AN INVESTIGATION INTO THE MOTIVATION BEHIND THE ABNORMAL
BEHAVIOUR OF FEATHER PECKING IN LAYING HENS
A Thesis
Presented to
The Faculty of Graduate Studies
of
The University of Guelph
By
LAURA M. DIXON
In partial fulfillment of requirements
for the degree of
Doctor of Philosophy
February, 2008
© Laura M. Dixon, 2008 Library and Bibliotheque et 1*1 Archives Canada Archives Canada Published Heritage Direction du Branch Patrimoine de I'edition
395 Wellington Street 395, rue Wellington Ottawa ON K1A0N4 Ottawa ON K1A0N4 Canada Canada
Your file Votre reference ISBN: 978-0-494-47588-1 Our file Notre reference ISBN: 978-0-494-47588-1
NOTICE: AVIS: The author has granted a non L'auteur a accorde une licence non exclusive exclusive license allowing Library permettant a la Bibliotheque et Archives and Archives Canada to reproduce, Canada de reproduire, publier, archiver, publish, archive, preserve, conserve, sauvegarder, conserver, transmettre au public communicate to the public by par telecommunication ou par Plntemet, prefer, telecommunication or on the Internet, distribuer et vendre des theses partout dans loan, distribute and sell theses le monde, a des fins commerciales ou autres, worldwide, for commercial or non sur support microforme, papier, electronique commercial purposes, in microform, et/ou autres formats. paper, electronic and/or any other formats.
The author retains copyright L'auteur conserve la propriete du droit d'auteur ownership and moral rights in et des droits moraux qui protege cette these. this thesis. Neither the thesis Ni la these ni des extraits substantiels de nor substantial extracts from it celle-ci ne doivent etre imprimes ou autrement may be printed or otherwise reproduits sans son autorisation. reproduced without the author's permission.
In compliance with the Canadian Conformement a la loi canadienne Privacy Act some supporting sur la protection de la vie privee, forms may have been removed quelques formulaires secondaires from this thesis. ont ete enleves de cette these.
While these forms may be included Bien que ces formulaires in the document page count, aient inclus dans la pagination, their removal does not represent il n'y aura aucun contenu manquant. any loss of content from the thesis. Canada ABSTRACT
AN INVESTIGATION INTO THE MOTIVATION BEHIND THE ABNORMAL BEHAVIOUR OF FEATHER PECKING IN LAYING HENS
Laura M. Dixon Co-Ad visors: University of Guelph, 2008 Dr. I.J.H. Duncan & Dr. G.J. Mason
This thesis investigates the motivation behind the feather pecking behaviour common in the laying hen industry. The severe version of this pecking is a welfare concern, since feather removal is painful for the pecked bird and current solutions to this problem, such as beak-trimming and housing in wire cages to limit contact with conspecifics, are also detrimental to well-being. It has been hypothesized that feather pecking stems from frustrated motivation to forage or to dustbathe. However, the evidence to date is not conclusive, since foraging and dustbathing share similarities; it has thus not always been clear which motivation is being fulfilled when birds are interacting with a substrate. The objectives of this thesis were to (1) determine whether early experience or current environment had a greater influence on feather pecking; (2) determine the diurnal rhythm of feather pecking and compare this to dustbathing and foraging; (3) determine the effects of different types, of enrichment on feather pecking;
(4) validate a new method of studying abnormal (stereotypic) behaviour patterns; and (5) apply this to uncover the underlying motivation behind feather pecking. In addition, (6) the methods used here, and in previous research on similar problems, were surveyed to analyse how future work should best test hypotheses about the motivational bases of abnormal behaviour. Results showed that current environment was most influential on the levels of feather pecking shown (P<0.05) and this over-rode any potential positive effects of early substrate access (P>0.05). Feather pecking was evenly distributed throughout the light hours, as was foraging (P>0.05 for both), whereas dustbathing peaked around four to seven hours after lights on (P<0.05). The provision of any of the enrichments used decreased the amount of feather pecking, but foraging substrates decreased it the most
(P<0.05). A novel method of studying abnormal behaviour using Fixed Action Patterns was validated, with the motivation behind a behaviour pattern influencing the motor patterns involved. The motor patterns (head angles, durations of movements and fixation) involved in foraging and dustbathing pecks proved to be different (P<0.05), and severe feather pecks were found to be similar to foraging pecks (P>0.05) but different from dustbathing pecks (P<0.05). Overall, it appears that severe feather pecking stems from frustrated foraging behaviour and the design of industry-useable forages should be encouraged. ACKNOWLEDGEMENTS
First I would like to thank Ian Duncan, whose support and interest brought me to
Guelph and Georgia Mason, whose instruction and encouragement helped me to improve as a scientist and a writer. Both this thesis and I, personally, have benefitted from your expertise and knowledge and both are better for it. Also thanks to Tina Widowski and Suzanne
Millman for great comments and advice throughout my PhD.
Second, I would like to thank all my friends and family who have supported me through my PhD. Special thanks to my fiance Bruce Halstead, parents Linda and Lloyd
Dixon and grandparents Gladys and Albert Legge for hugs, words of comfort and delicious cookies! Additional thanks to Emily Toth-Tamminga, Jenn Brown, Maria Diez Leon,
Allison Bechard, Becky Meagher, Uta vonBorstel, Colleen Doherty, Val and Les Toth, the
Halstead's, the Clark's, the Clifford-Rakush's, and the Animal Behaviour and Welfare group at the University of Guelph. Included in my family, I'd also like to thank the various animal companions I've had the privilege to know over the past few years: especially my 'daughter'
Jingle, Falcon (Pointy Roozle), Tenna, Harpie and Panther, plus the rest of the non-human
Glenire acres gang.
Third, I'd like to thank the hundreds of chickens who participated in various experiments throughout my PhD. At the risk of being anthropomorphic, I would like to acknowledge their help and the sacrifice of their lives to further our knowledge of chicken behaviour and welfare.
Last, I'd like to dedicate this thesis to my dog, Kuno (2000-2008) who passed away shortly before my thesis defence. He was without a doubt the 'bestest' dog and he will be missed terribly. TABLE OF CONTENTS
List of Tables p.iii
List of Figures p.v
Chapter 1: A review of feather pecking literature and welfare issues associated with it p.l
Chapter 2: Changes in substrate access did not affect early feather pecking behaviour in
two strains of laying hen chicks, p.38
Chapter 3: A comparison of the diurnal rhythms involved in feather pecking, foraging
and dustbathing in laying hen chicks, p.61
Chapter 4: The effects of four types of enrichment on feather pecking behaviour in
laying hens housed in barren environments, p.81
Preface to Chapter 5 p. 100
Chapter 5: What's in a peck? Using Fixed Action Pattern morphology to identify the
motivational basis of abnormal feather pecking behaviour, p. 101 Chapter 6: A review of the methods used to study the motivations underlying stereotypic behaviour patterns, p. 141
Chapter 7: General Discussion and conclusion, p. 188
References p.200
Appendix A: Substrate Validation for Chapter 5. p.250
n LIST OF TABLES
Table 2.1: The average frequencies (per twenty-minute observation period) of feather pecking, dustbathing, and ground pecking and scratching on wire and peat moss floor treatments, combining strains, for Period One. p.57
Table 2.2: The average frequencies (per twenty-minute observation period) of feather pecking, dustbathing, and ground pecking and scratching on wire and peat moss floor treatments for Period Two. p.58
Table 2.3: The average frequencies (per twenty-minute observation period) of feather pecking, dustbathing, and ground pecking and scratching for ISA White Leghorns and
ISA Brown Leghorns during Period One. p.59
Table 2.4: The average frequencies (per twenty-minute observation period) of feather pecking, dustbathing, and ground pecking and scratching for ISA White Leghorns and
ISA Brown Leghorns during Period Two. p.60
Table 3.1: The correlations between feather pecking compared to foraging and dustbathing for the three Periods the daylight hours were divided into. The data in each box represent the correlation value (r), the corresponding P-Value and the sample size
(n). The * symbol denotes a significant correlation, p.80 Table 4.1: Example of Latin Square Design used, p.98
Table 4.2: All combinations of the four enrichments were used as Latin Squares and each Latin Square combination was replicated. The last stimulus presented was repeated to account for potential carry-over effects (not shown in table). D = dustbath; F = forage;
Nov = novel object; Not = no enrichment, p.99
Table 5.1: A comparison of the morphology involved in different types of pecks, p. 136
Table 5.2: A comparison of various aspects of the peck morphology involved in gentle and severe feather pecks, p. 137
Table 5.3: Comparison of the Chicken-Shaped and Flat varieties of stimuli, p. 138
Table 5.4: Severe feather pecks compared to the other types of pecks, p. 139
Table 5.5: Gentle feather pecks compared to the other types of pecks, p. 140
iv LIST OF FIGURES
Figure 1.1: An example of severe feather pecking that led to bleeding, p.37
Figure 2.1: The experimental set-up for chicks in peat moss and wire floored pens for
Period One. p.55
Figure 2.2: The experimental set-up for chicks in peat moss and wire floored pens for
Period Two. p.56
Figure 3.1: The experimental set-up for chicks in peat moss and wire floored pens, p.74
Figure 3.2: The frequency of feather pecking, dustbathing and foraging over the three
Periods of the daylight hours. The data represent un-transformed values for the average number of feather pecking, dustbathing and foraging bouts per Period for eight thirty- minute observation sessions ± SEM (Feather pecking: F<2,89) = 2.25, P = 0.1076;
Dustbathing: F(2,89) = 4.90, P = 0.0218; Foraging: F(2,89) = 1.15, P = 0.3176). p.75
Figure 3.3: The average amount of feather pecking in each week for peat moss and wire floor treatments. The data represent un-transformed values for the average number of
feather pecking bouts per week for six thirty-minute observation sessions ± SEM. (F(3,89)
= 8.64, P <0.0001). p. 76
v Figure 3.4: The average amount of dustbathing in each week for each of the three
Periods of the daylight hours. The data represent un-transformed values for the average number of dustbathing bouts per period per week for two thirty-minute observation sessions ± SEM. (F(6,89) = 3.95, P = 0.0313). p.77
Fig.3.5: The average frequency of foraging bouts in each week for each treatment. The data represent un-transformed values for the average number of foraging bouts per week per treatment for two thirty-minute observation sessions ± SEM. (F(3,89) = 5.97, P =
0.0007). Different letters (a, b, c) represent significant differences (P<0.05) between weeks while similar letters denote no significance found (P>0.05). p. 78
Fig.3.6: The average frequency of feeding bouts in each week for each treatment. The data represent un-transformed values for the average number of feeding bouts per week per treatment for two thirty-minute observation sessions ± SEM. (F^, 89) = 4.85, P =
0.0029). Different letters (a, b, c) represent significant differences (P<0.05) between weeks while similar letters denote no significance found (P>0.05). p.79
Figure 4.1: The average amount of feather pecking performed when given various
stimuli. The data represents the un-transformed average amount of feather pecking per pen ± SEM during thirty minute observation sessions (F(3; i77) = 165, P <0.0001, power =
0.999, a = 0.05). p. 96
VI Figure 4.2: The number of pecks directed to the stimuli. The data represent the un- transformed average number of pecks per pen ± SEM during thirty minute observation sessions. (F(2,119) = 464.74, P <0.0001, power = 0.999, a = 0.05) p.97
Figure 5.1: An example of the pen and test area set-up. p.131
Figure 5.2: Vertical head angle (degrees): measured as the angle between a virtual line through the bird's eye down the middle of the beak to the vertical, p. 132
Figure 5.3: Horizontal head angle (degrees): the angle made by a virtual line through the middle of the bird's head, down the middle of the beak to the horizontal, p. 133
Figure 5.4: a) Duration (sec) of head fixation: the length of time that the head is kept still before the peck, b) Contact to stimulus: duration (sec) from the end of head fixation to beak contact with the stimulus, c) Duration (sec) of the whole peck: time from no head movement, through the peck, back to no head movement, p. 134
vn Figure 5.5: An example of one of the peck measures used (duration of the peck) to examine pecks to the different shapes (Flat and Chicken-Shaped (C-S)) and materials used as the novel object, foraging and dustbathing stimuli measured in seconds (± SE).
Each stimulus type had two shapes and three materials used whenever possible. All sub types of each stimulus were not significantly different (Dur: F<6,108) = 2.00, P = 0.0916) while the different types of stimuli were significantly different from each other (F(4,129) =
19, 219.8, P<0.0001). p.135
vin CHAPTER ONE
A review of feather pecking literature and welfare issues associated with it
(A version of this chapter has been accepted for publication in Avian Biology Research)
ABSTRACT
Feather pecking, the pecking at or removal of feathers from one bird by another,
is a problem in the poultry industry. The elimination of damaging feather pecking from
flocks is made especially difficult by the large number of factors that appear to influence
its prevalence. This review begins with a general discussion of how abnormal behaviour patterns can arise and then shifts to the specific factors that contribute to feather pecking,
organized around Tinbergen's four questions on causation, ontogeny, phylogeny and
function. There is growing evidence that feather pecking (especially severe feather
pecking) is related to foraging motivation and gut function. However, other factors, such
as improper early experiences, strain and individual differences, and perseveration of the
behaviour help explain its continued occurrence, even if the birds are kept in enriched
environments. To date, methods of dealing with feather pecking are inadequate and
involve welfare concerns of their own. The problems of excessive pelage/plumage
removal or redirected oral/foraging related behaviour are not unique to poultry and seem
to occur in other species in which foraging and forage intake is important Between
species comparisons of related behaviour patterns may also improve our understanding of
1 feather pecking and help to design effective solutions. In order to solve the problem of feather pecking, the factors discussed in this review need to be accounted for or we risk applying 'band-aid' solutions, which may outwardly appear to solve the problem, but leave the underlying cause(s) still present and the animal's welfare potentially still
compromised.
1. INTRODUCTION
This review will first begin with a brief discussion of modern poultry farming and
a description of feather pecking and the welfare problems associated with it. Next we
will move on to a discussion of the various definitions used in the study of abnormal behaviour, before describing potential motivational and non-motivational explanations
for abnormal behaviour. Motivational explanations will be considered in some detail
since they form the focus of the rest of the thesis. Third, the known factors that contribute
to feather pecking will be reviewed, organized around Tinbergen's four questions on
causation, ontogeny, phylogeny and function (Tinbergen, 1963). Fourth, the current
methods of controlling severe feather pecking will be reviewed, ending with a discussion
of how the 'four whys' fit together to explain feather pecking behaviour.
1.1 Poultry production
Modern day chickens (Gallus domesticus) were originally domesticated from the Red
Jungle fowl (Gallus gallus) over eight thousand years ago. Initially, they were used for
2 religious sacrifice or cockfighting until around Roman times when they began to be developed for food production (Wood-Gush, 1971). The modern poultry industry began in the early nineteenth century, with most producers having small flocks of dual purpose
(egg and meat) birds raised extensively with access to outdoors and with hens brooding their young. As technology advanced, this changed, leading to more intensive systems with all birds being raised indoors and brooded artificially. The use of dual purpose birds decreased, and selection for two different types of birds (high egg or meat production) increased (Appleby et ah, 1992).
In contrast to their free range ancestors, modern egg layers tend to be housed in small wire cages (battery cages) which increase ease of egg collection, decrease the incidence of certain diseases and allow for larger numbers of birds to be fed and housed than was possible before (CARC, 2003). However, these cage systems also prevent the birds from performing most normal behaviour patterns, such as foraging and dustbathing
(e.g. Dawkins, 1977; Baxter, 1994). It is thought that the inability to perform some behaviour patterns may lead to frustration (e.g. Duncan and Wood-Gush, 1970). Since
approximately five billion laying hens are used for egg production worldwide each year
(estimated from FAOSTAT, 2008), the welfare of a large number of individuals may be
compromised. Many countries are starting to ban the use of battery cages. For example, the EU plans to have battery cages banned from 2012. However, the banning of battery
cages does not eliminate all welfare problems and hens in alternative husbandry systems
also experience several of these (Appleby and Hughes, 1991; Baxter, 1994; Duncan
2001a; Appleby et al, 2004).
3 1.2. Feather-pecking as a welfare and production problem
A welfare problem that often occurs more frequently in alternative husbandry systems is the abnormal behaviour pattern of feather pecking (Blokhuis, 1986; 1989), the pecking at
and possible removal of feathers from one bird by another (Hoffmeyer, 1969). Feather pecking can be divided into two distinct types, gentle feather pecking and severe feather pecking. Gentle feather pecking involves the feathers being gently pecked at or nibbled
and can be sub-divided into pecks to various feather targets that are thought to be driven
by exploration and repeated pecks to a single location on a feather (Newberry et ah,
2007), while severe feather pecking involves the vigorous pecking at and possible
removal of feathers (Fig. 1.1) (McAdie and Keeling, 2002). To date, it has been unclear
whether gentle and severe feather pecking stem from different motivational systems (e.g.
Kjaer and Vestergaard, 1999). However, it is severe feather pecking that constitutes the
real welfare concern (e.g. Blokhuis and Wiepkema, 1998), and thus will be the main
focus of this review. Since feather pecking has been associated with barren and stressful
environments, its occurrence may be an indicator of reduced welfare. However, the
primary welfare concern is that it can lead to cannibalism. Cannibalism can be divided
into two types: 1) cloacal cannibalism in which the vent area of the bird is pecked and
which may lead to the pecking out of abdominal organs and 2) tissue pecking in denuded
areas where the exposed skin is pecked, leading to haemorrhage (Savory, 1995). With
both kinds of cannibalism, the damage tends to be severe enough that the bird has to be
euthanized. There is no significant correlation between cloacal cannibalism and feather
pecking (Hughes and Duncan, 1972). However, severe feather pecking leading to blood
4 on the feathers stimulates increased pecking of that area and potentially leads to other types of cannibalism (Allen and Perry, 1975; Cloutier et ah, 2000) and feathers that have been damaged but that are not bloody also attract increased pecking (McAdie and
Keeling, 2000).
1.3 Feather-pecking as an Abnormal Repetitive Behaviour
Feather pecking is abnormal in the sense that it is rarely found in a population deemed normal, such as in the wild (e.g. Mason, 1991). It is also repetitive, in that some individuals tend to perform the behaviour numerous times throughout a day, day after day and week after week, throughout the lives of the birds. The term Abnormal Repetitive
Behaviour (ARB) has been suggested to cover a multitude of strange-seeming behaviour patterns in captive animals, from plumage and pelage-removal in a host of species to oral
stereotypic behaviours in ungulates. This term is descriptive and implies nothing specific
about causation; it does not even necessarily imply that there is problem associated with performance of the behaviour (e.g. Garner, 2006).
In the body of this review, feather pecking will therefore be referred to as ARB.
However, this classification will be revisited and re-evaluated in the discussion, to assess
whether feather-pecking should more properly be termed a stereotypic behaviour. The
classic definition of stereotypic behaviour or stereotypy is: 'unvarying repetitive
behaviour patterns with no apparent goal or proximate function' (e.g. Fox, 1964;
• dberg, 1978). Feather-pecking clearly does not fit this definition. However, more
5 recent definitions have focused on biological causation, instead of on these superficial phenotypic properties of a repeated behaviour. Garner (2006) suggested that 'stereotypy' be defined as a repetitive behaviour induced by striatal dysfunction; this relates to its usage in clinical psychology and contrasts with disorders such as obsessive compulsive disorder or impulsive behaviour (characterized by motor patterns that appear to be goal directed but are inappropriately repeated, and are caused by malfunctions in other, non- striatal brain regions). Mason (2006) agreed with Garner's proposed new definition of stereotypy, but followed Mills & Luescher (2006) in suggesting that the term 'stereotypic behaviour' be broader. She thus re-defined stereotypic behaviour (SB) as 'repetitive behaviour induced by frustration, repeated attempts to cope and/or C.N.S. (brain) dysfunction (regardless of its degree of variation or repetition). I will consider whether
feather-pecking has the properties of SB as defined in this way at the end of this paper.
Now that the terminology involved in the study of abnormal behaviour and ARB
has been defined, the next step is to discuss why ARBs occur, beginning with a general
discussion of motivation, then moving on to how highly motivated behaviour can become
abnormal, and ending with an overview of non-motivational explanations for abnormal
behaviour.
6 2. MOTIVATIONAL CAUSES OF ABNORMAL REPETITIVE BEHAVIOURS IN
CAPTIVE ANIMALS
Motivation explains why an animal chooses to perform a particular behaviour pattern over another at any point in time, and the amount of effort it puts into performing that behaviour (e.g. Toates, 1986; McFarland, 1999; Mason et al, 2001). As we will see later, it is a type of 'causal' explanation for behaviour.
The strength of a motivation is typically influenced by both the animal's internal state, such as gut fill, and also by external stimuli, such as the sight of food (Hogan,
1994), although whether internal or external factors are more important varies between different types of motivated behavior. The behaviour pattern with the strongest motivation is the one that the animal will direct the most time and effort in trying to achieve (e.g. McFarland and Sibley, 1975; Dawkins, 1990). As motivation increases, activity and appetitive (goal seeking) behaviour tend to increase (thus increased motivation leads to higher rates of behaviour that should lead to fulfillment of that motivation). Levels of this flexible, goal-searching appetitive behaviour are increased to attempt to achieve the consummatory behaviour, which is 'goal-completing' and often less flexible than appetitive behaviour (comprising Fixed Action Patterns [FAPs] - unlearned, species-specific movements that are relatively unaffected by feedback: Lorenz and Tinbergen 1938; see also Chapter 5). Consummatory behaviour also increases in rate as motivation increases (Craig, 1918 in Berridge, 2004). The threshold for the stimuli that elicit both appetitive and consummatory behaviour decreases with increasing
7 motvation: a larger number of stimuli become acceptable as triggers or outlets for the behaviour, or in other words, the animal becomes less 'choosy' (e.g. Hinde, 1970).
Positive feedback processes may act to increase motivation, and thence the amount of and intensity of behavior; e.g. the performance of appetitive behavior has the potential to increase motivation for the consummatory behavior (e.g. Hughes & Duncan 1988b); while performing a very small amount of consummatory behavior (e.g. eating a small
amount of food when hungry; the early phases of intromission during sexual behavior:
Wiepkema, 1971; Toates, 1980) also involves positive feedback that increases
motivation. Typically, negative feedback from the consequences of completing the
consummatory response (e.g. ingesting food) then decreases motivation (e.g. Le Magnen
1968 in McFarland 1971). However in some instances performing the appetitive or
consummatory actions per se (regardless of functional, e.g. physiological, consequences)
may reduce motivation. For example, hens perform nest building sequences during pre-
lay, even if the nest used previously is still available (Hughes et ah, 1989).
Motivational processes are relevant to welfare because motivation is related to
emotion: states elicited by rewards (things animals will work for) and by punishments
(things animals will work to avoid) (Rolls, 2000). Motivation causes the animal to seek
out and perform behaviour patterns that elicit positive emotions and to avoid events or
behaviours that elicit negative emotions (Rolls, 2000; see also Dawkins, 1990 and
Ferguson, 2000). Being in a state of high motivation that cannot be reduced is thus
8 unpleasant, stressful and aversive (Dawkins, 1988), while addressing and consummating a highly motivated behavior is pleasurable (e.g. Cabanac, 1979).
The various known properties of motivated behavior have together led to an ethological explanation for ARBs: that they derive from 'redirected', 'vacuum' or
'intention' movements. When animals are highly motivated to perform an activity, but a suitable outlet for that behavior is not available, they redirect it to something else: here the behaviour is transferred to a less preferred substitute (e.g. Breland and Breland,
1963). They may even perform vacuum activities: here, motor patterns of the behaviour are mimed without the appropriate stimulus present (e.g. Hinde, 1970). For example, sows kept in barren production systems will vacuum nest-build, going through some of the movements involved in nest building shown by extensively kept sows but without any nesting material (Baxter, 1982 in Cronin et ah, 1994). If a whole sequence is not possible, 'intention movements' may occur: incomplete or preparatory movements often
shown at the beginning of an activity (Hinde, 1970). Examples include caged canaries performing motor patterns that usually precede flight - a behaviour not possible in the
restrictive cage environment (e.g. Dawkins, 1988). Redirected, vacuum or intention
behaviours have therefore often been suggested to be the 'source behaviours' (Mason
1991a) for ARBs (see also Clubb et ah, 2006). ARBs derived in this way may therefore
involve similar motor patterns, diurnal rhythms, or eliciting stimuli as the normal
behaviour that was originally motivated.
9 But why should captive animals repeat redirected, vacuum or intention behaviours
(including escape attempts) so very often within a bout, and so consistently from day to day, that they come to be termed abnormal? One potential explanation is that the internal/external stimuli eliciting motivation are unusually strong or sustained in captivity and may be 'super normal' (Mendl, 1997). Internal factors may be high either because the animal is in a severe psychological deficit (e.g. feed restricted sows: Terlouw and
Lawrence 1993), or because its highly motivated to perform a natural behaviour irrespective of homeostasis (e.g. nest-building in hens: Hughes et al, 1989); and/or external factors can be very high (e.g. the sounds of food preparation and delivery can increase performance of ARB in some animals, such as pigs and mink). In terms of feedback, performance of the behaviour may have no consequences, therefore negative feedback never happens and the motivation to perform the behaviour pattern remains high (e.g. vacuum-chewing in sows may be an attempt to increase forage intake and/or gut fill but as forages are not provided, there is no negative feedback and motivation to ingest forages remains high, Horrell et al, 2001); or alternatively it is possible that the behaviour does have consequences and so is repeated to obtain reward (e.g. cross suckling in calves: de Passille and Rushen, 1997; see also Mason and Latham, 2004).
2.2 Non-motivational influences on ARB
If captive animals are motivated to perform natural behaviours patterns that are not possible (see e.g. Hughes and Duncan, 1988a; Mason et al, 2001), this may thus both cause stress and lead to ARB (see also Mason, 2006). Further possible explanations,
10 however, involves non-motivational changes, such as 'perseveration': the continuing of an activity without the appropriate stimulus present' (e.g. Sandson and Albert, 1984).
Perseveration can be induced by stress (e.g. Francis et al., 1995), and/or by lasting
CNS changes caused by e.g. rearing without motivationally significant stimuli (rearing infant primates without their mothers: Novak and Harlow, 1975 in Novak et al, 2006).
In cases like this, frustration may therefore lead to abnormal behavior indirectly, by mediating changes in the brain that alter how behavior is controlled, thence causing animals to be more generally prone to abnormal repetition. An animal whose ARB is perseverative may be abnormally persistent in other aspects of its behaviour as well (e.g.
Sandson and Albert, 1984; Garner, 2006; Mason, 2006). These changes may be due to alterations to motivational areas in the brain (e.g. the limbic loop, cf. Cabib, 2006) or to areas associated with motor control (e.g. the striatum and motor cortex (cf. Lewis et al.,
2006).
Furthermore, as a behaviour pattern is repeatedly performed, neural control of the behaviour may shift into a type of automatic processing referred to as central control or habit formation. This allows behaviour patterns to be performed with minimal cognitive processing or sensory feedback (for reviews of these processes see e.g. Dantzer, 1986;
Mason and Turner, 1993; Toates, 2001; Mason 2006). Further non-pathological factors which can affect the performance of ARB are properties of the animal and/or its circumstances, such as the animal's stamina or the free time if has available to perform
ARBs. Thus, animals that are adapted to travel large distances may have higher physical
11 endurance and may be able to sustain the ARB (such as pacing) longer than animals adapted to have small home ranges. Animals with limited behaviour patterns available to them, such as tethered sows, also have more free time in which to perform high levels of
ARB than animals with a larger range of behaviour patterns available, such as group housed sows who have the opportunity to move around their enclosure and perform a variety of social behaviour patterns (e.g. Blackshaw and McVeigh, 1984 in Rushen,
1984).
3. WHY DO DOMESTIC FOWL FEATHER PECK?
Over 40 years ago Tinbergen wrote an influential paper describing methods that could be used to investigate behaviour patterns by dividing the study of behaviour into four questions relating to causation, ontogeny, phylogeny and function (Tinbergen, 1963).
These levels of explanation are still thought to be useful today (e.g. Olsson and Keeling,
2005) and will be used to organize this review of factors affecting feather pecking behaviour. This paper will therefore now review the known factors that contribute to feather pecking behaviour organized around Tinbergen's four questions on causation, ontogeny, phylogeny and function. Next, the current methods of controlling severe feather pecking will be reviewed, ending with a discussion of how the 'four whys' fit together to explain feather pecking behaviour. The main focus will be on severe feather pecking as opposed to gentle feather pecking, since severe feather pecking is detrimental to welfare. There are a variety of factors affecting feather pecking at each of these levels.
Some may be very specific to feather pecking; some probably have very general effects;
12 some may affect the feather peckers; some may affect the recipients of the pecks. Here I will not attempt to disentangle these, but simply present all the factors that can influence this phenomenon.
3.1. Causation
Causation refers to what produces the behaviour pattern, taking into account both internal and external stimuli and the mechanisms by which these produce the behaviour
(Tinbergen, 1963). The two main motivational hypotheses behind feather pecking involve re-directed dustbathing motivation or re-directed foraging motivation due to lack of appropriate litter substrates. Dustbathing and foraging behaviour patterns would normally (i.e. in free living birds) be directed to substrates but instead may get re-directed or transferred to an alternative stimulus (feathers of other birds) when the animal's response is inhibited or frustrated (e.g. Dawkins, 1990). In this thesis frustrated will refer instances where the behaviour pattern is not possible or is being thwarted in some way
(i.e. bird housed with no substrate or can see but not access a substrate), while, as mentioned above, re-directed will refer to instances where the behaviour pattern is being directed to an inappropriate stimulus, such as dustbathing or foraging behaviour being transferred to the feathers of other birds (see 2.1, above). This transference can continue
even if the original motivated behaviour is no longer frustrated (e.g. the bird has access to
a substrate) due to habit formation or perseveration of the behaviour (see 2.2, above).
13 Vestergaard (1994) proposed that feather pecking is re-directed dustbathing due to lack of appropriate dustbathing substrates and lack of appropriate experience with these substrates early in life. In other words, dustbathing behaviour that would normally be directed to a substrate is transferred to the feathers of other birds (alternate stimulus) when there is no appropriate dustbathing substrate present and if chicks do not receive early exposure to dustbathing substrates (by about day 10 of life), this may encourage the long-term mis-identification of feathers as appropriate dustbathing substrates through habit formation or perseveration (Vestergaard et ah, 1990). Birds dustbathe for about Vi hour every other day (Vestergaard, 1982a). Dustbathing starts with the bird pecking,
scratching and bill-raking at the substrate. Next, the bird squats in the substrate and performs head rubbing, bill-raking, vertical wing shaking and leg scratching interspersed by periods of side-lying or side-rubbing. A dustbathing bout generally ends when the bird rises and vigorously shakes (e.g. Larsen et ah, 2000). Birds without access to litter
will occasionally still go through the movements of dustbathing, suggesting that there
may be a build up of internal motivation (Vestergaard, 1982a; Olsson and Keeling, 2005).
The provision of a dustbathing substrate does decrease rates of feather pecking
compared to that of birds housed on wire flooring (e.g. Vestergaard et ah, 1993; Johnsen
and Vestergaard, 1996; Norgaard-Nielsen, 1997). Early experience with a substrate may
also be protective against feather pecking later in life, even if the substrate is no longer
available (Johnsen et ah, 1998): an issue I return to in the 'Development' section. Birds
reared on sand and peat moss feather pecked less than birds reared on straw, and this
continued even after all birds were switched to straw alone (Norgaard-Nielsen et ah,
14 1993). Also, exposure to wood shavings, even for a short time (a few hours a day for 10 days), was found to be protective against later feather pecking (Nicol et ah, 2001). Thus there is some evidence consistent with the dustbathing hypothesis.
An alternate hypothesis states that feather pecking is foraging behaviour directed at feathers when appropriate substrates are absent (Hoffmeyer, 1969; Hughes and
Duncan, 1972; Blokhuis, 1986; 1989). In other words, foraging behaviour that would normally be directed to substrates, gets transferred to the feathers of other birds (alternate stimulus) when an appropriate foraging substrate is not present. Foraging is appetitive feeding behaviour, involving ground pecking and scratching. Foraging generally leads to feeding, which is the consummatory behaviour and involves ingestion of feed particles
(Duncan, 1980). Birds feed frequently throughout the day, at least once per hour
(Duncan et al., 1970) and given the space and opportunity will also spend a lot of time foraging. For example, when housed outdoors, junglefowl spend about 60% of their time ground pecking and 34% of their time ground scratching during the active part of the day
(Dawkins, 1989). In early life, a number of different objects, both nutritious and non- nutritious are pecked and the chick quickly learns to classify objects correctly (Hogan-
Warburg and Hogan, 1981; Hogan, 1994). However, under commercial conditions for laying hens, the consumption of food is easily achieved without the need to forage, and there are generally no foraging substrates present. Nevertheless, foraging behaviour may have its own internal motivational factors which drive the bird to perform the behaviour even when this is not necessary for food consumption (Hughes and Duncan, 1988a).
Feathers may thus be perceived as forages when no other substrates are present (e.g.
15 Blokhuis, 1986; Huber-Eicher and Wechsler, 1997a, 1998). Further support for this hypothesis includes evidence of an inverse relationship between ground pecking and feather pecking, with birds housed on litter ground pecking more and feather pecking less than those housed on wire (Blokhuis and Arkes, 1984; Huber-Eicher and Wechsler,
1997a). The quality of a forage, as measured by the amount of pecking and scratching directed to it, also affects feather pecking since birds with higher quality substrates (e.g. long cut straw) perform less feather pecking than birds with lower quality substrates (e.g. shredded straw) (Huber-Eicher and Wechlser, 1998). Thus this hypothesis, too, has some empirical support.
But which hypothesis is correct? And how can both be supported by empirical evidence? The key problem with these two hypotheses is that the results found for one could easily be substituted to fit the other. Dustbathing and foraging both include pecking and scratching at the ground, so observations of "ground pecking behaviour" could actually be the start of either activity (Savory, 1995). There is also an overlap between substrate classification as dustbaths or forages. For example, in one study designed to investigate feather pecking as re-directed dustbathing, straw and sand were used as the dustbathing substrates (Johnsen et al., 1998). However, another study intending to look at feather pecking, also used straw, but this time as a foraging substrate
(Aerni et al., 2000). In cases like this it would be hard to tell if the pecking directed at the straw was motivated by foraging or dustbathing.
16 Regardless of which hypothesis is correct, the inability to properly fulfill a motivation (either to forage or dustbathe) may lead to frustration; and frustration per se may play a role in feather-pecking. For example, birds whose feeding behaviour was frustrated by covering feed rewards will increase their pecking of the environment
(Duncan and Wood-Gush, 1972), and it has been proposed that by increasing arousal, aggression and fear, frustration may cause occasional pecks to other birds' feathers to develop into damaging feather pecking (Lindberg and Nicol, 1994). However, when
Rodenburg et al. (2005) induced frustration in High Feather Pecking (HFP) and Low
Feather Pecking (LFPs) birds (see more on HFPs and LFPs in the 'Genetics' section below) by training them to peck a key for a food reward and then preventing physical but not visual contact with the food by covering the food dish with clear Perspex, levels of feather pecking did not increase.
Also, regardless of which motivational hypothesis is correct, the type of flooring available, such as wire or slatted floors compared to floors with a substrate, does affect feather pecking behaviour, with birds housed on shavings consistently performing less feather pecking than those housed on wire (e.g. Nicol et al., 2001). Providing enrichment for the birds, such as string or novel objects, also decreases feather pecking compared to controls, although this may be due to providing the birds with more to occupy their time
(McAdie et al., 2005). Enrichment does not solve all feather pecking problems, however, as some levels of feather pecking are still found when birds are housed with a substrate or moved to enriched environments (e.g. Nicol et al., 2001; van Hierden et al., 2002). This may be due to perseveration of the behaviour (see below for discussion).
17 A number of other internal and external factors are known to influence feather pecking behaviour, such as the nutritional content of the feed and the feed form (mash, crumble or pellet). There is more pecking damage to feathers when chickens are provided with pelleted feed compared to mash and even dilution of the pelleted food with water does not affect feather pecking (e.g. Savory et al., 1999; El-Lethey et al, 2000).
These higher levels of feather pecking appear to be due to the shorter time required to eat pelleted food, even if the pellets have expanded in size from being soaked in water
(Savory et al, 1999). Thus, the time available appears to influence the amount of feather pecking performed. Proper feather development requires certain nutrients, such as the
amino acids methionine and cysteine (vanKrimpen et al, 2005), without which the number of improperly developed or abnormal looking feathers may increase, and these
abnormal or damaged feathers receive higher levels of feather pecking than normal
feathers (McAdie and Keeling, 2000; vanKrimpen et al, 2005). Supplementing diets
with protein, such as casein, gelatin or soybean oil meal, or the amino acid, L-
trypotophan decreases feather pecking levels compared to when birds are fed a standard
diet (Schaibnle et al, 1947; Savory et al, 1999). Since feathers are made up of 89-97%
protein it is possible that the levels of protein in a standard poultry diet may not be
sufficient to ensure that all feathers develop properly and these abnormal feathers may
receive more feather pecks. L-tryptophan has also been shown to reduce aggression in
males by causing sedative-like effects (Shea et al, 1990) and this general reduction of
activity may be the cause of the decreased feather pecking found. However, low levels of
serotonin (which is synthesized from tryptophan) have also been implicated in other
18 behavioural disorders (in humans and animals), such as depression and obsessive compulsive disorder, and supplementation with tryptophan (which leads to increased levels of serotonin) has been shown to improve these conditions (e.g. Sandyk, 1992;
Weld et ah, 1998). Chickens selected for high levels of feather pecking also have lower serotonin turnover than those selected for low levels of feather pecking (as discussed below). Thus, tryptophan and associated serotonin turnover appear to be in some way
associated with feather pecking. Fibre levels also affect feather pecking. When chickens
are fed high fibre diets (which had low energy) they feather peck less than controls.
However they also spent more time at the feeder and consumed more feed, suggesting
that the effect may reflect a difference in activity distribution (i.e. they spent more time
eating so had less time available to feather peck) (e.g. Bearse et ah, 1940; vanKrimpen et
ah, 2005). The addition of roughages, such as clover or maize-silage, also decreased
feather pecking (along with mortality rates) but did not affect feed efficiency (e.g.
Steenfeldte/a/.,2001).
