UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Understanding the behaviour of transition dairy cows Proudfoot, Kathryn Louise 2013

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
24-ubc_2013_fall_proudfoot_kathryn.pdf [ 1.84MB ]
Metadata
JSON: 24-1.0103338.json
JSON-LD: 24-1.0103338-ld.json
RDF/XML (Pretty): 24-1.0103338-rdf.xml
RDF/JSON: 24-1.0103338-rdf.json
Turtle: 24-1.0103338-turtle.txt
N-Triples: 24-1.0103338-rdf-ntriples.txt
Original Record: 24-1.0103338-source.json
Full Text
24-1.0103338-fulltext.txt
Citation
24-1.0103338.ris

Full Text

  UNDERSTANDING THE BEHAVIOUR OF TRANSITION DAIRY COWS  by  Kathryn Louise Proudfoot  B.Sc., The University of California, Los Angeles, 2004 M.Sc., The University of British Columbia, 2008   A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Applied Animal Biology)    THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)    September 2013  ? Kathryn Louise Proudfoot, 2013 ii  Abstract Many concerns over the welfare of dairy cattle occur during the time around parturition. As cows transition from a pregnant to a lactating state, they are at high risk of disease and other painful conditions. In most intensive housing systems, these ?transition? cows are also kept in environments designed for the ease of management, with little consideration given to the expression of natural behaviours. This thesis addresses two main themes that are currently missing from the transition cow literature: 1) using knowledge of behaviour to improve management and housing practices, and 2) using behaviour as an indicator of poor health. Chapters 3, 4 and 5 address the first theme, and provide evidence that common management practices that disturb cows during parturition may interfere with calving; cows moved from a group pen into an individual pen during a late stage of labour spent more time standing in the hour before calving and experienced prolonged stage II labour compared to those moved earlier. Next, two preference studies were used to determine the type of environments that cows prefer during parturition. Results suggest that cows prefer to be in an undisturbed, secluded environment during labour and calving. To address the second theme, Chapter 2 describes the growing evidence in the human and laboratory animal literature that social behaviour can be useful as both an indicator of illness, as well as an early predictor of disease. Yet, there is little research to date making this link in farm animals.  The remaining chapters describe studies that used behaviour to identify cows with three major health problems: infectious disease, dystocia and lameness. Cows with infectious diseases ate less, spent more time lying and secluded themselves from a nearby group pen, all common sickness behaviours in other species. Feeding, social and standing behaviours were also found to predict cows at-risk for dystocia and lameness well before diagnosis. Collectively, these results provide evidence that a better understanding of transition cow behaviour can be useful to both improve housing and management, as well as identify cows at-risk for poor health. iii  Preface In all of the experiments described in the chapters of this thesis, animals were cared for according to a protocol approved by the University of British Columbia?s Animal Care Committee (A10-0163). For the experiments conducted in Denmark (Chapters 3 and 5), animals were cared for according to a protocol approved by the Danish Animal Experiments Inspectorate, Ministry of Justice, Copenhagen, Denmark.   A version of Chapter 2 has been published: Proudfoot, K. L., D. M. Weary, and M. A. G. von Keyserlingk. 2012. Linking the social environment to illness in farm animals. Appl. Anim. Behav. Sci. 138:203-215. The paper was co-authored by Kathryn?s main supervisor, M. A. G. von Keyserlingk and supervisory committee member, D. M. Weary. Co-authors supervised, helped interpret material, and edited drafts. The main concepts and ideas outlined in the manuscript were developed by Kathryn L. Proudfoot.  A version of Chapter 3 has been published: Proudfoot, K. L., M. B. Jensen, P. M. H. Heegaard, and M. A. G. von Keyserlingk. 2013. Effect of moving dairy cows at different stages of labour on behaviour during parturition. J. Dairy Sci. 96: 1638-1646. This paper was co-authored by Kathryn?s main supervisor, M. A. G. von Keyserlingk, a member of her supervisory committee, M. B. Jensen, and a colleague the National Veterinary Institute, Technical University of Denmark, P. M. H. Heegaard. The research was conducted at Aarhus University?s Research facilities in Foulum, Denmark, and was overseen by M. B. Jensen. Co-authors supervised, helped interpret material, and edited drafts. The main ideas for the manuscript were researched by Kathryn L. Proudfoot.  A version of Chapter 4 has been submitted for publication: Proudfoot, K. L., D. M. Weary, and M. A. G. von Keyserlingk. Seeking privacy: maternal isolation behaviour of Holstein dairy cows. This paper was co-authored by M. A. G. von Keyserlingk, Kathryn?s main supervisor and D. M. Weary, a member of Kathryn?s supervisory committee. Co-authors supervised, helped interpret material, and edited drafts. The main ideas for the manuscript were researched by Kathryn L. Proudfoot. A version of Chapter 5 has been submitted for publication: Proudfoot, K. L., M. B. Jensen, D. M. Weary, and M. A. G. von Keyserlingk. Dairy cows prefer an isolated area to calve and when ill. This paper was co-authored by Kathryn?s main supervisor M. A. G. von Keyserlingk, and two members of her supervisory committee, D. M. Weary and M. B. Jensen. The research was conducted at Aarhus University?s Research facilities in Foulum, Denmark, and was overseen iv  by M. B. Jensen. Co-authors supervised, helped interpret material, and edited drafts. The main ideas for the manuscript were researched by Kathryn L. Proudfoot.  A version of Chapter 6 has been published: Proudfoot, K. L., J. M. Huzzey, and M. A. G. von Keyserlingk. 2009. The effect of dystocia on dry matter intake and behaviour of Holstein cows. J. Dairy Sci. 92: 4937-4944. This paper was co-authored by Kathryn?s main supervisor,   M. A. G. von Keyserlingk and a colleague at the University of British Columbia, J. M. Huzzey. Co-authors supervised, helped interpret material, and edited drafts. The main ideas for the manuscript were researched by Kathryn L. Proudfoot.  A version of Chapter 7 has been published: Proudfoot, K. L., D. M. Weary, and M. A. G. von Keyserlingk. 2010. Behaviour during transition differ for cows diagnosed with claw horn lesions in mid-lactation. J. Dairy Sci. 93: 3970-3978. This paper was co-authored by Kathryn?s main supervisor, M. A. G. von Keyserlingk and supervisory committee member D.M. Weary at the University of British Columbia. Co-authors supervised, helped interpret material, and edited drafts. The main ideas for the manuscript were researched by Kathryn L. Proudfoot.   The first pages of these chapters have similar information in the footnotes. v  Table of Contents  Abstract ......................................................................................................................................... ii Preface .......................................................................................................................................... iii Table of Contents ...........................................................................................................................v List of Tables ............................................................................................................................... vii List of Figures ............................................................................................................................ viii List of Abbreviations ................................................................................................................... ix Acknowledgements ........................................................................................................................x Dedication .................................................................................................................................... xii 1. Introduction ...............................................................................................................................1 1.1. Background ...............................................................................................................................1 1.2. Using knowledge of behaviour to improve management and housing practices .....................2 1.3. Using behaviour as an indicator of poor health ........................................................................7 1.4. Specific research questions of this thesis ...............................................................................10 2. Linking Social Behaviour to Illness in Farm Animals .........................................................11 2.1. Introduction ............................................................................................................................11 2.2. The effect of the social environment on disease risk..............................................................12 2.3. The effect of disease on social behaviour ...............................................................................24 2.4. Social factors and disease in farm animals: where to go from here? .....................................26 2.5. Conclusions ............................................................................................................................28 3. The Effect of Management during Parturition on Behaviour and Labour  ......................30 3.1. Introduction ............................................................................................................................30 3.2. Materials and methods ............................................................................................................31 3.3. Results ....................................................................................................................................37 3.4. Discussion ...............................................................................................................................41 3.5. Conclusions ............................................................................................................................43 4. Preference for a Secluded Calving Site .................................................................................44 4.1. Introduction ............................................................................................................................44 4.2. Materials and methods ............................................................................................................44 4.3. Results ....................................................................................................................................47 4.4. Discussion ...............................................................................................................................50 4.5. Conclusions ............................................................................................................................51 vi  5. Preference for a Secluded Area during Calving and Infectious Illness ..............................52 5.1. Introduction ............................................................................................................................52 5.2. Materials and methods ............................................................................................................53 5.3. Results ....................................................................................................................................58 5.4. Discussion ...............................................................................................................................61 5.5. Conclusions ............................................................................................................................63 6. Behavioural Indicators of Dystocia ........................................................................................64 6.1. Introduction ............................................................................................................................64 6.2. Materials and methods ............................................................................................................65 6.3. Results ....................................................................................................................................68 6.4. Discussion ...............................................................................................................................74 6.5. Conclusion ..............................................................................................................................76 7. Behavioural Indicators of Lameness .....................................................................................77 7.1. Introduction ............................................................................................................................77 7.2. Materials and methods ............................................................................................................78 7.3. Results ....................................................................................................................................83 7.4. Discussion ...............................................................................................................................90 7.5. Conclusions ............................................................................................................................91 8. General Conclusions and Discussion .....................................................................................92 8.1. Summary of thesis ..................................................................................................................92 8.2. Using knowledge of behaviour to improve housing and management ..................................92 8.3. Using behaviour as an indicator of poor health ......................................................................96 8.4. Final conclusions ..................................................................................................................101 Bibliography ...............................................................................................................................102 vii  List of Tables Table 2.1. Models of social stress used in laboratory and farm animals .......................................15 Table 2.2. The effect of social stressors on immunity in studies using farm animals ...................23 Table 3.1. Definitions of calving signs ..........................................................................................34 Table 3.2. Signs of calving used to create movement categories ..................................................35 Table 3.3. Ethogram of behaviours occuring before calving ........................................................36 Table 6.1. Discriminant analysis of behaviours used to predict dystocia......................................74 Table 7.1. Feed composition of pre- and post-calving diets ..........................................................79 Table 7.2. Feeding behaviour of cows with and without hoof lesions ..........................................86 Table 7.3. Behaviours predictive of hoof lesions ..........................................................................89     viii  List of Figures Figure 2.1. Linkages between management practices and disease risk .........................................13 Figure 2.2. Individual differences in disease risk ..........................................................................21 Figure 3.1. Labour length and abdominal contractions .................................................................38 Figure 3.2. Lying behaviour in the hour before calving ................................................................40 Figure 4.1. Design of maternity pens ............................................................................................46 Figure 4.2. Choice of calving site ..................................................................................................48 Figure 4.3. Use of shelter and distance from partner before calving.............................................49 Figure 5.1. Design of individual maternity pens ...........................................................................54 Figure 5.2. Choice of calving site in individual maternity pens ....................................................58 Figure 5.3. Time spent in secluded corner.....................................................................................58 Figure 5.4. Behaviour of ill and health cows .................................................................................60 Figure 6.1. Feeding behaviour of cows with and without dystocia ...............................................69 Figure 6.2. Cumulative DMI of cows with and without dystocia .................................................70 Figure 6.3. Meal size after calving for cows with and without dystocia .......................................71 Figure 6.4. Water consumption of cows with and without dystocia .............................................72 Figure 6.5. Standing behaviour of cows with and without dystocia .............................................73 Figure 7.1. Standing behaviour of cows with and without hoof lesions .......................................84 Figure 7.2. Standing location of cows with and without hoof lesions...........................................85 Figure 7.3. Pattern of daily feed intake for cows with and without hoof lesions ..........................87 Figure 7.4. Social behaviour of cows with and without hoof lesions ............................................88  ix  List of Abbreviations   ADF = acid detergent fiber ANOVA = analysis of variance AUC = area under the curve BW = body weight CI = confidence intervals CP = crude protein DM = dry matter DMI = dry matter intake HPA = hypothalamic-pituitary-adrenal axis IL = interleukin  LPS = lipopolysaccharide LS = least-squares means N:L = neutrophil to lymphocyte ratio NDF = neutral detergent fiber NEL = net energy of lactation NK = natural killer cells NSAID = non-steroidal anti-inflammatory OR = odds ratio R2 = coefficient of determination ROC = receiver operating characteristics SAM = sympathetic-adrenal-medullary pathway SE = pooled standard error SEM = standard error of the mean SNS = sympathetic nervous system STD = standard deviation Th = T-lymphocyte helper cells TMR = total mixed ration ??F? = tumor necrosis factor alpha WBC = white blood cells x  Acknowledgements   First and foremost I thank my supervisor, Dr. Nina von Keyserlingk. I would not be the scientist, the teacher and the person that I am today without the inspiration and support that she has provided me over the last 8 years. I could not have asked for a better supervisor, and feel extremely privileged to have the opportunity to work with and learn from her day after day.   I also sincerely thank my supervisory committee members, Drs. Dan Weary and Margit Bak Jensen. Dan?s role in my thesis was much like a supervisor, as he was always willing to provide feedback and help. His mentorship has profoundly shaped the way I think, the way I write and the way I do research. Margit was a wonderful addition to my committee, as she provided incredible support and knowledge during the last part of my Ph.D. During the 6 months I spent with her in Denmark, she became both a colleague and a dear friend. Thanks also to Greg Miller, a former committee member, for his valuable insights on my review paper.  I must acknowledge two other members of the Animal Welfare Program: Dr. David Fraser and Chris McGill. Words cannot describe the influence that David has had on me. The impact that he has made on the Animal Welfare Science community is unmatched, and it is my hope that he never retires. Chris was always able to point me in the right direction when I was lost, which usually happened daily. I feel extremely lucky to have had them both in my life.  I thank all of the professors, adjunct professors and farm staff at the UBC Dairy Education and Research Centre for teaching me everything I know about dairy cattle. Nelson Dinn started with a city-girl and transformed her into a dairy cow expert with his wisdom and mentorship. Jim Thompson will always be an inspiration to me for his dedication to students and to moving the dairy industry forward step by step. Dr. Doug Veira?s thoughtful comments and questions never failed to make me think about the world slightly differently. Drs. Jeff Rushen and Anne-Marie de Passill? enriched my time at the centre with their insights and critical comments. Many thanks must also go to my trusty assistant and friend, Kathleen Ma, who was always excited to jump into a new research adventure with me. And finally, Barry, Brad, Ted and Bill: you are my second family, and you can definitely count on me to come back and visit!  