Implications

Nutrition is an important aspect of welfare. However, in intensive livestock production systems animals receive monotonous rations or plant monocultures, frequently with excesses or imbalances of nutrients which compromise welfare. Exploring the feeding behavior of ruminants when offered a variety of feeds has the potential to improve their health and well-being by reducing levels of stress and fear and providing early detection of pain and illness. Management strategies in ruminant production systems should benefit by acknowledging the animals’ role as active players in feeding systems and by incorporating feeding behavior as a measure on how well homeostasis is achieved.

Behavior by consequences and homeostatic regulation

Food preference (behavior) in herbivores develops because of the dynamic interplay between flavor and post-ingestive feedback (consequences), which is determined by an animal’s physiological condition and a plant’s chemical characteristics (Provenza, Reference Provenza1996; Provenza and Villalba, Reference Provenza and Villalba2006). Taste, as well as smell and sight, are labels which animals use for discriminating among foods. Post-ingestive feedback calibrates a food’s flavor with its homeostatic utility (consequences). If a particular food provides chemicals, which are required by an animal at a certain point in space and time, then preference for that food will increase. In contrast, if a food supplies chemicals which provide negative effects to the body or compounds which are not needed in that particular point in space and time (e.g. excess nutrients), preference for the food will decline, although selection becomes less efficient in complex feeding environments (Duncan et al., Reference Duncan, Elwert, Villalba, Yearsley, Pouloupoulou and Gordon2007). On this basis, herbivores preferentially ingest foods high in nutrients (Kyriazakis and Oldham, Reference Kyriazakis and Oldham1993; Villalba and Provenza, Reference Villalba and Provenza1997a) and low in toxins (Provenza et al., Reference Provenza, Burritt, Clausen, Bryant, Reichardt and Distel1990), but they avoid specific nutrients when needs are met (Villalba and Provenza, Reference Villalba and Provenza1999) and they eat toxins when toxins provide medicinal (e.g. antiparasitic) effects (Lisonbee et al., Reference Lisonbee, Villalba, Provenza and Hall2009; Villalba et al., Reference Villalba, Provenza, Hall and Lisonbee2009a). Thus, the process of food selection in herbivores can be interpreted as the search for chemicals in the environment that restore homeostasis allowing animals to stay well (Villalba and Provenza, Reference Villalba and Provenza2007).

Varied diets and animal welfare

Generalist herbivores achieve homeostasis by consuming a diverse diet. This is in part because of the fact that no one plant species can provide for all of the animals’ nutrient demands (Westoby, Reference Westoby1978), and because generalist herbivores are unable to detoxify large amounts of chemically similar plant secondary metabolites (Freeland and Janzen, Reference Freeland and Janzen1974). Yet, herbivores consume a diversity of foods even when nutritional needs can be met with single foods or when toxins are not a concern (e.g. Wilmshurst et al., Reference Wilmshurst, Fryxell and Hudson1995; Provenza et al., Reference Provenza, Scott, Phy and Lynch1996). Other reasons for varied diets are sampling to update the information about available foods (Westoby, Reference Westoby1974) and increases in hedonics associated with switching to alternative foods (Rolls et al., Reference Rolls, van Duijvenvoorde and Rowe1983). It has also been argued that the preference for variety may result from aversions to any food eaten too frequently or in excess (Provenza, Reference Provenza1996). The process of ingesting a food causes animals to satiate on that food, and the satiety hypothesis attributes decreases in food preference because of transient food aversions caused by flavors, nutrients and toxins interacting along with the concentration gradients (Provenza, Reference Provenza1996).