Numerous other environmental conditions affect feather pecking behaviour. For
example, feather pecking is present in both floor pens and battery cages. Some studies,
such as Hughes and Duncan (1972), found the frequency of feather pecking to be higher
in cages compared to pens, but more commonly it is shown that feather pecking rates are
higher in floor pens than cages (e.g. Tauson, 2005). This is thought to be due to the
stable social groups in cages and the larger group sizes housed in floor pens, as these
larger groups increase the number of birds primary feather peckers have access to and, as
laying hens can modify their own behaviour when watching the behaviour of others
19 (Nicol, 1995), the primary feather peckers can encourage more secondary feather pecking
(i.e. a small number of primary birds first draw attention to the feathers of the victim bird and this encourages other birds start pecking at the feathers as well) (e.g Hughes and
Duncan, 1972; Savory et al, 1999). For example, Zeltner et al. (2000) introduced chicks showing either high levels or low levels of feather pecking to groups of chicks that were performing only low levels of feather pecking behaviour. Groups in which the high feather pecking chicks were added increased their levels of feather pecking while groups of chicks that had birds showing low levels of feather pecking did not. However, when a later study differentiated between severe and gentle feather pecks, only gentle feather pecking, which does not cause feather damage, seemed to be socially transmitted
(McAdie and Keeling, 2002). Some studies have found a relationship between a bird's ranking in the dominance order and its tendency to feather peck others, with birds higher
in the dominance order more likely to feather peck birds lower in the dominance order
(Hughes and Duncan, 1972; Vestergaard, et al, 1993). However, these authors suggest
that this is not a direct effect but simply due to the fact that a more dominant bird is more
likely to approach a subordinate bird than the converse, and so is more likely to be in a
position to feather peck. Higher levels of feather pecking are also shown with higher
light intensities and decreased light has been recommended to control feather pecking
outbreaks (Hughes and Duncan, 1972; Kjaer and Vestergaard, 1999). This increase in
feather pecking is most likely due to the increased visibility of particles on the plumage
of other birds, thus increasing motivation to peck, as dark feathered birds also receive
more pecks compared to light feathered birds, most likely due to the darker feathers
having a greater colour contrast with the particles than light coloured feathers, increasing
20 particle visibility on the feathers (Savory and Mann, 1999) or due to non-motivational explanations, such as a general increase in activity.
Feather pecking also appears to have a diurnal rhythm, at least in older birds: 38 and 50 week old laying hens showed increased levels of feather pecking in the later part of the daylight hours (10 hours or more after lights on) (Preston, 1987; Kjaer, 2000).
However, Bright (2007) found higher levels of severe feather pecking in the early part of the day in similarly aged birds. This difference may be due to the division of feather pecks into gentle and severe pecks in Bright (2007)'s study while Preston (1987) and
Kjaer (2000) lumped both types of feather pecking together. Thus, it is possible that the
increase in feather pecking found in later parts of the day may be due to higher levels of
gentle feather pecks.
Neuropeptides (short links of amino acids), like dopamine and endogenous
opioids, are thought to play a role in self mutilating and stereotypic behaviour in birds
(e.g. Goodman et al, 1983; Kjaer et al, 2004). One possible explanation is that these
neuropeptides are associated with pleasure systems in the brain and reinforce motivation
for certain activities that stimulate these systems (Arias-Carrion and Poppel, 2007). For
example, chickens given a low dose of haloperidol (a dopamine antagonist) decreased
their feather pecking behaviour compared to that of birds injected with saline (Kjaer et
al., 2004), thus implicating involvement of the dopaminergic system. Hormones also
influence feather pecking levels. There is an increase in feather pecking behaviour at the
start of lay and this is endocrine mediated: this increase can be stimulated by implanting
21 gonadal hormones in 12-week-old pullets (Hughes, 1973). This research revealed that progesterone alone causes an increase in pecking, a combination of progesterone and oestrogen causes a greater increase, while oestrogen implantation alone has no effect.
However, the increase of damaging feather pecking can be blocked with testosterone.
Chickens kept in barren conditions (no substrate) feather pecked more and also had an increase in their heterophil/lymphocyte ratio, an indicator of avian stress (for review of this measure see Maxwell, 1993), than birds housed with a substrate (El-Lethey et al.,
2000). Corticosterone supplementation in the feed also increases feather pecking levels of birds housed on litter compared to similarly housed control birds (El-Lethey et al,
2001)., suggesting a possible causal relationship between stress and feather pecking.
As discussed in the introduction, another causal explanation of feather pecking may the perseveration of behaviour. Thus, ARBs like feather pecking may not stem from normal behaviour alone, but instead from changes that make the animal abnormally persistent in all types of behaviour (cf. Garner, 2006). For example, Harlander-
Matauschek et al. (2006a) found that HFP hens would peck a key at significantly higher amounts than LFP hens to access feathers in a trough, suggesting that HFP birds are highly motivated to eat feathers, which is consistent with the foraging hypothesis.
However, HFP birds also worked harder to access an empty trough, indicating that once they had learnt to key-peck for food, they then continued to peck persistently regardless
of the reward on offer. Such changes in the brain and the fundamental control of behaviour may explain why some levels of feather pecking are still found when birds are
housed with a substrate or moved to enriched environments (e.g. Nicol et al., 2001;
22 vanHierden et ah, 2002) or even why birds continue to feather peck after beak trimming when removing feathers becomes difficult (no feedback from feather removal or improved gut function from feather pecking). Thus, proper early experience and enriched housing conditions will be extremely important in controlling or eliminating damaging feather pecking behaviour.
3.2. Ontogeny
Ontogeny refers to the development of a behaviour pattern in an individual through integration of learned and innate aspects of the behaviour (Tinbergen, 1963). 'Down pecking' can be found in chicks from day one of life (e.g. Savory and Mann, 1999; Chow and Hogan, 2005). Most pecks are initially gentle rather than severe; but the incidence of severe feather pecks increases as the birds age (Newberry et ah, 2007). As mentioned previously, there is a marked peak in feather pecking at the onset of lay, around 18 to 24 weeks of life that appears to be due to increased hormone levels and perhaps in nest- building motivation (in natural situations, forages would be used as nesting material as well) (Hughes, 1973).
The development of feather pecking may also be influenced by early experience with or without a substrate for foraging or dustbathing as discussed above. The
perception of a substrate as appropriate for feeding or dustbathing occurs during trial-
and-error pecking of nutritious and non-nutritious objects during the first few days of life
(Hogan-Warburg and Hogan, 1981). Jensen et ah (2006) found that providing dark
23 brooding areas where chicks can rest, results in long-term reductions of feather pecking in adulthood compared to when the chicks were given traditional heat lamps. The likely mechanism for this reduction in feather pecking is simply by decreasing the availability of inactive chicks which are preferred to active chicks as targets for pecking (Jensen et al, 2006). Chicks reared by a broody hen show more ground pecking and less feather pecking than non-brooded chicks. It is thought that the hen encourages the chicks to explore the environment and direct more pecks at the ground (Riber et al, 2007).
However, another explanation for the higher levels of feather pecking in non-brooded chicks is possibly that brain development is somehow impaired by the absence of a mother-like presence (see Mason, 2006, Box 6.2.). Improper early experience, such as continually frustrated motivation (Novak et al, 2006) or barren and predictable captive environments (Lewis et al, 2006) may also lead to perseveration of the behaviour.
As with most populations of captive animals, there are animals that perform
ARBs and there are those that do not, even when reared and housed in the same conditions (e.g. Mason, 2006). While work has been done to identify factors that predispose a bird to become a feather pecking victim (such as plumage colour, Keeling et al, 2004), it has not been determined if these are the birds who do not perform feather pecking. Not all birds develop damaging feather pecking behaviour: for example in a survey of 64 flocks, 40% of those raised in aviaries and 35.9% of those raised in deep litter systems developed severe feather pecking (Huber-Eicher and Audige, 1999). Thus, research examining differences between flocks that do and do not develop severe feather
24 pecking, as well as individuals who do and do not develop this behaviour could lend insight into other factors that influence the development of feather pecking.
3.3. Function
Function refers to the survival value of a behaviour pattern, either in a proximate
(short-term) or ultimate (long-term) manner (Tinbergen, 1963). The function of feather pecking is still unknown; and if the behaviour is in fact pathological, it may have no true function. Alternatively, it has been suggested that feather pecking may be a type of reaction style that allows chickens to manage with a less than ideal environment (i.e. a response to help reduce aversion or stress in undesirable conditions) (e.g. Cooper and
Nicol, 1993; Korte et al, 1997; Rodenburg et al. 2004). It is thought that the benefits associated with performing certain behavioural reactions may reinforce performance of the behaviour pattern (Wiirbel et al, 2006). There are two general types of reacting or responding styles distinguished: 1) Proactive responding involves the animal actively
trying to deal with the situation and 2) Reactive responding involves suppression of
environmentally directed activities (e.g. Bolles, 1970; Koolhaas et al, 1999). There is
some evidence that proactive responders may be more vulnerable to developing ARBs
than reactive responders (e.g. Koolhaus et al, 1999). Correspondingly, the greater levels
of gentle and severe feather pecking found in High Feather Pecking lines are thought to
be evidence of a proactive responding style, while the increased levels of foraging and
dustbathing in Low Feather Pecking lines may be evidence of a reactive responding style
(Johnsen and Vestergaard, 1996; Korte et ah, 1997; van Hierden et ah, 2002). In
25 addition, high feather pecking chicks had stronger immune responses when challenged
(Buitenhuis et al, 2006), lower plasma corticosterone levels, and lower dopamine and serotonin turnover levels in the forebrain than low feather pecking chicks (van Hierden et al, 2002), and these characteristics have been shown to be associated with proactive responding strategies in rodents and pigs (e.g. deBoer et al, 1990). However, the above results only indicate a possible correlation, not a causal relationship between reaction style and feather pecking.
Because feather pecking involves ingestion, another possible function of feather pecking involves improving gut function and digestion. Insoluble fibres stimulate the gizzard and are beneficial to nutrient digestion (Hetland et al, 2003). Commercial chicken feed generally does not contain much insoluble fibre, even though insoluble fibre may be necessary to maintain normal gizzard function (Hetland et al, 2005). Chicken feathers are considered non-nutritive (McCasland and Richardson, 1966) but are often consumed by feather peckers after being pecked from the victim bird (McKeegan and
Savory, 1999) or if provided in a bowl to feather pecking birds (Harlander-Matauschek et al, 2007). Feather consumption has been shown to increase the speed of feed passage in
a manner similar to insoluble fibres (Harlander-Matauschek et al, 2006b). Feather pecking chickens will also work to obtain feathers they can consume by pecking a key an
increasing number of times, demonstrating that they are motivated to access and eat the
feathers (Harlander-Matauschek et al, 2006a). It is not yet known, however, if providing
birds with feathers for consumption decreases levels of feather pecking or if the flooring
substrate most successful at decreasing feather pecking is also consumed the most,
26 although these are both logical predictions. It is known that birds given free-choice with a variety of feed/forage types adjust consumption levels to obtain sufficient dietary components, such as insoluble fibres; however, this is not possible in production systems
(e.g. Emmans, 1979; Hetland et ah, 2003). Thus, it is possible that the function of severe feather pecking, where feathers are removed and consumed, is to improve gut function by increasing dietary insoluble fibres, while the function of gentle feather pecking, where feathers are nibbled, may be related to other behaviour patterns, such as preening (van
Hierden et ah, 2002) or social exploration (Riedstra and Groothuis, 2002). As mentioned above, provision of fibrous substrates does not eliminate feather pecking entirely. Thus, it is possible that feather pecking has become divorced from its original function and becomes perseverative in birds that continue with this stereotypic behaviour even if their conditions improve.
3.4. Phylogeny
Phylogeny refers to the evolutionary development of a behaviour pattern and thus focuses mainly on genetics (Tinbergen, 1963). Feather pecking has a heritable component with heritability estimates ranging from h =0.14 to 0.56 (e.g. Cuthbertson, 1980; Kjaer and
Sorensen, 1997; Rodenburg et ah, 2003). In addition, different strains of birds have different levels of feather pecking (e.g. Hughes and Duncan, 1972) and within one strain of birds, High and Low Feather Pecking lines (HFP and LFP) have been selected for and developed i.e. there are individual differences in feather pecking levels (e.g. Kjaer et ah,
2001), and these strain and individual differences may contribute to the high variability
27 found in the heritability estimates. In lines selected for different levels of feather pecking, the amount of both gentle and severe feather pecking was higher in the High
Feather Pecking line (van Hierden et ah, 2002) while Low Feather Pecking lines performed more dustbathing (Johnsen and Vestergaard, 1996), foraging and feeding (van
Hierden et al, 2002), indicating a potential inverse relationship between feather pecking and ground pecking activities. HFP and LFP also react differently to frustration. As mentioned above, birds have been operantly trained to peck a key for a food reward but access to the reward can be frustrated by covering of the food dish. During this frustration, LFPs pecked more at the environment (especially at the covered food dish) which was thought to indicate higher frustration (Rodenburg et ah, 2002) while HFPs performed more gentle feather pecking and aggressive pecking to their housing companion and more ground scratching (Rodenburg et ah, 2005). Research also being done in molecular genetics has identified possible Quantitative Trait Loci (QTLs: allele variations associated with phenotypic variations) involved in gentle feather pecking on
DNA micro-satellite markers GGA1 and GGA2 and a QTL for severe feather pecking on
GGA2 (Buitenhuis et ah, 2003) and birds whose plumage colour is due to melanization controlled by the wild recessive allele PMEL17 may be genetically predisposed to become victims of feather pecking (Keeling et ah, 2004).
Feather pecking is also found in other types of avian species, such as broiler chickens (e.g. Kristensen et ah, 2007), turkeys (e.g. Sherwin et ah, 1999), ducks (e.g.
Gustafson et ah, 2007), pheasants (e.g. Hoffmeyer, 1969), and parrots (e.g. Meehan et ah,
2003). The factors contributing to feather pecking have been most extensively studied in
28 laying hens (approximately 235 hits in a Web of Science search for feather pecking and laying hen vs. less than 100 hits for feather pecking in all other types of birds); however, feather removal in these other types of birds appears analogous to feather pecking in laying hens due to behavioural similarities between these species (e.g. they all forage, preen, etc.) and the performance of feather pecking occurring primarily in barren captive conditions with the provision of forages or other types of enrichment reducing this pecking (e.g. Martrenchar et ah, 2001; Meehan et ah, 2003; Gustafson et ah, 2007). In addition, similar abnormal pelage removal has also been reported in non-avian species, such as chimpanzees (Smith et ah, 2004), mice (Garner et ah, 2004), rabbits (Jackson,
1991), sheep (Marsden and Wood-Gush, 1986), cats and dogs (Sawyer et ah, 1999) and even humans (Schlosser et ah, 1994). Research into these different types of integument removal may lend insight into the causation, development, function and phylogeny of stereotypic feather pecking behaviour in poultry.
Oral, foraging-related abnormal behaviour patterns are also very common in herbivores - even though they do not all involve pelage ingestion. Similar to chickens, these species also normally have high fibre diets that are not satisfied in captivity and this could lead to the performance of stereotypic behaviour. Thus, aside from similarities with other types of stereotypic pelage removal, such as wool-chewing in sheep and fur removal in rabbits, feather pecking may also be similar to the types of stereotypic oral behaviour found in ungulates with similar foraging requirements, some of which involve
ingestion such as wood-chewing in horses and slat-chewing in sheep, and some which do
not, such as tongue rolling in cattle and crib biting in horses. Like feather pecking, the
29 main hypotheses for the motivation behind these stereotypic behaviour patterns are that they derive from natural foraging and can be reduced in captivity by the provision of more naturalistic food with a higher fibre content and increased volume (Bergeron et al,
2006). In addition, the performance of these stereotypic behaviours are thought to either improve gut function by, for example, increasing nutrient digestion or saliva production
(e.g. McGreevy and Nicol, 1998) and/or attempting to satisfy natural foraging (e.g.
Bashaw et al, 2001). There are also genetic components involved in these stereotypic behaviours. For example, female chinchillas are prone to higher levels of fur-chewing than males (Ponzio et al., 2007). Garner et al. (2006) found that different families of parrots performed different levels of feather picking, suggesting there might be a genetic component to this particular behaviour. Stress has been implicated in the performance of other stereotypic behaviours, with animals that are more stressed (e.g. close to a noisy, high traffic area vs. in a more quiet, less traveled area) performing higher levels of stereotypic behaviour (Garner et al, 2006) with anti-depressants and serotonin reuptake inhibitors recommended for treatment (Seibert, 2007).
Thus, motivation to forage and/or to ingest high levels of fibre are a potential cause of stereotypic behaviour found in other taxa - making the hypothesis that chickens feather peck as re-directed foraging to increase fibre intake and improve gut function
(e.g. Harlander-Matuaschek et al., 2006a; 2007) particularly plausible. Further studies into this type of stereotypic behaviour could also increase our understanding of feather pecking and review of multi-species literature when examining behavioural problems, like stereotypic behaviours, should be encouraged.
30 4. CURRENT SOLUTIONS TO FEATHER PECKING
Current solutions to feather pecking are to beak trim the birds and/or keep them in battery cages. These practices are reviewed here to demonstrate that 1) these solutions do not actually succeed in stopping feather pecking; instead, they merely decrease the noticeable damage; and 2) that the use of these practices is questionable as they involve welfare problems of their own. Instead we should advocate ethologically-based methods that solve the problems underlying feather pecking and thus improve bird welfare.
Beak trimming or debeaking is the removal of approximately the last third of the upper and lower beak. This can be done using a heated blade that cuts through and cauterizes the beak or with electronic beak trimmer which uses a high voltage electrical current to burn a small hole in the upper beak (Gentle et al, 1990). The cells in the beak tip die and slough off after three to seven days (Cunningham et al, 1992). A more recently developed method is the use of an infrared beam directed at the tip of the beak, which damages the underlying tissue, causing the tip to fall off in about two weeks
(Gentle and McKeegan, 2007). Beak trimming does not stop the action of feather pecking but since the hooked end of the beak is removed, the amount of feather damage from pecking is reduced (e.g. Blokhuis and van der Haar, 1989). Since studies involving feather pecking tend to use non-beak trimmed birds, it is difficult to say if the levels of feather pecking are even decreased. However, chickens that were beak trimmed pecked less at a bunch of feathers than non-beak trimmed chickens, indicating that rates of
31 feather pecking may decrease after beak trimming (Martinec et ah, 2002). In addition, beak trimming has many negative side effects. Extensive neuromas, an overgrowth of nerves associated with an injury, may form on the stump of the beak (Breward and
Gentle, 1985) and acute and chronic pain have been associated with beak trimming
(Duncan et ah, 1989; Gentle et ah, 1990). Feed intakes falls temporarily after hot blade beak trimming (Gentle et ah, 1982), and there is a reduction in use of the beak for non
essential behaviour, such as preening and exploration (Duncan et ah, 1989). Also, after
trimming, the beak continues to grow and may get to a length where another trim is
required, this time when the bird is older and the beak is larger (Gentle and McKeegan,
2007).
Battery cages, cages made primarily of wire-mesh and including a feeder and
drinker were originally implemented to separate hens from their faeces (e.g. Appleby et
ah, 2001). They also have the advantage of allowing larger numbers of birds to be kept
in a building (CARC, 2003) and decreasing the incidence of cannibalism (Huber-Eicher
and Sebo, 2001). Access to a limited number of other hens and the restricted movement
available in cages may decrease the amount of damaging feather pecking behaviour
(especially since these birds also tend to be beak trimmed) when compared to floor or pen
housed birds which are kept in larger numbers with more space available for movement
and targeting of birds with damaged feathers (e.g. Hughes and Gentle, 1995). Keeping
birds in smaller groups also reduces the risk of social transmission from primary peckers
to secondary peckers (Hughes and Duncan, 1972). However, birds are still able to
feather peck in battery cages, and in some cases at higher rates than in other systems such
32 as enriched cages (Appleby et al., 2002), which may be related to the lack of substrate.
Hens in alternative systems can also perform a larger number of behaviour patterns than those in cages, including foraging and nesting, which in addition to potentially satisfying motivations, may give the birds more to do, leaving less time available for feather pecking (Appleby et al., 1992). However, there is also the possibility that feather pecking may become perseverative, thus high rates of feather pecking may still be found in extensive conditions with little effect of additional enrichment (e.g. Garner, 2006).
Battery cages in general are coming under criticism and there is an increased
demand for eggs that are not produced in them (Huber-Eicher and Sebo, 2001). Other welfare problems associated with battery cages include lack of space (Nicol, 1987),
dustbaths, and perches (Duncan, 2001b). Also, the reduced space does not allow for
much movement, which can contribute to bone weakness (Leeson and Morrison, 1978).
Thus, it seems clear that alternative husbandry systems or the addition of enrichment to
conventional cages are needed to improve the problem of feather pecking.
5. 'FOUR WHYS': HOW DO THEY FIT TOGETHER? AND SHOULD FEATHER-
PECKING BE CLASSIFIED AS A STEREOTPIC BEHAVIOUR?
Based on the above review, it has been shown that the motivation behind feather
pecking is still uncertain and past attempts have tended to reduce but not eliminate the
problem, indicating that there is still much to learn about this behaviour. However, there
is strong evidence that feather pecking (especially severe feather pecking) may be related
33 to foraging behaviour and gut functioning, while the motivation behind gentle feather pecking is less clear and may relate to other behaviour patterns not often focused on in feather pecking studies, such as motivation to preen or allo-groom. In addition to the causal basis of feather pecking, other factors also influence its prevalence. Improper early experiences, such as having no substrate to peck at and no hen to encourage ground pecking, may exacerbate the problem and in some instances perseveration of the behaviour may occur. If this happens, the behaviour may persist, even if the bird is moved to enriched, more natural conditions. Feather pecking also has a genetic component and it is possible that there has been some indirect selection for it, either because it is linked to a production trait or because birds with very good feather cover
(and therefore with a higher risk of performing the pecking rather than being the pecked birds) have been selected for breeding. If feather pecking is actually a reaction style to help deal with continually stressful environments, then birds performing this behaviour
(proactive responders - actively deal with problems) may actually have better welfare than those that do not (reactive responders - hypothetically in 'depression'-like states).
Thus, proactive responders may be trying to satisfy foraging motivation or gut
abnormalities through feather pecking while reactive responders are doing nothing to
alleviate the problems and, thus, may potentially also be a welfare concern. However,
experimental evidence is still needed to prove or disprove these claims.
At the beginning of this review, we classified feather pecking as Abnormal
Repetitive Behaviour, which conveyed nothing about the cause of the behaviour.
However, based on the evidence presented in the above review, it appears that feather
34 pecking fits the definition of stereotypic behaviour: 'repetitive behaviour induced by frustration (frustrated foraging motivation), repeated attempts to cope (with barren and stressful environments) and/or C.N.S. (brain) dysfunction (brought about by chronic stress and/or improper early experiences)' (cf Mason, 2006). It is recommended that future research classify feather pecking behaviour as such, and past methods that have been used to study stereotypic behaviour are implemented in the study of feather pecking behaviour.
To date our methods of dealing with feather pecking are inadequate: feather pecking is still found in battery cages and beak trimming does not seem to prevent feather pecking, although it does decrease damage from feather removal. It is possible that beak trimmed birds continue to feather peck with little or no feedback from feather ingestion because of habit-formation or perseveration of this behaviour (appropriate stimuli are no
longer necessary). It is also possible that the act of feather pecking itself is reinforcing -
foraging behaviour is thought to have its own internal motivational factors which drive
the bird to perform the behaviour even when this is not necessary for food consumption
(Hughes and Duncan, 1988a), and feather pecking may behave similarly. The problems
of stereotypic pelage/plumage removal or stereotypic oral/foraging related behaviour are
not unique to poultry. Similar stereotypic behaviour occurs in other species in which
foraging and forage intake is still important even though they are provided concentrated
feed which they can consume quickly. Comparisons of these stereotypic behaviour
patterns may help improve our knowledge of feather pecking in chickens.
35 In order to completely solve the problem of feather pecking, the above factors need to be accounted for and determination of the underlying motivation must be known or we risk applying 'band-aid' solutions, which may only work through distraction or increases in opportunity for activity and not by actually solving the problems. These
'band-aid' solutions are a risk when implementing any enrichment program because while they appear outwardly to be solving the problem (decreasing stereotypic behaviour), the underlying cause may still be present and thus, the animal's welfare may
still be compromised.
In conclusion, the purpose of this thesis is to investigate the underlying motivation of feather pecking through testing of the hypothesis that feather pecking stems
from redirected foraging motivation.
36 nm m |||[|B
m
(Photo by L.M. Dixon)
Figure 1.1: An example of severe feather pecking that led to bleeding.
37 CHAPTER TWO
Changes in substrate access did not affect early feather pecking behaviour in two
strains of laying hen chicks
(A version of this chapter has been accepted for publication in the Journal of Applied
Animal Welfare Science.)
ABSTRACT
Feather pecking is detrimental to bird welfare, but is commonly found in flocks of laying hens (Gallus gallus). Layers are normally housed without a floor substrate, which is thought to cause this problem. Some evidence suggests that early substrate access decreases later feather pecking. However, there has been little research on the immediate effects of a change in substrate availability on bird welfare, though environmental modifications like this are often done when brooding and rearing laying hen chicks. To investigate this, the behavior of two strains of laying hen chicks was recorded for four weeks. The birds were kept on either wire or peat moss until 14 days and then half the chicks were switched to the other flooring. Early feather pecking was not significantly different for birds started on peat moss and switched to wire than for birds only on wire
(P>0.05). Since moving chicks from peat moss to wire did not cause additional welfare problems, it is recommended that chicks be kept on a substrate when young as feather pecking levels are lower and immediate welfare is improved compared to birds kept only on wire.
38 1. INTRODUCTION
Abnormal behavior patterns are commonly found in smaller, barren environments, and may indicate poor welfare (Mason, 1991; Mason and Latham, 2004), as they are generally detrimental to the animal itself or to the group the animal is kept in (Garner,
2005). Feather pecking in laying hens can be classified as stereotypic behaviour: repetitive behaviour induced by frustration, repeated attempts to cope and/or brain dysfunction, regardless of the degree of variation or repetition {cf. Mason, 2006). It is known to be influenced by current environment (Huber-Eicher & Audige, 1999; Nicol et ah, 2001); however, the effects of early experience with a substrate on later behavior are still disputed as different studies have found different results. For example, Vestergaard and colleagues (1993) found early substrate experience to be protective against later feather pecking, while Huber-Eicher & Wechsler (1997a) and Newberry and colleagues
(2007) did not. Past studies have mainly focused on the effect of early experience on behavior later in life and did not examine any immediate consequences of changing environments on stereotypic behavior. Laying hens are often switched between environments with and without substrates after hatching, brooding, or rearing (e.g. after hatching, chicks may be moved to a brooder with wood-shavings, then switched to wire cages three weeks later until point of lay) (Sainsbury, 2000). Since these environmental changes are recommended industry practice (e.g. North & Bell, 1990; Gillespie, 2004), it is important to know if there are any immediate consequences on bird welfare.
39 Feather pecking also has a genetic component (Cuthbertson, 1980) and levels of feather pecking have been studied in a number of poultry strains. However, many strains tend to be dual purpose (selected for meat and eggs), or Red Junglefowl (ancestor of domestic fowl), and thus are not commonly used as commercial egg production strains
(e.g. Hughes & Duncan, 1972; Vestergaard et ah, 1993; Savory & Mann, 1999). There has been work done using commercial strains to estimate feather pecking heritability but
these studies focused more on feather pecking heritability among different ages of birds
(Kjaer & Sorensen, 1997) or used lines differing in feather pecking behavior (high and
low feather pecking lines) to determine other behavioural differences (Rodenburg &
Koene, 2003) and did not focus on the effects of certain early experiences on different
commercial strains. It is possible that different strains may respond differently to the
addition or removal of substrates and a line better suited to these production practices
could be found. In Canada, two commonly used egg laying strains are ISA White
Leghorns and ISA Brown Leghorns (Al Dam, Poultry Specialist, OMAFRA, personal
communication). Thus, we will be examining the feather pecking in these two strains so
the results will be more applicable to Canadian industry.
The objectives of this study were threefold. The first objective was to test what
effect early experience with or without a preferred natural substrate (peat moss) would
have on immediate bird welfare. The second objective was to investigate the effects of
early environment (with or without a substrate) on feather pecking in a subsequent
environment (with or without a substrate). The third objective was to examine the
40 difference in feather pecking behavior early in life between two commonly used strains of egg laying chicks in these different environments.
We tested the hypothesis that high levels of feather pecking are the result of early experience without exposure to a loose substrate. We predicted that early experience with peat moss (a much preferred dustbathing substrate: Petherick and Duncan, 1989) during weeks one and two of life would be protective and decrease feather pecking during weeks three and four of life compared to wire reared birds and that birds with appropriate early experience of peat moss during weeks one and two then switched to wire (change of environment) for weeks three and four should feather peck less than birds reared and housed on wire (no change of environment). We also hypothesized that different strains of birds would feather peck different amounts due to the genetic component of feather pecking and predicted that ISA White Leghorns would feather peck more than ISA Brown Leghorns, as ISA Whites have been anecdotally reported to have very high levels of feather pecking.
2. METHODS
2.1. Animals
Day-old female non-beak trimmed ISA White Leghorns (white egg layers) and ISA
Brown Leghorns (brown egg layers) were obtained from a local hatchery (Bonnie's
Hatchery, Elmira, Ontario) and housed at Arkell Poultry Research Facility (Arkell,
41 Ontario) for the duration of the study. Chicks were randomly distributed throughout the pens, at twelve birds per pen, 576 birds per strain in total and 48 pens per treatment in total. Strains of chicks were kept separate throughout the experiment.
Power tests were performed before beginning this experiment using estimates from previous experiments (Blokhuis & Arkes, 1984 and Johnsen, Vestergaard, & Norgaard-
Nielsen, 1998) that compared feather pecking on wire and substrate reared birds. A
sample size of n = 48 per treatment (wire or peat moss flooring), at a power = 90% and a
= 0.05, was used to detect a 30% difference between the treatment averages, which is a
level equal or superior to that of previous work.
2.2. Housing
The chicks were kept in wooden frame floor pens measuring 127cm x 107cm x
60cm with sides of 2.5cm x 2.5cm grid chicken wire. Half the pens had 1cm x 1cm grid
wire flooring and half had solid floors covered with approximately two inches of peat
moss. The sides of the pens were covered with black cloth, so chicks could not see into
other pens. The pens had a fine mesh cover over top to prevent chicks from escaping.
The chicks were fed ad libitum chick starter crumbles (Protein 20%>, ME 1290 kcal/lb)
and water from automatic cup drinkers. The feed was sifted before being fed to the
chicks to remove as much dust as possible and discourage dustbathing at the feeder.
Lights were left on 24 hours a day for the first three days, then reduced to an 10L:14D
schedule. The temperature started at 34°C and was reduced to 28°C over the four weeks
42 of the trial. Lighting was provided by incandescent bulbs and measured 60 lux at approximately the head level of the birds. This remained consistent for the duration of the trial.
2.3. Observations
Observations of the chicks were conducted from one to four weeks of age. This was divided into two periods. Period One covered weeks one and two of life and Period
Two was weeks three and four of life. There were 48 pens per treatment during both
Period One and Period Two. There were two observers collecting the data both live and from video. There was no significant difference between observers (P<0.05) or between live and video observations (P>0.05). A twenty minute observation session was performed twice weekly on each pen (once between 730 and 1200 hours and once between 1200 and 1630 hours).
During the twenty minute observation session, focal-animal, all-occurrence sampling was used. Each pen was divided into six equal-sized, unmarked sections and a table of random numbers was used to determine which section to observe. The chick closest to the center of that section was observed. If there was no chick in this section, the next numerical section was used until a chick was found. The focal animal was switched every minute, for a total of 20 animals per session and it was possible for a bird to be observed more than once in an observation session.
43 For Period One, there were twelve chicks in each pen (Fig.2.1). At the end of this period, each pen was split in two by a plastic board to maintain recommended space allowances and this board was also a novel environmental stimulus that all birds experienced. At this point, six chicks from each pen remained in that pen, while the other six were moved to a pen in the other treatment (Fig.2.2). For example, with a wire pen, six birds remained in the wire floor treatment, and six birds were removed and placed in a peat moss pen. These birds were not mixed with the other treatment birds and moving was balanced across treatment. The six birds moved remained as a group and were placed in an empty half of a peat moss pen. Birds that did not switch treatments were handled and then returned to their own pen with the original flooring. This resulted in
four treatments of birds: 1) Birds that had originally been on wire and were kept on wire
(P1P2 Wire), 2) Birds that were originally on peat moss and were kept on peat moss
(P1P2 Peat Moss), 3) Birds that were on wire and were switched to peat moss (PI Wire
P2 Peat Moss), and 4) Birds that were on peat moss and switched to wire (PI Peat Moss
P2 Wire).
2.4. Behavioral Measures
The following behavior patterns were recorded: Dustbathing was defined as the
chick squatting on the ground, kicking its legs, rubbing its head and side on the ground
and performing a vertical wing shake. The number of dustbathing bouts were counted
and considered to be the time from when the birds were squatting on the ground kicking
its legs until a vertical wing shake was performed. Chickens perform these same
44 movements when "vacuum" dustbathing on wire (Vestergaard, 1982a). Feather pecking was defined as the pecking at or removal of feathers or down feathers from one bird by the pecking of another bird. Ground scratching was defined as the chick scratching the ground with one leg at a time, alternating between legs. Ground pecking was defined as the chick directing a peck at the ground or particles on the ground (e.g. Hoffmeyer,
1969). A behavior pattern was considered to last until there was a four second or longer pause in the performance of this behavior. At this point, the next behavior the focal animal performed was recorded.
2.5. Ethical note
The use of all animals and methods in the following experiments were approved by the University of Guelph Animal Care Committee which adheres to Canadian Council on
Animal Care guidelines.
The birds used in this study were not beak trimmed, so to prevent potential
cannibalism, the birds were inspected a minimum of three times a day (once at 800 hours,
once at 1200 hours and once at 1600 hours). Any feather pecking injuries found were
coated in pine tar to discourage further pecking and the bird was checked once an hour
for the rest of the light period to ensure the injury did not get worse. If the pine tar was
unsuccessful in discouraging pecking, the bird would be removed from the pen, however,
this measure was not needed in this study.
45 2.6. Statistical Analysis
A mixed model variance component analysis was used to determine the effect of pen floor type (wire or peat moss) and previous experience with a floor type on chick behavior (SAS, Ver.8). The model had Strain as a blocking factor, Pen nested in Trial as
a random factor and all interaction terms between the independent variables were
included. Data were normalized with a V(x + 0.375) transformation. Differences of the
Least Square Means adjusted by Tukey were used to determine treatment differences.
3. RESULTS
3.1. Feather Pecking
Period One (PI): In the first two weeks of life, all chicks kept on wire feather pecked
more than chicks kept on peat moss (F(i;4o) = 44.2, PO.0001), although these levels
were quite low (~3 feather pecking bouts/20 minutes for wire housed, ~1 feather pecking
bout/20 minutes for peat moss housed) (Table 2.1). ISA Whites feather pecked more
overall than ISA Browns in Period One (Strain: F(1,4o) = 11 -07, P = 0.0019) (Table 2.3).
There was no interaction between strain (WL and 1SAB) and treatment (wire or peat
moss) (F(i,4o) = 0.07, P = 0.7869).
Period Two (P2): Again, for both ISA White and ISA Brown chicks, those that were
on wire feather pecked more than those on peat moss (6 feather pecking bouts/20 minutes
46 vs 2 feather pecking bout/20 minutes), regardless of early experience with a substrate, thus chicks housed on peat moss then switched to wire did not feather peck significantly
less from those only on wire (F(3j80) = 38.09, PO.0001) (Table 2.2). ISA White chicks feather pecked more overall than ISA Brown chicks (F(i, 80) = 70.62, PO.0001) (Table
2.4) and there was no significant interaction found between these factors (F^ go) = 2.22,
P = 0.0913).
3.2. Dustbathing
All birds dustbathed at the same frequency throughout the trial, regardless of their previous experience (F(i,4o) = 6.46, P = 0.0948) or the current environment they were
being housed on (F^ so) = 1-53, P = 0.2133) (Tables 2.1 and 2.2). There was no
difference in the amount of dustbathing performed by ISA Brown and ISA White chicks
(PI: F(i>4o) = 0.01, P = 0.9034; P2: F(1> 8o) = 2.89, P = 0.0884) (Tables 2.3 and 2.4), nor
were there any interactions of treatment and strain (PI: Trt*Strain: F(i,4o) = 0.02, P =
0.9010; P2: Trt*Strain: F(3,8o) = 1-34, P = 0.2676).
3.3 Ground Pecking and Scratching
Period One (PI): Both ISA White and ISA Brown chicks housed on peat moss
performed more ground pecking and scratching than those on wire floor (F(i;4o) =
342.93, P<0.0001) (Table 2.1). ISA Brown chicks had a higher overall frequency of
47 these behavior patterns than ISA Whites (F(ij4o) = 12.67, P = 0.002) (Table 2.3) but there was no significant interaction of strain and substrate access (F(i,4o) = 0.01, P = 0.9078).
Period Two (P2): Regardless of previous floor type, both strains of birds ground pecked and scratched more when kept on peat moss than when kept on wire floor (F(3) so)
= 277.36, PO.0001) (Table 2.2). There were also no differences in the ground pecking and scratching between strains of birds on their current floor types (F(i, so)= 0.69, P =
0.409) (Table 2.4) and no significant interactions of between strain and treatment (F(3; 80)
= 3.1,P = 0.0911).
4. DISCUSSION
The presence of peat moss decreased the amount of feather pecking regardless of the chicks' early experience with it, and the prediction that early experience (during weeks one and two of life) with a substrate would decrease feather pecking behavior after substrate removal (during weeks three and four of life) was not met. This is not likely to be a Type II error - with this sample size (n = 48 pens per treatment) we can say with confidence that there cannot have been an effect bigger than a 30% difference between the means (a level less than or equal to past research where differences in early experience were found). Chicks are thought to show preferences for dustbathing on or pecking and scratching at a particular substrate around eight to ten days of life
(Vestergaard, 1982b; Sanotra et ah, 1995). In this experiment, chicks were kept on their original floor treatment until fourteen days of life, giving them time to develop
48 preferences for either the peat moss or the feathers of other birds as dustbathing substrates or forages. These findings do not fit the hypothesis that the feather pecking shown early in life stems from lack of early experience with a substrate. Instead, only the current environment had an impact on feather pecking in the first few weeks after the environment was changed.
There has been extensive research done on the importance of early life experiences on later behavior (e.g. Rentier and Rosenweig, 1987). In laying hens, it has been demonstrated that early experience with an appropriate substrate protects against later feather pecking (Vestergaard, 1994; Johnsen et al., 1998). For example, birds that were reared only on straw feather pecked more than those reared on sand + peat, even after all were switched to straw alone (Norgaard-Nielsen et al., 1993). There are, however, instances where the current environment has a greater influence on behavior than previous experience (e.g Nicol et al, 2001). For example, some studies on feather pecking only found an effect of current environment: early access to 'litter' did not reduce the later rates of feather pecking in commercial flocks of floor-housed poultry
(Gunnarsson et al, 1999). Sand, which was protective in some studies, did not prevent injurious feather pecking when chicks were kept on it from day one of life but chicks kept with sand and straw did not develop this problem (Huber-Eicher & Wechsler, 1997a).
More recently, Newberry et al. (2007) did not find any association on an individual bird basis between low levels of early foraging behavior on a substrate and adult feather pecking behavior, in fact, birds that foraged more when young were more likely to perform severe feather pecks when older.
49 A reason for the lack of protective effects from early substrate experience in this
study may be that the observations ended after the fourth week of life, while in other
experiments they were later in life, even past when birds came into lay. Unfortunately, we did not have appropriate housing for older (and larger) birds when this experiment was being conducted, so early experience effects could not be examined later in life.
