HAL Id: hal-00889798https://hal.archives-ouvertes.fr/hal-00889798

Submitted on 1 Jan 1999

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

A behavioural approach to feeding broilersMichel Picard, Michèle Plouzeau, Jean Michel Faure

To cite this version:Michel Picard, Michèle Plouzeau, Jean Michel Faure. A behavioural approach to feeding broilers.Annales de zootechnie, INRA/EDP Sciences, 1999, 48 (4), pp.233-245. <hal-00889798>

Review

A behavioural approach to feeding broilers*

Michel Picard Michèle Plouzeau, Jean Michel Faure

Station de recherches avicoles, Inra, 37380 Nouzilly, France

(Received 24 August 1998; accepted 28 January 1999)

Abstract&mdash; Chicks use their beaks in the hatching process and pecking is one of the chick’s first activ-ities after hatching. Pecking is a precisely controlled action modulated by visual and tactile cues ofthe physical characteristics of feed particles. Post-ingestive nutritional effects of the feed are mem-orised and are coupled with sensory cues to let the chicken identify the feed. Some interactionsbetween nutritional factors (amino acid and energy balance, feeding regimen) and other environ-mental factors (temperature, light) are discussed in this review. Both the nutritional content andphysical characteristics of the diet in concert with other constraints such as climate modulate feed intakeand the rhythm and efficiency of pecking at feed by broilers. Further developments in feed millingprocesses that control the physical characteristics of the feed particles might facilitate the efficiencyof pecking (’feed prehension’) in general, and feed identification after a new feed delivery. Under com-mercial conditions, feed intake behaviour of chickens can be measured by focal sampling of broilersbetween two periods of rest or by scanning techniques. In the laboratory, detailed observations of peck-ing may complement our knowledge of the effects of the physical structure of feed on pecking andthe identification of the feed by broilers. The necessary validation of behavioural methods in com-mercial farms would benefit from increased interactions between field research and laboratory trials.(&copy; Elsevier / Inra)

broiler / feed intake / behaviour / nutrition / growth / environment / learning / adaptation

Résumé &mdash; Une approche comportementale de l’alimentation du poulet de chair. Les poussinsutilisent leur bec dès la naissance pour briser la coquille de l’oeuf et picorer est l’une de leurs premièresactivités de vie autonome. Le picorage est un ensemble cordonné d’actions précises modulées par lescaractéristiques physiques visuelles et tactiles des particules alimentaires. L’identification d’un ali-ment se construit par association de ses caractéristiques sensorielles avec ses effets post-ingestifsnutritionnels qui sont mémorisés. Quelques interactions entre facteurs nutritionnels (équilibres durégime en acides aminés et énergie, rationnement) et environnementaux (température, lumière) sontdiscutés à partir de travaux récemment publiés. Par exemple, le poulet peut s’adapter en plusieurs joursà des concentrations énergétiques très différentes du régime, si les aliments sont granulés ( 950 à

* An abbreviated version of this paper was presented at the 1998 European Poultry Congress inJerusalem, Israel.** Correspondence and reprints.Tel.: (33) 02 47 42 78 40; fax: (33) 02 47 42 77 78; e-mail: picard@tours.inra.fr

3 212 kcal-kg-1), en revanche, en climat chaud il reste incapable d’ajuster son ingéré énergétiquequelle que soit la teneur en acides aminés de l’aliment. Le bilan énergétique est la base de régulationde l’ingéré mais l’équilibre en acides aminés indispensables a plusieurs effets sur l’ingéré qui nesemblent pas explicables par la production de chaleur métabolique. Les poulets de chair ajustentleur budget temps en fonction des caractéristiques physiques et de la valeur nutritionnelle de l’alimenten équilibre avec d’autres contraintes environnementales comme le climat. Par exemple, les rythmesd’activité des poulets dépendent de l’équilibre en acides aminés indispensables de l’aliment maisaussi de la dureté des granulés et des fluctuations de la température. Dans les conditions de l’élevage,les caractéristiques biochimiques du régime sont calculées pour être nutritionnellement équilibréeset permettre une croissance musculaire rapide et efficace. Les caractéristiques physiques optimalesdes particules alimentaires sont connues avec moins de précision et l’observation du comportementdes animaux pourrait conduire à des ajustements technologiques pour, par exemple, mieux contrô-ler le temps passé à la mangeoire des animaux ou faciliter les transitions entre deux livraisons d’ali-ment. De même, les techniques de distribution des aliments (heure, rythme) peuvent faciliter l’adap-tation des poulets aux contraintes climatiques. Par exemple, un retrait de l’aliment avant les heuresles plus chaudes de la journée couplé à un éclairement nocturne, limite la mortalité en finition. Dansces deux domaines, technologie et rythmes de distribution de l’aliment, l’utilité des méthodes d’étudedu comportement doit être validée au niveau de l’élevage. Cela nécessite un développement desinteractions entre essais de laboratoire et mesures de terrain. (&copy; Elsevier / Inra)

poulet de chair / consommation d’aliment / comportement / nutrition / croissance / environnement /apprentissage / adaptation

1. INTRODUCTION

Feeding chickens has now reached a highdegree of precision in determining animalrequirements and biological availability ofnutrients. The major instrument used to for-mulate a well-balanced diet is the bio-chemical evaluation of the feed. However,chickens peck mainly according to the phys-ical characteristics of feed particles (meal,crumbs, pellets) that their sensory organsperceive.