Gustatory, olfactory and visual neurons stop responding to the taste, odor and sight of a particular food eaten to satiety, yet they continue to respond to other foods (Critchley and Rolls, Reference Critchley and Rolls1996). Preference for a specific food declines when the food is eaten repeatedly, presumably because the taste of food consumed to satiety becomes less pleasurable (Swithers and Hall, Reference Swithers and Hall1994). Aversions become pronounced when foods contain high concentration of toxins or nutrients or nutrient imbalances. Thus, although undue restrictions of nutrients are recognized as important adverse influences on animal welfare (Ewing et al., Reference Ewing, Lay and von Borell1999), excesses of nutrients and nutrient imbalances are also stressful and compromise animal well-being.

Ruminants can encounter pasture forages excessive in concentrations of protein, which degrades rapidly in the rumen, and so far as 73% of the ingested nitrogen may be excreted (Van Vuuren and Meijas, Reference Van Vuuren and Meijas1987). Excessive nitrogen concentrations can lead to ammonia build-up in the blood resulting in loss of appetite, infertility and death (Lobley and Milano, Reference Lobley and Milano1997). Likewise, feedlot rations with high concentrations of grain produce acidosis with negative effects on intake and animal health (Huber, Reference Huber1976). Lambs prefer flavored-straw paired with low-to-moderate doses of energy (e.g. the volatile fatty acids propionate or acetate) or nitrogen (urea and casein) delivered to the rumen, which do not cause satiety. However, at higher doses of energy or nitrogen, they become averse to that flavored straw (Villalba and Provenza, Reference Villalba and Provenza1996, Reference Villalba and Provenza1997a and Reference Villalba and Provenza1997b).

Aversions also result when foods are deficient in nutrients and they occur even when a food is nutritionally adequate because satiety and surfeit are on a continuum from mild to stronger food aversions. Sensory-specific satiety refers to the decrease in preference for the taste of food as it is consumed too frequently or in excess (Rolls et al., Reference Rolls, Rolls, Rowe and Sweeney1981; Rolls, Reference Rolls1986). For instance, cows and sheep were fed with the same foods but with distinctive flavors one day (e.g. coconut) prefer a different flavor (e.g. maple) the next day (Early and Provenza, Reference Early and Provenza1998; Scott and Provenza, Reference Scott and Provenza1998; Atwood et al., Reference Atwood, Provenza, Wiedmeier and Banner2001a). The reward experienced by the hedonic value of a specific food (i.e. ‘liking’), involves different neurotransmitter systems than the reward experienced by its nutrient value (i.e. ‘wanting’, Berridge and Robinson, Reference Berridge and Robinson1998). This hedonic value is important in influencing food choice, and may play a role in animal welfare. A decrease in liking because of repeated sensory stimulation by a single food may stimulate animals in the search for food diversity. Likewise, sensory-specific satiety seems to be a reflection of both a reduction in liking and in wanting, (Balleine and Dickinson, Reference Balleine and Dickinson1998) which primes animals to search for alternative food sources.