xi  I was very privileged to be able to spend 6 months conducting research in Denmark during my Ph.D., so must also thank the faculty, staff and students at The Aarhus University Animal Science Department, as well as the The Aarhus University Research Foundation for partially supporting my stay. In particular, I thank John Misa Obidah and Erik Luc Decker for their assistance with my data collection, as well as Lene Munksgaard and Peter L?vendahl for their warm hospitality.  Over the years I have gained friendships within the Animal Welfare Program, the Dairy Centre and Vancouver that will last a lifetime. Speacial thanks to my dear friend Gosia for her love and friendship throughout the years. Thanks also go out to Joanna Makowska, Tan Lee, Miriam Gordon, Alejandra Barrientos, Kiyomi Ito, Pablo Restrepo, Julie Huzzey, Erin Ryan, Amelia Macrae, Liv Baker, Kristen Walker, Xavier Sandoval and Nuria Chapinal. Thanks for the humor, the love, and all of the great parties.   I could never have completed this Ph.D. without the love and support of my family. My parents always inspired me to do what I loved most, and, Mom and Dad: this is definitely it! Without their support I would not be able to spend the rest of my life doing what I am most passionate about. Alisa, Holly and James, thanks for making me laugh for 31 years. And James, your statistics advice was very much appreciated. I owe you a lot of sushi.   Lastly, I acknowledge and thank my loving partner, Magnus, for waking me up every morning with fresh coffee and a big smile. You make me happier then you will ever know.  Katy Proudfoot, September 2013 xii  Dedication           For my wonderful family and loving partner, Magnus   1  1. Introduction  1.1. Background Concern over the welfare of food animals in North America and Europe has grown considerably in recent decades, leading to legislation in a number of countries (Croney and Millman, 2007; Olynk, 2012). The science of animal welfare has developed as one approach to resolve some of these concerns (e.g., Brambell, 1965 in the UK). However, sufficiently defining good and poor welfare for animals has been difficult, as it involves interpreting scientific findings within cultural values about animals (Fraser et al., 1997; Ohl and Staay, 2012). To help direct research in this area, Fraser et al. (1997) provided a scientific framework for the study of animal welfare that integrated science with three common ethical concerns about the quality of life of animals: 1) that animals should function properly and be healthy, 2) that animals should feel well and be free of negative affective states, and 3) that animals should live naturally according to their adaptations and capabilities.  The chapters of this thesis address concerns over the welfare of dairy cattle during the sensitive ?transition? period from 3 wk before to 3 wk after calving. During transition, all three concerns outlined by Fraser et al. (1997) may be compromised. For example, cows are at high risk of infectious and metabolic disease after calving, leading to poor health and biological functioning (Ingvartsen, 2006). Many cows also experience painful conditions such as dystocia and lameness, affecting their health but also contributing to negative affective states (Huxley and Whay, 2006). Moreover, a variety of management and housing practices also occur during this period, potentially affecting a cow?s ability to perform her natural maternal behaviours (von Keyserlingk and Weary, 2007).  These welfare concerns are also often associated with production and economic loss for dairy producers. For instance, cows with infectious diseases, lameness and dystocia have lower milk production compared to healthy cows, and are at higher risk of being culled (Enting et al., 1997; Seegers et al., 2003; Wittrock et al., 2011; Barrier and Haskell, 2011). These conditions also often require veterinary assistance and medication that add even greater cost to the farm (e.g., Kossaibati and Esslemont, 1997). Given the clear impact of this time period on the cow and the farm, decades of research has focused on the transition dairy cow (see review by LeBlanc et al., 2006); but despite advancements in epidemiology, nutrition and immunity, many welfare concerns persist. A review 2  of the literature, covered in the following sections and in Chapter 2, illustrates some of the major gaps in transition cow research. These gaps centre on a lack of understanding about the behaviour of transition cows, and can be separated into two main themes: 1) Using knowledge of behaviour to improve management and housing practices 2) Using behaviour as an indicator of poor health  1.2. Using knowledge of behaviour to improve management and housing practices There is great variation in management practices during the transition period. Many of these practices were designed for the ease of management, with little consideration given to the expression of the cow?s natural behaviour. In a recent review, von Keyserlingk and Weary (2007) suggested that maternal behaviours might provide great insight into improving the care and management of transition cows. To address this topic, this section will begin with a definition of maternal behaviour, and a description of these behaviours in dairy cows. Then, an overview of the management that occurs on many commercial dairy farms will be given, followed by a discussion of two methods that can use knowledge of behaviour to improve management decisions. 1.2.1. Maternal behaviour in cattle For most mammals, the survival of young depends on the care given by mothers, especially in early life.  Although there are a number of definitions of ?maternal behaviour?, here it will be defined as ?the suite of behaviours expressed by mothers in late gestation? and directed towards the offspring throughout lactation until the young are weaned? (Dwyer, 2008). For the purpose of this thesis, this includes both behaviours of the dam directly aimed at the calf, as well as those expressed in preparation of, and during, calving. For beef and dairy cows housed in extensive, outdoor systems, maternal behaviour begins when the cow seeks isolation from the herd to find a suitable place to calve (Lidfors et al., 1994).  These cows prefer a birth-site that is dry and has some cover overhead such as tree branches (Lidfors et al., 1994). Few other behaviours before calving have been reported in extensively-housed cattle; however, the primary research about maternal behaviours directed at the calf has come from beef cattle, as they raise their calves until weaning at approximately 6 to 7 months (e.g., Lidfors and Jensen, 1988; St?hulov? et al., 2013). In the first few hours of birth, dams spend most of their time licking the calf and the calf begins to nurse (Lidfors and Jensen, 1988); they then spend progressively more time further away, yet nurse at about the same frequency (St?hulov? et al., 2013). As in other mammals, there is a high variation in care provided to calves, 3  which is partially influenced by aspects of the dam such as body condition and parity, as well as aspects of the calf such as sex and weight (St?hulov? et al., 2013). For dairy cattle housed in intensive, indoor systems, the calf is removed soon after birth, limiting the amount of time maternal care is provided after calving. However, the dam expresses a number of behaviours during late gestation and early lactation that are quite different from other stages of lactation. These behaviours begin in the first stage of labour when the calf is moving into its appropriate position for birth. During this stage, the cow becomes restless, characterized by a high number of position changes (Huzzey et al., 2005; Miedema et al., 2011; Jensen, 2012). She also spends more time paying attention to her abdomen (Jensen, 2012), increasingly raises her tail while standing (Miedema et al., 2011), and spends less time eating and drinking during this period (Jensen, 2012). The second stage of labour begins with abdominal contractions that help propel the calf through the birth canal (Noakes et al., 2001). During this stage, the cow is often recumbent to help facilitate calf delivery (Schuenemann et al., 2011).  After calving, the calf is removed from the dam, sometimes within a few minutes to hours of birth. To investigate the maternal behaviour of dairy cows directed at the calf, Jensen (2012) kept the dam with the calf for the first 24 h after birth. Cows spent most of the first few hours intensively licking the calf. Thereafter, they progressively decreased the time spent directing behaviours toward the calf, and spent more time lying and feeding. The topic of maternal behavior in cattle is covered briefly here, but has been extensively reviewed elsewhere (von Keyserlingk and Weary, 2007). 1.2.2. Management of transition dairy cows For the purpose of this thesis, the transition period can be separated into three phases: 1) the pre-calving, ?dry? or ?close-up? phase beginning 3 wk before calving and ending when calving is imminent, 2) the parturient phase in the hours just before and during labour, and 3) the post-calving or ?fresh? phase lasting from calf delivery until 3 wk later.  Management of cows during these phases is often dependent on the size and the nutritional regime used on the farm (Overton and Waldron, 2004). For instance, the industry-standard for nutritional management in the pre-calving phase in North America consists of a two-group scheme, meaning that cows are moved from a low-energy diet to a high-energy diet approximately 3 wk before calving (Overton and Waldron, 2004). Each diet change also reflects a pen change; cows are usually regrouped at least once before calving (Cook and Nordlund, 2004).  When signs of calving are present, or the cow has reached her expected calving date, she is either left in a group pen or moved to an individual pen to calve (Cook and Nordlund, 2004). 4  On many large farms, cows are housed in groups for calving, as it would be impractical to move a large number of cows that are expected to calve each day into individual pens. On smaller farms, and on all farms in Denmark where individual maternity pens are compulsory (Danish Ministry of Justice Law no. 520, May 26 2010), cows are moved from the group into an individual maternity pen to calve. Regardless of the type, these pens are often in high traffic areas allowing producers to frequently monitor the progress of labour. Post-calving, cows are switched to a high-energy diet to support their lactation. On some farms, cows are moved into a ?fresh? pen for the first few days after calving; the time spent in this pen often coincides to when their milk is not saleable, thus easing the work of the milker (Cook and Nordlund, 2004). Cows are then either moved into a post-fresh monitoring pen, or directly into the lactating group. According to Cook and Nordlund (2004), the ?traditional? movement of cows during transition can include as many as 5 pen changes during the 6 wk period. 1.2.3. The effect of management and housing on behaviour The high variation in management practices during transition is partly due to aspects of the farm, but may also be due to a lack of understanding about the optimal environment for cows during parturition. One way to improve housing and management practices on dairy farms is to gain a better understanding of how these practices affect cow behaviour and health. Much of the research in this area has focused on the effect of management practices on feeding behaviour and optimal feed intake, particularly during the pre- and post-calving phases of transition (reviewed by Grant and Albright, 1995; Drackley, 1999). Ensuring optimal feed intake is important, as it is related to lower health risks after calving (Huzzey et al., 2007; Goldhawk et al., 2009). However, practices such as overstocking and regrouping cows during transition and other stages of lactation have been shown to reduce feed intake and alter feeding behaviour (von Keyserlingk et al., 2008; Proudfoot et al., 2009; Schirmann et al., 2011).  Research using mid-lactation cows has also found that management and housing can influence lying time. Lying time decreases when the stocking density of the feed bunk and free stalls increase (Fregonesi et al., 2007b; Proudfoot et al., 2009), and when the stall surface is uncomfortable, such as when the bedding material is moist or insufficient (Tucker and Weary, 2004; Fregonesi et al., 2007a).  Most of the research in this area has focused on mid-lactation cows, or those in the pre- and post-calving phases of transition. To my knowledge, no research had identified the effect of housing and management practices on cow behaviour during the parturient phase. This phase may be particularly important, as disturbing cow behaviour may also prolong labour, which is thought 5  to be very painful (Mainau and Manteca, 2011). It is also important to consider the unique management practices that occur during this phase compared to other stages in a cow?s life cycle, such as moving cows from a group pen into an individual pen while labour is in progress.  1.2.4. Improving practice using preference and motivation testing A second method that may help improve housing and management practices during transition is to measure which aspects of the environment are important to the cow. To do this, researchers can give cows some control over their environment and observe the choices they make. In dairy cows and other animals, this has been done using preference and motivation testing, where animals are given various options in their environment and are able to choose, or work for, the desired option (Kirkden and Pajor, 2006).  Kirkden and Pajor (2006) review preference and motivation tests as they apply to animal welfare; as this is an extensive topic, it will be briefly reviewed here and re-addressed in the conclusion (section 8.2). Although there are many different definitions of motivation, these authors define it ?as the tendency for an animal to perform a behaviour?, but also recognize that motivations can reflect the animal?s desire to perform the behaviour (Kirkden and Pajor, 2006). Motivations can vary in strength, and can be caused by both internal (e.g., neural and hormonal) and external (e.g., environment) sources. A ?preference? refers to an animal having a higher motivation for one option over others.  ?Preference tests? are used to determine the preference of one resource over one or several alternatives (e.g., Jensen et al., 2008). To determine preference, researchers commonly use ?choice tests? that give animals the option to choose between various environments or resources. For example, this can be done instantaneously using T- or Y-mazes, where animals enter the maze and make choices between the options at either arm of the maze. If an animal chooses one arm over another, then it can be said that the animal prefers the option in that arm (e.g., Pajor et al., 2003). A choice test can also be longer in duration; in these cases, animals are housed in more natural ?home? environments, such as their own pen or cage, and are given two options with ample time to use both. The duration of time spent using each option is measured, and the option with the most use is considered to be preferred (e.g., Fregonesi et al., 2009). Although no research has used choice tests with parturient cows, these tests have been used on mid-lactation cows to address questions regarding preferred management practices and housing features. For instance, instantaneous choice tests have been used to determine cow preference for gentle or no talking over shouting (Pajor et al., 2003), and feeding over milking (Prescott et al., 1998). In addition, choice tests of longer duration have been used to determine 6  aspects of stall design that cows prefer, such as an open pack over a free stall (Fregonesi et al., 2009), a deep-bedded stall surface over mattresses (Tucker et al., 2003), and dry bedding over wet bedding (Fregonesi et al., 2007a). Preference tests provide valuable insight into environments that animals prefer; however, these tests are unable to provide information about the importance of the resource to the animal (Fraser, 1993). To address this problem, ?motivation tests? were developed to measure the strength of the motivation or preference (Jensen and Pedersen, 2008). The strength of a motivation can be determined by making the animal pay a ?price? before they are allowed access to the resource, and then increasing the price until the animal stops paying for the resource (drawing from consumer demand theory used in microeconomics, Dawkins, 1990).  ?Operant tests? are often used to determine the strength of motivations; animals are taught to perform a task before given access to a resource and the difficulty of the task is then increased according to a schedule (Kirkden and Pajor, 2006). The price of the task can be time (e.g., pressing a lever or panel, Holm et al., 2002), weight (e.g., pushing through a weighted door, Olsson and Keeling, 2002) or distance travelled (Schutz et al., 2006), amongst others. The animals? response to the task can then be used to infer the relative strength of the motivation or importance of the resource to the animal (e.g., by calculating the price elasticity or maximum price paid, see Jensen and Pedersen, 2008 for a review). Motivation tests have been used to determine the strength of motivations in mid-lactation cows to obtain resources such as feed (Schutz et al., 2006), shade (Schutz et al., 2008) and access to pasture (Charlton et al., 2013), and to perform behaviours such as lying (Jensen et al., 2005). For example, Schutz (2006) found that cows will walk longer distances to obtain food when they are food-deprived compared to when they were not, suggesting that they are more motivated (and more willing to work) to obtain food as hunger increases. However, no research to my knowledge has used preference and motivation testing to determine preferred or important housing features for transition cows. 1.2.5. Gaps in the literature The previous sections have described two methods for using behaviour to improve management: 1) by studying the effect of management on behaviour, and 2) by measuring the importance of aspects of the environment to the cow. However, there has been limited research using these methods to improve the management and housing of transition cows, particularly during the parturient phase just before calving.  7  One major gap in the literature is a lack of research on the effect of management practices on maternal behaviour and labour progress. Chapter 3 addresses this question by moving cows from a group pen into an individual maternity pen at different stages of labour and measuring behaviour and length of the second stage of labour. Another gap in the literature is a lack of research addressing cow preference and motivation for features of her calving environment. Chapter 4 and 5 take the first steps to address this topic. Chapter 4 describes a preference test whereby cows were given two options for a calving site: an open area exposed on all sides and a sheltered environment covered on all sides except for a door allowing free entry and exit. These environments were chosen based on evidence that wild ungulates, as well as dairy and beef cows housed on pasture, seek a covered calving site when it is available (Lidfors et al., 1994; Barbknecht et al., 2011). Cows were housed in this pen for a few days before calving, and the duration of time they spent in each area was recorded before and during calving. Chapter 5 describes a practical individual pen that allowed cows to have access to both a secluded corner and an open window to a group pen and measured where cows chose to be during and after labour.   1.3. Using behaviour as an indicator of poor health In addition to aiding in management and housing decisions, behaviour is becoming an important tool to identify ill animals. Behaviours have been associated with ill health since the ancient Egyptians (Griffith, 1898) and Aristotle (Treatise: ?On the Sacred Disease? 