SEARCH

SEARCH BY CITATION

Keywords:

  • Calluna;
  • grazing;
  • fragmentation;
  • management

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

1. Heather moorland is an internationally important resource; it is valued as food and habitat for a range of herbivores as well as for landscape, conservation and recreation. In many parts of its range, grazing by large herbivores has impacted greatly on its current status and distribution. The interactions between two widespread herbivores, sheep and red deer, and the vegetation were examined within a naturally fragmented heather/grass mosaic in north-east Scotland from 1991 to 1996. This paper reports their foraging behaviour and inter-species interactions, and discusses the implications for grazing management within this vegetation type.

2. Both species foraged widely in this highly fragmented mosaic (red deer more widely than sheep) and consumed similar proportions of grass and heather. Neither species showed any major dietary shift as grass availability declined during experimental periods.

3. Grass patch size affected foraging behaviour differentially. Sheep spent more time grazing smaller grass patches (1–6 m2), whereas red deer showed no consistent grass patch size preferences (range: 1 m2 to over 200 m2). When lying down, red deer lay within heather almost exclusively, whereas sheep spent equal amounts of time within heather and small grass patches.

4. Faecal distribution also differed in a similar way, with concomitant implications for the spatial distribution of both physical and chemical impacts on the vegetation.

5. Contrary to previous hypotheses, the overall patterns of foraging behaviour by sheep and deer were little affected by the presence or absence of the other species. This suggests that range overlap between these species may have much less of an impact on their vegetation use than previously assumed.

6. Heather moorland is frequently found on sloping ground, which was shown to have a major effect on foraging, with a strong preference by both species for grazing facing either uphill or across the slope. This, and other slope-related impacts, resulted in spatially different patterns of heather use around the edges of grass patches.

7. The primary conclusions are that, although the overall use of grass and heather by sheep and red deer was similar within this fragmented mosaic, vegetation pattern clearly affected their foraging behaviour differentially. Sheep appeared to be much more affected by the scale of heather fragmentation than were deer and their habitat use was more closely focused on paths and grass patches, which contrasts with the more even use of the mosaic by deer. The implications of these behavioural differences for the management of upland grazings under one or both of these herbivore species are discussed.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Large herbivores have a major influence on the function and dynamics of most terrestrial biomes (Hodgson & Illius 1996). Through their grazing, trampling, defecation and urination they affect nutrient flows, vegetation community dynamics and the responses of associated fauna. In turn, ecosystem characteristics such as resource composition, productivity and distribution determine the individual and population nutrition of the herbivore (Clutton-Brock & Albon 1989; Langvatn et al. 1996). The spatial distributions of different plant communities can thus have a strong influence on the foraging behaviour of free-ranging herbivores (Senft et al. 1987; Cougenhour 1991; Gordon & Illius 1992; Gross et al. 1995; Owens et al. 1995). Although such differences are known to be important, for many herbivores and vegetation types the interrelationships between vegetation pattern, foraging behaviour and plant community use are still poorly understood, leading to poorly developed management strategies for agricultural or other land use objectives (Bailey, Dumont & Wallis de Vries 1998). Many studies have focused primarily on the preferred vegetation ‘patch’ (Illius, Clarke & Hodgson 1992; Jiang & Hudson 1993a; Langvatn & Hanley 1993), yet the more fragmented or widely distributed that preferred vegetation is, the greater the impacts on the surrounding vegetation are likely to be, which has important implications for management. In semi-natural vegetation, food is generally distributed in variably spaced patches of different size and quality (reflected by nutrient status or biomass, for example) which can also change seasonally. Thus, the management of semi-natural vegetation by grazing animals depends fundamentally on an adequate understanding of the interrelationships between the animal and the complexities of the vegetation (Hanley 1997; Milne et al. 1998). Furthermore, the greater the impact of vegetation pattern on the foraging behaviour of a particular species, the greater may be the divergence in impact of that species on the vegetation between areas which have different distributions of the same vegetation types (Cougenhour 1991; Clarke, Welch & Gordon 1995b; Bailey, Dumont & Wallis de Vries 1998).

The system selected for study here represents a fragmented example of an internationally important vegetation type, heather moorland (Gimingham 1972), grazed by two widespread herbivore species, Scottish Blackface sheep and red deer Cervus elaphus L. Heather moorland (dominated by Calluna vulgaris L. Hull) is found in varying amounts throughout north-western Europe (de Smidt 1995), yet there has been considerable fragmentation and contraction of its range in many areas, partly as a result of grazing impacts (Anderson & Yalden 1981; Froment 1981; Bokdam 1995). Northern Britain represents a major stronghold of this community, with heather-dominated vegetation still covering about 30% of the land area of Scotland, for example (Gimingham 1972; Gauld et al. 1991; Thompson, Hester & Usher 1995). However, even here, there has been widespread grazing-related fragmentation, loss and conversion to grass-dominated vegetation (Sydes & Miller 1991; National Countryside Monitoring Scheme 1993; Staines, Balharry & Welch 1995; Tudor & Mackay 1995).

Domestic sheep and wild red deer range freely in large numbers over extensive areas of heather-dominated vegetation in northern Britain, and are also found within this vegetation type in other parts of Europe (Ahlen 1965; Schmidt 1983; Kottman, Schwöppe & Willers 1985; Staines, Balharry & Welch 1995). Both these herbivores are mixed strategy feeders, with preferences for grass species such as Agrostis/Festuca over woody species such as heather (Mitchell, Staines & Welch 1977; Hofmann 1989). Grass patches within heather moorland will therefore act as attractants, leading to increased pressure on the heather immediately adjacent to the areas of grass (Grant et al. 1978; Clarke, Welch & Gordon 1995a,b), thus affecting the rates and patterns of any subsequent heather decline. In many such areas the impacts of these herbivores on the vegetation have been marked, and the need for more appropriate grazing management has long been recognized (e.g. Anderson & Yalden 1981; Sydes & Miller 1991; Milne et al. 1998). Within the UK, for example, heather management through reduced grazing forms one of the requirements of management within the government's Environmentally Designated Areas Scheme (Henderson et al. 1994; Nolan et al. 1994). However, monitoring of the success of this scheme has already revealed some of the inadequacies of generalized herbivore stocking rate prescriptions that do not specifically take vegetation pattern into account (e.g. Nolan et al. 1994). There is a clear need to improve our understanding of how these herbivores interact with vegetation pattern in heather-dominated systems to enable robust predictions to be made about their likely impacts, and to guide current and future herbivore management prescriptions.

The main objective of this research, therefore, was to test how foraging behaviour and grass patch selection by sheep and red deer was affected by the size and distribution of natural grass patches within an area of fragmenting heather moorland. Naturally fragmenting moorland contains a variety of sizes, shapes and distributions of patches of grass, which we hypothesized would differentially affect the foraging behaviour of sheep and red deer. This experiment was set up to complement the work of Clarke, Welch & Gordon (1995a,b), who artificially created large, uniformly distributed squares of grass within heather and compared the foraging behaviour of sheep and red deer when presented with one of three different patch sizes (one size of patch per plot) within a heather mosaic.

