Do endogenous seasonal cycles of food intake influence ... - BES journal

FEC462.fm Page 614 Thursday, October 12, 2000 2:22 PM

Functional Ecology 2000 14, 614 – 622

Do endogenous seasonal cycles of food intake influence foraging behaviour and intake by grazing sheep?

12

Blackwell Science, Ltd

G. R. IASON, D. A. SIM and I. J. GORDON Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen AB15 8QH, UK

Summary 1. Large herbivores living in temperate regions show different degrees of seasonal biological variation, including voluntary food intake (VFI). The decline of VFI in winter has been hypothesized to be an evolved response to lower food availability or quality, which can act as an internal constraint on food intake. 2. The hypotheses were tested that (i) animals that have a greater inherent seasonal variation of VFI, measured indoors under ad libitum conditions, would also have a greater seasonal variation in intake and grazing behaviour under field conditions, and (ii) greater seasonal variation in intake and grazing behaviour under field conditions would be expressed at a higher level of food availability. 3. The intake and grazing behaviour in summer and winter, of three breeds of sheep, were compared at two levels of food availability (at pasture heights of 3·7 and 5·4 cm). The breeds were known to have contrasting degrees of seasonal variation in food intake when fed ad libitum; the VFI of the Shetland (SH) and Scottish Blackface (BF) sheep varies greatly between seasons whereas that of the Dorset Horn (DH) is less seasonally variable. 4. All three breeds consistently increased their rates of biting and duration of grazing activity in the winter, taking many more smaller bites each day than in the summer, and both digestibility and intake were lower in winter than in summer. 5. Contrary to expectation, the DH ewes had the highest seasonal difference of dry matter intake at pasture, whereas the SH breed had the lowest variation of intake between seasons. 6. This experiment provides no evidence that differences between seasons in intake and foraging behaviour in the field vary with the animals’ degree of endogenous seasonal variation in VFI. Variation between seasons was consistent at both levels of resource availability, suggesting that it resulted from seasonal changes in food quality (digestibility) rather than biomass availability. It is not easy to extrapolate from laboratory feeding studies, where animals’ own physiological constraints apply, to foraging ecology in the field, where constraints imposed by the environment may be more important. Key-words: Food intake, foraging, resource availability, seasonality, sheep Functional Ecology (2000) 14, 614 – 622

Introduction

© 2000 British Ecological Society

Growth (Leader-Williams & Ricketts 1981), voluntary food intake (VFI: Barry et al. 1991) and basal metabolic rate (BMR; Silver et al. 1969; Blaxter & Boyne 1982) of large herbivores show annual cyclicity in seasonal environments. The seasonal variability of these phenomena is thought to be an evolved response by animals to reduce the possibility of failure to meet their requirements during predictable seasonal declines in food availability (Kay 1985). At high latitudes, the decline in winter of VFI and BMR is under photoperiodic control and mediated by decreasing autumn day-lengths (Kay 1979, 1985). This seasonal variation

is thus expected to persist even under conditions of high food availability (Sibbald et al. 1993). Several studies have recorded seasonal variation in VFI of large herbivores fed ad libitum, including several species of deer as well as domestic herbivores (see Arnold 1985). The significance of this inherent seasonal variation in intake or food requirements, to foraging behaviour and intake under field conditions has been rarely explored, although Heydon et al. (1993) suggested that expression of endogenous seasonal changes in appetite of Red Deer were dependent on food availability. We test the hypothesis that animals’ seasonal variation of food requirements and intake, as measured by voluntary food intake, are a fixed, internally derived constraint 614


615 Grazing seasonality

© 2000 British Ecological Society, Functional Ecology, 14, 614 – 622

and influence foraging behaviour and intake under field conditions (internal constraint hypothesis). The interaction between the animal’s seasonal variation in intake and the availability of food resources is also examined. Specifically, we test the hypothesis that animals with a greater degree of inherent seasonal variation in VFI would express greater seasonal differences in intake and foraging behaviour, under field conditions of greater resource availability, when foraging parameters are not constrained by limited food availability. Comparison of genotypes of sheep with contrasting degrees of endogenous variability of VFI between seasons, provides the opportunity to investigate, within a single species, the extent to which this is related to intake realized under field conditions. Non-breeding Shetland and Scottish Blackface sheep, which are found at high latitudes and in upland areas of Britain, have a greater seasonal variation in VFI, measured under indoor conditions of ad libitum food availability, compared with the lowland Dorset Horn breed (Iason et al. 1994). In this work we hence refer to Scottish Blackface and Shetland sheep as inherently seasonal, and to Dorset Horn sheep as inherently aseasonal. In this study we compare the daily food intake of these three breeds of sheep grazing at pasture in summer and winter. Daily intake can be considered as the product of time spent grazing, the rate of biting and the size of the bites (Allden & Whittaker 1970; Hodgson 1985). Hence, we also examine the behavioural basis for seasonal variation in intake, hypothesizing that the inherently seasonal genotypes would vary one or more of these behaviours between winter and summer, to a greater degree than the inherently aseasonal genotype. We also aimed to compare in the same experiment, the effect of food availability represented by height of the grass swards, on seasonal variation in intake and foraging behaviour. When food availability is low, for example when grazing on short grassy swards, the mean mass of food ingested in each bite is reduced (Black & Kenney 1984; Gordon, Illius & Milne 1996). Bite mass is one of the key determinants of rate of intake of grazing large herbivores (Spalinger & Hobbs 1992; Laca, Ungar & Demment 1994). Therefore, in order to meet its requirements, a grazing animal must compensate for smaller bite masses by either increasing daily foraging time (Fryxell 1991; Iason et al. 1999), or the rate of biting (Penning 1986). We would hence predict that seasonal variation in foraging behaviour would be constrained by sward characteristics, being less for bite mass on shorter swards, but varying more between seasons for variables that may compensate for restricted bite mass on shorter swards. Greater inherent seasonality of intake would be more likely to be expressed when food resources were abundant. If an endogenously driven reduction in intake occurs in the winter months, then it may do so regardless of the level of food availability. However, an increase in daily