Although birds have been shown to feather peck early in life and this feather pecking is
affected by the presence of a substrate (e.g. Savory & Mann, 1999; Nicol et al, 2001;
Jensen et al, 2006), the protection from early substrate experience may not be shown
until the birds are older, so three to four week old chicks may not exhibit any benefit of
early substrate experience. However, we feel a more likely explanation would be the
quality of the substrate or enrichment given to our chicks compared to other studies. The
quality of the enrichment influences how effective it will be in preventing stereotypic
behaviors, e.g. a chewing toy did not decrease tail biting in pigs but straw did (Van de
Weerd et al, 2005). For chickens, the quality of the enrichment also influences the
amount of time spent interacting with it, with higher quality stimuli having more
interactions (Huber-Eicher & Wechsler, 1998). The peat moss provided to the birds in
this experiment is a preferred substrate for pecking and scratching (Petherick & Duncan,
1989) and it may have been more effective at preventing feather pecking than some of the
less preferred substrates used in other experiments (e.g. straw, sand, wood-shavings).
Finally, the lack of protection from early substrate experience may be the result of half
the chicks being moved to a new environment (confound of move and new treatment),
while others remained in their original pens, even though birds not being moved or
50 changing treatment were handled in a similar way to those that were. It is possible that the experience of a novel environment superseded any beneficial early experience.
However, we feel this is unlikely, as at the point in the experiment where half the chicks were moved, all pens had a novel divider included. This divider would be novel for all birds and would change the look and the size of the environment in the original pens as well as the new pens. Furthermore, birds given the opportunity to interact with novelty perform less feather pecking than birds that did not (Chow and Hogan, 2005; Chapter 4).
Thus, the birds moved to the novel environments should have shown less feather pecking compared to those that did not move (e.g. birds moved from wire to peat moss should have feather pecked less than birds who did not move and were only kept on peat moss), but this was not the case. It was the current environment that influenced the amount of feather pecking performed, at least early in life.
Although the exact benefits of being reared with a substrate are not clear, the provision of a substrate does decrease feather pecking and the short term consequences of
changing environments does not exacerbate this behavior. It appears that the most reliable way to have low feather pecking rates in a flock is to provide some form of
substrate with enough room for the birds to interact with it in a natural way. Not only are
the consequences of feather pecking harmful but the performance of this behavior alone
indicates that there is a problem with how chickens are being kept. The results found
here add to the evidence that continued research is needed to develop enriched cages or
floor systems that improve poultry welfare (see Chapter 4).
51 In this experiment, current environment influenced ground pecking and scratching, with birds housed on a substrate performing significantly more pecking and scratching at the peat moss than birds reared on wire, regardless of early substrate experience.
However, the number of dustbathing bouts remained similar between treatments (peat moss and wire) and there was no effect of early experience with or without a substrate.
Similar to ground pecking and scratching, the amount of feather pecking was also affected by provision of a substrate, with birds on wire feather pecking more than those on peat moss, again, with no affect of early experience. This complements the existing evidence of a relationship between feather pecking and pecking and scratching at a substrate. Since the numbers of dustbathing bouts were not affected by the presence of a substrate, while ground pecking and scratching and feather pecking were, it does not appear that dustbathing motivation is associated with feather pecking.
Based on the above evidence, it appears that re-directed foraging motivation plays more of a role in feather pecking than re-directed dustbathing. However, the data presented here were not measures of durations but the number of occurrences of behaviour patterns, so it is possible that these data may not tell the whole story.
Examining the daily time budgets of chicks, and determining how much time is spent performing various behavior patterns (dustbathing, foraging, feather pecking, etc.) in various environments (barren, with substrates) would help confirm the redirected
foraging hypothesis (see Chapter 3). Due to the difficulties in determining the difference between dustbathing and foraging pecks (which are both directed at the ground and look
52 similar), a different technique may need to be developed to differentiate between the two and determine how much time is actually spent doing each activity (see Chapters 5).
There was a strain difference in the amount of feather pecking performed between the ISA Whites and ISA Browns, with ISA Whites feather pecking more than twice as much as ISA Browns. By the end of the trial, there were no strain differences in the amount of dustbathing or ground pecking and scratching. While having darker feathers predisposes chickens to becoming victims of feather pecking (Keeling et ai, 2004), the highest feather pecking rates were shown in groups of mixed light and dark feathered birds, with the white-feathered birds doing more of the pecking (Savory and Mann,
1999). Alternately, ISA Browns were found to have higher feather pecking levels than two other types of White Leghorn (Lohmann Selected Leghorn and Norbrid 41) (Kjaer,
2000). However, we used ISA White Leghorns and as feather pecking has a heritable component (e.g. Cuthbertson, 1980; Kjaer & Sorensen, 1997), it appears that different strains exhibit different amounts of feather pecking. Thus, as ISA Browns and ISA
Whites are commonly used in Canadian industry, and ISA Browns show lower levels of
feather pecking early in life, their use should be encouraged and they could potentially be crossed with higher feather pecking strains to reduce overall levels performed.
5. CONCLUSIONS
Our results do not support the hypothesis that feather pecking early in life stems
from lack of early experience with a substrate. Early experience with the substrate during
53 the first two weeks of life did not protect the birds from feather pecking during weeks three and four; it was the current environment that influenced the amount of feather pecking performed. However, there were no negative consequences found by housing the birds with a substrate early in life and switching them to a more barren environment
(wire), other than those already found in wire housed birds. As birds have better welfare when kept with a substrate due to the ability to perform a greater range of behaviour patterns, this practice is recommended. Also, ISA Browns perform overall less feather pecking in the first four weeks of life than ISA Whites and their increased use over ISA
Whites should be encouraged both in industry and research. Finally, these results emphasize the need for enriched cages or pens and indicate that the ability to forage may be more important in preventing feather pecking than dustbathing. However, due to the
similarity of dustbathing and foraging pecks, new techniques are needed to determine definitively if the motivation behind feather pecking relates to foraging or dustbathing.
54 ^HM i- ^^^i M > ••• ' • u :G»??V' ,-,v,v '-"'*•• • •. ?>J^ft*W",\ywa^i*** • - •••- " .^....
•• ri t I-., t .u - . .1 •«• *• "** fafavstfr •--..?•'.-, •1 h - ; . - • i * 1 ' *;•*. * .--•',:,'• •; _-M- 1 >.*.* ** - i^ - »j - '.-' •-•"•••i -."^ I «• ^ .«»••> | J ; - - • . 1 •"
"<" • • s O , t.'jl - . .1-
Peat Moss Wire
=Feeder Waterer Q = Chicks
Figure 2.1: The experimental set-up for chicks in peat moss and wire floored pens for
Period One.
55 •^•!f»K.«"V.3| . ." -. «:-•.
Peat Moss Wire
Pen Divider Feeder = Waterer = Chicks
Figure 2.2: The experimental set-up for chicks in peat moss and wire floored pens for
Period Two.
56 Table 2.1: The average frequencies (per twenty-minute observation period) of feather pecking, dustbathing, and ground pecking and scratching on wire and peat moss floor treatments, combining strains, for Period One.
Period 1 Peat Moss Wire F(df) P-Value (mean ± SE) (mean ± SE)
Feather Pecking 1.38 ±0.12 2.7±0.18 F(l,40) = 44.2 O.0001 *
Dustbathing 0.304 ± 0.07 0.075 ± 0.022 F(l,40) = 6.46 0.0948
Ground Pecking and 17.23 ±0.88 2.97 ±0.23 F(l, 40) = 342.93 O.0001 * Scratching
57 Table 2.2: The average frequencies (per twenty-minute observation period) of feather pecking, dustbathing, and ground pecking and scratching on wire and peat moss floor treatments, combining strains, for Period Two. Differences in letters (a and b) denote significant differences while similarities in letters denote no significant difference.
Period 2
Peat Moss Wire then Peat Moss Wire Only F(df) P-Value Only Peat Moss then Wire (mean ± SE) (mean ± SE) (mean ± SE) (mean ± SE)
Feather 1.79±0.15a 2.75 ±0.25 a 5.64 ± 0.35 b 5.48 ± 0.38 b F(3, 80) = 38.09 <0.0001 Pecking
Dustbathing 0.245 ± 0.048 0.195 ±0.047 0.18 ±0.033 0.225 ± 0.04 F(3, 80) =1.53 0.2133
Ground Pecking and 20.18±1.06a 18.6 ±0.64 a 4.23 ± 0.23 b 4.45 ± 0.23 b F(3, 80) = 277.36 O.0001 Scratching
58 Table 2.3: The average frequencies (per twenty-minute observation period) of feather pecking, dustbathing, and ground pecking and scratching for ISA White Leghorns and
ISA Brown Leghorns during Period One, combining Treatments.
Period 1
WL ISAB F(df) P-Value (mean ± SE) (mean ± SE)
Feather Pecking 2.35±0.18 1.72±0.14 F(l, 40) = 11.07 0.0019 *
Dustbathing 0.243 ± 0.06 0.246 ± 0.08 F(l,40) = 0.01 0.9305
Ground Pecking and 8.92 ± 0.75 12.49 ±0.97 F(l, 40) =12.67 0.002 * Scratching
59 Table 2.4: The average frequencies (per twenty-minute observation period) of feather pecking, dustbathing, and ground pecking and scratching for ISA White Leghorns and
ISA Brown Leghorns during Period Two, combining Treatments.
Period 2 WL ISAB F(df) P-Value (mean ± SE) (mean ± SE)
Feather Pecking 4.88 ±0.16 2.42 ± 0.27 F(l, 80) = 70.62 <0.0001 *
Dustbathing 0.20 ±0.05 0.22 ± 0.06 F(l,80) = 2.89 0.0884
Ground Pecking and 11.3 ±0.53 13 ±0.7 F(l,80) = 0.69 0.409 Scratching
60 CHAPTER THREE
A comparison of the diurnal rhythms involved in feather pecking, foraging
and dustbathing in laying hen chicks
ABSTRACT
Feather pecking has been hypothesized to stem either from re-directed dustbathing or foraging behaviour. Dustbathing and foraging involve some similar motor patterns, but are however, quite distinct in their diurnal distributions. Dustbathing is performed every other day for about half an hour, while birds will spend 60% of the light period foraging if given the opportunity. The aims of this study were to examine the diurnal distribution of feather pecking in chicks during their first month of life and compare this to the distributions of dustbathing and foraging. The light hours were divided into three, three-hour periods and 24 pens with six birds in each were observed twice a week during each period for the first four weeks of life. Half the birds were kept on wire flooring and half were housed on peat moss, as substrate access is known to affect feather pecking behaviour and, as a result, may influence diurnal rhythm. Focal animal sampling was used with all-occurrences of bouts of feather pecking, foraging, dustbathing and feeding recorded, bouts being separated by four second 'still' intervals.
Mixed model analysis found no effect of time of day on feather pecking (P = 0.1076),
foraging (P = 0.3176) or feeding behaviour (P = 0.35). After the first week of life,
dustbathing frequency was highest in period two (from four to seven hours after lights
61 on) (P = 0.0313). More research into the underlying motivation is needed to draw a definitive conclusion; however, based on timing throughout the day, feather pecking appears to be similar to foraging not dustbathing behaviour.
1. INTRODUCTION
Feather pecking is the pecking of feathers or down of one bird by another
(Hoffmeyer, 1969). It is hypothesized to be caused by either re-directed motivation to
dustbath (Vestergaard and Lisborg, 1993; Vestergaard et al, 1993) or re-directed motivation to forage (Blokhuis 1986; 1989), due to lack of appropriate substrates to perform these behaviours on. Dustbathing and foraging are similar in some respects,
both involving pecking and scratching at the ground, and low levels of feather pecking
have been associated with high levels of ground pecking (e.g. Blokhuis 1989). However,
due to the similarities in dustbathing and foraging motor patterns, it is not clear what the
underlying motivations behind these ground pecks are.
One major difference between dustbathing and foraging is their diurnal rhythms.
When birds have continual access to litter, they do most of their dustbathing at about five
hours after the lights come on, or five hours after the sun rises if they are kept in more
extensive environments. A dustbathing bout will last approximately half an hour and be
performed every other day (Vestergaard, 1982a). Foraging does not have this distinct
62 kind of rhythm. Birds feed and forage frequently throughout the day, at least once per hour, and thus, these behaviour patterns can be observed throughout the day (Duncan et al, 1970).
The diurnal rhythm of feather pecking is not as clear as that of dustbathing and has only been looked at in older birds. Studies of 38 week-old and 50 week-old layer strains found an increase in the amount of feather pecking in the later part of the daylight hours (38 week: Kjaer, 2000; 50 week: Preston, 1987), which suggests that feather pecking follows a different diurnal rhythm than foraging or dustbathing. However, one study divided the day into two periods and found an increase in feather pecking in the second period of the day (Preston, 1987), while the other study looked at feather pecking every hour and found an ongoing increase over the light period (Kjaer, 2000). A later experiment recorded the incidence of feather pecking throughout the day with 25 and 73 week-old birds and found that the highest rates of severe feather pecking were found in morning (25 week old: 100-1200 hours; 73 week old: 500-1200 hours) compared to the afternoon (25 week old: 1300-1600 hours; 73 week old: 1300-2000 hours) (Bright, 2007).
The environments were also different for the birds, with one study housing birds in wire cages (Preston, 1987), one keeping them on sand in a loose housed system (Kjaer, 2000) and one using free-range birds (Bright, 2007). As the frequency of feather pecking is influenced by the provision of a substrate, the diurnal distribution of feather pecking may also change depending on substrate access. The birds from these experiments were in lay and kept with long daylight hours. None of the studies looked at the early morning light hours; both instead started their observations five to six hours after lights on. From this,
63 it is not clear what the diurnal rhythm of feather pecking is or if chicks kept with shorter daylight hours will have similar distributions of feather pecking throughout the day as older birds.
To determine the underlying motivation of a stereotypic behaviour, such as feather pecking, aspects of the stereotypic behaviour can be compared to aspects of normal behaviour patterns (e.g. Mason, 1993). Thus, a comparison of the diurnal rhythms involved in feather pecking, dustbathing and foraging may indicate which normal behaviour pattern feather pecking stems from.
The objectives for this study were threefold. The first was to look at the temporal distribution of feather pecking over the day light hours in laying hen chicks. The second was to compare the distribution of feather pecking with that of dustbathing, foraging and feeding. The third was to examine the effect of housing with and without a substrate to determine if there was an effect on the daily distribution of feather pecking behaviour, as substrate access is known to influence the amount of feather pecking performed; it may also have an effect on diurnal rhythm.
It was hypothesized that feather pecking stems from a motivation to forage. From this, it is predicted that feather pecking behaviour will follow a similar daily distribution as foraging. Alternately, it was hypothesized that feather pecking stems from a motivation to dustbath and is predicted that feather pecking will follow a similar daily distribution as dustbathing.
64 2. METHODS
2.1 Animals and Housing
A total of 154 female day-old, non beak trimmed Black Rock chicks (Jarvis
Chicks, Ontario) were used in order to maximize feather pecking results, as they have been anecdotally reported by local producers to have problems with high levels of feather pecking.
Chicks were randomly distributed to wooden-framed floor pens measuring 127cm x 107cm x 60cm, with sides of 2.5cm x 2.5cm chicken wire. The pens were all divided in two by a board and six chicks were housed in each half of the pen (24 halves in total).
This was done to increase sample size, each half pen being one unit, as there were not enough whole pens for the number of chicks needed. Half the pens had 1cm x 1cm wire flooring and half were floored with approximately 5cm of peat moss. The sides of the pens were covered with black cloth to prevent visual contact with other chicks. The chicks were fed ad libitum chick starter crumbles in slide top feeders and water from automatic cup drinkers (Fig.3.1). Feed was added to the feeders after observations for the day had concluded, so feeding bouts stimulated by the addition of fresh feed were not recorded. Lights were left on 24 hours a day for the first three days, then reduced to an
10L:14D schedule. The temperature started at 34°C and was reduced to 28°C over the
four weeks of the trial.
65 Power tests were performed before beginning this experiment using estimates from previous published experiments (Preston, 1987; Kjaer, 2000) that compared feather pecking during different times of day. A sample size of n = 12 per floor treatment (all pens were observed in all three light periods), at a power = 90% and a = 0.05, was needed to detect a 25% difference between the averages for time of day within each floor type (Berndtson, 1991), which is a level superior to that of previous work where differences in diurnal rhythms were successfully detected.
2.2. Observations
Live observations were conducted by one observer from one to four weeks of age.
The light hours were divided into three periods, each three hours long, with a half hour free at the beginning and end of the day to allow for room maintenance. Period one was from 730 to 1030 hours, period two was from 1030 to 1330 hours and period three was from 1330 to 1630 hours. A thirty minute observation session was performed twice weekly on each pen for each time period. During the thirty minute observation session, all-occurrence sampling and focal animal sampling were used. Each pen was divided into six equal-sized, unmarked sections and a table of random numbers was used to determine which section to observe. The chick closest to the center of that section was observed. If there was no chick in this section, the next numerical section was used until
66 a chick was found. The focal animal was switched every minute, for a total of 30 animals per session and it was possible for a bird to be observed more than once in an observation session modified from Martin and Bateson, 2007).
2.3. Behavioural Measures
The following behaviour patterns were recorded: Dustbathing was defined as the bird squatting on the ground, kicking its legs, rubbing its head and side on the ground until it stands and performs a vertical wing shake. A dustbathing bout was considered to start when the bird squatted on the ground and kicked its legs to move (or attempt to move) dust onto its feathers and ended when the bird stood, shook its feather and remained still (i.e. no vertical wing shakes or ground pecks) for 4s or more. Feather pecking was defined as the pecking at or removal of feathers or down feathers from one bird by the pecking of another bird. Ground scratching was defined as the chick
scratching the ground or wire floor with one leg at a time, rotating between legs. Ground pecking was defined as the chick directing a peck at ground or particles on the ground or wire floor. Feeding was when the chick was standing with its head in the feeder and was pecking at feed particles (e.g. Hoffmeyer, 1969). A behaviour pattern was considered to be performed once until there was a four second or longer pause in the performance of
the behaviour. At this point, the next behaviour that the focal animal performed was
recorded.
67 2.4. Ethical note
The use of all animals and methods in the following experiments were approved by the University of Guelph Animal Care Committee which adheres to Canadian Council on
Animal Care guidelines.
The birds used in this study were not beak trimmed, so to prevent potential cannibalism, the birds were inspected a minimum of three times a day (once at 800 hours, once at 1200 hours and once at 1600 hours). Any feather pecking injuries found were coated in pine tar to discourage further pecking and the bird was checked once an hour
for the rest of the light period to ensure the injury did not get worse. If the pine tar was unsuccessful in discouraging pecking, the bird would be removed from the pen, however, this measure was not needed in this study. In fact, the use of pine tar itself was rarely needed - throughout the study there were only two occurrences where pine tar was
needed to discourage further pecking of a bloody injury.
2.5. Statistical Analysis
Ground scratching and pecking bouts were combined for the analysis and this
combination was labeled as foraging. A dustbathing bout started when the bird squatted
on the ground and ended when the bird stood and shook its feathers. A mixed model
variance component analysis was used to compare between and determine the effects of
time of day (daylight hours broken into three periods) and housing, with or without a
68 substrate, on feather pecking, dustbathing and foraging bouts, using pen nested in treatment as the subject and week as the repeated measure (SAS, Ver.8). Data were normalized using a V(x + 0.375) transformation. Differences of the Least Square Means adjusted by Tukey were used to further examine treatment differences. Pearson correlation coefficients, partialled for pen, were calculated to determine any potential relationships between dependent variables.
3. RESULTS
3.1 Distribution of Feather Pecking
There was no significant period difference in the amount of feather pecking over the three light periods (Fp, 89)= 2.25, P = 0.1076) (Fig.3.2). Birds on wire floor feather pecked more after week one and, except for week one, feather pecked more than birds on a peat moss substrate (F(3) 89) = 8.64, P <0.0001) (Fig. 3.3). However, floor type did not affect the distribution of feather pecking throughout the day (F(lj89) = 2.81, P = 0.1074).
3.2 Distribution of Dustbathing
There was a week by period effect for dustbathing (F(6,89) = 3.95, P = 0.0313). In the first week, there was no difference in the amount of dustbathing done throughout the day. However, from weeks two to four, dustbathing activity was - as expected - highest in the middle of the light period (between four to seven hours after lights came on) (Fig.
69 3.4). Floor type did not significantly affect dustbathing behaviour, with birds on wire floors sham dustbathing at similar rates to birds dustbathing in peat moss (F(i> 89)= 1.19,
P = 0.3487).
3.3 Distribution of Foraging and Feeding
There was no period effect on the amount of foraging performed throughout the day (F(2,89)= 1.15, P = 0.3176) (Fig. 3.2). Birds kept on peat moss did more ground pecking and scratching than birds on wire and these behaviour patterns increased after week one of life (F(3( 89) = 5.97, P = 0.0007) (Fig.3.5).
Similarly, no period effect was found for the number of feeding bouts during the day (PI: 11.26 ± 0.66, P2: 10.15 ±0.69, P3: 11.31 ±0.80, F(2,89)= 1.06, P = 0.35). Birds on wire and peat moss performed the same number of feeding bouts during week one, after that, and wire housed birds were observed at the feeder more than birds on peat moss (F(3> 89) = 4.85, P = 0.0029) (Fig. 3.6).
3.4. Correlations
There were significant negative correlations between feather pecking and foraging
(P<0.05). In other words, levels of feather pecking were high when levels of foraging were low and vice versa. There were no significant correlations found, however, between feather pecking and dustbathing (P>0.05) (see Table 3.1).
70 4. DISCUSSION
The results of this study supported our prediction that feather pecking behaviour would follow a similar diurnal distribution as foraging and reject the prediction that feather pecking would follow a similar diurnal distribution as dustbathing. After the first week of life, dustbathing activity was highest in the middle of the day, from 1030 to 1330 hours, and is consistent with Vestergaard's (1982a) finding that dustbathing peaks about five hours after lights on. In addition, treatment of wire or peat moss floor did not affect dustbathing bout frequency. This is similar to the findings of Vestergaard and colleagues
(1990), in which birds kept on wire floors dustbathed at similar amounts to those kept on silica sand. It is possible that birds are so highly motivated to dustbathe that less preferred stimulus (such as wire floor) release the behaviour in an attempt to satisfy the motivation.
Feather pecking behaviour was distributed evenly throughout the daylight hours and did not show a marked peak. Similarly, ground pecking and scratching and feeding behaviour were spread evenly throughout the day, as was expected from previous research on foraging and feeding behaviour (e.g. Duncan et ah, 1970). However, frequency of feeding bouts was higher in wire housed than peat moss housed birds. It may be that birds are attempting to forage at the feeder as no other substrate is present.
As feather pecking levels are higher in wire housed birds, pecking at the feed does not seem to be able to substitute for foraging behaviour. This idea is further supported by the
71 positive correlation found between feather pecking and feeding behaviour - birds that spent more time at the feeder performed more feather pecking, thus pecking at the feed does not seem to be able to satisfy frustrated foraging motivation.
The feather pecking results found in the current study are somewhat different than what has been previously found. Preston (1987) and Kjaer (2000) found that feather pecking increased in the later part of the day. However, they were looking at birds in lay with long daylight hours and did not observe the early part of the day, including 5-6 hours after light on when the frequency of dustbathing has been shown to peak
(Vestergaard, 1982a). In contrast, Bright (2007) found an increase in severe feather pecking in the morning, coinciding with the peak dustbathing time. These differences may be due to the separation of gentle and severe feather pecking in Bright (2007)'s study or due to the flocks used being free-range with access to outdoor stimulation and natural light cycles. The data reported here show that birds in their first month of life evenly distribute feather pecking over the daylight hours, even as the overall behaviour increases. It is possible that as birds age and daylight hours are increased, various hormones influence reproductive ability and feather pecking frequency (Hughes, 1973), and the diurnal distribution of feather pecking may change, thus accounting for the results found previously. It is also possible that as daylight hours were grouped into three periods in the current study and not reported as individual hours, variations of feather pecking behaviour within the three time periods were missed. In addition, bout frequencies were reported in this study as opposed to individual event frequencies and this may also contribute to differences found between this study and others, as well as for
72 differences between past experiments. Longitudinal studies of gentle and severe feather pecking of birds in different housing systems may help to determine changes in daily distribution of the behaviour over the life of the birds and to clear up the discrepancies found in the diurnal rhythm of feather pecking found in past research.
5. CONCLUSIONS
Feather pecking bouts are spread evenly throughout the day during the first month of life while dustbathing bouts show a diurnal rhythm that peaks mid-way through the light period. However, foraging and feeding behaviours show a similar daily distribution to feather pecking. The hypothesis that feather pecking stems from foraging motivation is supported, while the hypothesis that feather pecking stems from dustbathing is rejected.
However, more research is needed into the motivation behind feather pecking to conclusively determine its cause.
73 ••'. B. / **• •
- •! .• — - T*-*^-* _i _ T? - .yqr. p -O
- ' "' - .v -. I.-0. _, 'P.-^V'^l *' -" -• »- f*fev 'V ** 1 -.
Peat Moss Wire
= Pen Divider = Feeder = Waterer o =Chick s
Figure 3.1: The experimental set-up for chicks in peat moss and wire floored pens.
74 23 i / / 4 fr I Period 1 C b -EE- 0 feather pecking dustbathing foraging Figure 3.2: The frequency of feather pecking, dustbathing and foraging over the three periods of the daylight hours, combining weeks and treatments. The data represent un- transformed values for the average number of feather pecking, dustbathing and foraging bouts per period for eight thirty-minute observation sessions ± SEM (Feather pecking: F(2,89) = 2.25, P = 0.1076; Dustbathing: F(2, 89) = 4.90, P = 0.0218; Foraging: F(2> 89) = 1.15, P = 0.3176). Different letters (a, b, c) represent significant differences (P<0.05) between periods. 75 9 Peat Moss Wine Figure 3.3: The average bout frequency of feather pecking in each week for peat moss and wire floor treatments. The data represent un-transformed values for the average number of feather pecking bouts per week for six thirty-minute observation sessions ± SEM. (F(3> 89) = 8.64, P <0.0001). Different letters (a, b, c) represent significant differences (P<0.05) between periods while similar letters denote no significance found (P>0.05). 76 • Period 1 • Period 2 • Period 3| 2 Week 3 Figure 3.4: The average frequency of dustbathing bouts in each week for each of the three periods of the daylight hours, combining treatment. The data represent un- transformed values for the average number of dustbathing bouts per period per week for four thirty-minute observation sessions ± SEM. (F(6,89) = 3.95, P = 0.0313). Different letters (a, b, c) represent significant differences (P<0.05) between periods while similar letters denote no significance found (P>0.05). 77 • Weekl 0 Week 2 • Week 3 H Week 4 Peat Moss Wire Fig.3.5: The average frequency of foraging bouts in each week for each treatment. The data represent un-transformed values for the average number of foraging bouts per week per treatment for two thirty-minute observation sessions ± SEM. (F(3,89) = 5.97, P = 0.0007). Different letters (a, b, c) represent significant differences (P<0.05) between weeks while similar letters denote no significance found (P>0.05). 78 20 18 >> 16 u 14 0) • Weekl: 12 o> 0 Week 2 10 D Week 3 8 H Week 4 6 f 4 2 0 Peat Moss Wire Fig.3.6: The average frequency of feeding bouts in each week for each treatment. The data represent un-transformed values for the average number of feeding bouts per week per treatment for two thirty-minute observation sessions ± SEM. (F(3> 89) = 4.85, P = 0.0029). Different letters (a, b, c) represent significant differences (P<0.05) between weeks while similar letters denote no significance found (P>0.05). 79 Table 3.1: The correlations between feather pecking compared to foraging and dustbathing for the three periods of daylight hours. The data in each box represent the correlation value (r), the corresponding P-Value and the sample size (n). The * symbol denotes a significant correlation. Foraging Feeding Dustbathing Periods 1 2 3 1 2 3 1 2 3 Feather 1 -0.22713* -0.36922* -0.16170* 0.3371* 0.4310* 0.3151* 0.01417 -0.01956 -0.11936 pecking 0.0498 0.0014 0.1779 0.0037 0.0002 0.0074 0.9059 0.8705 0.3215 72 72 72 72 72 72 72 72 72 2 -0.23582* -0.29257* -0.26184* 0.4152* 0.4393* 0.4355* -0.09565 0.11609 -0.04015 0.0461 0.0126 0.0274 0.0003 0.0001 0.0001 0.4242 0.3315 0.7396 72 72 72 72 72 72 72 72 72 3 -0.40681* -0.39975* -0.42214* 0.5187* 0.47024* 0.3884* -0.02139 -0.06851 0.09253 0.0004 0.0006 0.0002 O.0001 <0.0001 0.0008 0.8595 0.5703 0.4428 72 72 72 72 72 72 72 71 72 80 CHAPTER FOUR The effects of four types of enrichment on feather pecking behaviour in laying hens housed in barren environments ABSTRACT Feather pecking, an stereotypic behaviour in chickens, can be reduced by providing enrichment. However, there is little comparative information available on how effective different types of enrichment are. Providing forages to birds is likely to decrease feather pecking behaviour the most, as it is generally thought that feather pecking stems from re-directed foraging motivation. Yet, other types of enrichment, such as dustbaths and novel objects, have also been shown to reduce feather pecking. In order to develop a practical and effective enrichment, these different possibilities must be examined. Using a Latin Square Design, birds were given each of four treatments: 1) forages, 2) novel objects, 3) dustbaths or 4) no enrichment. The amount of feather pecking behaviour and the number of pecks to the enrichments were recorded. Results show feather pecking to be highest when no enrichment was present and lowest when the forages were present (P<0.05). However, the number of pecks birds gave to the forages and dustbaths were the same (P>0.05), suggesting they are similarly used. Thus, we suggest here that forage enrichments are most effective at alleviating feather pecking and 81 attempts should be made to develop poultry housing that allows for natural foraging behaviour. In lieu of this, providing any kind of enrichment will increase bird welfare and is therefore still beneficial. 1. INTRODUCTION Feather pecking, the pecking at or removal of feathers of one bird by another (e,g, Hoffmeyer, 1969), is an on-going welfare problem in the poultry industry, as removal of feathers is painful (Gentle and Hunter, 1990) and bleeding associated with feather loss can lead to cannibalism (Allen and Perry, 1975). There are two types of feather pecking, gentle and severe, however only severe feather pecking involves removal of feathers and potentially causes a decrease in bird welfare (McAdie and Keeling, 2002). The most common method used to manage feather pecking is to trim the beaks of the birds, making feather removal more difficult. However, the beak trimming procedure has been associated with acute and chronic pain in chickens and is itself a welfare concern (Gentle et ah, 1990). It is therefore important to build on previous research to better understand feather pecking and its underlying motivation in an attempt to develop more appropriate management practices (for a review of the various factors that affect feather pecking see Chapter 1 and Dixon, submitted). A number of studies have reported reduced feather pecking as a result of adding some type of enrichment to the pen or cage (e.g. Blokhuis, 1989; Huber-Eicher and Wechsler, 1998; Aerni et al, 2000; McAdie et ah, 2005). In this paper, the term 82 enrichment refers to an object added to the environment that would not usually be present under commercial housing conditions. Although there is now general agreement that foraging behaviour is redirected into feather pecks (Blokhuis, 1989; Harlander- Matauschek, et ah, 2007; Dixon et ah, in press) a number of studies have investigated the association between feather pecking and dustbathing behaviour (e.g. Vestergaard, 1994; Johnsen et ah, 1998). Many different types of substrates or combinations of substrates have been used in attempt to satisfy motivations that reduce feather pecking. Feather pecking was decreased by the provision of forages compared to wire housed birds, such as straw (El-Lethey et ah, 2000), polystyrene blocks (Wechsler and Huber-Eicher, 1998), and woodshavings (Blokhuis and Arkes, 1984). Substrates given as dustbaths, like sand (Huber-Eicher and Wechsler, 1997a) and peat moss (Norgaard-Nielsen et ah, 1993), also decreased the amount of feather pecking performed. Other items have been given to determine their effect on feather pecking: string enrichment devices were successful at decreasing feather pecking (McAdie et ah, 2005) and birds that used perches were less likely to become feather pecking victims (Wechsler and Huber-Eicher, 1998). The quality of the enrichment, or the amount of time and number of pecks given to the object, influenced its effectiveness. For example, long cut straw decreased feather pecking more than shredded straw or wood shavings, as did polystyrene blocks compared to polystyrene beads (Huber-Eicher and Wechsler, 1998). While it is agreed that enrichment does decrease feather pecking, there is no definite consensus as to what type of enrichment works best. Feather pecking is thought to stem from foraging motivation, but it is not clear if the provision of forages decreases 83 feather pecking more than the provision of dustbaths or the exploration of new objects. Previous studies observed various ages of birds that were kept and reared in different environments, making comparisons between studies unreliable. For example, Aerni et al. (2000) used birds reared in an aviary system and provided long cut straw in week 18, Norgaard-Nielsen (1997) reared birds on wire and provided sand from day two and McAdie et al. (2005) reared birds on litter and presented string enrichments at various days (on ages 1, 22, 52 days or never) and for different time periods (continuous or 4 hour periods from day 1). The most effective type of enrichment can be determined by comparing the provision of enrichments within a group of birds that have had similar rearing environments, experiences and enrichment schedules. Once this is known, a more practical model can be designed to improve the welfare of the billions of birds in the laying industry (estimated from Stats Canada, 2007). The objectives of this study were to assess different categories of enrichment by measuring their effect on feather pecking behaviour and comparing their relative use. It was hypothesized that feather pecking stems from a motivation to forage and thus predicted that forages will decrease feather pecking and be pecked at more than any other type of enrichment. Alternately, it was hypothesized that feather pecking stems from a motivation to dustbath and is predicted that dustbaths will decrease feather pecking and be pecked at more than any other type of enrichment. 84 2. METHODS 2.1. Animals and Housing Female, non beak trimmed White Leghorns were kept in groups of 4 or 5 and randomly distributed to forty-eight pens (230 birds total). Birds were obtained from another experiment at 14 weeks of age and had all experienced the same rearing and experimental conditions (see Chapter 5 for details). The birds were kept in wooden frame floor pens measuring 127cm x 107cm x 60cm with sides of chicken wire, 2.5cm x 2.5cm, and a wire mesh floor, 1cm x 1cm. Each pen had an automatic cup drinker and a metal 12" slide top feeder. The sides of each pen were covered with black cloth to prevent visual contact with birds from other pens. The chickens were fed ad lib. chicken grower diet until the end of the trial which was topped up every morning before observations began. The lights were set on a 10L:14D schedule and the temperature was 22°C . 2.2. Experimental Treatments The birds were observed with each of four enrichment categories. To avoid pseudo-replication, three exemplars of each enrichment category were used when possible, (see also Chapter 5) 85 1) Forages: Commercial wire bird-feeder suet holders (10cm x 10cm x 4cm) were each filled with one of the three types of substrate: i) peanut butter suet (peanut butter was added to suet at a lc suet to l/4c peanut butter ratio), ii) seeds in suet (seeds were a mix of shelled and unshelled sunflower seeds added in the same ratio as the peanut butter) and iii) cabbage leaves. Forages were able to have pieces pecked at and torn off but were not able to be dustbathed on. The forages were designed using nutritive substrates, as foraging is an integral part of feeding behaviour. However, the birds also had access to ad libitum feed for the duration of the study and it has been previously demonstrated that satiated birds direct a large number of pecks to a nutritive substance but only ingest a small amount (as measured by changes in weight of the substance) (Keeling and Hurnik, 1996). Thus, we would expect most pecks directed to the forages to be appetitive in nature and fewer to lead to consumption. Additionally, there seems to be no agreement in the literature as to whether 'foraging' has to be directed to a non-nutritive substance. For example, Andersson and colleagues (2001) only considered pecking at non-nutritive rocks to be foraging, while pecks to the chicken feed hidden underneath were thought to be feeding related. In contrast, Vaisanen and Jensen (2003) considered all pecks directed at commercial chicken feed to be foraging, even though the birds had been food-deprived. Based on these studies, we feel our choice of nutritive forages is justified, however, feeding motivation and potential feedback from ingestion of forages should be taken into account in interpretation of the data. 86 2) Novel Objects: Flat wooden blocks (10cm x 10cm x 4cm) were covered by one of the three materials: i) tin foil, ii) tissue paper or iii) felt to create three different exemplars of novelty. The birds had no experience with the material they were presented with. 3) Dustbaths: Dustbaths were presented in large pans (30cm x 15cm x 6cm), filled with one of three substrates: peat moss, white sand or grey sand. Dustbathing substrates were non-nutritive and could not have pieces torn off. 4) Control: No enrichment was added to the pen. The exemplars were chosen based on a previous study showing that pecks to the exemplars in each category are similar and implies that the motivations involved in pecking the exemplars in each category are the same (see Appendix A). The enrichments were presented to the birds in a Latin Square Design (Table: 4.1). The birds were given three habituation sessions to everything but the novel objects, thus the forages and dustbaths should not be elicit exploratory pecks, as the birds have had experience foraging and dustbathing with them previously. The enrichment was added to a pen thirty minutes before observations began to allow the birds to settle after being disturbed by the addition of the enrichment. This was done so the exemplars in the novel object category could be observed while they were still fairly new to the birds but after the birds had been allowed to settle from being disturbed by the addition of the object. Treatments were applied on alternate days (e.g. after the birds had one category 87 of enrichment, they got the next day off and the following day they got another enrichment category). All order combinations of the four enrichments were used as Latin Squares (24 in total) and each square was replicated (forty-eight in total) (Table: 4.2). The last enrichment presented was repeated after a day off. This was done to account for any carry-over effects: the potential lasting effect a treatment may have on the following treatment (Moore and McCabe, 1999). If carry-over effects are not a problem, the results for the last enrichment given should be the same both times it was presented. In this design, each enrichment was preceded an equal number of times by all other enrichments and potential carry-over effects were balanced for. 2.3. Behavioural Measures The birds were observed for thirty minute sessions on a pen basis during 730 and 1630 hours. The following measures were recorded: 1) Number of feather pecks: All occurrences of feather pecking were recorded. Feather pecking was defined as the pecking at of feathers of one bird by another. Due to the number of birds in each pen and conducting all observations live (as opposed to video recorded), it was not possible to accurately distinguish all feather pecks as gentle or severe. However, from a previous experiment in which these birds were used, all birds 88 2) had been observed performing both severe and gentle feather pecking. As severe feather pecking is a welfare concern (compared to gentle feather pecking), it is this behaviour that we hope to decrease. 