As sheep and red deer commonly graze within the same upland areas (Clutton-Brock & Albon 1989), and both species seem to occupy similar foraging niches in these areas (Hunter 1962; Clutton-Brock, Guinness & Albon 1982; Osborne 1984), the second objective of this research (not tested by Clarke, Welch & Gordon 1995a) was to quantify the extent of interaction and feeding overlap between sheep and red deer when grazing together under controlled experimental conditions. This would enable us to test current hypotheses about differences in their observed feeding strategies (Osborne 1984; Clutton-Brock, Iason & Guinness 1987; Illius & Gordon 1987). Circumstantial evidence from field observations (Osborne 1984) suggested, for example, that red deer may feed more on heather in the presence of sheep than when alone. This may have been an indirect effect caused by sheep grazing the grass to a lower height than that required to sustain red deer; alternatively direct behavioural factors, such as avoidance of the other herbivore species, may have been important in this particular study. If such interactions occur between sheep and red deer, their impacts on heather/grass mosaics could differ greatly according to whether both sheep and red deer or just one of the species are present. Any such differences would clearly affect the design of appropriate management strategies for this vegetation type, depending on the herbivore species using the area.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Treatments

Six 1-ha plots were fenced within an area of building/ mature heather-dominated moorland (Gimingham 1972) at the MLURI's Glensaugh Research Station in north-east Scotland (National Grid reference NO677782). The heather moorland was dissected by a natural mosaic of numerous patches of semi-natural grassland; this fragmentation was the result of many years of grazing (Nicholson & Robertson 1958; S. Grant, personal communication). The area of grass patches totalled ≈ 15% of each plot (Fig. 1), comprising predominantly Agrostis spp. (60%), Deschampsia flexuosa (20%), Festuca spp. and a range of forbs. The heather-dominated matrix (≈ 85% of each plot, Fig. 1) comprised over 95% cover of Calluna vulgaris; the only other species greater than 0·5% cover was Vaccinium myrtillus L. (mean < 5% cover per plot). The hillside was west-north-west facing at an altitude of 200–250 m.

image

Figure 1. Layout of the experimental plots and close-up of one plot showing the vegetation mosaic.

Download figure to PowerPoint

The plots were arranged in two replicate blocks of three, with one plot in each block containing either sheep only (12 per plot), red deer only (8 per plot) or sheep (6) and red deer (4) together. Two smaller (replicate) plots were fenced as ungrazed controls. Numbers of animals put on the plots were selected to give a similar total offtake of herbage per plot. Numbers were also limited by the availability of grass in the plots, to avoid excessive consumption of heather and loss of condition of the animals. The decision made on the balance between numbers of animals per plot and time spent grazing also considered the effects of group size on the behaviour of sheep and red deer. Clarke, Welch & Gordon (1995a; unpublished data) found that within heather/grass mosaics, the mean group size of red deer was three, whereas the mean group size of Scottish Blackface sheep was five to seven animals, but sheep group size declined strongly with reducing size of grass patches within the mosaic. Penning et al. (1993) also suggested that minimum group sizes of three or four sheep (Scottish Halfbred) were adequate for studies of grazing behaviour. Thus, by keeping the minimum group size above three for deer and five for sheep, we aimed to minimize any effects of group size on their behaviour. Similarly, our choice of plot size aimed to minimize the impact of enclosure as far as possible, as discussed later.

The deer:sheep ratio used, 1:1·5, was calculated on the basis of their predicted intake (Milne et al. 1978) and the live weights of the animals used. The red deer used each year were yearling hinds (mean age 14 months, mean live weight 52 kg) and the sheep were yearling Scottish Blackface ewes (mean live weight 40 kg). All animals were born and reared at the farm and had similar nutritional histories. Yearling animals were selected for the following reason: as the red deer were farm reared, the yearling weights more closely matched those of mature wild hinds (Blaxter et al. 1974; Mitchell, Staines & Welch 1977). As increasing maturity does not appear to change diet selection, but intake does generally increase as a function of body weight (Clutton-Brock, Guinness & Albon 1982; Gillingham, Parker & Hanley 1997), selecting appropriate body weights above age was considered the most desirable approach for this experiment. Yearling sheep are already considered close to maturity (D. Nelson, personal communication). It is not known whether farmed red deer are more or less tolerant of other herbivores than wild red deer, but the animals used had been brought up separately and as such should have no reason to be more or less tolerant of each other's presence (P. Goddard, personal communication) than free-grazing sheep and wild deer on the open hill.

The experiment was run from 1991 to 1996, with animals placed in the plots for an 8-week experimental period each year. In 1991, the grazing period was from mid-August to mid-October (autumn grazing), in 1992 from 1 June to 1 August (summer grazing), and in 1993–95 the 8-week grazing periods were split into two 4-week blocks, the first in June and the second from mid-August to mid-September (split summer grazing). In 1996 the animals were placed into the plots for only the first 4-week grazing period in June, to collect supplementary behavioural data (see below). The animals were intensively observed in 1991, 1992 and 1993, with additional measurements taken in 1996 as detailed in the next section. Vegetation measurements were made every year throughout the 6 years of the experiment and are reported elsewhere (Hester & Baillie 1998).

Animal measurements

Observations of all the animals were made by telescope during daylight hours (04.00–22.00 hours in mid summer, 07.30–18.30 hours by mid October), from a distance of ≈ 0·5 km. From previous work on 24-hour grazing patterns (Hester et al. 1996), it was known that most grazing time by sheep and red deer takes place during daylight hours, although red deer tend to spend a slightly greater proportion of darkness hours grazing than sheep. From this and other work we considered this daytime sample to be sufficiently representative of the habitat use by these two herbivores (Clutton-Brock, Guinness & Albon 1982; Penning, Rook & Orr 1991; Hester et al. 1996).

Each animal was individually marked to enable identification: red deer with colour-coded collars, and sheep with large numbers spray-painted on their body sides, plus colour-coded tapes on their horns. Data were collected (all animals) using two different observation methods. The first used spot scans (Altmann 1974), recording and mapping the behaviour, exact location and orientation of each animal every half hour during daylight hours. The behavioural activities recorded were: grazing (head down biting vegetation, ± simultaneous moving), standing (inactive or ruminating), lying (inactive or ruminating), moving (head up), interaction (aggressive or passive), other (drinking, scratching, etc.). The second method used was to observe a single focal animal continuously for 15 min, recording (by dictogram) the timing and location of each behaviour, and marking the route taken on a scaled map of the plot. Both spot-scans and 15-min focal animal observations were made on two (different) days per week (unless visibility was too poor) through each experimental period; the former in 1991, 1992 and 1993; the latter in 1991 and 1996. Bite rates of individual animals on grass or heather were measured in 1996 by observing an actively grazing animal for 1 min and counting the number of bites taken. As all animals used had previous experience of upland grassland but not heather-dominated vegetation, the first week of each experimental period was considered a settling-in period for the animals to become accustomed to the plots. Therefore, no data from any first week was used in the analyses presented here. One week was considered to be sufficient time, both from other work (Gluesing & Balph 1980; Scott, Banner & Provenza 1996) and from data collected in 1991 (A. Hester & G. Baillie, unpublished data).

All grass patch sizes were measured and categorized into three size classes which were related to the use by the animals: ‘small’ was defined as 1–6 m2, where no more than one animal was observed grazing at any one time; ‘medium’ was defined as 6–30 m2, where several animals were sometimes observed to be grazing but never the whole group; and ‘large’ was all patches over 30 m2, where all animals could graze together on the same patch.

Faecal deposition patterns by sheep and red deer can also impact significantly upon the vegetation, therefore a subsample of two of the single species plots (one plot per species) was sampled in 1993 to examine this question. A random sample of 120, 50 × 50-cm quadrats was taken in each plot. For quadrats on grass, the size of the patch was also noted; for quadrats in heather, quadrats were classified as adjacent to grass (< 1 m) or far (> 1 m) from grass. Within each quadrat, the number of faecal pellets or the volume of clumps (some sheep faeces) of faeces present after 4 weeks grazing were recorded. Ten, randomly collected samples (pellets and clumps) per plot were dried and weighed, giving a mean weight per volume and/or per pellet. This was used to convert numbers of pellets or clumps to mean weights of faeces deposited.