In summary, animals experience flavor-nutrient toxin-specific satieties, satiety to single foods is mildly to strongly aversive, and that causes them to eat a variety of foods (Provenza, Reference Provenza1996; Provenza et al., Reference Provenza, Villalba, Dziba, Atwood and Banner2003). Thus, monotonous diets typically used in intensive livestock feeding systems such as total-mixed rations (TMRs) or monocultures – even when nutritionally balanced – have the potential to cause aversions and consequently be stressful, which could depress intake and negatively impact animal welfare. There is evidence that psychological stress affects eating patterns in humans, with shifts in preferences toward more palatable foods of higher energy density (Oliver and Wardle, Reference Oliver and Wardle1999; Lim et al., Reference Lim, Norman, Clifton and Noakes2009), suggesting a variety of foods may help people to cope with the stress. In contrast, restriction of variety per se may be stressful. Rats exposed to attractive, but inaccessible food cues manifest higher levels of stress as measured by concentration of corticosterone (Polivy et al., Reference Polivy, Herman and Coelho2008). Generalist herbivores evolved eating a variety of foods in diverse landscapes and this scenario is typically not replicated in confinement or at pasture. Needs of individuals differ because of environment and physiological state. The nutrient content of plants also varies during the growing season. Allowing animals to select their own diet may improve their productivity and well-being as single foods and monotonous rations may not completely meet the needs of individuals within the herd on a daily basis, and may even provide excesses of nutrients (Manteca et al., Reference Manteca, Villalba, Atwood, Dziba and Provenza2008). Beef calves offered a mixed ration of rolled barley, rolled corn, corn silage and alfalfa hay, or a choice among these foods offered individually manifest differences in intake and performance (Atwood et al., Reference Atwood, Provenza, Wiedmeier and Banner2001b). During the trial, no two animals consistently chose the same ingredients, and none of them selected a diet of similar composition to the nutritionally balanced mixed ration, yet animals offered the choice that had similar protein:energy ratios. On average, as animals offered the mixed-ration and they gained weight at a similar rate and gain per unit food consumed (Atwood et al., Reference Atwood, Provenza, Wiedmeier and Banner2001b). When grain and silage were offered as a choice to feedlot cattle, they consistently chose lower ratios of grain to silage (75% to 80% grain) than those offered to controls in a TMR (85% to 90% grain), but they had significantly lower dry matter intake, and higher overall feed efficiency than animals given the TMR (Zobel et al., Reference Zobel, Schwartzkopf-Genswein, Veira, Genswein, Zobel, Gibb and von Keyserlingk2006).

There may be other reasons besides flavors, nutrients and toxins for herbivores to display varied diets. Sheep, cattle and goats show a consistent diurnal pattern of food selection, with a strong preference for legumes in the morning, and an increased preference for grass over the course of the day (reviewed by Rutter, Reference Rutter2006; Hill et al., Reference Hill, Chapman, Cosgrove and Parsons2009). One explanation for this pattern of behavior has been attributed to a predator avoidance strategy aimed at reducing food consumption at night. Such a strategy involves eating a bulky feed rich in fiber with a low rate of passage (i.e. grass) in the evening, helping to maintain rumen fill and reducing the need to eat to maintain rumen fill at night (Rutter, Reference Rutter2006). Sheep appear to target fiber in the evening (Dumont et al., Reference Dumont, D’hour and Petit1995), and restricting the allowance of roughage in cattle significantly increases the development and frequency of oral stereotypies (Redbo and Nordbald, Reference Redbo and Nordblad1997). It has also been argued that deviations from the animal’s optimal supply of fiber generate discomfort, which the animal tends to minimize to improve well-being (Forbes, Reference Forbes2003). It has also been proposed that ruminants will choose a diet associated with the maintenance of optimal conditions in the rumen (Cooper et al., Reference Cooper, Kyriazakis and Oldham1996). Thus, availability of roughage (and other feeds) at certain points in space and time may allow animals to express their foraging preferences, to minimize the discomfort and enhance their welfare.