intake during the summer months, brought about by changes in the associated foraging behaviours, is more likely to be constrained by available biomass on the shorter swards (Hodgson 1985).

Materials and methods   The experiment consisted of a factorial design containing three breeds of sheep (Dorset Horn, Scottish Blackface, Shetland) each grazing experimental plots at two target sward heights (short, 3·5 cm; tall, 5·5 cm). These six treatments were replicated and the experiment conducted on the same 12 experimental plots in each of the two seasons, summer (July) and winter (November). Therefore there were four plots for each breed (two target sward height treatments × two replicates) in each season. Each plot was 0·43 ha of a homogeneous sown ryegrass (Lolium perenne) and Timothy (Phleum pratense) sward, at the Macaulay Land Use Research Institute’s Glensaugh Research Station, Kincardineshire (latitude 56°55′ N; longitude 2°35′ W). Twenty-four ewes of each breed were prevented from mating during the previous autumn and six ewes of a single breed were allocated to each plot, 5 weeks prior to the summer measurements of intake and behaviour which took place during a 12-day period from 23 July to 3 August 1990. Ewes were weighed prior to allocation, which was carried out in a stratified manner equalizing the body mass on the four plots containing each breed. The ewes were then held elsewhere as a single group and the same individuals were allocated to the same experimental plots and treatments one week prior to the winter measurement period (16–27 November 1990). During the two 12-day measurement periods, the condition of each sheep was scored according to the back-fat condition score system, which uses a scale 0–5 with subdivisions of 0·25, corresponding to total fat content of the animal (Russel, Doney & Gunn 1969). The sheep were shorn immediately after the summer measurement period. Sward heights within each plot were recorded twice weekly by measuring height at 40 points (Barthram 1986), and target heights were maintained by addition of non-experimental sheep. These sheep were of different breeds and did not interact with the experimental ewes, and thus provided minimal behavioural disturbance to the experimental sheep. Water was available throughout. The relationship between sward height and standing crop biomass was investigated by cutting seven 100 cm × 20 cm quadrats in each plot during each intensive measurement period. Quadrats were positioned according to a stratified random procedure which meant that within each plot, two quadrats were placed on patches that were visibly judged to be at a taller than average sward height, two on patches judged to be at a shorter than average sward height, and three were


616 G. R. Iason et al.

randomly chosen. Prior to cutting the vegetation with electric hand shears, 30 measurements of sward height were made within each quadrat. Vegetation samples were oven dried to constant mass at 80 °C.

     Daily dry matter intake and digestibility of dry matter were measured in summer and winter in each of the 72 ewes using the n-alkane marker technique (Mayes, Lamb & Colgrove 1986; Mayes et al. 1995). Each ewe was orally dosed daily for the first 11 days of the 12day measurement period, with a paper pellet impregnated with 1·3 g of each of C32 and C36 n-alkane, using a standard veterinary dosing gun. Dosing was always carried out at the same time each morning when the ewes were gathered. On each of the final 5 days of the period, a faecal grab sample was collected and pooled within each ewe. These were subsequently freeze-dried and analysed for naturally occurring oddchain n-alkanes from plant waxes and dosed alkanes using gas chromatography, following heptane extraction. Comparison of ratios of dosed and natural faecal n-alkanes permits calculation of dry matter intake and faecal output. From these, dry matter digestibility was also calculated. An estimate of the alkane composition of herbage ingested was made following collection of herbage samples from each plot (see Experimental design) by oesophageally fistulated non-breeding ewes of the same breed as the experimental sheep on that plot. These sheep were given access to the plots to be sampled for at least 8 h free grazing with the experimental ewes prior to the collection of oesophageal extrusa. Insufficient faeces were obtained from one ewe in the summer, and one was removed from the experiment prior to the winter period because of illness. Hence no measurement of intake and digestibility was possible for one ewe in summer and for another in winter.