Data analysis

All analyses of behavioural data were carried out using the Residual Maximum Likelihood (REML) method (Genstat 5 Committee 1993). Scan data was averaged per plot per week before analysis (i) because individual animal behaviour within a plot was not considered independent, and (ii) to remove between-day variation as it was not relevant to the aims of this paper. In view of the natural variation between plots in the vegetation mosaics, plot was included as a random factor in all the analyses. Percentage data were angular-transformed before analysis. Fifteen-min individual animal behavioural observation data were also analysed (log-transformed) using REML, using individual animal measurements, with plot and animal number as random factors. Faecal data were also analysed using REML on log-transformed data. Data in all tables have been back-transformed for clarity.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Time spent on each activity: half-hourly scan data

Figure 2 shows the mean percentage of daylight hours spent by sheep and red deer in each activity. There were no differences in total time spent in each activity between sheep grazing alone or with red deer, or between red deer grazing alone or with sheep. This was the case for all scan data and all 15-min observation data presented here; therefore, all these data have been amalgamated for each species. Both sheep and red deer spent more daylight hours grazing than undertaking other activities, followed by lying ruminating and lying inactive. Sheep consistently spent significantly more daylight time grazing than red deer (P < 0·001). Red deer spent more daylight time lying down (ruminating plus inactive), moving and standing (inactive) than did sheep (P < 0·001, P < 0·05, P < 0·05, respectively). Very few animal interactions were observed (< 0·2% of all observations) between sheep or red deer, and no interactions between different species were observed. For both sheep and red deer, almost all of the interactions observed were aggressive (biting/kicking/chasing) and occurred either on small grass patches or on heather. There were insufficient interactions to analyse these data statistically.

image

Figure 2. Percentage daylight hours spent by sheep or red deer on each activity (mean ± SEM, n = 288).

Download figure to PowerPoint

For all subsequent analyses of activities, the following three categories, grazing, lying (ruminating plus inactive) and moving, have been selected as the most frequent and/or the most important in relation to their effects on the vegetation.

Time spent grazing

Overall, both sheep and red deer spent about half of their daylight grazing time on heather and half on grass (Table 1). Both herbivores spent significantly more time grazing heather in 1991 (autumn grazing) than in 1992 (summer grazing) or 1993 (split summer grazing) (P < 0·01). The mean areas covered by the different sizes of grass patch and heather are given in Table 1. Sheep consistently spent more time grazing small grass patches than did red deer (P < 0·01). There were also significant differences between years (P < 0·01). In 1992 and 1993, sheep spent more time grazing on small patches than on medium (1992: P < 0·01; 1993: P < 0·05) or large patches (1992: P < 0·05; 1993: NS). But in these years there were no significant differences in patch size use for red deer. In 1991 both sheep and red deer spent less time overall on small grass patches than in the other two years (P < 0·01). Therefore, in 1991 red deer spent significantly less time on small grass patches than medium or large patches (P < 0·01), but there were no significant patch size differences for sheep. Converting these patch use results to selection indices, i.e. by dividing percentage grazing time by the proportional areas of each grass patch size within each plot, gave similar results, i.e. consistent preferences for small grass patches by sheep and no consistent patch size preferences by red deer. There were no significant changes in daylight time spent grazing for either sheep or red deer through the weeks of each experimental period, but there was a trend for both species to increase their proportional grazing time on heather slightly as grass sward heights decreased, especially in 1991 (red deer: P < 0·05; sheep: NS). Week was therefore kept in the analysis as a random factor (covariate).

Table 1.  Percentages of daylight time spent (a) grazing, (b) lying, and (c) moving, by sheep and red deer on heather or grass (REML adjusted back-transformed means), plus mean areas (m2 ± SEM) of grass and heather within the 1-ha plots. For all data comparisons within each behaviour category, both between and within years, numbers with the same superscript are not significantly different at P < 0·05
 Small grassMedium grassLarge grassHeather
  • (a) SED*(within species) = 0·0274; SED (between species) = 0·0271; residual m.s. d.f. = 6.

  • (b) SED*(within species) = 0·0195; SED (between species) = 0·0207; residual m.s. d.f. = 6.

  • (c) SED*(within species) = 0·0068; SED (between species) = 0·0101; residual m.s. d.f. = 6.

  • *

    SEDs are not back-transformed.

(a) Grazing (n = 24)
1991
Sheep7·6bc10·2cd8·0bc36·0 h
Red deer3·3a10·3cd7·0bc32·7gh
1992
Sheep15·3e8·2bc6·8bc28·1fg
Red deer9·4bcd6·8bc5·9ab24·8f
1993
Sheep15·0e10·3cd12·4de22·7f
Red deer7·2bc7·4bc6·8bc23·9f
(b) Lying (n = 72)
Sheep9·3e6·0cd3·8b7·4cde
Red deer8·2de4·5bc1·5a24·2f
(c) Moving (n = 72)
Sheep0·14a0·10a0·07a0·80c
Red deer0·06a0·06a0·03a2·74d
Mean areas (± SEM) of grass and heather (no of plots per species = 4):Plots grazed by:
Sheep280 ± 42574 ± 148493 ± 1558653 ± 284
Red deer233 ± 19601 ± 162641 ± 1268524 ± 286

Time spent lying (ruminating plus inactive) and moving

Overall, red deer spent almost twice as much daylight time lying in heather than in grass (P < 0·01); this was the opposite for sheep (Table 1). Both red deer and sheep spent more time lying in small grass patches than in medium or large grass patches (P < 0·05). There were no significant differences between years or weeks.

Both sheep and red deer spent most of their moving time within heather rather than grass, and red deer spent much more time moving over heather than did sheep (P < 0·001) (Table 1). There were no significant differences in time spent moving over any of the different sizes of grass patch for either sheep or red deer. These effects were consistent between years, even though the magnitudes of some of the differences changed.

Faecal distribution within grass and heather

Faecal deposition by sheep and red deer differed greatly in relation to the vegetation (Table 2). Sheep deposited 73% of pellets within grass patches, whereas red deer deposited 63% of pellets within heather (P < 0·001). In view of the different proportions by area of grass and heather in the plots, mean weight of faeces per m2 was used as the unit for analysis. There were no differences in faecal deposition between heather at the edge of grass patches or further away, therefore data for all heather quadrats have been combined. Sheep consistently deposited a greater weight of faeces than red deer overall (P < 0·05) and within all sizes of grass patch (P < 0·01). Sheep also deposited a lesser weight of faeces per m2 within heather than within grass (P < 0·001). The weights of red deer faeces found were not significantly different between any grass patch sizes or heather. Variability between quadrats was high (Table 2).

Table 2.  Faecal pellet distribution by sheep and red deer over 4 weeks (REML adjusted back-transformed means), n = 120. For each measure, numbers with the same superscript are not significantly different at P < 0·05
 SheepRed deer
  • % pellets: SED*(within species) = 0·5046; SED (between species) = 0·5092; residual m.s. d.f. = 232.

  • Mean weight: SED*(within species) = 0·3214; SED (between species) = 0·3711; residual m.s. d.f. = 236.

  • *

    SEDs are not back-transformed.

% pellets on grass74·2b36·6a
% pellets on heather25·8a63·4b
Mean weight of one pellet (g)0·194a0·213a
Mean weight of faeces deposited (g m–2)
Small grass14·3bc7·3ab
Medium grass31·7c10·2ab
Large grass13·3bc3·2a
Heather3·0a5·6ab

Foraging bouts and movement rates: 15-min continuous observations

There were no significant differences between 1991 and 1996 for any measurements made in the 15-min observations, therefore both years of data were amalgamated. For none of the measurements made was there sufficient data to analyse different sizes of grass patch separately, so all analyses consider simply grass, heather, or both vegetation types together. Data were split into behavioural subsets for analysis as follows:

Grazing bouts

A grazing bout was defined as a period of continuous grazing (i.e. head not up for more than 30 s at a time, plus or minus simultaneous moving). Mean bout length (under 15 min) was calculated using the subset of bouts which started and finished within the 15 min observation period (Table 3). Bout lengths of sheep and red deer were not significantly different, therefore the data are combined in Table 3. Grazing bouts that included both grass and heather were significantly longer than those spent on one vegetation type only (P < 0·001). Bouts on grass also became longer as the experiment progressed (P < 0·05). Movement rates (distance/bout duration) during these bouts varied greatly, differing significantly between vegetation types (P < 0·01) but not between animal species (Table 3). Both species moved most slowly when grazing on grass for the whole bout, but this difference was only significant for red deer (P < 0·01). There were no consistent significant differences in movement rates between weeks.