Feed intake can be viewed as themedium-term result of feed pecking, whichis a learned process modulated by memoryof previously pecked particles. Observationof the behaviour of a broiler while eating ison a different time scale than feed intake asnutritionists measure it. Short-term varia-tions in feed intake are sometimes consid-ered as minor incidents. Nevertheless, fieldobservations suggest that sometimes chick-ens under-eat for several hours (e.g. after anew delivery of feed is distributed), orovereat during certain times of the day (e.g.in the morning in warm climates). Theseshort-term discrepancies have health impli-

cations such as diarrhoea (when broilers eattoo much litter material because theymisidentify the new feed delivery) or mor-tality (when broilers overeat before a thermalpeak).

Most of the current practical problemsthat exist in broiler production derive frominteractions between genetic, nutrition,pathology and housing conditions. Some ofthese are difficult to measure at researchstations because the reactions of the broilers

may depend on environmental factors (e.g.the feed delivery system differs largely inmost research situations from the farm). Itmay therefore be useful to develop tech-niques that might apply to feeding chickensat the farm level to analyse quantitativelythe problems and evaluate solutions.Behavioural observations may be part ofthese techniques. One aim of this paper is toprovide a framework of the study of feedingbehaviour of broiler considering succes-sively the various components of feed intake(figure 7).

Recent reviews on control of feed intake

[ 15] and feed intake behaviour of chickens

[58] substantiate the call made by Morris[45] for development of &dquo;adequate theorieswithout waiting for a fundamental under-standing of the complex and multidimen-sional system of appetite control in birds&dquo;.Using a behavioural perspective, somerecent publications on broiler nutrition arerevisited in the present review. New direc-tions of applied research will be based onspeculative interpretations quoted at the endof each section as ’implications’ to sepa-rate them from the review of published sci-entific results.

2. PECKING

Chicks hatch using their beaks and one oftheir first activities after hatching is pecking(see [62] for review). In the absence of aparent, pecking can be stimulated by otherchicks and even by an oscillating arrow [68].Using nutritive and non-nutritive stimuli,Hogan [33] showed that pecking is, at thebeginning of life, independent of hunger.Pecking followed by ingestion leads to theestablishment of a feeding experience, and

positive and negative feedback from theitems pecked are responsible for early feedselection [30]. Many studies of early feedpecking development have been done usingslow-growing chickens. Compared to junglefowl or leghorn chicks, fast-growing broilerchicks have earlier and more rapid devel-opment of feed intake behaviour and diges-tive capabilities [53] associated with a fasterresorption of residual yolk [69].

Pecking provides rewards in the form ofnutrients which help to maintain body tem-perature and allow growth of tissues. Peck-ing also provides sensory information tocomplement visual cues, and recent reportsshow that olfactory cues may be moreimportant in chicken nutrition than has beencommonly assumed (see [37] for review).Pecking may be viewed as one of the meansof sensory perception which are linked toingestion. It is a precisely adjusted and rapidmechanism which can only be fully under-stood when video-taped pictures of themovements made by the skull and the beakare examined in slow motion [5]. ] .

At normal speed, the observation of apecking suggests a continuous and indis-tinct repeated movement. In fact, the head ofa young broiler remains stationary for 75 %of the time during a continuous pecking ses-sion [77]. However, this relatively ’long’period for close observation of feed parti-cles between pecks lasts less than 1 s. This

period not only allows swallowing or‘mandibulations’, which occur only whena sufficient amount of feed has been grasped,but it is also a time when the chicken seemsto gather additional information about thefeed particles and chooses which particle topeck next. Yo et al. [77] showed that twoof three pecks of young chickens did notaim at or seize a feed particle, suggestingthat pecking is also associated with touchingand thereby ’exploring’ the feed [55].

Chickens go to the feed trough at irregu-lar intervals which can hardly be describedas meal feeding [64]. Feeding bout patterns(size, duration, intervals) can be biased by

experimental design and the experimenter’s sdefinition of a feeding bout, i.e. parametersused to define what is the maximum dura-tion of an eating pause allowed within thesame bout [10]. Feeding bouts are depen-dent both on the necessary energy home-ostasis of the body detected by productionand dissipation of heat [75] and on envi-ronmental factors. Under operant condi-

tioning conditions, when an animal has towork to obtain a reward, environmental con-straints, such as the necessity to peck severaltimes at a key for access to the feeder, reducethe randomness of the occurrence of feedingbouts [65]. In practice, social factors (com-petition and imitation, size of the group),feed management (frequency of filling ofthe trough, feed restriction), temperaturefluctuations and lighting programmes canchange feeding patterns [9, 51].

Analyses of organisation of feed peck-ing bouts over time may be useful for a moreprecise evaluation of the effects of envi-ronmental factors on time budgets of chick-ens. For example, when a diet is deficientin an essential amino acid, the time spentapparently eating increases but the feedingbouts are shorter [56], and when the tem-perature increases in the middle of the day ina poultry shed, the duration of the feeding

bouts is reduced [61 Thus, the measure-ment of the duration and frequency ofaccesses to a trough may indicate early ina poultry shed if broilers perceive and howthey adapt to an environmental constraint.Methods that can be used in practice arebriefly described at the end of section 5.

Implications

Detailed observations of peckingbehaviour at the laboratory level, using slow-motion video [77], may bring about a newevaluation of the feed itself, of its sensoryperception by chickens and of the conse-quences of different physical characteris-tics of the particles. If a change in the diet isdetected by chicks, they increase the dura-tion of the observation between pecks andthe number of effective pecks resulting inseizing a particle is reduced. A more detailedmeasurement of pecking in relation to thephysical characteristics of the feed particlesmight lead in the future to a model predict-ing the broiler responses to the feed struc-ture. A draft representation of a peckingmodel can be adapted for several domesticanimals on a similar time-scale when an ele-

mentary act (here one peck) is considered(figure 2).

3. FEED PARTICLES

The specificity of pecking in chickensand the sensitivity of the beak to differencesin the structure of the particles seized ortouched suggest that more attention needsto be given to the physical structure of feedparticles. Such information has direct impli-cations on the practical feed technology.