  

© 2000 British Ecological Society, Functional Ecology, 14, 614– 622

Foraging behaviour was measured by electronic bite meters (Penning 1983; Penning, Steel & Johnson 1984) which consisted of a transducing noseband fitted around the sheep’s muzzle, the signals from which were recorded using a small harness-mounted audio cassette recorder. A bite meter was mounted on each of three randomly chosen sheep in each plot of one complete replicate set of plots, for the first 4 days of the summer sampling period. Bite meters were then transferred to three sheep per plot of the second replicate series for the final 4 complete days of the sampling period. This was repeated in the winter season, when the ewes fitted with bite meters were those that had not received them during the summer period. Behavioural data were hence collected on only half of the 72 ewes in each season. The first day of each 4-day period was treated as a

training run for the ewes, and all data on these occasions were discarded. Bite meter signals were decoded by computer software programmed according to Penning (1983). Jaw movements were described as being either prehension bites, mastication bites or chews during rumination. Each minute recorded was allocated to time spent grazing, ruminating or idling, the latter being defined as time with less than 40 jaw movements per minute. The proportion of the day in each activity was hence calculated, as were the mean rates of the different types of jaw movement per grazing or ruminating minute. Activity or biting data were included only in analyses where more than 960 min were available in each 24-h period. Owing to failures in bite-recording equipment, behavioural data were available for 57 of a possible 72 individuals across both seasons. The mean numbers of full days of data were 2·5 days per ewe per plot (SE 0·29) in summer and 2·0 days per ewe per plot (SE 0·21) in winter.

  The statistical analysis of intake, digestibility and behavioural data consisted of a three-way analysis of variance with treatments consisting of two seasons (summer and winter), three breeds (DH, SH, BF) and two sward heights (short and tall). Plots and season within plots were treated as random effects, which structures the analysis to test the treatment effects against the residual variation between plots, except the effects of season and its interaction with other variables which were measured within plots. The Residual Maximum Likelihood (REML) procedure of Genstat Committee (1993) was used for all analyses. Statistical testing was carried out by dividing the Wald statistics by their degrees of freedom, and referencing this to the appropriate F-distribution (Elston 1998). Standard errors of difference (SED) are quoted for all statistical comparisons. The statistical design of pooling individuals within plots means that residual degrees of freedom (df ) were unaffected by the missing data values for intake, digestibility and behavioural measurements. Where comparisons were made between means pooled across treatments, then the SED specific to this comparison is quoted in the text, without units, which are the same as the corresponding means. Although the relationship between intake and body mass may be allometric when comparing across species, preliminary graphical inspection of the data showed that across the relatively small range of body masses within the sheep in this experiment, the relationship was approximately linear. Addition of a quadratic term to the model did not explain significantly more variation in any of the intake variables. All quoted measurements of intake, plus estimated bite mass, were adjusted to the mean body mass of ewes in all treatments by entering this as a covariate into the REML analyses.


617 Table 1. The actual sward heights (cm) achieved on each treatment. Means (and pooled Grazingstandard errors) of six values are quoted for each treatment. These were obtained on three separate days at 4-day intervals during intensive measurement seasonality periods, on each of the two replicate plots for each treatment. Each value was the mean of 40 measurements per plot Shetland

Scottish Blackface

Dorset Horn

Season

Short

Tall

Short

Tall

Short

Tall

Summer

3·6 (0·20) 3·8 (0·29)

6·5 (0·26) 5·8 (0·20)

3·7 (0·15) 3·5 (0·26)

5·7 (0·21) 4·7 (0·41)

3·9 (0·13) 3·8 (0·27)

4·9 (0·15) 4·5 (0·37)

Winter

Table 2. Body mass and condition score of sheep of each breed. SEDw is the SED appropriate for comparisons within breeds, and SEDb is for comparisons between breeds

Season

Shetland

Scottish Blackface

Dorset Horn

SEDw

SEDb

Body mass (kg)

Summer Winter

38·1 38·1

60·9 60·4

63·4 65·2

1·00

1·23

Condition score

Summer Winter

0·034

0·066

2·39 2·16

2·75 2·58

2·86 2·82

Results Sward height (cm) was correlated with standing crop biomass (SCB, kg ha–1), explaining a large proportion of variation according to the following relationship (F = 439·2, df 1,155, P < 0·001; r 2 = 0·725): SCB = 271·0 (SE 12·9) sward height + 66 (SE 85·4). eqn 1 This relationship was consistent between seasons and breeds, indicated by the absence of any significant interaction effects (sward height × season: F = 0·08, df 1,155, NS; sward height × breed F = 1·43, df 2, 155, NS). Actual sward heights achieved broadly matched the target sward heights (Table 1). A significant difference between tall and short sward height treatments was maintained (short 3·69 cm, tall 5·25 cm, SED 0·174, F = 272·5, df 1,6, P < 0·001). However, the Dorset Horn sheep on the tall target sward height treatments had relatively short sward heights of 4·7 cm. Sward heights were slightly but significantly affected by breed (SH 4·9 cm, BF 4·3 cm, DH 4·2 cm; SED 0·12; F = 19·7, df 2,6, P < 0·01, r 2 = 0·067) as well as season (summer 4·7 cm; winter 4·2 cm; SED 0·17; F = 11·5, df 1,52, P < 0·01; r 2 = 0·064).