Table 3.  Grazing bouts and movement rates (metres per min) by sheep and red deer (a) during discrete grazing bouts of under 15 min, (b) during grazing bouts of over 15 min, and (c) whilst actively foraging for 15 min (REML adjusted back-transformed means). Sample numbers in parentheses. Within each category, values with the same superscript are not significantly different at P < 0·05
 GrassHeatherGrass & heather
  • (a) Bout length: SED*(within weeks) = 0·1516; SED (between weeks) = 0·1561; residual m.s. d.f. = 42.

  • Movement rates: SED*(within species) = 0·4002; SED (between species) = 0·4401; residual m.s. d.f. = 28.

  • (b) SED*(within species) = 0·1724; SED (between species) = 0·1627; residual m.s. d.f. = 32.

  • (c) SED*(within species) = 0·0859; SED (between species) = 0·0843; residual m.s. d.f. = 52.

  • *

    SEDs are not back-transformed.

(a) Mean lengths of grazing bouts (seconds) which both started and finished within a 15-min observation period (sheep & deer combined)
Weeks 2–350a (55)47a (57)175c (29)
Weeks 4–558ab (38)71ab (48)214c (25)
Weeks 6–7106bc (11)72ab (7)186c (14)
(a) Movement rates during (above) grazing bouts under 15 min
Sheep1·12a (15)4·06b (7)2·40ab (7)
Red deer2·37ab (9)2·99b (4)2·42ab (6)
(b) Movement rates during continuous grazing bouts of over 15 min
Sheep0·80a (3)0·94a (9)1·14a (29)
Red deer0·75a (3)1·42a (4)1·16a (16)
(c) Movement rates whilst actively foraging for 15 min (grass and heather combined)
Sheepweeks 2–3 2·71bc (26)
 weeks 4–5 2·28b (23)
 weeks 6–7 1·45a (6)
Red deerweeks 2–3 3·50c (18)
 weeks 4–5 2·96bc (22)
 weeks 6–7 2·95bc (7)

Grazing bouts that lasted the whole 15 min period were analysed separately to give an indication of the maximum intensity at which the animals were willing to feed in a specific area or vegetation type. When continuously grazing for 15 min, there were no significant differences in rates of movement between species, vegetation, or time, with both sheep and red deer moving at about 1 metre per min (Table 3). However, there was a trend for both species to move more slowly over grass than heather, which was significant (P < 0·05) when analysed as mean distance moved over 15 min (rather than per min as presented in Table 3). Only 10% of animals observed spent the whole of the 15 min grazing on grass, 20% on heather, and 70% over a mixture of grass and heather. Mean numbers of vegetation boundaries crossed (1·2 per 15 min) did not differ significantly between species or time.

15-min active foraging

This was defined as where an animal was grazing at the start and end of the 15 min period, but may have briefly interrupted its grazing by spells of moving (head up) or standing, within the 15 min observation period. As this ‘active foraging’ measure includes search time, it gives an indication of the response of animals to microscale food dispersion; thus a large difference between this movement rate and ‘continuous grazing’ suggests a sparsely dispersed resource (Bailey et al. 1996; Hanley 1997). As 95% of animals observed as continuously foraging moved over both grass and heather within the 15 min observed, community was not included as a factor in this analysis. Red deer moved faster whilst actively foraging for 15 min than did sheep (P < 0·01, Table 3), and this difference was greatest at the end of the experiment (P < 0·001) when grass heights were at their lowest. Sheep movement rates decreased significantly through time (P < 0·01). Red deer movement rates were also fastest at the start of the experimental periods, but this difference was not significant. Mean numbers of transitions over grass:heather boundaries were greater for red deer than sheep (6·2 and 3·8 per 15 min, respectively, P < 0·01). There were no significant differences between weeks in the numbers of boundaries crossed by sheep or red deer.

Bite rates

There was a significant difference in bite rates between grass and heather (Table 4), with both species having significantly faster bite rates on grass (P < 0·001). Bite rates changed through time: for sheep, bite rates on grass were slowest at the start of the experimental period (P < 0·01), but bite rates on heather did not change. For red deer, bite rates on grass did not change, but bite rates on heather were fastest at the start of the experimental period (P < 0·01). Bite rates by sheep on grass were therefore slower than those by red deer at the start of the experimental period only (P < 0·05), and red deer bite rates on heather were slower than those of sheep after week 2 of the experimental period (P < 0·01).

Table 4.  Bite rates (number of bites per minute) of sheep and red deer on grass or on heather (REML adjusted back-transformed means). Sample numbers in parentheses. Values with the same superscript are not significantly different at P < 0·05
GrassHeather
 Week 2Week 3Week 4Week 2Week 3Week 4
  • SED*(within species) = 0·0373; SED (within veg) = 0·0388; SED (within time) = 0·0382; residual m.s. d.f. = 37.

  • *

    SEDs are not back-transformed.

Sheep44cd (4)56e (9)54e (9)37bc (8)36b (15)37b (15)
Red deer53e (7)48de (10)55e (7)40bcd (2)28a (8)29a (12)

Animal orientation

The orientation of animals on sloping ground strongly affects the spatial distribution of their impact (Hester & Baillie 1998), particularly when grazing and moving. Whilst grazing (Table 5), both sheep and red deer spent most time facing across the line of slope (P < 0·001) and the least time facing downhill (P < 0·001). This was the case on both grass and heather. In addition, sheep grazed facing uphill more often than did red deer (P < 0·05). When moving, there were no species or community differences; both sheep and red deer spent 53% of the time walking across the line of slope, and only 22 or 25% of the time walking uphill or downhill, respectively (P < 0·001).

Table 5.  Percentage of grazing time spent facing uphill, downhill or across the slope by sheep or red deer (REML adjusted back-transformed means). Within each category, values with the same superscript are not significantly different at P < 0·05
 UphillAcross slopeDownhill
  • SED*(within species) = 0·0409; SED (between species) = 0·0340; residual m.s. d.f. = 36.

  • *

    SEDs are not back-transformed.

Sheep27·6c60·8d6·6a
Red deer21·2b65·9d4·8a

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Grazing preferences between grass and heather

Both sheep and red deer spent similar amounts of time grazing grass and heather within this highly fragmented vegetation mosaic, even though grass only covered 15% of the area of each plot (cf. Clarke, Welch & Gordon 1995a; Hester et al. 1996). The known preference by these species for grass over heather, especially in the growing season (Milne et al. 1978; Osborne 1984; Armstrong et al. 1997b), was clearly apparent. Whilst grazing grass, sheep and red deer moved more slowly, grazing bouts were longer and the bite rates of both species were faster than on heather. However, we propose that the degree of fragmentation of vegetation will have a strong influence on the proportion of time spent by sheep grazing grass and heather, but not necessarily by red deer, for the following reason: when grazing the mosaic of small grass patches in our experiment, both sheep and red deer spent about half their grazing time on grass; comparing this with larger patches of grass, from the results of Clarke, Welch & Gordon (1995a) (20% grass, 80% heather in their plots), the time spent grazing grass by sheep increased significantly as grass patch size increased (from 55% on 400 m2 patches to 90% time grazing grass on 5000 m2 patches), whereas red deer grazing behaviour was unaffected by grass patch size. It is likely that sheep, being smaller, find it more difficult to move through mature heather than do red deer and therefore if grass patches are large, there will be less need to move as frequently through heather. The greater distances moved by red deer than sheep whilst actively foraging across grass and heather support this assertion. In addition, red deer have been shown to digest heather more efficiently than do sheep (Milne et al. 1978). We conclude from this that red deer appear to select a mixed grass/heather diet which is relatively unaffected by the pattern of vegetation, whereas the strong preference for grass by sheep only appears to be reduced as the costs of foraging for grass increase under increasing vegetation fragmentation (more, smaller grass patches within the heather), i.e. when the animals become ‘encounter-limited’ rather than ‘food-processing limited’ (Spalinger & Hobbs 1992; Farnsworth & Illius 1998). This has important implications for management, in that the more fragmented the vegetation, the greater is likely to be the proportional intake of heather by sheep.