400 BC). More recently, research has determined that an animal?s behaviour can indicate pain and malaise associated with poor health, but can also predict those at risk (reviewed by Weary et al., 2009). The following sections will focus on the usefulness of behaviour as both an indicator, as well as an early predictor of poor health in transition cows. 1.3.1. Sickness behaviour Changes in behaviour have long been associated with illness, but were considered to be a debilitating outcome until the late 1980s (Hart, 1988; Dantzer and Kelley, 1989). Hart (1988) and Dantzer and Kelley (1989) suggested that these behavioural changes (or ?sickness behaviours?) were an adaptive and strategic response meant to help the animal fight the illness. Later, it was suggested that these behaviours reflect motivational changes occurring during illness (Aubert et al., 1997). Since that early work, researchers have determined the physiological pathway linking illness and behaviour (reviewed by Dantzer, 2001). Briefly, as summarized by Dantzer (2001), at 8  the first sign of pathogens the body mounts an immune response. Cytokines, the communicators of the immune system, are released in the brain and signal an increase in body temperature (a febrile response). As both the febrile and immune responses require a great deal of energy, these cytokines also signal an increase in energy-conserving behaviours such as rest and inactivity. Energy-consumptive behaviours are also reduced, including feeding, grooming, reproductive behaviour and other social behaviours. The combination of these behavioural and physiological changes creates an environment that can effectively help to fight the pathogen.  1.3.2. The behaviour of ill cows In the past 10 years, researchers have begun to link behaviours with illnesses of transition cows. In general, the results have followed the classic signs of sickness behaviour in other species. For instance, Huzzey et al. (2007) found that cows with clinical and subclinical metritis ate less, spent less time at the feeder and visited the feeder less often compared to healthy cows. Cows with ketosis, mastitis and inflammatory hoof lesions show a similar decline in feeding behaviour and intake (Galindo and Broom, 2002; Almeida et al., 2008; Goldhawk et al., 2009; Siivonen et al., 2011; Fogsgaard et al., 2012). Cows with mastitis also decreased self-grooming (Fogsgaard et al., 2012), and cows with lameness increased their time spent lying and resting (Almeida et al., 2008). Some research has also found non-classical signs of illness. For example, cows with both naturally-occurring and experimentally-induced mastitis spent more time standing (though they stand ?idly?) during the disease compared to baseline (Siivonen et al., 2011; Fogsgaard et al., 2012; Cyples et al., 2012). This behaviour, though not a classic sign of illness, is likely reflective of motivational changes associated with increased pain; cows with mastitis chose to stand and relieve pain and pressure from their udder rather than of lie down and rest. 1.3.3. The predictive power of behaviour In addition to identifying ill animals, behaviour can also be an early predictor of animals that are at-risk for poor health (Weary et al., 2009). In a recent review, Sep?lveda-Varas et al. (2013) described the growing evidence that feeding, standing and social behaviours can be useful predictors of common diseases occurring during transition.  Perhaps one of the most well studied of these behavioural predictors of health is feeding behaviour. Cows normally decrease their intake as calving approaches (VandeHaar et al., 1999), but these declines appear to be more severe in cows at-risk for disease. For example, Huzzey et al. (2007) found that cows diagnosed with metritis after calving had lower feed intake, spent less time at the feed bunk and visited the feed bunk less often than healthy cows beginning 2 wk 9  before calving and 3 wk before clinical signs of disease. In a similar study, Goldhawk et al. (2009) found that cows with subclinical ketosis after calving showed declines in intake and feeding behaviour during the week before calving.  Standing and lying behaviour may also be a useful indicator of poor health. For example, Jawor et al. (2012) found that cows diagnosed with hypocalcaemia on the day of calving spent nearly 3 h longer standing during the day before calving compared to healthy cows. Finally, social behaviours might also be useful predictors of disease. Little research to date has addressed this relationship in dairy cows, but there is an expanding literature in humans, laboratory animals and other farm animals that illustrate the link between social behaviour and disease risk (see Chapter 2 for a detailed review of this topic). 1.3.4. Gaps in the literature  The research described in the preceding sections has broadened our understanding of the behaviour of ill cows and those at-risk of becoming ill. However, there remain some potentially useful areas of research missing from the transition cow literature. Firstly, there is a lack of research relating social behaviours to illnesses common after calving. In other species, social exploration is reduced during early stages of experimentally induced illness (Crestani et al., 1991; Bluth? et al., 1999). Changes in social behaviour may provide a more comprehensive picture of illness in dairy cows, yet, no work had determined if dairy cows would reduce social behaviour when clinically ill. Chapter 5 describes the first study to address the effect of illness on social behaviour. Cows with signs of infectious diseases were housed in an individual pen outfitted with a secluded area and an area that gave them visual and head to head contact with their group. The use of the secluded area was then compared to that of healthy cows. Secondly, there are some clear linkages between behaviours predictive of metritis, ketosis and hypocalcaemia, however, no work prior to my PhD research had attempted to identify behaviours predictive of two painful health conditions: dystocia and lameness. Dystocia, defined as difficult calvings that require farmer or veterinarian assistance, has been estimated to occur in 2 to 23% of calvings (Mee, 2008). Dystocia is considered painful (Huxley and Whay, 2006) and is associated with high risk of disease and reproductive complications after calving (Oltenacu et al., 1988). Mee (2008) reviewed the known risk factors for dystocia (e.g. calf size, calf position and pelvic size of the dam), but many of these provide little information to a producer before calving. Although behaviour could provide an earlier indication that a cow is at-risk for dystocia, only descriptive data had been available in the literature prior to the start of my thesis research. Thus, 10  Chapter 6 describes an experiment that determined the behavioural differences between cows with and without dystocia in the few days before and after calving. Lameness, defined as abnormal gait caused by multiple painful conditions in the hooves, feet and legs, is considered to be one of the largest welfare concerns facing the dairy industry today (Huxley and Whay, 2006). In a recent on-farm study in North America, von Keyserlingk et al. (2012) found the average prevalence of clinical lameness in free stall-housed cows to be 30% and 55%, depending on the region. A major cause of lameness is claw horn lesions, whereby the proper development of the hoof horn is disturbed, leading to an abscess or ulceration of the hoof (Bicalho et al., 2009). The highest risk of lesions occurs in the few months after calving, but the start of lesions is thought to occur during transition (Leach et al., 1998). There are likely multiple factors that cause cows to become lame after calving, including physiological changes, external influences on the hoof and cow behaviour (Webster, 2002; Tarlton et al., 2002; Bicalho et al., 2009). No research had identified behavioural risk factors for lameness during the transition period; Thus, Chapter 7 describes a study used to determine if behaviours during the transition period could identify cows that develop hoof lesions in the months after calving.  1.4. Specific research questions of this thesis The research presented in this thesis addresses some of the gaps outlined in the preceding sections by focusing on developing a better understanding of transition cow behaviour. Specifically, the hypothesis of this dissertation is that behaviour can be used to both develop suitable housing and management practices for cows during parturition, as well as identify cows with or at-risk for poor health. The following specific research questions were used to address this hypothesis:   1) Chapter 3: Does moving cows to a maternity pen during labour affect behaviour and labour progress?  2) Chapters 4 and 5: Do cows prefer a secluded or open environment for calving?  3) Chapter 5: When cows are ill, do they show common signs of sickness behaviour?  4) Chapter 6: Can behaviour be used to identify cows at-risk for dystocia?  5) Chapter 7: Can behaviour during transition be used to identify cows at-risk for lameness after calving?  11  2. Linking Social Behaviour to Illness in Farm Animals1  2.1. Introduction This chapter provides a critical review of the available literature linking social behaviour and illness in farm animals, as this was identified as a major gap in the transition cow literature. Although some of this research has been done using farm animals such as pigs, cattle and chickens, a majority of the research has come from the human and laboratory animal literature. Thus, the discussion will mainly focus on this body of literature, but will end with specific recommendations for future research using farm animals such as the transition dairy cow.  Despite decades of research, disease remains a major challenge of farm animal care. Ill animals have poorer welfare and are less productive than healthy animals, and the products from ill animals may be less safe for human consumption (CAST, 2012). Disease prevalence varies tremendously by species, region, year, management system and farm. For instance, in a survey of studies conducted in Europe and North America, Ingvartsen (2003) found that the lactational incidence of mastitis in dairy cows ranged from 2.8 to 39%. Similarly, over a 14-year period (1987 to 2001), Snowder et al. (2006) recorded a 5 to 44% range in respiratory disease incidence following transportation in feedlot cattle.   Research to date has advanced our knowledge of the causes of specific diseases and their treatments. Recently, researchers have shifted their focus from treatment to early disease detection and prevention (e.g., LeBlanc et al., 2006). Disease prevention strategies have been developed, utilizing vaccines and diets formulated to help protect against disease (VandeHaar and St-Pierre, 2006; LeBlanc et al., 2006). Technological advancements, as well as research in animal behaviour and veterinary epidemiology, have also enabled us to identify early predictors of disease (Stark, 2000; Mulligan et al., 2006; Weary et al., 2009).  One challenge in disease detection and prevention may be the structural change of animal industries during the last few decades. Animals are increasingly being housed in fewer, larger farms in both developed and developing countries (Fraser, 2008; MacDonald and McBride, 2009). However, it is unclear whether this increase in farm size has an impact on the incidence of disease. Some studies have reported lower morbidity and mortality with increasing farm size (Jenny et al., 1981; Hill et al., 2009; Hybschmann et al., 2011) and others report a higher                                                  1 A version of Chapter 2 has been published: Proudfoot, K. L., D. M. Weary, and M. A. G. von Keyserlingk. 2012. Linking the social environment to illness in farm animals. Appl. Anim. Behav. Sci. 138:203-215.  12  morbidity and mortality (Brooks-Pollock and Keeling, 2009; Gay and Barnouin, 2009; Gulliksen et al., 2009a). One reason for this difference is the method used to collect disease records; for instance, Hill et al. (2009) relied on producer accounts of disease occurrence, which may decrease with farm size due to a lowered capacity to detect sick individuals within a large group.  Farm size alone may be a poor indicator of morbidity, as many other management practices employed on farms can also influence disease risk. Epidemiologists have identified practices including health-related protocols (e.g., vaccine use), and grouping strategies (e.g., movement of animals into new pens) as determinants of disease. Grouping practices linked to high morbidity tend to involve moving animals between groups (i.e., ?regrouping?, ?remixing? or ?comingling?), and housing animals of variable ages in the same pen (Ribble et al., 1995; Lo Fo Wong et al., 2004; Svensson et al., 2006; Sanderson et al., 2008; Hultgren and Svensson, 2009). A limitation to this epidemiological work is the inability to determine the cause of illness, as management practices such as regrouping likely have both physical and social implications on the animals. For instance, finishing-pig herds consisting of pigs from more than three suppliers have a higher risk of testing positive for salmonella compared to those herds with their own replacement stock (Lo Fo Wong et al., 2004). Pigs in the herds with multiple suppliers likely have a greater exposure to pathogens, and must also cope with an unstable social environment compared to the pigs housed with familiar herdmates.  To determine the specific role of the social environment on management-related disease, some new research has determined the biological linkage between a farm practice, social stressors and the biological intermediaries of diseases. This review will describe the research to date assessing the role of an animal?s social environment on its disease risk.   2.2. The effect of the social environment on disease risk Figure 2.1 depicts the potential relationship between farm management practices and disease. Farm practices that act as social stressors can impact the ability of an animal to defend against disease. Animals that succumb to illness change their social behaviour as an adaptive mechanism to prevent secondary infections. The following sections will describe this relationship drawing on a conceptual framework developed in the human and laboratory animal literature.   13    Figure 2.1. Linkages between management practices and disease risk Flow diagram of the potential linkage between farm management practices and disease risk, including influences from both the social and physical environment.  2.2.1. Social stressors and health: a conceptual framework The primary experimental research linking the social environment with disease has tied ?stress? to biological intermediaries of disease (e.g., reduced immunity and resistance to infection) and consequent disease risk. The concept of stress in mammals has been reviewed extensively (e.g., Moberg, 2000; Sapolsky et al., 2000). For this review, a ?stressor? is a stimulus in an animal?s environment that exceeds the natural regulatory capacity of the animal, and is usually unpredictable and/or uncontrollable. The ?stress response? is the animal?s behavioural and neuroendocrine reaction to the stressor (Moberg, 2000; Koolhaas et al., 2011). Two systems work in conjunction to establish this state: the sympathetic nervous system (SNS) and the hypothalamic-pituitary-adrenal (HPA) axis. The complexities of the stress response and its various physiological indicators have been covered elsewhere; for this review we will only 14  mention glucocorticoids (i.e., cortisol and corticosterone) as one hormonal indicator of the HPA response, and catecholamines (epinephrine and norepinephrine) as indicators of the SNS activity, as we mention these in later sections. Since Selye (1956), researchers have determined that a variety of stressors can cause different stress responses in mammals. In a meta-analysis of human research, Dickerson and Kemeny (2004) discovered that HPA activity in response to a laboratory stressor was most pronounced during situations of social threat. ?Social stressors? (also called ?psychosocial stressors?) are factors of an animal?s social environment that initiate a stress response. The concept of social stress has also been considered a potential concern for animal agriculture since at least the 1970s (Wood-Gush et al., 1975). 2.2.1.1. Methodological approaches In the human literature, both short-term (?acute?) experimental stressors (e.g., a math test) and long-term  (?chronic?) ?real-life? stressors (e.g., bereavement over a loved one or care for an ill family member) have been used to study the influence of stress responses on disease outcomes. Laboratory animal models of social stressors have also been developed to study the effect of the social environment on human health (Table 2.1). In these models, a stress response is generally achieved by causing agonistic or conflict behaviours between two or more animals, by isolating an animal from its parent(s) or group, or by manipulating social status. These stressors can be acute (e.g. resident-intruder tests when two unfamiliar animals meet for the first time), chronic (e.g. weeks of social isolation), or intermittent (e.g. 2 h of social defeat over multiple days). Some researchers also use multiple paradigms simultaneously to increase the severity and duration of the animal?s stress response (e.g. intermittent social defeat combined with overcrowding). Decades of research using human subjects and animal models have established a relationship between social stressors and biological intermediaries of disease (i.e., biological processes that predict disease, such as compromised immunity), as well as disease onset and progression. This research has received much scepticism, primarily due to interpretations about health made using a few biological markers (e.g., a single immune component such as lymphocyte proliferation). In the human literature, methodological approaches have been suggested to address these criticisms. For instance, Cohen et al. (2007) reviewed the relationship between stress and disease with emphasis on the role of stressors and stress responses on specific biological determinants of four diseases (i.e., clinical depression, cardiovascular disease, human immunodeficiency virus and cancer).  15  Miller et al. (2009) suggests that researchers start with a disease and work backwards to link the disease to stressors in the environment. In this ?disease-centred? approach, researchers begin with an established linkage between a social factor and morbidity, for instance, through epidemiological studies that include social stressors as determinants of disease (e.g., social threat predicts AIDS progression). Researchers can then examine the influence of a particular social stressor on disease risk by first identifying the biological determinants of the diseases (e.g., decreased antiviral capabilities of the individual, Cole et al., 1996), and then assessing the impact of the stressor and stress response (e.g., appraisal of social threat and epinephrine response) on these biological determinants (e.g., epinephrine enhancement of viral replication, Cole et al., 2003). Cohen et al. (2007) and Miller et al. (2009) review a variety of diseases including depression, asthma, as well as autoimmune, cardiovascular and infectious disease. The following section will briefly describe the available research to date linking social stressors to changes in immunity and risks of infectious disease, as these are the most common in farm animals.  Table 2.1. Models of social stress used in laboratory and farm animals Models of social stress used in laboratory and farm animals adopted from Blanchard et al. (2001). ?Test? animal refers to the animal(s) experiencing the social stress. Stressors can be acute (e.g., one-time social defeat), intermittent (e.g., bouts of social defeat separated by time) or chronic (e.g., constant recurring social defeat)   Social stressor Definition Source of stress  Social defeat   Test animal exposed to confrontation with aggressive conspecific   Agonistic behaviour  Social instability Introducing test animal(s) into a group of unfamiliar conspecifics Agonistic behaviour Social disruption Introducing an aggressive animal (usually a male) into stable group of test animals Agonistic behaviour Crowding Group of test animals housed with confined physical space per animal Proximity to conspecifics/agonistic behaviour Social isolation Test animal is separated from the group and individually housed Separation from conspecifics  16  2.2.1.2. Social stressors, immunity and infectious disease Researchers have shown that not all social stressors cause biological harm to the animal; indeed, an initial stress response is beneficial and adaptive (McEwen, 1998). However, social stressors that persist for a long duration in an either chronic or intermittent fashion, or occur at a sensitive time of life, can cause pathological harm to the animal. Minutes to hours after exposure to a social stressor, leukocytes (i.e., immune cells that protect the body from infection) are redistributed from the circulation to lymph nodes and the skin (Dhabhar and McEwen, 1997). This change is thought to be an adaptive response to an acute stressor, as animals may need to protect their skin from infection following injuries that can occur during a confrontation. An acute stressor causes a reduction in circulating lymphocytes (lymphophilia), as these move from the blood stream to the lymph nodes and skin, and an increase in circulating granulocytes (granulocytosis, Stefanski and Engler, 1998).  