   © 2000 British Ecological Society, Functional Ecology, 14, 614 – 622

The mean masses of the three different breeds of sheep were Shetland, 38·1 kg, Scottish Blackface, 60·5 kg and Dorset Horn, 64·5 kg (SED 1·01, F = 400·3, df 2,6, P < 0·001; Table 2) and these remained constant

between seasons (summer 54·3 kg, winter 54·4 kg, SED 0·58, F = 0·03, df 1,6, NS; breed–season interaction, F = 1·47, df 2,6, NS). Body condition score also varied significantly among breeds (SH 2·3, BF 2·7, DH 2·8; SED 0·06, F = 40·6, df 2,6, P < 0·001, Table 2), and a small effect of season occurred only in Scottish Blackface and Shetland sheep (breed–season interaction, F = 8·45, df 2,6, P < 0·05, Table 2). Daily dry matter intake increased with body mass of the sheep by 0·079 kg DM d–1 (SE 0·0025) per kg (F = 46·2, df 1,117, P < 0·001), and this relationship did not vary between breeds (breed × body mass interaction F = 0·7, df 2,115, NS). Body mass was hence used as a covariate in the analysis of daily dry matter intake, instantaneous rate of intake and mean bite mass (Tables 3 and 4). Daily dry matter intake, adjusted for body mass was significantly (P < 0·001) lower in winter than in summer (summer: 1·03 kg DM d–1, winter: 0·64 kg DM d–1, SED 0·034), significantly lower (P < 0·001) on the shorter sward treatments (short 0·73 kg DM d–1; tall 0·94 kg DM d–1, SED 0·064) and varied significantly (P < 0·05) between breeds (SH 0·68 kg DM d–1; BF 0·81 kg DM d–1; DH 1·01 kg DM d–1, SED 0·092, Table 3). Daily intakes adjusted for body mass were significantly lower in Shetland and Scottish Blackface breeds compared with the Dorset Horn breed (P < 0·05, LSD test, Snedecor & Cochran 1980). There was a significant season breed interaction effect on dry matter intake (F = 5·25, df 2,6, P < 0·05), for Shetland, Scottish Blackface and Dorset Horn the mean values being 0·81, 1·04 and 1·26 kg DM d–1 in summer, and 0·56, 0·59 and 0·76 kg DM d–1 in winter, respectively (interaction SED 0·093). The dry matter digestibility of the ingested diet was significantly (P < 0·01) greater in summer than in winter (summer 0·781; winter 0·670; SED 0·0228), but digestibility did not differ significantly between breeds (SH 0·723; BF 0·716; DH 0·737; SED 0·0406), nor between tall and short swards (short 0·694; tall 0·757; SED 0·0331, Table 3). The digestible dry matter intake, which is the product of dry matter intake and digestibility expressed as a proportion, showed a similar pattern of variation with experimental treatments as dry matter intake (Table 3).

  The sheep spent a significantly greater percentage of time eating in winter than in summer (summer 28·7%; winter: 38·6%; SED 2·38, P < 0·001; Table 3), and overall the Shetland sheep ate for less time than the other breeds (SH 29·4%; BF 36·9%; DH 34·9%; SED 2·21, P < 0·05). In contrast, the proportion of time spent ruminating was significantly lower in winter (summer 19·7%; winter 12·4%; SED 1·055, P < 0·001). This seasonal reduction in percentage of time ruminating was particularly pronounced on the short swards

Shetland

Scottish Blackface

Dorset Horn

SED

F

Season

Short

Tall

Short

Tall

Short

Tall

bseason

wbreed

wsward ht

Season df 1,6

Breed df 2,6

Sward height df 1,6

Significant interactions‡

Daily intake† (kg DM d–1)

Summer Winter

0·73 0·50

0·87 0·62

0·87 0·46

1·20 0·72

1·12 0·71

1·41 0·81

0·084

0·128

0·144

130·6***

5·6*

10·5*

Season × Breed

Digestible dry matter intake† (kg DM d–1)

Summer Winter

0·58 0·34

0·70 0·48

0·66 0·26

0·94 0·52

0·87 0·45

1·14 0·54

0·079

0·128

0·139

137·1***

3·1

9·1*

Season × Breed

Digestibility

Summer Winter

0·76 0·61

0·79 0·74

0·76 0·60

0·78 0·73

0·78 0·66

0·82 0·69

0·056

0·070

0·070

23·8**

0·2

3·6

*P < 0·05, **P < 0·01, ***P < 0·001. †Adjusted for ewe mass. ‡See text.

Table 4. The mean biting and chewing parameters and instantaneous intake rate of ewes of each breed on each target sward height in summer and winter SEDs and total sample sizes are as in Table 3 Shetland

Total daily prehension bites (bite d–1) –1

Mean prehension biting rate per grazing minute (bites min ) –1

Mean mastication biting rate per grazing minute (bites min ) –1

–1

Mean instantaneous intake rate† (mg DM min kg ) Mean prehension bite mass† (mg DM bite-1)

*P < 0·05, **P < 0·01, ***P < 0·001. †Adjusted for body mass.