For the reasons given above, we expected that a decline in grass availability through the experimental periods (to swards as low as 3–4 cm: see Hester & Baillie 1998) would affect foraging behaviour, particularly of red deer, such as to increase the proportion of their time spent grazing and moving through heather (Colquhoun 1970; Osborne 1984; Illius & Gordon 1987; Clarke, Welch & Gordon 1995a). In both 1991 and 1992 animals spent slightly more time on the heather as grass heights fell, but no major dietary shifts were apparent. Faecal cuticle analysis (P. Cuartas, J. Gordon, A.J. Hester & I.R Hulbert, unpublished data) suggested that there was about a 10% increase in the proportion of heather in the diet of both sheep and red deer by the end of the experimental period in 1992. When grazing grass, individual bout lengths also became longer for both sheep and red deer, and sheep bite rates on grass increased as the grass heights fell. In addition, both sheep and red deer moved faster whilst actively foraging at the start of the experiment than subsequently. This concurs with the suggestions of Hanley (1982) that animals move faster whilst foraging on food that is either abundant or of a higher quality (i.e. when less search effort is required). Similar responses of elk and white-tailed deer to forage depletion were found by McCorquodale (1993) and Kohlmann & Risenhoover (1994) within woody/herbaceous vegetation mixes. This suggests that the animals attempted to compensate for declining resource availability by reducing movement rates and increasing their feeding rates (i.e. increases in feeding time and bite rates) (cf. Alados & Escos 1987; Sorensen & Taylor 1995).

This experiment did not specifically consider the effects of time of year on vegetation use by sheep and red deer, but the greater amounts of time spent grazing heather in the autumn grazing treatment (1991) are considered to reflect the seasonal shift towards heather from grass in autumn and winter, as the relative availability and nutritive value of grass declines (Bakker et al. 1983; Armstrong & Milne 1995; Armstrong et al. 1997a). Such a seasonal change in grass/heather use has been observed on the open hill with red deer (Colquhoun 1970; Charles, McGowan & East 1977; Gordon 1989).

Enclosure of animals (especially deer) in plots can influence their behaviour. Therefore, the animals were observed closely for any indications of behavioural responses to enclosure that might reduce the applicability of our results to open hill situations. After the first few days of each experimental period, when the animals explored the plots, we found no evidence of any major impacts on their foraging behaviour. The animals moved freely throughout the plots with no obvious influence of fenced edges; boundary pacing was relatively infrequent and nearest neighbour distances of both sheep and red deer were comparable to those found in a range of other studies of free-ranging sheep and red deer (Clutton-Brock, Guinness & Albon 1982; Thouless 1990; Cruickshank 1992; Lynch, Hinch & Adams 1992; Goddard, Gordon & Hamilton 1996; A.J. Hester & G.J. Baillie, unpublished data). Thus, we feel that our results should also be applicable to free-grazing situations.

The influence of vegetation pattern on animal distribution and behaviour

Grass patch size differentially affected choice of grazing locations by sheep and red deer. In view of the apparent preference by sheep for small grass patches and the lack of consistent patch size preferences by red deer, we would expect their impact on the vegetation to differ. Since the ratio of patch edge to patch area decreases with increasing size of grass patch, then the impact of a given herbivore density on edge heather would be expected to be greater for larger patches than for smaller ones. This being the case, we would not expect the preference for smaller patches by sheep to result in a greater impact on edge heather around these patches but we might expect the more even use of all patch sizes by red deer to result in increased heather utilization around larger patches than smaller patches. Our data on heather utilization (Hester & Baillie 1998) confirmed that there was an increase in the utilization of patch edge heather by red deer with increasing size of grass patch, and there were no consistent patch size differences in utilization of patch-edge heather by sheep. The implications of these findings are that in areas of fragmenting heather with a range of different sizes of grass patches, damage to heather at the edges of grass patches will increase with increasing size of grass patch under grazing by red deer, but not by sheep, leading to faster increases in size of larger grass patches under heavy grazing by deer. Differences in grouping behaviour between sheep and red deer were observed in our experiment, and we suggest that this may have influenced the preference for small grass patches by sheep and the more even use of different sizes of patch by red deer. Sheep often grazed singly and were even observed to ‘defend’ small grass patches from other sheep (A.J. Hester, personal observation). Red deer, on the other hand, grazed more frequently in a group than singly, thus it was rare to see all the individual animals in any one group grazing on grass at the same time (as many patches were too small to accommodate all animals). Thus, one ‘cost’ for red deer of grouping within highly fragmented vegetation is that some individuals may be forced onto the less preferred vegetation (cf. Valone 1993; Ruxton 1995). As sheep frequently grazed alone, this ‘cost’ of grouping was less likely to occur.

Heather, especially when mature, is much less resilient to damage than grass (Gimingham 1972; Palmer 1997) and in addition to grazing impacts, animals lying in the heather can cause significant damage through stem breakage. From our results, we would thus predict greater physical damage to heather by red deer than by sheep, due to the greater proportions of time spent lying within heather by red deer. Over the 6 years of this experiment we observed significant damage and even death of heather plants where the deer regularly lay. This preference for lying within heather presumably reflects the greater sensitivity of red deer than sheep to wind-chill (Staines 1976), as heather canopies significantly reduce windspeed (Grace & Easterbee 1979; Wallace et al. 1984).

The greater amount of time spent moving through heather by red deer than by sheep is also likely to result in more widespread physical (trampling) damage to the heather by red deer. From observation, red deer were also much more likely to roam freely through the heather without using existing paths, unlike sheep, which would spread the distribution of heather damage by deer more widely than that by sheep. Vegetation measurements relating to path use reflected this observed difference between sheep and red deer (Hester & Baillie 1998; Oom & Hester 1999).

Differences in faecal distribution are also likely to contribute to differential effects of the two herbivores on the vegetation. As animals tend to defecate when getting up after a period of rumination (Leuthold 1977; Taylor et al. 1987), the concentration of sheep faeces within grass patches rather than heather, unlike that of red deer (widely dispersed), reflected the different preferences of these herbivores for lying within grass or heather, as discussed above. A concentration of faeces within grass patches could lead to increased rates of nutrient cycling within the soils under grass, accelerating the succession towards more mesophytic soil conditions and vegetation assemblages (Bakker et al. 1983; Welch 1985; Marrs, Rizand & Harrison 1989; Hester & Baillie 1998), which could reduce the likelihood of any subsequent return to heather even if grazing pressures were decreased. If this is the case, the likely success of any heather regeneration scheme may decline more rapidly with time (since heather loss) under sheep grazing than under red deer grazing, but this assertion has not been tested to date.

Interactions between sheep and red deer

When examining the interactions between sheep and red deer, we considered it important to make a distinction between those interactions that would change the overall impact of either herbivore on the vegetation, and those that would not. Within our experimental set-up, the observed effects of one species on another fell into the latter category. For example, although the two species generally avoided each other (usually > 30 m apart), they freely used the same patches of vegetation, but almost always at different times of day (in only 0·4% of all observations were individuals of each species on the same grass patch). Therefore, we found no measurable difference in their overall use of the vegetation when grazing alone or together. The implications of this finding are that the presence or absence of one species will not greatly affect the overall impact of the other species on the vegetation. Our results do not support the suggestions of Colquhoun (1970) and Osborne (1984) that when both herbivores are present in an area, red deer spend more time on the less nutritious vegetation due to feeding competition with sheep. Their suggestion relates to the hypothesis that sheep, having smaller mouths and smaller daily herbage intake requirements than red deer (Milne et al. 1978) are better able to sustain themselves on shorter swards than are red deer (Illius & Gordon 1987). Within our grass:heather mix there is no evidence of such competitive exclusion having taken place, even at the end of each experimental period when grass swards were as low as 3 cm. We suggest that the nutritional difference between Agrostis-dominated grassland and heather within the growing season is so great (see Armstrong et al. 1997b) that the red deer still choose to eat similar proportions of the two vegetation types, even when the grass swards become very short. When considering the wider implications of these results, it is important to note that the animals were within 1-ha fenced plots in this experiment, whereas on the open hill they might have the option of using different areas of ground, depending on the scale and pattern of the vegetation present in any area. However, even on the open hill, Osborne (1984) found significant overlaps in habitat use between sheep and red deer (especially hinds). Though it has been suggested by other studies that the presence of one herbivore can affect foraging by the other (e.g. Nolan & Connolly 1977; Kie 1996), many of these studies have been confounded with changes in herbivore densities.