Intermittent or chronic exposure to social stressors lasting days or weeks can have a detrimental impact on the immune system. Animals exposed to an intermittent and chronic social stressor are thought to undergo a ?generalized immune dysfunction? (Schmidt et al., 2010). This is characterized by a reduction in both circulating lymphocytes (Ruis et al., 2001a; Engler et al., 2004) and antibodies (Turner et al., 2000; Bartolomucci et al., 2003) as well as a reduction in the reactivity of immune cells to a pathogen challenge, such as an endotoxin or a viral vaccine (Hopster et al., 1998; de Groot et al., 2001). Elenkov and Chousos (1999) described a model that links the chronic stress-induced changes in immunity with clinical disease risk. This model relied on changes in the distribution of T-helper cells (Th), messengers that direct the distribution and stimulation of other immune cells. During a stress response, glucocorticoids and catecholamines signal the immune system to shift toward anti-inflammatory, humoral activity (i.e., a ?Th2? shift; responsible for eliminating extracellular bacterial challenges) and away from pro-inflammatory, cellular activity (i.e., a ?Th1? shift; responsible for eliminating intracellular viral challenges). These changes are thought to contribute to reductions in host resistance to viruses and the onset of several infectious diseases in rodents (e.g., Leishmania major and Theiler's virus, Johnson et al., 2004; Ehrchen et al., 2008) and HIV progression in humans (Clericia and Gene M. Shearer, 1993).  Social stressors persisting weeks or months results in a down-regulation of HPA activity and glucocorticoid output via negative feedback mechanisms in the brain (i.e., ?glucocorticoid resistance?, Miller et al., 2007). Glucocorticoid resistance results in a desensitization of lymphocytes to the anti-inflammatory actions hormone (Miller et al., 2002; Cole et al., 2009) and 17  a shift of the immune system back to an inflammatory Th1 response (Elenkov and Chrousos, 1999). Without the aid of anti-inflammatory glucocorticoids, chronic stress may lead to a state of mild inflammation that has been linked to coronary heart disease (reviewed by Black and Garbutt, 2002) and autoimmune disease (reviewed by Elenkov and Chrousos, 2002). For instance, 30 wk of social isolation in rats caused a shift in the immune response towards Th1 and glucocorticoid resistance, and these changes increased inflammation and exacerbated existing autoimmune disease (Chida et al., 2005). Although the Th1/Th2 model is well-developed and aids in our understanding of the relationship between social stress and immunity, it is likely not the sole mechanism responsible for stress-related disease risk (Kidd, 2003).  A second line of research provides evidence that social stressors impact the body differently depending on life stage; social stressors that occur during early life may have more profound impacts on disease risk compared to those that occur later in life (reviewed by Coe and Lubach, 2003). Much of this research has used the social isolation paradigm. Social isolation in neonatal monkeys caused fundamental changes in their immune responses as adults, and early maternal separation reduced resistance to viral infections of adult mice (Avitsur et al., 2006). Piglets isolated for two hours per day over the first few weeks of life showed reduced lymphocyte stimulation to LPS up to 45 d after the last isolation period (Kanitz et al., 2004). Maternal social stress during gestation can also affect the immunity of offspring. For example, repeated regrouping of gilts (female pigs) with unfamiliar animals during late gestation reduced the immune response of offspring up to at least 60 d of life (Couret et al., 2009). Research across multiple disciplines has aided our understanding of the influence of social stress on immunity and disease. Although chronic, intermittent and early life stressors have been discussed here, there are other aspects of the stressor that likely influence an animal?s risk of illness. The type of social stressor, degree to which the stressor is controllable and an individual?s perception of the stressor can all influence the subsequent neuroendocrine response (Dickerson and Kemeny, 2004). In the next section we discuss individual variation in stress responses and why some individuals may be better able to avoid social stress-related disease.  2.2.2. Individual variation in response to stressors and disease risk Individual animals differ in many ways, including their response to a particular social stressor. Stress responses vary between species, breed, sex, as well as individual characteristics such as personality, coping style and social status within a group. Although species and breed differences in response to social stressors exist (Bartolomucci, 2007; Fahey and Cheng, 2008), there has been little work using these differences to explain disease risk. Males and females have 18  been found to differ in their perception and physiological response to stressors (de Groot et al., 2001; Rohleder et al., 2001). In general, males tend to be commonly used as animal models of social stress, and these models rely heavily on the social defeat paradigm. Female models of social stress are beginning to emerge in the animal literature, and these models are relying more on social instability and isolation, rather than defeat (Haller et al., 1999; Taylor et al., 2000; Palanza et al., 2001). However, little evidence has linked these sex differences in response to disease risk. For this review we focus on current areas of research that have attempted to link individual stress responses with disease susceptibility; this research has focused on constructs of personality, coping styles and social status. 2.2.2.1. Defining personality, coping styles and social status  Personality is a main construct used by psychologists to classify individual differences in humans and animals. Broadly, personality refers to the characteristics of individuals that describe consistent patterns of affect, cognition, and behaviour (Pervin and John, 1997). These characteristics are sometimes categorized into discrete dimensions (e.g., Eysenck?s 3-part model or the Big Five Personality Dimensions, Gosling et al., 2003; John et al., 2008); within each dimension, individuals vary along a continuous scale. For instance, along the sociability dimension individuals vary from being not sociable to very sociable.  A second construct for classifying individuals focuses on the variation in the capacity to cope with stressors in the environment. Coping styles are roughly defined as persistent and correlated physiological and behavioural responses of animals to a number of stressors (Koolhaas et al., 1999; Veenema et al., 2003). Two main coping styles have been described in the literature: individuals that ?proactively? or actively cope, and those who ?reactively? or passively cope. Proactive coping is characterized by high SNS activity, low HPA response and more aggressive behavioural tendencies in response to stressors, whereas reactive coping describes individuals with behavioural inhibition, low-aggression and a high HPA response to stressors.  Researchers also classify individuals by their social status relative to other group members (e.g., dominance hierarchies, Stricklin and Mench, 1987). Social organization is different for every species; however, individual differences in social status are considered an emergent property of any group of animals (Chase et al., 2002). Individuals are typically delineated into high, middle and low status, depending on the number of wins versus losses in agonistic interactions.  19  These three constructs have many overlaps. For instance, animals of low social status show similar behavioural and neuroendocrine responses to those with reactive coping styles, and animals of high social status share traits with proactive individuals (Sapolsky, 1990; Koolhaas, 2008). Mendl et al. (1992) investigated the relationship between an aggressive personality trait and social status in gilts and found that both high ranking and middle ranking pigs showed the same amount of aggressiveness. Low ranking animals, however, showed no or very low aggressiveness. Lindberg (2001) suggests that social status is more flexible than coping style and personality; individuals can shift their social status depending on context, whereas personalities and coping styles are generally elemental to the animal and thus consistent across situations.  2.2.2.2. Individual differences in response to social stressors and disease susceptibility All three constructs described in the previous section explain some variation in disease risk. Several reviews have addressed the relationship between personality traits and disease risk (Segerstrom, 2000; Friedman, 2008). Notably, personality types predict the capacity to which an individual is able to cope with a stressor. Because individuals with different personality types respond to stressors differently, they are thus at risk for different pathological outcomes. For instance, people with ?A? type personalities ? characterized by bold, aggressive traits ? are at higher risk for coronary heart disease than those with less aggressive personalities (Suls and Wan, 1993). Humans and rhesus macaques scoring low on the sociability dimension have higher risk of infectious disease (Capitanio, 1999; Cohen et al., 2003; Capitanio et al., 2008). For instance, Capitanio (1999) showed that rhesus macaques with lower sociability that are inoculated with simian immunodeficiency virus had a slower decline in plasma glucocorticoids following the inoculation and fewer antibodies to the virus.  The coping style literature has defined individuals partly by their neuroendocrine response to stressors, and thus has been able to identify a possible biological linkage between an individual?s style and their risk of disease. Figure 2.2 illustrates this linkage. Animals with reactive styles and more pronounced HPA responses to social stressors are likely at higher risk of infection, whereas rats with proactive styles and stronger SNS responses to stressors may be at higher risk of cardiovascular disease such as hypertension. Though little experimental work is available to support this idea, some work with rats has identified a possible link between coping style and shifts in the immune response. Kavelaars et al. (1997) found that rats with reactive coping styles and higher HPA activity shifted toward a Th2 response. This shift was associated with higher susceptibility to infectious disease such as periodontitis (Breivik et al., 2000). In contrast, rats with proactive coping styles and hypoactive HPA activity showed a shift toward a 20  Th1 response, and a higher susceptibility to inflammatory or autoimmune disease (Kavelaars et al., 1997; Kavelaars et al., 1999).  An individual?s social status within a group can also explain some variability in stress response and disease risk. In animals, Sapolsky (2005) argued that the impact of social stressors on individuals depends largely on the organization of the species; generally, the individuals that are most susceptible to disease are those who have the greatest exposure and physiological response to social stressors such as defeat (i.e., a ?loss?). In addition, the ability of individuals to cope with stressors - for instance, the capacity to which a low status individual can avoid a high status individual following conflict - can also influence the physiological response to stressors.  Generally, laboratory animal research has found both changes in immunity and disease risk in both high and low status individuals, depending largely on the stressor paradigm. In a stable group, caged mice with low status generally have lower innate immunity than high status individuals (Sa-Rocha et al., 2006), and a higher incidence of leukaemia following exposure to Moloney virus (Ebbesen et al., 1991). Similarly, rates of respiratory disease after exposure to a virus were higher in low status macaques, regardless of whether these animals were housed in a stable or unstable group (Cohen et al., 1997). Morbidity and mortality following Aujeszky disease virus were also higher for low status pigs housed in groups (Hessing et al., 1994).  It seems that in a stable group, individuals facing the highest number of losses, and thus deemed low status, show reductions in innate immunity and risk of infectious disease compared to those individuals that experience more wins. The number of losses an animal undergoes can also be experimentally controlled. It would follow that a situation where a high status individual begins to lose should also cause impairments in immunity. Indeed, paradigms that reverse status roles (i.e., force high status individuals to lose and low status individuals to win) have also found changes in immunity. High status mice ?made subordinate? by experimentally increasing the number of times they are defeated (i.e., ?defeat?induced subordination?) have greater HPA responses than control animals that experienced no change in status (Ely and Henry, 1978), and experience various reductions in immunity, including a shift toward a Th1 response (i.e., reduced splenocyte proliferation and IL-4 and IL-10 production, Bartolomucci et al., 2001). Moreover, Devoino (2003) discovered that mice ?made dominant? (i.e., increased number of times they defeated other mice) showed immunostumulation compared to mice without status changes. Thus far the discussion has bridged multiple disciplines to build a comprehensive framework for the relationship between social stressors and disease risk in humans and animals. The following section describes research to date which applies this concept to farm animals.  21    Figure 2.2. Individual differences in disease risk Potential differences between animals with proactive and reactive coping styles in their behaviour, neuroendocrine response to stressors and subsequent health risk. Adapted from Korte et al. (2005).  2.2.3. Farm animals Some research using farm animals has been mentioned in previous sections when the experiment was novel and relevant to developing the conceptual framework linking social stress and disease. We will now comment on the breadth and quality of research that has focused on farm animals. Table 2.2 shows 24 studies conducted since 1980 that have experimentally imposed social stress paradigms and have measured biological intermediaries of disease with or without the presence of clinical disease. These studies were identified using searches on Web of Science, IngentaConnect and Google Scholar, and search terms including the various farm animal species, ?social stress?, ?housing?, ?grouping?, ?remixing?, ?regrouping? or ?isolation? in combination with the terms ?immunity? or ?health?. This list is not meant to be exhaustive, but attempts to show the variety of experiments conducted to address the relationship between social stress and disease risk in farm animals.  A substantial majority (19 of 24) of these studies were conducted using pigs, and the most common social stressors were instability in group composition (i.e., regrouping) and social isolation. Our search only identified one study that has tested the effect of social stressors on the 22  health of dairy cows, two studies using chickens and one with veal calves. One study tested the effect of a single mixing event and overcrowding on the immune status of beef cattle; a second study using beef cattle (Fisher et al., 1997) was not included due to pseudoreplication.  A number of these studies reviewed here rely on short-term changes in immunity, such as lymphocyte proliferation and N:L ratio, without clear links to disease risk. These measurements must be viewed with caution, as some studies using acute stress paradigms (e.g., Hopster et al., 1998; Tuchscherer et al., 1998) may have recorded decreases in blood lymphocyte profiles resulting from normal movements from the portal circulation into the extremities (Dhabhar et al., 1995). These declines in immune components are typically transient and rapidly reversed when the stressor is removed, and are thus unlikely to be detrimental to the animal. More relevant are those studies that measured immune response to an endotoxin or vaccine challenge (Schrama et al., 1997; Hopster et al., 1998; Turner et al., 2000; de Groot et al., 2001; Bolhuis et al., 2003; Sutherland et al., 2007; Couret et al., 2009), as these provide a better reflection of host resistance. We also note that some of the studies rely on acute rather than chronic stressors such as one-time mixing (Moore et al., 1994; Tuchscherer et al., 1998; de Groot et al., 2001; Merlot et al., 2004) and brief periods of social isolation (Tuchscherer et al., 2009). Although these studies provide some insights on immune states, they may not adequately reflect the impact of management practices that often involve prolonged and multiple stressors. Intermittent and chronic models of social stress may better reflect farm animal management practices (Hopster et al., 1998; Turner et al., 2000; Ruis et al., 2001a; Ruis et al., 2001b; Sutherland et al., 2006; Couret et al., 2008; Fahey and Cheng, 2008).  The timing of measurements relative to application of the stressor may also influence the results. For instance, Van Reenen et al. (2000) attempted to measure the effect of social isolation on viral shedding of veal calves inoculated with bovine herpes virus 1 (BHV1).  Calves were isolated on the same day as inoculation; viral shedding and stress response was measured for 14 d thereafter. Although researchers found reductions in viral shedding of socially isolated calves in the first few days after inoculation, signs of a chronic stress response (i.e., glucocorticoid resistance) appeared much later. As veal calves are commonly socially isolated for months at a time, a more relevant approach would have been to isolate the calves days or weeks prior to inoculation, rather than the same day.  23  Table 2.2. The effect of social stressors on immunity in studies using farm animals  Species Social stressor Duration of stressor Immunity Individual variation Author Piglets Social isolation  2 h/d for 8 d  ? Lymphocytes? ? IL??? -- Kanitz et al. (2004)   2 h/d for 8 d  ? ??F?? -- Tuchscherer et al. (2004)   Single (4 h) ? Lymphocytes? Th1 shift? -- Tuchscherer et al. (2009)  Mixing at weaning Single No effect on WBC -- Merlot et al. (2004) Pigs Social isolation 3 wk ? N:L3 Coping style Ruis et al. (2001)  Social instability  2/wk for 4 wk No effect on WBC2 -- Couret et al. (2008)   Single ? Lymphocytes2 Th shift2 Social status de Groot et al. (2001)    Single No effect on N:L3 Social status Moore et al. (1994)   Single ? Lymphocytes?  Social status Morrow-Tesch et al. (1994)   Single ? Lymphocytes? ? Antibodies? Social status Tuchscherer et al. (1998)  Social defeat + isolation  Single + 4 h  ? N:L3 -- Ruis et al. (2001)  Crowding 6 wk ? Antibodies2 -- Turner et al. (2000)  Crowding + instability 14 d + single  ? NK cytotoxicity2 Social status Sutherland et al. (2006)  Low status (stable group)  ? lymphocyte proliferation2 Social status Hessing et al. (1994)    ? Antibodies2 ? Lymphocytes2 Social status Rudine et al. (2007)    ? NK cytotoxicity Social status Sutherland et al. (2007)  Reactive coping style (stable group)  ? Lymphocytes2  ? Antibodies2 Coping style Bolhuis et al. (2003)    ? Lymphocytes Coping style Hessing et al. (1995)    ? Antibodies2 Coping style Schrama et al. (1997) Chickens Social instability Repeated  ? N:L ? Antibodies2 -- Gross (1984)  Crowding 60 wk ? Lymphocytes Strain Fahey and Cheng (2008) Dairy cows Social isolation 55 h ? Lymphocytes2 Coping style Hopster et al. (1998) Beef calves Mixing + crowding Single + 88 d ? N:L3 -- Gupta et al. (2007) Veal calves Social isolation 14 d Delayed viral shedding2 -- van Reenen et al. (2000)  1Measured after samples were exposed to mitogen challenge 2Measured after animals were exposed vaccination or endotoxin challenge 3Neutrophil:Lymphocyte ratio showing lymphophilia and granulocytosis 24  Eight studies with pigs considered the effect of social status on immunity; three of these measured basal immunity in high and low status individuals housed in stable groups. Generally, low social status is related to poorer immunity; however, there were discrepancies in the response of individuals to social stressors. For instance, de Groot et al. (2001) found that regrouping caused a greater reduction in lymphocyte proliferation after vaccination in high status compared to low status individuals. In contrast, Sutherland et al. (2006) found that 14 d of overcrowding caused lower immunity (i.e., natural killer cell cytotoxicity and neutrophil phagocytosis) in low status rather than high status individuals. This discrepancy may be partially explained by the type of stressor: high status individuals may be more susceptible to regrouping than crowding, as there is a greater threat to status when animals are mixed with unfamiliar animals. In contrast, low status individuals may be at higher risk of experiencing defeat in crowded conditions. Classifying animals by personality or coping style may be more valuable than by social status alone, as the former is less sensitive to changes in situation. Five studies considered the effect of coping style on immunity. These studies found poorer lymphocyte and greater antibody responses of reactive individuals, consistent with a Th2 shift in the model described earlier in the review. No studies attempted to assess the influence of personality on disease risk in farm animals.  