Scottish Blackface

Dorset Horn

SED bseason

wbreed

wsward ht

Season df 1,6

Breed df 2,6

Sward ht df 1,6

3380

6029

6893

104·0***

1·2

2·2

8·41

10·52

11·42

24·8**

3·0

2·0

8·91

12·00

12·06

28·0**

6·1*

1·8

42·4***

3·0

7·8*

52·8***

3·7

4·1

Season

Short

Tall

Short

Tall

Short

Tall

Summer Winter Summer Winter Summer Winter Summer Winter Summer Winter

24 418 39 937 65·4 80·9 70·3 51·2 2·13 1·19 28·2 12·4

19 545 29 276 70·8 75·5 70·7 53·7 2·87 1·67 39·7 21·1

22 628 44 747 57·8 83·9 82·5 57·0 2·01 0·77 37·8 9·9

21 331 41 960 45·3 76·0 98·5 64·9 2·21 1·32 50·3 18·1

20 839 41 411 48·5 75·2 88·2 51·5 2·56 1·01 58·8 19·5

23 827 34 930 58·7 64·8 78·9 69·7 4·12 1·71 78·6 34·2

0·483 8·79

F

0·630 14·64

0·768 18·41




Table 3. The daily dry matter (DM) intake, dry matter digestibility and daily digestible dry matter intake of ewes of each breed on each target sward height in summer and winter (total N = 142 measurements in both seasons). SEDs are as follows: bseason is for comparisons between seasons within breeds and within sward heights, wbreed is for comparisons within breeds between sward heights within or between seasons, and wsward ht is for comparisons within sward heights between breeds, within or between seasons


Table 619 5. The percentage of total time in each 24 h spent eating, ruminating and idling by ewes of each breed on each target sward height in summer and winter (N = 57 observations across both seasons). Means and SEDs are derived from untransformed data, while statistical testing was based on Grazing angular transformed proportions. SEDs are as in Table 3

seasonality

Shetland

Eating (%) Ruminating (%) Idling (%)

Scottish Blackface

Dorset Horn

F SED

Season

Short

Tall

Short

Tall

Short

Tall

Summer Winter Summer Winter Summer Winter

28·3 39·6 22·3 9·0 49·4 51·0

20·8 28·8 13·5 11·1 65·8 59·7

31·0 41·9 17·4 7·8 52·0 50·3

32·4 42·1 18·6 11·2 48·8 46·6

32·8 38·8 23·6 15·9 44·1 45·5

27·2 40·6 23·0 19·3 49·8 39·8

bseason

wbreed

wsward ht

Season df 1,6

Breed df 2,6

Sward height df 1,6

4·63

6·34

6·34

15·9**

5·6*

4·3

2·25

3·97

4·30

48·5***

3·7

0·0

4·70

7·35

7·98

4·3

1·3

1·1

*P < 0·01, **P < 0·01, ***P < 0·001.

(sward height season interaction: F = 5·2, df 1,6; P < 0·1, see Table 3), although the main effect of sward height on all of the measurements of foraging activity was weak and not statistically significant. There were no significant differences between breeds in the percentage of time spent ruminating (SH 14·0%, BF 13·8%, DH 20·4%; SED 2·26, NS, Table 3). The percentage of time spent idling did not vary significantly with any of the experimental treatments (P > 0·05, Table 5).

 


Data for ingestive behaviour are given in Table 4. The main effect of season was significant for all variables. Breed of sheep did not significantly explain any of the variation in any of the variables, and none of the interactions between any of the explanatory variables was statistically significant. There was a considerable increase between summer and winter in both the rate of biting (summer 57·8 bites min–1; winter 76·1 bites min–1; SED 3·87, P < 0·01) and the total number of bites per day (summer 22 098 bites d–1; winter 38 710 bites d–1; SED 1728, P < 0·001). However, the bites were significantly much smaller in winter than in summer (summer 48·9 mg DM bite–1; winter 19·2 mg DM bite–1; SED 4·35, P < 0·001) and the instantaneous intake rate per grazing minute was commensurately lower in winter than in summer (summer 26·4 g DM min–1, winter 12·8 g DM min–1; SED 2·20, P < 0·001). All these seasonal effects were consistent for all breeds on short and tall sward heights. Total number of daily bites was on average lower on the taller sward height (short 32 330 bites d–1, tall 28 478 bites d–1; SED 2640, NS), although like prehension bite rate (short 68·6 bites min–1, tall 65·2 bites min–1; SED 3·99, NS), and mean bite mass adjusted for body mass (short 27·7 mg bite–1, tall 40·3 mg bite–1; SED 6·30, NS), this did not differ significantly between sward height treatments. However, these weak effects combined to result in significantly greater instantaneous rates of intake on the tall swards (short 16·1 g DM min–1, tall 23·2 g DM min–1; SED 2·56, P < 0·05).