The colour [6, 74] and odour [70] of afeed are precisely perceived by broilers. Inchickens, each eye may have separate func-tions in feed identification and feed pre-hension due to brain lateralization [62, 71]. ].Size, shape and hardness (i.e. resistance tocrush) of the feed particles are importantfactors determined by feed milling. Youngbroilers prefer coarse particles of grain [50,52], and the size and hardness of feed par-ticles not only change the feed intake ofbroilers but also the development of theirdigestive tract, which may have severalmetabolic consequences (see [48] forreview). Providing feed as pellets changesnot only the particle size but the availabilityof the nutrients, such as carbohydrate in thefeed, may be affected by the temperatureincreases in the meal during processing [31]. ] .Contrary to early reports, however, physicalform appears to be the major reason forincreased feed intake and feed efficiency inbroilers fed pellets rather than mash, becausethese improvements are abolished by thegrinding of pellets [59]. When particles orcrumbs of the same batches of maize and

peas screened to the same size were intro-duced into a balanced diet (46 % maize and30 % peas), broilers immediately identifiedthe difference between particles and crumbs.They initially preferred the screened parti-cles, but after 2 days they progressivelychanged their preference to the crumbs [57].

Implications

The identification of a feed particle is theresult of sensory cues and associated post-ingestion information. Major traits in thesensory perception of the feed are: accurate

vision of detail and precise tactile sensitiv-ity of the beak within a complex and rapidpecking sequence [77]. In this context, asingle, complete feed might be consideredby chickens as the sum of feed particleswhose homogeneity at the trough level isquestionable.

Today the size of particles and quality ofthe manufactured pellets or crumbs aremainly designed to resist damage from thefeed distribution system in order to delivera homogeneous mix to the animal through.Feed technology of broiler diets require fur-ther consideration for various reasons:

- during their short life span, the beaks ofyoung broilers grow and hence their per-ception of the physical characteristics offeed particles may change with age;

- characteristics of feed particles such ashardness, roughness and elasticity may influ-ence the prehension of the feed and thus thetime spent eating and the feed intake levelunder production conditions [58];

- feed identification or recognition bybroilers seem mainly based on the physicalcharacteristics (visual and tactile) of the feedparticles (see section 7).

4. FEED LEARNING

Chickens have survived in a complexenvironment for millions of years by test-ing and selecting feed particles, none ofwhich individually provide a balanced nutri-tional supply, and some of which are toxic.Chickens, like most animals, learned thecharacteristics of edible materials in theirenvironment without chemical analyses andcomputerised calculations. For this purpose,they probably used detailed visual imagesof feed particles, repeated tactile experi-ences via the beak sensors, olfactory cuesand various sensory messages from the

digestive tract.Collier [13] described feed intake and

feed selection in animals as the result of a

dialogue between the &dquo;house economist&dquo;

and the &dquo;resident physiologist&dquo;. The &dquo;houseeconomist&dquo; represents the body of infor-mation gathered by observation and testingof feed sources available in the environmentand the effort necessary to ingest them. The&dquo;resident physiologist&dquo; symbolises the bio-logical demand for nutrients resulting inhunger. In modem broilers, the house econ-omy is reduced to a single balanced diet thatis provided constantly and the resident phys-iologist demand has been increased bygenetic selection for rapid growth.

Several foraging theories propose opti-misation of the benefit (intake)/cost (effortto eat) ratio in the medium term. However,domestic fowl work by pecking a disc togain access to a feed which is simultane-ously freely available in the same cage [ 16].This situation is called ’contrafreeloading’and has also been observed in rats and can-not be simply explained ’economically’ [26].The necessity for chicks to test frequentlypossible alternative feed resources or thenecessity to &dquo;find something to do&dquo; by cap-tive chickens kept in a homogeneous envi-ronment are candidate explanations to con-trafreeloading. More generally, the timewindow in which the eating decisions areoptimised by birds is still largely unknown[14].

Learning is easily demonstrated experi-mentally when broilers are given a choice ofdiets [24]. Free-choice feeding allows the useof whole grain cereals with a complemen-tary diet [22, 39, 54]. For example, the pro-portions of maize and complementary dietsconsumed by broilers from 15 to 42 days ofage varied between 73/27 when both werefed as a mash and 63/37 when maize wasfed as whole grain and the complementarydiet was pelleted [76]. These results sug-gest that broilers may not balance their dietonly on the basis of metabolic feedback, butalso on sensory stimuli and pecking cost.

Broilers provided with a choice of pro-tein-rich and energy-rich diets did not eatto maximise growth or feed efficiency. Atmarket weight they had consumed more

energy and less protein than required formaximum muscle deposition and they werefatter than broilers fed a complete diet [66].However, maximisation of meat yields withlow-fat content in the carcass is a motivationfor farmers and human consumers, but donot necessarily correspond to that of achicken at the beginning of its life.

Broiler chicks fed alternately a sub-defi-cient diet in essential amino acids and a diet

supplemented, by changing their feed everyday, needed 1 week to identify the diets (orthe rhythm of distribution). Although thereduction of intake observed after distribu-tion of a sub-deficient diet initially takesmore than 4 h, after 1 week of adaptation,the broilers reduced their intake within the1 st h of the sub-deficient diet being offered[56]. Similar results were obtained withalternate low- and high-protein feeds, lead-ing to the conclusion that broiler chicks cancompensate for unbalanced diet if &dquo;they areable to gauge the sensory properties of thefeeds&dquo; [25].