An interesting line of research in the farm animal literature has identified the effect of chronic early life social stress on long-term immunity and subsequent disease risk in swine (Kanitz et al., 2004; Tuchscherer et al., 2004; Tuchscherer et al., 2010). These studies used relevant models of social stressors (i.e., intermittent social isolation over a period of 8 d) and measured Th cell distributions as well as indicators of glucocorticoid resistance, both viewed as good indicators of a chronic stress response. These studies have contributed to both the farm animal literature, as well as a general understanding of potentially harmful effects of pre- and neonatal social stressors in humans and animals.   2.3. The effect of disease on social behaviour  Thus far our discussion has focused on social stress as a risk for disease, but changes in social behaviours can also be a consequence of disease. When animals become ill, the brain initiates an orchestrated set of behavioural changes meant to aid recovery from the infection. These ?sickness behaviours? have been well documented (Dantzer and Kelley, 2007). At the first signs of infection, communicators from the immune system (i.e., pro-inflammatory cytokines such as IL-1 and IL-6) signal the brain to increase body temperature (i.e., a febrile response) to create a 25  detrimental environment for the pathogen. Simultaneously, a coordinated set of behavioural changes facilitates a conservation of energy to fuel the febrile response and to mount an immune response. Following an infection, cytokines initiate a state of energy-conservation similar to a state of psychological depression in humans; infected animals appear fatigued and reduce general activity, including withdrawal from social behaviours. Declines in social exploration are often used as an indicator of sickness (Crestani et al., 1991; Kent et al., 1992; Bluth? et al., 1999; Arakawa et al., 2010). In a study using a common animal model of sickness behaviour, an adult rat was isolated in a home cage, an unfamiliar juvenile introduced for a short period, and the duration of social exploration (e.g., sniffing, grooming and chasing) was measured. When an immune response was stimulated (e.g., injecting IL-1), rats reduced social exploration in a dose-dependent manner (Crestani et al., 1991). Animals may specifically benefit from avoiding conspecifics when they are ill. Avoiding group members prevents ill animals from being introduced to secondary infections and simultaneously reduces exposure of pathogens to kin and other group mates (Loehle, 1995). Individuals respond to illness differently (Dantzer and Kelley, 2007). For example, when injected with LPS, high status rats decrease agonistic behaviours, whereas low status individuals maintain these behaviours (Cohn and de Sa-Rocha, 2006). There may be a higher cost to the low status animal in relinquishing agonistic behaviours, as these behaviours likely protect them from injury, whereas high status individuals can better afford to reduce these behaviours.  In humans and laboratory animals, sickness-driven social avoidance is well documented. There is less evidence of social avoidance in farm animals, perhaps because intensive housing systems generally allow little opportunity for ill individuals to isolate themselves or to avoid conspecifics. Gregory et al. (2009) monitored behavioural changes in hens housed in free-range versus caged systems. Following an LPS injection, free-range hens were observed to express a more comprehensive set of sickness behaviours than their caged counterparts, likely because the free-range environment allowed birds to express a greater range of behaviours.  Most research on sickness behaviour in farm animals has focused on measures of feeding behaviour, feed intake and activity rather than social behaviours (Waldron et al., 2006; Borderas et al., 2008), but two recent studies on dairy cattle have shown that cows that become ill after calving perform fewer social interactions (i.e., physically displacing one another from feeder space) up to 2 wk before showing signs of clinical and subclinical disease (Huzzey et al., 2007; Goldhawk et al., 2009). It is unclear whether cows that become ill after calving were already feeling ill weeks before and were, thus, less active, or were behaving this way because of other 26  factors that may have predisposed them to infection, such as low social status or coping style.  Additionally, lameness (i.e., hoof and leg injuries often caused by infection) can influence social behaviour of dairy cows; lame cows displayed fewer agonistic behaviours but were more often groomed than non-lame cows (Galindo and Broom, 2000). More work identifying sickness behaviours, particularly the relinquishment of social behaviours at earlier stages of infection, may aid in the identification of early signs of disease in farm animals. Environments that allow for greater expression of sickness behaviours may facilitate this disease detection, such as those that provide more space or access to resources that allow self-isolation.  2.4. Social factors and disease in farm animals: where to go from here? We have reviewed research using multiple species to describe current understanding of the relationship between the social environment and health. We suggest that this line of research is important in understanding, preventing and detecting disease risk in animal agriculture. Social stressors, including group instability, disruption, isolation, and crowding are common farm animal management practices. Reducing or eliminating practices that act as potent social stressors may be an important component of disease prevention, particularly when disease risk is highest (e.g., early life and parturition). Moreover, including social ?sickness behaviours? as early indicators of disease in farm animals may allow for earlier treatment. Although farm animal research in these areas is developing, we believe that there are methodological and conceptual tools that can improve this research in the future.  Drawing on the human and laboratory animal literature, we suggest four directions for further farm animal research: 1) adopt the ?disease-centred? approach proposed by (Miller et al., 2009) using relevant social stressors, 2) develop tools to better categorize individual animals based on personality and coping style rather than relying on social status alone, 3) identify long-term effects of early life social stressors on disease risk later in life and, 4) determine if changes in social behaviour can be used as early indicators of infection. A major shortcoming of the current farm animal literature is the reliance on some immune components (i.e., lymphocyte proliferation) as indicators of health, without clear links to disease. Collaboration between veterinary epidemiology and the basic animal sciences (Duffield et al., 2009) using the ?disease-centred? approach may help avoid this problem. For instance, epidemiologists have already established some management-related determinants of common diseases in dairy cows during the calving period (Svensson et al., 2006; Hultgren and Svensson, 27  2009). Further experimental research can assess the effects of these practices on established biological intermediaries of these common diseases. The dairy cow literature has many biological intermediaries of disease to draw from (e.g., a decline in neutrophil cytotoxicity before calving predicts retained placenta and metritis, Kimura et al., 2002; Hammon et al., 2006). The biological chain connecting these intermediaries to relevant social stressors and stress reactivity (e.g., the impact of regrouping before calving on neutrophil cytotoxicity) now needs to be examined.  Using relevant social stressors is an important component of this ?disease-centred? approach. We encourage research assessing the impact of regrouping on the immunity and disease risk of beef calves and finishing pigs, as there is an epidemiological linkage between mixing and morbidity (Ribble et al., 1995; Lo Fo Wong et al., 2004; Sanderson et al., 2008). We also encourage epidemiological research to draw from the laboratory animal research and include social stressors as farm-level effects on morbidity and mortality, such as overcrowding, social isolation and regrouping. However, we do caution against relying on models of social stress established using laboratory animals of social stressors, as there are likely many species differences in stress responses (Bartolomucci, 2007). Moreover, farm settings provide dynamic stressors (e.g., pigs are socially isolated for months during gestation and until weaning, then moved into a group of unfamiliar animals), thus, future work should use chronic, intermittent and/or a combination of stressors reflective of these farm practices.  A second shortcoming of the farm animal literature is an inconsistent categorization of animals based on individual variation in response to social stressors and susceptibility to illness. Identifying at-risk animals within a group will aid in the development of better disease prediction models for farm animals (Ingvartsen et al., 2003). Some studies have categorized animals by social status, but it is clear that an individual?s status can be highly influenced by its social environment. Identifying status does give some insight into the effects of farm animal management practices on group health risks. For instance, regrouping and overcrowding may influence the number of agonistic behaviours and defeats experienced by all members of the group. Identifying the influence of these practices on agonistic behaviours and status-reversals may provide a better understanding of how these might influence morbidity of the group.  Classifying animals by coping style and personality can provide an additional layer to our understanding of the at-risk individual, as these definitions are also linked to disease but are more persistent over situations and management practices. Hessing et al. (1994) and Van Reenen (2005) provide good methodology for defining coping styles in pigs and dairy calves, respectively. Identifying the personality traits of individual animals (e.g., sociability) may also be promising, as 28  tools are being developed to evaluate these differences on-farm (Gibbons et al., 2010). Moreover, pigs, beef cattle and chickens are often housed as mixed-sex groups, but dairy cattle provide a unique all-female model where some of these concepts may not directly apply but more work is encouraged. Thirdly, the effect of early life stressors on long-term changes in immunity and disease risk is an intriguing and highly relevant area of research for farm animals. Evidence for the harmful effects of early social isolation has begun to emerge in the swine literature, where multiple studies have assessed both prenatal and neonatal responses to social stress (Kanitz et al., 2004; Tuchscherer et al., 2004; Couret et al., 2009). We also encourage research in this area on dairy calves, as calves are typically separated from their dam immediately and socially isolated until weaning (e.g., 6 to 8 wk). Housing calves individually has some health benefits compared to large groups (Svensson et al., 2006; Gulliksen et al., 2009b), however, this practice can also impair social learning and weight gain (de Paula Vieira et al., 2010; Duve and Jensen, 2012), and can have long-term effects on HPA activity (Creel and Albright, 1988). Future research is needed to determine the potential immunological detriment of individual housing compared to small group housing for young dairy calves.  Lastly, work with humans, laboratory animals, and to some degree farm animals, has provided evidence that social behaviours are reduced as part of a general response to infection. Research identifying early changes in social behaviour during the course of infection could be integrated into herd health programs for disease detection. As the expression of these behaviours is dictated by the animal?s environment, we also encourage future research to test the effect of practical changes to housing (i.e., increasing space in the pen or cage-free systems) that allow for behaviours such as self-isolation.   2.5. Conclusions Disease remains a major challenge in animal agriculture. Epidemiologists have identified linkages between farm animal grouping practices and morbidity. To understand the role of the social environment on disease risk, researchers have begun testing the effects of management practices on biological intermediaries of disease. Aspects of the social environment can cause a stress response, and this response varies among animals with varying individual characteristics, such as social status, personality and coping style. Individual responses to social stressors include changes in the immune system, such as lowered host resistance, inflammation and changes in Th cell distributions. Depending on the duration and period in life when these stressors occur, these 29  biological intermediaries can increase the risk of clinical disease. Once animals become ill, they reduce socializing to aid in their recovery and avoid spreading the infection to kin.   Growing farm sizes and other management practices provide challenges for identifying ill animals using sickness behaviours such as self-isolation. A greater understanding of the relationship between social factors and disease in farm animals, utilizing established methodologies from the human and laboratory animal literature, will add an important dimension to both disease detection and prevention. Some of the following research chapters will describe experiments using transition dairy cows that address the gaps listed above. For instance, the second part of Chapter 5 determined the effect of infectious disease on social behaviour, or avoidance of conspecifics, in dairy cows after calving. In Chapter 7, one measurement of social behaviour during transition is used as a predictor of lameness after calving. The remaining chapters will address gaps outlined in Chapter 1, beginning with the influence of management practices during the parturient phase on cow behaviour and labour progress.  30  3. The Effect of Management during Parturition on Behaviour and Labour 2  3.1. Introduction One of the main gaps in the literature identified in Chapter 1 was the lack of information about the effect of management practices during the parturient phase on cow behaviour and labour progress. This chapter explores the effect of moving cows from a group to an individual pen during different stages of labour. Moving cows to individual pens just before calving is still a common practice on many farms, yet there is a lack of data available in the literature to determine when cows should be moved to this pen relative to calving. Often this decision is based on calving signs; cows are moved when calving is imminent. A combination of physical and behavioural cues before and during labour provides a dairy producer some indication of when calving is imminent. These signs occur just before and during the first two stages of labour. Labour is traditionally described using three stages (described in detail in Noakes et al., 2001; Jackson, 2004; and reviewed by Mainau and Manteca, 2011), although there is no clear stop and start to these stages as they progress gradually. Thus, it is often difficult to determine which stage of labour an animal is in. There is also a high variation in the timing of labour signs; for instance, Jackson (2004) suggests the first stage can last from 4 to 24 h. Briefly, we will describe the stages of labour using signs that are relevant to this study. A first indicator of impending labour is the relaxation of pelvic ligaments, as this reflects the opening of the cervix (Noakes et al., 2001). An engorged udder may also be present at this time, which will gradually enlarge as labour and calving approaches. The start of the first stage occurs when the cow?s cervix begins to dilate and the calf begins moving into position for delivery. Verification of this first stage can only be achieved by palpating the cervix; however, the cow also begins to show changes in behaviour during this stage. The cow has myometrial (uterine) contractions and becomes restless (often characterized by an increase in position changes, Huzzey et al., 2005; Miedema et al., 2011; Jensen, 2012), she also begins to pay attention to her stomach (Jensen, 2012), and increasingly raises her tail while standing (Miedema et al., 2011). The onset of rhythmic abdominal contractions and the release of the amniotic sac or ?water-bag? are two                                                  2 A version of Chapter 3 has been published: Proudfoot, K. L., M. B. Jensen, P. M. H. Heegaard, and M. A. G. von Keyserlingk. 2013. Effect of moving dairy cows at different stages of labour on behaviour during parturition. J. Dairy Sci. 96: 1638-1646. 31  prominent landmarks of the beginning of the second stage of labour (Noakes et al., 2001). During this stage, the cow is often recumbent, as the forces of the uterine and abdominal contractions help to expel the calf (Schuenemann et al., 2011). During the third and final stage of labour, the placenta is released and uterine involution occurs. These physical and behavioural cues before labour provide a dairy producer some indication that parturition is approaching, but it remains unclear when cows should be moved to a maternity pen relative to these signs. Moving cows to a new pen too close to calving may increase the number of cows calving in unwanted areas (e.g., the free stall in some cases) or may disturb normal behaviour and interrupt the progression of labour. The effect of changing pens too close to parturition has not been tested in cows, but Pedersen and Jensen (2008) found that primiparous sows moved to farrowing crates on the day of farrowing experienced longer labour and more stillbirths compared to those moved 20 d earlier. Moving cows to an isolated, individual pen too early before calving may also be detrimental to the cow. Previous work indicates that multiparous cows subjected to even short-term (15 min) complete social isolation increased vocalizations, heart rate and hypothalamic-pituitary-adrenocortical axis activity (Rushen et al., 1999). To our knowledge no research to date has attempted to determine any changes in behaviour or disruptions in the progression of labour in management practices where cows are routinely moved to a maternity pen during different stages of labour. Thus, the objective of this study was to assess the effect of moving cows at different stages of labour on behaviour and the progression of labour.  3.2. Materials and methods 3.2.1. Animals, housing and diet This study was conducted at the Aarhus University?s Research facilities in Foulum, Denmark between September 2011 and February 2012. Cows were cared for according to a protocol approved by the Danish Animal Experiments Inspectorate, Ministry of Justice, Copenhagen, Denmark and the University of British Columbia?s Animal Care Committee.   A total of 79 multiparous Danish Holstein dairy cows were used for this study. Before calving, cows were grouped into 1 of 6 blocks of 14.7 ? 1.5 (mean ? STD; min = 12, max = 16) cows based on expected calving date. Cows within a block were moved into 1 of 2 group pens (each 9 m ? 15 m) approximately 2 wk before the first cow was due to calve. Group pens had deep straw bedding and 12 individual feeding bins (each 0.75 m wide). Adjacent to each group pen were 10 individual maternity pens (each 3.0 ? 4.5 m) also with deep-bedded straw. New 32  bedding was added daily to the group and maternity pens. Bedding was completely changed in the group pen between blocks, and in the maternity pens after each cow was moved to the milking herd. Composition of the group pen was dynamic as cows left the pen to calve, but there was at least one non-experimental animal (pregnant cows or heifers) in the group pen when the last experimental cow of the block calved.  