www.elsevier.com / locate / livprodsci
How forage characteristics influence behaviour and intake in
small ruminants: a review
a ,
*
a b cR. Baumont
, S. Prache , M. Meuret , P. Morand-Fehr
a
´ `
Unite de Recherches sur les Herbivores, INRA Centre de Clermont-Fd /Theix, 63122 Saint-Genes-Champanelle, France
b
´ ´
Unite d’Ecodeveloppement, INRA Domaine Saint-Paul-Site Agroparc, 84914 Avignon Cedex 9, France
c
Laboratoire de Nutrition et Alimentation, INA-PG, 16, Rue Claude-Bernard, 75231 Paris Cedex 05, France
Abstract
´ ` ´
l’ingestion des petits ruminants dans differents milieux sont brievement discutees. 2000 Elsevier Science B.V. All rights reserved.
Keywords: Intake; Feeding behaviour; Forage; Foraging; Sheep; Goat
1. Introduction depends on nutritional requirements and body
re-serves (Faverdin et al., 1995). Feed factors act
Small ruminant production systems vary widely, mainly on the short-term control.
from intensive feeding with zero grazing to the utilisation of rangelands in arid areas. In all
pro-2.1. Meal behaviour and satiation process duction systems, it is generally economically
sens-ible to maximise the proportion of forage in the diet
When fed indoors main meals are scheduled by to minimise feeding costs. Furthermore, as people
the feed distribution, usually two per day. At pasture, are becoming more and more sensitive to the image
two main grazing periods usually occur at sunrise of animal products, maximising forage utilisation is
and sunset. Post-ingestive signals coming from feed an increasingly important tool in animal production.
(rumen fill, fermentation products and nutrients) In extensive systems, grazing also contributes to
contribute to the satiation process. These feedback resource preservation. However, efficient utilisation
signals are mainly integrated in the control of intake of forage resources and control of animal impact on
to prevent excess. vegetation need thorough knowledge of what
de-Sixty to 80% of daily intake is eaten during two termines feeding behaviour and dietary choices.
main meals following distributions. Thus daily for-This paper reviews feed factors that influence
age intake is closely related to the amount eaten feeding behaviour, dietary choices and ultimately
during main meals (Jarrige et al., 1995). During a nutrient intake in sheep and goats. First, the main
main meal, the rate of intake is highest at the characteristics of feeding behaviour are described in
beginning and then decreases continuously as satia-terms of the satiation process and motivation to eat.
tion proceeds until satiety. Simple exponential Then, the ways in which vegetation characteristics
models accurately fit cumulative intake during meals influence ingestion of forage, feeding behaviour and
in both cows (Faverdin, 1985) and sheep (Baumont foraging decisions are reviewed. At pasture, two
et al., 1989). Initial rate of intake represents the different situations are considered: (i) situations
motivation to eat and the constant of deceleration of where animals are free to develop their own foraging
the exponential function the satiation process. Kinet-strategy; and (ii) situations where shepherds’ and
ics of intake were also modelled on fresh leafy animals’ strategies interact.
branches offered indoors to goats (Meuret, 1989) and can also be recorded at pasture by direct observation of feeding bouts. Rate of intake, especially at the
2. Feeding behaviour and control of intake beginning of the meal, seems to be a key factor for
understanding variations in voluntary intake between
Since the 1950s the feeding behaviour of rumin- forages (Moseley and Antuna Manendez, 1989)
ants fed indoors or at pasture has been extensively When fed indoors with two distributions and free
studied (see reviews by Jarrige et al., 1995; Ungar, access to forage, rumen fill, measured as the amount
1996). Intake is influenced primarily by hunger, of digesta in the rumen reaches a first maximum after
which is distressing, and by satiety, which is general- the morning main meal and the daily maximum after
ly pleasurable (Forbes, 1995). Regulation of feed the evening main meal (Baumont et al., 1988). In
intake and dietary choices combines short-term grazing sheep, Thomson et al. (1985) also observed
control of feeding behaviour related to the body a first maximum at 09:00 h and the daily maximum
capacity of the rumen is involved in the control of feeds allows anticipation of nutritional and
physio-intake is supported by two types of physiological logical consequences of intake. This is essential to
evidence: (i) stretch- and mechano-receptors are determine motivation to eat and feed preferences.
present in the rumen wall (Leek, 1977); and (ii)
increasing rumen fill with indigestible material by 2.2.1. Learning the post-ingestive effects of feeds
the equivalent of 1 kg dry digesta decreases intake The senses that are stimulated in the presence of
by 0.6 kg dry matter on average (review by Faverdin feed enable the animal to anticipate the postingestive
et al., 1995). When rumen fill is increased with effects of feed. The affective system integrates the
indigestible material, animals increase the number of taste of a feed with postingestive feedback and the
meals and the time spent ruminating per kg ingested cognitive system integrates the odour and appearance
(Baumont et al., 1990a). This stimulation of rumina- of the feed with its taste (Provenza et al., 1992).