Implications

When the feeding experiences of broilersare varied by dietary choice or changes indiet, they are able to adapt, memorise and re-evaluate new items or feed particles whichhad negative consequences. When given achoice, chickens usually continue to sam-ple all edible materials available even if theyeat quantitatively less feed of low nutritivevalue or with toxic effects [34]. Suchbehaviour was probably adaptive under con-ditions where feed resources were incon-sistent, and of varying nutritional value ortoxicity level and chickens had to remaininformed of the available dietary alterna-tives. Contrafreeloading results might beexplained by such a quest for information.&dquo;Adequate&dquo; choices and/or feed intake lev-els require precise feed identification [23],and a paradigm where such &dquo;adequacy&dquo; canbe expressed [17]. In his review, Kyriazakis[38] considered that feeding behaviour offarm animals is &dquo;goal-orientated&dquo; and that

this assumption might bring a better under-standing of the way chicken, when given achoice, perceive the consequences of theirdiet on their comfort... if humans were ableto identify these chicken’s &dquo;goals&dquo;?

At present, free choice is not sufficientlyunderstood to risk allowing broilers todetermine their own diet, as they may notmatch the farmer’s goal to optimise profits.However, feed learning may be a majorissue because feed intake, at least on a short-term basis, may be based on memorised rep-resentations of a feed and its nutritionalvalue. This is considered in the followingthree sections on feed intake regulation, feedmanagement and feed neophobia.

5. FEED INTAKE

Broilers have a propensity to rest whichincreases with age [7]. This ’propensity’may be due to their high metabolic rate andheat production and/or inactivity linked tophysical constraints to locomotion. Eatingrepresents the main non-resting activity ofbroilers, and given their bulimic-like intakeassociated with rapid growth, factors whichpredominantly regulate their feed intakerequire further consideration.

Broilers can adapt their medium-termfeed intake to a wide range of feed types.For example, feeding pelleted diets vary-ing in metabolizable energy from 1 950 to3 212 kcal.kg-I did not change weight gainof broilers up to 49 days of age [40, 41]. Inseveral experiments with various nutritionalchallenges, an interval of approximately 1week was necessary for chickens to adjusttheir feed intake level [56, 64].

In other instances, broilers exposed tohigh environmental temperature were unableto adjust their feed intake regardless of theamino acid content of the diet [1, 11, 12].Net energy yield or heat production havebeen suggested as explanations for varia-tions in feed intake related to the proteinbalance of broiler diets [8]. Heat produc-

tion was not increased by higher dietary pro-tein levels in lean and fat broilers exposed to32 °C [27]. Macleod [43] could not mea-sure stimulation of heat production by thedegradation of excess amino acids and sug-gested a direct effect of the amino acid bal-ance on energy intake.

Two-week-old broilers rested for 76 %of the time when fed a balanced diet ad libi-tum and only 56 % when the same diet wassub-deficient in essential amino acids [56].The feed intake adjustment of broilers tothe amino acid balance of their diet is finelytuned and probably beyond what can be pre-cisely measured under current experimentalconditions. For example, using sound mod-els of interpretation, the requirements forsulphur amino acids [3] and for lysine [72]are higher for maximal feed efficiency thanfor fastest growth. Given the definition offeed efficiency (growth/intake ratio), suchdifferences can only be explained by a vari-ation in feed intake.

Implications

Although there are numerous models thatpredict growth responses from the nutrientintake, the rate of feed intake needs to beaccurately predicted and it is still a majorarea for future research [18]. The amountof feed ingested by a chicken depends onseveral factors in addition to the energy con-centration of the diet, which would benefitfrom a closer evaluation. In recent unpub-lished results it appeared that variations inthe fat content of broiler feed seem to inducealmost no feed intake adaptation, while adiluent such as fibre lead to a more preciseadjustment of the daily energy intake ofbroilers. In practice, feed intake of broilersis currently mostly regulated by environ-mental factors such as light, although, a fewyears ago, the introduction of a 5- to 6-h

night would have been almost completelycompensated for by overeating during thelight phase.

For technical reasons, short-term varia-tion in feed intake is not accurately con-

trolled on most broiler farms. However,behaviour of broilers within large commer-cial flocks can be measured by simplifiedor adapted methods [46, 47, 61]. Repeatedshort (10 min) scanning of the number ofchicken in a given activity, ie, eating fromone trough, give relatively consistent eval-uations of the time spent in one activity.Given the complexity of pecking [see sec-tion 2], this evaluation of the time spent atthe trough cannot be accurately correlatedwith feed intake. However, within a farm, achange in regularly registered values canprovide an early indication of a problem.Similarly, floor-oriented behaviours suchas pecking or scratching at the litter vary inintensity, depending on the nutritional bal-ance and the physical adequacy of the feed.

Focal observations of a single bird, takenat random within a large flock, were usedin order to follow the duration of its standuntil it once again rested. The duration of arecorded bout of activity depends on the ageof the broiler, on the type of activities per-formed during that bout and on environ-mental factors (bird density, temperature,equipment). Consistent results were obtainedwhen a sufficient number (n > 100) of boutswere recorded [61]. When practical deci-sions have to be taken, observation of thefocal birds performed systematically at sev-eral farms of one integration seems to be atool of potential value. The resultant behav-ioural data might complement records offeed input into a broiler shed (when avail-able) by providing early signals of non-adap-tation of the flock to its diet and/or its envi-ronment.

6. FEED MANAGEMENT

Feed management is a part of the envi-ronment of broilers and it has become aneffective way to modulate growth andimprove health. In practice, eating feed, themajor non-resting activity of broilers, canbe affected by factors such as flock density,the type of trough and the mode and fre-quency of feed distribution.