3) Number of pecks to the enrichment: All occurrences of the pecks being directed to the enrichment - not measured in the control category. 2.4. Ethical note The use of all animals and methods were approved by the University of Guelph Animal Care Committee which adheres to Canadian Council on Animal Care guidelines. The birds used in this study were not beak trimmed, so to prevent potential cannibalism, the birds were inspected a minimum of three times a day (once at 800 hours, once at 1200 hours and once at 1600 hours). Any feather pecking injuries found were coated in pine tar to discourage further pecking and the bird was checked once an hour for the rest of the light period to ensure the injury did not get worse. If the pine tar was unsuccessful in discouraging pecking, the bird would be removed from the pen, however, this measure was not needed in this study. Similar to other chapters, the use of pine tar itself was rarely needed - throughout the study there were only three occurrences where pine tar was needed to discourage further pecking of a bloody injury. 89 2.5. Statistical A nalysis A mixed model variance component analyses was used to determine if there were any potential carry-over effects with carry-over effects nested in category of stimuli. A mixed model analysis was also used to determine the effect of the enrichment on feather pecking with exemplars of each enrichment nested in the category of enrichment and pen nested in replicate (SAS, v.8). Data was normalized using V(x +0.375) transformations and homogeneity of variance was achieved. Differences of the Least Square Means adjusted by Tukey were used to determine treatment differences. 3. RESULTS 3.1 Carry-Over Effects There were no carry-over effects from any of the enrichment stimuli on the number of feather pecks (F(4,8s) = 0.18, P = 0.9465) or pecks to the stimulus (F(3> 64) = 2.13, P = 0.105). 3.2 Feather Pecks There was a Pen nested in Replicate effect, with Pen Jb feather pecking more overall than Pen 2b (F(4,8s) = 1 -59, P = 0.0178). The rest of the pens were not significantly different from each other. 90 Frequency of feather pecking was highest in the treatment with no added enrichment and lowest in the forage enrichment. The dustbathing and novel object enrichments had similar levels of feather pecking (t(m) = -0.94, P = 0.7809) and were statistically different from and numerically in between the no enrichment and foraging categories (Treatment: F(3,177) = 165, P O.OOOl) (Fig:4.1). 3.3 Pecks to the Enrichment There were a similar number of pecks directed at the forages and dustbaths (t(H9) = -0.71, P = 0.7590) with fewer pecks directed at the novel objects (Treatment: F(2, 119) = 464.74, P O.0001 ) (Fig: 4.2). 4. DISCUSSION There were no significant carry-over effects observed for any of the stimuli used in the Latin Square design. Feather pecking levels were highest when no enrichment was given and, as predicted, lowest when given forages. This agrees with literature showing that when forages were provided, feather pecking levels were lower than in wire housed birds (e.g. Blokhuis, 1989; Huber-Eicher and Wechsler, 1998; Areni et al, 2000; Nicol et ah, 2001) and comparisons of diurnal rhythms also found feather pecking to be similar to foraging (Chapter 3). One way to study the underlying motivations of stereotypic behaviour is to manipulate aspects of the external environment, such as the addition of 91 enrichment, and determine the effects on the stereotypic behaviour (e.g. Mason et al, 2007; Chapter 6). As forages were most effective at decreasing feather pecking, while appearing to be just as attractive as dustbaths, this adds evidence to the hypothesis that feather pecking stems from re-directed foraging motivation. As mentioned previously, there is the possibility that feedback from ingestion of forages may have influenced the results. However, similar to Keeling and Hurnik (1996) there did not seem to be much ingestion of the forages. The suet bird feeders were filled with forages prior to be presented to the birds, and when removed they were still basically full. While there is some evidence that feather pecking may be a method to increase fibre in the diet and improve gut function (e.g. Hetland et al, 2003; Harlander-Matauschek et al, 2006a; 2006b; see Chapter 1), there is as yet no evidence of a relationship between feather pecking and high fat/high calorie diets. Thus, post-ingestion feedback from the forages used in this study is unlikely to be a confounding factor in the results. Providing the birds a dustbath or novel object to peck also decreased feather pecking when compared to birds living in barren environments. Similarly, previous research has assessed the provision of enrichment, such as sand (Norgaard-Nielsen, 1997), novel environments (Chow and Hogan, 2005) or string (McAdie et al, 2005), and found a decrease in the amount of feather pecking compared to those not given enrichment. Commercial laying hens are kept in barren conditions and it is possible that any addition to their environment may have positive consequences. This may explain why re-directed dustbathing was thought to motivate feather pecking (e.g Vestergaard, 1994) or why exploration has been thought to be related to feather pecking levels (e.g. 92 Riedstra and Groothuis, 2002). However, the consistent result of a decrease in feather pecking may be due to the birds having more to occupy their time, leaving less time available for feather pecking, rather than due to the properties of the enrichment or underlying motivation. Future work measuring the durations of the time spent feather pecking and performing other behaviour patterns, with and without enrichment; and determining how motivated birds are to gain access to these enrichments would help to clarify this issue (see Chapter 7). The number of pecks given to the forages and dustbaths were not significantly different. This is interesting as feather pecking levels were lower when given forages as compared to dustbaths, even though both enrichments received similar attention from the birds. This suggests that it was not just the opportunity for increased activity that resulted in decreased levels of feather pecking but that the opportunity to forage decreased motivation to feather peck, potentially through fulfillment of foraging motivation. This also adds increased support for the feather pecking as re-directed foraging hypothesis compared to the re-directed dustbathing hypothesis. Unfortunately gentle and severe feather pecks were not able to be reliably distinguished during the live observations in this trial due to simultaneous observation of the pecking behaviour of all birds in a pen (all-occurrence sampling), as the increased feather pecking levels when given access to dustbaths may have been gentle feather pecking at birds with dusty plumage. It is known that the relative amount of severe feather pecking increases as birds age (Huber-Eicher and Sebo, 2001), so we would still expect a fair percentage of the overall pecks given with the dustbathing enrichment to be severe. The amount of feather pecking shown with 93 the novel object enrichment was also similar to the amount given when provided a dustbath. As none of the novel objects were dusty or left particles on the feathers of the birds, it seems unlikely that the higher levels of feather pecking in these categories of enrichment compared to the provision of forages was due to gentle feather pecking at particles on the plumage alone. One last issue that needs to be addressed is the inability of any of the enrichments to abolish feather pecking altogether. There are a few potential reasons as to why this may occur: as a behaviour pattern is repeatedly performed, neural control of the behaviour may shift into a type of automatic processing referred to as central control or habit formation (e.g. e.g. Dantzer, 1986; Mason, 2006). Alternately, the feather pecking may continue because the chickens themselves have been altered in some profound way, potentially through permanent changes occurring in Central Nervous System (CNS) functioning and/or behavioural control processes that inhibit appropriate responses to stimuli and eventually perseveration (the repetition of a behaviour pattern without the appropriate stimulus present) may occur (e.g. e.g. Sandson and Albert, 1984; Garner, 2006; see Chapter 7). These explanations may also help to explain why feather pecking can be found in highly enriched or free range environments where the birds have a number of different behavioural opportunities available to them. 94 5. CONCLUSIONS The hypothesis that feather pecking stems from re-directed foraging behaviour is supported as feather pecking levels were lowest when provided foraging substrates. This confirms the importance of providing birds with forages and encourages future research to develop an inexpensive, easy to use substrate that still decreases feather pecking. For example, hay or straw could be provided to poultry kept in free-run or floor systems while a modified version of a hanging straw bag that is available for horses could be hung on the sides of cages for poultry housed in these systems. However, these foraging opportunities should be provided before the behaviour becomes habit-like or perseverative to be most effective. This research also indicates that providing any of the enrichments used in this experiment, not just forages, would benefit laying hens, as all enrichments reduced feather pecking behaviour and thus improved bird welfare. 95 • No Stiirulus • Forage • Dustbath 0 Novel Figure 4.1: The average amount of feather pecking performed when given various stimuli. The data represents the un-transformed average amount of feather pecking per pen ± SEM during thirty minute observation sessions (F^ 177) = 165, P <0.0001, power = 0.999, a = 0.05). Different letters (a, b, c) denote significant differences (P<0.05), while the same letter denotes no significant difference (P>0.05). 96 tJ - a a 40 T 35 'l^^l v^ 30 O D Forage Q_ 25 ^H • Dustbath 'o 20 HHHHHH 0 Novel 15 HH^^R^^^ 10 ^^H^R^^^ 5- H^BII^^ Figure 4.2: The number of pecks directed to the stimuli. The data represent the un- transformed average number of pecks per pen ± SEM during thirty minute observation sessions. (F(2,119) = 464.74, P O.OOOl, power = 0.999, a = 0.05). Different letters (a, b, c) denote significant differences (P<0.05). 97 Table 4.1: Example of Latin Square Design used Day Pen 1 3 5 7 9 - Last stimulus repeated 4 Control Forage Dustbath Novel Novel 15 Novel Control Forage Dustbath Dustbath 8 Dustbath Novel Control Forage Forage 24 Forage Dustbath Novel Control Control 98 Table 4.2: All combinations of the four enrichments were used as Latin Squares and each Latin Square combination was replicated. The last stimulus presented was repeated to account for potential carry-over effects (not shown in table). D = dustbath; F = forage; Nov = novel object; C = Control D-F-Nov-C F-D-Nov-C Nov-F-D-C C-F-D-Nov D-F-C-Nov F-D-C-Nov Nov-F-C-D C-F-Nov-D D-Nov-F-C F-Nov-D-C Nov-D-F-C C-D-F-Nov D-Nov-C-F F-Nov-C-D Nov-D-C-F C-D-Nov-F D-C-F-Nov F-C-D-Nov Nov-C-F-D C-Nov-F-D D-C-Nov-F F-C-Nov-D Nov-C-D-F C-Nov-D-F 99 Preface to Chapter Five At this point in the thesis, it had been shown that current environment had the most effect on feather pecking behaviour, thus there is no need to ensure certain early experiences (either positive or negative) to obtain accurate feather pecking results. However, the motivation behind feather pecking still needed to be determined. As the two main causes of feather pecking were hypothesized to be redirected foraging or dustbathing, which both involve pecking and scratching at the ground, gross observations of these behaviour patterns would not clarify this issue, though the differential effects of substrate on foraging and feather pecking in parallel (Chapter 2), the diurnal rhythms of feather pecking, foraging and dustbathing (Chapter 3) and the provision of different types of enrichment (Chapter 4) supported the foraging hypothesis. Thus, a new method had to be developed. 100 CHAPTER FIVE What's in a peck? Using Fixed Action Pattern morphology to identify the motivational basis of the abnormal behaviour of feather pecking (A version of this chapter has been accepted for publication in Animal Behaviour) ABSTRACT Like many captive animals, hens {Gallus gallus) used for agricultural production perform abnormal behaviour patterns. They are particularly prone to stereotypic feather pecking, the severest form of which involves the pecking at and removal of feathers, which can cause bleeding and even stimulate cannibalism. The two main hypothesized explanations for feather pecking concern re-directed motivation to forage, or alternately re-directed motivation to dustbathe, leading to redirected behaviour in the form of pecks to plumage. Previous work on pigeons has shown that the detailed morphology of pecks involved in drinking and feeding, or in working for food or water, involves motivationally distinct 'Fixed Action Patterns' (FAPs). We therefore used similar methods to these FAP studies to conduct a preliminary trial to determine 1) if pecks given by chickens have Fixed Action Patterns and 2) if different types of pecks (e.g. pecks to forages, pecks to feathered models) have different motor patterns involved. It was found that pecks to the stimuli could be measured in fine detail and appeared to be FAPs as 101 within and between bird variation was quite low. Also, the motor patterns involved in different types of pecks, such as pecks to the head of the feathered model and pecks to the forages, were different for most measures (P<0.05), however, measures were similar for severe feather pecks and pecks to the forages ( P>0.05). At this point, we used the same techniques to quantify the motor patterns involved in foraging and in dustbathing pecks, for comparison to feather-pecking. Sixty chickens were video recorded while pecking at a variety of stimuli, including forages and dustbaths and the durations of different aspects of the pecks were recorded. Mixed models assessed whether the motivation underlying a peck affected its morphology; and whether severe feather pecks resembled or differed from either dustbath or foraging pecks (or other types of pecks). The motor patterns involved in pecks to forages, dustbaths, novel objects and water all varied significantly; importantly, the motor patterns involved in pecking during dustbathing and foraging differed (P<0.0001 for all measures). Severe feather pecks proved similar to foraging pecks (P>0.05; power >0.95), but different from all other pecks including dustbathing (P<0.0001 for all measures). These results indicate that severe feather pecking derives from re-directed motivation to forage, not to dustbathe. More broadly, they suggest that finely analyzing 'Fixed Action Pattern' morphology can help elucidate the motivational bases of puzzling stereotypic behaviours in captive animals. 102 1. INTRODUCTION 1.1. Abnormal Behaviour When animals are kept in restrictive conditions, they often show abnormal behaviour, with estimates of over 85 million farm, lab and zoo animals displaying these behaviour patterns (Mason and Latham, 2004). Behaviour is generally classified as abnormal if it is not seen in a population deemed normal (e.g. in the wild) or if it is caused by a known underlying pathology (e.g. brain dysfunction) or if it causes pathology (e.g. skin lesions) (Mason, 1991; Garner, 2005). Abnormal behaviour often takes the form of stereotypic movements which are repetitive and fixed in form (Fox, 1968), may be induced by frustration, repeated attempts to cope and/or brain dysfunction (Mason, 2006), and may indicate poor welfare as stereotypic behaviour patterns are prevalently found in smaller, barren environments that cause other signs of stress (Mason, 1991; Mason and Latham, 2004). The performance of stereotypic behaviour in captive conditions often results in reduced welfare or indicates reduced welfare, and as a result, there is great interest in determining their causal factors. Once this is done, the conditions the animals are kept in can be changed, the incidence of stereotypic behaviours reduced or eliminated, and the welfare of the animals improved. As animal use continues to grow, understanding an animal's behaviour becomes extremely important in improving the human-animal relationship (Timberlake, 1997). However, there are numerous examples of stereotypic 103 behaviour patterns found in captive animals whose underlying motivations are still uncertain: e.g. belly nosing in piglets (e.g. Widowski et ah, 2008); barbering in rodents (e.g.Garner et al. 2004); crib-biting in horses (e.g.Bachmann et al., 2003); pacing in carnivores (Clubb and Vickery, 2006) and feather pecking in laying hens (e.g. Harlander- Matauschek et al, 2006a). Our goal was to use a novel approach to help determine the motivational bases of stereotypic behaviour, using feather pecking in laying hens as a model. 1.2. Feather Pecking in Laying Hens Feather pecking is an extensive problem in the poultry industry, with 77% of surveyed commercial poultry operations reporting feather pecking behaviour (Huber- Eicher and Sebo, 2001); this likely affects a large number of individuals since approximately five billion laying hens are used for production worldwide each year (estimated from Stats Canada, 2007). Feather pecking involves pecking at and possibly pulling out of the feathers of one bird by another (Hoffmeyer, 1969). This behaviour can be divided into two categories: gentle feather pecks, where the feather is nibbled and pecked at but not pulled out, and severe feather pecks, where the feather is vigorously pulled or removed (McAdie and Keeling, 2002). Severe feather pecking is considered a welfare problem for poultry, since feather removal is painful for the recipient (Gentle and Hunter, 1990) and blood associated with feather loss can lead to cannibalism (Hughes and Duncan, 1972). 104 Previous hypotheses propose that feather pecking stems from either re-directed dustbathing or re-directed foraging motivation. Both hypotheses suggest, in general terms, that the lack of particulate matter on the floor (in which to dustbathe or forage) leads to the development of feather pecking in these birds who are often housed in cages with wire mesh floors (Blokhuis, 1986; 1989; Vestergaard and Lisborg, 1993); however, different strains of birds do perform different levels of feather pecking and there are even individual differences in the rates of feather pecking within one strain of birds, indicating differences in the levels of the underlying motivation of feather pecking (Kjaer et al, 2001; see Chapter 1 for further discussion). In addition, this stereotypic behaviour can persist even after the birds are moved to more enriched conditions due to habit formation or perseveration of the behaviour (see Chapters 1 and 6 for further discussions). Both foraging and dustbathing behaviour have a ground-pecking component. If there is no natural substrate on the ground to peck at, it is thus hypothesized that the feathers of cagemates are used as a substitute (Blokhuis, 1986; Johnsen and Vestergaard, 1996). However, the evidence for either motivational hypothesis is not entirely conclusive as past studies do not differentiate between dustbathing ground pecks and foraging ground pecks (e.g. Huber-Eicher and Wechsler, 1997; Savory and Mann, 1999; Blicik and Keeling, 2000) and instead make assumptions regarding the motivation underlying the pecking behaviour directed to naturalistic flooring substrates. Our aim is to resolve this issue by exploring the potential use of Fixed Action Patterns to unravel the motivational bases of feather pecking in laying hens. If useful, this technique could then be applied to other stereotypic behaviour in other species. 105 1.3. Fixed Action Patterns Fixed Action Patterns (FAP) are unlearned (innate), repetitive movements in response to external stimuli that are relatively unaffected by feedback, influenced by underlying motivation, and species-specific (Lorenz and Tinbergen, 1938). They are sometimes referred to as Modal Action Patterns, a term which takes into account variations both within and between individuals in performance of the behaviour by referring to the most frequent measurement value (mode) shown in the distribution of the behaviour pattern (Barlow, 1968). One well-known example of a FAP is egg rolling in the Greylag goose: if an egg rolls out of the nest, the goose uses specific, predictable head and neck movements to roll the egg back into it, and will continue these rolling motions even if the egg stimulus is removed (Lorenz and Tinbergen, 1938). More recently studied examples include the head bob displays of Anole lizards - which are expressed in adult form right from hatching, regardless of whether animals are raised in the field or laboratory (Lovern and Jenssen, 2003); and grooming movements by mice - which contain certain repetitions of head, arm and paw movements that have been categorized into four distinct phases of grooming (Berridge, 1990), and which once started, will continue, even if feedback from physical contact with the body is prevented (Golani and Fentress, 1985). The motivation behind a behaviour pattern will influence precisely how the response to a stimulus is performed (Breland and Breland, 1968). Thus, in pigeons, the 106 stereotyped pecks involved in eating and drinking differ because motivation affects FAP morphology (Zweers, 1982; 1992). This was further demonstrated in an autoshaping experiment, autoshaping being when the shape or movements directed at a predictive stimulus, such as a button or light, resemble those that are elicited by the reward. For example, rats direct social behaviour to another rat whose presence predicts a future food reward, as rats include social aspects in their feeding behaviour (Timberlake, 1975). Thus, pigeons conditioned to peck a key for a food or water reward will also show 'eating' and 'drinking' pecks to the key, respectively, further demonstrating that the motivation behind each movement affects the motor patterns expressed (Jenkins and Moore, 1973). Based on these characteristics, we planned to use FAPs in an attempt to uncover the motivation behind stereotypic feather pecking. The objectives of this preliminary experiment were: 1) to measure aspects of the morphology involved in feather pecks and other types of pecks to determine if these are FAPs based on consistent movements with low within and between bird variability. The degree of variability of the physical components that make up the behaviour or the different elements of the behaviour (e.g. duration, speed, position) is considered the most important identification tool of FAPs (Schleidt, 1974). 2) to measure the motor patterns involved in differently motivated types of pecks to determine if these different pecks reliably have different motor patterns (i.e. FAP pecks with different motivations should consistently look different). It was hypothesized that chickens' pecks would involve FAPs and predicted that pecks to each type of stimulus would be consistent and show little variation within and between birds. 107 After this had successfully been done, another experiment was conducted, quantifying aspects of pecking morphology during feather pecking, dustbath pecking and foraging pecking. The aims of this study were to compare the motor patterns of the different types of feather pecks, as well as the motor patterns of feather pecking pecks to dustbathing and foraging pecks. If severe feather pecking stems from re-directed foraging motivation then feather pecks should resemble foraging pecks, and nothing else; however, if severe feather pecking stems from re-directed dustbathing motivation, then feather pecks should resemble dustbathing pecks and nothing else. Finally if gentle and severe feather pecking stem from different motivations, then gentle feather pecking pecks should be different in morphology from severe feather pecking pecks. 2. EXPERIMENT 1: PRELIMINARY TRIAL 2.1 METHODS 2.1.1. Animals and housing All chicks were non beak trimmed, one-day old female ISA White Leghorns, received from a commercial hatchery (Bonnie's Hatchery, Elmira, Ont); this strain is known to exhibit feather pecking. The chicks, housed 10 to a pen, were randomly distributed between two identical rooms, with 12 pens in each room (i.e. 24 total) at the Arkell Poultry Research Facility (University of Guelph). For these experiments, birds 108 that performed feather pecking were needed. To ensure this and identify subjects, birds were kept in a restrictive environment typical of commercial conditions with no access to a dustbath or forages (cf. e.g. El-Lethey et al, 2000; van de Weerd and Elson, 2006). The pens were wooden framed and measured 127cm x 107cm x 60cm with wire grid (2.5cm x 2.5cm) sides and a wire grid (1cm x 1cm) floor. Each pen had two automatic drinking cups and two metal 12" slide top feeders. The chicks were fed ad libitum Chick Starter Fine Crumbles, until week 6 of life and Chicken Grower diet from week 6 thereafter. The lights were kept on a 23L: ID schedule for the first 3 days, then switched to 10L:14D for the remainder of the trial. Temperature at chick level started at 33°C and was gradually reduced to 20°C by the end of 20 weeks. After four weeks, all birds were wing-tagged with coloured tags for individual identification. At one end of each pen, a triangular test area was created (35cm x 35cm x 24cm) using a sheet of plastic board covered with black cloth to prevent visual contact with other birds (Fig. 5.1). 2.2.1. Home Pen Observations During week five of life, birds were observed continuously for three 30-minute periods (with at least one period between 700 and 1200 hours and at least one period between 1200 and 1700 hours). All incidences of feather pecking and the identity of the bird performing the behaviour were recorded. The ten birds that performed the most feather pecking were selected to be Test Birds for the next part of the experiment. 109 2.2.2. Test Pen Observations At the end of week five, the Test Birds were given three 20-minute habituation sessions to the test area. After this, all birds were given all peck stimuli on weeks 6, 8, and 10 of life. 2.2.3. Peck Stimuli 1) Feathered Model: Birds of similar age to the Test Birds were euthanized with CO2, placed in a standing position and frozen to maintain this posture (as per Millman and Duncan, 2000). These models were presented to the Test Birds with either i) their backs facing outward where feather pecks could be directed or ii) with their heads exposed, where pecks to the beaks and combs could be directed. The type of feather peck, gentle or severe, given to the back of the model was recorded. The Feathered Model was presented with its head exposed as a control to ensure that all pecks did not look the same. If peck morphology does in fact differ among motivational systems, then the pecks to the head of the Feathered Model should be different than those directed to feathers or forages, as pecks to the head area (beak, comb, wattles) are not part of feather pecking nor are they a part of foraging behaviour, thus the motivation behind these pecks should be different and the motor patterns involved should be different. 110 2) Forages: Two materials were presented, grass and crumpled tissue paper, in the form of a square 30cm2 in area. The criteria for the forages were that they could be pecked and scratched at, have pieces torn off and be consumed. Tissue paper was used as a comparison to the grass to determine if foraging pecks are consistent, or if they change depending on substrate. Ideally, all stimuli would be presented to the test birds so that the area being pecked was approximately at the same height above the cage floor. However, the head and the back of the Feathered Model could not be presented at the same height (as the model was designed to stand on the cage floor) but the forages were presented attached to the side of the pen so that they spanned the difference between these two heights, and thus it was never the case that all stimuli but one were presented at a certain height above the cage floor. In addition, it has been shown that pigeons reliably orientate their head at certain distances from a target during two head fixations, the first is where the decision to peck the target is made and the second allows the bird to calculate the size and depth of the target (Goodale, 1983) and if chickens behave similarly, the height from the cage floor the stimulus was presented at will not influence the measures taken. 2.2.4. Peck Measures Measures were chosen from previous studies that quantified various aspects of pecking (Goodale, 1983; Dawkins, 1995). Ill 1) Vertical head angle (degrees): measured as the angle between a virtual line through the bird's eye down the middle of the beak to the vertical (Fig.5.2). 2) Horizontal head angle (degrees): the angle made by a virtual line through the middle of the bird's head, down the middle of the beak to the horizontal (Fig. 5.3). 3) Duration (sec) of head fixation: the length of time that the head is kept still before the peck (Fig. 5.4a). 4) Contact to stimulus: duration (sec) from the end of head fixation to beak contact with the stimulus (Fig. 5.4b). 5) Duration (sec) of the whole peck: time from no head movement, through the peck, back to no head movement (Fig. 5.4c). Sessions where the birds were exposed to the stimuli were video recorded in real time at 60fps with a Panasonic CCD camera (WV-BL90A), monitor (WV-CM110A) and time lapse VCR (TL 950). Recording at this high speed enabled us to play back pecks frame by frame and determine accurate durations of the peck morphology with little variation within measures. Each Test Week the peck stimuli were each presented twice to the Test Bird for 30-minute sessions in a randomly determined order. The sessions were video recorded 112 from the side in order to view the vertical head angle and from above to view the horizontal head angle. Head angles were measured off the monitor and durations were calculated. The types of feather pecks, gentle or severe, directed to the back of the Feathered Model were also noted. 2.2.6. Ethical Note The use of all animals and methods in the both Experiment 1 and 2 were approved by the University of Guelph Animal Care Committee which adheres to Canadian Council on Animal Care guidelines. The birds used in this study were not beak trimmed, so to prevent potential cannibalism, the birds were inspected a minimum of three times a day (once at 800 hours, once at 1200 hours and once at 1600 hours). Any feather pecking injuries found were coated in pine tar to discourage further pecking and the bird was checked once an hour for the rest of the light period to ensure the injury did not get worse. If the pine tar was unsuccessful in discouraging pecking, the bird would be removed from the pen, however, this measure was not needed in this study. 2.2.7. Statistical Analysis A mixed model variance component analysis function was used with Bird as the subject and Test Week as the repeated measure (SAS, v.8) to determine the effect of peck 113 stimuli on peck measurements. Transformations were made with the natural log function when necessary. Significant results were further examined using the LS Mean function adjusted by Tukey and Pearson correlations were performed to determine potential relationships between measures (e.g. those to the two types of forages). 2.3. RESULTS 2.3.1. Motor patterns of different types of pecks All five peck measures were successfully recorded for each stimulus and each bird, and each had low inter- and between bird variability (max. 1 degree and 0.018 sec). In addition, for all measures recorded pecks to the Tissue Paper and Grass were similar but different from pecks to the head of the Feathered Model (Table 5.1). 2.3.2. Motor patterns of gentle and severe feather pecking The motor patterns involved in both severe and gentle feather pecking directed to the back of the feathered model were successfully recorded for each measure and had low inter- and between bird variability (again max. 1 degree and 0.018 sec). There were no differences between gentle and severe feather pecking for vertical head angle, horizontal head angle or duration to contact stimuli from head fixation. However, severe feather pecks did have a longer overall duration and have longer head fixations before the pecks than gentle feather pecks (Table 5.2). 114 2.3.3. Correlations For vertical head angle, horizontal head angle and duration of the peck, pecks that were similarly motivated (i.e. peck to the Tissue Paper and Grass forages) were highly correlated (r = 0.8, P O.OOOl; r = 0.56, P = 0.0493 and r = 0.62, P = 0.0008, respectively), that is, as the measures for one increased, the measures for the other acted similarly. However, there were no correlations between differently motivated pecks (i.e. pecks to the foraging substrates and pecks to the head of the Feathered Model) (P>0.1 for all). This demonstrates that similarly motivated pecks (i.e. foraging motivated) have similar movement patterns and this is not affected by the type of substrate. 2.4. DISCUSSION For the preliminary study, the prediction that the motor patterns involved in each kind of peck (e.g. foraging, feather pecking) would be consistent was met for all measures: within and between bird variation for each peck type was small (max. 1 degree and 0.018 sec). These results were also consistent with the previous pigeon literature; differently motivated pecks (e.g. foraging pecks and pecks to the head of the Feathered Model) had different motor patterns involved while similarly motivated pecks had similar motor patterns (e.g. pecks to the two types of forages). Pecks to the back of the Feathered Model and foraging pecks were not significantly different; however, the power of this experiment is too low to make any conclusions about this result. As mentioned above, low variability within a motor pattern is a key indicator of FAPs (Schleidt, 1974), 115 thus the motor patterns involved in chicken pecks appear to involve FAPs and the motor patterns are different depending on the underlying motivation. 3. EXPERIMENT 2 3.1. METHODS 3.1.1. Stimulus design When choosing and designing appropriate stimuli to be pecked, we wanted to ensure that any potential differences in peck morphology were due to underlying motivation and not any other aspect of the stimulus, such as its shape or material. We also wanted to address the potential problem of pseudoreplication: if only single exemplars from a type of stimulus are used to test hypotheses about the type itself, it is not clear if results found reflect that particular exemplar only or the broader type of stimulus (e.g. Kroodsma et al., 2001). For example, in one study of the responses of voles and mice to owl calls, a single tawny owl call was recorded and played back to test behavioural and hormonal reaction to a perceived predator (Eilam et al., 1999). However, as Kroodsma et al. (2001) points out, only one tawny owl called was used, making it unclear if effects were due to owls in general or that particular one tawny owl recording. Thus, we used two shapes and three materials for each stimulus type, wherever possible. 116 1) Feathered Model a) Chicken-Shaped Feathered Model: as per Preliminary b) Flat Feathered Model: Chickens were euthanized and their skin removed with feathers attached. This "feather-skin" was then attached to a wooden block (10cm x 10cm x 4cm). The birds used for the feathered models were approximately the same age as the test birds and were the same breed of chicken, obtained from the same breeder. Pecks were classified as gentle when the feather was nibbled at and severe when the feather was vigorously pecked and may be removed (cf. e.g. McAdie and Keeling, 2002). 2) Forages a) Chicken-Shaped Forage: A toy American football (30cm x 10cm x 10cm) was chosen for its resemblance in size and shape to the back and tail area of a chicken. Three types of forage were applied to create three different sub-types of stimulus: i) peanut butter suet (peanut butter was added to suet at a lc suet to l/4c peanut butter ratio), ii) seeds in suet (seeds were a mix of shelled and unshelled sunflower seeds added in the same ratio as the peanut butter) and iii) cabbage leaves. The suet was evenly spread over the football, while the leaves were glued in place with non-toxic glue. b) Flat Forage: Commercial bird-feeder suet holders (10cm x 10cm x 4cm) were each filled with one of the three types of substrate as used for the Chicken-Shaped Forage. 117 3) Novel Object a) Chicken-Shaped Novel Object: Footballs identical to those used in the Chicken- Shaped Forage were covered with one of three different materials: i) tin foil, ii) tissue paper or iii) felt, to create three different exemplars of novelty. b) Flat Novel Object: Flat wooden blocks of the same size as used for the Flat Feathered Model were covered by one of the three materials used for the Chicken-Shaped Novel Object. 4) Dustbath: Dustbaths were presented in large pans (30cm x 15cm x 6cm), filled with one of three substrates: peat moss, white sand or grey sand. It was not possible to create a Chicken-Shaped version of this stimulus. 5) Water: A clear dish of water (10cm x 10cm x 4cm) was provided, so the beak was still visible through the side when the bird was drinking. Again, it was not possible to create a Chicken-Shaped version of this stimulus. Each of the test birds was video recorded pecking at one exemplar (sub-type) of each of the five stimulus types. All birds were recorded pecking at both the Chicken- Shaped and Flat varieties of each sub-type. As with the preliminary experiment, all stimuli were presented on the pen floor and so that the surface to be pecked was approximately at the same height from the ground. With the 'Flat' versions of the stimuli, this was easily achieved and all the 118 'Chicken-Shaped' stimuli were also approximately the same height from the ground. However, it was not possible to present the 'Flat' and 'Chicken-Shaped' varieties at the same height from the ground due to the curved shape of the 'Chicken-Shaped' stimuli. As mentioned above, pigeons reliably orientate their head at certain distances from a target during two head fixations (Goodale, 1983) and results from the Preliminary Trial demonstrate that chickens consistently perform the second head fixation, thus stimuli height from the cage floor should be compensated for by the chickens as they orientate themselves to the stimulus. In addition, the two shapes of stimuli were statistically compared to determine if their differences had an influence on the peck measurements. 3.1.2. Selection of subjects As per Preliminary 3.1.3. Data Collection The birds selected as Test subjects were given three 20 minute habituation sessions to the test area over a 4 day period. After habituation, the birds were given each of the five categories of stimuli to peck at in the Test Area. Each stimulus was presented once to each bird for thirty minutes, with the order of presentation randomized, and the session video recorded. From the video recording of each stimulus type for each bird, the peck measures for 20 pecks to the stimulus were recorded. The test birds housed together (2 or 3 test birds per pen) got the same sub-type of forage, dustbath and novelty and this sub- 119 type was the same for both the Chicken-Shaped and Flat exemplars. During trials dustbathing pecks were only measured after a dustbathing bout had begun - i.e. the bird was squatting in the dust and kicking dust onto its feathers (Hoffmeyer, 1969). Sessions where the birds were exposed to the stimuli were video recorded at the same speed and with the same equipment as in the Preliminary Study. 3.1.4. Peck measures The measures of peck morphology used in this experiment were based on the measures in the Preliminary study which had the lowest variations. 1) Duration of head fixation 2) Duration from fixation to contact 3) Duration of the whole peck 3.1.5. Choice of 'N' Using Preliminary Trial Data Data from the preliminary study (see Tables 5.1 and 5.2) were used to calculate the sample size with a power high enough to make Type II errors unlikely. Increasing the power of a test increases the probability of correctly finding no significant difference 120 between treatments (cf. e.g. Moore and McCabe 1999). We aimed for a power of 95%, giving us only a 5% chance of mistakenly finding no difference where one actually exists. The results from the preliminary study were used to determine the smallest differences between the means of pecks that were not significantly (nor biologically) different from each other (pecks to the two types of Forages) (P>0.05) and pecks that were significantly different (pecks to the back and head of the Feathered Model) (P<0.05) and these differences were used to calculate sample sizes using the techniques of Berndtson (1991). From these preliminary/pilot data, the smallest difference between means was calculated to be 7% (compared to differences between the means of 10-30% in the pigeon literature, e.g. Goodale, 1983) and a sample size of 60 birds was needed for the next part of this experiment to have a power of 95%, thus making the potential for Type II errors unlikely. These 60 subjects were chosen from the pool of feather peckers. 3.1.6. Ethical note As per Preliminary 3.1.7. Statistical analyses A mixed model variance component analysis blocked by Sub-Type of stimulus, with Pen nested in Sub-Type, Pen as the random factor, and Bird Nested in Pen and Sub- 121 Type as the subject, was used to determine the effect of peck stimuli on peck measurements (SAS, v.8). Data were normalized using Natural Log transformations where necessary, and treatment differences were determined using Least Square Means adjusted by Tukey. 3.2. RESULTS 3.2.1. Chicken-Shaped vs. Flat stimuli There were no significant differences between the Chicken-Shaped and Flat versions of any stimulus for any of the peck motor pattern measures (see Table 5.3 and Fig. 5.5). Because of this, we could safely assume that the properties of dustbathing, feather pecking and other pecks were not mere products of the stimulus shape involved but really did reflect motivational state. The same held for the different materials/sub types used (e.g. the different substrates offered as dustbaths, etc.) - see below. 3.2.2. Foraging vs. dustbathing and other types of pecks Pecks to the dustbaths, forages, novel objects and water all significantly differed from each other in terms of peck duration (F(4> 129) = 19, 219.8, P <0.0001). In contrast, the sub-types of each dustbath, forage or novel object involved did not differ (Sub-Type: Dur: F(6, ]08) = 2.00, P = 0.0916; Contact: F(6,108) = 1.92, P = 0.085; HF: F(6> ,08) = 0.58, P = 0.7495) (see Fig.5.5). Thus the motivation behind the peck - not the exact sub-type of 122 substrate involved - significantly affected at least this aspect of peck morphology. Somewhat similar findings emerged when comparing foraging, dustbathing and other types of pecks for the other two measures of peck morphology, but values for contact duration were not different between dustbathing and novel object pecks (dustbathing vs. novel object: t(128) = 2.37, P = 0.1305; Peck Stimuli: F(4,129) = 612.24, PO.0001), while values for head fixation were not different between foraging and novel object pecks (forage vs. novel object: t(]28) = 0.38, P = 0.9955; Peck Stimuli: F(4> 129) = 106.97, P<0.0001). However, even for these measures, where there were significant differences between motivationally distinct types of stimulus: the subtype (e.g. forage sub-type, dustbath sub-type, etc.) had no effect (see Table 5.4 and 5.5 for means ± SE). 3.2.3. Feather pecks: gentle vs. severe Gentle and severe feather pecks had different morphologies. Severe feather pecks took longer to perform (F(i; 119) = 757.25, P <0.0001), had a longer time to contact the stimulus (F(i;n9) = 52.21, P <0.0001) and involved a longer head fixation (FO, 119) = 12.99, P = 0.0005) than gentle feather pecks (see Tables 5.4 and 5.5 for means ± SE). 3.2.4. Severe feather pecks Severe feather pecks were similar to foraging pecks, but differed from all other peck types in duration of the whole peck and time to contact the stimuli (severe feather pecks vs. foraging: t(i28) = -0.18, P = 0.9997). However, for the duration of head fixation, 123 severe feather pecks differed only from dustbathing and drinking pecks, with severe feather pecking taking more time and less time, respectively (severe feather pecks vs. foraging: t(128) = -0.31, P = 0.9979; Peck Stimuli: F(4,129) = 106.97, PO.0001) (Table 5.4). 