairs test
- **Data type:** New field data
- **Observer qualification:** Not reported
- **Discipline:** Behaviour and ecology; Wildlife tracking
- **Year published:** 1995

## Detailed keywords

- **Species:** Studied: European roe deer (Capreolus capreolus). Mentioned for comparison or as possible predators, not studied: white-tailed deer (Odocoileus virginianus), elk (Cervus canadensis), mule deer (Odocoileus hemionus), moose (Alces alces) and lynx (Lynx lynx); hunting dogs. Trees and shrubs recorded: Norway spruce (written “Norwegian spruce”, Picea abies), Scots pine (Pinus sylvestris), hoary alder (Alnus incana), bird cherry (Prunus padus), elm (Ulmus glabra), linden (Tilia cordata) and yew (Taxus baccata), an important winter browse plant for roe deer in parts of the area.
- **Location and habitat:** Lier valley, southern Norway (59°52’ to 59°58’N, 10°14’ to 10°20’E), a 740 ha study area in the boreonemoral region. Commercially managed forest dominated by Norway spruce mixed with Scots pine, with clearcuts; deciduous forest and cultivated fields along the valley bottom. Hilly terrain rising from Lake Holsfjorden (63 m above sea level) to over 400 m, with clay-eroded ravines 10 to 100 m deep. Estimated 25 to 40 roe deer in the study area, heavily hunted from August to December.
- **Topics:** winter bed-site selection; thermal cover and energy conservation; canopy cover versus ground cover; “forage beds” and “cover beds”; distance between feeding and bedding sites; bedding duration and temperature; wind and operative temperature; microtopography (bedding on small tops); bed scraping; group bedding and spacing between beds; reuse of bed sites; antipredator strategies; view of the incoming track.
- **Statistics or models:** Paired comparisons of each bed with a random point 50 m away (and 10 m away for nearest tree species): Wilcoxon’s paired test for continuous data, sign test for categories; Mann–Whitney U test (reused versus not reused sites, scraping); Spearman’s rank correlation (canopy cover versus ground cover, -0.491); two-way ANOVA with canopy cover (arcsine square root transformed) and with bed-to-feeding distance as dependent variables, bedding time and temperature as factors (the models explained 26% and 23% of the variation). Software: SYSTAT. Small inconsistency: the abstract and discussion say deer in groups bedded 2 to 4 m apart, but the results give an average median minimum distance of 4.3 m for two deer (2.2 m for three, 2.0 m for four).
- **Data type:** New field data: beds located by tracking in snow, November 1992 to January 1993 (winter 1992-1993). Weather data from a nearby meteorological station (Tryvasshøgda).

## Synopsis

### 1. What they did

In winter, deer face a balancing act: food is often found in open areas, but resting under conifers can save energy in the cold. Roe deer rest for 60 to 65% of the time in winter, yet no one had studied in detail where they choose to bed. The authors set out to describe roe deer winter beds, especially their cover and position in the landscape, and to test whether deer bed under heavier canopy when it is colder and when they rest for longer.

- Finding beds. From November 1992 to January 1993, on snow no more than 3 days old, they walked lines about 150 m apart through randomly chosen parts of a 740 ha study area in the Lier valley. When they found a roe deer track, they followed it both backwards and forwards to the edge of the area, marking every bed. They found 294 beds. Beds within 30 m of each other were counted as one bed site (118 sites), and one bed per site was chosen at random for measurement.

- Comparing each bed with a random point. At each bed and at a point 50 m away in a random direction, they recorded the nearest tree (conifer or deciduous), canopy cover (with a spherical densiometer 30 cm above the ground, a bedded deer’s view), ground cover (how far away a person had to stand before a 40 cm cylinder in the bed was hidden), slope and position relative to small-scale topography (top, upper, middle or lower third, bottom, or flat).

- Linking beds to feeding. They backtracked from each bed to the first sign of feeding, measured the distance and the canopy cover there, and grouped beds as close to (under 30 m) or far from (over 30 m) the feeding area.

- Bedding time and weather. From the amount of snow melted in each bed, they judged whether the deer had lain there a short or long time (only when this was clear, 67 beds), and classed nights as cold or mild by whether the minimum temperature fell below -5°C.

- Other behaviour. They recorded whether beds were scraped clear, whether sites were reused, how many deer had bedded together and how far apart, and how far the deer could see along their own incoming track.

### 2. What they found

Sharing and reuse

- 22.9% of bed sites were reused within 30 m, most of these (77.8%) within 5 m. Reused and not reused sites did not differ in canopy cover or topographic position.

- At 61% of sites, more than one roe deer had bedded: 2 deer at 41.6% of these, 3 at 30.6%, 4 at 13.9%, and an unknown number at 13.9%. Beds in a group were close together, on average 2.0 to 4.3 m apart.