Feed restriction has been introduced inorder to reduce leg disorders and circula-tory problems of broilers by slowing earlygrowth [49, 5 1 ]. These increasingly frequentproblems seem linked to a genetic selection,which has been oriented towards the utili-sation of eaten nutrients for the growth ofspecific tissues such as gut and muscles [67J.Intermittent lighting [9] during the first2 weeks after hatching induces an earlydelay in growth which is followed by com-pensatory growth [79]. The compensatorygrowth period seems almost independent ofthe nutritional balance of the diet offered

[42], mainly during the week following thechange in environment as a result ofincreased feed intake. In some instances, adlibitum feeding of a high energy diet is notthe most efficient method to feed broilers.Feed restriction during the daylight can limitpeaks of metabolic heat production duringthe warmest part of the day and help broil-ers to adapt to hot climatic conditions [75,78].

Implications

The diet of the jungle fowl was proba-bly never a single feed, constantly avail-able, nor highly concentrated in nutrients.Feed management in the future mightinvolve diets composed of several distinctfeeds given according to a programme ofdistribution that would be adapted to envi-ronmental conditions and to the genotyperequirements. Practical examples exist inEurope where diets including whole graincereals simultaneously or sequentiallyoffered to broilers with a complementaryfeed are being used [63]. More precise andsystematic control of the behaviour of thechicken at the farm level would be usefulfor further developments of such feedingtechniques. Changes in farming practicesmight also be validated by a more carefulobservation of individual and flockbehaviours using the techniques brieflydescribed in section 5.

7. FEED NEOPHOBIA OR LACKOF FEED IDENTIFICATION?

In their 40 days of life a broiler pecksmore than 100 000 times at pellets (manymore if the diet is in mash form), and accu-mulate a significant sensorial experience oftheir feed. A minor change of diet can some-times induce periods of non-eating (severalhours). Are these transitory reactions dueto the fear induced by the change, a rela-tively well-described mechanism called neo-phobia (fear of novelty) or do the chickensnot eat for a time because they do not recog-nise or identify the feed as being an ediblematerial, a reaction which could be distinctfrom fear?

A new feed can be at the same timeattractive and induce reactions of fear asdemonstrated by Hogan when jungle fowlchicks received a meal worm [32]. Thegenetic origin of the chicken can changetheir propensity to express hesitancy to eatunfamiliar feed. In this experiment, pub-lished by Jones [35], the heavier line testedexpressed greater neophobia to an unusu-ally coloured diet than the lighter line. Thereis limited published work comparing neo-phobic responses of genotypes and there-fore it is difficult to generalise about theoccurrence of feed neophobia. However, itseems to be observed more frequently infast-growing birds such as turkeys and broil-ers than in slower-growing hybrids. Famil-iarisation to visual cues such as red-colouredwater facilitates the latter consumption ofa solution of vinegar [20, 21]. This suggeststhat feed identification is based on a com-bination of learned cues or that a familiarcue has a reassuring effect which mightreduce fear [36].

The reaction towards a new feed orunlearned aversion [44] is different fromthe learned aversion for a feed which hasbeen associated with an induced malaise,such as via an injection of lithium chloride(i.e. [28]). Broadening the sensorial expe-rience of a chicken by prior exposure to a

variety of feed colours, for example [35],can reduce the latency to peck to a new diet.This enrichment of the environment or ofthe sensorial experiences of a chicken isknown to have fear-reducing capacities [36].Under commercial conditions, stimulatingeffects of environmental factors on learn-

ing in broilers are not obvious. Changingthe environment by adding hanging mobileshas failed to improve feed intake and growthin broilers but does improve tibia strength[4]. Other experiments, however, suggestthat enrichment of the environment of broil-ers can improve their performance [29].

With increasing age, broiler chicks reducetheir exploration of the environment andtheir ability to detect hidden feed [73].Whether this reduced ability is due to areduction of locomotion linked to fast growthor to a insufficiently stimulating environ-ment that favours routine behavioural pat-terns is open to debate. Red or white lighthas stimulated activity, while blue or greenlight has had a calming effect and seemed tobe preferred [60]. In this experiment, as inothers, chickens preferred the conditionswith which they were familiar. Familiari-sation with the colour of the feed has beendescribed as &dquo;feed imprinting&dquo; by Bessei[6]. Familiarisation with a feed may rein-force the cues used by chickens to cate-gorise, which is mainly controlled by theleft hemisphere of the brain, although theright hemisphere is involved in response tonovelty which determines neophobia [2, 55].

Implications

Reactions of broilers to a new feed mightinclude a combination of simple non-iden-tification modulated by the familiarity withthe cues of the new feed and a fear reactionmodulated by the habituation of the chickento a homogeneous or enriched environment.

The chickens’ awareness of their envi-ronment is probably ’global’, e.g. simulta-neously built by different facets corre-sponding to combinations of various cues

of their environment, with feed being justone part. Given their sensory capabilities,a single complete feed which appears homo-geneous to humans might be considered bybroilers as an aggregate of different feedparticles. Environmental enrichment mightconsist of increasing the variety of feed par-ticles and sensorial cues offered to broilers

during their short life span. Given the impor-tance of early experiences in feed learning,special attention should be paid to the starterdiet that is given during a sensitive periodwhen there may be long-term memory ofthe sensorial experiences (Lynch J., per-sonal communication).

As the broiler ages, its perception of theenvironment focuses on its major non-rest-ing activity: eating. Thus, the aspect of thefeed, the ’feed image’, may become moredominant with age. Familiarisation and

learning have distinct roles in the feedingof broilers, with learning being essential tothe identification and ingestion of a feed.Familiarity with a long established eatingscheme in a homogeneous environmentmight hypothetically reduce the ability ofthe birds to accept a new feed [55].

8. CONCLUSION

The aim of nutritionists is to predict whata feed must contain in order to achieve opti-mal production. Medium-term feed intakeis the essential variable taken into accountfor this purpose. However, under produc-tion conditions, the short-term reactions ofbroilers to their diet may involve distinctmechanisms of adaptation, such as feed iden-tification and feed prehension (figure 3).Short-term rejection or non-identificationof a perfectly formulated diet are sometimesobserved. Behavioural criteria provide away to evaluate the adaptation of chickens totheir total environment (feed included), andtherefore, a behavioural approach to feedingbroilers may provide a complementary toolfor the understanding of feed identificationand feed prehension.