Foraging behavior, plant secondary compounds and animal health

Plants provide herbivores with an array of chemicals with the potential to improve health and well-being. For instance, phytochemicals – known as secondary compounds – which protect plants from consumers and pests adversely affect cellular and metabolic processes in herbivores (Rosenthal and Janzen, Reference Rosenthal and Janzen1979; Cozier et al., Reference Cozier, Clifford and Ashihara2006), but at low doses and in appropriate mixtures, they can have beneficial effects on animal nutrition and health (Lozano, Reference Lozano1998; Engel, Reference Engel2002). For instance, eating plants high in tannins is a way for herbivores to reduce internal parasites (Min and Hart, Reference Min and Hart2003) and tannins alleviate bloat by binding to proteins in the rumen (Waghorn, Reference Waghorn1990). By making plant protein unavailable for digestion and absorption until it reaches the more acidic abomasum, tannins also enhance the nutrition by providing high-quality protein to the small intestines (Barry et al., Reference Barry, McNeill and McNabb2001). Plant-derived alkaloids, terpenes and phenolics also have antiparasitic properties (Kayser et al., Reference Kayser, Kiderlen and Croft2003; Hocquemiller et al., Reference Hocquemiller, Cortes, Arango, Myint, Cave, Angelo, Munoz and Fournet1991), and sesquiterpene lactones have anti-tumorigenic, anti-amebic, anti-bacterial, anti-fungal and cardiotonic properties (Picman, Reference Picman1986; Robles et al., Reference Robles, Arguellin, West and Rodriguez1995; Huffman et al., Reference Huffman, Ohigashi, Kawanaka, Page, Kirby, Gasquet, Murakami and Koshimizu1998). Some complex carbohydrates (i.e. galactomannans, glucans and arabinogalactans) and flavonoids stimulate macrophage function and have immunomodulatory properties (Tizard et al., Reference Tizard, Carpenter, McAnalley and Kemp1989; Middleton et al., Reference Middleton, Kandaswami and Theoharides2000; Schepetkin and Quinn, Reference Schepetkin and Quinn2006). The biochemical diversity of plants thus offers animals the opportunity to enhance their health and well-being, provided such diversity is present in their diets. For instance, of the 192 recognized plant food species routinely consumed by Mahale chimpanzees in Africa, many have ethnomedicinal uses (Huffman et al., Reference Huffman, Ohigashi, Kawanaka, Page, Kirby, Gasquet, Murakami and Koshimizu1998). Eating diverse diets thus provides herbivores all the advantages of bioactive compounds such as anti-parasitic agents and immunomodulators on a daily basis with health-maintenance effects.

Herbivores exposed to varied diets may also learn about the benefits of specific plant secondary compounds for alleviating certain maladies. In their quest for substances in the environment, which contribute to homeostasis, animals may experience the medicinal properties of specific foods. For instance, wild chimpanzees suffering from parasite-related diseases with overt signs and symptoms of sickness consume the bitter pith of the anti-parasitic plant Vernonia amygdalina, even when they have available ad libitum species of much higher nutritive value (Huffman and Seifu, Reference Huffman and Seifu1989; Huffman, Reference Huffman2003). Consistent with this observation, sheep also increase intake and preference for tannin-containing foods when they are infected with endoparasites relative to when parasitic loads are not present (Lisonbee et al., Reference Lisonbee, Villalba, Provenza and Hall2009; Villalba et al., Reference Villalba, Provenza, Hall and Lisonbee2009a). Other studies suggest livestock self-medicate by grazing specific medicinal plants when ill (Grade et al., Reference Grade, Tabuti, Van Damme and Arble2007 and Reference Grade, Tabuti and Van Damme2009).

Foraging behavior and experience with food

Past experiences in utero and early in life have life-long influences on behavior, but we seldom notice because we know or remember so little about the history of any animal. Experiences with food have life-long influences on diet, health and disease in humans and herbivores by causing neurological, morphological and physiological changes, which influence on foraging behavior (Provenza and Balph, Reference Provenza and Balph1990; Distel et al., Reference Distel, Villalba and Laborde1994 and Reference Distel, Villalba, Laborde and Burgos1996). By interacting with the genome during growth and development, social and biophysical environments influence gene expression and behavioral responses (LeDoux, Reference LeDoux2002; Fish et al., Reference Fish, Shahrokh, Bagot, Caldji, Bredy, Szyf and Meaney2004; Meaney, Reference Meaney2009). Each individual emerges from the ongoing interplay between its genes and social and the biophysical environments, and much of this dialog occurs during development in utero and early in life. An important form of this interaction, termed predictive adaptive responses, refers to responses that are (i) induced by the environment early in life, (ii) cause changes neurologically, morphologically and physiologically, and (iii) confer survival advantages when the environment of rearing matches the environment where a young animal then lives (Gluckman et al., Reference Gluckman, Hanson and Spencer2005). Predictive adaptive responses act through developmental plasticity early in life to modify the phenotype so it matches the environment of rearing, which is expected to be inhabited later in life. This will be so (i) if the behavior of mother is appropriate for the post-weaning environment and (ii) if that environment does not change too drastically during the life of the offspring.