The influence of slope

Previous work has identified that slope can significantly affect patterns of herbivore use and impact on the vegetation (e.g. Weaver & Dale 1978; Ganskopp 1987) and this was confirmed by the results of our experiment. Heather-dominated vegetation is frequently found on sloping ground; in the UK slopes of similar magnitude to the incline of our study site (17°) are common (Bakker et al. 1983; MLURI Soils Inventory database: see Langan, Paterson & Taylor 1996) and thus the impacts described here will be pertinent to many other areas of heather-dominated vegetation. The reluctance of animals to face downhill is as expected because it is probably the least comfortable or practical position on sloping ground (cf. Jiang & Hudson 1993b). Because of this, there was less grazing damage to the heather on downhill edges of grass patches (Hester & Baillie 1998), which has clear implications for the likely direction of fragmentation under grazing. When moving, both sheep and red deer tended to forage across the slope rather than up or down, presumably because of the higher energy costs of movement uphill rather than across-slope (Clapperton 1964; Brockway & Gessaman 1977; Parker, Robbins & Hanley 1984). This was mirrored by the preponderance of paths contouring the slope and increased damage from trampling at downhill edges of these paths (Hester & Baillie 1998; Oom & Hester 1999).

In conclusion, from the experimental results presented here, together with those of Clarke, Welch & Gordon (1995a) and Hester et al. (1996), it is clear that the degree of heather fragmentation can have a profound effect on how Scottish Blackface sheep, in particular, will forage within heather moorland. The implications are that differences in vegetation pattern between sites will not only change sheep foraging patterns but will also change the relative proportions of grass and heather consumed. Red deer, on the other hand, seem much less affected than sheep by changes in vegetation pattern. Access to large areas of grass, which could be natural or re-seeded, is therefore likely to totally change the habitat use and diet selection of sheep but not red deer. Such a difference has important implications for the management of upland grazings where sheep and/or red deer are present. Hester & Baillie (1998) show that the herbivore foraging differences recorded were indeed reflected in the spatial patterns of heather utilization over the duration of this experiment, even at these relatively low herbivore densities and short, mainly summer, grazing periods. These factors need to be taken into account when designing grazing management prescriptions for the maintenance of heather cover in different areas (Nolan et al. 1994; Armstrong et al. 1997a; Hester & Baillie 1998; Milne & Sibbald in press). For example, in areas of highly fragmented heather (i.e. containing many small grass patches), in view of the fact that both sheep and red deer are likely to graze widely and consume similar proportions of heather and grass, further fragmentation of the heather should be minimal under low densities of either species (except on sloping ground, see Hester & Baillie 1998). However, as stated earlier, the use of ‘resting areas’ within the heather by red deer could still lead to significant heather damage in very localized areas. Under high densities of either or both species, we concur with the predictions of Clarke, Welch & Gordon (1995a) that the utilization of heather will be heavy and widespread, but still heaviest at patch edges. In addition, we would expect a stronger focus on path creation and use (when moving within the heather) by sheep than by red deer, leading to further fragmentation of the heather (Hester & Baillie 1998; Oom & Hester 1999). Conversely, if only a few, large grass patches (natural or re-seeded areas) are present within the heather, the utilization of heather will be mostly confined to the few grass patch edges (and paths, by sheep). Thus, even under high densities of sheep and/or red deer, the majority of the heather should be relatively little utilized, but utilization at the (few) grass patch edges will be extremely heavy. Thus, there is a trade-off between severity and distribution of heather utilization under different herbivore densities according to the degree of existing fragmentation. This needs to be considered in the design of management strategies for different areas, in accordance with the specific aims for each area. Furthermore, in view of the impacts of vegetation pattern on the foraging behaviour of sheep (see also Clarke, Welch & Gordon 1995a), where there are only a few, large grass patches present, sheep are likely to consume considerably less heather overall than red deer. Thus, under those conditions we predict that, for a given density of herbivores, the overall impact of deer on the heather will be greater than that of sheep, and any control of herbivore numbers should take this species difference into account. Other impacts of these herbivores on the vegetation and their implications for heather moorland management are discussed in Hester & Baillie (1998).