As calving approached, cows were moved into an individual maternity pen. The timing of this movement relative to calving was dependent on treatment (described below). In half of the maternity pens, the sides were made from vertical tubular metal bars and a 3.0 m wide side separated cows from the group pen and a 4.5 m wide side separated cows from neighboring individual pens. In the other maternity pens, a plywood barrier was built around the pen so that cows could only see into the group pen through a 1.5 m wide window. The pen dimensions were the same in all individual pens. This pen treatment was used to test a different hypothesis (Chapter 5), however, to be sure that this pen treatment had no effect on standing behaviour and length of stage II labour we included pen type as an categorical variable in our first model. We found no effect of pen type on standing time or standing bouts in the 1 h before calving (P = 0.48 and P = 0.70, respectively), length of abdominal contractions during this same hour (P = 0.76) or length of stage II labour (P = 0.25). There were also no interactions between pen type and movement category (see below for a description of these categories) for any variable. Thus, the pen treatment was not included in the final analysis.  In both the group and maternity pens, cows were fed a pre-calving TMR ad libitum containing a dietary forage to concentrate ratio of 79:21 on a DM basis. Diets were balanced for dietary cation-anion difference, however this was not subjected to verification through nutritional analysis. After calving, cows were fed a post-calving TMR ad libitum with a 60:40 forage to concentrate ratio on a DM basis. Feed was allocated twice daily at 1000 and 1700 and fresh straw was provided at 1000. After calving, cows were milked twice daily at 0600 and 1800 using a manual milking machine. Water was available ad libitum in the group and maternity pens via water bowls. 3.2.2. Experimental design and inclusion criteria Before entering the group pen cows within block were assigned to 1 of 2 treatments. Within parity (first, second or later) cows were randomly assigned to treatment: 1) moved to a maternity pen 3 d before expected calving date, or 2) moved to a maternity pen when signs of calving were present. Before the experiment, training on the signs of calving (listed below) was provided to all farm staff. For a week before the first cow of the experiment was due, the farm 33  staff, farm manager and experimenters monitored cows close to calving together, and discussed each sign of calving to reach a consensus about each definition (Table 3.1). Cows assigned to treatment 2 were monitored for signs of calving 7 times daily (0500, 0800, 1000, 1300, 1500, 1900 and 2200). During each check, farm staff first visually inspected cows from a distance for signs of abdominal contractions, raised tail or suddenly enlarged, tense udder. Farm staff then entered the pen and physically inspected (i.e., palpation of the udder and pelvic ligaments) each cow while she was in a standing position. Farm staff recorded any signs they saw or felt on the cow, and looked at the previous records during each check. Farm staff moved cows to a maternity pen if one or more of the signs on Table 3.1 were noted. Calving signs and time of moving were recorded for each cow. After calving, cows were kept in the maternity pen with their calf for 3 d before being moved to the milking herd.  Cows were not included in the study if they calved in the group pen (n = 10), were in the individual pen for more than 5 d (n = 10), had a difficult calving or twins (n = 7) or were subjected to disturbing management practices (e.g. delivery of fresh bedding) by the farm staff during stage II labour (n = 3).  The remaining 49 cows were then categorized by the stage of labour they were in when they were moved from the group pen to the maternity pen. To do this, we used 1) the time relative to calving when the cow was moved (collected retrospectively from video), and 2) the sign(s) used by the farm staff to move each cow (Table 3.2). Although most cows showed more than one sign of calving when they were moved (e.g., visible sac and raised tail), final category assignment was determined using the latest sign.  For example, in the case where a cow was identified as having both a raised tail and a visible sac, the presence of the sac normally occurs after the raised tail and thus the cow was placed in the Stage II category. The order of signs listed in Table 3.2 reflects our expectation as to the order of the signs relative to calving. Sixteen cows were moved before any signs of labour, or showed signs of approaching labour (relaxed pelvic ligaments) and were moved > 24 h before calving (?Before Labour?). Seven cows were moved when signs of the second stage of labour were obvious (amniotic sac or calf legs were visible outside the vulva; ?Stage II?). The remaining 28 cows were moved just before or during the first stage of labour. We separated these cows into two categories: 1) those moved between 2 and 24 h before labour with a raised tail, relaxed pelvic ligaments or suddenly enlarged, tense udder (?Early Stage I?; n = 17) and 2) those moved between 1 and about 4 h before calving with viscous, bloody mucous on the outside of the vulva or the first signs of abdominal contractions (?Late Stage I?; n = 9). The first category we call ?Early Stage I?, although relaxed 34  pelvic ligaments and suddenly enlarged udders may occur before labour, or during the first and second stage of labour (Noakes et al., 2001). We are confident that cows in this category are not in the second stage of labour, as this stage is recognizable by landmarks such as abdominal contractions and the amniotic sac becoming visible. However, we could not be certain that these cows were not transitioning into the first stage of labour, or were not yet at this stage, as the start of stage I is detectable only by palpation of the cervix. We call the second group ?Late Stage I?; although these cows are likely transitioning between the first and second stage of labour, as abdominal contractions are considered a first sign of the second stage, and viscous bloody mucous is pushed through the vulva as the calf begins to move through the cervix.  All cows moved during Stage II (n = 7) calved within 1 h of being moved, which prevented any meaningful behavioural data from being collected from the maternity pens; thus, these cows will not be included in the analysis.    Table 3.1. Definitions of calving signs Definition of the signs of calving used to move cows from a group to an individual maternity pen  Sign of Calving Definition Tense and enlarged udder The udder is enlarged considerably relative to the previous inspection and is tense. The cow may pay attention to the udder by turning the head toward it, or by licking it Relaxed pelvic ligaments The pelvic ligament near the tail head appears sunken and feel soft and flexible when palpated Raised tail The tail is raised for longer than 1 minute without any urination or defecation Viscous mucous with blood Viscous mucous mixed with blood appearing outside the vulva Abdominal contractions Standing, lying on side, or partially lying on side and the abdominal muscles contract and release in a rhythmic motion Visible amniotic sac The unbroken amniotic sac appears outside the vulva  Visible calf legs The calf?s feet and possibly part of the legs, appears outside the vulva    35  Table 3.2. Signs of calving used to create movement categories The signs of calving used to move cows from a group pen to an individual maternity pen for each movement category, and a description of when cows calved relative to these signs                 3.2.3. Behaviour and body weight measurements A digital video camera was fixed above the group pen (MONACOR, TVCCD-460 fitted with a wide-angle lens) and each of the individual maternity pens (MONACOR, TVCCD-140IR). Video was used to record the detailed behaviour of each cow beginning 4 h before calving by one trained observer. Table 3.3 shows the ethogram of behaviours and postures collected for each cow. The length of the second stage of labour was estimated by subtracting the time that the first abdominal contractions were recorded from the time of calf delivery. Abdominal contractions are one of the first signs of the second stage of labour (Noakes et al., 2001), and are easy to detect with video compared to other signs, such as appearance of the amniotic sac (e.g., if the posterior of the cow was facing away from the camera when the sac became visible outside the vulva). In the case of 2 cows classified as Late Stage I, the onset of abdominal contractions was observed when they were in the group pen; this record was used as the start time for abdominal contractions.     Time spent in maternity pen (h) Sign of calving for each category N Median Mean Min Max Before Labour       No signs 16 68.2 74.9 37.2 117.9 Early Stage I       Tense and enlarged udder 4 10.5 10.0 3.6 15.2 Relaxed pelvic ligaments 8 10.8 12.4 4.4 20.5 Raised tail 5 12.5 10.5 2.7 19.3 Late Stage I       Viscous mucus with blood 7 2.8 2.6 1.3 4.1 Abdominal contractions 2 1.7 1.7 1.4 2.1 Stage II       Visible amniotic sac 4 0.6 0.5 0.1 0.8 Visible calf legs 3 0.3 0.4 0.3 0.6 36  To collect meaningful behavioural data, we only included cows that were in the maternity pen for more than 1 h before calving. Lying time, number of times the cow changed from standing to lying (i.e., lying bouts), lying bout duration (lying time divided by the number of lying bouts) and the length of abdominal contractions were summarized for the 1 h before calving. During the trial we noticed that other cows in the group pen were approaching the cows moved during Late Stage I, even after they were moved into the individual pen. Thus, we also monitored the behaviour of cows in the group pen during the 1 h before calving. Newborn calves were removed from the pen and weighed using a scale (Danvaegt, model 4301) within 12 h after birth.  Table 3.3. Ethogram of behaviours occurring before calving Ethogram of cow postures, behaviours and calving events recorded in the individual maternity pens 4 h before calving  Variable Level Definition Standing Posture Body is supported by four legs, standing or walking Lying Posture Lying on sternum or side, head may be rested or raised Abdominal contractions Behaviour Standing, lying on side, or partially lying on side and the abdominal muscles contract and release in a rhythmic motion Attention from other cows Group behaviour A cow in the group pen puts any part of her head over the bars of the individual maternity pen where the focal cow is housed for at least 1 minute Start of abdominal contractions Calving event First time cow is lying on side, or partially on side and the abdominal muscles contract and release in a rhythmic motion Calf delivery Calving event The calf's hips are fully expelled from the cow  3.2.4. Blood samples and haptoglobin measurement A blood sample was taken from the jugular vein (vena jugularis) at 0800 from a sub-sample of 23 cows 3 to 27 h after they had calved; although this seems like a large variation in sampling time, there is evidence of a protracted response of bovine haptoglobin (e.g., lasting from 24 to 96 h after LPS injection, Jacobsen et al., 2004). Samples were analyzed for content of haptoglobin, an indicator of systemic inflammation typically present during labour, and highest in 37  cows with uterine disease (Huzzey et al., 2009). Haptoglobin was measured using a sandwich ELISA originally developed by Godson et al. (1996) as described by Heegaard et al. (2000) and having a detection limit of 7.8 microg/ml. Briefly, all samples were run in three dilutions together with a dilution row of calibrated bovine serum as standard, calculating the haptoglobin concentrations in the samples by interpolation using a fitted standard curve (Godson et al., 1996). As our categorization of cows into movement categories was done retrospectively, we were not able to balance these samples for categories (Late Stage I: n = 4; Early Stage I: n = 11; Before Labour: n = 8).  3.2.4. Statistical analysis All statistical analyses were performed with SAS software (version 9.2; SAS Inst. Inc., Cary, NC) using the cow as the experimental unit. Data were scanned for normality, and residual plots were created to determine outliers using the UNIVARIATE and GLM procedures. One extreme outlier was found in the haptoglobin data, and was removed. An ANOVA (PROC GLM) was used to determine the effect of moving cows at different stages of labour on the length of the second stage of labour, the duration of abdominal contractions, lying time, lying bouts and lying bout duration recorded in the 1 h before calving. The initial model included block, but this had no effect on any variable and was removed. The final model included parity and calf weight as covariates, and movement category (Late Stage I, Early Stage I or Before Labour) as a main effect. The PDIFF statement was used to compare the LS means of each movement category.  We had insufficient power to test the effect of movement category on haptoglobin, we instead tested an alternative hypothesis that prolonged labour might result in higher systemic inflammation. We tested the relationship between length of stage II labour and haptoglobin using a model that included hour of sampling relative to calving and length of stage II labour. To determine the effect of movement category on attention from group-mates, data were transformed into binomial data (0 = no attention, 1 = at least one cow had her head over the pen for at least 1 minute). A logistic model (PROC LOGISTIC) was used with category as the only explanatory variable; Late Stage I was compared to the 2 other categories using a contrast statement.   3.3. Results  3.3.1. Length of stage II labour and abdominal contractions  Figure 3.1 shows the effect of movement category on the length of stage II labour, and the length of abdominal contractions in the 1 h before calving. After controlling for calf weight and 38  parity, cows that were moved to an individual pen during Late Stage I had the longest stage II labour (P < 0.001), but did not differ in the duration of abdominal contractions in the hour before calving (P = 0.52).     Figure 3.1. Labour length and abdominal contractions  LS means and pooled SE of length of stage II labour and duration of abdominal contractions in the 1 h before calving for cows moved from a group to an individual pen Before Labour (n = 16), during Early Stage I labour (n = 17), or during Late Stage I labour (n = 9). Columns with different letters are significantly different at P < 0.05.    39  3.3.2. Behaviour in the hour before calving Figure 3.2 shows the effect of movement category on lying time, the number of lying bouts and the length of lying bout duration occurring in the 1 h before calving. Cows that were moved to an individual pen during Late Stage I spent the least amount of time lying (P = 0.002) but did not differ in the number of lying bouts (P = 0.20). There was a tendency for shorter lying bout in cows moved during Late Stage I (P = 0.09). There was a tendency for cows that were moved during Late Stage I to have a higher likelihood of receiving attention from group cows compared to the other categories (P = 0.06). Six of 9 cows in Late Stage I, 5 of 17 cows in Early Stage I and 4 of 16 cows moved Before Labour had attention from cows in the group pen.  3.3.3. Haptoglobin After controlling for time of sampling, haptoglobin moderately increased as length of the second stage of labour increased (R2 = 0.42; P = 0.01).    40     Figure 3.2. Lying behaviour in the hour before calving LS means and SE of time spent lying, number of lying bouts and a lying bout duration during the 1 h before calving for cows moved from a group to an individual maternity pen Before Labour (n = 16), during Early Stage I labour (n = 17), or during Late Stage I labour (n = 9). Columns with different letters are significantly different at P < 0.05.    41  3.4. Discussion The aims of this study were to determine if moving cows from a group pen to an individual maternity pen at different stages of labour affected behaviour or the progression of labour. Cows moved during the late part of stage I labour (or perhaps transitioning from the first to the second stage) showing viscous, bloody mucous outside the vulva or abdominal contractions had the longest second stage of labour, spent the least amount of time lying and received slightly more attention from other cows during labour.  Most notable was that the total duration of the second stage of labour was about 30 min longer for cows moved during the late part of the first stage of labour compared to those moved during the earlier part of the first stage and those moved before labour (91 vs. 58 and 60 min, respectively). Two recent studies reported estimates of normal length of the second stage of labour in cows with unassisted calvings. Barrier et al. (2012) estimated the median time from signs of calf feet to calving to be 54.7 min, and Schuenemann et al. (2011) estimated the mean time from signs of amniotic sac to calving to be 45.1 min. The time from first abdominal contractions to calving recorded in our study for cows moved before labour and during Late Stage I labour is slightly longer than these estimates, presumably because there is some time between the first abdominal contractions and the appearance of the amniotic sac or calf feet. It is, however, somewhat concerning that the cows in our experiment that were moved during Late Stage I labour had stage II labour that was much longer than unassisted calvings in both of these studies.  Due to low sample size, we were unable to detect any potential effect of prolonged stage II labour on the health of the dam or the calf. In a study using a much larger sample size, Gundelach et al. (2009) found that cows with stage II labour > 120 min had a higher risk of stillbirth compared to those with shorter stage II labour. We did not record any cases of stillbirths in this study, likely because with our small sample size we did not encounter any cows with stage II labour as long as those reported in Gundelach et al. (2009). Prolonged stage II labour is also associated with dystocia (Schuenemann et al., 2011; Barrier et al., 2012).  The reasons for prolonged stage II labour of cows moved late remains unclear. At this stage the calf is in its birthing position and is beginning to move into the birth canal (Noakes et al., 2001). Although no research has been done using cattle, there is some evidence that disturbances that cause a stress response during labour can prolong labour in humans (reviewed by Johnson et al., 2003), and impair uterine motility in sheep and rabbits (Bontekoe et al., 1977). Perhaps moving cows into a novel environment causes an acute stress response that disrupts normal hormonal changes during labour; more research is encouraged in this area. Moving a cow 42  into a novel environment during this sensitive period may also cause deviations from the normal calving behaviour. Indeed, normal behaviour did seem to be disturbed in cows moved during Late Stage I labour. These cows spent half as much time lying in the hour before calving compared to both of the other categories. Previous work has reported that when cows calved without assistance they became recumbent at the start of abdominal contractions and remained in this position until birth (Schuenemann et al., 2011). The low lying time recorded in the cows moved during Late Stage I labour is, thus, likely a deviation from normal behaviour. The lower lying time of cows moved during Late Stage I may be related to the cow responding to the novel and slightly more restrictive environment of the maternity pen. Pre-parturient sows housed in restrictive farrowing crates that do not allow for the expression of normal pre-parturient behaviours will increase the number of postural changes, and reduce standing time in the hours before farrowing (Jarvis et al., 2001). In our case, cows moved to a maternity pen just before calving (Late Stage I) did not differ in postural changes in the hour before calving compared to the other categories, although the power may have been too low to detect differences because of our low sample size. Instead, they had slightly shorter lying bouts and were more reluctant to become recumbent, perhaps spending more time exploring their new pen to ensure that it is a safe place to give birth.  