Given the importance of experience on animal physiology and behavior, it follows that it may also have an impact on the welfare of animals. For instance, exposure to novelty (i.e. lack of experience) is one of the most potent experimental conditions to cause negative emotions of fear and anxiety (Boissy, Reference Boissy1998). This is probably why animals typically sample novel foods cautiously, a phenomenon known as food neophobia (Chapple and Lynch, Reference Chapple and Lynch1986). Excessive fearfulness can cause chronic stress and reduce productivity and therefore reducing the level of fearfulness has advantages for both animal welfare and production (Jones, Reference Jones1997). Selecting less fearful livestock improves maternal behavior and animal survival (Kilgour and Szantar-Coddington, Reference Kilgour and Szantar-Coddington1997). Fearfulness may also affect feeding behavior and the acceptance of new foods, and three lines of evidence suggest such a relationship in ruminants. First, the neural substrate responsible for the fear response is also involved in diet selection. Neuronal networks in the amygdala and the hippocampus are involved in both fear and diet selection responses (Bechara et al., Reference Bechara, Tranel, Damasio, Adolph, Rockland and Damasio1995). Second, when sheep are placed in a new environment likely to elicit a stress response, they show a greater reluctance to eat novel foods compared with the same animals offered new food in a familiar environment (Burritt and Provenza, Reference Burritt and Provenza1997). Finally, work with wild and farm animals shows environmental enrichment that allows animals to show more flexible foraging behavior decreases chronic stress (Carlstead and Shepherdson, Reference Carlstead and Shepherdson2000). A study carried out in our laboratory to determine whether there was a link between general fearfulness (as measured by the open field test) and stress-induced hyperthermia (SIH) and the readiness of sheep to eat new foods suggests that individuals less responsive to social isolation (lower number of bleats) may be less cautious at accepting novel foods than individuals more responsive to social isolation, but no correlation was found between SIH and food neophobia (Villalba et al., Reference Villalba, Manteca and Provenza2009b).

From a management and welfare perspective, it is important that animals readily accept new foods. Overcoming food neophobia by adding a familiar flavor to food (Launchbaugh et al., Reference Launchbaugh, Provenza and Werkmeister1997) or ameliorating neophobia through repeated exposure to new foods, or exposing pregnant mothers (Nolte and Provenza, Reference Nolte and Provenza1992a and Reference Nolte and Provenza1992b), or individuals early in life (Distel et al., Reference Distel, Villalba and Laborde1994 and Reference Distel, Villalba, Laborde and Burgos1996) are all ways to improve intake of novel foods and reduce fearfulness and stress within the herd.

Feeding behavior as criteria for measuring welfare

Feeding behavior may also represent an important tool for detecting pain or sickness in livestock, before overt symptoms are displayed. Early detection of pain, as well as its quantification could substantially contribute to its mitigation for improving the welfare of animals (Schwartzkopf-Genswein et al., Reference Schwartzkopf-Genswein, González, Gibb and McAllister2009). For instance, self-selection of analgesic drugs by animals considered to be in pain may be an important indication that they are in pain (Smith and Boyd, Reference Smith and Boyd1991). If the amount of analgesic consumed is proportional to the level of pain, then the amounts of analgesic selected could be used as an objective measurement of pain. There is some evidence that animals self-select their analgesics. Rats suffering from adjuvant-induced arthritis consumed more of an analgesic agent than control animals (Colpaert et al., Reference Colpaert, De Witte, Mondi, Awouters, Niemegeers and Jansen1980). Lame broiler chickens also learn to self-administer the analgesic carprofen in their feed and the amount of the analgesic consumed increased with the severity of lameness (Danbury et al., Reference Danbury, Weeks, Chambers, Waterman-Pearson and Kestin2000).