Discussion The question ‘does an animal’s endogenous seasonal variation in physiology constrain its foraging behaviour and intake?’ is fundamental to the ecology and evolution of large herbivores in relation to their environment. The seasonal biological cycles in large herbivores of temperate latitudes are controlled by fluctuating photoperiod, the mechanisms associated with which appear to be very strongly conserved by natural selection (Loudon 1985). Such rigid physiological systems are considered likely to limit species’ behaviour and ecological niche (Karasov & Diamond 1988). However, because such systems have presumably evolved in concert with relevant selective forces, it is difficult to uncouple their causes and effects under natural circumstances. The adaptive function of a winter decline in VFI and MMR is that it would result in less of a shortfall in nutritional requirements at a time when food availability or quality is at its lowest. Growth, reproduction and activity are also restricted at this time. Conversely, a summer increase in intake would facilitate the productive processes. Artificial selection for continuous production has led to contrasting degrees of endogenous seasonality, in breeds of sheep whose overall biology is otherwise similar (Kay 1985). In this study, the consistency of the results across all breeds of sheep, regardless of their inherent seasonal variability, suggests that inherent seasonal variation of the animals’ VFI per se plays only a small part in their foraging ecology in the field. All observed effects can be explained by the dominant effect of the seasonal decline in quality of the available vegetation. The breeds of sheep compared in this experiment were chosen for their contrasting degrees of seasonal variation in VFI. Under an ad libitum feeding regimen, VFI (expressed as the seasonal difference as a percentage of the overall mean for the breed) varied by 9·1% for DH ewes, 17·8% for BF ewes and 13·7% for SH ewes between August and November (Iason et al. 1994). The DH ewes maintained a more constant VFI between seasons than the other two breeds, mainly because of a lower propensity to increase intake in




the summer; all breeds reached a similar nadir to their cycle of VFI in the winter months. If this pattern of seasonal variation in intake persisted in the field, then we would have expected significant breed–season interaction effects in the statistical analysis of intake and foraging behaviour in this field experiment. The results did show a significant interaction between breed and season for daily intake of dry matter and digestible dry matter, measured under field conditions. However, this interaction was not in the direction consistent with the seasonal variation in intake being determined, by the breeds’ known inherent seasonal variation observed under ad libitum feeding. In fact the proportional difference between summer and winter intakes, measured in the field, were much greater than those measured under ad libitum conditions, and was greatest for the DH (49·7%) and BF (55·0%) breeds, and least for the SH breed (35·3%). The actual daily intakes of the two predicted inherently seasonal breeds, measured in the field in winter were low, and consistent with the hypothesis that some internal constraint may determine a nadir in an annual cycle of intake. Contrary to this internal constraint hypothesis, the sheep’s daily intakes in the field in summer did not exceed the very high values of the DH, which also increased markedly between winter and summer. Similarly, there was no indication that the predicted more seasonal BF and SH expressed their endogenous seasonal variation to a greater extent than the DH, at greater levels of food availability, represented by taller swards. The sward surface heights offered to the sheep in this experiment reflected the sward biomass, and spanned the range that ought to be expected, from previous work, to produce differences in foraging behaviour and daily intake of sheep (Arnold 1975; Hodgson 1985; Penning et al. 1995). Lower daily intakes were realized on shorter swards, despite longer daily grazing times and faster prehension bite rates, which do not fully counteract the smaller bite masses (Allden & Whittaker 1970). The pasture was only lightly grazed over the autumn with minimal grazing during the winter measurement period, when active growth would have been negligible. This lack of growth of the sward during the winter measurement period may have led to the ratio of the grass leaf to stem material declining slightly between the summer and winter, which in turn could have influenced seasonal variation in biting behaviour. However, the measured relationship between standing crop biomass and sward height remained constant in the two seasons. The consistency of the grazing response between seasons among all breeds and at both target sward heights confirms that it is not due to slight departures in actual sward height from the target sward heights. Overall, the much smaller bite masses of sheep in winter cannot be explained by seasonal variation in sward structural characteristics, which were maintained by the design of the experiment to be constant.

Body mass and condition score are correlated (Russel et al. 1969), but the seasonal variation in intake of different breeds cannot be explained by either. Although we did not measure intake in the field on a year-round basis, our results support the suggestion that the seasonal variation in body reserves of large herbivores do not drive foraging behaviour, rather, body reserves are the result of variation in foraging efficiency (Parker et al. 1996). Our results are consistent with the lower quality of the available winter forage driving foraging behaviour of all breeds and over-riding any effects due to differences in inherent seasonality in the animals’ food intake or requirements. The sheep of all breeds showed profound seasonal variation in feeding behaviour; there were no significant interactions involving breed for any of the measurements of foraging activity or ingestive behaviour, regardless of the degree of inherent seasonal variation in VFI expressed by the breeds, under ad libitum-fed conditions. Evidence for the strong decline in forage quality in winter is provided by the lower digestibility of the ingested forage by all breeds. When these same breeds of sheep were fed on the same quality of forage under controlled conditions, there was a 4% greater DM digestibility in summer than winter (Iason, Sim & Foreman 1995). This small difference contrasts with the 11% greater DM digestibility in summer found in this study, a large proportion of which must be due to the changing quality of the vegetation available. Since intake in ruminants is positively related to digestibility (Minson 1982), the seasonal difference in digestibility explains the large magnitude of the seasonal difference in intake measured in the field in this study, compared to that measured under ad libitum conditions with constant diet quality (see above). Regardless of the height of the sward and breed of sheep, the sheep bit at a much faster rate in winter, and grazed for longer each day, the result of which was that they took almost twice as many bites each day in winter than in summer. The bites were, however, much smaller in winter than in summer, on both sward height treatments and in all three breeds of sheep. This consistency suggests that the smaller bites taken in winter reflect a greater selectivity of foraging in this season. This is interpreted as an attempt to maintain a high quality of diet in winter, when the fibrousness of the available food is greatest, and its digestibility is lowest. It appears that sheep in this experiment were not attempting to maximize short-term or daily intake regardless of forage quality. Maximization of intake rate could best be achieved in both seasons by maximization of bite size, by biting deeper into the sward (Black & Kenney 1984; Spalinger & Hobbs 1992), which clearly did not occur in winter, when bite masses were very small. The relationship between the animal’s physiology, intake and foraging behaviour is of central importance to models of optimal diet choice by herbivores, which