3.2.5. Gentle feather pecks Gentle feather pecks were different from pecks to the other stimuli for the duration of the whole peck and time to contact stimuli. However, for the duration of head fixation, gentle feather pecks were similar to pecks to the forages (t(B8) = 1.15, P = 0.7819) and novel objects (t(!38) = -0.89, P = 0.9017) (Table 5.5). 3.4. DISCUSSION Previous literature had shown that pigeons will give 'feeding' or 'drinking' pecks to buttons they have been trained to peck for a food or water reward, respectively (Jenkins and Moore, 1973), thus showing that motivation affects peck morphology, independent of the target being pecked. Our results for poultry were consistent with this. There were no differences in the peck motor patterns directed to the chicken-shaped and flat versions of each stimulus; nor were there differences in the peck motor patterns directed to the different sub-types within each stimulus (the three forages; the three dustbaths; the three novel objects). In contrast, the motivations involved did affect peck morphology, since peck motor patterns between the different stimulus types were different from each other 124 (forages vs. dustbaths vs. novel objects vs. drinking). Furthermore, between bird variance for each peck type was very low, as is typical for Fixed Action Patterns (Barlow, 1968). The rest of the discussion will therefore talk about forages, novel objects, etc. without differentiating their shape or sub-type. Our results showed that severe feather pecks take longer to perform than gentle feather pecks. When observing the birds, severe feather pecks appear very quick; however, the physical range of head movement is much greater for severe feather pecks than for gentle feather pecks (which only involves a slight movement of the head and neck). Contact duration and the duration of head fixation were also longer. These differences suggest that severe and gentle feather pecks stem from different motivational systems. Furthermore, severe feather pecks are harmful to other birds and thus are a welfare concern, while gentle feather pecks are positively associated with preening in chickens, that is, as the incidence of gentle feather pecks increases, so does the amount of preening (Newberry et al, 2007) and gentle feather pecks may be similar to allo-preening found in other species of birds (e.g. Zebra Finches: Adkins-Regan and Robinson, 1993; Arabian Babblers: Pozis-Francois et al, 2004). We therefore consider them separately. Severe feather pecks proved similar in morphology to foraging pecks, but significantly different from all other types of pecks including pecks to dustbaths. In contrast, gentle feather pecks were typically different in morphology from other types of pecks. Therefore our data indicate that severe feather pecking stems from a re-directed motivation to forage and not re-directed motivation to dustbathe. Gentle feather pecking 125 in contrast had a different basis - possibly rooted in preening or in some other motivation not investigated here, for example the potential differences in appetitive and consummatory pecks discussed below. Given the opportunity, birds will spend a large amount of their time ground pecking (Dawkins, 1989) and the provision of a substrate has been known to decrease feather pecking (e.g. Savory and Mann, 1999; Bilcik and Keeling, 2000; Nicol et ah, 2001). Recent work has also shown there is a relationship between feather pecking and feather eating, with high feather pecking birds eating more feathers than low feather pecking birds (Harlander-Matauschek et ah, 2006b). The ingested feathers stimulate feed passage in the digestive tract in a manner similar to insoluble fibres, such as that found in litter substrates (Harlander-Matauschek et ah, 2006a). This is all consistent with feather pecking deriving from re-directed foraging. There is, admittedly, another potential explanation for our results. It is possible that the mere biomechanics of severe feather pecking and foraging are similar in that they both involve picking up an object (feather or food): thus our findings could reflect simply the morphologies of pecks involving pulling and lifting. However, we feel this is unlikely for three reasons. First, birds pecking at the dustbathing substrate sometimes picked particles up and sometimes consumed them. Similarly, birds pecked pieces off the novel objects and at times ate them. Despite this, novel objects and dustbathing pecks differed from foraging; and the within and between bird variances in peck morphology was still very low for novel object and dustbathing pecks, suggesting that this sporadic 'picking up' did not affect peck morphology. Second, we have shown that sub-type of stimulus and shape does not affect motor patterns, suggesting that a motivational not 126 biomechanical explanation is correct. Third, pigeon autoshaping data have demonstrated that the motivation behind a peck influences the motor patterns involved. For example, pigeons trained to peck a button to receive a feed reward will use the same feeding motor patterns on the button as they do when they peck the feed reward itself (Jenkins and Moore, 1973). A definitive test might thus be to now require poultry to peck operant buttons to reach different types of reward and look at the morphology of this peck to a completely standardized stimulus. Our prediction is that working for a feather peck stimulus will involve operant pecks that are spontaneously similar to those elicited when foraging is the reward, and different from pecks elicited to the operant for other rewards. Another potential source for error is the inability to separate the appetitive and consummatory aspects of the pecks. For example, chickens spend time searching for feed by pecking at and ingest numerous particles on the ground before finding appropriate food (e.g. Hogan-Warburton and Hogan, 1991). However, it is often very difficult to make the distinction between when appetitive pecking ends and consummatory pecking begins and in fact they may even over lap. For example, when dealing with appetitive and consummatory feeding behaviour, past studies have covered feed with grit and categorized the pecking at the grit to uncover the feed as appetitive behaviour while the pecking at the grain was categorized as consummatory behavior (however, actual ingestion of the feed was not recorded) (Andersson et al, 2001), while other studies considered all pecks directed at commercial feed to be foraging, even though the birds had been food deprived and their behavior could reflect hunger as well (e.g. Vaisanen and Jensen, 2003). Due to the set-up of the experiments conducted in this 127 chapter, we were not able to distinguish on a consistent enough basis whether the chickens only pecked at the stimuli or whether they also consumed pieces to make it a practical measure. Also, the chickens consumed pieces of stimuli that were non-edible, making the criteria of eating a piece of the stimuli equal to feeding-related consummatory behaviour unreliable. However, the variability within pecks and within birds was quite low, either indicating that we were reliably measuring the same response (either appetitive or consummatory) or that there is no difference with regards to morphology with these measures in chickens. It is possible that the differences found in severe and gentle feather pecks relate to differences in appetitive and consummatory feeding behaviour. Thus, perhaps severe feather pecks (which were consistently similar to foraging pecks) are actually similar to the consummatory pecks of feeding behaviour while gentle feather pecks (which were not similar to foraging pecks) may be similar to the appetitive pecks of feeding behaviour. Future research examining the finite details involved in appetitive and consummatory pecking and their comparison with gentle and severe feather pecking should help resolve this issue. Overall, the data thus concur with recent views that re-directed foraging leads to feather pecking. In terms of poultry welfare in commercial production systems, where processed feed is readily available from a trough, it does not appear that foraging motivation is being satisfactorily fulfilled. The ingestion of high fibre forages have also been found to aid digestive processes and nutrient digestion, as fibrous material stimulates activity in the gizzard (Hetland et ai, 2005). Thus, incorporation of forages into poultry housing, such as straw, hay or silage, is highly encouraged. 128 The techniques used in this paper provide a model to improve our ability to distinguish specific avian behaviour patterns that, although superficially similar in phenotype, are actually quite distinct. For example, foraging and dustbathing both involve ground pecks, but these pecks are different in their underlying motivation and associated motor patterns. It is possible for a substrate to be used for both foraging and dustbathing depending on the dominant motivation but analyzing the motor patterns involved in ground pecks is one way to differentiate between two behaviour patterns and thus to reduce the ambiguity of reporting 'ground pecks' in future studies. There are potentially other motor patterns involved in pecking that could be used in future studies to help distinguish between different peck types. For example, if chickens - like pigeons - have two distinct head fixations before pecking (Goodale, 1983), not just the one we were able to measure, then this additional measure may be useful when comparing peck morphology. Such additional motor patterns measures would be useful if one was found that gives good statistical power with a smaller sample size. Future work using FAPs could also test the idea that gentle feather pecks are related to preening. Turning to other species, future work with FAPs might help identify for instance whether piglet belly- nosing represents infantile udder-massage or adult-like rooting (cf. Torrey and Widowski, 2006), whether carnivore pacing represents patrolling, hunting, or escape (Clubb and Vickery, 2006), and so on. 129 4. GENERAL CONCLUSIONS Similar to pigeons, chickens appear to have FAPs involved in their pecking, based on the low variability within peck measures and these measures can be used to study motivation. The results from the second experiment indicate that severe and gentle feather pecking stem from different behaviour systems. The motor patterns involved in severe feather pecking resemble those of foraging pecks but gentle feather pecks largely differed in morphology from all other pecks measured. This supports the hypothesis that feather pecking stems from a re-directed motivation to forage, not to dustbathe; and emphasizes the need for laying hens to be housed in an environment that allows full expression of foraging behaviour. Finally, this paper demonstrates that analyzing 'Fixed Action Pattern' morphology can help determine the motivational bases of stereotypic behaviour in captive animals. 130 Door Feeder Figure 5.1: An example of the pen and test area set-up. 131 Angle of beak to vertical Figure 5.2: Vertical head angle (degrees): measured as the angle between a virtual line through the bird's eye down the middle of the beak to the vertical. 132 1 90 Figure 5.3: Horizontal head angle (degrees): the angle made by a virtual line through the middle of the bird's head, down the middle of the beak to the horizontal. 133 Figure 5.4: a) Duration (sec) of head fixation: the length of time that the head is kept still before the peck, b) Contact to stimulus: duration (sec) from the end of head fixation to beak contact with the stimulus, c) Duration (sec) of the whole peck: time from no head movement, through the peck, back to no head movement. 134 0.09! Novel Objects Forages Dustbaths Figure 5.5: An example of one of the peck measures used (duration of the peck) to examine pecks to the different shapes (Flat and Chicken-Shaped (C-S)) and materials used as the novel object, foraging and dustbathing stimuli measured in seconds (± SE). Each stimulus type had two shapes and three materials used whenever possible. All sub types of each stimulus were not significantly different (Dur: F^ 108) = 2.00, P = 0.0916) while the different types of stimuli (novel object, forages, and dustbaths) were significantly different from each other (F(4) 129) = 19, 219.8, P <0.0001). 135 Table 5.1: A comparison of the morphology involved in different types of pecks. Different letters denote significant differences (PO.05) while similar letters denote no significant difference (P>0.05) Grass Tissue Paper Head of Feathered Model Measure Mean ± SE Mean ± SE Mean ± SE F Value and dfs P Value VAngle (degrees) 47.07 ±1.09 a 49.15 ±1.19 a 65.7±1.5b F(2,56) = 4.63 0.0238 HAngle (degrees) 66.03 ± 0.94 a 65 ± 0.974 a 56.4 ±1.51 b F(2,56) = 5.31 0.0164 Duration of the 0.201 ± .0085 a 0.2 ±0.01 la 0.153 ± 0.012 b F(2, 56) =11-87 0.0005 peck (sec) Time to contact 0.069 ± 0.0026 a 0.07 ± 0.00325 a 0.059 ± 0.0045 b F(2,56) = 7.14 0.0052 the stimulus (sec) Duration of Head 0.19 ±0.0172 a 0.18±0.017a 0.14±0.018b F(2j56)= 18.89 0.0001 Fixation (sec) 136 Table 5.2: A comparison of various aspects of the peck morphology involved in gentle and severe feather pecks (FP). * denotes a P-value of <0.05 Gentle FP Severe FP Measure Mean ± SE Mean ± SE F Value and dfs P Value VAngle (degrees) 47.370 ±3.6500 50.160 ± F(i, 18) = 0.55 0.4771 1.2200 HAngle (degrees) 57.330 ±2.8300 62.850 ± F(i,i8) = 3.89 0.0800 2.5300 Duration of the peck 0.204 ±0.0410 0.238 ± F(i,i8) = 48.85 <0.0001* (sec) 0.0210 Time to contact the 0.064 ± 0.0070 0.069 ± F(i,i8) =1.82 0.1974 stimulus (sec) 0.0035 Duration of Head 0.130 ±0.0174 0.190 ± F(I,i8) = 24.06 0.0002* Fixation (sec) 0.0170 137 Table 5.3: Comparison of the Chicken-Shaped and Flat varieties of stimuli, ns denotes a P-value >0.05 Duration (sec) Contact (sec) Head Fixation (sec) Chicken-Shaped 0.1760 ±0.0025 0.0751 ±0.0015 0.1650 ±0.0045 Flat 0.1750 ±0.0030 0.0746 ±0.0015 0.1690 ±0.0050 F(i, 190) 0.47 0.71 1.73 p 0.4941 ns 0.4017 ns 0.1797 ns 138 Table 5.4: Severe feather pecks compared to the other types of pecks. Different letters denote significant differences (P<0.05) while similar letters denote no significant difference (P>0.05). Duration (sec) Contact (sec) Head Fixation (sec) Severe FP 0.190 ±0.005 a 0.078 ± 0.003 a 0.170 ±0.009 a Foraging Pecks 0.190 ±0.004 a 0.077 ± 0.003 a 0.170 ±0.008 a Dustbathing 0.140 ± 0.003 b 0.071 ± 0.002 b 0.140 ± 0.006 b Pecks Novel Object 0.150 ± 0.004 c 0.070 ± 0.002 b 0.160 ±0.008 a Pecks Water 0.960 ±0.018 d 0.120 ± 0.004 c 0.220 ± 0.010 c f(4,129) 19219.80 612.24 106.97 P O.0001 O.OOOl 0.0001 139 Table 5.5: Gentle feather pecks compared to the other types of pecks. Different letters denote significant differences (P<0.05) while similar letters denote no significant difference (P>0.05) Duration (sec) Contact (sec) Head Fixation (sec) Gentle FP 0.130 ±0.006 a 0.065 ± 0.003 a 0.150 ± 0.009 ab Foraging Pecks 0.190 ± 0.004 b 0.077 ± 0.003 b 0.170 ±0.008 a Dustbathing 0.140 ± 0.003 c 0.071 ± 0.002 c 0.140 ± 0.006 b Pecks Novel Object 0.150 ± 0.004 d 0.070 ± 0.002 c 0.160 ±0.008 a Pecks Water 0.960 ± 0.018 e 0.120 ± 0.004 d 0.220 ± 0.010 c F(4, 104) 17359.00 446.00 65.12 P O.0001 O.0001 O.0001 140 CHAPTER SIX A review of the methods used to study the motivations underlying stereotypic behaviour patterns (Manuscript based on this chapter to be submitted to Applied Animal Behaviour Science, with GJM as co-author) ABSTRACT Frustrated motivations to perform natural behaviour may cause stress and stereotypic behaviour (SB) in captive animals. The causal factors of SBs have been investigated using various research techniques, many designed to eliminate the behaviour. For example, the morphology, location or timing of the SB can be compared to normal behaviour, risk factors for developing the SB can be identified, external and internal factors can be manipulated, and/or predisposing factors can be determined. However, no research method is without problems and using only one method will not definitively determine the original motivation being frustrated. To further complicate matters, non-motivational factors can also play a role in SB development and maintenance. Biological or even management differences between animals (e.g. differences in stamina), may explain differences in SB levels. Performance of the SB may become habit-like, making the behaviour harder to interrupt, or the SB may continue due to perseverance, changes in CNS functioning that inhibit appropriate responses to 141 stimuli. Thus at times, performance of the SB may be unrelated to the 'source behaviour'. To attempt to account for all involved factors we put forward these four methodological points: 1) 2 or 3 complementary research methods should be used; 2) non-motivational explanations should also be considered; 3) young animals should be used; 4) pseudo-replication of resources/stimuli should be avoided whenever possible. The following of these methods will help to provide complementary results that will support or contradictory results that will reject the motivational hypothesis under test. If the original frustrated motivation can be determined, housing conditions can be improved and the welfare of the numerous animals performing stereotypic behaviour in captivity can be increased. 1. INTRODUCTION Motivation explains why an animal performs a certain behaviour pattern at any point in time and the amount of effort that is given to perform the behaviour. The frustration or thwarting of strongly motivated behaviour patterns is aversive and can lead to behavioural changes and ultimately stereotypic behaviours. There is evidence for two types of motivational explanations and one non-motivational explanation for stereotypic behaviour (Mason, 2006): that they derive from redirected or vacuum activities performed by frustrated or thwarted animals; that they derive from escape attempts shown by frustrated or thwarted animals; and/or that housing or rearing animals without 142 motivationally significant stimuli causes stress and/or abnormal CNS development which alter how behavior is controlled, e.g. making it prone to preservation. There has been much interest in determining the motivational origins of stereotypic behaviours. The reasons behind this interest include both applied and fundamental aims. For example, environmental enrichments provided are often only partially successful at alleviating stereotypic behaviours, suggesting perhaps that the wrong enrichment has been used (applied reason); the 'mystery' of why, at the individual level, performing some stereotypic behaviours seems related to relatively improved welfare, compared to non-stereotypic animals in the same poor environment (fundamental reason); and the desire to understand how non-motivational factors like 'perseveration' inter-relate with motivational explanations for stereotypic behavior (fundamental reason). However, researchers working with different species, or coming from different disciplines, have often used different approaches. Thus, our aim was to review the various techniques available. We devote the majority of the paper to discussing the pros and cons of various research approaches for investigating stereotypic behaviours' motivational origins. These include analyzing the morphology, timing and location of stereotypic behaviour; identifying — via epidemiology or experimental manipulation (including the use of environmental enrichments) — the external eliciting stimuli and internal physiological states that affect the behavior; and investigating the traits - including strength of 143 preference for various resources or natural behaviours — that predispose individuals, genotypes or species to the developing the behavior. 2. MOTIVATIONAL AND NON-MOTIVATIONAL EXPLANATIONS 2.1. Motivational explanations for stereotypic behavior These properties of motivated behavior have together led to two main ethological explanations for stereotypic behaviours. As discussed in Chapter 1, the first is that they derive from 'redirected', 'vacuum' or 'intention' movements and if a whole behaviour sequence is not possible, 'intention movements' may be shown (e.g. Breland and Breland, 1963; Hinde, 1970; Chapter 1). Thus, redirected, vacuum or intention behaviours have often been suggested to be the 'source behaviours' (Mason 1991a) for stereotypic behaviours (see also Clubb et al., 2006) and may contain similar motor patterns, diurnal rhythms, or eliciting stimuli as the originally motivated behaviour. The second proposed link between motivation and stereotypic behaviour is that instead of attempting to perform normal behaviour patterns that are not possible in the captive environment, animals may simply try to remove themselves from the frustrating situation through repeated escape attempts (cf. Mason, 2006). For example, when Duncan and Wood-Gush (1972) prevented hens from feeding, they began to vocalize and repeatedly pace against the exit door of the cage in an attempt to escape the aversive environment. Repeated attempts to escape from motivationally frustrating environments 144 may therefore be another 'source behaviour'. For example, Keiper (1969) suggested this as the source of route-tracing in wild-caught canaries. Because escape is impossible, these should be described as 'intention movements'. Stereotypic behaviours derived in this way may therefore involve similar diurnal rhythms or eliciting stimuli as the normal behaviour that was originally motivated, but not similar motor patterns. 2.2. Non-motivational explanations for stereotypic behavior It is important to consider non-motivational explanations as to why stereotypic behaviour patterns may occur or continue in long bouts, and perhaps why the performance of these behaviours can be so difficult to extinguish, because these factors may need to be controlled for when designing experiments aimed to investigate motivation. As discussed in Chapter 1, animals may become stereotypic because they themselves have been altered in some profound way, potentially through permanent changes in CNS functioning or neural control of the behaviour may shift into central control or habit formation with repeated performance (for reviews of these processes see e.g. Dantzer, 1986; Mason and Turner, 1993; Toates, 2001; Mason 2006). This may also make the animal less responsive to external stimuli, in a way which could affect some experiments. Other general properties of the animal and/or its circumstances which can affect SB are non-pathological effects such as the animal's stamina, and its available free time. (e.g. Blackshaw and McVeigh, 1984 in Rushen, 1984; see Chapter 1). Thus, measurements of the time an animal spends performing SB must ideally take into account 145 the time spent performing other behaviour patterns as well, to give an accurate representation of the level of SB shown (see e.g. Vickery, 2006). Thus overall, while some SBs may (at least originally) stem from frustrated motivations, others may be due to, or exacerbated by, abnormal brain function, or by habit formation, or even by physical strength or 'spare time'. These various causal factors may help explain why SB can be so hard to reduce or abolish and can still be found in enriched environments. They are all important to control for in experiments aimed at assessing the original source behavior or motivational origin of a stereotypic behaviour. 3. RESEARCH METHODS USED TO INVESTIGATE THE MOTIVATIONS BEHIND STEREOTYPIC BEHAVIOUR PATTERNS 3.1. The form of the stereotypic behaviour Some motivated actions can have distinctive morphologies, especially the consummatory phases of behavioural sequences which tend to be more species-typical and unvarying (see Chapter 1). This can potentially help us identify the motivational origin of a redirected, vacuum or intention movement. For example, in pigs, raccoons and other species, species-typical foraging movements may be redirected at inedible objects predicting food reward (the 'misbehaviour' of Breland and Breland 1961), while in pigeons, the distinctive Fixed Action Patterns that comprise feeding and drinking pecks 146 (Zweers 1982, 1992; see also Chapter 5) are also expressed when birds peck a key to gain respectively food or water (Jenkins and Moore, 1973). The form/morphology of SB can therefore also be useful for generating or even testing hypotheses regarding the underlying motivation. For example, Morris (1964), observing caged parrots displaying repeated swinging or rolling head movements, suggested they resemble the initial take off movements that would be performed if the birds had the space available to fly, and thus may stem from frustrated motivations to fly (Morris, 1964). The stereotypic tongue-rolling of dairy cows appears so similar to the movements made in grazing, where the tongue is rolled around grass to pull it into the mouth (Sambraus, 1985, in Bergeron et al, 2006), that this led to hypotheses that it stems from unfulfilled foraging motivation (e.g. Redbo, 1990; Sato et al, 1992). Form/morphology can also rule out some possible motivational explanations, for example the squeaking, head-rolling and 'huffing' sometimes seen when carnivores pace (Mason 1993, Poulsen and Tesky, 2006) suggest that pacing cannot be derived in a simple way from hunting; while the form of bar-mouthing displayed by caged mice, in which the bar is held in the diastema, shows that the behavior is not derived from gnawing to aid tooth-wear (Wurbel 2006a). One obvious problem with this rather subjective, qualitative approach is that several normal behaviour patterns may involve similar movements, making it hard to judge which one underlies a similar-looking SB. For example, bar-biting in pigs could 147 represent aggressive biting or foraging-related biting (e.g. Rushen, 1984); locomotory movements in carnivores could represent prey-search, prey-capture, mate search, patrolling, migrating or exploration (e.g. Mason, 1993; Carlstead, 1998; Clubb and Vickey, 2006); and the repetitive digging movements performed in cage corners by captive Mongolian gerbils looks similar to natural burrow-creating movements (Wiedenmayer, 1997), but could equally represent an attempt to escape from the cage. Thus, looking only at the superficial morphology of the SB is generally not enough to determine its motivational cause. Another potential problem is that SB may alter in form over time. For example, caged chaffinchs show intention movements of nest material collection which become abbreviated to tic-like head movements (Hinde, 1962 in Mason, 1993). The loss of elements from the behaviour may be responsible for the reported transition from environment-directed to self-directed stereotypic behaviour that is sometimes seen. For example, crib biting in horses (which requires the horse to grasp a fixed object by their incisor teeth) has been reported to change to wind sucking (no grasping of object) (Meyer-Holzapfel, 1968). Not all SBs become abbreviated as they progress: the sequence of movements may become more complex over time, even if the movements involved are abbreviated. Sows perform increasingly longer stereotypic behaviour sequences, with new added components, the longer they are restrained (Cronin et al, 1984 in Mason, 1993). Thus, as the SB continues, its movement patterns become increasingly dissimilar to those involved in the source behaviour. 148 For responses involving Fixed Action Patterns, it may however be possible to collect quantitative data that do allow the proper test of competing motivational hypotheses. There is only one example of their use in studying SB to date: that of feather pecking (Dixon et ah, in press; see Chapter 5). The FAPs involved in different types of pecks, such as foraging, dustbathing, drinking and novel object pecks were recorded and compared to feather pecking pecks. The motor patterns involved in severe feather pecks were similar to foraging pecks, but different from all other pecks including dustbathing, indicating that severe feather pecking derives from re-directed motivations to forage, not to dustbathe. This approach has a lot of promise for the future, but we should point out its potential disadvantages too. First, the measurement of FAPs requires the behaviour patterns under test to be recorded with high resolution and using controlled viewing angles, etc. This may not be possible if, say, data are needed from wild animals in non- captive environments or even from large enclosures. The analysis of the images can be slow and time-consuming. Furthermore, there are experimental design issues that need be dealt with: if the aim is to assess similarity, a large enough sample size needs to be used such that a non-significant difference is unlikely to be merely a Type II error; and when behavior patterns are recorded, it should be ensured that any potential similarities or differences in morphology are due to underlying motivation and not any other aspect of the stimulus or situation (such as substrate/object shape or material) (see also Chapter 5). 149 3.2. The location and timing of the stereotypic behavior Some stereotypic behaviour patterns occur during certain times of day or in specific locations within the cage/enclosure. Examining this timing and/or location can help identify the internal and external causal factors that influence the behaviour (e.g. Mason, 1993; Vickery and Mason, 2004); suggest what the animal's 'intention' is, in the case of suspected intention movements (e.g. Meyer-Holzapfel, 1968); and allow comparison with the temporal patterns or locations of those normal behaviours that have been hypothesized to have a common motivational basis (e.g. Terlouw et al, 1991). Again, this is largely a qualitative approach leading to the generation rather than test of hypotheses, but it has also sometimes been used been more formally and quantitatively. One early example comes from Winkelstraeter (1960, in Meyer-Holzapfel, 1968) who reported circular running in a female ocelot when expecting food. This and many other cases like it lead to the hypothesis that locomotor SB in carnivores derives from hunting (e.g. Terlouw et al, 1991; Mason and Mendl, 1997; Toates, 2000). However, further timing data exist from these and other species, collected in more contexts, that cast doubt with this hypothesis. Pacing and similar by carnivores is not exclusively seen pre-feeding; thus in ad libitum fed mink it is most common at dawn and dusk when animals are most generally active (Hansen et al, 1994), and in some carnivores, such as the Persian leopard and snow leopard (Lyons et al., 1997) or fennec foxes seeking somewhere to cache food items (Carlstead, 1991) pacing actually increases after feeding. Furthermore, poultry and ungulates have also been shown to locomote before feeding 150 (reviewed by Mason and Mendl, 1997); for example, broiler breeders begin to pace their cages as feeding time approaches (Blokhuis, 1986; Savory and Maros, 1993). This suggests motivations other than hunting are at work, e.g. general appetitive or food searching, a general anticipation of reward (van der Harst and Spruijt, 2007), or even frustrated attempts to escape (reviewed by Clubb and Vickery, 2006) - a hypothesis we return to below. Related studies demonstrated that circadian patterns were also found in oral forms of SB: patterns that were more consistent than those found in locomotory forms, like pacing. Poultry, ungulates, and even pandas and walruses, locomote before feeding but then switch to stereotypic oral behaviour patterns after consuming their daily meal (reviewed by Mason and Mendl, 1997; Mason 2006, Box 2.2). For example, broiler breeders stereotyped pecking at the feeder or floor, and also feather pecking, become more frequent their daily meal (Blokhuis, 1986; Savory and Maros, 1993). Well- designed experiments by Terlouw and colleagues (1993) showed that, in pigs at least, it is the ingestion of food that stimulates the SB rather than other external cues (e.g. increased noise) occurring at feeding time. This led to ideas that such behaviours are related to either post-ingestive physiological changes or enhanced motivations to forage (Bergeron et al. 2006). For example, in the wild, pigs are patch feeders (food is in small clusters), thus consuming a small amount of food should actually increase foraging motivation (positive feedback) (Mason & Mendl 1993). 151 Our final example of the use of timing to test a motivational hypothesis comes from feather-pecking in chickens, which has been hypothesized to stem from redirected foraging or redirected dustbathing behaviour (e.g. Blokhuis, 1986; Vestergaard, 1994). The amount of feather pecking performed throughout the day was recorded and compared to that of foraging, which occurs regularly throughout the day (e.g. Duncan et al, 1970), and dustbathing, which peaks about 5-6 hours after light on (e.g. Vestergaard 1982a). It was found that feather pecking followed a similar daily rhythm as foraging, not dustbathing, and adds support for the hypothesis that feather pecking stems from redirected foraging (see Chapter 2). Using the timing of SBs to suggest or even test ideas about their motivational origins has several advantages. This type of experiment does not need to involve complex (or expensive) equipment, and the behavioural measures needed are fairly straight forward and easy to identify (i.e. durations/timing of various behaviour patterns) and thus would make a useful 'first step' into determining the motivational basis for a SB. Its disadvantages are as follows: to be done well, animals should be observed over the full 24 h; the factors underlying any circadian or ultradian rhythms may well still need elucidating with further experiment; and several normal behavior patterns may share similar timing, especially if they are all similarly affected by changes in arousal/activity. Furthermore, for some SBs, perseveration may have occurred, emancipating the behaviour from its original motivating factors, making measures of timing and/or location irrelevant. On the basis of anecdotal evidence, it has been suggested that SB may 152 sometimes continue even after the frustrated or thwarted behaviour is able be performed. For example, as mentioned in Chapter 1, when hens are prevented from feeding, they repeatedly pace against the exit door of the cage in an attempt to escape the aversive environment; however, when the exit door was left open for two of the hens, they continued pacing in the cage, suggesting that the stereotypic behaviour had become emancipated from the source behaviour (Duncan and Wood-Gush, 1972). Schloeth (1954, in Meyer-Hozapfel, 1968) moved circus polar bears displaying pacing SB into larger zoo enclosure; however, some bears maintained their pacing at the same dimensions of their circus enclosure, again showing emancipation of the SB from the source behaviour. These results suggest that stereotypic behaviour may become emancipated from its original motivational factors (e.g. escape) as animals age and these effects cannot be completely prevented by moving the animals to a better environment. In 'old' SBs, timing maybe therefore be misleading. SB location gave weight to suggestions that at least some locomotor forms derive from attempts to escape. Early reports were largely anecdotal. For example, when a dingo was isolated from its pack in an adjacent enclosure at an Amsterdam zoo, it immediately started pacing back and forth along the separating fence, eventually leading to 'figure eight' type paths (Meyer-Hozapfel, 1968). Mason (1993) also describes the locations of pacing in mink as if trying to escape from human or an unwanted mate. Shepherdson (1989 in Clubb and Vickcry, 2006) suggested the 'escape hypothesis' for all carnivore pacing on the ground that it typically occurs at enclosure edges, or against locked doors animals wish to pass through (although this suggestion, that 'edges are 153 typical' never seems to have been tested statistically - to do this careful data collection would be needed on enclosure zones, their relative areas, etc., to see if 'edge use' occurs more often than one would expect by chance). This idea has, however, been tested experimentally for caged rodents (Nevison et al. 1999a; Lewis and Hurst, 2004). Laboratory mice perform a number of SBs directed to the bars of their cages, such as bar gnawing and bar circling (e.g. Wiirbel et al., 1998; Nevison et al., 1999a). The gaps between the bars represent the only contact the mice get with their external environment and due to the high levels of SB associated with this location: do these behaviour patterns therefore signify frustrated escape attempts? To test this hypothesis, Nevison et al., (1999) housed mice in cages with two sets of metal bars, one which could open and served as the entrance/exit to the cage and one which remained closed. In addition, half of entrance/exit bars and half of the closed bars had a clear Perspex backing to eliminate external cues. It was predicted that if the mice were motivated to escape, they would perform most of their SB by the exit bars and/or away from the Perspex backed bars, depending on the combination of variables in the cage. Overall this was confirmed: mice were found to interact more with the metal bars that allowed external cues (no Perspex) than with the bars that were blocked with Perspex and more with the bars providing a potential exit/escape route than with the metal bars that never opened — thus supporting the hypothesis that the SB stems form frustrated escape attempts. A potential problem with these methods is that the mice may have been attracted to the entrance/exit doors in general because they want to escape, but ended up spending most of their time near these doors so that everything they did (both stereotypic and non 154 stereotypic behaviour) seemed to be located near the entrance/exit, thus increased stereotypic behaviour directed towards these doors would not reveal much in terms of motivation underlying the behaviour. This is an elegant example of using stereotypic behaviour location, along with the manipulation of cues/expectation at those locations, to test a motivational hypothesis. However, it only tests an idea about the source behavior of the stereotypic behaviour: it does not, without further experiment, tell us exactly why the mice (or carnivores in other studies) might be so motivated to escape. Several different motivations (e.g. to seek shelter, to avoid aggression, to find mates, to forage, to explore, to range) may underlie motivations to escape (cf. Wiirbel, 2006a; Clubb & Vickery 2006) - and it is these underlying motivations which need to be addressed if we are to tackle welfare and stereotypic behaviour effectively. A second potential drawback with relying on 'location' data is that stereotypic behaviour may become divorced from its original motivation; thus, the location and timing of the SB may not be similar to that of the original behaviour. For example, when behaviour is redirected or transferred to a less preferred substitute, the location of the substitute stimuli is not necessarily similar to the original or preferred one (e.g. Breland and Breland, 1963). This is quite evident in the case of feather pecking: the behaviour may originally stem from foraging or dustbathing motivation, both of which are directed to the ground; however, feather pecks are directed to the feathers of other birds (e.g. Blokhuis, 1986; Vestergaard, 1994). Thus, based only on comparisons of the locations of stereotypic and non-stereotypic behaviour, it could be concluded that feather pecking derives from preening. 155 3.3. External factors affecting stereotypic behaviour External factors, such as eliciting stimuli, can help determine which behaviour patterns are most motivated, and how strong the motivation is. Thus assessing and manipulating the external stimuli that influence a stereotypic behavior is an obvious way of investigating its motivational basis. Researchers have typically done this in three ways: assessing aspects of housing or diet that make SB more or less prevalent in a population, using an epidemiological approach; manipulated aspects of housing or diet experimentally, to observe its effect on both prevalence and individual time budgets (sometimes with the practical aim of reducing SB rather than with the scientific aim of understanding its origins); or observing the onset and cessation of bouts of SBs more closely to try and ascertain the environmental triggers that initiate performance. One of the first studies to screen a population to identify aspects of environmental risk factors for the development of stereotypic behaviour was by McGreevy and colleagues (2001). However, these results were not easy to interpret, perhaps due to the huge variety of ways in which housing, diet, training, and so on, can vary across stables. Later studies yielded some clearer findings; for example, in a survey of over 2000 Swiss horses, animals fed primarily concentrate with no pasture access had increased odds of displaying cribbing, weaving or stall-walking behaviour (Bachmann et al, 2003). This work was thus consistent with the hypothesis that frustrating natural foraging plays a role in equine stereotypic behaviours, and also led to practical recommendations that 156 reduction can be achieved through the provision of low concentrate/high forage diets and pasture access. Further surveys of another ungulate, zoo-housed giraffes, again suggested that diet and its presentation particularly influenced oral stereotypic behaviours, while other aspects of husbandry (e.g. size of indoor enclosure) were better predictors of pacing (Bashaw et al, 2001). Together these studies suggest that different forms of ungulate SB are affected by different aspects of housing and husbandry, and thus have different motivational bases. Similar epidemiological approaches have been taken by several other authors. Green et al. (2000) surveying chickens, found that, for example, feather pecking was more likely in flocks that did not go outside much even when the weather was sunny, and in flock with bell drinkers (as opposed to nipple drinkers); while in clouded leopards, the availability of high enclosures/hiding resting places seemed linked to SB, and leopards provided these enclosures also had decreased stress levels (Wielebnowski et al, 2002). Other similar epidemiological studies include those of Garner et al. (2004: mice; 2006: parrots); and Tarou et al. (2005: prosimians). Epidemiological techniques can be useful for determining practical recommendations and as a starting point for generating hypotheses. They also have the advantage of using pre-existing variation, which may allow large data-sets to be built up for relatively little cost and also removes the need to house additional animals in poor conditions in order to research their stereotypic behaviour. However, epidemiological studies are generally designed to look for sources of variation and not to test for causation. In addition, due to the involvement of so many inter-correlated independent 157 variables (not all of which will be reported in a questionnaire) it can become hard to tell what factors actually correlate. False results may be caused by confounding variables if care is not taken (e.g. types of husbandry that systematically co-vary with genotype): thus if an important cause of stereotypic behaviour is the same across all husbandry systems, epidemiological approaches will not be able to identify it. For example, under natural conditions, chicks are brooded by the mother hen (e.g. Wood-Gush and Duncan, 1976); however chicks in commercial settings are never given contact with the hen. Thus, the lack of the hen and the early experiences associated with her may affect stereotypic behaviour in chickens but as this practice does not occur, with the possible exception of some small, 'back yard' flocks, it would not be identified as a risk factor for SB. Alternately, if two aspects of husbandry always co-occur (e.g. eating grass and having time out of the stable for horses) or even aspects of husbandry and genetics (e.g. some breeds of horses are always kept in particular ways - see also 'Predisposing traits' section, below), it will be impossible to tell which is most important. Furthermore, the risk factors they identify may not always act via motivation, i.e. by causing/alleviating specific frustrations. For example, general stress and even disease may be amongst the measured risk factors (e.g. being in a flock that has egg peritonitis is one risk factor for feather pecking: Green et al, 2000); while husbandry systems that increase the number of behaviour patterns available to be performed (something not typically measured in such studies) may decrease SB simply because animals have more to occupy their time, and so have less time in which to perform stereotypic behaviour (e.g. Swaisgood et al, 2001). Lastly, epidemiological approaches are correlational only: they reveal what correlates with the behaviour once it has emerged, but not what actually causes it to emerge. A 158 good example comes from Tarou et ah (2005) on prosimians (e.g. lemurs); they found that the most stereotypic animals also had the most enriched environments! Potential explanations were that enrichment induces SB (perhaps the enrichments roused the animals out of a depressed stupor, or perhaps they were stressful rather than truly enriching); however an alternative explanation is that the SB causes increased enrichment attempts on the part of the keepers — thus demonstrating that correlations are not able to show cause and effect. Thus it seems near impossible to actually test motivational hypotheses with epidemiology. Experimentally manipulating external factors is an alternative approach that avoids many of these problems. Stimuli thought to relate to unfulfilled motivated behaviour (e.g. forages, perches, nestboxes, cover, and other forms of'environmental enrichment') can be added, removed or changed and the effect this has on the prevalence or frequency of SB can be measured. These environmental enrichments are typically designed to attempt to either abolish the unfulfilled motivation by providing the end result of the behaviour (e.g. a tunnel leading to a nest chamber), or to allow the motivated behaviour to be re-directed elsewhere in a more natural way (e.g. hiding food in the enclosure to allow foraging) (see Swaisgood and Shepherdson, 2006 for a review of enrichment techniques). It is thought that if the stereotypic behaviour is reduced by this manipulation, then this may indicate post-hoc that the motivation that was not being fulfilled (Mason et al, 2007). 