tion behaviour, related to increased stimulation of Learned food aversions towards toxic plants or feeds
tactile receptors in the rumen wall, speeds up digesta experimentally laced with several compounds that
outflow and tends to reduce rumen fill. cause malaise have been clearly established in
During the main meals rapid fermentation of the ruminants (du Toit et al., 1991). Learned preferences
soluble fraction of feed increases osmotic pressure based on positive nutritional postingestive feedback
and VFA concentration of rumen fluid and lowers are undoubtedly also important for ruminants. After
´
pH (Remond et al., 1995). VFA infusion in the 10 days adaptation, lambs develop strong preferences
rumen decreases feed intake in the short term, more for non-nutritive flavours paired with glucose over
so as the molecular weight of the infused VFA is low flavours paired with saccharin (Burrit and Provenza,
(Faverdin et al., 1995). The influence of molecular 1992). Thus ruminants, like other mammals, develop
weight indicates involvement of osmotic pressure preferences for feeds that are richer in energy
(Carter and Grovum, 1990). Signals sensed by (Provenza, 1995). However, in a free-choice
situa-chemoreceptors in the rumen wall and / or in the liver tion, diet selection does not always maximise energy
enable the animal to avoid excess and nutritional density in the diet. Sheep eat some straw to prevent
disorders. This may explain the low duration of main rumen disorders, even when a more concentrated
meals with grass silages that contain large amounts feed is also on offer (Cooper et al., 1995). Sheep fed
of organic acids and ammonia, especially when they with a ‘long-fibre-free’ diet will eat 10-mm
poly-are poorly preserved (Van Os et al., 1995). In this ethylene fibre to restore normal rumination activity
case, satiation occurs before rumen fill reaches a (Campion and Leek, 1997). It was even observed
maximum. The same kind of adaptive phenomena that dairy goats could buffer seasonal variations in
may explain the choice of the feeds lowest in salt available vegetation composition. For example, at
when goats were given a diet rich in sodium chloride turn out, goats seek herbage species relatively low in
(Morand-Fehr et al., 1996). During a meal, the protein and rich in fibre; it is likely that animals
signals that contribute to the satiation process act select vegetation to reduce the variation in ingesta
simultaneously and probably additively as indicated composition as far as possible in the face of large
by the effects of an increase of rumen fill and an seasonal variations of vegetation composition
acetate infusion, performed separately or at the same (Fedele et al., 1993). Post-ingestive stimuli need to
time (Adams and Forbes, 1981). The different sig- be periodically reinforced, and so the animal has
nals that are sensed in the digestive tract and the regularly to re-evaluate the nutritional benefit of the
liver are integrated in the central nervous system and different choices.
balanced with other stimuli (Forbes, 1996). Dietary experience, particularly early in life,
modulates feeding behaviour and diet selection
2.2. Motivation to eat and feed preferences (Provenza, 1995). For example, naive sheep, cattle
and goats may graze up to 40% less than experienced
Satiation factors, mainly those related to rumen animals in the same environment (Provenza and
fill, have been favoured to explain forage intake. Balph, 1987). When grazing, animals adapt to their
plant species and their grazing skills (Flores et al., only 0.4 kg / day when they ate straw and received
1989), and by learning and memorising the dis- grass in the rumen but rose to 0.9 kg / day in the
tribution of resources (Dumont and Petit, 1998). reverse situation, although the digestibility of the
Learning from the dam and social partners is im- total diet was similar (Greenhalgh and Reid, 1971).
portant in the acquisition of dietary habits (Thor- Unpleasant sensations when previously eating straw
hallsdottir et al., 1990). Lambs accustomed to graze may explain its very low hedonic value. Hedonic
either clover or grass with their dam have a stronger behaviour can explain voluntary intakes greatly in
post-weaning preference for the species they had excess of requirements in wethers fed good quality
previously experienced (Ramos and Tennessen, forages (Baumont et al., 1997). The sensory
motiva-1992). After weaning, lambs reared by goats spent tion induced by a second distribution of fresh hay
less time eating clover than those reared by ewes, will override the satiety signals associated with the
consistent with the stronger preference for clover of first distribution (Baumont et al., 1990b). However,
ewes compared with goats (Orr et al., 1995). For the size of the second meal depends on the relative
adults, short-term neophilia may temporarily modu- quality of the two hays distributed. In goats, feed
late preferences. Sheep that had grazed either clover refusals may be explained by hedonic behaviour
or grass swards for 3 weeks subsequently showed a even when this selectivity prevents them meeting
stronger preference for the species previously lacking their energy requirements (Morand-Fehr et al.,
in their diet; nevertheless, after 3 days, they reverted 1991b). At pasture, the pleasant experiences
associ-to a stronger preference for the species they had been ated with the consumption of a new food have been
previously accustomed to eat (Parsons et al., 1994). suggested to explain preference for novelty or for
rarity (Newman et al., 1992). Hedonic behaviour,
2.2.2. Hedonic behaviour however, competes with the effort that has to be
Mechanisms of brain reward can to some extent expended to earn the reward. When animals have to
induce hedonic feeding behaviour, which competes walk to obtain a good forage, the preference for the
with physiological factors controlling intake. Sensory good forage depends on the amount offered the
properties of the feed will stimulate hedonic be- animals in reward (Dumont et al., 1998).
haviour to some extent. Total intake of sheep was In conclusion, Fig. 1 summarises the main
tionships between forage characteristics, short-term creases the amount of cell content, which is soluble,
control of intake, feeding behaviour and finally rapidly degraded and has almost no fill effect, and
forage intake. Modelling intake behaviour is a good increases the amount of cell walls. Consequently,
tool to test our knowledge of what controls intake. forage retention time in the rumen and thus fill effect
Forbes (1980) developed the first mechanistic model increases. In addition, tissue lignification increases
of intake in ruminants. More recently, Sauvant et al. the undegradable fraction of the cell walls and
(1996) proposed a mechanistic model of intake and decreases the degradation rate of the degradable
chewing activities that integrates relationships be- fraction (Grenet and Demarquilly, 1987). The time
tween feeding behaviour and digestive processes. needed to reduce particle size before ruminal escape
Decisions between eating, ruminating and resting are is also increased. Retention time in the rumen
taken according to the relative values of the func- depends mainly on the degradation rate of the
tions of motivation to eat and of satiety, which degradable fraction and on the proportion of the
integrate the signals described above. The forage is undegradable fraction, since its outflow rate does not
characterised by the cell wall content and its po- vary widely among plant species (Baumont et al.,
tential digestibility, and by the proportion of large 1997). Residence time of dry matter in the rumen is
particles. A palatability index and a coefficient of closely related to forage ingestibility (Baumont et al.,
heterogeneity take into account non-nutritional 1996).