Pecking, a precise behaviour, requiresfurther detailed observation using slowmotion video techniques to identify the pre-cise characteristics of the feed that aredetected by the chickens. The duration ofthe observation period of the feed betweentwo pecks, the ratio between exploratory

and effective pecks, eating speed and tem-poral structures of feeding bouts are poten-tially new variables to explore. Also impor-tant to note is the sensorial cues of the feedand how broilers perceive them. Visual andtactile cues based on the physical charac-teristics of feed particles relevant to feedtechnology determine feed identificationand prehension.

Feed memorisation appears to be a learn-

ing process that interacts with other envi-ronmental factors conditioning the senso-rial experience of a chicken. Creatingfamiliarity with a known cue may be a wayto facilitate recognition of a feed and/or toreduce fear. Enrichment by increasing thediversity of the sensorial experiences ofbroilers at an early age may also decreasefear and aid adaptation to new diets. Feedand environment management offer an areaof investigation to facilitate adaptation ofthe future broiler genotypes. The improve-ments resulting from such investigationswill require intensified field research, includ-ing behavioural observations coupled withlaboratory trials.

Nutritionists have greatly improved thechemical composition of feeds. However,the physical characteristics of the feed par-ticles are also very important for the bird,and a research effort in this area may bebeneficial to short- and medium-term accept-ability of feed by broilers.

ACKNOWLEDGEMENTS

We are very grateful to C. Beaumont,G. Hinch, J. Lynch, I. Nir, J. Nolan and P. Siegelfor their essential comments on early versionsof the manuscript and to D. Raine for correctionof the English language.

REFERENCES

[ I ] Alleman F., Leclercq B., Effect of dietary pro-tein and environmental temperature on growthperformance and water consumption of malebroiler chickens, Br. Poult. Sci. 38 (1997)607-610.

!2j Alonzo Y., Lateralization of visual guidedbehaviour during feeding in zebra finches (Tae-niopygia gutiata), Behav. Process. 43 ( 1998)257-263.

! 3 ! Baker D.H., Fernandez S.R., Webel D.M., Par-sons C.M., Sulfur amino acid requirement andcystine replacement value of broiler chicks dur-ing the period three to six weeks posthatching,Poult. Sci. 75 (1996) 737-742.

[4 ! Balog J.M., Bayyari G.R., Rath N.C., Huff W.E.,Anthony N.B., Effect of intermittent activity onbroiler production parameters, Poult. Sci. 76(1997)6-12.

[51 Bermejo R., Zeigler H.P., Prehension in thepigeon. II. Kinematic analysis, Exp. Brain Res.75 (1989)577-585.

[61 Bessei W., Eine Untersuchung zur Futterpr5-gung bei Hiihnerkuken, in: Proceedings of theEuropean Poultry Science Congress, Hamburg,Germany, 1980, vol. 4, pp. 145-152.

[71 Bessei W., Das Verhalten von Broilern unterintensiven Haltungsbedingungen, Arch.

Gelflugelkd. 56 (1992) 1-7.[81 Boorman K.N., Ellis G.M., Maximum nutri-

tional response to poor-quality protein and aminoacid utilisation, Br. Poult. Sci. 37 (1996)145-156.

191 Buyse J., Simons P.C.M., Boshouwers F.M.G.,Decuypere E., Effect of intermittent lighting,light intensity and source on the performanceand welfare of broilers, World Poult. Sci. J. 52(1996) 121-130.

[ 10! Castonguay T.W., Kaiser L.L., Stern J.S., Mealpattern analysis, artifacts, assumptions and impli-cations, Brain Res. Bull. 17 (1986) 439-443.

[111 Cheng T.K., Hamre M.L., Coon C.N., Effect ofenvironmental temperature, dietary protein andenergy levels on broiler performance, J. Appl.Poult. Res. 6 (1997) 1-17.

[121 Cheng T.K., Hamre M.L., Coon C.N.,Responses of broilers to dietary protein levelsand amino acid supplementation to low proteindiets at various environmental temperatures,J. Appl. Poult. Res. 6 ( 1997) I 8-33.

[131 Collier G., The dialogue between the houseeconomist and the resident physiologist, Nutr.Behav. 3 (1986) 9-26.

[ 14! Collier G., Johnson D.J., The time-window offeeding, Physiol. Behav. 48 ( 1990) 771-777.

[ )5j Denbow D.M., Appetite and its control, Poult.Sci. Rev. 5 (1994) 209-229.

! 16J Duncan I.J.H., Hughes B.O., Free operant feed-ing in domestic fowls, Anim. Behav. 20 ( 1972)775-777.

[) 7] Emmans G.C., Diet selection by animals: theoryand experimental design, Proc. Nutr. Soc. 50( 1991 ) 59-64.

[181 Emmans G.C., Problems in modelling thegrowth of poultry, World Poult. Sci. J. 51 ( 1995)77-89.

[19] Faverdin P., Picard M., Pr6hensibilit6 des ali-ments par les herbivores et les volailles. Avant-propos, Inra Prod. Anim. 10 (1997) 375.

[20] Franchina J.J., Mere exposure to telereceptivecues facilitates intake of novel flavor in chickens(Gallus domesticus), Behav. Neural. Biol. 56(1991) 108-112.

[21] Franchina J.J., Slank K.L., Role of telerecep-tive and interoceptive (Taste) cues in ingestionalneophobia in chicks (Gallus domesticus), Behav.Neural. Biol. 52 (1989) 116-122.