Cover

- Beds were more often under conifers than random points (77 of 113) and closer to a conifer trunk (median 85 cm versus 180 cm at random points).

- Overall, beds had more canopy cover than random points (median 48% versus 21%), but less ground cover.

- Feeding took place in open areas (median canopy cover 1%).

- Beds split clearly into two kinds. Beds close to the feeding area had little canopy cover (median 10%), no more than random points. Beds far from the feeding area had heavy canopy cover (median 59%) and were mostly under conifers.

- Canopy cover and ground cover were negatively related: where the canopy was dense, ground cover was thinner.

Temperature and bedding time

- Beds used for a long time, and beds made in cold periods, had more canopy cover than beds used briefly or made in mild periods.

- This was because the deer travelled farther from the feeding area before bedding down when they rested longer and when it was colder.

- In open habitat (under 10% canopy cover), beds had more ground cover on the side the wind came from than on the opposite side (13 beds).

Topography and other behaviour

- Beds were far more often on small tops than random points were (64 of 118 beds versus 13 of 118 random points), and on gentler slopes. The median height of these tops was 10 m.

- From the bed, deer had a slightly better view back along their own incoming track than in the opposite direction (median 27 m versus 23 m).

- 48.3% of beds were scraped. Scraped beds were more often under conifers (81.5% versus 55.9%) and more often used for a long time (82.8% versus 28.9%).

- No signs of predation on roe deer were found, but there had been intense hunting with dogs before the study.

### 3. What they concluded

- In winter, roe deer prefer to bed under dense canopy, which gives thermal and possibly other benefits. Because they feed in open areas, they must balance feeding in the open with sheltering in conifer stands.

- This gives two kinds of beds: “forage beds” with little cover close to feeding areas, and “cover beds” under heavy canopy farther away. How long the deer will rest and how cold it is help decide whether they leave the feeding area before bedding down, though the simple model explained only about a quarter of the variation.

- Wind was a factor in bed choice in open habitat, so wind probably mattered for how cold the deer actually felt.

- Scraping beds may be comfort behaviour, and may also save energy by removing snow and rime that would draw heat from the body.

- Bedding on small tops is probably an antipredator strategy: a deer on a top can hear better and escape downhill out of sight, while a predator must attack uphill. Bedding 2 to 4 m apart in groups may make it harder for a predator to approach unseen. The better view of the incoming track could be a defence against predators following the track, or simply a result of deer moving efficiently along ridges; this needs closer study.

- Studies under different predation pressures are needed to confirm bed choice as an antipredator strategy, and clearer ways to measure cover and especially topography are needed.

### 4. How tracking was or was not used

Tracking in snow was the whole basis of the study. The researchers searched for roe deer tracks along lines through the study area, then trailed each track both backwards and forwards to find and mark every bed. They read the beds themselves (melted snow to judge how long a deer had lain there, scraped or not, how many deer had bedded together) and used tracks to tell apart a site reused over time from one used by a group at the same time. They also backtracked from each bed to the first feeding signs to measure the distance between feeding and bedding. The authors note that where roe deer were dense and tracks crossed, they sometimes could not follow a single track that far, so long distances between beds and feeding sites may be underrepresented.

### 5. Tracker or observer certification

Not reported. The paper does not say who did the tracking or what their experience was. This matters because the results depend on following single tracks, telling group sites from reused sites, and judging bedding time from snow melt, which the authors themselves describe as uncertain.

### 6. Other technologies, skills or innovations

- Lemmon spherical densiometer (model C), read 30 cm above the ground to match a bedded deer’s view of the canopy.

- A ground cover index: a 30 cm cylinder in the bed, marked 40 cm above the ground, and the distance at which it was first fully hidden from the north, south, east and west.

- Bedding time judged from the amount of snow melted in the bed.

- Weather data from a nearby meteorological station.

### 7. Implications for conservation (from the paper)

The paper does not discuss conservation directly. It points out that both canopy cover (thermal shelter) and ground cover matter to wintering roe deer and should be described separately, and that dense conifer cover near open feeding areas is part of how roe deer manage their energy in winter.

### 8. Implications for tracking (Original Wisdom's observations)

_These are Original Wisdom's observations, not the authors'._

- The study shows how much a set of winter beds can tell a tracker: how long a deer rested, whether it scraped the bed, how many deer lay together, and, by backtracking, how far it travelled from where it fed.

- For trackers looking for roe deer beds in winter, the patterns reported here are useful: forage beds close to open feeding areas, cover beds under conifers farther away (especially on cold nights), beds on small tops, and group beds only a few metres apart.

## Cite this summary

Lawrence, K. 2026. Summary of: Bed-site selection by European roe deer (Capreolus capreolus) in southern Norway during winter (Mysterud and Østbye, 1995). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/1995-mysterud-ostbye-bed-site-selection-european/