characteristics and selection possibility in the forage
and are combined to estimate instantaneous 3.2. Sensory properties
palatability of the forage. This model accurately
predicted intake kinetics of sheep fed different types The sense of touch plays a role in the response of
of hay indoors. the animal to the feed (Arnold, 1966). Physical
characteristics of the forage such as dry matter content and particle size, and resistance to fracture
3. Role of plant characteristics on forage are known to affect ease of prehension and thus
´
ingestibility intake rate (Inoue et al., 1994). Sheep generally
prefer the feeds they can eat faster. With dried
Forage ingestibility is defined as the maximum forages, relative preferences for mixtures with
vary-quantity of the feed that can be eaten by the animal ing proportions of long and short particles were
when this is supplied ad libitum as the sole feed. closely related to the differences in intake rates
When given indoors, ingestibility of green forage (Kenney and Black, 1984).
depends mainly on its nutritive value and fill effect Small ruminants are also sensitive to the four
and on its sensory properties, assuming it does not primary tastes: sweet, salty, bitter and sour (Goatcher
contain toxic compounds (Fig. 1). Taking into and Church, 1970). However, animals’ response to
account the main variations in forage ingestibility added chemicals varies with feed management
allowed the development of the fill unit system to (Grovum and Chapman, 1988) and animals in the
predict feed intake for ruminants fed indoors (Jarrige same flock display a very wide range of response to
et al., 1986). the four taste components (Morand-Fehr et al.,
1993). The effects of various odoriferous compounds
3.1. Nutritive value and fill effect naturally present in plants were analysed by Arnold
et al. (1980) by sprinkling the chemicals onto cotton
For a given plant, ingestibility, like digestibility, is wool pads placed in the manger. Effects of
odorifer-dependent on the vegetation stage and the number of ous compounds are difficult to interpret because they
the vegetation cycle. During the first vegetation can vary in amplitude and sometimes in sign
accord-cycle, ingestibility decreases with the age of the ing to whether the animals are in choice situations or
plant (Demarquilly et al., 1981). The decrease in not. Moreover, several compounds recognised to
ingestibility with age of forage is the consequence of decrease intake by their odour or by their taste
the increase in its fill effect. As the plant ages, its (tannic acid, gramine, etc.) were found to decrease in
Low intake of silage is often attributed to low 1996). When an animal begins to eat, it selects a
palatability, since digestibility is only slightly differ- feeding site, and then a patch within that site.
ent from that of green forage. Effects of smell and
taste on silage intake were studied with anosmic and 4.1.1. Representation of the grazing process and
agustatory sheep (Michalet-Doreau, 1975). The in- theoretical bases of foraging behaviour
crease in silage intake by anosmic compared with Two approaches to foraging decisions have been
normal sheep was more pronounced with poorly proposed. Synthetic approaches assume animals
preserved (133%) than with well preserved silages organise their behaviour towards an objective,
(16.4%). Silage intake by agustatory sheep was not whereas analytical approaches assume that
behav-modified. Acetic acid added to the silage had a clear iours arise from cause–effect relationships. The basic
negative effect on intake (Buchanan-Smith, 1990). axiom of the main synthetic approach, Optimal
Amines are suspected to decrease palatability, be- Foraging Theory (OFT), is that present-day animals
cause in sheep initial eating rate at the beginning of forage optimally as a result of natural selection,
the meal was depressed by addition of amines in because optimal foraging enables the animal to
silage (Van Os et al., 1995). Low palatability of maximise its reproductive output (‘fitness’) (Krebs
silages probably results from learning the negative and McCleery, 1984). Fitness maximisation has
post-ingestive signals due to high amounts of fer- often been translated into efficiency of foraging,
mentation end products. which, for practical reasons, has often been equated
with short-term dry matter intake rate. The recent model of Newman et al. (1995) base foraging
4. Role of vegetation characteristics on grazing decisions on maximisation of fitness and include a
behaviour simplified mechanistic sub-model of digestion.
Analytical approaches explain foraging behaviour
Behavioural constraints are different and more in a complementary manner. For example, patch
diversified at pasture than indoor. Animals have to choice may be motivated by sensory stimuli, by
search for feed and they have to harvest the plants. post-ingestive feedback resulting from previous
Even so, animals are free at pasture to develop their choices, and by dietary experience. Synthetic and
own foraging strategy. They exploit environmental analytical approaches should be perceived as
com-heterogeneity by grazing selectively. Prediction of plementary rather than exclusive. Although OFT
intake and of the impact of animals on vegetation stresses the importance of natural selection in
de-needs an understanding of their foraging decisions. termining behaviour, it cannot exclude the
impor-When shepherds control their flocks on grazing tance of short-term dietary experiences.
Optimisa-circuit, their strategy of resource utilisation influ- tion-based predictions should, therefore, be
de-ences the animal’s behaviour and interacts with the veloped to include animals’ dietary experiences. The
foraging strategy of the flock. difficulty in dissociating the different approaches is
well illustrated by the question posed by Illius et al.
4.1. Foraging strategy and diet selection (1999): do animals eat a plant species faster because
they prefer it (sensory stimulus) or do they prefer it
On homogeneous swards, the importance of pre- because they can eat it faster (optimisation of
hension constraints first caused herbage intake to be behaviour)?
represented as the product of intake rate and grazing
time (Allden and Whittaker, 1970), assuming spatio- 4.1.2. How do vegetation characteristics influence
temporal stability of intake rate. As this approach is foraging decisions?
inappropriate on heterogeneous swards and
shrub-lands, a hierarchical approach is now preferred, that 4.1.2.1. Patch level
organises the grazing process into different levels A patch is defined as a spatial aggregation of bites
corresponding to animal decisions, which are then over which instantaneous intake rate (IIR) remains
integrated over wider spatio-temporal levels (Senft et relatively constant (Illius and Hodgson, 1996). Thus
paddock. IIR on the patch is determined by the mass the animal can graze the preferred patch without
of the bite and the functional relationship relating having to search for it, and (ii) those where
prefer-bite mass to the time required to handle it. Time per ences are modified by a cost of searching.