159 There are many, many examples of enrichments being used to tackle stereotypic behaviour (for example, an ISI Web of Knowledge search of 'enrich' + 'stereotyp' yielded about 300 hits), so we will just give a few here. When horses known to wood chew were provided with forages, their wood chewing was decreased compared to those fed a concentrated grain. This once again suggests that horses are motivated to forage but this motivation is not being fulfilled with a diet of concentrated grains (Redbo et al, 1998). In a follow up to the epidemiological study of clouded leopards, provision of enclosures where the leopards could hide decreased fecal corticoid concentrations (Shepherdon et al., 2004). Enriching zoo carnivores by enlarging their cage sizes may reduce SB by providing more space to e.g. roan or alternatively, it may allow the animals to retreat further from visitors (e.g. escape aversive), again leaving the motivation behind the SB unclear (Clubb and Vickery, 2006). In fact, when a whole diverse array of environmental enrichments affect pacing and other locomotor SBs that may relate to escape, it is possible that they are all working by helping to make the cage a less aversive place to be - with every change helping to do this (cf. Clubb and Vickery, 2006).). Chickens provided with a substrate that they could peck and scratch on, performed less feather pecking than those housed in wire (e.g. Aerni et al., 2001; Nicol et al., 2001) and an inverse relationship has even been demonstrated between levels of ground pecking and feather pecking (Blokhuis, 1986). However, a problem with this example is that this leaves the motivation behind ground-pecking unclear (is it to forage, to dustbathe or even to explore — as all involve ground pecks). 160 At around the same time, some other research, this time on gerbils, showed how very specific the enrichments may be which successfully target and abolish an SB. In an attempt to determine the motivation behind stereotypic digging in Mongolian gerbils, Wiedenmeyer (1996, 1997) manipulated various aspects of their environment. Gerbils were housed in either standard laboratory sized cages, or in cages four times the size of the standard model; however, there was no effect on the development of stereotypic behaviour (Wiedenmayer, 1996). Gerbils were also provided with a substrate to dig in (sand) to determine if it was the act of normal digging (appetitive behaviour) which was frustrated, but again this had no effect. However, if gerbils were provided with an artificial burrow leading to a nest chamber, they did not develop this stereotypic behaviour. This shows that a naturalistic stimulus resembling the end result of burrow formation (consummatory behaviour) is needed to terminate digging motivations in gerbils. Different elements of the complete burrow, such as the tube leading to the enclosed chamber or just the chamber alone, were then provided separately to determine if there was a key feature that decreased digging motivation. It was found that the tube alone (attached to the nest chamber) was the necessary stimulus to prevent the appearance of stereotypic digging; gerbils provided only the chamber still developed this behaviour. Enrichment experiments like these can thus reveal practical ways to abolish stereotypic behaviour, and also test competing motivational hypotheses. Sometimes, however, interpreting 'enrichment effects' is not straightforward. One problem with manipulating external factors in this way is that assumptions about the motivations being 161 satisfied by a resource may not be accurate. For example, in one study of mink, a raised look-out platform was provided which the researchers termed an 'elevated look out' - but on close inspection of its actual use, it became evident that the mink instead used it as a place for drying themselves off after swimming (Mason, personal communication). In other cases the resources provided may actually be able to be used to satisfy a number of motivations. For example, the provision of all substrates to chickens may allow foraging and dust-bathing behaviour (and also other activities like exploration and even nest- building). A related problem is that an enrichment may have general, non-motivational effects, like taking up an animal's time (e.g. Swaisgood and Shepherd, 2006; Vickery, 2006 Box 9.4), or even reducing its general stress, for instance, by providing hiding places (with stress reduction then affecting tendencies to perseverate). These last effects could help explain why stereotypic behaviour can sometimes be reduced by a number of very diverse enrichments, which seem to fulfill widely different motivations. For example, novel objects and foraging devices both successfully reduced pacing, somersaulting and swaying in captive Giant Pandas (Swaisgood et al, 2001); providing extra spaces with climbing structures, or instead foraging enrichments (food scattered in sawdust), both reduced stereotypic pacing in rhesus macaques to the same extent (Gimpel, 2005, in Novak et al., 2006 Box 9.3 ); and indeed that very diverse enrichments have very similar effects on SB seems to be rather typical, at least in the zoo animal literature (see meta-analysis by Swaisgood and Shepherd, 2006). 162 One last stumbling block in the use of enrichments is that animals may become resistant to them as they age. Thus older voles (14 months old), for example, were more persistent (or harder to interrupt) while engaged in SB than younger voles (2 months old). Additionally, after being moved to an enriched environment (including a larger cage), older voles still performed SB, while young bank voles did not (Cooper et al., 1996).. We do not currently know whether this means that they come to find them less motivationally satisfying, or instead whether other, non-motivational effects are at work, e.g. the SB becomes a fixed habit, or the animals becomes increasingly perseverative with age. A final use of external stimuli to investigate the motivational basis of SB is to investigate which release or terminate bouts of the behaviour. To illustrate such effects in captivity, thrown inanimate objects like stones would be chased by one captive coati, and retrieved (also pushed back through the wire back to the thrower!); it was speculated that the thrown pebble released natural motivations to chase prey items (Gilbert Manley, in Morris, 1964). As mentioned previously, Terlouw et al. (1993) demonstrated that it was the ingestion of food that stimulated SB rather than other external cues. Similarly, de Passille et al. (1992; 1993) performed an elegant series of experiments that demonstrated that non-nutritive sucking in calves is elicited by the taste of milk, not the volume of milk consumed or the hunger state of the calves. In a final example, exposure to stressors was shown to elicit bouts of stereotypic and self-injurious behaviour in caged primates. In a laboratory, the presence of a technician wearing black 'catch gloves' (worn when primates were to be removed from the cage for a procedure) increased the primate's 163 stereotypic behaviour, such as pacing and rocking, and self injurious behaviour, such as self-biting.(Cross and Harlow, 1965 in Novak et al, 2006). One explanation was that the acute stressors motivated behaviours which somehow act to reduce that stress, and later work (reviewed Novak et al, 2006) did provide evidence for that idea for self-biting (as we discuss in the next section), albeit by mechanisms that are only partially understood. 3.4. Internal factors affecting stereotypic behaviour If external stimuli typically interact with internal stimuli (e.g. physiological signals relevant to homeostasis; pre-existing psychological states like fear or other forms of expectation to determine motivation, then assessing or manipulating specific internal stimuli is another way of investigating the motivational bases of stereotypic behaviour. Experiments have typically taken two forms: seeing how changes in internal state affect the prevalence or frequency of SB in an experimental group, and monitoring individual animals during or just after the performance of a bout of stereotypic behaviour, to ascertain if there are particular internal changes which bring the behaviour to a close. The effects of internal state on the prevalence or frequency of SB may be investigated by altering the resources available for an animal to consume (e.g. feed/water availability); by using invasive techniques to assess or alter physiological state more directly, perhaps via drugs or hormones; and by seeing how age, breeding status (e.g. sex and oestrous state) or previous experience affect the stereotypic behaviour induced by a standardized change in environment. Much work on pigs, for example, has investigated 164 the effects of the restriction of food, or conversely diets which supply more energy or greater gut fill, on their post-feeding oral stereotypic behaviours. For example, decreasing food rations increases SB (Appleby and Lawrence, 1988) while providing high fibre diets that lead to gut fill and promote short term satiety, also reduce stereotypic behaviour and feeding motivation in the first few hours after feeding (Lepionka et ah, 1997; Rushen et ah, 1999). However, pre-feeding SB returns before the next meal, indicating that fibre alone is insufficient to reduce hunger long term and an increase in energy intake may also be important (cf. Bergeron et ah, 2006). The manipulations of these internal factors in pigs suggest involvement of hunger and feeding motivation in their stereotypic behaviour. However, these experiments alone cannot disentangle the effects of hunger caused by low calorie rations, and the frustration of natural foraging behaviour caused by the small amounts of food that are presented simply in a trough. Physiological means of investigating the role of internal state more directly might be to manipulate signals of homeostatic need/satiety by means of intra-venous nutrition, and/or administration of CCK, leptin or ghrelin (see review by Geary, 2004) but such work seems not to have been conducted yet on pigs. Invasive work on cows has, however, allowed the direct manipulation of gut fill without involving feed intake behaviours. Fistulated cows that had rumen contents added directed into their stomach but were not allowed to feed did not show increased levels of SB compared to cows who were food deprived only (Lindstrom and Redbo, 2000). While tests of the role of fear in e.g. escape-related pacing, have been performed by dosing animals with anxiolytic agents. An early 165 example comes from laying hens showing stereotyped pacing being dosed with a tranquillizer which subsequently decreased their pacing behaviour; however, this may have been due to the sedative-like effects of the tranquillizer (Duncan and Wood-Gush, 1974). Additionally, a polar bear in the Calgary zoo spending about 70% of her time pacing was treated with Prozac (SSRI), which gradually reduced her SB over a number of weeks (Poulsen and Teskey, 2006 Box. 10.4). Companion animals are often treated with drugs to decrease their incidence of SB. Dogs that had been diagnosed with a variety of SB, from self licking to tail chasing, were treated with clomipramine (tricyclic antidepressant) and this was four times more likely to reduce the SB than the placebo (Hewsone/a/., 1998). Finally, ethologists have sometimes investigated the effects of internal state indirectly by varying the age, sex, reproductive status or past experience of animals exposed to conditions that induce stereotypic behavior. For example, piglets removed from their mothers and moved to new enclosures react differently to this manipulation if weaned at different ages, showing more belly-nosing if this occurs earlier - an effect consistent with the behavior deriving from frustrated suckling (Worobec et al, 1999). Keiper (1969) showed that canaries differing in past experience also varied in their responses to being caged. For example, wild-caught birds showed far more route-tracing than captive-bred animals, an effect he hypothesized was due to greater motivation to escape confinement. More recently, Latham & Mason (in prep.) experimentally manipulated motivation to escape by raising lab. mice in two groups. One was always housed in standard cages, while the other raised with enrichments but then transferred to 166 standard cages, the hypothesis being that the latter would be more frustrated at confinement due to contrast effects. This hypothesis was tested by assessing the strength of motivation the mice had to reach an enriched cage (via the maximum weight they would push). Enriched-reared mice moved to standard cages were -just as predicted ~ both more stereotypic and more motivated to reach enrichments than standard-raised mice, suggesting motivations to escape underlie their SBs in standard cages. The other broad group of studies of internal stimuli have attempted to see how they change over the course of performance of a bout of SB. Ethologists have done this rather indirectly, treating the animal as a 'black box' in which internal stimuli (inferred rather than measured) relevant to the behavior pattern in question may build up if behavior is prevented, but subside if the behavior is allowed. One question tackled with this type of approach is whether a stereotypic behaviour is able to substitute for certain more naturalistic behaviours. If it can substitute, this would help identify the likely source behavior and help to explain why the SB is performed so frequently: if the stereotypic behaviour results in motivationally significant feedback, then its performance should decrease an animal's motivation to perform the 'real behaviour' (e.g. Hughes, 1980, Lindberg and Nicol, 1997). One experiment has sought to investigate this, though unfortunately not with significant results. Could vacuum dustbathing, as performed by hens without access to litter, at least partially satisfy the motivation to dustbathe? This question was investigated by Olsson and colleagues (2002) by allowing litter deprived hens to either dustbathe, vacuum dustbathe, or see dust but not dustbathe. The birds were then given litter, and their latencies to dust-bathe and the amount performed was 167 measured. Dustbathing was reduced by previous true dust-bathing, but no reduction in dustbathing was found in hens that had previously only vacuum dustbathed - indicating that the motivation to dustbathe was not reduced by a bout of vacuum dustbathing. A few other studies have used invasive measures to collect direct data on physiological change. Calves allowed to perform non-nutritive sucking after drinking milk had increased insulin and CCK (related to post-ingestion satiety) compared to calves only given milk to drink (de Passille et al, 1993; Rushen and dePassille, 1995). This suggest calves are motivated to perform non-nutritive sucking because it resembles teat- sucking (which they are deprived of) and it has beneficial effects on their digestive physiology. Additionally, self-injurious behaviour in primates is associated with a decrease in heart rate, and an inverse relationship to Cortisol levels, suggesting that this behaviour may reduce anxiety and may serve as a coping mechanism to alleviate acute episodes of stress (Marinus et al, 2000; Novak, 2003). 3.5. Predisposing factors to developing stereotypic behavior Motivational explanations can be both generated and tested by research investigating the traits that either predispose individuals, genotypes or species to developing the behavior, or that correlate with the behavior once it has developed. This research may exploit naturally occurring individual or strain/species variation, or instead generate/enhance that variation through differential rearing or genetic selection. The basic reasoning here is that if a particular motivation underlies a stereotypic behaviour, 168 then individuals or genotypes developing high levels of that stereotypic behaviour should display other behavioural signs of that elevated motivation. These signs might include high levels of appetitive or goal seeking behaviour, a larger number of stimuli begin to elicit the behaviour, preferences for resources which could satisfy the motivation, or working increasingly hard for access to a resource. However, one complexity with this approach, as we will see, is that some stereotypic behaviour may have motivational consequences (see Chapter 1) that help remedy the underlying frustration. This may then reduce or even abolish any motivational differences that might have been evident before development of the stereotypic behaviour. A second complexity is that it is often hard to disentangle motivational differences from other, non-motivational behavioural differences (e.g. in general activity levels, or in tendencies to perseverate). One example comes from work on foraging, showing that feather pecking negatively correlates with ground-pecking: chickens who performed more ground pecking behaviour, performed less feather pecking, while chickens displaying high levels feather pecking did not perform much ground pecking (Blokhuis, 1986). Another recent study of individual variation comes from Nicol and Badnell-Waters (2005), who looked at the behavioural traits predicting the development of stereotypic oral behaviour, such as crib-biting and wood chewing, in foals after weaning. They found that before weaning, foals that would later develop stereotypic behaviour were characterized by increased suckling and teat nuzzling than foals that did not go on to develop these behaviour patterns, adding to theory that oral stereotypic behaviour may be associated with digestive function. However, piglets that belly nosed (rhythmic up-and-down movement 169 of one piglet with its snout on the belly of another) more after weaning performed less suckling behaviour on the sow compared to other pigs and these piglets tended to be smaller both before and after weaning and grew more slowly after weaning than piglets that did not go on to develop belly nosing, indicating that performance of belly nosing may be associated with a piglets nutritional need (Torrey and Widowski, 2006). Finally, a recent study used individual differences to test and then reject a widely-accepted idea about excessive allo-grooming in laboratory mice. Laboratory mice sometimes pluck out patches from each others' fur, or even pull out the whiskers of their cage-mates - a behaviour termed 'barbering'. Barbering had long been assumed to derive from a form of social grooming in which dominant animals aggressively groom subordinates to assert their status. Garner et al. (2004) tested this hypothesis by using a standardized test to measure the dominance status of barbers and recipients of barbering. They were unable to find any differences in status, suggesting that dominance grooming is not actually the motivational origin of this behaviour. Other studies of individual variation have instead assessed the motivational correlates of stereotypic behaviour by quantifying animals' strength of preferences for particular resources or behavioural opportunities. As mentioned previously, Cooper and Nicol (1991) investigated whether bank voles with stereotypic behaviour find standard cages particularly aversive. Preference tests can be used to determine how important particular resources are to animals (e.g. Duncan and Hughes, 1972; Dawkins, 1977). Cooper and Nicol (1991) therefore offered bank voles a choice of a standard laboratory cage and an enriched cage containing hay at three different ages post weaning, to see if 170 stereotypic and non-stereotypic animals differed. The results, however, were rather surprising. Non-stereotypic voles preferred the enriched environment consistently throughout the experiment and always more than stereotypic voles. Stereotypic voles started off preferring the enriched cage; however, as stereotypic behaviour emerged in these individuals, they actually decreased their preference for the enriched environment! As mentioned previously, Latham and Mason (in prep) measured motivational strength of stereotypic laboratory mice to escape from standard cages into enriched cages compared to less stereotypic mice, using mice that had been reared in either standard or enriched cages. While enriched-reared mice that were moved to standard cages were overall both more stereotypic and more motivated to reach enrichments than standard-raised mice, the results were somewhat more complex (Mason and Latham in prep.). Furthermore, within each rearing group, individual differences in SB did not correlate with individual motivations to escape (at least as measured by this technique). Overall, the results of these two rodent experiments are thus hard to interpret. They may indicate that individual differences in motivation do not translate simply into individual differences in stereotypic behaviour, because other factors like activity levels and routine formation (Mason 1991b) or perseveration (see Chapter 1) also play a role. Additionally, individual differences in preference as measured here are themselves affected by traits other than motivation, such as perseveration or physical strength, which then either act as confounds, and/or once stereotypic behaviours have developed, they have feedback effects that reduce the levels of the causal motivation. This is a difficult problem to avoid, as individual differences in stereotypic behaviour can only be determined after the behaviour develops. One potential, though time-consuming, solution would be to screen a large number of young 171 animals before stereotypic behaviour develops and follow them along through development to determine who becomes stereotypic and who does not. The individual variation in stereotypic behaviour central to the experiments above has allowed the selection and breeding of animals exhibiting differences in their stereotypic behaviour. For example, chickens can be selected for High Feather Pecking (HFP) and Low Feather Pecking (LFP) lines by using number of feather pecking bouts as selection criteria. A significant difference in levels of feather pecking emerged after just three generations of selection (Kjaer et al, 2001). These lines differed in other motivationally-relevant characteristics as well. When given a choice to dustbathe on either sand or feathers, HFP significantly preferred the feathers while the LFP preferred the sand (Johnsen and Vestergaard, 1996): this is consistent with them being motivated to direct dust-bathing-related behaviours at feathers. However, they also spend more time foraging and feeding than LFP hens, consistent with HFP birds also being more motivated to forage (van Hierden et al., 2002), and when given a choice they are more likely to help themselves to loose feathers which they then ingest (Harlander-Matauschek et al, 2006a; 2006b). Harlander-Matauschek et al. (2006a) built on these findings and quantified motivations further by giving birds from these two lines the opportunity to perform an operant response to gain a trough of food, feathers, wood-shavings or "no reward" (an empty trough). They had to peck a key to access each of these, this response experimentally being made harder and harder via an increasing number of pecks required to obtain the reward to assess how strong each bird's motivation was for each resource (cf e.g. Dawkins, 1983; Dawkins, 1990; Mason et al. 2001). The highest number of pecks 172 the birds were willing to make occurred when feed was the reward, and the lines showed no difference here. However, HFP hens worked significantly harder than LFP hens to access feathers in a trough, which they usually then ingested (Harlander-Matauschek et al, 2006a). These data suggest that HFP birds are highly motivated to eat feathers, which is consistent with the foraging hypothesis. Unfortunately, another potential explanation is simply that the HFP birds were more perseverative in their behavior (cf. e.g. Garner, 2006): thus once they had learnt to key-peck for food, they then continued to peck persistently regardless of the reward on offer. Thus to illustrate, they also worked harder than LFP birds even just to access the last resource, the empty trough. Another research approach has been to compare lines, breeds or strains that spontaneously differ in SB (in contrast to lines that have been selected to differ). For example, breeds of horses that have been bred for high levels of activity (Warmbloods and Thoroughbreds, both characterized as being highly energetic and reactive to stimuli) also display higher levels of stereotypic behaviour (Bachmann et al, 2003). This may be due to higher levels of motivation to perform certain behaviour patterns (leading to e.g. more escape behaviour). However, these breed differences could instead stem from differences in husbandry, which would need factoring out (see epidemiology part of 'External stimuli' section); or they may have non-motivational explanations, such as being caused by differences in stamina: thus the higher activity and endurance levels generally associated with Warmbloods and Thoroughbreds could simply allow them to maintain the stereotypic behaviour for longer periods of time. Other similar comparisons have been made for high- and low feather pecking commercially available breeds of hen 173 (e.g. Huber-Eicher and Sebo, 2001; Nicol et ai, 2003), and mouse strains that differ in how stereotypic they are in standard laboratory cages (Nevison et al, 1999b). Dogs of different breeds also vary in the forms of stereotypic behaviour they are most prone to. Dogs that have been bred to show increased vigilance, such as guard dogs, tend to develop self-directed behaviours like flank-sucking or persistent licking, while dogs that have been bred to be highly active, such as herding dogs, being more prone to locomotor behaviours like circle-spinning and tail-chasing (Mills and Luescher, 2006). Again, the type of stereotypic behaviour shown may reflect different underlying motivations, depending on the breed characteristics (e.g. motivations to remain still and vigilant while guarding, or motivations to chase moving objects while hunting) but these ideas are still speculative. A more systematic quantitative comparison of genetically different carnivores, including some canids, was conducted by Clubb & Mason (2003, 2007) — although they looked at multiple species rather than breeds within a species, and focused on a single form of stereotypic behaviour: pacing. They tested competing hypotheses that pacing arises from frustrated hunting, ranging, patrolling or general activity, by comparing how much repetitive pacing each of 33 Caraivora species showed when captive with aspects of their behavioural biology in the wild (especially the style and time-budgets of their natural foraging/hunting behaviour, their home range sizes, their average daily travel distances, and territoriality). Clubb and Mason found that the amount of stereotypic pacing was significantly predicted by high levels of natural ranging (both large home- 174 range sizes and great daily travel distances), but not by foraging or hunting behaviour. These findings once again suggest that frustrated hunting is not the motivational origin of pacing. However, why home range size and travel distances are better predictors remains unclear without further data: are these species more motivated to travel, and thus more frustrated by being confined? Or - as in our horse breeds example above - are there non- motivational explanations? For instance, do naturally wide-ranging species instead have naturally high physical endurance, such that they are then just able to perform stereotypic behaviours for longer? Or are they rendered more stereotypic by captivity? (See Clubb & Vickery 2006, Clubb & Mason 2007, for further discussion). This powerful technique could probably be successfully applied to other species comparisons. For example, laboratory rodents tend to be housed in similar ways, but while about 50% of mice show stereotypic behaviour, rats rarely do - and other differences occur across other caged rodents species. What causes this variation? Comparisons of differences in natural behavioural biology could help reveal the risk factors (cf. Wiirbel, 2006) (see also Novak, Box 3.3 in Clubb and Vickery, 2006) on species differences across captive primates.). Similarly, it could be used to statistically identify the reasons for differences in SB between strains, breeds and sub-species and strains (Nevison, 1999b). Like epidemiological approaches, these techniques have the advantages of using pre-existing variation, which may allow large data-sets to be built up for relatively little cost and also removes the need to house additional animals in poor conditions in order to research their stereotypic behaviour. However, also like epidemiological approaches, false results may be caused by confounding variables if care 175 is not taken (e.g. types of husbandry that systematically co-vary with genotype), and these techniques only identify general risk factors — they do not necessarily distinguish motivational factors from other influential biological differences. To specifically test motivational hypotheses one would need more ethological data, such as on species differences in preference. Finally, to conduct such studies well, data from a large number of species or strains are required, and when traits are shared due to the close relatedness of sister species ('phylogenetic signal': e.g. Blomberg et al, 2003; see also Clubb and Mason, 2004), this needs to be controlled for to ensure that species means in an analysis are statistically independent (see e.g. Harvey and Pagel, 1991 in Clubb and Mason, 2004). 4. FINAL CONCLUSIONS Our aim for this paper was to review the various techniques available and here we summarize these techniques into a final recommendation. No method is without problems and there has yet to be one ideal, agreed-upon technique - but we argue that carefully selecting 2 or 3 complementary approaches should generally yield strong tests of the motivational hypotheses being investigated. Thus if a combination of techniques is used, agreement from the results of these techniques will lend the most support to the motivational hypothesis being tested (indeed as in this thesis). In contrast, disagreement of results may indicate more than one motivation is behind the behaviour. 176 4.1. Review of pros and cons So, let us give an overview of the pros and cons of the different techniques available, before moving onto more specific recommendations. First, we will briefly summarize the pros and cons of observing form and timing. Observing the form or morphology of the stereotypic behaviour, in a non-quantitative way, is relatively easy and can help generate or even reject hypotheses on the source behaviour. However, if several behaviour patterns have the same form or the behaviour pattern alters over time (e.g. become abbreviated and/or more complex), this measure will not be useful. Using the form of FAPs provides a quantitative method that allows for proper testing of competing motivational hypotheses. This approach can be quite difficult - high quality equipment is needed to record the behaviour patterns at high resolution and with controlled viewing angles; data collection from the videos is time intensive; and testing for similarities between SB and normal behaviour patterns means the sample size must be large enough to ensure high statistical power and decrease the risk of Type II errors. Timing measurements of SB compared to normal behaviour is another relatively simple and inexpensive method that can be used to generate/test hypotheses. The behaviour should be recorded for the entire 24-hour period, as it is possible that the SB occurs at a certain time in a daily rhythm because that is when the majority of behaviour is performed and does not actually give an indication of underlying motivation. Other factors underlying the circadian/ultradian rhythm still need to be determined as well, i.e. what is causing the behaviour to occur at that particular time? Finally, several behaviour patterns may share similar timing, 177 making it potentially difficult to determine which behaviour is actually the source behaviour for the SB. Examining/manipulating external factors can also help determine which behaviour patterns are most motivated, and how strongly. Environmental risk factors can be identified through epidemiological approaches - which have the advantage of using pre existing variation, allowing for large data-sets to be built up for relatively little cost, but these identified factors may not always act by motivational means (e.g. disease may be a risk factor for certain SBs), nor do they take time budgets into account; thus certain environmental factors may appear related to a decrease in SB but this decrease is do to non-motivational factors. Epidemiological studies are also limited by the variation already occurring across different husbandry systems and they only reveal what correlates with the emergence of SB, not what causes the emergence. Experimentally manipulating external factors can help avoid many of these problems. Stimuli thought to relate to unfulfilled motivation can be added, removed or changed and the effect of this on SB can be measured. These manipulations can have practical advantages in that they can be used to devise the best methods to decrease/abolish the SB, in addition to testing competing motivational hypotheses. However, assumptions are being made about what motivation is being satisfied by the resource that may not be accurate; in fact the enrichment may decrease SB but due to non-motivational effects, such as taking up an animal's time or even by reducing general stress. It is also possible that as the SB changes over time, they become less dependent on external environmental factors, thus stimuli which may once have affected performance of the SB, may no longer be as 178 effective. A combination of physiological and behaviour data can add strong evidence to a motivation hypothesis; however, this has been done rather indirectly, and inferences regarding the internal stimuli are made, which may be incorrect. In addition, some measures need to be made carefully, in relation to behavioural data - for example, corticoids can be measured as an indicator of stress but also may be elevated when the animal is excited (e.g. when mating). Similar to an epidemiological approach, determining traits that predict or accompany SB can exploit naturally occurring individual or strain/species variation, or instead generate/enhance that variation through differential rearing or genetic selection. However, some stereotypic behaviour may have motivational consequences that help remedy the underlying frustration. This may then reduce or even abolish any motivational differences that might have been evident before development of the stereotypic behaviour and it is often hard to disentangle motivational differences from other, non-motivational differences. In addition, false results may be caused by confounding variables if care is not taken (e.g. animal husbandry may always be done in a certain way with certain species or strains). 4.2. Recommendations i. Selecting two-three appropriate, complementary techniques: Given the above, our first broad methodological pointer is that the various pros and cons of each method should be considered, before choosing two or three to use. One 179 should select methods that are not prone to the same confounds, i.e. that are subject to complementary problems rather than the same problems. Choice should also be based on the type of stereotypic behaviour it is suspected one is dealing with. To illustrate, only measuring the form of the SB and manipulating external factors to determine the source behaviour is not advisable, since both have the potential to be affected by changes in the stereotypic behaviour over time (i.e. abbreviation of the SB and the SB becoming less dependant on external factors over time). An approach that does not suffer from this problem should be used instead of (or in addition to) one of these methods, such as determining the risk factors involved in developing the SB. Different research approaches are also differentially suited to SBs with different motivational explanations. As discussed in the Introduction, there seem to be three types of motivational explanation for SB (Mason, 2006): that they derive from redirected or vacuum activities performed by frustrated or thwarted animals; that they derive from escape attempts shown by frustrated or thwarted animals; and/or that housing or rearing animals without motivationally significant stimuli causes stress and/or abnormal CNS development which alter how behavior is controlled. This indicates that different research approaches will yield more accurate results for different types of stereotypic behaviour. In the first group, we can learn a lot from a stereotypic behaviour's form and location, because the SB's source behaviour also reflects the causal frustration (Mason, 2006). For stereotypic behaviour of the second group, an SB's form and location, would tell us the animal wants to escape, but not why, that is, the source behaviour of the SB will not reveal the original cause of frustration. For SBs of the third group, motivational 180 frustration will have lead to stereotypic behavior only indirectly, and perhaps a long time earlier in the animals' developmental history; furthermore, in animals with SBs of this type, there may even be two motivational explanations, the first accounting for what made them perseverative, the second accounting for the type of SB they currently perform. Perseveration may also be a confound we need to control for, as discussed further below. ii. Dealing with non-motivational confounds: Our second methodological recommendation is to always consider - and ideally address experimentally — alternative, non-motivational mechanisms involved in SB, such as (to give just three examples from our review) general activity, time budgets and perseveration. Thus factors that increase or decrease general activity or arousal may affect SB in a very non-specific way that says little about specific motivation. This could be an issue when, for example, using 'timing' to identify the motivational basis of SB: using circadian rhythms alone to identify the motivational basis of stereotypic behaviour could be too simplistic, as discussed above, it is possible that most behaviour happens at these certain times, thus the fact that the SB also occurs then is not very revealing. Additionally, if SB is most likely when animals are most awake and active, then this could also act as a confound when using the comparisons of individuals, strains/breed and species to test motivational ideas. General activity changes could also be an issue in some environmental enrichment experiments; enrichments which paradoxically increase SBs 181 often seem to be ones which make animals more active (e.g. Korhonen et ah, 2001). Finally, this may also be an issue in some research using pharmacological compounds, e.g. if a compound presented as an anxiolytic also acts as a sedative, and thus reduces general activity. Solutions include controlling for activity during experiments (e.g. using it as a covariate in statistical models, or expressing SB as a % of total active time); and/or running supplementary studies to check it is not the cause of any results. Another, related non-motivational confound is time-budget. Many environmental enrichments applied experimentally, or aspects of housing/husbandry identified as important in epidemiological work, may impact on SB by increasing or decreasing the amount of 'spare time' an animal has, rather than by affecting the motivation underlying the SB. Thus when using enrichments experimentally this needs to be taken into account (however, this will be difficult in epidemiological approaches where the animals are not actually observed). Solutions include subtracting the amount of time spent interacting with an enrichment before calculating time-budgets devoted to SB (Vickery, 2006); or using different categories of enrichments designed to address different motivations, with multiple exemplars from each category (not just one of each 'motivational type'), with the exemplars in each category addressing the same motivation but differing in the amount of time they take. This avoids pseudo-replication (see fourth recommendation) and allows motivational effects to be teased out from mere time budget ones. The third important confound is perseveration. If stereotypic individuals are more perseverative, this will affect a number of different aspects of the behaviours being 182 measured to research motivation. Different husbandry practices could lead to different levels of SB, and different individuals, strains and species may be more or less vulnerable to SB, not for motivational reasons but because they involve or cause different degrees of perseveration. Furthermore, in experiments, perseverant animals may be slow to react to manipulations - thus slow, for example, to respond to changes in where cues enter a cage; slow to respond to the movement of enrichments between locations; and slow to respond to the presentation of enrichments (see Mason et ah, 2007 for further discussion). This may make the initial findings of an experiment look as though they are rejecting certain motivational explanations, when in fact in given more time and experience with these changes, the animals would react and alter their SBs (see Mason et ah, 2007 for further discussion). Perseveration is also an important potential confound to consider when assessing strength of motivation in operant and similar tasks: because perseverant animals are slow to modify responses, they may be slow to react to increasing costs (thus looking more motivated than they truly are), and slow to react to changes in the presented reward (e.g. see criticism of Harlander-Matauschek et ah, 2006a). Finally, as discussed in the Introduction, perseverant animals may show SB in response to the immediate presentation of motivationally significant stimuli (e.g. appearance of vets!), but be perseverant (the risk factor for SB) for past reasons that cannot be ascertained now (see third recommendation) or in ways that can be gradually reduced by any factors that reduce stress. Solutions to account for perseveration include presenting the various costs and rewards in different orders to different animals in an experiment, so that the order is not the same for all, and/or assessing subject's perseverative tendencies in experiments designed to do so (Garner, 2006; Mason, 2006), so that they can then be factored out 183 statistically (e.g. used as covariates). An alternative solution to deal with non- motivational factors is, as suggested in our first recommendation, to run a complementary investigation of the motivational hypothesis under test using a technique that does not suffer from this potential problem (e.g. assessing strength of preference; analyzing SB FAP morphology for comparisons with normal behaviour). iii. The benefits of working with young, even pre-stereotvpic, animals: Our third recommendation is to work with young animals that have not yet developed SBs and/or whose SBs are still 'young'. If SBs have effects that feedback and reduce frustration, this - as we have argued - could reduce our abilities to tell what the underlying motivation was that originally caused the behaviour to develop. This suggests that looking for motivational precursors of SB development, in young animals, could be more helpful than looking at motivational correlates in individuals who already have the behaviour. The use of families, strains/breeds and/or species already known to develop high/low levels of SB could facilitate collecting data from young animals inherently likely to develop high/low SB. Furthermore, if SBs have properties that change over time (a possibility we have raised at several times in this review), then looking at SBs early in their development might well have many advantages over studying them in older animals that have been performing them for a long time. 184 iv. Avoiding pseudo-replication: Our fourth recommendation is to avoid pseudo-replication of resources/stimuli/subjects. Pseudo-replication is the treating of non-independent data- points as if they are independent and thus truly sample a larger population. One form of pseudoreplication is the use of one exemplar from a category (assessed multiple times) as if it represents multiple exemplars from that category: an error that often occurs when single enrichment items are given labels as if they represent a broad class of enrichment item (e.g. one single toy is used to assess the value of toys). This is a problem because if only a single exemplar from a category of stimulus is used to test motivational hypotheses about the category itself, it is not clear if results found reflect only that particular exemplar or if they indeed reflect the broader category of stimulus (e.g. Kroodsma et ah, 2001). For example, if straw was the only factor manipulated to test the hypothesis that feather pecking stems from redirected foraging, the results found may only be applicable to straw and would not necessarily be representative of all forages. Thus, a number of different exemplars from each motivational category must be used whenever stimuli/resources are needed to test a hypothesis — such as when using approaches designed to measure FAPs, the effectiveness of enrichments, or assess preferences. This will ensure that the results obtained are not due to idiosyncratic properties of the resource that are not important (e.g. its shape, size, texture, its ability to take up time, etc) and are instead due to the motivation of interest. 