621 Grazing seasonality

assumes that the animal chooses an optimal diet within a set of constraints. These constraints may be imposed by factors that are largely internal to the animal, such as food requirements, limited capacity for digestive processing and daily foraging time. They may also be due to interactions between plant and animal factors, such as rates of encounter with and consumption of plants (Belovsky 1986; Spalinger & Hobbs 1992; Owen-Smith 1993). Despite their theoretical importance, very little is known about the relative positions of these internal and environmental constraints. In conclusion, seasonal variation in feeding behaviour under field conditions was similar in all three breeds of sheep, regardless of the degree of seasonality expressed by the breeds, under ad libitum-fed conditions indoors. Although intake varied differently between seasons for different breeds, the effects were not commensurate with the expectation that those with strongly seasonal variation in intake under ad libitum conditions would show greater seasonal variation when grazing under field conditions. Nor was a greater degree of seasonal variation of intake or foraging behaviour expressed under conditions of greater food availability. We suggest that variation in the nutritional ecology of large herbivores between seasons is determined primarily by seasonally variable environmental constraints acting under field conditions. In particular, the lower quality of the available winter forage drives foraging behaviour and over-rides any effects due to inherent seasonal variation in the animals’ appetite. This experiment also suggests that there may be no simple basis for extrapolation from indoor experiments with ad libitum-fed animals, to their intake and foraging behaviour in the field which is strongly influenced by the structure and quality of available vegetation.

Acknowledgements We thank Bob Mayes for help and advice with the alkane technique, David Elston of Biomathematics and Statistics Scotland for statistical advice, Murray Beattie, Elaine Foreman and Iain Thompson for technical assistance and Phillippe Cherdel and Ewen Robertson for help in processing data. We also thank Andrew Illius, Norman Owen-Smith, Angela Sibbald, John Milne and Jeff Maxwell for critical comments. This research was supported by the Scottish Executive Rural Affairs Department.

References


Allden, W.G. & Whittaker, R.H. (1970) The determinants of herbage intake by grazing sheep: the interrelationship of factors influencing herbage intake and availability. Australian Journal of Agricultural Research 21, 755 – 766. Arnold, G.W. (1975) Herbage intake and grazing behaviour in ewes of four breeds at different physiological states. Australian Journal of Agricultural Research 26, 1017 –1024. Arnold, G.W. (1985) Regulation of forage intake. Bioenergetics of Wild Herbivores (eds R. J. Hudson & R. G. White), pp. 81–101. CRC Press, Boca Raton, FL.