bite can be split into prehension and mastication When searching costs are negligible, animals
times. The animal is considered as subject to two generally prefer patches where they can eat rapidly
constraints: (i) the time required to prehend a bite, (Black and Kenney, 1984). However, preference is
which is independent of bite mass; and (ii) the time not absolute and intake rate is often less than
required to masticate the harvested material, which is predicted from optimisation theory. For example, on
proportional to bite mass (Newman et al., 1994; adjacent monocultures of grass and clover, the
Prache, 1997) and varies with plant species (New- proportion of clover in the diet of sheep is about
man et al., 1994). Bite mass is determined by the 70%, even though sheep generally eat clover faster
ease with which the sward can be gathered into the than grass (Newman et al., 1994). Why do animals
mouth and sheared (see review by Prache and express partial preference and choose mixed diets? In
Peyraud, 1997). On vegetative swards, sward height the short-term, the necessity for the animal to
and bulk density are the main determinants of bite consume the different patches to evaluate their
mass and IIR (Black and Kenney, 1984; Burlison et profitability, difficulty or lack of interest of the
al., 1991). On mature swards, bite mass is also animal in discriminating, and discrimination errors,
affected by the presence of ‘barrier’ components in have been proposed as explanations (Illius et al.,
the sward such as stems and dead material. Green 1999). On a daily scale basis, diurnal pattern in
leaf mass per unit area has been shown to be the best preferences, search for a balance of nutrients, avoid
predictor of bite mass and IIR across different excess of protein, and digestive constraints have
phenological stages of the sward (Penning et al., been proposed as explanations of partial preferences
1989; Prache et al., 1998). Sward height or green and mixed diets (Newman et al., 1995).
leaf mass do not take into account all the factors Preference is sensitive to relative vertical
availa-involved, because they are one- (sward height) or bility in different plant species. Animals may trade
two-dimensional (green leaf mass) descriptors, while off quality for quantity and switch to the less
bites are taken in three dimensions (Ungar, 1996). preferred item, when greater benefit is obtained on it
(Harvey and Orr, 1996; Prache et al., 1996). If one
4.1.2.2. Patch departure assumes that animals seek to maximise intake rate,
Forage depletion in the patch and expectation of the switch to the less preferred patch may be
intake opportunities in other patches will motivate predicted from potential intake rate on each patch.
the animal to move on. The animal has to make a For example, sheep rotationally grazing swards
trade-off between continuing to graze a patch where containing reproductive patches in a vegetative
back-it is experiencing diminishing marginal rewards, and ground, should consume the vegetative patches as
moving to another patch, thereby incurring a time long as their green leaf mass will allow a higher
cost. If the animal seeks to maximise intake rate, the intake rate than the reproductive ones (Fig. 2). Thus,
Marginal Value Theorem (MVT, Charnov, 1976) if animals seek to maximise intake rate, they should
predicts that the animal will leave the patch when switch to reproductive patches when the green leaf
intake rate within the patch equals the average intake mass on vegetative patches becomes lower than 300
21
rate for the whole environment. Demment et al. kg DM ha (i.e., 9 cm height in this study). This
(1993) and Laca et al. (1993) have globally validated result is in agreement with the preference
measure-MVT, but Bazely (1988) observed a longer residence ments of Dumont et al. (1995).
time than predicted by MVT. Searching constraints may limit intake rate. They
induce the animal to consume less preferred but
4.1.2.3. Multi patch level easier-to-reach food patches (Clarke et al., 1995) and
Vegetation characteristics, distance to water, need to develop search strategies: it may walk faster
for shelter, and social factors will then influence (Roguet et al., 1998), increase daily grazing time
patch choice. Considering vegetation characteristics, (Prache et al., 1998), or learn about the location of
quency to resources that had the same edible biomass relative to the distance to the patch (Dumont et al., 1998).
4.2. When a shepherd organises patch selection
Shepherding consists in interacting with sponta-neous animal’s decisions. The herder’s interventions could be considered simply as new constraints to the expression of the behavioural trends of the flock. An experienced shepherd will identify spontaneous be-havioural trends and avoid countering them too often. While building a strategy for both animal feeding and resource management, the shepherd has to some extent to follow the movements of the flocks to maintain the confidence of the animals. Effective management can upgrade dietary motivation on heterogeneous pastures (Meuret, 1993a,b).
Fig. 2. Prediction of diet choices based on dry matter intake rate
Shepherding interacts with the animals’ foraging
maximisation: an example on swards containing reproductive
patches (dotted curve) in a background of vegetative patches (solid response patterns (Senft et al., 1987), through
trans-curve) (from Prache et al., 1998). humance and nomadism practices, that both have
developed a wide range of informal and formal rules in seasonal forage uses. For animal nutrition
sci-cues (Edwards et al., 1997; Dumont and Petit, 1998). entists, there is now a need to devote more attention
The cost of searching is difficult to assess. Indirect to how a shepherd can manoeuvre an animal’s
indicators may be used, such as the animal’s degree appetite, taking full advantage of forage spatial
of selectivity expressed as the difference between heterogeneity and variability. That could help to
diet and sward composition, intake per distance reconsider the leader paradigm that ignores livestock
walked and proportion of potential bites encountered mobility characteristics and promotes settling process
removed (Prache et al., 1998; Roguet et al., 1998). A from land carrying-capacity evaluation (Scoones,
decrease in the horizontal availability may affect 1994).
encounter rate and therefore the proportion of pre- Rangelands are patchy environments. Within a
ferred species in the diet (Parsons et al., 1994). single day, a flock may well come across extremely
The effect of scale of patchiness may have marked diverse vegetation, and each of its members will
effects on animals’ foraging decisions, although it distribute thousands of bites over contrasted plant
has not, as yet, received much attention. Two types parts, from a tiny grass regrowth to a large mature
of patchiness may involve searching costs: small- leafy branch. Such broad diversity can motivate
scale, with constraints due to selection of preferred feeding. Firstly, it offers the animal many ways to
from less preferred food items, and larger scale with solve the usual conflicting problems: obtaining
maxi-constraints due to moving and locating alternative mal quality and adequate quantity (Owen-Smith and
patches. Fine mixtures reduce the opportunity for Novellie, 1981). Secondly, shrubs and leafy branches
selection. Sheep grazing grass / clover swards select can be easily prehended, allowing small ruminants
less clover when the species are offered in an high intake rate (Meuret, 1997). Finally, motivation
intimate mixture than in separate strips (Clark and is increased if the diversity of plant material on offer
Harris, 1985). Larger scale heterogeneity may im- during a grazing bout is high, especially when the
pose constraints on the visual perception of alter- material is of medium or low potential palatability
native patches and spatial memory. In an operant (Meuret and Bruchou, 1994).