The tendency for cows moved during Late Stage I to draw more attention from cows within the group is novel. These cows were moved within 4 h of calving, so perhaps the changes in their behaviour or the odor of amniotic fluid while they were in the group pen may be been appealing for other cows in the pen. Both pre-parturient cows and ewes are attracted to newborns (Edwards, 1983), as well as to amniotic fluid alone before parturition in cows and during parturition in ewes (Arnould et al., 1991; Pinheiro Machado et al., 1997). Given the small number of cows used in this study we strongly encourage future work in this area. Although we did not have the power to test the effect of movement category on early post-calving haptoglobin, we do provide some evidence that cows with longer stage II labour have higher inflammation after calving. Haptoglobin, a marker of systemic inflammation, is normally elevated just after calving and declines within the first 6 to 9 d of calving in healthy cows (Huzzey et al., 2009). Despite the variation in the timing of the blood samples (blood samples were taken between 3 and 27 h after calving), we found that cows with longer second stage of labour had higher levels of serum haptoglobin post-calving. High haptoglobin during this period may be a sign of stress or tissue damage during calving, and may also predict later uterine disease (Huzzey 43  et al., 2009). Future research using larger sample sizes and more precisely timed blood samples are encouraged to determine the effect of prolonged stage II labour on uterine health outcomes.  Only one other study has examined the effect of moving cows from a group to a maternity pen on health outcomes (stillbirths, Carrier et al., 2006). Similar to our findings, these authors found a detrimental effect of moving cows during the late part of the first stage of labour (viscous mucous with or without blood), although these cows were compared to those moved during the second stage, which we were unable to measure. Specifically, they found a 2.5-fold increase in stillbirths in cows moved during the late part of the first stage of labour. Based on their findings, Carrier et al. (2006) recommended moving cows with signs of the second stage of labour. However, moving cows during the second stage of labour, or ?just in time?, presents its own challenges.  Waiting to move cows at this late stage will certainly increase the number of cows that calve in the group pen (or non-maternity pens). In our case, 10 of 79 cows calved in the group pen, despite farm staff checking cows 7 times per day for early signs of calving. Little research has been done on the effect of calving cows in groups. In one study, Edwards (1983) discovered that cows housed in groups spent less time licking their calves compared to cows housed individually. This reduction in licking time was likely due to other cows in the group licking the newborn calves (Edwards, 1983). There is also some evidence that cows prefer seclusion to calve; cows housed in semi-natural environments distance themselves from herd mates in the few hours before calving (Lidfors et al., 1994). In the case where producers elect to move cows to an individual maternity pen just before calving, there may be an optimal time for this movement to take place. Although our study was unable to determine when the best time to move is, it is clear that there seems to be a sensitive period near the end of the first stage of labour, or the transition between stage I and II where moving a cow would disrupt normal behaviour and the progress of labour. We should also note that we found no differences between cows moved before signs of labour (up to 5 d before calving) and those moved during early signs of labour. We highly encourage more work in this area utilizing larger sample sizes.   3.5. Conclusions When cows were moved from a group to an individual maternity pen during the late part of stage I labour (i.e., signs of contractions or viscous, bloody mucus), their behaviour was disturbed and the length of the second stage of labour prolonged. These cows spent less time lying in the hour before calving and tended to get more attention from group mates.  44  4. Preference for a Secluded Calving Site3  4.1. Introduction The study described in Chapter 3 determined the effect of a common management practice on behaviour and labour. In this chapter, we take a step back and ask the cow to help us determine what types of management practices and housing features she prefers during parturition. To do this, we first determined the options that might be suitable for cows nearing partition drawing from evidence in wild ungulates as well as dairy and beef cows housed on pasture.  When cover is available and high quality forage is nearby, wild ungulates such as elk and bison will seek shelter to calve (Lott and Galland, 1985; Barbknecht et al., 2011). When dairy cows are kept in semi-natural environments, some will hide to calve, but much like their wild counterparts this behaviour only occurs when suitable conditions are present (e.g., tall grass or tree cover with appropriate grazing sites are nearby, Lidfors et al., 1994). Dairy cattle housed indoors are provided with protection from predation and typically provided energy-dense and easily accessible feed ad libitum. Management practices around parturition differ among farms, but cows are generally housed in either a group pen or moved to an individual maternity pen within hours or days before calving. Calving pens are often located in high traffic areas to allow frequent observations by farm staff to better detect calving difficulties.  It has been suggested that a better understanding of the maternal behaviour of parturient cows may help producers improve the care and management of these animals (von Keyserlingk and Weary, 2007). Yet, no research to date has assessed if, and under what conditions, indoor-housed dairy cows will hide during calving when cover is made available. The objective of this study was to determine if indoor-housed dairy cows seek shelter to calve if given the opportunity, and if this behaviour is influenced by the time of day or the presence of another cow in the pen.  4.2. Materials and methods 4.2.1. Animals and housing  Eighty-eight Holstein dairy cows (62 multiparous and 26 primiparous) were used in the experiment over a period of 6 months (July to December, 2011). The study was conducted at the University of British Columbia?s Dairy Education and Research Centre in Agassiz, British                                                  3 A version of Chapter 4 has been submitted for publication: Proudfoot, K. L., D. M. Weary, and M. A. G. von Keyserlingk. Submitted. Seeking privacy: maternal isolation behaviour of Holstein dairy cows. J. Anim. Sci. 45  Columbia, Canada. Animals were cared for according to the guidelines provided by the Canadian Council on Animal Care (2009). Cows were housed in one pre-calving pen 21 d before their expected calving date. This pen had 12 lying stalls and 12 headlocks; lying stalls contained a mattress with a layer of sand. Stocking density was maintained at 12 cows in this pen (10 m2 per cow), but was dynamic as cows came and left the pen depending on their expected calving date.    Cows were paired based on calving date, and the pair was moved into 1 of 4 maternity pens 6.6 ? 3.0 d before calving, and remained in this pen until they calved. This pen was the same size as the pre-partum pen, but contained either one cow (120 m2 of space per cow) or two cows (60 m2 of space per cow).  Each maternity pen contained 2 bedded sawdust areas (2.4 m ? 7.3 m). Manure and urine were raked from each area 4 times per day (0800, 1200, 1600 and 2000). Sawdust was changed or added daily or when necessary to ensure that each lying space was clean and dry. After a cow calved on an area, the bedding of that area was removed and fresh bedding was provided. The dam and calf were removed from the pen immediately after calving. After the first cow of the pair calved, the second cow remained alone in the pen until she gave birth. 4.2.2. Preference test and inclusion criteria Each maternity pen contained one ?sheltered? area with a 2.4 m tall plywood barrier around 85% of the width of the area (excluding the ceiling) except for a 2.4 m wide opening for cows to freely enter or exit, and one ?open? area with no barrier (Figure 4.1). Cows were introduced to the maternity pen at least 24 h before calving to gain exposure to both areas before calving. To eliminate the effect of any side bias, the shelter was built on alternating sides of the four pens.  If a cow calved within 24 h of entering the pen it was assumed that she was not familiar with the new pen and was not included in the analysis (n = 4). Cows were also excluded from the study if they had a very difficult, assisted calving (n = 9), if they calved in the alley (n = 2), or if they had hypocalcaemia (a disease that restricts mobility) before calving (n = 1). The final analysis included 72 cows (50 multiparous and 22 primiparous; single-housed: n = 34; pair-housed: n = 38).     46   Figure 4.1. Design of maternity pens Design of maternity pens built with a shelter and an open area for calving. Both areas were fitted with a mattress and a layer of sawdust bedding. The shelter was built using 1.2 ? 2.4 m plywood slats. Four maternity pens were used; in 2 pens the sides of the sheltered and open areas were reversed.  4.2.3. Behavioural measurements  Four digital video cameras (Panasonic WV-CW504SP) were mounted above each maternity pen and continuously recorded video using a GeoVision 1480 digital recorder (USA Vision Systems, CA, USA); each camera had a view of one area. Red lighting was used to facilitate night viewing.   Video was used to determine the time and location of calf delivery (i.e., the calf?s hips were fully expelled from the dam). The video was monitored for 12 h period before calving, and the same 12 h period on the day before calving (i.e., -36 to -24 h before calving) using 5-min scan sampling. At each scan we measured the location of the focal cow (open or sheltered area; lying or standing with at least the 2 front hooves in the area) and whether or not she was greater than 1 cow-length from their partner if she was housed in a pair. These data were used determine when cows began to selectively use the shelter and separate from their partners before calving; video data were collected for a sub-sample (n = 11) of single-housed cows that calved in the shelter, and a sub-sample (n = 20) of cows that were housed with a partner. These sub-samples were used because digital video data were lost for the remaining animals in these categories. 4.2.4. Statistical analysis 47   All statistical analyses were performed with SAS software (version 9.2; SAS Inst. Inc., Cary, NC) using the cow as the experimental unit. In a preliminary analysis, parity (primiparous or multiparous) was included in all models to test for any differences between these groups. No differences were detected, so the following tests include both parities together. Two-tailed Chi-square tests (PROC FREQ) were used to determine the probability that cows would calve in the open or sheltered area, and if this was dependent upon time of day of calving (day = 0800 to 1959, night = 2000 to 0759) or whether or not there was another cow in the pen at the time of calving.   To determine if and when cows began to selectively use the shelter and separate from their partners before calving, we calculated a ? hour by subtracting the amount of time cows spent in the shelter or spent away from their partner during the 12 h baseline period (by hour) from the 12 h before calving.  If cows performed these behaviours the same amount of time in both periods, ? hour would be zero; if cows performed these behaviours more during the 12 h before calving, ? hour would be positive.  Based on visual inspection of both graphs, it was clear that the data were non-linear and segmented into 2 distinct lines. Non-linear regression (PROC NLIN) showed that the optimal breakpoint (i.e., the hour that the pattern began to change) was 8 h before calving for time spent in the shelter and 8.3 h for time spent away from partner; for simplicity we used a breakpoint of -8 h for both behaviours in the following models. A piece-wise random coefficients model (PROC MIXED) was used to determine the parameters of each line (the first line included data from -12 h to -9 h, the second line included data from -8 h to -1 h). The intercept and slopes for each cow were considered random, and the data was modelled with autoregressive covariance.   4.3. Results 4.3.1. Calving site selection Choice of calving site was dependent on whether or not there was another cow in the pen (Figure 4.2; ?271 = 5.47, P = 0.02). Twenty-one of the 34 single-housed cows calved in the shelter, whereas a majority of pair-housed cows calved in the open area (25 of 38).  Calving site selection was also dependent on the time of day of calving for single-housed cows (?233 = 4.90, P = 0.03), but not pair-housed cows (?237 = 0.01, P = 0.91). Of the single-housed cows that calved during the daytime, 13 of 16 used the shelter; when calving occurred at night cows were equally as likely to use each area.   48   Figure 4.2. Choice of calving site Number of pair-housed (n = 38) and single-housed (n = 34) cows that calved in the sheltered or open area during the day (0800 to 1959) or night (2000 to 0759).   4.3.2. Change in behaviour as calving approached Figure 4.3A shows the change in time that single-housed cows spent using the shelter as calving approached. Before the breakpoint (8 h before calving), cows decreased their use of the shelter over time (slope = -4.5, SE = 0.8, P < 0.001); after the breakpoint, cows increased their use of the shelter over time (slope = 7.2, SE = 1.2, P < 0.001).  Figure 4.3B shows the change in time that pair-housed cows spent greater than 1 cow length away from their partner as calving approached. Before the 8 h breakpoint cows decreased their time spent away from their partner over time (slope = -1.9, SE = 0.6, P = 0.004); after the breakpoint cows increased their time spent away from their partner as calving approached (slope = 3.0, SE = 1.1, P = 0.01).   49    Figure 4.3. Use of shelter and distance from partner before calving (A) The use of the shelter by single-housed cows that calved in that area (n = 11), and (B) the time that pair-housed cows spent away from their partner (n = 20). Differences are between the 12 h before calving and the same 12 h period the day before calving. Data were segmented using a breakpoint of 8 h. The solid lines represent the predicted values for each segment, and the dashed lines represent the upper and lower 95% confidence limits. 50  4.4. Discussion The objective of this study was to determine if, and under what conditions, indoor-housed dairy cows would use a shelter to calve. The cows tested showed a preference to calve in the shelter, but only if they were alone in the pen and if they calved during the day. The effect of time of day may be driven by high activity occurring in the barn during the day (e.g. feeding, cleaning, movement of cows to and from the milking parlour) compared to the relative quiet at night. There are few other studies that have measured the effect of time of day or human disturbance on the calving site selection of ungulates. In one study, Dzialak et al. (2011) sought to determine if pre-parturient elk living near human activity (a gas field) would change their birth-site depending on time of day. Similar to our results, authors found that when calving occurred during the daytime when human activity was highest, calving site was characterized by cover and general avoidance of the activity. In contrast, when calving occurred at night, the elk showed no aversion to the gas field and instead selected calving areas characterized by valley-bottoms and foraging resources.  In our study, cows began to increase their use of the shelter about 8 h before calving. This period coincided with the time that pair-housed cows began to spend more time away from their partners. It is likely that these behaviours occur during the first stage of labour, when cows also begin to show other behavioural signs of calving. It is thought that the first stage of labour (i.e., the dilation of the cervix and start of uterine contractions) in cattle begins approximately 12 h before calving, but it remains unclear exactly when this stage starts (Noakes et al., 2001). The only clear indication of this stage is a dilated cervix, but this can only be detected with palpation and this is rarely done for fear that it will interfere with labour (Wehrend et al., 2006). A slew of behavioural changes are associated with this first stage of labour, including increased restlessness, decreased feeding time, increased tail-raises and increased pawing or licking at the ground (Huzzey et al., 2005; Wehrend et al., 2006; Miedema et al., 2011; Jensen, 2012). Miedema et al. (2011) measured changes in behaviour as calving approached, and discovered that a reduction in feeding has the earliest breakpoint (15.4 h before calving), although there was high variation in this behaviour. Less variable behaviours began to change between 3 to 6 h before calving (tail raising: -6.3 h, lying frequency: -4.2 h, and ground-licking: -3.3 h).  To our knowledge, this is the first study to report an estimation of the time that indoor-housed dairy cattle begin seeking a birth-site and separating from herd mates. In a study using pre-parturient dairy cows housed on pasture, Lidfors et al. (1994) found that cows began to distance themselves from the herd on the day of calving, but the time that this behaviour began to 51  change was unknown. This separation behaviour and a reduction in time spent grazing with the herd were the first behavioural signs of calving recorded by these authors. The breakpoints for the behaviours recorded in our study occurred slightly earlier than many of the other behaviours reported previously. These behaviours may be associated with an early part of the first stage of labour, allowing the cow to find a safe place to calve before calving becomes imminent. When pair-housed, cows were more likely to calve in the open area rather than the secluded shelter. We speculate that the avoidance of the shelter was socially mediated, and may be associated with reduced resource holding potential around parturition, as there were likely cows within the pairs with dominance over the shelters. In several cases we noted that the partner was using the shelter when the focal cow was calving (5 of 14 of the sub-sample of pair-housed cows calving in the open area), but in other cases there was no clear reason why cows chose the open area compared to the shelter. On dairy farms where cows are housed indoors, there is a growing trend for cows to be housed in groups at calving, likely limiting the ability of cows to find seclusion. Group housing also increases the chance that the newborn calf will be the recipient of licking and other maternal behaviour from both the mother and other pre-parturient cows. Approximately 30% of calves were nursed by a cow other than their mother when they were born in a group maternity pen (Edwards, 1983; Illmann and Spinka, 1993). Colostrum quality declines rapidly after calving, so such ?mis-mothering? can interfere with passive transfer of antibodies. Work on sheep has found that adding cubicles to a group pen allowed for periparturient animals to find seclusion from the group and helped to eliminate the interference of the mother-young contact caused by other sheep housed in the pen (Gonyou and Stookey, 1985). Our results indicate that there may be social influences on the use of these cubicles.  4.5. Conclusions Indoor-housed dairy cows preferred to use a secluded calving site when housed alone in the pen, and if calving occurred during the day when human activity was highest. Shelter use and separation from other cows began about 8 h before calving, coinciding with the onset of first stage of labour.   52  5. Preference for a Secluded Area during Calving and Infectious Illness4  5.1. Introduction Results from Chapter 4 suggest that cows, under certain conditions, will calve in a secluded shelter. However, the large shelter used in that study could arguably be impractical for a commercial farm to implement. This chapter describes a study that tested a similar hypothesis, but used a much simpler, practical maternity pen design. The second part of this section also introduces the next theme of this thesis: using behaviour as an indicator of illness.  Dairy cows are gregarious animals with complex social lives (Fraser and Weary, 2009). As herd animals, cows synchronize their behaviours and do not often stray from the protection of the group (Miller and Wood-Gush, 1991). The only documented evidence of self-isolation in wild ungulates and domesticated cattle kept on pasture occurs during the calving period, when dams seek seclusion and separate from the herd (Lidfors et al., 1994; Barbknecht et al., 2011).  