185 A similar problem arguably occurs when a single High Feather Pecking (HFP) and a single Low Feather Pecking (LFP) strain are compared (for example): differences between them might indeed reflect general differences between all possible HFP birds and all possible LFP birds, but instead they could just reflect differences between these two particular strains, with no broader significance beyond that. Solving this problem would this involve using multiple strains or species which vary in the behaviour of interest, e.g. multiple HFP lines and multiple LFP lines. If this approach is taken, however, an important point to then consider is the relatedness of animals within a species, sometimes called phylogenetic signal (Blomberg et ah, 2003), or similarities in how animals are housed and managed in captivity compared to animals from different species. If these are not considered, calculated strain or species values may not be statistically independent, again leading to pseudo-replication (Harvey and Pagel, 1991 in Clubb and Mason, 2004). 4.3. In conclusion In conclusion, the presence of stereotypic behaviour has welfare consequences and is a good indication that we are not housing and managing animals in an acceptable manner. Animals in captivity are most likely highly motivated to perform behaviours that are not possible and may also be experiencing negative emotions (frustration, stress) as a result. As stated in Chapter 1, to completely solve the problems of stereotypic behaviour, we must determine the underlying motivation or we risk applying 'band-aid' solutions, which may only work through distraction or increases in opportunity for activity and not 186 by actually solving the problem. As there are still many stereotypic behaviour patterns whose underlying motivations are unknown, further research into determining why, on a motivation basis, stereotypic behaviours occur and improving the welfare of these animals is important. 187 CHAPTER SEVEN GENERAL DISCUSSION AND CONCLUSIONS 1. Discussion of thesis results The overall objective of this thesis was to determine the underlying motivation behind feather pecking behaviour that is commonly found in laying hens. The two main hypotheses explaining the motivational basis of feather pecking involve re-directed foraging and re-directed dustbathing behaviour. However, no definitive conclusion has yet been reached, owing partly to the difficulties involved in distinguishing between foraging and dustbathing ground pecking, a measure commonly used in the motivational analysis of feather pecking. This difficulty was overcome by using past techniques designed to assess the motivation of stereotypic behaviour (Chapter 6) and by designing a novel method to assess motivation that solves the problem of determining the motivation involved in ground pecking (Chapter 5). The first objective of this thesis was to determine the importance of rearing conditions and current environment on feather pecking in early life. There has been conflicting evidence on the effects of early experience on later feather pecking behaviour. Some studies have found early experience with an appropriate litter substrate to decrease later feather pecking compared to birds that have only experienced wire flooring, while others have found that early experience is less important than the current environment 188 (e.g. Vestergaard et al, 1993; Nicol et al, 2001; Newberry et al, 2007). This may be due to a variety of factors, such as type of housing (floor or cage), type of substrate used, age of bird when moved between environmental conditions (e.g. Johsen et al, 1998; Nicol et al, 2001) and, since feather pecking has a genetic component, the strain of bird used (e.g. Rodenburg et al, 2003). Additionally, the immediate effects of changing environments have not been determined, and since it is common practice with commercial laying hens to move chicks between environmental conditions early in life, the effects of this change on the chicks is another important aspect to consider. This research project used two common strains of commercial egg layers, so the results not only help determine if rearing conditions and environmental changes influence the welfare of commercial laying hen chicks but also indicate the appropriate rearing and housing conditions needed for accurate feather pecking results with these strains. It was found that current environment influenced feather pecking the most, with no immediate significant adverse effects of housing shifts. Thus, for these strains of laying hen chicks, rearing on a substrate should be encouraged since their immediate welfare will be improved during this time. The second objective of this thesis was to determine the diurnal rhythm of feather pecking and compare this to dustbathing and foraging rhythms. Feather pecking can be observed from day one of life (e.g. Savory and Mann, 1999; Chow and Hogan, 2005) but previous research into the diurnal distribution of feather pecking has only been done with older birds that have already begun to lay (Preston, 1987; Kjaer, 2000; Bright, 2007). Preston (1987) and Kjaer (2000) found an increase in feather pecking in the later part of 189 the day, while Bright (2007) found an increase in severe feather pecking early in the day; thus, it was not clear whether the results would be similar for younger birds, who do not have the hormones associated with nest building and egg laying behaviour influencing their feather pecking behaviour and who may not yet have become perseverative in their behaviour. It was found that chicks up to a month in age evenly distributed their feather pecking, foraging and feeding behaviour throughout the day, while dustbathing mainly occurred 4-7 hours after lights on. Since foraging was likewise evenly spread over the day, this suggests that feather pecking is related to re-directed foraging and not re directed dustbathing behaviour. The third objective of this study was to determine the effects of different types of enrichment on feather pecking behaviour. Past studies have used a variety of enrichment objects with different housing systems and ages and strains of birds (e.g. Norgaard- Nielsen 1997; Aerni et al. 2000; McAdie et al. 2005), making comparison between results difficult. A Latin Square design was implemented using forage, dustbaths, novel objects and no enrichment to determine their effects on feather pecking and their relative use by the birds. Whereas a similar number of pecks were directed at forages and dustbaths and these were more than at the novel object, there was less feather pecking when the forage was provided than in the other treatments. This adds additional evidence to the hypothesis that feather pecking stems from re-directed foraging behaviour. However, as these were count data and not durations, it may be that the forages actually occupied more of the birds' time, leaving less time available to feather peck. Foraging is time consuming behaviour, since birds will forage about 60% of the daylight hours 190 (Dawkins, 1989), whereas, dustbathing only occupies about half an hour every other day (Vestergaard, 1982a). Therefore, birds given the opportunity to forage would have less time available to feather peck. Future research on the duration of time spent attending to the enrichment objects may help to clarify this issue. Nonetheless, it still indicates that developing practical forages may be the best method to control feather pecking behaviour. The fourth and fifth objectives of this research project were to validate a new method of studying stereotypic behaviour patterns and apply this method to uncover the underlying motivation behind feather pecking. A number of different behaviour patterns have been shown to have Fixed Action Patterns (FAPs), such as egg rolling in grayleg geese (Lorenz and Tinbergen, 1938) and grooming in mice (Berridge, 1990). Work with pigeons has shown that the motivation behind a behaviour pattern influences the motor patterns involved (Jenkins and Moore, 1973). Using methods adapted from this past research, it was found that chickens peck at various stimuli with repeated, consistent motor patterns, with very low variation within measures and within and between birds, thus indicating FAPs in their pecking behaviour. It was found that severe feather pecks and foraging pecks had similar motor patterns for all measures, while dustbathing pecks and severe feather pecks were different. In addition, ground pecks given while foraging and ground pecks given during a dustbath were different. This adds strong evidence to the hypothesis that feather pecking stems from re-directed foraging and not from re-directed dustbathing. Gentle 191 feather pecks were generally different from the other pecks measured and thus appear to be motivated differently from severe feather pecks. It is possible that gentle feather pecks represent a form of allo-preening, shown by other types of birds (e.g. Zebra Finches: Adkins-Regan and Robinson, 1993; Arabian Babblers: Pozis-Francois et al., 2004) and occasionally by hens towards roosters. Future research examining and comparing the FAPs in other types of pecking and preening behaviour may help clarify this issue. 2. Do foraging enrichments always abolish feather pecking? A problem that remains regarding feather pecking is that some birds given the opportunity to fulfill foraging motivation, by housing them on litter (usually wood shavings) or even with access to outdoor runs, still feather peck (e.g. Chapter 4). In addition, the average amount of feather pecking can be decreased by the provision of other enrichments (e.g. string to peck, novel objects), at least in some birds (e.g. McAdie et al., 2005; Chapter 4), but again feather pecking is not abolished. There are a few possible reasons for this. It is possible that the enrichment provided doesn't completely satisfy the motivation (i.e. the enrichment is an inadequate substitute), thus frustration continues as does performance of the stereotypic behaviour (Swaisgood and Shepherd, 2006). Another possibility is that the feather pecking shown in these housing conditions may be gentle in nature and possibly part of a normal allo-preening behaviour. Not all trials have distinguished between gentle and severe feather pecks (e.g. Kjaer, 2000) and thus reports of feather pecking on litter-housed or free range/free run birds may be a 192 product of this. It has also been shown that the quality of a substrate, determined by the amount of pecking and scratching directed to it, affects the amount of feather pecking shown. Higher quality substrates, substrates that have an increased amount of time spent with them (more pecking and scratching) compared to a lower quality substrate, also have less feather pecking associated with them compared to lower quality substrates (Huber- Eicher and Wechsler, 1998). It is possible that the birds that have been observed feather pecking while housed on substrates were kept with a low quality variety, and thus foraging motivation may not be satisfied. Finally it is possible that as the birds age, feather pecking becomes perseverative or habit-like and continues even if the appropriate stimulus is provided to release the behaviour (e.g. foraging motivation) or if the birds are housed in near-to-nature conditions, such as free range (e.g. Swaisgood and Shepherd, 2006). Commercial poultry housing can be quite stressful for the birds and aside from foraging, a number of 'natural' behaviour patterns, such as nesting, often cannot be performed (Duncan, 2001a). This chronic stress may lead to a pathological condition or even a change in the brain of the birds (Garner, 2005), whereby feather pecking cannot be controlled through fulfillment of its original motivation. The degree of perseveration shown can be affected by a number of factors, such as stress and early social environment. As chickens, even those kept in free range situations, rarely experience natural early social conditions, that is, being brooded by a hen, and together with their siblings, learning about the environment from her, this may help explain reports of feather pecking in more enriched conditions (Mason, 2006 Box 6.2; Riber et ai, 2007). 193 Another reason for the inability to abolish feather pecking which is commonly found in past research may be due to differences in genetics. Feather pecking has a genetic component (Cuthbertson, 1980) and different strains of chickens have been shown to perform different levels of feather pecking (Hughes and Duncan, 1972). This is demonstrated in Chapter 2, where ISA White chicks feather peck more than ISA Browns. There have not been behavioural profiles done comparing these strains in terms of general activity levels. If some strains of birds tend to forage more than others, they may have a higher motivation to forage and when this gets frustrated, they may feather peck more than birds with lower foraging motivation. Alternatively, if some birds spend more time foraging than others, this may take up more of their time, leaving less time available to feather peck. There may also be genetic or even individual differences in what constitutes a 'good quality' forage, i.e. differences in individual perception. Thus, substrates that satisfy foraging motivation in some strains, or in some individuals, may not be as satisfying for others. There is fairly clear evidence that the provision of some type of foraging enrichment decreases feather pecking behaviour (e.g. Huber-Eicher and Wechsler, 1997; Chow and Hogan, 2005; Chapter 4), however, the decrease may vary between strains. These genetic differences may also help explain the conflicting results on the effects of early experience on later feather pecking. Some strains of birds may benefit from proper early substrate experience and/or being brooded by the mother hens and this may have lasting effects in some strains, while in others it may not. Moreover, what constitutes as 'proper early experience' may differ between strains or even individuals. 194 Based on the many known differences between different strains of birds, or even the differences between individuals in one strain, it is possible that some birds may feather peck primarily due to frustrated foraging motivation while others may feather peck primarily due to chronic stress and potential changes in CNS function. It has been demonstrated that birds showing high levels of feather pecking (HFP) have lower serotonin turnover than those selected for low levels of feather pecking (LFP) (van Hierden et al., 2005). Low levels of serotonin turnover have also been implicated in other behavioural disorders (in humans and animals), such as depression and obsessive compulsive disorder, and supplementation with tryptophan (which leads to increased levels of serotonin) has been shown to improve these conditions (e.g. Sandyk, 1992). Thus, there appears to be changes in brain function between these birds. 3. Future research on hens - both applied and fundamental Future applied work should aim to address some of the questions raised above. Multi-factorial experiments can be designed that use different strains of birds on an individual basis to investigate i) the effects of early experience with and without a certain substrate, ii) the effects of being brooded and reared by the mother hen on later feather pecking levels, iii) later housing conditions with or without the substrate (does substrate influence feather pecking if given later in life) and iv) the types of substrates that reduce feather pecking the most. Answering these questions will help determine if designing an effective, practical forage will need to be done on a strain basis or if certain early 195 experiences need to be ensured to decrease the risk of perseveration of the behaviour. These forages could be as simple as scattered feed on the ground in floor systems to increase opportunities to forage or in cage systems, 'hay bags', like those used for horses, can be hung along the sides of the cages and filled with high fibre forages, which again increase foraging opportunities while also potentially improving gut function due to increased fibre intake. The use of strains exhibiting lower levels of feather pecking, such as the ISA Browns (Chapter 2) should also be encouraged and these strains could be analyzed to see which foundation lines have the desirable trait in an attempt to genetically decrease feather pecking. As discussed in Chapter 5, a future definitive test for the use of FAPs would be to require poultry to peck operant buttons to reach different types of reward, including a stimulus to feather peck. The morphology of this peck could then be compared to a completely standardized stimulus, as has been done in past pigeon research (e.g. Jenkins and Moore, 1973). In addition, continued research into the differences in brain function of feather peckers and non-feather peckers, such as differences in neurotransmitters, turnover time and reuptake rates, may add evidence to the idea that feather pecking is similar to behavioural disorders in humans. These potential differences in brain function may help explain why feather pecking can be decreased but not completely abolished in some birds. Birds differing in their feather pecking levels can also be tested for their degree of perseveration. For example, animal that have become perseverative will continue performing a behaviour in the absence of the appropriate stimuli. Thus, a measure of perseveration could be the time it takes individuals to stop responding to a 196 previously rewarded behaviour, as perseverative individuals should take longer to stop this behaviour (e.g. Garner, 2006). Additionally, the effects of early environment, such as substrate access or being brooded by the mother hen can be manipulated and the effects on perseveration measured. These results should help determine the best way to house chickens in order to reduce the risk of their behaviour becoming perseverative and thus, their feather pecking behaviour harder to abolish. 4. Future suggestions for similar behaviour in other species This insight into the motivation behind feather pecking lends ideas to the motivation behind other types of stereotypic plumage or pelage removal or abnormal oral behaviour. Animals who have evolved to have similar digestive requirements as poultry, that is high fibre/forage diets, and are fed concentrates in captivity, may remove and consume pelage for similar reasons as chickens or may perform other stereotypic oral behaviour to help buffer high levels of stomach acid that would not normally be a problem because of forage intake (cf. Bergeron et al, 2006). Again, similar to the methods used in chickens, the provision of different types of forages to these species, taking into account breed/strain or even individual differences and early experience may help reduce the incidence of stereotypic behaviour. The methods used in Chapter 5 could easily be applied to other birds that are known to feather peck (e.g. turkeys, ducks, parrots) to determine definitively if their underlying feather pecking motivation is the same as chickens; although it does appear to 197 be, since factors similar to those in chickens influence feather pecking in these species (e.g. Martrenchar et al, 2001; Meehan et al, 2003). The motivation behind other stereotypic behaviour patterns could also be determined with these techniques. For example, other types of pelage removal, such as barbering in mice or wool removal in sheep, could be video recorded and the elements making up the behaviour compared to other normal behaviour patterns, such as grooming in mice (which is already very well documented) or foraging in sheep. Other stereotypic behaviour, such as tongue rolling in cattle or pacing in carnivores, could also be studied using the FAPs involved in the stereotypic behaviour and comparing them to normal behaviour patterns, such as grazing behaviour for cattle or different types of locomotory behaviour (e.g. foraging, mate searching, ranging) in carnivores. Finally differences in brain function and perseveration of behaviour can be investigated. As mentioned previously, these results should help determine why the stereotypic behaviour is so hard to stop in some individuals and what early experiences may be needed to decrease risk of perseveration. 5. Final conclusions In conclusion, the results of this research support the hypothesis that feather pecking stems from re-directed foraging behaviour and not re-directed dustbathing behaviour. Feather pecks have a similar diurnal rhythm to foraging and both are primarily influenced by current environment. Severe feather pecks are morphologically similar to foraging ground pecks but are different from dustbathing ground pecks. Finally, the provision of forages decreased feather pecking more than the provision of a 198 dustbath, even though both received a similar number of pecks. More generally, the use of Fixed Action Patterns can help distinguish specific behaviour patterns that, although superficially similar in phenotype, are actually quite distinct, such as dustbathing and foraging pecks. Lastly, the study of FAPs is a useful tool that can help elucidate the motivation behind puzzling stereotypic behaviour patterns found in captive animals. 199 REFERENCES Adkins-Regan, E.; Robinson, T.M. 1993. Sex differences in aggressive behavior in Zebra Finches (Poephila guttata). Journal of Comparative Psychology, 107: 223-229. Aerni, V.; El-Lethey,H.; Wechsler,B. 2000. Effect of foraging material and food form on feather pecking in laying hens. British Poultry Science, 41: 22-28. Aerni, V.; Brinkhof, M.W.G.; Wechsler, B.; Oester, H.; Frohlich, E. 2005. Productivity and mortality of laying hens in aviaries: a systematic review. World's Poultry Science, 61:130-142. Ainslie, G. 1975. Specious reward: A behavioural theory of impulsiveness and impulse control. Psychological Bulletin, 82: 463-496. Allen, J.; Perry, G.C. 1975. Feather pecking and cannibalism in a caged layer flock. British Poultry Science, 16: 441 -451. Anderrsson 2001 Appleby, M.C.; Hughes, B.O. 1991. Welfare of laying hens in cages and alternative systems: environmental, physical and behavioural aspects. World's Poultry Science, 47: 109-126. 200 Appleby, M.C.; Lawrence, A.B. 1987. Food restriction as a cause of stereotypic behaviour in tethered gilts. Animal Production, 45: 103-110. Appleby, M.C.; Hughes, B.O.; Elson, H.A. 1992. Poultry Production Systems: Behaviour, Management and Welfare. CAB International, Wallingford, Oxon, UK pp.150-153. Appleby, M.C.; Mench, J.A.; Hughes, B.O. 2004. Poultry Behaviour and Welfare. CABI Publishing, Cambridge, Massachusetts. Appleby, M.C.; Walker, A.W.; Nicol, C.J.; Lindberg, A.C.; Freire, R.; Hughes, B.O.; Elson, H.A. 2002. Development of furnished cages for laying hens. British Poultry Science, 43: 489-500. Arias-Carrion, O.; Poppel, E. 2007. Dopamine, learning and reward-seeking behaviour. Acta Neurobiologiae Experimentalis, 67: 481-488. Bachmann, I.; Audige, L.; Stauffacher, M. 2003. Risk factors associated with behavioural disorders of crib-biting, weaving and box-walking in Swiss horses. Equine Veterinary Journal, 35: 158-163. 201 Barlow, G.W. 1968. Ethological units of behavior. In: The Central Nervous System and Fish Behavior (Ed. by D. Ingle), pp. 217-232. Chicago: University Press of Chicago. Bashaw, M.J.; Tarou, L.R.; Maki, T.S.; Maple, T.L. 2001. A survey assessment of variables relating to stereotypy in captive giraffes and okapi. Applied Animal Behaviour Science, 73: 235-247. Bateson, M. 2004. Mechanisms of decision-making and the interpretation of choice tests. Animal Welfare, 13: SI 15-120. Baxter, M.R. 1994. The welfare problems of laying hens in battery cages. Veterinary Record, 134: 614-619. Bearse, G.E., Miller, V.L. and McClary, C.F. 1940. The cannibalism preventing properties of fibre fraction of oat hulls. Poultry Science, 19: 210-215. Bechara, A.; van der Kooy, D. 1992. A single brain substrate mediates the motivational effects of both opiates and food in non-deprived but not in deprived animals. Behavioural Neuroscience, 106:351-363. Bergeron, R.; Badnell-Waters, A.J.; Lambton, S.; Mason, G. 2006. Stereotypic oral behaviour in captive ungulates: foraging, diet and gastrointestinal function. In: 202 Stereotypic Animal Behaviour, 2nd edn. G. Mason and J. Rushen, eds. CAB International, Cambridge, pp. 58-85. Beradtson, W.E. 1991. A simple, rapid and reliable method for selecting or assessing the number of replicates for animal experiments. Journal of Animal Science, 69: 67-76. Berridge, K.C. 1990. Comparative fine-structure of action-rules of form and sequence in the grooming patterns of 6 rodent species. Behaviour, 113: 21-56. Berridge, K.C. 1996. Food reward: brain substrates of liking and wanting. Neuroscience and Biobehavioural Reviews, 20: 1 -5. Berridge, K.C. 2004. Motivational concepts in behavioural neuroscience. Physiology and Behaviour, 81: 179-209. Berridge, K.C; Valenstein, E.S. 1991. What psychological process mediates feeding evoked by electrical stimulation of the lateral hypothalamus? Behavioural Neuroscience, 105:3-14. Bilcik, B.; Keeling, L.J. 2000. Relationship between feather pecking and ground pecking in laying hens and the effect of group size. Applied Animal Behaviour Science, 68: 55-66. 203 Blokhuis, H.J. 1986. Feather-pecking in poultry: its relation with ground pecking. Applied Animal Behaviour Science, 16: 63-67. Blokhuis, H.J. 1989. The effect of a sudden change in floor type on pecking behaviour in chicks. Applied Animal Behaviour Science, 22: 65-73. Blokhuis, H.J.; Arkes, J.G. 1984. Some observations on the development of feather- pecking in poultry. Applied Animal Behaviour Science, 12: 145-157. Blokhuis, H.J.; van der Haar, J.W. 1989. Effects of floor type during rearing and of beak trimming on ground pecking and feather pecking in laying hens. Applied Animal Behaviour Science, 22: 359-369. Blokhuis, H.J.; Wiepkema, P.R. 1998. Studies of feather pecking in poultry. Veterinary Quarterly, 20: 6-9. Blomberg, S.P.; Garland, T.; Ives, A.R. 2003. Testing for phylogenetic signal in comparative data: behavioural traits are more labile. Evolution, 57: 717-745. Bolles, R.C. 1970. Species-specific defense reactions and avoidance learning. Psychological Review, 77: 32-48. 204 Breland K.; Breland, M. 1961. The mis-behaviour of organisms. American Pyschologist, 16: 681-684. Breward, J.; Gentle, M.J. 1985. Neuroma formation and abnormal afferent nerve discharges after partial beak amputation (beak trimming) in poultry. Experientia, 41: 1132-1134. Buitenhuis, A.J.; Rodenburg, T.B.; van Hierden, Y.M.; Siwek, M.; Cornelissen, S.J.B.; Nieuwland, M.G.B.; Crooijmans, R.P.M.A.; Groenen, M.A.M.; Koene, P.; Korte, S.M.; Bovenhuis, H.; van der Poel, J.J. 2003. Mapping Quantitative Trait Loci Affecting Feather Pecking Behavior and Stress Response in Laying Hens. Poultry Science, 82: 1215-1222. Buitenhuis, A.J.; Rodenburg, T.B.; Wissink, P.H.; Visscher, J.; Koene, P.; Bovenhuis, H.; Ducro, B.J.; van der Poel, J.J. 2004. Genetic and phenotypic correlations between feather pecking behavior, stress response, immune response, and egg quality traits in laying hens. Poultry Science, 83: 1077-1082. Buitenhuis, A.J; Kjaer, J.B.; Labouriau, R.; Juul-Madsen, H.R. 2006. Altered circulating levels of serotonin and immunological changes in laying hens divergently selected for feather pecking behaviour. Poultry Science, 85: 1722-1728. 205 CARC (Canadian Agri-Food Research Council). 2003. Recommended code of practice for the care and handling of pullets, layers and spent fowl. Canadian Agri-Food Research Council, Ottawa, Ontario. Cabanac, M. 1979. Sensory pleasure. The Quarterly Review ofBiology, 54: 1-29. Cabib, S. 2006. The neurobiology of stereotypy II: the role of stress. In: Stereotypic Animal Behaviour, 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 227-255. Carlstead, K. 1991. Husbandry of the fennec fox Fennecus zerda: environmental conditions influences stereotypic behaviour. International Zoo Yearbook, 30: 202-207. Carlstead, K. 1998. Determining the causes of stereotypic behaviours in zoo carnivores: towards developing appropriate enrichment. In: Second Nature: Environmental Enrichment for Captive Mammals. D.J. Sheppardson, J.Mellen and M. Hutchins, eds. Smithsonian Institution Press, Washington. Chow, A.; Hogan, J.A. 2005. The development of feather pecking in Burmese red junglefowl: the influence of early experience with exploratory-rich environments. Applied Animal Behaviour Science, 93: 283-294. 206 Cloutier, S.; Newberry, R.C.; Forster, C.T.; Girsberger, K.M. 2000. Does pecking at inanimate stimuli predict cannibalistic behaviour in domestic fowl? Applied Animal Behaviour Science, 66: 119-133. Clubb, R.; Mason, G.J. 2003. Captivity effects on wide-ranging carnivores. Nature, 425: 473-474. Clubb, R.; Mason, G. 2004. Pacing polar bears and stoical sheep: testing ecological and evolutionary hypotheses about animal welfare. Animal Welfare, 13: S33-40. Clubb, R.; Mason, G. 2007. Natural behavioural biology as a risk factor in carnivore welfare: How analysing species differences could help zoos improve enclosures. Applied Animal Behaviour Science, 102: 303-328. Clubb, R.; Vickery, S. 2006. Locomotory stereotypies in carnivores: does pacing stem from hunting, ranging or frustrated escape? In: Stereotypic Animal Behaviour, 2nd edn, G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 58-85. Clubb, R.; Vickery, S.; Latham, N. 2006. Box 1.1. Motivation and motivational explanations for stereotypies. In: Stereotypic Animal Behaviour, 2nd edn, G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 7-8. 207 Cooper, J.J.; Nicol, C.J. 1991. Stereotypic behaviour affects environmental preferences in bank voles, Clethrionomys glareolus. Animal Behaviour, 41: 971-977. Cooper, J.J.; Nicol, C.J. 1993. The 'coping' hypothesis of stereotypic behaviour: a reply to Rushen. Animal Behaviour, 45: 616-618. Cooper, J.J.; Nicol, C.J. 1996. Stereotypic behaviour in wild caught and laboratory voles. Animal Welfare, 5: 245-257. Cooper, J.J.; Udberg, F.; Nicol, C.J. 1996. Limitations on the effectiveness of environmental improvement in reducing stereotypic behaviour in bank voles {Clethrionomys glareolus). Applied Animal Behaviour Science, 48: 237-248. Cronin, G.M.; Smith, J.A.; Hodge, F.M.; Hemsworth, P.H. 1994. The behaviour of primiparous sows around farrowing in response to restraint and straw bedding. Applied Animal Behaviour Science, 39: 269-280. Cunningham, D.L., Buhr, R.J., and Mamputu, M. 1992. Beak trimming and sex effects on behaviour and performance traits of large white turkeys. Poultry Science, 71: 1606- 1614. Cuthbertson, G.J. 1980. Genetic variation in feather-pecking behaviour. British Poultry Science, 21: 447-450. 208 Dantzer, R. 1986. Behavioural, physiological and functional aspects of stereotyped behaviour: a review and reinterpretation. Journal of Animal Science, 62: 1776-1786. Dawkins, M.S. 1977. Do hens suffer in battery cages? Environmental preferences and welfare. Animal Behaviour, 25: 1034-1046. Dawkins, M.S. 1983. Battery hens name their price: consumer demand theory and the measurement of ethological 'needs'. Animal Behaviour, 31: 1195-1205. Dawkins, M.S. 1988. Behavioural deprivation: a central problem to animal welfare. Applied Animal Behaviour Science, 20: 209-225. Dawkins, M.S. 1989. Time budgets in red junglefowl as a baseline for the assessment of welfare in domestic fowl. Applied Animal Behaviour Science, 24: 77-80. Dawkins, M.S. 1990. From an animal's point of view: motivation, fitness, and animal welfare. Behavioral and Brain Sciences, 13: 1-61. deBoer, S.F.; Slangen, J.L.; van der Gugten, J. 1990. Plasma Catecholamine and Corticosterone Levels During Active and Passive Shock-Prod Avoidance Behavior in Rats: Effects of Chlordiazepoxide. Physiology and Behaviour, 47: 1089-1098. 209 dePassille, A.M.; Rushen, J. 1997. Motivational and physiological analysis of the causes and consequences of non-nutritive sucking by calves. Applied Animal Behaviour Science, 53: 15-31. dePassille, A.M.B.; Metz, J.H.M.; Mekking, P.; Wiepkema, P.R. 1992. Does drinking milk stimulate sucking in young calves? Applied Animal Behaviour Science, 34: 23-36. dePassille, A.M.B.; Christopherson, R.; Rushen, J. 1993. Non-nutritive sucking by the calf and postprandial secretion of insulin, CCK and gastrin. Physiology and Behaviour, 54: 1069-1073. Dixon, L.M. submitted. A review of feather pecking literature and welfare issues associated with it. Avian Biology Research Dixon, L.M.; Duncan, J.J.H. in press. Changes in substrate access did not affect early feather pecking behaviour in two strains of laying hen chicks. Journal of Applied Animal Welfare Science. Dixon, L.M.; Duncan, I.J.H.; Mason, G.J. in press. What's in a peck? Using Fixed Action Pattern morphology to identify the motivational basis of abnormal feather pecking behaviour. Animal Behaviour 210 Dixon, L.M.; Duncan, I.J.H. & Mason, G. in prep. A comparison of four types of enrichments and their effects on feather pecking behaviour in laying hens. Duncan, I.J.H. 1978. The interpretation of preference tests in animal behaviour. Applied Animal Ethology, A: 197-200. Duncan, I.J.H. 1980. The ethogram of the domestic hen. Current Topics in Veterinary Medicine and Animal Science, 8: 5-18. Duncan, I.J.H. 2001a. Animal welfare issues in the poultry industry: is there a lesson to be learned? Journal of Applied Animal Welfare Science, 4: 207-221. Duncan, I.J.H. 2001b. The pros and cons of battery cages. World's Poultry Science, 57: 381-390. Duncan, I.J.H. and Wood-Gush, D.G.M. 1970. Frustration and aggression in the domestic fowl. Animal Behaviour, 19: 500-504. Duncan, I.J.H.; Wood-Gush, D.G.M. 1972. Thwarting of feeding behaviour in the domestic fowl. Animal Behaviour, 20: 444-451. Duncan, I.J.H.; Wood-Gush, D.G.M. 1974. The effects of a rauwolfia tranquillizer on stereotyped movements in domestic fowl. Applied Animal Ethology, 1: 67-76. 211 Duncan, I.J.H.; Home, A.R.; Hughes, B.O.; Wood-Gush, D.G.M. 1970. The pattern of food intake in brown leghorn hens as recorded in a Skinner box. Animal Behaviour, 18: 245-255. Duncan, I.J.H.; Slee, G.S.; Seawright, E.; Breward, J. 1989. Behavioural consequences of partial beak amputation (beak trimming) in poultry. British Poultry Science, 30: 479- 488. Eilam, D.; Dayan, T.; Ben-Eliyahu, S.; Schulman, I.; Shefer, G. & Hendrie, C.A. 1999. Differential behavioural and hormonal responses of voles and spiny mice to owl calls. Animal Behaviour, 58, 1085-1093. El-Lethey, H.; Aerni, V.; Jungi, T.W.; Wechsler, B. 2000. Stress and feather pecking in laying hens in relation to housing conditions. British Poultry Science, 41: 22-28. El-Lethey, H.; Jungi, T.W.; Huber-Eicher, B. 2001. Effects of feeding corticosterone and housing conditions on feather pecking in laying hens (Gallus gallus domesticus). Physiology and Behaviour, 73: 243-251. Emmans, G. C. 1979. Free-choice feeding of laying poultry. In: W. Haresign & D. Lewis D., Eds. Recent Advances in Animal Nutrition, London, UK. pp. 31-39 212 FAOSTAT. 2008. Agribusiness Handbook vol.6. Poultry/Eggs and Poultry Meat. ftp://ftp.fao.org/docrep/fao/007/ae374e/ae374e00.pdf Fentress, J.C. 1976. Dynamic boundaries of patterned behaviour: interaction and self- organization. In: Growing points in ethology. P.P.G. Bateson and R.A. Hinde, eds. Cambridge University Press, Cambride, UK. Ferguson, E.D. 2000. Motivation: A biosocial and cognitive integration of motivation and emotion. Oxford University Press, Oxford, UK. Pp. 81-105. Fox, M.W. 1968. Abnormal behavior in animals. W.B.Saunders Company, London, UK. Francis, D.D.; Zaharia, M.D.; Shanks, N.; Anisman, H. 1995. Stress-induced disturbances in Morris water-maze performances: interstrain variability. Physiology and Behaviour, 58: 57-65. Fraser, D.; Duncan, l.J.H. 1998. 'Pleasures', 'pains' and animal welfare: toward a natural history of affect. Animal Welfare, 7: 383-396. Fraser, D.; Matthews, L.R. 1997. Preference and motivation testing. In: Animal Welfare. M.C. Appleby and Hughes, B.O., eds. CABI Publishing, New York, NY. 213 Garner, J.P. 2005. Stereotypies and other abnormal repetitive behaviors: potential impact on validity, reliability, and replicability of scientific outcomes. ILAR Journal, 46: 113- 117. Garner, J.P. 2006. Perseveration and stereotypy - systems-level insights from clinical psychology. In: Stereotypic Animal Behaviour, 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 121-152. Garner, J.P; Dufour, B.; Gregg, L.E.; Weisker, S.M.; Mench, J.A. 2004. Social and husbandry factors affecting the prevalence and severity of barbering ('whisker trimming') by laboratory mice. Applied Animal Behaviour Science, 89: 263-282. Garner, J.P.; Meehan, C.L.; Famula, T.R.; Mench, J.A. 2006. Genetic, environmental, and neighbor effects on the severity of stereotypies and feather picking in Orange-winged Amazon parrots (Amazona amazonica): An epidemiological study. Applied Animal Behaviour Science, 96: 153-168. Geary, N. 2004. Endocrine controls of eating: CCK, leptin, and ghrelin. Physiology and Behaviour, 81: 719-733. Gentle, M.J.; Hunter, L.N. 1990. Physiological and behavioural responses associated with feather removal in Gallus gallus var domesticus. Research in Veterinary Science, 50:95-101. 214 Gentle, M.J.; McKeegan, D.E.F. 2007. Evaluation of the effects of infrared beak trimming in broiler breeder chicks. Veterinary Record, 160: 145-148. Gentle, M.J.; Hughes, B.O.; Hubrecht, R.C. 1982. The effect of beak trimming on food intake, feeding behaviour and body weight in adult hens. Applied Animal Ethology, 8: 147-159. Gentle, M.J.; Waddington, D.; Hunter, L.N.; and Jones, R.B. 1990. Behavioural evidence for persistent pain following partial beak amputations in chickens. Applied Animal Behaviour Science, 27: 149-157. Gillespie, J.R. 2004. Modern Livestock & Poultry Production. 7th ed. Clifton Park, NY: Delmar Learning. Golani, I; Fentress, J.C. 1985. Early ontogeny of face grooming in mice. Developmental Psychobiology, 18: 529-544. Goodale, M.A. 1983. Visually guided pecking in the pigeon (Columba livid). Brain Behavioural Evolution, 22: 22-41. 215 Goodman, I., Zachny, J., Osman, A., Azzaro, A., Donovan, C, 1983. Dopaminergic nature of feeding-induced behavioral stereotypies in stressed pigeons. Pharmacology Biochemistry and Behaviour, 18: 153-158. Green, L.E.; Lewis, K.; Kimpton, A.; Nicol, C.J. 2000. Cross-sectional study of the prevalence of feather pecking in laying hens in alternative systems and its associations with management and disease. Veterinary Record, \A1: 233-238. Grill, H.J.; Norgren, R. 1978. The taste reactivity test. I. Mimetic responses to gustatory stimuli in neurologically normal rats. Brain Research, 143: 263-279. Gunnarsson, S.; Keeling, L.J.; Svedberg, J. 1999. Effect of rearing factors on the prevalence of floor eggs, cloacal cannibalism and feather pecking in commercial flocks of loose housed laying hens. British Poultry Science, 40: 12-18. Gustafson, L.A.; Cheng, H.W.; Garner, J.P.; Pajor, E.A.; Mench, J.A. 2007. Effects of bill-trimming Muscovy ducks on behavior, body weight gain, and bill morphopathology. Applied Animal Behaviour Science, 103: 59-74. Hansen, S.W.; Hansen, B.K.; Berg, P. 1994. The effect of cage environment and ad libitum feeding on the circadian rhythm, behaviour and feed intake of farm mink. Acta Agriculturae Scandinavica A -Animal Science, 44: 120-127. 216 Harlander-Matauschek, A.; Baes, C; Bessei, W. 2006a. The demand of laying hens for feathers and wood shavings. Applied Animal Behaviour Science, 101: 102-110. Harlander-Matauschek, A.; Piepho, H.P. & Bessei, W. 2006b. The effect of feather eating on feed passage in laying hens. Poultry Science, 85, 21-25. Harlander-Matauschek, A.; Benda, I.; Lavetti, C; Djukic, M.; Bessei, W. 2007. The relative preferences for wood shavings or feathers in high and low feather pecking birds. Applied Animal Behaviour Science, 107: 78-87. Hetland, H.; Svihus, B.; Krogdahl, A. 2003. Effects of oat hulls and wood shavings on digestion in broilers and layers fed diets based on whole or ground wheat. British Poultry Science, AA: 275-282. Hetland, H.; Svihus, B; Choct, M. 2005. Role of insoluble fibre on gizzard activity in layers. Journal of Applied Poultry Research, 14:38-46. Hewson, C.J.; Parent, J.M.; Conlon, P.D.; Luescher, U.A.; Ball, R.O. 1998. Efficacy of clompramine in the treatment of canine compulsive disorder: a randomized, placebo- controlled, double blind clinical trial. Journal of the American Veterinary Medical Association, 213: \160-1166. 217 Hinde, R. A. 1970. Animal Behaviour: A Synthesis of Ethology and Comparative Psychology. McGraw-Hill, Inc., New York, USA. Hoffmeyer, I. 1969. Feather pecking in pheasants - an ethological approach to the problem. Danish Review of Game Biology, 6: 1-36. Hogan, J.A. 1994. The concept of cause in the study of behaviour. In: Causal Mechanisms of Behavioural Development. J.A. Hogan and J.J. Bolhuis, eds. Cambridge University Press, Cambridge, UK. Hogan-Warburg, A. J.; Hogan, J.A. 1981. Feeding strategies in the development of food recognition in young chicks. Animal Behaviour, 29: 143-154. Huber-Eicher, B.; Audige, L. 1999. Analysis of risk factors for the occurrence of feather pecking in laying hen growers. British Poultry Science, 40: 599-604. Huber-Eicher, B.; Sebo, F. 2001. Reducing feather pecking when raising laying hen chicks in aviary systems. Applied Animal Behaviour Science, 73: 59-68. Huber-Eicher, B.; Wechsler, B. 1997'a. Feather pecking in domestic chicks: its relation to dustbathing and foraging. Animal Behaviour, 54: 757-768. 