Pain not only may affect selection and intake of analgesics but feeding patterns. Monitoring feeding patterns is a useful tool for the early detection of sickness in livestock. For instance, pain associated with castration results in calves making fewer visits to the feeder while daily feeding duration does not differ from control animals (Schwartzkopf Genswein et al., Reference Schwartzkopf-Genswein, González, Gibb and McAllister2009). Sick steers in feedlots, identified as having bovine respiratory disease, spend 30% less time feeding than healthy steers (Sowell et al., Reference Sowell, Bowman, Branine and Hubbert1998).

Assessing the links between food choice and animal welfare

Our thesis is that food variety allows animals to express their foraging preferences, which in turn enhances their welfare. Behavioral measurements may represent an important tool to assess such an effect in livestock. For instance, it would be possible to explore the willingness or motivation of an animal to get access to multiple foods. Livestock are prepared to do extra work (i.e. pay a cost such as walking) to gain access to preferred foods (Dumont et al., Reference Dumont, Dutronc and Petit1998). Thus, it may be possible to quantify the motivation of animals (e.g. distance walked; number of lever-press responses) to obtain a diversity of feeds in their diets.

Lack of food alternatives could lead to frustration behavior which in extreme situations may develop into stereotypies (Broom, Reference Broom1986). These stereotyped behaviors are persistent, recurrent actions (i.e. fence-pacing and head-weaving) which can be linked to types of automatism or repetitive comfort behaviors (Gregory, Reference Gregory2004). Compulsive water drinking can be a sign of frustration or anxiety (Gregory, Reference Gregory2004); a behavior which could also give insights into the welfare of animals fed monotonous diets. Increases in exploratory behavior (e.g. rooting in pigs) may also be indicative of the effects of stress caused by boredom (Van Roijen, Reference Van Roijen1991), and potentially food monotony. Play behavior (e.g. running) has also been suggested as a useful measure of how the welfare of ruminants is affected by feeding practices (e.g. milk allowance) (Krachun et al., Reference Krachun, Rushen and de Passille2010).

Physiological measurements such as estimation of corticotropic responses are useful in monitoring fear and stress in farm animals and may represent an important tool in the future for complementing behavioral parameters to improve the assessment of welfare in animals fed with different types of diets. The hypothalamic pituitary-adrenal (HPA) axis is activated during aversive situations. The level of cortisol in the blood increases in pigs, cattle and sheep when they were subjected to a painful procedure like castration (e.g. Coetzee et al., Reference Coetzee, Gehring, Bettenhausen, Lubbers, Toerber, Thomson, Kukanich and Apley2007). The HPA is also activated in animals separated from familiar peers, mixed with unknown animals, transported and restrained in a shoot (Boissy and Le Neindre, Reference Boissy and Le Neindre1997). In addition, corticotropin-releasing hormone produced by the hypothalamus in response to a stress situation has a direct effect on reducing food intake (Matteri et al., Reference Matteri, Carroll and Dyer2000), suggesting a link between feeding behavior and activation of the HPA axis. The HPA axis is also activated during frustration and anxiety in mammals (Dantzer et al., Reference Dantzer, Arnone and Mormède1980; Landgraf et al., Reference Landgraf, Wigger, Holsboer and Neumann1999) and lack of food alternatives may lead to such states in herbivores. Glucocorticoid hormones can be measured in plasma, saliva, urine and feces (Berman et al., Reference Berman, Johnson, Phillips and Barry1980). The compound 11,17-dioxoandrostanes, a group of cortisol metabolites, have been used with fecal sample to evaluate the adrenocortical activity (Fell and Shutt, Reference Fell and Shutt1986) and cortisol metabolites increase in cattle feces after transport and after adrenocorticotropin administration (Palme et al., Reference Palme, Robia, Messmann, Hofer and Mostl1999).