Barry, T.N., Suttie, J.M., Milne, J.A. & Kay, R.N.B. (1991) Control of food intake in domesticated deer. Physiological aspects of digestion and metabolism in ruminants. Proceedings of the Seventh International Symposium on Ruminant Physiology (eds T. Tsuda, Y. Sasaki & R. Kawashima), pp. 385 – 401. Academic Press, New York. Barthram, G.T. (1986) Experimental Techniques: The HFRO Sward Stick, pp. 29 – 30. Biennial report 1984–85. Hill Farming Research Organisation, Penicuik. Belovsky, G.A. (1986) Optimal foraging theory and community structure: implications for a guild of generalist grassland herbivores. Oecologia 70, 35–52. Black, J.L. & Kenney, P.A. (1984) Factors affecting diet selection by sheep. II Height and density of pasture. Australian Journal of Agricultural Research 35, 565–578. Blaxter, K.L. & Boyne, A.W. (1982) Fasting and maintenance metabolism of sheep. Journal of Agricultural Science, Cambridge 99, 611– 620. Elston, D.A. (1998) Estimation of denominator degrees of freedom of F-distributions for assessing Wald statistics for fixed-effect factors in unbalanced mixed models. Biometrics 54, 1085 –1096. Fryxell, J.M. (1991) Forage quality and aggregation by large herbivores. American Naturalist 138, 478–498. Genstat Committee (1993) Genstat-5, Release 3, Reference Manual. Clarendon Press, Oxford. Gordon, I.J., Illius, A.W. & Milne, J.D. (1996) Sources of variation in the foraging efficiency of grazing ruminants. Functional Ecology 10, 219–226. Heydon, M.J., Sibbald, A.M., Milne, J.A., Brinklow, B.R. & Loudon, A.S.I. (1993) The interaction of food availability and endogenous physiological cycles on the grazing ecology of red deer hinds (Cervus elaphus). Functional Ecology 7, 216 – 222. Hodgson, J. (1985) The control of herbage intake in the grazing ruminant. Proceedings of the Nutrition Society 44, 339 – 346. Iason, G.R., Sim, D.A., Foreman, E., Fenn, P. & Elston, D.E. (1994) Seasonal variation in voluntary food intake and basal metabolic rate in three contrasting breeds of sheep. Animal Production 58, 381–387. Iason, G.R., Sim, D.A. & Foreman, E. (1995) Seasonal changes in intake and digestion of chopped timothy hay by three breeds of sheep. Journal of Agricultural Science 125, 273 – 280. Iason, G.R., Mantecon, A.R., Sim, D.A., Gonzalez, J., Foreman, E., Bermudez, F.F. & Elston, D.A. (1999) Can grazing sheep compensate for a daily foraging time constraint? Journal of Animal Ecology 68, 87–93. Karasov, W.H. & Diamond, J.M. (1988) Interplay between physiology and ecology in digestion. Bioscience 9, 602–611. Kay, R.N.B. (1979) Seasonal changes of appetite in deer and sheep. ARC Research Reviews 5, 13–15. Kay, R.N.B. (1985) Seasonal variation of appetite in ruminants. Recent Advances in Animal Nutrition 1985 (eds W. Haresign & D. J. A. Cole), pp. 199 –210. Butterworths, London. Laca, E.A., Unger, E.D. & Demment, M.W. (1994) Mechanisms of handling time and intake rate of large mammalian grazer. Applied Animal Behaviour Science 39, 3–19. Leader-Williams, N. & Ricketts, C. (1981) Seasonal and sexual patterns of growth and condition of reindeer introduced into South Georgia. Oikos 38, 27–39. Loudon, A.S.I. (1985) A gland for all seasons. New Scientist 1466, 40 – 43. Mayes, R.W., Lamb, C.S. & Colgrove, P.M. (1986) The use of dosed and herbage n-alkanes as markers for the determination of herbage intake. Journal of Agricultural Science 107, 161–170. Mayes, R.W., Dove, H., Chen, X.B. & Guada, J.A. (1995) Advances in the use of faecal and urinary numbers for




measuring diet composition, herbage intake and nutrient utilization in herbivores. Recent Developments in the Nutrition of Herbivores (eds M. Journet, E. Grenet, M.-H. Force, M. Theriez & C. Demarquilly), pp. 381–406. INRA, Paris. Minson, D.J. (1982) Effects of chemical and physical composition of herbage eaten upon intake. Nutritional Limits to Animal Production for Pasture (ed. J. B. Hacker), pp. 167 –182. CAB, Slough. Owen-Smith, N. (1993) Evaluating optimal diet models for an African browsing ruminant, the kudu: how constraining are the assumed constraints? Evolutionary Ecology 7, 499 – 524. Parker, K.L., Gillingham, M.P., Hanley, T.A. & Robbins, C.T. (1996) Foraging efficiency: energy expenditure versus energy gain in free-ranging black-tailed deer. Canadian Journal of Zoology 74, 442 – 450. Penning, P.D. (1983) A technique to record automatically some aspects of grazing and ruminating behaviour of sheep. Grass and Forage Science 38, 89 – 96. Penning, P.D. (1986) Some effects of sward conditions on grazing behaviour and intake by sheep. Grazing Research at Northern Latitudes (ed. O. Gudmundsson), pp. 219 – 226. Plenum/NATO Science Affairs Division, London. Penning, P.D., Steel, G.L. & Johnson, R.H. (1984) Further

development and use of an automatic recording system in sheep grazing studies. Grass and Forage Science 39, 345 –351. Penning, P.D., Parsons, A.J., Orr, R.J., Harvey, A. & Champion, R.A. (1995) Intake and behaviour responses by sheep in different physiological states when grazing monocultures of grass or white clover. Applied Animal Behaviour Science 45, 63 – 78. Russel, A.J.F., Doney, J.M. & Gunn, R.G. (1969) Subjective assessment of body fat in live sheep. Journal of Agricultural Science, Cambridge 72, 451–454. Sibbald, A.M., Fenn, P.D., Kerr, W.G. & Loudon, A.S.I. (1993) The influence of birth date on the development of seasonal cycles in red deer hinds (Cervus elaphus). Journal of Zoology 230, 593 –607. Silver, H., Colovos, N.F., Holter, J.B. & Hayes, H.H. (1969) Fasting metabolism of white-tailed deer. Journal of Wildlife Management 33, 490 – 498. Snedecor, G.W. & Cochran, W.G. (1980) Statistical Methods. Iowa State University Press, Iowa. Spalinger, D.E. & Hobbs, N.T. (1992) Mechanisms of foraging in mammalian herbivores: new models of functional response. American Naturalist 140, 325–348. Received 8 December 1999; revised 22 March 2000; accepted 31 March 2000