218 Huber-Eicher, B. & Wechsler, B. 1997b. The effect of foraging material and perch height on feather pecking and feather damage in laying hens. Applied Animal Behaviour Science, 58, 131-141. Huber-Eicher, B.; Wechsler, B. 1998. The effect of quality and availability of foraging materials on feather pecking in laying hen chicks. Animal Behaviour, 55: 861-873. Hughes, B.O. 1973. The effect of implanted gonadal hormones on feather pecking and cannibalism in pullets. British Poultry Science, 14: 341-348. Hughes, B.O. 1980. Feather damage in hens caged individually. British Poultry Science, 21: 149-154 Hughes, B.O.; Duncan, I.J.H. 1972. The influence of strain and environmental factors upon feather pecking and cannibalism in fowls. British Poultry Science, 13: 525-547. Hughes, B.O.; Duncan, I.J.H. 1988a The notion of ethological 'need', models of motivation and animal welfare. Animal Behaviour, 36: 1696-1707. Hughes, B.O.; Duncan, I.J.H. 1988b Behavioural needs: Can they be explained in terms of motivational models? Applied Animal Behaviour Science, 19: 352-355. 219 Hughes, B.O.; Gentle, M.J. 1995. Beak trimming for poultry: its implications for welfare. World's Poultry Science, 51: 51-61. Hughes, B.O.; Duncan, I.J.H.; Brown, M.F. 1989. The performance of nest building by domestic hens: is it more important than construction of a nest? Animal Behaviour, 36: 1696-1707. Hutson, G.D. 1984. Animal welfare and consumer demand theory: are preference tests a luxury we can't afford? Animal Behaviour, 32: 1260-1261. Jackson, G. 1991. Intestinal stasis and rupture in rabbits. Veterinary Record, 129: 287- 289. Jenkins, H.M.; Moore, B.R. 1973. The form of the auto-shaped response with food or water reinforcers. Journal of the Experimental Analysis of Behavior, 20: 163-181. Jensen, P.; Toates, F.M. 1993. Who needs 'behavioural needs'? Motivational aspects of the needs of animals. Applied Animal Behaviour Science, 37: 161-181. Jensen, A.B.; Palme, R.; Forkman, B. 2006. Effect of brooders on feather pecking and cannibalism in domestic fowl (Gallus gallus domesticus). Applied Animal Behaviour Science, 99: 287-300. 220 Johnsen, P.F.; Vestergaard, K.S. 1996. Dustbathing and pecking behaviour in chicks from a high and a low feather pecking line of laying hens. Applied Animal Behaviour, 49: 237-246. Johnsen, P.F.; Vestergaard, K.S.; Norgaard-Nielsen, G. 1998. Influence of early rearing conditions on the development of feather pecking and cannibalism in domestic fowl. Applied Animal Behaviour Science, 60: 25-41. Keeling, L.J. 1994. Inter-bird distances and behavioural priorities in laying hens: the effect of spatial restriction. Applied Animal Behaviour Science, 39: 131-140. Keeling, L.J.; Andersson, L.; Schutz, K.E.; Kerje, S.; Fredriksson, R.; Carlborg, O.; Cornwallis, C.K.; Pizzari, T.; Jensen, P. 2004. Chicken genomics: Feather-pecking and victim pigmentation. Nature, 431: 645-646. Keiper, R.R. 1969. Causal factors of stereotypies in caged birds. Animal Behaviour, 17: 114-119. Kirkden, R.D.; Pajor, E.A. 2006. Using preference, motivation and aversion to ask scientific questions about animals' feelings. Applied Animal Behaviour Science, 100: 29-47. 221 Kirkden, R.D.; Edwards, J.S.S.; Broom, D.M. 2003. A theoretical comparison of the consumer surplus and the elasticities of demand as measures of motivational strength. Animal Behaviour, 65: 157-178. Kjaer, J.B. 2000. Diurnal rhythm of feather pecking behaviour and condition of integument in four strains of loose housed laying hens. Applied Animal Behaviour Science, 65: 331-347. Kjaer, J.B.; Sorensen, P. 1997. Feather pecking behaviour in White Leghorns, a genetic study. British Poultry Science, 38: 333-341. Kjaer, J.B. and Vestergaard, K.S. 1999. Development of feather pecking in relation to light intensity. Applied Animal Behaviour Science, 62: 243-254. Kjaer, J.B.; Sorensen, P.; Su, G. 2001. Divergent selection on feather pecking behaviour in laying hens (Gallus gallus domesticus). Applied Animal Behaviour Science, 71: 229- 239. Kjaer, J.B.; Hjarvard, B.M.; Jensen, K.H.; Hansen-Moller, J.; Naesbye Larsen, O. 2004. Effects of haloperidol, a dopamine D2 receptor antagonist, on feather pecking behaviour inlaying hens. Applied Animal Behaviour Science, 86: 77-91. 222 Koolhaas, J.M.; Korte, S.M.; De Boer, S.F.; Van Der Vegt, B.J.; Van Reenen, C.G.; Hopster, H.; De Jong, I.C.; Ruis, M.A.W.; Blokhuis, H.J. 1999. Coping styles in animals: current status in behavior and stress-physiology. Neuroscience and BiobehaviouralReviews, 23: 925-935. Korhonen, H.T.; Niemela, P.; Jauhiainen, L. 2001. Effect of space and floor material on the behaviour of farmed blue foxes. Canadian Journal of Animal Science, 81:189-197. Korte, S.M.; Beuving, G.; Ruesink, W.; Blokhuis, H.J. 1997. Plasma catecholamine and corticosterone levels during manual restraint in chicks from a high and low feather pecking line of laying hens. Physiology and Behaviour, 62: 437-441. Kristensen, H.H.; Prescott, N.B.; Perry, G.C.; Ladewig, J.; Ersboll, A.K.; Overvad, K.C.; Wathes, CM. 2007. The behaviour of broiler chickens in different light sources and illuminances. Applied Animal Behaviour Science, 103:75-89. Kroodsma, D.E.; Byers, B.E.; Goodale, E.; Johnson, S. & Liu, W.C. 2001. Pseudoreplication in playback experiments, revisited a decade later. Animal Behaviour, 61, 1029-1033. Larsen, B.H.; Vestergaard, K.S.; Hogan, J.A. 2000. Development of Dustbathing Behavior Sequences in the Domestic Fowl: The Significance of Functional Experience. Developmental Psychobiology, 37: 5-12. 223 Lawrence, R.A.; Appleby, M.C.; Macleod, H.A. 1988. Measuring hunger in the pig using operant conditioning - the effect of food restriction. Animal Production, 47: 131- 137. Leeson, S.; Morrison, W.D. 1978. Effect of feather cover on feed efficiency in laying birds. Poultry Science, 57: 1094-1096. Lepionka, L.; Malbert, C.H.; Laplacce, J.P. 1997. Proximal gastric distension modifies ingestion rate in pigs. Reproduction Nutrition Development, 37: 449-457. Lewis, R.S.; Hurst, J.L. 2004. The assessment of bar-chewing as an escape behaviour in laboratory mice. Animal Welfare, 13: 19-25. Lewis, M.H.; Presti, M.F.; Lewis, J.B.; Turner, C.A. 2006. In: Stereotypic Animal Behaviour, 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 190-226. Lindberg, A.C.; Nicol, C.J. 1994. An evaluation of the effect of operant feeders on welfare of hens maintained on litter. Applied Animal Behaviour Science, 41: 211-227. Lindberg, A.C.; Nicol, C.J. 1997. Dustbathing in modified battery cages: Is dustbathing an adequate substitute? Applied Animal Behaviour Science, 55: 113-128. 224 Lindstrom, T.; Redbo, I. 2000. Effect of feeding duration and rumen fill on behaviour in dairy cows. Applied Animal Behaviour Science, 70: 83-97. Lorenz, K. 1950. The comparative method in studying innate behaviour patterns. Symposia of the Society for Experimental Biology (Sym. Soc. Exp. Biol.), 4: 221-268. Lorenz, K.; Tinbergen, N. 1938. Taxis and instinctive behaviour pattern in egg-rolling by the Greylag goose. In: Studies in Animal and Human Behaviour vol A (Ed. by K.Lorenz; Translated by R. Martin), pp. 316-350. Massachusetts: Harvard University Press. Lovern, M.B.; Jenssen, T.A. 2003. Form emergence and fixation of head bobbing displays in the green anole lizard {Anole carolinensis): A reptilian model of signal ontogeny. Journal of Comparative Psychology, 117: 133-141. Lyons, J.; Young, R.J.; Deag, J.M. 1997. The effects of physical characteristics of environment and feeding regime on behavior of captive felids. Zoo Biology, 16: 71-83. Marinus, L.M.; Chase, W.K.; Novak, M.A. 2000. Self biting behaviour in rhesus macaques {Macaca mulatto) is preferentially directed to body sites associated with acupuncture analgesia. American Journal of Primatology, 51: 71-72. 225 Marsden, M.D.; Wood-Gush, D.G.M. 1986. A note on the behaviour of individually- penned sheep regarding their use for research purposes. Animal Production, 42:157-159. Martrenchar, A.; Huonnic, D.; Cotte, J.P. 2001. Influence of environmental enrichment on injurious pecking and perching behaviour in young turkeys. British Poultry Science, 42: 161-170. Mason, G. J. 1991a. Stereotypies: A critical review. Animal Behaviour, 41: 1015-1037. Mason, G.J. 1991b. Stereotypies and suffering. Behavioural Processes, 25: 103-115. Mason, G.J. 1993. Age and context affect the stereotypies of caged mink. Behaviour, 127: 191-229. Mason, G. 2006. Stereotypic behaviour in captive animals: fundamentals and implications for welfare and beyond. In: Stereotypic Animal Behaviour, 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 121-152. Mason, G.J.; Latham, N.R. 2004. Can't stop, won't stop: Is stereotypy a reliable animal welfare indicator. Animal Welfare, 13: S57-69. 226 Mason, G.J.; Mendl, M. 1997. Do the stereotypies of pigs, chickens and mink reflect adaptive species differences in the control of foraging? Applied Animal Behaviour Science, 53: 45-58. Mason, G.; Turner, M. 1993. Mechanisms involved in the development and control of stereotypies. In: Perspectives in Ethology 10, P. Bateson, ed. Plenum Press, New York, USA. Pp. 53-85. Mason, G.J.; McFarland, D.; Garner, J. 1998. A demanding task: using economic techniques to assess animal priorities . Animal Behaviour, 55: 1071-1075. Mason, G.J.; Cooper, J.; Clarebrough, C. 2001. Frustrations of fur-farmed mink. Nature, 410: 35-36. Mason, G.; Clubb, R.; Latham, N.; Vickery, S. 2007. Why and how should we use environmental enrichment to tackle stereotypic behaviour? Applied Animal Behaviour Science, 102: 163-188. Maxwell, M.H. 1993. Avian blood leucocyte responses to stress. World's Poultry Science, 49: 34-43. 227 McAdie, T.M.; Keeling, L.J. 2000. Effect of manipulating feathers of laying hens on the incidence of feather pecking and cannibalism. Applied Animal Behaviour Science, 68: 215-229. McAdie, T.M., Keeling, L.J. 2002. The social transmission of feather pecking in laying hens: effects of environment and age. Applied Animal Behaviour Science, 75: 147-159. McAdie,T.M.; Keeling, L.J.; Blokhuis, H.J.; Jones, R.B. 2005. Reduction in feather pecking and improvement of feather condition with the presentation of a string device to chickens. Applied Animal Behaviour Science, 93: 67-80. McCasland, W.E.; Richardson, L.R. 1966. Methods for determining the nutritive value of feather meals. Poultry Science, 45: 1231-1236. McFarland, D. 1971. Feedback Mechanisms in Animal Behaviour. Academic Press, London, England. Pp.66-69. McFarland, D. 1999. Animal Behaviour, 31 ed. Pearson Education Limited, Essex, UK. Pp.246-283. McFarland, D.; Sibly, R.M. 1975. The behavioural final common path. Philosophical Transactions of the Royal Society of London B, 270: 265-293. 228 McGreevy, P.; Nicol, C.J. 1998. Physiological and behavioural consequences associated with short-term prevention of crib-biting in horses. Physiology and Behaviour, 65: 15- 23. McGreevy, P.D.; Webster, A.J.F.; Nicol, C.J. 2001. Study of the behaviour, digestive efficiency and gut transit times of crib-biting horses. Veterinary Record, 148: 592-596. McKeegan, D.E.F.; Savory, C.J. 1999. Feather eating in layer pullets and its possible role in the aetiology of feather pecking damage. Applied Animal Behaviour Science, 65: 73-85. Meehan, C.L.; Millam, J.R.; Mench, J.A. 2003. Foraging opportunity and increased physical complexity both prevent and reduce psychogenic feather picking by young Amazon parrots. Applied Animal Behaviour Science, 80: 71-85. Mendl, M. 1997. Responses to supernormal stimuli and their implications for animal welfare: can animals have too much of a good thing? Applied Animal Behaviour Science, 54:47-51. Meyer-Hozapfel, M. 1968. Abnormal behaviour in zoo animals. In: Abnormal Behaviour in Animals. M.W. Fox, ed., W.B.Saunders Company, Ontario, Canada. Pp. 476-503. 229 Millman, S.T.; Duncan, I.J.H. 2000. Strain differences in aggressiveness of male domestic fowl in response to a male model. Applied Animal Behaviour Science, 66: 217- 233. Mills, D.; Luescher, A. 2006. Veterinary and pharmacological approaches to abnormal repetitive behaviour. In: Stereotypic Animal Behaviour, 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 286-324. Moberg GP. 2000. Biological response to stress: Implications for animal welfare. In:. The Biology of Animal Stress: Basic Principles and Implications for Animal Welfare. G.P. Moberg and J.A. Mench, eds., CAB International, Wallingford, UK p. 1-22. Moore, D.S. & McCabe, G.P. 1999. Introduction to the Practice of Statistics. 3rd ed. NewYork: W.H. Freeman and Company. Morris, D. 1964. The response of animals to a restricted environment. Symposium of the Zoological Society, London, 13: 99-120. Mroz, Z.; Partridge, I.G.; Mitchell, G.; Keal, H.D. 1986. The effects of oat hulls, added to the basal ration for pregnant sows, on reproductive performance, apparent digestibility, rate of passage and plasma parameters. Journal of the Science of Food and Agriculture, 37: 239-247. 230 Nader, K.; Bechara, A.; van der Kooy, D. 1997. Neurobiological constraints on behavioural models of motivation. Annual Review of Psychology, 48: 85-114. Nevison, CM.; Hurst, J.L.; Barnard, C.J. 1999a. Why do male ICR(CD-l) mice perform bar-related (stereotypic) behaviour? Behavioural Processes, 47: 95-111. Nevison, CM.; Hurst, L.J.; Barnard, C.J. 1999b. Strain-specific effects of cage enrichment in male laboratory mice (Mus musculus). Animal Welfare, 8: 361-379. Newberry, R.C; Keeling, L.J.; Estevez, I.; Bilcik, B. 2007. Behaviour when young as a predictor of severe feather pecking in adult laying hens: The redirected foraging hypothesis revisited. Applied Animal Behaviour Science, 107: 262-274. Nicol, C.J. 1987. Behavioural responses of laying hens following a period of spatial restriction. Animal Behaviour, 35: 1709-1719. Nicol, C.J. 1995. The social transmission of information and behaviour. Applied Animal Behaviour Science, 44: 79-98. Nicol, C.J.; Badnell-Waters, A.J. 2005. Suckling behaviour in domestic foals and the development of abnormal oral behaviour. Animal Behaviour, 70: 21-29. 231 Nicol, C.J.; Lindberg, A.C.; Phillips, A.J.; Pope, S.J.; Wilkins, L.J.; Green, L.E. 2001. Influence of prior exposure to wood shavings on feather pecking, dustbathing and foraging in adult laying hens. Applied Animal Behaviour Science, 73: 141-155. Nicol, C.J.; Potzsch, C; Lewis, K.; Green, L.E. 2003. Matched concurrent case-control study of risk factors for feather pecking in hens on free-range commercial farms in the UK. British Poultry Science, 44: 515-523. Norgaard-Nielsen, G. 1997. Dustbathing and feather pecking in domestic chickens reared with and without access to sand. Applied Animal Behaviour Science, 52: 99-108. Norgaard-Nielsen, G.; Vestergaard, K.S.; Simonsen, H.B. 1993. Effects of early experience and stimulus enrichment on feather damage in laying hens. Applied Animal Behaviour Science, 38: 345-352. North, M.; Bell, D. 1990. Commercial Chicken Production Manual. Westport, Connecticut: AVI Publishing Company. Novak, M.A. 2003. Self-injurious behaviour in rhesus monkeys: new insights into its etiology, physiology, and treatment. American Journal of Primatology, 59: 3-19. 232 Novak, M.A.; Meyer, J.S.; Lutz, C; Tiefenbacher, S. 2006. Deprived environments: developmental insights from primatology. In: Stereotypic Animal Behaviour, 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 153-189. Ddberg, F.O. (1978) Abnormal behaviours: stereotypies. In: First World Congress on Ethology Applied to Zootechnics, Industrias Grafices Espana, Madrid. Pp.475-480. Ddberg, F.O. 1986. The jumping stereotypy in the bank vole (Clethrionomys glareolus). Biology of Behaviour, 11: 130-143. • dberg, F.O. 1987. The effect of cage size and enrichment on the development of stereotypies in bank voles {Clethrionomys glareolus). Behavioural Processes, 14: 155- 173. Olsson, I.A.S.; Keeling, L.J. 2005. Why in earth? Dustbathing behaviour in jungle and domestic fowl reviewed from a Tinbergian and animal welfare perspective. Applied Animal Behaviour Science, 93: 259-282. Olsson, I.A.S.; Keeling, L.J.; Duncan, I.J.H. 2002. Why do hens sham dustbathe when they have litter? Applied Animal Behaviour Science, 76: 53-64. 233 Pecina, S.; Cagniard, B.; Berridge, K.C.; Aldridge, J.W.; Zhuang, X. 2003. Hyper- dopaminergic mutant mice have higher "liking" but not "wanting" for sweet rewards. Journal ofNeuroscience, 23: 9395-9402. Perry, G.C.; Allen, J. 1976. The effects of caging conditions on the incidence of feather pecking and cannibalism. In: 5th European Poultry Conference (Pp. 1019-1029) Malta: World Poultry Science Association. Petherick, J.C.; Duncan, I.J.H. 1989. Behaviour of young domestic fowl directed towards different substrates. British Poultry Science, 30: 229-238. Ponzio, M.F.; Busso, J.M.; Ruiz, R.D.; de Cuneo, M.F. 2007. A survey of the incidence of fur-chewing in commercial chinchilla {Chinchilla lanigera) farms. Animal Welfare, 16: 471-479. Poulsen, E.M.B.; Teskey, G.C. 2006. Box 10.4. Pacing, Prozac and a polar bear. In: Stereotypic Animal Behaviour, 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 309. Pozis-Francois,0.; Zahavi, A. & Zahavi, A. 2004. Social play in Arabian Babblers. Behaviour, 141, 425-450. 234 Preston, A.P. 1987. Location in the cage and diurnal distribution of feather pecking by caged layers. British Poultry Science, 28: 653-658. Ramonet, Y.; Meunier-Salaun, M.C.; Dourmad, J.Y. 1999. High-fiber diets in pregnant sows: digestive utilization and effects on the behaviour of the animals. Journal of Animal Science,!!: 591-599. Redbo, I. 1990. Changes in duration and frequency of stereotypies and their adjoining behaviours in heifers, before, during and after the grazing period. Applied Animal Behaviour Science, 26: 57-67. Redbo, I.; Redbo-Tortensson, P.; Ddberg, F.O.; Hedendahl, A.; Holm, J. 1998. Factors affecting behavioural disturbances in race-horse. Animal Science, 66: 475-481. Reinhardt, V. 2005. Hair pulling: a review. Laboratory Animals, 39: 361-369. Riber, A.; Forkman, B. 2007. A note on the behaviour of the chicken that receives feather pecks. Applied Animal Behaviour Science, 108:337-341. Riber, A.; Mench, J.A. in press. Effects of feed- and water-based enrichment on activity and cannibalism in Muscovy ducklings. Applied Animal Behaviour Science, (2008), doi:10.1016/j.applanim.2008.03.005 235 Riber, A.B.; Wichman, A.; Braastad, B.O..; Forkman, B. 2007. Effects of broody hens on perch use, ground pecking, feather pecking and cannibalism in domestic fowl {Gallus gallus domesticus). Applied Animal Behaviour Science, 106: 39-51. Riedstra, B.; Groothuis, T.G.G. 2002. Early feather pecking as a form of social exploration: the effect of group stability on feather pecking and tonic immobility in domestic chicks. Applied Animal Behaviour Science, 11: 127-138. Robert, S.; Matte, J.J.; Farmer, C; Girard, C.L. Martineau, G.P. 1993. High-fibre diets for sows: effects on stereotypies and adjunctive drinking. Applied Animal Behaviour Science, 37: 161-171. Rodenburg, T.B.; Zimmerman, P.H., Koene, P. 2002. Reaction to frustration in high and low feather pecking laying hens. Behavioural Processes, 59: 121-129. Rodenburg, T.B.; Koene, P. 2003. Comparison of individual and social feather pecking tests in two lines of laying hens at ten different ages. Applied Animal Behaviour Science, 81: 133-148. Rodenburg, T.B.; Buitenhuis, A.J.; Ask, B.; Uitdehaag, K.A.; Koene, P.; van der Poel, J.J.; Bovenhuis, H. 2003. Heritability of feather pecking and open-field response of laying hens at two different ages. Poultry Science, 82: 861-867. 236 Rodenburg, T.B.; van Hierden,Y.M.; Buitenhuis,A.J.; Riedstra,B.; Koene,P.; Korte,S.M.; van der Poel,J.J.; Groothuis,T.G.G.; Blokhuis,H.J. 2004. Feather pecking in laying hens: new insights and direction for research? Applied Animal Behaviour Science, 86: 291- 298. Rodenburg, T.B.; Koene, P.; Bokkers, E.A.M.; Bos, M.E.H.; Uitdehaag, K.A.; Spruijt, B.M. 2005. Can short-term frustration facilitate feather pecking in laying hens? Applied Animal Behaviour Science, 91: 85-101. Rolls, E.T. 2000. Precis of the brain and emotion. Behavioural and Brain Sciences, 23: 177-234. Rushen, J. 1984. Stereotyped behaviour, adjunctive drinking and the feeding periods of tethered sows. Animal Behaviour, 32: 1059-1067. Rushen, J. 1993. The'coping'hypothesis of stereotyped behaviour. Animal Behaviour, 45:613-615. Rushen, J.; de Passille, A.M. 1995. The motivation of non-nutritive sucking in calves, Bos taurus. Animal Behaviour, 49: 1503-1510. 237 Rushen, J.; Mason, G.J. 2006. A decade-or-more's progress in understanding stereotypic behaviour. In: Stereotypic Animal Behaviour, 2nd edn. G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. Pp. 1-18. Rushen, J.; Robert, S.; Farmer, C. 1999. Effects of an oat-based high fibre diet on insulin, glucose, Cortisol and free fatty acid concentrations in gilts. Animal Science, 69: 395-401. Sainsbury, D. 2000. Poultry Health and Management Ames, Iowa: 4' ed. Blackwell Publishing. Sandson, J.; Albert, M.L. 1984. Varieties of perseveration. Neuropsychologia, 6:715- 732. Sandyk, R. 1992. 1-Tryptophan in neuropsychiatric disorders: a review. International Journal of Neuroscience, 67: 127-144. Sanotra, G.S.; Vestergaard, K.S.; Agger, J.F.; Lawson, L.G. 1995. The relative preferences for feathers, straw, wood-shavings and sand for dustbathing, pecking and scratching in domestic chicks. Applied Animal Behaviour Science, 43: 263-277'. 238 Sato, S.; Kubo, T.; Abe, M. 1992. Factors influencing tongue-rolling and the relationship between tongue-rolling and production traits in fattening cattle. Journal of Animal Science, 70 (Suppl. 1): 157. Savory, C.J. 1995. Feather pecking and cannibalism. World's Poultry Science, 51: 215- 219. Savory, C.J.; Mann, J.S. 1999. Feather pecking in groups of growing bantams in relation to floor litter substrate and plumage colour. British Poultry Science, 40: 565-572. Savory, C.J.; Mann , J.S; Macleod, M.G. 1999. Incidence of pecking damage in growing bantams in relation to food forms, group size, stocking density, dietary tryptophan concentration and dietary protein source. British Poultry Science, 40: 579-584. Savory, C.J.; Maros, K. 1993. Influence of degree of food restriction, age, and time of day on behaviour of broiler breeder chickens. Behavioural Processes, 29: 179-190. Sawyer, L.S.; Moon-Fanelli, A.A.; Dodman, N.H. 1999. Psychogenic alopecia in cats: 11 cases (1993-1996). Journal of the American Veterinary Medical Association, 214: 71-74. 239 Schaibnle, P.J.; Davidson, J.A.; Bandemer, S.L. 1947. Cannibalism and feather pecking in chicks as influenced by certain changes in a specific ration. Poultry Science, 26: 651- 656. Schleidt, W.M. 1974. How "fixed" is the Fixed Action Pattern? Zeitschrift fur Tierpsychologie, 36: 184-211. Schlosser, S.; Black, D.W.; Blum, N. 1994. The demography, phenomenolgy, and family history of 22 persons with compulsive hair pulling. Annals of Clinical Psychiatry, 6: 147-152. Schoenecker, B.; Heller, K.E. 2001. The involvement of dopamine (DA) and serotonin (5-HT) in stress-induced stereotypies in bank voles (Clethrionomys glareolus). Applied Animal Behaviour Science, 73: 311-319. Seibert, L.M. 2007. Pharmocotherapy for behavioural disorders in pet birds. Journal of Exotic Pet Medicine, 16: 30-37. Shea, M.M.; Mench, J.A.; Thomas, O.P. 1990. The effect of dietary tryptophan on aggressive behaviour in developing and mature broiler breeder males. Poultry Science, 69: 1664-1669. 240 Shepherdson, D.J.; Carlstead, K.C.; Wielebnowski, N. 2004. Cross-institutional assessment of stress responses in zoo animals using longitudinal monitoring of faecal corticoids and behaviour. Animal Welfare, 13: SI05-113. Sherwin, CM.; Nicol, C.J. 1998. A demanding task: using techniques to assess animal priorities. A reply to Mason et al. Animal Behaviour, 55: 1079-1081. Sherwin, CM.; Lewis, P.D.; Perry, G.C 1999. Effects of environmental enrichment, fluorescent and intermittent lighting on injurious pecking amongst male turkey poults. British Poultry Science, 40: 592-598. Skinner, B.F. 1938. The Behaviour of Organisms: An Experimental Analysis. D. Appleton Century, New York, NY. Smith, K.; St Claire, M.; Shaver, C; Olexa, P. .2004. Use of a shredded paper substrate to ameliorate abnormal self-directed behaviour of a chimpanzee {Pan troglodytes). American Journal of Primatology, 62: 94—95. Steenfeldt, S.; Engberg, R.M.; Kjaer, J.B. 2001. Feeding roughage to laying hens affects egg production, gastro-intestinal parameters and mortality. In Proceedings of the 13 th European symposium on poultry nutrition. Blankenbergen, Belgium. Stolba, A.; Wood-Gush, D.G.M. 1989. The behaviour of pigs in a semi-natural environment. Animal Production, 48: 419-425. 241 Su, G.; Kjaer, J.B.; S0rensen, P. 2006. Divergent selection on feather pecking behavior in laying hens has caused differences between lines in egg production, egg quality, and feed efficiency. Poultry Science, 85:191-197. Swaisgood, R.; Shepherdson, D. 2006. Environmental enrichment for mitigating stereotypies in zoo animals: a literature review and meta-analysis. In: Stereotypic Animal Behaviour, 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 256-285. Swaisgood, R.; Ellis, S.; Forthman, D.L.; Shepherdson, D.J. 2001. Commentary: Improving Well-Being for Captive Giant Pandas: Theoretical and Practical Issues. Zoo Biology, 22: 347-354. Tarou, L.R.; Bloomsmith, M.A.; Maple, T.L. 2005. Survey of stereotypic behavior in Prosimians. American Journal of Primatology, 65: 181-196. Tauson, R. 2005. Management and housing systems for layers - effects on welfare and production. World's Poultry Science, 61: 477-490. Teitelbaum, P. 1966. The use of operant methods in the assessment and control of motivational states. In: W.K. Honig, ed. Operant behaviour: areas of research and application. Appleton Century Crofts, New York, NY. Pp. 565-608. 242 Terlouw, E.M.C.; Lawrence, A.B. 1993. Long-term effects of food allowance and housing on development of stereotypies in sows. Applied Animal Behaviour Science, 38: 103-126. Terlouw, E.M.C.; Lawrence, A.B.; Illius, A.W. 1991. Influences of feeding levels and physical restriction on the development of stereotypies in sows. Animal Behaviour, 42: 981-991. Terlouw, E.M.C.; Wiersma, A.; Lawrence, A.B.; Macleod, H.B. 1993. Ingestition of food facilitates the performance of stereotypies in sows. Animal Behaviour, 46: 939-950. Tinbergen, N. 1951. The study of instinct. Oxford University Press, London. Tinbergen, N. 1963. On aims and methods of ethology. Zeitschrift fur Tierpsychologie, 20:410-433. Toates, F. 1980. A systems approach to sexual behaviour. In: Analysis of Motivational Processes. F.M. Toates and T.R. Halliday, eds. Academic Press, London, UK. Toates, F. 1986. Motivational Systems. Cambridge University Press, Cambridge, UK. 243 Toates, F. 2000. Multiple factors controlling behaviour: implications for stress and welfare. In: The Biology of Animal Stress. G. Moberg and J. Mench, eds. CAB International, Wallingford, UK. Pp. 199-226. Toates, F. 2001. Biological Psychology: An Integrative Approach. Pearson Education Limited, Essex, UK. Torrey, S.; Widowski, T.M. 2006. Is belly nosing redirected suckling behaviour? Applied Animal Behaviour Science, 101: 288-304. van de Weerd, H.A.; Docking, CM.; Day, J.E.L.; Edwards, S.A. 2005. The development of harmful social behaviour in pigs with intact tails and different enrichment backgrounds in two housing systems. Animal Science, 80: 289-298. van der Harst, J.E.; Spruijt, B.M. 2007. Tools to measure and improve animal welfare: reward-related behaviour. Animal Welfare, 16(S): 67-73. vanHierden, Y.M.; Korte, S.M.; Ruesink, E.W.; vanReenen, C.G.; Engel, B.; Koolhaas, J.M.; Blokhuis, H.J. 2002. The development of feather pecking behaviour and targeting of pecking in chicks from a high and low feather pecking line of laying hens. Applied Animal Behaviour Science, 11: 183-196. 244 vanHierden, Y.M.; Koolhaas, J.M.; Kost'al, L.; Vyboh, P.; Sedlackova, M.; Rajman, M.; Jurani, M.; Korte, S.M. 2005. Chicks from a high and low feather pecking line of laying hens differ in apomorphine sensitivity. Physiology and Behaviour, 84: 471-477. vanKrimpen, M.M.; Kwakkel, R.P.; Reuvekamp, B.F.J.; van der Peet-Schwering, C.M.C.; den Hartog, L.A.; Verstegen, M.W.A. 2005. Impact of feeding management on feather pecking in laying hens. World's Poultry Science, 61: 663-686. Vestergaard, K.S. 1982a. Dust-bathing in the domestic fowl - diurnal rhythm and dust deprivation. Applied Animal Ethology, 8: 487-495. Vestergaard, K.S. 1982b. The significance of dustbathing for the well-being of the domestic hen. Tierhaltung, 113: 109-118. Vestergaard, K.S. 1994. Dustbathing and its relation to feather pecking in the fowl: Motivational and developmental aspects. PhD Thesis. Royal Veterinary and Agricultural University, Copenhagen. Vestergaard, K.S., Lisborg, L. 1993. A model of feather pecking development which relates to dustbathing in the fowl. Behaviour, 126: 291. 245 Vestergaard, K.S.; Hogan, J.A.; Kruijt, J.P. 1990. The development of a behaviour system: dustbathing in the Burmese Red Junglefowl I. The influence of the rearing environment on the organization of dustbathing. Behaviour, 112: 99-116. Vestergaard, K.S.; Kruijt, J.P.; Hogan, J.A. 1993. Feather pecking and chronic fear in groups of red junglefowl: their relations to dustbathing, rearing environment and social status. Animal Behaviour, 45: 1127-1140. Vickery, S. 2006. Evaluating stereotypic frequency in enrichment studies: different methods lead to different conclusions. In: Stereotypic Animal Behaviour, 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 268. Vickery, S.; Mason, G. 2004. Stereotypic Behavior in Asiatic Black and Malayan Sun Bears. Zoo Biology, 23: 409-430. vonHolst, E.; Mittelstaedt, H. 1950. Das Reafferenzprinzip. Naturwissenschaften 37: 464-476. Wechsler, B. 1991. Stereotypies in polar bears. Zoo Biology, 10: 177 - 188. Weld, K.P.; Mench, J.A.; Woodward, R.A.; Bolesta, M.S.; Suomi, S.J.; Higley, J.D. 1998. Effect of tryptophan treatment on self-biting and central nervous system serotonin 246 metabolism in Rhesus monkeys (Macaca mulatto). Neuropsychopharmacology, 19: 314— 321. Wechsler, B.; Huber-Eicher, B. 1998. The effect of foraging material and perch height on feather pecking and feather damage in laying hens. Applied Animal Behaviour Science,5S: 131-141. Widowski, T.M.; Torrey, S.; Bench, C.J.; Gonyou, H.W. 2008. Development of ingestive behaviour and the relationship to belly nosing in early-weaned piglets. Applied Animal Behaviour Science, 110: 109-127. Wiedenmayer, C. 1996. Effect of cage size on the ontogeny of stereotyped behaviour in gerbils. Applied Animal Behaviour Science, 47: 225-233. Wiedenmayer, C. 1997. Causation of the ontogenetic development of stereotypic digging in gerbils. Animal Behaviour, 53: 461-470. Wielebnowski, N.C.; Fletchall, N.; Carlstead, K.; Busso, J.M.; Brown, J.L. 2002. Noninvasive assessment of adrenal activity associated with husbandry and behavioral factors in the North American Clouded Leopard population. Zoo Biology, 21: 77-98. Wiepkema, P.R. 1971. Positive feedbacks at work during early feeding. Behaviour, 39: 266-273. 247 Wise, R.A.; Rompre, P.P. 1989. Brain dopamine and reward. Annual Review of Psychology, 40: 191-225. Wood-Gush, D.G.M. 1971. The Behaviour of the Domestic Fowl. Heinemann Educational, London, UK. Worobec, E.K.; Duncan, I.J.H.; Widowski, T.M. 1999. The effects of weaning at 7, 14 and 28 days on piglet behaviour. Applied Animal Behaviour Science, 62: 173-182. Wilrbel, H. 2006a. The motivational basis of caged rodents stereotypies. In: Stereotypic Animal Behaviour. 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 86-120. Wiirbel, H. 2006b. Box 5.4. Stereotypies and abnormal perseveration - a unifying theory? In: Stereotypic Animal Behaviour. 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. pp. 142. Wiirbel, H.; Bergeron, R.; Cabib, S. 2006. Box 1.3. The coping hypothesis of stereotypic behaviour. In: Stereotypic Animal Behaviour. 2nd edn., G. Mason and J. Rushen, eds., CAB International, Cambridge, UK. Pp. 14-15. 248 Wurbel, H.; Freire, R.; Nicol, C.J. 1998. Prevention of stereotypic wire-gnawing in laboratory mice: effects on behaviour and implications for stereotypy as a coping response. Behavioural Processes, 42: 61-72. Zeltner, E., Klein, T., and Huber-Eicher, B. 2000. Is there social transmission of feather pecking in groups of laying hen chicks? Animal Behaviour, 60: 211-216. Zweers, G.A. 1982. Pecking of the pigeon {Columbia livia I.). Behaviour, 81: 173-230. Zweers, G.A. 1992. Behavioural mechanisms of avian drinking. Netherlands Journal of Zoology, 42: 60-84. 249 APPENDIX -A Validation of substrate choices used in Chapter Five, Experiment 1 1. INTRODUCTION The use of substrates is quite common in poultry research; however, multiple substrates have been used to fulfill the same function (e.g. as forages or dustbaths) without considering how the bird reacts to these materials. For example, birds have been presented wood shavings, sand and peat moss as dustbaths (eg. Petherick and Duncan 1989; Nicol et al, 2001; and Lundberg and Keeling 2003) but it is not known if there is a difference in the way the birds perceive these substrates as dustbaths. This problem occurs with forages as well: birds have been given wood shavings, long and short cut straw and "litter" as forages in various studies (eg. Blokhuis 1989; Huber-Eicher and Wechsler 1997a; Newberry et al. 2007). There is also an overlap in which substrates are classified as forages and which are classified as dustbaths, and it appears that a substrate is classified based on what the experiment is studying and not necessarily on how the birds perceive them. Both foraging and dustbathing on a substrate involve pecking and scratching and it is not yet known if there is even a difference between pecking at forages and pecking at a dustbath. To complicate matters, a number of forages can be dustbathed on and all dustbaths can be foraged on. 250 Before continuing to compare feather pecks with foraging and dustbathing pecks, it must be ensured that forages and dustbaths were pecked at differently and that the type of substrate did not affect the motor patterns directed to it. The objectives of this experiment are threefold. First, was to compare different types of forages to determine if the motor patterns directed to these substrates are similar. Second, was to compare pecks to different types of dustbathing substrates to determine if the pecking motor patterns are similar. Third, was to compare pecking motor patterns of forages and dustbaths to establish if there were any differences. It was hypothesized that because foraging and dustbathing stem from different behavioural systems the pecking behaviour associated with each would take on a different form. From this, it was predicted that pecks to various forages would be similar and pecks to various dustbaths would be similar, but that the motor patterns involved in foraging would be different than those involved in dustbathing. 2. METHODS 2.1. Animals and Housing The 24 non-beaked trimmed, one-day old female ISA White Leghorns, received from a commercial breeder (Bonnie's Hatchery, Elmira, Ont.) were randomly distributed between two identical pens with twelve birds to a pen. They were housed at the Arkell Poultry Research Facility (University of Guelph) in wooden framed pens measuring 251 127cm x 107cm x 60cm with 2.5cm x 2.5cm grid chicken wire sides and a 1cm x 1cm wire mesh floor. On one end of each pen, a test area was divided off with a sheet of plastic board with a removable door. The test area formed a triangle, 35cm x 35cm x 24cm with black cloth covering the sides to prevent visual contact with other birds. Each pen had two automatic cup drinkers and two metal 12" slide top feeders. The chicks were fed ad lib. Chick Starter Fine Crumbles from Floradale Feed Mill Ltd., Floradale, Ontario, until week 6 of life. The lights were kept on a 23L:1D ratio for the first 3 days, then switched to 10L:14D ratio for the remainder of the trial. Temperature started at 33°C and was gradually reduced to 28°C by the end of the trial. For these experiments, birds that performed feather pecking were needed. To ensure this, birds were kept in a restrictive environment with no access to a dustbath or forages. After four weeks, all birds were wing-tagged with coloured tags in the skin flap at the base of the wing for individual identification. 2.2. Home Pen Observations Chicks were given habituation sessions to various stimuli in their home pens at four weeks of age by presentation of each stimulus three times for thirty minutes periods (at least once between 800-1200 hours, once between 1200-1600 hours). The order of the 252 stimuli was randomly determined and six birds from each pen were randomly chosen to be test subjects. 2.3. Test Area Observations The test birds were given habituation sessions to the Test Area three times for thirty minute sessions (at least once between 800-1200 hours, once between 1200-1600 hours). Two categories of stimuli were presented in the Test Area, forages and dustbaths, and for each category of stimuli, three different types of substrate (sub-types) were used. Birds were video recorded once with each sub-type of each stimulus from both categories in the Test Area, in a randomly determined order. 2.4. Peck Stimuli 1) Forages: Commercial bird-feeder suet holders (10cm x 10cm x 4cm) were each filled with one of the three types of substrate to create three different types of stimulus: i) peanut butter suet (peanut butter was added to suet at a lc suet to l/4c peanut butter ratio), ii) seeds in suet (seeds were a mix of shelled and unshelled sunflower seeds added in the same ratio as the peanut butter) and iii) cabbage leaves. These forages are not able to be dustbathed on, are not dusty, and are able to have pieces torn off and consumed. They were presented to the birds on the floor of the pens. 253 2) Dustbaths: Dustbaths were presented in large pans (30cm x 15cm x 6cm), filled with one of three substrates: peat moss, white sand or grey sand. Pecks were only measured after a dustbathing bout had started and the bird was squatting on the ground. These were presented to the birds on the floor of the pens. As birds were required to be sitting in the dustbath but tend to be standing when pecking forages, it was difficult to present the stimuli to the birds so that the distance from the stimuli to the birds was the same in both cases. However, as mentioned in Chapter 5, it has been shown that pigeons reliably orientate their head at certain distances from a target during two head fixations (Goodale, 1983) and results from the Preliminary Trial in Chapter 5 demonstrate that chickens consistently perform the second head fixation, thus stimuli height from the cage floor should not affect the measures being recorded. 2.5. Peck Measurements The measures of peck morphology used in this experiment were based on previous studies of both pigeons and poultry that quantified various aspects of pecking and compared pecks with different motivational basis (e.g. Jenkins and Moore, 1973; Goodale, 1983). 1) Duration of head fixation (sec): The length of time that the head is kept still before the peck. 254 2) Duration from fixation to contact (sec): Duration from the end of head fixation to beak contact with the stimulus. 3) Duration of the whole peck (sec): Time from no fixation (head movement), through the peck, back to no head movement. Sessions where the birds were exposed to the stimuli were video recorded in real time at 60fps with a Panasonic CCD camera (WV-BL90A), monitor (WV-CM110A) and time lapse VCR (TL 950). Recording at this high speed enabled us to play back pecks frame by frame and determine accurate durations of the peck morphology with little variation within measures. 2.6. Statistical Analysis A mixed model variance component analysis was used with Bird nested in Pen as the subject and Sub-Type (e.g. white sand, peat moss or gray sand) nested in Category (forage, dustbath) (SAS, v.8). Data was normalized using Natural Log transformations when necessary. Differences of the Least Square Means adjusted by Tukey were used to further examine treatment differences. 255 2.7. Ethical Note The use of all animals and methods in the following experiments were approved by the University of Guelph Animal Care Committee which adheres to CCAC guidelines. 3. Results 3.1. Different Sub-Types of Forages and Dustbaths There was no effect of the sub-type of forages or the sub-type of dustbaths on any of the peck motor patterns measured (Fig. A-C). 3.2. Dustbathing versus Foraging Pecks The durations of the whole pecks to the forages were longer and had longer head fixations than pecks to the dustbaths; however, there were no significant differences in the time to contact with the stimulus (Fig. A-C). 4. DISCUSSION Pecks to the different sub-types of forages and pecks to the different sub-types of dustbaths were similar for all motor pattern measures. The time it took to contact the stimuli from the end of head fixation was similar between substrate categories (forages or 256 dustbaths), however, pecks directed to the forages took overall longer to complete and had longer head fixation. The predictions that pecks to different sub-types of forages would be similar and pecks to different sub-types of dustbaths would be similar in motor patterns were met. As well, the motor patterns involved in pecking at the ground during a dustbath and pecking at the ground while foraging were distinct from each other. This emphasizes the differences in motivational systems behind these behaviour patterns. This also complements previous research which demonstrates that the motivations behind behaviour influence the motor patterns involved and that differently motivated behaviours have different motor patterns (i.e. pigeons show feeding or drinking pecks to a key they were trained to peck for a food or water reward, Jenkins and Moore, 1973) These results also help to justify the use of various "dustbathing" and "foraging" substrates in previous experiments. Researchers may be making correct assumptions if they use non-ambiguous substrates (e.g. forages that can't be dustbathed on) and take dustbathing measures only after the start of a dustbathing bout. Of course, this is just a brief comparison of these two behaviour patterns and future work may discover more differences or similarities. As with Chapter 5, we were not able to distinguish on a consistent enough basis whether the chickens only pecked at the stimuli or whether they also consumed pieces to make it a practical measure. Also, the chickens were also observed to consume pieces of the dustbathing material, making the criteria of eating a piece of the stimuli equal to feeding-related consummatory behaviour unreliable. However, the variability when 257 recording pecks to the forages was quite low, both within pecks and within birds, indicating that either we were reliably measuring the same response (either appetitive or consummatory) or that there is no differences with regards to morphology with these measures in chickens. Now that differences in dustbathing and foraging pecks have been confirmed, comparison of those with feather pecks can be made and underlying motivation determined. 5. CONCLUSIONS The hypothesis that that because foraging and dustbathing stem from different behavioural systems the pecking behaviour associated with each would take on a different form was supported, as there were different motor patterns in the pecks directed to dustbaths and forages. Future work comparing the motor patterns and underlying motivations of feather pecking with dustbathing and foraging is justified. 258 0.25 i • Gay Sand 0 Peat Moss • White Sand B Cabbage Leaves • Peanut Butter Suet H Seed Suet Dustbaths Forages Fig. A: The average duration of a peck to the different types of forages and dustbaths. The data represent the average duration of a peck ± SEM.(Categories: FO, 24) = 30.93, P O.0001; Types: F(4>24) = 0.93, P = 0.4625). * denotes PO.05 259 0.08 0.07 r—f— J T 0.06 L H^l •mi • day Sand 0.05 *W$wJ)ffl ^ 0 Peat Moss D White Sand 0.04 'W#s w §§lm • Cabbage Leaves 0.03 4 ^m ,miif g • Peanut Butter Suet 0.02 D Seed Suet 0.01 •mH 0 I yvvvwv Dustbaths Forages Fig.B: The average time from head fixation to contact with the stimulus for forages and dustbaths. The data represent the average time to contact the stimulus ± SEM. (Categories: F(i>24) = 0.14, P = 0.7134; Types: F(4>24) = 2.98, P = 0.0594) 260 0.25 • Cray Sand H Peat Moss 0.15 • White Sand H Ckbbage Leaves Q Peanut Butter Suet El Seed Suet 0.05 Dustbaths Forages Fig. C: The average duration of head fixation for forages and dustbaths. The data represent the average duration of head fixation ± SEM. (Category: F(i;24) = 16.12, P 0.0005; Type: F(4,24) = 2.27, P = 0.091). * denotes P<0.05 261