succession of encounters of the different patches. A distinct roles during a circuit, and their use value for
model of shepherded circuit design, based on farm a specific phase during the meal is assessed
in-surveys together with intake kinetics analysis has dividually according to two simple criteria: the
been developed (Meuret, 1993a,b). The grazing relative abundance and the ‘instantaneous
palatabili-circuit may be viewed as an ordered sequence of ty’ (Sauvant et al., 1996) of the resources. A
patches offered during one main meal, the objective moderation-patch, with abundant but not highly
being to optimise the animal’s feeding motivation palatable plants, could be used at the beginning of
(Fig. 3). The ‘menu’ model is based on two rules: (i) the circuit to reduce the appetite of the flock. In
creation of synergetic sequences within the foraging contrast, an appetite promotion patch, with highly
bout; and (ii) follow spontaneous behavioural trends palatable but rare resources, has to be used at the
while developing flock’s confidence in shepherd’s beginning to stimulate a flock that has a low appetite.
interventions. When intake must be stimulated in a The target-patch, with medium plant abundance and
particular patch (‘target-patch’) with rough and less palatability, is then used as a main-course, for the
palatable resources (e.g., a patch to be cleared of bulk of the meal. When animals are losing interest in
scrub), the shepherd has to detect and mobilise this patch, a booster-patch (see Fig. 3) has to be used
complementary and sometimes contiguous patches in to add diversity and revive appetite. When the
the feeding site. The different patches could play booster-action is successful, another target-patch
with slightly better instantaneous palatability than the 5. Conclusions
main course-patch is used for a second course. At the
end, if the shepherd considers there is not enough From the different research fields reviewed in this
time for a new booster-course sequence, a dessert- paper (control of intake, determinants of forage
patch, with high plant abundance and palatability, ingestibility, feeding behaviour and diet selection at
can be used. pasture, shepherding practices) we can propose a
The organisation of a ‘menu’ by a shepherd, that framework for a mechanistic model of intake
be-takes full advantage of a patchy environment and haviour and patch choice at the feeding site level
emphasises the concept of instantaneous palatability, (Fig. 4). The balance between motivation to eat and
seems consistent with the role of aversions in diet satiation controls eating decision. The expression of
selection at pasture hypothesised by Provenza this eating decision, patch choice and intake rate,
(1996). This shows that the feed value of such an depends on the patch characteristics and among them
environment results largely from proper feeding mainly the potential intake rate and the spatial
management at the level of the feeding site and distribution.
feeding bout that could manoeuvre the appetite. Meal size and diet composition is mainly
con-Experienced shepherds observe their flock’s attitudes trolled by anticipation of postingestive effects to
to evaluate initial hunger, intermediate disaffection avoid nutritional excesses or deficiencies. The role
for food and signs of satiety. A shepherd who played by sensory properties of the diet
(‘palatabili-succeeds in designing a menu is able to generate a ty’) is still not completely understood as it is partly
‘build-up’ pastoral value from a heterogeneous and confounded with the role of nutritive value after
variable vegetation (Meuret, 1996). learning, but it is essential in explaining the control
of feeding behaviour. Ruminants generally develop reduce further decline in forage quality. Predicting
preferences for feeds that provide a high satiety level diet selection is further complicated in more complex
rapidly. Feeds that can be ingested fast and that are situations such as rangelands, with a large diversity
rapidly and highly digested are very palatable pro- on offer, making the environment more difficult for
vided they do not contain toxic compounds. Never- the animal to perceive, and offering it opportunities
theless, for a given nutritive value, sensory properties for trade-offs between behavioural options.
of the feed per se can stimulate or depress hedonic behaviour and thus intake. Offering choice to
ani-mals may stimulate forage intake level and so References
forestall some pathological problems. The nutrient
requirements of each individual could be more Adams, G.B., Forbes, J.M., 1981. Additivity of effects of ruminal
precisely met. However, research is needed to de- acetate and either portal propionate or rumen distension on
food intake in sheep. Proc. Nutr. Soc. 40, 44A.
termine appropriate conditions for choice feeding.
Allden, W.G., Whittaker, A.M.C., 1970. The determinant of
A more efficient feed utilisation necessitates
im-herbage intake by grazing sheep: the interrelationship of factors
proved prediction of forage ingestibility. Residual influencing herbage intake and availability. Aust. J. Agric. Res.
variation in predictive models of voluntary intake 21, 755–766.
based on nutritional characteristics (i.e., energy and Arnold, G.W., 1966. The special senses in grazing animals. 2.
Smell, taste and touch and dietary habits in sheep. Aust. J.
nitrogen values, fill effect) remains generally high.
Agric. Res. 17, 531–542.
Fill effect of forage can be expressed as its ruminal
Arnold, G.W., De Boer, G., Boundy, 1980. The influences of
retention time and can be predicted using in situ odour and taste on food preferences and food intake of sheep.
degradability or laboratory methods. A so-called Aust. J. Agric. Res. 31, 571–585.
‘hedonic value’ of feed is probably much more Bailey, D.W., Gross, J.E., Laca, E.A., Rittenhouse, L.R.,
Coughenour, M.B., Swift, D.M., Sims, P.L., 1996. Mechanisms
difficult to predict, and to a first approximation can
that result in large herbivore grazing distribution patterns. J.
be assessed by the difference between observed and
Range Manage. 49, 386–400.
predicted intake. Baumont, R., Dulphy, J.P., Andrieu, J.P., 1988. Comportement
At pasture, intake, diet composition and the impact alimentaire et etat de repletion du reticulo-rumen chez le´ ´ ´ ´
` ´
of grazing on the vegetation is the result of a mouton nourri a volonte de foin de prairie ou de luzerne avec
` ´ ˆ
acces en continu ou limite: incidences sur le controle physique
complex interaction between the animal and the
de l’ingestion. Reprod. Nutr. Dev. 28, 573–588.
vegetation. The livestock farmer has to organise this
Baumont, R., Brun, J.P., Dulphy, J.P., 1989. Influence of the
interaction within a pastoral strategy adapted to local nature of hay on its ingestibility and the kinetics of intake
conditions. Through selective grazing, the animal during large meals in sheep and cow. In: Jarrige, R. (Ed.),
consumes a diet of a higher nutrient quality than that XVIth International Grassland Congress, Nice, France. French
Grassland Society, pp. 787–788.
on offer, and distributes its impact on the vegetation.