[22] Forbes J.M., Covasa M., Application of dietselection by poultry with particular reference towhole cereals, World Poult. Sci. J. 51 (1995)149-165.

[23] Forbes J.M., Kyriazakis I., Food preferences infarm animals: why don’t they always choosewisely? Proc. Nutr. Soc. 54 (1995) 429-440.

[24] Forbes J.M., Shariatmadari F., Diet selectionfor protein by poultry, World Poult. Sci. J. 50(1994) 7-24.

[25J Forbes J.M., Shariatmadari F., Short-term effectsof protein content on subsequent diet selectionby chickens and the consequences of alternatefeeding of high- and low-protein foods, Br.Poult. Sci. 37 (1996) 597-607.

[26] Gallo A., Cuq C., « Contrafreeloading » et loca-lisation de 1’agent renforcateur, Behav. Process.14 (1987) 11-20.

[27] Geraert P.A., Guillaumin S., Leclercq B. Aregenetically lean broilers more resistant to hotclimate? Br. Poult. Sci. 34 (1993) 643-653.

[28] Gillette K., Martin G.M., Bellingham W.P., Dif-ferential use of food and water cues in the for-mation of conditioned aversions in domesticchicks (Gallus gallus), J. Exp. Psych. Anim.Behav. Proc. 6 (1980) 99-111.

[29] Gvaryahu G., Cunningham D.L., Van Thien-hoven A., Filial imprinting, environmentalenrichment, and music application effects onbehavior and performance of meat strain chick-ens, Poult. Sci. 68 (1989) 2l 1-217.

[30] Hale C., Green L., Effects of early ingestionalexperiences on the acquisition of appropriatefood selection by young chicks, Anim. Behav.36 (1988) 211-224.

[31] Hamilton R.M.G., Proudfoot F.G., Ingredientparticle size and feed texture: effects on the per-formance of broiler chickens, Anim. Feed Sci.Technol. 51 (1995) 203-210.

[32] Hogan J.A., An experimental study of conflictand fear: an analysis of behavior of young chickstoward a mealworm. 1. The behavior of chickswhich do not eat the mealworm, Behaviour 25(1965)45-95.

[33] Hogan J.A., Pecking and feeding in chicks,Learn. Motiv. 15 (1984) 360-376.

[34] Hogan-Warburg A.J., Hogan J.A., Feedingstrategies in the development of food recognitionin young chicks, Anim. Behav. 29 (1981)143-154.

[35] Jones R.B., Responses of domestic chicks tonovel food as a function of sex, strain and pre-vious experience, Behav. Process. 12 (1986)261-271. 1 .

[361 Jones R.B., Fear and adaptability in poultry:insights, implications and imperatives, WorldPoult. Sci. J. 52 (1996) 131-174.

[37] Jones R.B., Roper T.J., Olfaction in the domes-tic fowl: a critical review, Physiol. Behav. 62(1997)1009-1018.

[38] Kyriazakis I., The nutritional choices of farmanimals: to eat or what to eat? In: Forbes J.M.,Lawrence T.L.J., Rodway R.G., Varley M.A.(Eds.), Animal Choices, Occasional Publicationof the British Society of Animal Science, 20(1997) 55-65.

[39] Leeson S., Caston L.J., Production and carcassyield of broilers using free-choice cereal feeding,J. Appl. Poult. Res. 2 (1993) 253-258.

[40] Leeson S., Caston L., Summers J.D., Broilerresponse to energy or energy and protein dilutionin the finisher diet, Poult. Sci. 75 (1996)522-528.

[41] Leeson S., Caston L., Summers J.D., Broilerresponse to diet energy, Poult. Sci. 75 (1996)529-535.

[42] Leeson S., Zubair A.K., Nutrition of the broilerchicken around the period of compensatorygrowth, Poult. Sci. 76 (1997) 992-999.

[43] Macleod M.G., Effects of amino acid balanceand energy:protein ratio on energy and nitro-gen metabolism in male broiler chickens, Brit.Poult. Sci. 38 (1997) 405-411. 1 .

[44J Marples N., Roper T.J., Effects of novel colourand smell on the response of naive chickstowards food and water, Anim. Behav. 51 (1996)1417-1424.

[45] Morris T.R., Poultry science: the next 20 years?Br. Poult. Sci. 37 (1996) 5-14.

[46J Murphy L.B., Preston A.P., Food availabilityand the feeding and drinking behaviour of broilerchickens grown commercially, Br. Poult. Sci.29 (1988) 273-283.

[47J Murphy L.B., Preston A.P., Time-budgeting inmeat chickens grown commercially, Br. Poult.Sci. 29 (1988) 571-580.

[48J Nir I., The effects of food particle size and hard-ness on performance: nutritional, behavioral andmetabolic aspects, in: Proceedings of the 20thWorld’s Poultry Science Congress, New Delhi,India, 2-5 September 1996, vol. 2, pp. 173-184.

[49] Nir I., Interaction of genetic stocks, growth rate,feeding regime and metabolic diseases, in: Pro-ceedings of the lOth European Poultry Confer-ence, Jerusalem, Israel, 21-26 June 1998, vol. l,pp. 105-I12.

[50] Nir I., Melcion J.P., Picard M., Effect of particlesize of sorghum grains on feed intake and per-formance of young broilers, Poult. Sci. 69 (1990)2177-2184.

[51 ] Nir I., Nitsan Z., Dunnington E.A., Siegel P.B.,Aspects of food intake restriction in youngdomestic fowl: metabolic and genetic consid-erations, World Poult. Sci. J. 52 ( 1996) 251-266.

[52] Nir I., Shefet G., Aaroni Y., Effect of particlesize on performance. 1. Corn, Poult. Sci. 73(1994) 45-49.