In the wider context, the system studied here is relatively simple, with only two major vegetation types present. In most hill areas, vegetation assemblages are more complex and there is a need for more research on how the spatial pattern and distribution of different vegetation communities affects the behaviour of herbivores and their impacts on the vegetation (Armstrong et al. 1997a,b; Milne et al. 1998). The consequences for vegetation dynamics and herbivore management are likely to be substantial.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We are grateful to David Hirst, David Elston and Betty Duff of BioSS for their invaluable statistical advice. Jo Clarke and David Welch contributed useful advice at early stages of the experiment. Grateful thanks are extended to the many MLURI staff who helped with animal handling at different stages (special thanks to John Black and Duncan Murray), and to Isla Cruickshank who helped with animal observations and analysis in 1993. Angela Sibbald and John Milne gave helpful comments on the manuscript. This project was funded by SOAFD, the initial 2 years being part of the Joint Agriculture and Environment Programme (JAEP).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Ahlen, I. 1965 Studies on the red deer, Cervus elaphus L., in Scandinavia. III. Ecological investigations. Viltrevy, 3, 208–351. Swedish Sportsmen's Association, Stockholm.
  • Alados, C.L. & Escos, J. 1987 Relationships between movement rate, agonistic displacements and forage availability in Spanish ibexes (Capra pyrenaica). Biology of Behaviour, 12, 245255.
  • Altmann, J. 1974 Observational study of behaviour: sampling methods. Behaviour, 49, 222265.
  • Anderson, P. & Yalden, D.W. 1981 Increased sheep numbers and the loss of heather moorland in the Peak District, England. Biological Conservation, 20, 195213.
  • Armstrong, H.M., Gordon, I.J., Grant, S., Hutchings, N., Milne, J. & Sibbald, A.R. 1997a A model of the grazing of hill vegetation by sheep in the UK. I. The prediction of vegetation biomass. Journal of Applied Ecology, 34, 166185.
  • Armstrong, H.M., Gordon, I.J., Grant, S., Hutchings, N., Milne, J. & Sibbald, A.R. 1997b A model of the grazing of hill vegetation by sheep in the UK. II. The prediction of offtake by sheep. Journal of Applied Ecology, 34, 186206.
  • Armstrong, H.M. & Milne, J.A. 1995 The effects of grazing on vegetation species composition. Heaths and Moorland: Cultural Landscapes (eds D.B.A.Thompson, A.J.Hester & M.B.Usher), pp. 162173. HMSO, edinburgh.
  • Bailey, D.W., Dumont, B. & Wallis de Vries, M.F. 1998 Utilization of heterogeneous grasslands by domestic herbivores: theory to management. Anneles de Zootechnie., 47, 321333.
  • Bailey, D.W., Gross, J.E., Laca, E.A., Rittenhouse, L.R., Coughenour, M.B., Swift, D.M. & Sims, P.L. 1996 Mechanisms that result in large herbivore grazing distribution patterns. Journal of Range Management, 49, 386400.
  • Bakker, J.P., De Bie, S., Dallinga, J.H., Tjaden, P. & De Vries, Y. 1983 Sheep-grazing as a management tool for heathland conservation and regeneration in the Netherlands. Journal of Applied Ecology, 20, 541560.
  • Blaxter, K.L., Kay, R.N.B., Sharman, G.A.M., Cunningham, J.M.M. & Hamilton, W.J. 1974Farming the Red Deer. HMSO, London.
  • Bokdam, J. 1995 Cyclic succession and shifting mosaics in a cattle grazed heathlands in the Netherlands. Proceedings of the Fifth International Rangeland Congress, Utah 1995 (ed. N.E.West), pp. 5859. Society for Range Management, Denver, Colorado.
  • Brockway, J.M. & Gessaman, J.A. 1977 The energy cost of locomotion on the level and on gradients for the red deer (Cervus elaphus). Quarterly Journal of Experimental Physiology, 62, 333339.
  • Charles, W.N., McCowan, D. & East, K. 1977 Selection of upland swards by red deer (Cervus elaphus L.) on Rhum. Journal of Applied Ecology, 14, 5564.
  • Clapperton, J.L. 1964 The energy metabolism of sheep walking on the level and on gradients. British Journal of Nutrition, 18, 4754.
  • Clarke, J.L., Welch, D. & Gordon, I.J. 1995a. The influence of vegetation pattern on the grazing of heather moorland by red deer and sheep. I. The location of animals on grass/heather mosaics. Journal of Applied Ecology, 32, 166176.
  • Clarke, J.L., Welch, D. & Gordon, I.J. 1995b. The influence of vegetation pattern on the grazing of heather moorland by red deer and sheep. II. The impact on heather. Journal of Applied Ecology, 32, 177186.
  • Clutton-Brock, T.H. & Albon, S.D. 1989Red Deer in the Highlands: the Ecology of a Marginal Population. Blackwell Scientific Publications, Oxford.
  • Clutton-Brock, T.H., Guinness, F.E. & Albon, S.D. 1982Red Deer: Behaviour and Ecology of Two Sexes. University of Chicago Press, Chicago.
  • Clutton-Brock, T.H., Iason, G.R. & Guinness, F.E. 1987 Sexual segregation and density related changes in habitat use in male and female red deer (Cervus elaphus). Journal of Zoology, 211, 275289.
  • Colquhoun, I.R. 1970 The grazing ecology of red deer and blackface sheep in Perthshire, Scotland. PhD Thesis, University of Edinburgh.
  • Cougenhour, M.B. 1991 Spatial components of plant–herbivore interactions in pastoral, ranching, and native ungulate ecosystems. Journal of Range Management, 44, 530542.
  • Cruickshank, I. 1992 The influence of vegetation pattern on the distribution and social behaviour of blackface sheep. BSc Thesis, University of Aberdeen.
  • Farnsworth. K.D. & Illius, A.W. 1998 Optimal diet choice for large herbivores: an extended contingency model. Functional Ecology, 12, 7481.
  • Froment, A. 1981 Conservation of Calluno-Vaccinietum heathland in the Belgian Ardennes, an experimental approach. Vegetatio, 47, 193200.
  • Ganskopp, D. 1987 Slope use by cattle, feral horses, deer and bighorn sheep. Northwest Science, 61, 7481.
  • Gauld, J.H., Bell, J.S., Towers, W. & Miller, D.R. 1991The Measurement and Analysis of Land Cover Change in the Cairngorms. Report to the Scottish Office Department and the Scottish Office Agriculture and Fisheries Department. Macaulay Land Use Research Institute, Aberdeen.
  • Genstat, 5 Committee 1993Genstat 5 Release 3 Reference Manual. Clarendon Press, Oxford.
  • Gillingham, M.P., Parker, K.L. & Hanley, T.A. 1997 Forage intake by black-tailed deer in a natural environment: bout dynamics. Canadian Journal of Zoology, 75, 11181128.
  • Gimingham, C.H. 1972Ecology of Heathlands. Chapman & Hall, London.
  • Gluesing, E.A. & Balph, D.F. 1980 An aspect of feeding behavior and its importance to grazing systems. Journal of Range Management, 33, 426427.
  • Goddard, P.J., Gordon, I.J. & Hamilton, W.J. 1996 The effect of post-capture management strategy on the welfare and productivity of wild red deer (Cervus elaphus) hinds introduced to farming systems. Animal Science, 63, 315327.
  • Gordon, I.J. 1989 Vegetation community selection by ungulates on the Isle of Rhum. II. Vegetation community selection. Journal of Applied Ecology, 26, 5364.
  • Gordon, I.J. & Illius, A.W. 1988 Incisor arcade structure and diet selection in ruminants. Functional Ecology, 2, 1522.
  • Gordon, I.J. & Illius, A.W. 1992 Foraging strategy: monoculture to mosaic. Progress in Sheep and Goat Research (ed. A.W.Speedy), pp. 15377. CAB International, Wallingford.
  • Grace, J. & Easterbee, N. 1979 The natural shelter for red deer (Cervus elaphus) in a Scottish glen. Journal of Applied Ecology, 16, 3748.
  • Grant, S.A., Barthram, G.T., Lamb, W.I.C. & Milne, J.A. 1978 Effects of season and level of grazing on the utilization of heather by sheep. I. Responses of the sward. Journal of the British Grassland Society, 33, 289300.
  • Gross, J.E., Zank, C., Thompson Hobbs, N. & Spalinger, D.E. 1995 Movement rules for herbivores in spatially heterogeneous environments: responses to small scale pattern. Landscape Ecology, 10, 209217.
  • Hanley, T.A. 1982 Cervid activity patterns in relation to foraging constraints: Western Washington. Northwest Science, 56, 208217.
  • Hanley, T.A. 1997 A nutritional view of understanding and complexity in the problem of diet selection by deer (Cervidiae). Oikos, 79, 209218.
  • Henderson, D.J., Madden, S., Still, M.J., Lilly, A. & Gauld, J.H. 1994The Environmentally Sensitive Areas Designated in Scotland. The Loch Lomond ESA Biological Monitoring Report, Years One to Five. 1989–93. Macaulay Land Use Research Institute, Aberdeen.
  • Hester, A.J. & Baillie, G.J. 1998 Spatial and temporal patterns of heather use by sheep and red deer within natural heather/grass mosaics. Journal of Applied Ecology, 35, 772–784.WP ***font***WP ***font***.
  • Hester, A.J., Mitchell, F.J.G., Gordon, I.J. & Baillie, G.J. 1996 Activity patterns and resource use by sheep and red deer grazing across a grass/heather boundary. Journal of Zoology, 240, 60920.