Baumont, R., Malbert, C.H., Ruckebusch, Y., 1990a. Mechanical
Nevertheless, the determinants of foraging decisions
stimulation of rumen fill and alimentary behaviour in sheep.
remain a matter of debate. In simple experimental Anim. Prod. 50, 123–128.
situations (very short-term, simple dietary choices), Baumont, R., Seguier, N., Dulphy, J.P., 1990b. Rumen fill, forage
animals tend to maximise their rate of food intake. palatability and alimentary behaviour in sheep. J. Agric. Sci.
115, 277–284.
However, short-term prediction of foraging
behav-Baumont, R., Barlet, A., Jamot, J., 1996. L’effet d’encombrement
iour is not sufficiently validated on a daily scale
´ ´ ruminal des fourrages: sa relation avec l’ingestibilite et etude
basis, which may be affected by digestive con- de sa prevision au laboratoire. Renc. Rech. Ruminants 3,´
straints, balance of nutrients, diurnal pattern of diet 313–316.
selection, sampling, social effects within the flock, Baumont, R., Dulphy, J.P., Jailler, M., 1997. Dynamic of
vol-untary intake, feeding behaviour and rumen function in sheep
flock-farmer interactions. Prediction of total intake is
fed three contrasting types of hay. Ann. Zootech. 46, 231–244.
further impaired by the difficulty of predicting
Bazely, D.R., 1988. Foraging behaviour of sheep (Ovis aries L.)
grazing time. General understanding of diet selection grazing on swards of perennial ryegrass (Lolium perenne L.),
may nevertheless be sufficient to predict switches PhD Thesis. Oxford University UK, 180 pp.
´
sheep. II. Height and density of pasture. Aust. J. Agric. Res. E., Farce, M.H., Theriez, M., Demarquilly, C. (Eds.), Proceed-35, 565–578. ings of the IVth International Symposium on the Nutrition of Buchanan-Smith, J.G., 1990. An investigation into palatability as Herbivores, Recent Developments in the Nutrition of
Herbi-a fHerbi-actor responsible for reduced intHerbi-ake intHerbi-ake of silHerbi-age by vores, INRA Editions, Paris, pp. 95–120.
sheep. Anim. Prod. 50, 253–260. Fedele, V., Pizillo, M., Claps, S., Morand-Fehr, P., Rubino, R., Burlison, A.J., Hodgson, J., Illius, A.W., 1991. Sward canopy 1993. Grazing behaviour and diet selection of goats on native structure and the bite dimensions and bite weight of grazing pasture in Southern Italy. Small Ruminant Res. 11, 305–322. sheep. Grass Forage Sci. 46, 29–38. Flores, E.R., Provenza, F.D., Balph, D.F., 1989. Role of ex-Burrit, E.A., Provenza, F.D., 1992. Lambs form preferences for perience in the development of foraging skills of lambs non-nutritive flavors paired with glucose. J. Anim. Sci. 70, browsing the shrub serviceberry. Appl. Anim. Behav. Sci. 23,
1133–1136. 271–278.
Campion, D.P., Leek, B.F., 1997. Investigation of a fibre appetite Forbes, J.M., 1980. A model of the short-term control of feeding in sheep fed a long fibre-free diet. Appl. Anim. Behav. Sci. 52, in the ruminant: effects of changing animal or feed
characteris-79–86. tics. Appetite 1, 21–41.
Carter, R.R., Grovum, W.L., 1990. A review of the physiological Forbes, J.M., 1995. In: Voluntary Food Intake and Diet Selection significance of hypertonic body fluids on feed intake and in Farm Animals, CAB International, Wallingford, UK, p. 532. ruminal function: salivation, motility and microbes. J. Anim. Forbes, J.M., 1996. Integration of regulatory signals controlling Sci. 68, 2811–2832. forage intake in ruminants. J. Anim. Sci. 74, 3029–3035. Charnov, E.L., 1976. Optimal foraging, the Marginal Value Goatcher, W.D., Church, D.C., 1970. Review of some nutritional
Theorem. Theor. Pop. Biol. 9, 129–136. aspects of the sense of taste. J. Anim. Sci. 31, 973–981. Clark, D.A., Harris, P.S., 1985. Composition of the diet of sheep Greenhalgh, J.F.D., Reid, G.W., 1971. Relative palatability to
grazing swards of differing white clover content and spatial sheep of straw, hay and dried grass. Br. J. Nutr. 26, 107–116. distribution. New Zealand J. Agric. Res. 28, 233–240. Grenet, E., Demarquilly, C., 1987. Rappels sur la digestion des
´
Clarke, J.L., Welch, D., Gordon, I.J., 1995. The influence of fourrages dans le rumen (parois) et ses consequences. In: ´
vegetation pattern on the grazing of heather moorland by red Demarquilly, C. (Ed.), Les Fourrages Secs: Recolte, Traite-deer and sheep. I. The location of animals on grass / heather ment, Utilisation, INRA Editions, Paris, pp. 141–162. mosaics. J. Appl. Ecol. 32, 166–176. Grovum, W.L., Chapman, H.W., 1988. Factors affecting the Cooper, S.D.B., Kyriasakis, I., Nolan, J.V., 1995. Diet selection in voluntary intake of food by sheep 4. The effect of additives sheep: the role of the rumen environment in the selection of a representing the primary tastes on sham intakes by diet from two feeds that differ in their energy density. Br. J. oesophageal-fistulated sheep. Br. J. Nutr. 59, 61–72. Nutr. 74, 39–54. Harvey, A., Orr, R.J., 1996. Dietary preference of sheep for grass
´
Demarquilly, C., Andrieu, J., Weiss, P., 1981. L’ingestibilite des and clover at contrasting sward surface heights. Br. Soc. Anim.