[53] Noy Y., Sklan D., Posthatch development inpoultry, J. Appl. Poult. Res. 6 ( 1997) 344-354.

[54] Olver M.D., Jonker A., Effect of choice feed-ing on the performance of broilers, Br. Poult.Sci. 38 (1997) 571-576.

[55] Picard M., Broiler behaviour and nutritionalconditions, in: Proceedings of the I lth Euro-

pean Symposium on Poultry Nutrition, Faaborg,Denmark, 24-28 August 1997, pp. 175-180.

[56] Picard M., Faure J.M., Siegel P.B., Dunnin-gton E.A., Uzu G., Dalibard P., Food Intake andAmino Acids in Poultry, Rh6ne Poulenc Ani-mal Nutrition, Antony, France, 1994, 44 p.

[57] Picard M., Giboulot B., Melcion J.P., Miettesou eclats de graines d’aspects semblables sontidentifies par le jeune poulet, in : Proc. 2e Jour-n6es de la recherche avicole, Tours, France,8-10 avril 1997, vol. 2, pp. 197-200.

[58] Picard M., Melcion J.P., Bouchot C., Faure J.M.,Picorage et prdhensibilit6 des particules ali-mentaires chez les volailles, Inra Prod. Anim.10 ( 1997) 403-414.

[59] Plavnik I., Wax E., Sklan D., Hurwitz S., Theresponse of broiler chickens and turkey poults tosteam-pelleted diets supplemented with fat orcarbohydrates, Poult. Sci. 76 (1997) 1006-1013.

[60] Prayitno D.S., Phillips C.J.C., Omed H., Theeffects of color of lighting on behavior and pro-duction of meat chickens, Poult. Sci. 76 (1997)452!57.

[6l] 1 Richard P., Vilarifio M., Faure J.M., Leon A.,Picard M., Etude du comportement du pouletde chair dans un elevage intensif tropical auVenezuela, Rev. Elev. Med. Vet. Pays Trop. 50(1997) 65-74.

[62] Rogers L.J., The Development of the Brain andBehaviour in the Chicken, CAB International,Wallingford, UK, 1995, pp. 95-I 10.

163] Rose S.P., Fielden M., Foote W.R., Gardin P.,Sequential feeding of whole wheat to growingbroiler chickens, Br. Poult. Sci. 36 (1995)97-111. 1 .

! 64] Savory C.J., Feeding behaviour, in: Boorman K.N.,Freeman B.M. (Eds.), Food Intake Regulation inPoultry, Br. Poult. Sci., Edinburgh, UK, 1979,pp. 277-323.

[65] Savory C.J., Responses of fowls to an operantfeeding procedure and its potential use for reduc-ing randomness of meal occurrence, Physiol.Behav. 45 ( I 989) 373-380.

[66] Siegel P.B., Picard M., Nir I., Dunnington E.A.,Willemsen M.H.A., Williams P.E.V., Responses

of meat-type chickens to choice feeding of dietsdiffering in protein and energy from hatch tomarket weight, Poult. Sci. 76 (1997) 1183-1192.

j67] Siegel P.B., Dunnington E.A., Ressource allo-cations: growth and immune responses, in: Pro-ceedings of the lOth European Poultry Confer-ence, Jerusalem, Israel, 21-26 June 1998, vol. I,pp.98-98.

[68J Suboski M.D., Bartashunas C., Mechanisms forsocial transmission of pecking preferences toneonatal chicks, J. Exp. Psychol. Anim. Behav.Process. 10 (1984) 182-194.

[69] Turro I., Dunnington E.A., Nitsan Z., Picard M.,Siegel P.B., Effect of yolk sac removal at hatchon growth and feeding behavior in lines of chick-ens differing in weight, Growth Develop. Aging58(1994)105-112.

1701 Turro I., Porter R.H., Picard M., Olfactory cuesmediate food selection by young chicks, Phys-iol. Behav. 55 (1994) 761-767.

171] Vallortigara G., Regolin L., Bortolomiol G.,Tommasi L., Lateral asymmetries due to pref-erences in eye use during visual discriminationlearning in chicks, Behav. Brain Res. 74 (1996)135-143.

[72] Vazquez M., Pesti G.M., Estimation of the lysinerequirement of broiler chicks for maximum bodygain and feed efficiency, J. Appl. Poult. Res. 6(1997) 241-246.

1731 Vilarino M., Leon A., Faure J.M., Picard M.,Ability of broiler chicks to detect hidden food,Arch. Gelflugelkd. 62 (1998) 156-162.

[74] Weeks C.A., Brooks C., Coe G., Danbury T.C.,Effect of food and feeder colour on food con-

sumption of young layers and broilers, Brit.Poult. Sci. 38 (Supp1.) ( 1997) S 17.

1751 Wiernusz C.J., Teeter R.G., Acclimation effectson fed and fasted broiler thermobalance duringthermoneutral and high ambient temperatureexposure, Br. Poult. Sci. 37 (1996) 677-687.

1761 Yo T., Siegel P.B., Gudrin H., Picard M., Self-selection of dietary protein and energy by broil-ers grown under a tropical climate: effect of feedparticle size on the feed choice, Poult. Sci. 76(1997 ) 1467-1473.

I77J Yo T., Vilarino M., Faure J.M., Picard M., Peck-ing in young chickens: new techniques of eval-uation, Physiol. Behav. 61 (1997) 803-810.

[78J Zhou W.T., Yamamoto S., Effects of environ-mental temperature and heat production due tofood intake on abdominal temperature, shankskin temperature and respiration rate of broil-ers, Br. Poult. Sci. 38 (1997) 107-1 14.

[79] Zubair A.K., Leeson S., Compensatory growthin the broiler chicken: a review, World Poult.Sci. J. 52 (1996) 189-201. 1.