  • Hodgson, J. & Illius, A.W. 1996 The Ecology and Management of Grazing Systems. CAB International, Wallingford, UK.
  • Hofmann, R.R. 1989 Evolutionary steps of ecophysiological adaptation and diversification of ruminants: a comparative view of their digestive system. Oecologia, 78, 443457.
  • Hunter, R.F. 1962 Hill sheep and their pasture: a study of sheep grazing in south-east Scotland. Journal of Ecology, 50, 561580.
  • Illius, A.W., Clarke, D.A. & Hodgson, J. 1992 Discrimination and patch choice by sheep grazing grass-clover swards. Journal of Animal Ecology, 61, 183194.
  • Illius, A.W. & Gordon, A.J. 1987 The allometry of food intake in grazing ruminants. Journal of Animal Ecology, 56, 989999.
  • Jiang, Z. & Hudson, R.J. 1993a Optimal grazing of wapiti (Cervus elaphus) on grassland: patch and feeding station departure rules. Evolutionary Ecology, 7, 488498.
  • Jiang, Z. & Hudson, R.J. 1993b Foraging postures of wapiti (Cervus elaphus). Applied Animal Behaviour Science, 36, 275287.
  • Kie, J.G. 1996 The effects of cattle grazing on optimal foraging in mule deer (Odocoileus hemionus). Forest Ecology and Management, 88, 131138.
  • Kohlmann, S.G. & Risenhoover, K.L. 1994 Spatial and behavioural response of white tailed deer to forage depletion. Canadian Journal of Zoology, 72, 506513.
  • Kottman, H.J., Schwöppe, W. & Willers, T. 1985 Heath conservation by sheep grazing: a cost–benefit analysis. Biological Conservation, 31, 6774.
  • Langan, S.J., Paterson, E. & Taylor, A.G. 1996 The Soil Resource of Scotland – Future Options for its Management. Soils, Sustainability and the Natural Heritage (eds A.G.Taylor, J.E.Gordon & M.B.Usher), pp. 6991. HMSO, edinburgh.
  • Langvatn, R., Albon, S.D., Burkey, T. & Clutton-Brock, T.H. 1996 Climate, plant phenology and variation in age at first reproduction in a temperate herbivore. Journal of Animal Ecology, 65, 653670.
  • Langvatn, R. & Hanley, T.A. 1993 Feeding patch choice by red deer in relation to foraging efficiency. Oecologia, 95, 164170.
  • Leuthold, W. 1977African Ungulates: a Comparative Review of Their Ethology and Behavioural Ecology. Springer, Berlin.
  • Lynch, J.J., Hinch, G.N. & Adams, D.B. 1992The Behaviour of Sheep. Biological Principles and Implications for Production. CAB International & CSIRO, Canberra, Australia.
  • Marrs, R.H., Rizand, A. & Harrison, A.F. 1989 The effects of removing sheep grazing on soil chemistry, above-ground nutrient distribution and selected aspects of soil fertility in long-term experiments at Moor House National Nature Reserve. Journal of Applied Ecology, 26, 647661.
  • McCorquodale, S.M. 1993 Winter foraging behaviour of elk in the shrub-steppe of Washington. Journal of Wildlife Management, 57, 881890.
  • Milne, J.A., Birch, C.P.D., Hester, A.J., Armstrong, H.M. & Robertson, A. 1998 The impact of Vertebrate Herbivores on the natural heritage of the Scottish uplands: a review. Scottish Natural Heritage Review, 95, 1127.
  • Milne, J.A., Macrae, J.C., Spence, A.M. & Wilson, S. 1978 A comparison of the voluntary intake and digestion of a range of forages at different times of the year by the sheep and the red deer (Cervus elaphus). British Journal of Nutrition, 40, 347357.
  • Milne, J.A. & Sibbald, A.C. in press Modelling of grazing systems at the farm level. Annals Zootechnie., in press.
  • Mitchell, B., Staines, B.W. & Welch, D. 1977Ecology of Red Deer: a Research Review Relevant to Their Management in Scotland. Institute of Terrestrial Ecology, Banchory.
  • National Countryside Monitoring Scheme 1993 NCMS Phase I (1940s–1970s). Summary Report for Scotland. Scottish Natural Heritage, Perth.
  • Nicholson, I.A. & Robertson, R.A. 1958 Some observations on the ecology of an upland grazing in north-east Scotland with special reference to Callunetum. Journal of Ecology, 46, 239270.
  • Nolan, T. & Connolly, J. 1977 Mixed grazing by sheep and steers – a review. Herbage Abstracts, 47, 367379.
  • Nolan, A.J., Still, M.J., Bell, J.S. & Gauld, J.H. 1994The Environmentally Sensitive Areas Designated in Scotland. The Breadalbane ESA Biological Monitoring Report, Years One to Five. 1989–93. Macaulay Land Use Research Institute, Aberdeen.
  • Oom, S. & Hester, A.J. 1999 Heather utilization along paths by red deer and sheep in a natural grass/heather mosaic. Botanical Journal of Scotland., in press.
  • Osborne, B.C. 1984 Habitat use by red deer (Cervus elaphus L.) and sheep in the West Highlands. Journal of Applied Ecology, 21, 497506.
  • Owens, M., Keith, M.J., Etzenhouser, D.E., Spalinger, D.E. & Murden, S.B. 1995 Grazing patterns of browsing ruminants in a heterogeneous landscape. Proceedings of the Fifth International Rangeland Congress, Utah, 199 (ed. N.E.West), pp. 424425. Society for Range Management, Denver, Colorado.
  • Palmer, S.C.F. 1997 Prediction of the shoot production of heather under grazing in the uplands of Great Britain. Grass and Forage Science, 52, 408424.
  • Parker, K.L., Robbins, C.T. & Hanley, T.A. 1984 Energy expenditures for locomotion by mule deer and elk. Journal of Wildlife Management, 48, 474488.
  • Penning, P.D., Parsons, A.J., Newman, J.A., Orr, R.J. & Harvey, A. 1993 The effects of group size on grazing time in sheep. Applied Animal Behaviour Science, 37, 101109.
  • Penning, P.D., Rook, A.J. & Orr, R.J. 1991 Patterns of ingestive behaviour of sheep continuously stocked on monocultures of ryegrass or white clover. Applied Animal Behaviour Science, 31, 237250.
  • Ruxton, G.D. 1995 Foraging on patches: are groups disadvantaged? Oikos, 72, 148150.
  • Schmidt, K. 1983 Winter ecology of non-migratory Alpine red deer. Oecologia, 95, 226233.
  • Scott, C.B., Banner, R.E. & Provenza, F.D. 1996 Observations of sheep foraging in familiar and unfamiliar environments: familiarity with the environment influences diet selection. Applied Animal Behaviour Science, 49, 165171.
  • Senft, R.L., Coughenour, M.B., Bailey, D.W., Rittenhouse, L.R., Sala, O.E. & Swift, D.M. 1987 Large herbivore foraging and ecological hierarchies. Bioscience, 37, 789799.
  • De Smidt, J.T. 1995 The imminent destruction of northwest European heaths due to atmospheric nitrogen deposition. Heaths and Moorland: Cultural Landscapes (eds D.B.A.Thompson, A.J.Hester & M.B.Usher), pp. 20617. HMSO, edinburgh.
  • Sorensen, V.A. & Taylor, D.H. 1995 The effect of seasonal change on the group size, group composition and activity budget of white-tailed deer (Odocoileus virginianus). Ohio Journal of Science, 95, 321324.
  • Spalinger, D.E. & Hobbs, N.T. 1992 Mechanisms of foraging in mammalian herbivores: new models of functional response. American Naturalist, 140, 325348.
  • Staines, B.W. 1976 The use of natural shelter by red deer in north-east Scotland. Journal of Zoology, 180, 18.
  • Staines, B.W., Balharry, R. & Welch, D. 1995 The impacts of red deer and their management on the natural heritage in the uplands. Heaths and Moorland: Cultural Landscapes (eds D.B.A.Thompson, A.J.Hester & M.B.Usher), pp. 294308. HMSO, edinburgh.
  • Sydes, C. & Miller, G.R. 1991 Range management and nature conservation in the British uplands. Ecological Change in the Uplands (eds M.B.Usher & D.B.A.Thompson), pp. 323338. Blackwell Scientific Publications, Oxford.
  • Taylor, J.A., Robinson, G.G., Hedges, D.A. & Whalley, R.D.B. 1987 Camping and faeces distribution by merino sheep. Applied Animal Behaviour Science, 17, 273288.
  • Thompson, D.B.A., Hester, A.J. & Usher, M.B, eds. 1995Heaths and Moorland: Cultural Landscapes. HMSO, edinburgh.
  • Thouless, C.R. 1990 Feeding competition between grazing red deer hinds. Animal Behaviour, 40, 105111.
  • Tudor, G. & Mackay, E.C. 1995 Upland land cover change in post-war Scotland. Heaths and Moorland: Cultural Landscapes (eds D.B.A.Thompson, A.J.Hester & M.B.Usher), pp. 2842. HMSO, edinburgh.
  • Valone, T.J. 1993 Patch information and estimation: a cost of group foraging. Oikos, 68, 258266.
  • Wallace, J.S., Lloyd, C.R., Roberts, J. & Shuttleworth, W.J. 1984 A comparison of methods for estimating aerodynamic resistance of heather (Calluna vulgaris L. Hull) in the field. Agricultural and Forest Meteorology, 32, 289305.
  • Weaver, T. & Dale, D. 1978 Trampling effects of hikers, motorcycles and horses in meadows and forests. Journal of Applied Ecology, 15, 451457.
  • Welch, D. 1985 Studies in the grazing of heather moorland in north-east Scotland IV. Seed dispersal and plant establishment in dung. Journal of Applied Ecology, 22, 461472.
Footnotes
  1. Fax: 01224 311556. E-mail: a.hester@mluri.sari.ac.uk

Received 18 October 1998; revision received 17 December 1998