´ ´
fourages verts et des foins et sa prevision. In: Prevision de la Sci., Winter Meeting, 161.
Valeur Nutritive des Aliments des Ruminants, INRA Publi- Illius, A.W., Hodgson, J., 1996. Progress in understanding the cations, Versailles, pp. 155–157. ecology and management of grazing systems. In: Hodgson, J., Demment, M.W., Distel, R.A., Griggs, T.C., Laca, E.A., Deo, G.P., Illius, A.W. (Eds.), The Ecology and Management of Grazing 1993. Selective behaviour of cattle grazing ryegrass swards Systems, CAB International, Wallingford, UK, pp. 429–458. with horizontal heterogeneity in patch height and bulk density. Illius, A.W., Gordon, I.J., Elston, D.A., Milne, J.D., 1999. Diet In: Proc. of the XVII Int. Grass. Congress, pp. 712–714. selection in grazing ruminants: A test of rate maximization. du Toit, J.T., Provenza, F.D., Nastis, A.S., 1991. Conditioned food Ecology 80, 1008–1018.
´
aversions: How sick must a ruminant get before it detects Inoue, T., Brookes, I.M., John, A., Kolver, E.S., Barry, T.N., toxicity in foods? Appl. Anim. Behav. Sci. 30, 35–46. 1994. Effects of leaf shear breaking load on the feeding value Dumont, B., Petit, M., 1998. Spatial memory of sheep at pasture. of perennial ryegrass (Lolium perenne) for sheep. 2. Effects on Appl. Anim. Behav. Sci. 60, 43–53. feed intake, particle breakdown, rumen digesta outflow and Dumont, B., D’Hour, P., Petit, M., 1995. The usefulness of animal performance. J. Agric. Sci. 123, 137–147.
grazing tests for studying the ability of sheep and cattle to Jarrige, R., Demarquilly, C., Dulphy, J.P., Hoden, A., Robelin, J., exploit reproductive patches of pastures. Appl. Anim. Behav. Beranger, C., Geay, Y., Journet, M., Malterre, C., Micol, D., Sci. 45, 79–88. Petit, M., 1986. The INRA ‘fill unit’ system for predicting the Dumont, B., Dutronc, A., Petit, M., 1998. How readily will sheep voluntary intake of forage-based diets in ruminants: a review. J.
walk for a preferred forage ? J. Anim. Sci. 76, 965–971. Anim. Sci. 63, 1737–1758.
Edwards, G.R., Newman, J.A., Parsons, A.J., Krebs, J.R., 1997. Jarrige, R., Dulphy, J.P., Faverdin, P., Baumont, R., Demarquilly, ´
Use of cues by grazing animals to locate food patches: an C., 1995. Activites d’ingestion et de rumination. In: Jarrige, R., example with sheep. Appl. Anim. Behav. Sci. 51, 59–68. Ruckebusch, Y., Demarquilly, C., Farce, M.H., Journet, M.
´ `
Faverdin, P., 1985. Regulation de l’ingestion des vaches laitieres (Eds.), Nutrition des Ruminants Domestiques, INRA Editions,
´ `
en debut de lactation. These de Doctorat INA Paris-Grignon, Paris, pp. 123–181.
France, 131 pp. Kenney, P.A., Black, J.L., 1984. Factors affecting diet selection by Faverdin, P., Baumont, R., Ingvartsen, K.L., 1995. Control and sheep. I. Potential intake rate and acceptability of feed. Aust. J.
Krebs, J.R., McCleery, R.H., 1984. Optimization in behavioural Owen-Smith, N., Novellie, P., 1981. What should a clever ungulate ecology. In: Krebs, J.R., Davies, N.B. (Eds.), Behavioural eat? Am. Nat. 119, 151–178.
Ecology: An Evolutionary Approach, Sinauer Associates, Parsons, A.J., Newman, J.A., Penning, P.D., Harvey, A., Orr, R.J., Sunderland, MA, pp. 91–121. 1994. Diet preference of sheep: effects of recent diet, physio-Laca, E.A., Distel, R.A., Griggs, T.C., Deo, G., Demment, M.W., logical state and species abundance. J. Anim. Ecol. 63, 465–
1993. Test of optimal foraging with cattle: the marginal value 478.
theorem successfully predicts patch selection and utilisation. Penning, P.D., Parsons, A.J., Hooper, G.E., Orr, R.J., 1989. In: Proc. of the XVII Int. Grass. Congress, pp. 709–710. Responses in ingestive behaviour by sheep to changes in sward Laca, E.A., Ortega, I.M., 1995. Integrating foraging mechanisms structure. In: Jarrige, R. (Ed.), XVIth Int. Grass. Congress,
across spatial and temporal scales. In: Fifth Int. Rangeland Nice, France French Grassland Society, pp. 791–792. Congress, pp. 129–132. Prache, S., 1997. Intake rate, intake per bite and time per bite of Leek, B.F., 1977. Abdominal and pelvic visceral receptors. Br. lactating ewes on vegetative and reproductive swards. Appl.
Med. Bull. 33, 163–168. Anim. Behav. Sci. 52, 53–64.
´ ´ ˆ ´
Meuret, M., 1989. Valorisation par des caprins laitiers de rations Prache, S., Peyraud, J.L., 1997. Prehensibilite de l’herbe paturee ´ ´
ligneuses prelevees sur parcours: feuillages, fromages et flux chez les bovins et les ovins. INRA Prod. Anim. 10, 377–390.
´ ´ ` ´
ingeres. These de Doctorat en Sciences Agronomiques, Faculte Prache, S., Roguet, C., Louault, F., Petit, M., 1996. Evolution des
´ ´ ´ ´
des Sciences Agronomiques de Gembloux, 229 pp. choix alimentaires d’ovins entre talles vegetatives et epiees au
ˆ ´ ´
Meuret, M., 1993a. Piloter l’ingestion au paturage. Et. Rech. Syst. cours de l’exploitation d’un couvert epie de dactyle. Renc.
Agric. Dev. 27, 161–198. Rech. Ruminants 3, 89–92.
`
Meuret, M., 1993b. Les regles de l’Art: garder des troupeaux au Prache, S., Roguet, C., Petit, M., 1998. How degree of selectivity ˆ
paturage. Et. Rech. Syst. Agric. Dev. 27, 199–216. modifies foraging behaviour of dry ewes on reprodutive Meuret, M., 1996. Organizing a grazing route to motivate intake compared to vegetative sward structure. Appl. Anim. Behav.
on coarse resources. Ann. Zootech. 45 (Suppl.), 87–88. Sci. 57, 91–108.
´ ´
Meuret, M., 1997. Prehensibilite des aliments chez les petits Provenza, F.D., 1995. Role of learning in food preferences of ruminants en landes et sous-bois. INRA Prod. Anim. 10, ruminants: Greenhalgh and Reid revisited. In: Engelhardt, W.V., 391–401. Leonhard-Marek, S., Breves, G., Giesecke, D. (Eds.),
Proceed-´
Meuret, M., Bruchou, C., 1994. Modelisation de l’ingestion selon ings of the Eighth International Symposium on Ruminant
´ ´ ´ `
la diversite des choix alimentaires realises par la chevre au Physiology, Ruminant Physiology: Digestion, Metabolism, ˆ
paturage sur parcours. Renc. Rech. Ruminants 1, 225–228. Growth and Reproduction, Ferdinand Enke Verlag, Stuttgart, Michalet-Doreau, B., 1975. Recherches sur les causes des varia- pp. 233–247.
´ ´ ´
tions des quantites d’ensilage d’herbe ingerees par les rumin- Provenza, F.D., 1996. Acquired aversions as the basis for varied
` ´ ´
ants. These Docteur-Ingenieur Universite de Nancy, 99 pp. diets of ruminants foraging on rangelands. J. Anim. Sci. 74, Morand-Fehr, P., Ben Ayed, M., Hervieu, J., 1993. Individual 2010–2020.
responses of goats to main taste components included in feeds. Provenza, F.D., Balph, D.F., 1987. Diet learning by domestic In: Proceedings FAO-CIHEAM-EU Workshop. Aristotle Uni- ruminants: theory, evidence and practical implications. Appl. versity, Thessaloniki (Greece) 24–26 Sept. 1993, Sheep and Anim. Behav. Sci. 18, 211–232.
Goat Nutrition, pp. 41–45. Provenza, F.D., Pfister, J.A., Cheney, C.D., 1992. Mechanisms of Morand-Fehr, P., Ben Ayed, M., Hervieu, J., Lescoat, P., 1996. learning in diet selection with reference to phytotoxicosis in
Relationship between palatability and rate of intake in goats. herbivores. J. Range. Manage. 45, 36–45.
In: Lindberg, J.E., Gonda, H.L. (Eds.), Recent Advances in Ramos, A., Tennessen, T., 1992. Effect of previous grazing
´ ´
Small Ruminants Nutrition. Options Mediterraneennes, Vol. 34, experience on the grazing behaviour of lambs. Appl. Anim. CIHEAM, Paris, pp. 121–123. Behav. Sci. 33, 43–52.
´ `
Moseley, G., Antuna Manendez, A., 1989. Factors affecting the Remond, B., Brugere, H., Poncet, C., Baumont, R., 1995. Le ´
eating rate of forage feeds. In: Jarrige, R. (Ed.), XVIth Int. contenu du reticulo-rumen. In: Jarrige, R., Ruckebusch, Y., Grassl. Congress, Nice, France French Grassland Society, pp. Demarquilly, C., Farce, M.H., Journet, M. (Eds.), Nutrition des
789–790. Ruminants Domestiques, INRA Editions, Paris, pp. 253–298.
Newman, J.A., Parsons, A.J., Harvey, A., 1992. Not all sheep Roguet, C., Prache, S., Petit, M., 1998. Feeding station behaviour prefer clover: diet selection revisited. J. Agric. Sci. 119, 275– of ewes in response to forage availability and sward
phenologi-283. cal stage. Appl. Anim. Behav. Sci. 56, 187–201.
Newman, J.A., Parsons, A.J., Penning, P.D., 1994. A note on the Sauvant, D., Baumont, R., Faverdin, P., 1996. Development of a behavioural strategies used by grazing animals to alter their mechanistic model of intake and chewing activities of sheep. J. intake rates. Grass Forage Sci. 49, 502–505. Anim. Sci. 74, 2785–2802.
Newman, J.A., Parsons, A.J., Thornley, J.H.M., Penning, P.D., Scoones, I., 1994. In: Living With Uncertainty: New Directions in Krebs, J.R., 1995. Optimal diet selection by a generalist Pastoral Development in Africa, Intermediate Technology grazing herbivore. Functional Ecol. 9, 255–268. Publications, IIED, London.
Thomson, B.C., Cruickschank, G.J., Poppi, D.P., Sykes, A.R., Ungar, E.D., 1996. Ingestive behaviour. In: Hodgson, J., Illius, 1985. Diurnal patterns of rumen fill in grazing sheep. Proc. A.W. (Eds.), The Ecology and Management of Grazing Sys-New Zealand Soc. Anim. Prod. 45, 117–120. tems, CAB International, Wallingford, UK, pp. 185–218. Thorhallsdottir, A.G., Provenza, F.D., Balph, D.F., 1990. Ability Van Os, M., Dulphy, J.P., Baumont, R., 1995. The effect of protein
of lambs to learn about novel foods while observing or degradation products in grass silages on feed intake and intake participating with social models. Appl. Anim. Behav. Sci. 25, behaviour in sheep. Br. J. Nutr. 73, 51–64.