Self-isolation in cattle may not be unique to parturition. In other social species, such as laboratory rodents, isolation is also a common sign of illness (Crestani et al., 1991; Arakawa et al., 2010). Dairy cows are at very high risk of disease after calving (Ingvartsen, 2006), and behaviour is becoming increasingly recognized as a useful tool for identifying ill animals (reviewed by Weary et al., 2009). Recent studies have determined that feeding and standing behaviours can predict post-partum disease (reviewed by Sep?lveda-Varas et al., 2013). Yet, to our knowledge, no research has determined if ill dairy cows will isolate if given the chance. In many indoor, commercial dairy farms, cows are given little opportunity to isolate both during calving as well as when they become ill after calving. Management of cows at calving and during illness is highly variable; cows are typically either housed in group pens or individual maternity or hospital pens (e.g., Fogsgaard et al., 2012b). These individual pens are often in high traffic areas, allowing producers easy access to monitor ill and calving cows. No work to date has determined whether cows would choose to seek isolation, if given the opportunity, when housed in an individual maternity or hospital pen. Here we designed a practical individual pen that allowed dairy cows to isolate from herd-mates during calving and the first few days after calving. Our objectives were to determine if: 1) cows were more likely to calve in a secluded area, and 2) cows that become ill after calving spent more time in the secluded area compared to healthy cows.                                                   4 A version of this Chapter has been submitted for publication: Proudfoot, K. L., M. B. Jensen, D. M. Weary, and M. A. G. von Keyserlingk. Submitted. Dairy cows prefer an isolated area to calve and when ill. J. Dairy Sci. 53   5.2. Materials and methods To address our research questions, we conducted an experiment at the Aarhus University?s Research facilities in Foulum, Denmark between September 2011 and February 2012. Cows were cared for according to a protocol approved by the Danish Animal Experiments Inspectorate, Ministry of Justice, Copenhagen, Denmark and the University of British Columbia?s Animal Care Committee (Canadian Council on Animal Care, 2009).   5.2.1. Animals and housing The experiment began with 79 multiparous Danish Holstein dairy cows. Before calving, cows were grouped into 1 of 6 blocks of 14.7 ? 1.5 cows (mean ? STD) based on expected calving date. Cows within a block were moved into one of two group pens (each 9 m ? 15 m) approximately 2 wk before the first expected calving date of cows within a block. Both group pens had deep straw bedding and 12 individual feeding bins (each 75 cm wide).  Either before or during signs of labour, cows were moved from the group pens into one of 10 individual maternity pens (each 3.0 ? 4.5 m) located directly adjacent to each group pen. The time that cows were moved into the individual pen was based upon treatments assigned to them to address a separate research question: the effect of time of movement into the individual pen on labour progress and behaviour (see Chapter 3 for a description of this treatment). To ensure that this treatment did not affect our results, we consider the variation in time that cows spent in the individual pen for both hypotheses tested in this experiment.  The composition of the group pen changed as cows were moved into maternity pens to calve. No cows were added to the group, but we ensured that there was at least 1 other non-experimental cow in the group pen while our cows were in the maternity pens. To test if the number of cows in the pen affected behavioural responses, our preliminary analysis including the number of cows in the group pen as a covariate; there was no effect of this variable on any measure so this variable will not be discussed further.  Figure 5.1 shows the design of the individual pens. These pens were surrounded by 3 ? 1.3 m (w ? h) sides made from horizontal tubular metal bars on three sides, and a feed bunk on the fourth side. In half of the pens, a 1.8 m high plywood barrier covered 2 pen sides. The side facing the group pen was covered half in plywood, creating one secluded corner and one 1.5 m-wide window that allowed visual contact with the group (?partially covered?; Figure 5.1A). The side of the window was alternated for each adjacent pen. In the other half of the individual pens, the pens were left ?uncovered? so that the pen was exposed on all sides, allowing visual and head to head 54  contact with the cows remaining in the group pen as well as cows in adjacent individual pens (Figure 5.1B). After calving, cows and their calves remained in the individual pens for 3 d.  Cows were fed a TMR ad libitum with a forage to concentrate ratio of 79:21 (% DM basis) before calving and 60:40 (% DM basis) after calving. Feed was allocated twice daily at 1000 and 1700. After calving, cows were milked twice daily at 0600 and 1800 using a manual milking machine. Water was available ad libitum in the group and maternity pens via water bowls.      Figure 5.1. Design of individual maternity pens Design of (A) 'partially covered? and (B) ?uncovered? individual maternity pens. Thick black lines indicate plywood barrier, and dashed lines indicate tubular metal bars separating pens. Cows were recorded as in the corner or window side of the pen if a majority of their bodies were on that side. The side of the corner was alternated for each partially covered pen, and the side of the corner for each uncovered pen was determined by matching these to partially covered pens.       55  5.2.2. Inclusion criteria  Inclusion criteria for each research question were defined separately. For both questions, cows were excluded from the experiment if they had: calved in the group pen (n = 10), experienced a difficult calving or had twins (n = 7), were disturbed by human activity (i.e., straw added to her pen) during labour (n = 1), or had milk fever on the day of calving (n = 2). Of the 59 remaining cows, 34 cows were moved to a partially covered maternity pen and 25 cows were moved to an uncovered maternity pen. 5.2.2.1. Location before and at calving.  To minimize the effects of novelty, we excluded cows that calved within 8 h of being moved from the group into the individual pen (n = 18). Two cows were also excluded because they were housed in an uncovered pen with a plywood barrier on one side (both calved next to the barrier). Thus, the final number of cows included to address this research question was 39 (partially covered: n = 19; uncovered: n = 20). 5.2.2.2. Location and health status post-partum. To determine the effect of illness on the use of partial cover, we included only cows that were housed in the partially covered pens (n = 34).  To determine health status, rectal body temperature was taken twice daily at milking for the first 3 d after calving. A health examination was performed 3 and 9 d after calving by one experienced student and one technician, or as needed by the herd veterinarian. During this examination, cows were checked for signs of metritis using a vaginal exam and a 4-point scoring system (see Huzzey et al., 2007 for a description of this scoring system), and signs of ketosis using a small blood sample from the tail vein (Precision Xtra Ketone Glucose and Ketone Monitoring System, Abbott Labouratories, USA, validated by Iwersen et al., 2009). Signs of mastitis were recorded by the milker daily based on the color and consistency of the milk. Cows were considered ill if they had at least 2 consecutive body temperatures > 39.0?C and showed signs of metritis (n = 2; a score of 2 or higher, all diagnosed on d 3 after calving), mastitis (n = 5; 3 diagnosed on d 2 after calving and 2 diagnosed on d 3 after calving), pneumonia (diagnosed by veterinarian on d 2 after calving; n = 1) or a combination of mastitis and metritis (n = 1; diagnosed on d 2 and 3 after calving, respectively). Cows with mastitis and pneumonia were treated on the day of diagnosis with an antibiotic given by the herd veterinarian according to standard operating procedures. This treatment was repeated for 3 d after diagnosis by farm staff. The cow with pneumonia on d 2 after calving and 4 cows with mastitis on d 2 after calving were also given a non-steroidal anti-inflammatory drug (NSAID) on the day after diagnosis (Flunixin 56  Megalumine, 50 mg/mL, 20 mL dosage; Finadyne Solution, MSD Animal Health, UK, or Metacam, 20 mg/mL solution, 15 mL dosage; Boehringer Ingelheim GmbH, Ingelheim am Rhein, Germany).  For these 9 ill cows we noticed a high variation in the amount of time they were kept in the maternity pen before calving (due to our previous treatment; min = 0.8 h max = 120 h). To determine if time in the maternity pen affected behaviour after calving, we ran a preliminary regression between time in the maternity pen and time spent in the corner after calving. There was a negative relationship, whereby cows that spent more time in maternity pen before calving used the corner less (R2 = 0.13, Slope = ?2.4, P = 0.04). To avoid any bias in our final analysis, we paired the 9 ill cows with 9 healthy cows (cows with no clinical signs of infectious or metabolic disease at anytime during the experiment including the 9 d after calving) by time spent in the maternity pen before calving, and included pair in the model as a block. 5.2.3. Behavioural data collection Behaviours were monitored using digital video cameras (MONACOR, TVCCD-140IR) mounted above each pen. The location of the cow when the calf was born  (i.e. when the calf?s hips were fully expelled from the dam) as well as the period beginning 6 h before calving to 72 h after calving was determined using 10-minute scan sampling.  One experienced observer collected all data from video. Figure 5.1 shows the two areas that were recorded for cows in the partially covered pens: 1) the ?corner? where the plywood barrier prevented visual and physical contact to the cows housed in the group pen, and the 2) ?window? where cows could have visual and nose to nose contact with the cows in the group pen. Cows were considered to be in one of the areas when the majority of her body was contained within the area. In the few exceptions where she was positioned directly in the centre of the pen and it was difficult to distinguish which side of the pen she was on, we recorded this as ?centre? (cows spent 11.0 ? 7.8 mean ? STD min/h in the centre of the pen). At the moment of calving, if cows were categorized in the centre we measured how much of their body was on each side of the pen and re-assigned them to be in either the corner or window side (partially covered: n = 1; uncovered: n = 3). We matched each partially covered pen with an uncovered pen to determine the ?side? of the window and corner for the uncovered pens.  To determine if cows changed their use of the corner during the calving period, we recorded data before and after calving. Observations began at 6 h before calving because previous work has determined that behaviour begins to change between 6 to 8 h before calving (Jensen, 2012). This also allowed us to include data from all cows, as the last cow to enter the maternity 57  pen did so at 8 h before calving. We also recorded cow location in the pen during the first 4 h after calf delivery when the intensity of maternal behaviours is highest (Jensen, 2012). To determine if behaviour changed when a cow became ill, we recorded data during the 72 h after calf delivery. Feed intake was measured after calving for ill and healthy cows housed in the individual pens. Feed was measured before the morning delivery at 1000 and any feed refusals were weighed and removed the next morning before fresh feed was delivered. Daily feed intake (on a % DM basis) was calculated as the difference between feed delivery and feed refusals. If a cow calved before 1000, DMI on the day of calving and 2 following days were included in the analysis (healthy: n = 4, ill: n = 5). If cows calved after 1000, DMI on the day after calving and the 2 following days were included in the analysis (healthy: n = 5; ill: n = 4). 5.2.4. Statistical analysis All statistical analyses were performed with SAS software (version 9.2; SAS Inst. Inc., Cary, NC) using the cow as the experimental unit. Two-tailed binomial tests were used to determine if cows were more likely to calve on one side or the other when housed in the partially covered or uncovered pens (with a reference of 50%).  To determine if cows changed their use of the corner in the hours around calving, video data were summed by hour for the 6 h before calving until the 4 h after calving. Data were analyzed with repeated measures ANOVA (PROC MIXED); the model included pen type (uncovered or covered), hour and a pen type ? hour interaction. The correlation between repeated hourly measures was modeled with compound symmetry. After visual inspection of the graph, specific contrasts were made between treatments in the 1 h before and 1 h after calving using the ESTIMATE statement of PROC MIXED. To determine if ill cows spent more time in the corner during the 72 h after calving compared to healthy cows, we first summed the data per 24 h period (0 to 23 h = ?24 h?, 24 to 47 h = ?48 h? and 48 to 71 h = ?72 h?). Video data were missing for 1 ill cow during the 48 h and 72 h, and for 1 ill cow during the 72 h after calving. Data were analyzed with repeated measures ANOVA (PROC MIXED); the model included pair, health (ill or healthy), 24 h period and a period ? health interaction. The correlation between repeated hourly measures was modeled with autoregressive covariance. To determine if ill cows ate less than healthy cows in the first 3 d after calving, data were analyzed with repeated measures ANOVA (PROC MIXED). The model included pair, health (ill or healthy), day and a period ? health interaction. After visual inspection of the graph, and a 58  period ? health interaction (P = 0.01), specific contrasts were made between health categories for all 3 d separately using the ESTIMATE statement of PROC MIXED.  5.3. Results 5.3.1. Location of cows at calving   Figure 5.2 shows the calving site for cows housed in uncovered (n = 20) and partially covered (n = 19) pens. Cows in the uncovered pen showed no preference; 10 calved on the corner side and 10 calved on the window side of the pen. Cows housed in the partially covered pens preferred to calve on the corner side of the pen (15 of 19 cows calved in the corner; P = 0.01). 5.3.2. Location of cows around calving  There was a pen type ? hour interaction for the time spent in the corner (Figure 5.3; P = 0.04). Cows housed in the partially covered pen used the corner more in the 1 h before (P = 0.003) and after (P = 0.02) calving compared to those housed in the uncovered pens.      Figure 5.2. Choice of calving site in individual maternity pens Percentage of cows that calved in the corner and window sides of partially covered and uncovered individual maternity pens.   59               *P < 0.05, **P< 0.01  Figure 5.3. Time spent in secluded corner Time spent in the corner side of the pen during the 6 h before calving until the 4 h after calving for cows housed in partially covered and uncovered individual maternity pens.   5.3.3. Dry matter intake and location of ill and healthy cows after calving   We detected an effect of health category (ill: 17.1 vs. healthy: 20.0 ? 0.6 kg DM/d; P = 0.02) as well as a day ? health interaction for DMI (P = 0.02). Ill cows ate less than healthy cows on the first day after calving (13.4 vs. 20.4 ? 1.1 kg DM/d, P < 0.001), but not on day 2 (18.5 vs. 20.1 ? 1.1 kg DM/d, P = 0.32) or day 3 (19.5 vs. 19.6 ? 1.2 kg DM/d, P = 0.95).  Figure 5.4 shows the lying time and time spent in the corner for ill and healthy cows. Ill cows tended to spend more time lying (P = 0.08) and spent more time in the corner compared to healthy cows during the 3 d after calving (P = 0.0003). No interactions between period and health were detected for lying time or time spent in the corner.    60    Figure 5.4. Behaviour of ill and health cows (A) Lying time and (B) time spent in the corner for ill (n = 9) and healthy (n = 9) cows during the 72 h after calving. 61  5.4. Discussion The objectives of this study were to determine if cows would seek seclusion during calving and during illness. We provided cows an individual pen that allowed the option to spend time in a secluded area or in a more open area with visual access to the group pen. Cows preferred to use a secluded area for calving, and began to use this area more in the 1 h before calving. After calving, cows that became ill continued to use the secluded area more than healthy cows.  Our finding that cows used a secluded area to calve agrees with results from Chapter 4 and those of an earlier study on cows kept on pasture (Lidfors et al., 1994). On pasture, cows distanced themselves from the herd on the day of calving, choosing to calve in a dry area surrounded by tree or bush cover (Lidfors et al., 1994). Like other ?hider? species, calves spent the first few days of life hidden in brush or tree cover in a similar location to where they were born while the dam grazed nearby (Vitale et al., 1986). If there was no suitable cover available, cows calved within close proximity of the herd (Lidfors et al., 1994). It has been suggested that hiding is an anti-predator strategy or a method of facilitating the dam-calf bond by preventing disturbance from other cows (Leuthold, 1977).  When housed indoors, cows are free from predation, are given ad libitum access to feed and are protected from inclement weather. Thus, the reason for cows hiding in this study remains unclear. However, we speculate that this behaviour was partly due to the presence of cows in the group pen adjacent to the maternity pens. The partially covered maternity pen allowed cows to isolate from cows in the adjacent group pen, while allowing some head contact between cows via the window. Perhaps cows were using the secluded area to avoid attention from cows in the group pen drawn to odors associated with calving (see Chapter 3 for an example of this behaviour).  It should also be noted that human activity in the barn was minimal; activity was limited to feeding and milking, or when cows were inspected for signs of labour. However, this activity may have also influenced cow behaviour. In our previous study (Chapter 4), indoor-housed cows primarily used a shelter to calve during the daytime; we speculated that this behaviour may have been partly driven by the high traffic of people and cows in the barn during that time. Unfortunately, in the present study we did not have enough cows calve on each side of the partially covered pen to assess differences in time of day of calving, however, it should be noted that all 4 cows that calved in the window side did so during the daytime. Thus, cows in this study may have been less affected by activity during the daytime, and potentially more affected by their close proximity to cows in the group pen.  62   Cows used the secluded area most in the 1 h before and 1 h after calving. In other studies, cows were reported to begin to isolate in the 24 and 8 h before calving (Lidfors et al., 1994 and Chapter 4, respectively). The difference in the present study could be driven by the rather small size of the individual pen compared to space allotted to cows in the other two studies. As cows approach calving they become restless, characterized by increased postural transitions from standing to lying (Huzzey et al., 2005; Jensen, 2012). In this study, the small size of the pen (4 ? 3 m) did not allow us to determine if a cow was switching sides of the pen or was simply restless, as cows changed their location within the pen (from the corner to the window side and vice versa) when they transitioned between lying and standing. However, cows increased their use of the secluded area in the hour before calving when restlessness is known to be high, which suggests that the motivation to hide may have increased as calving became imminent. Cows in the partially covered pens also spent more time in the secluded area in the hour after calving. This behaviour is likely due to the position of the calf at calving, as most cows in the partially covered pens calved in the secluded area. In studies of maternal behaviour, cows spend the first few hours after calving licking the calf (Edwards and Broom, 1982; Jensen, 2012). Licking is thought to be essential for the development of the dam-calf bond, as calves that are prevented from being licked are at greater risk of being rejected by the dam (Hudson and Mullord, 1977). Spending time in the secluded area of the pen with the calf would thus likely help strengthen the dam?s bond with the calf. For our second research question, we hypothesized that cows would also seek seclusion as a result of illness. When an animal becomes ill with an infectious disease, two key changes occur to help the animal fight the infection: 1) a febrile (fever) response is mounted, and 2) strategic behaviours are changed to conserve and re-direct energy to the immune system (Hart, 1988). Sickness behaviours are driven by signals from the immune system in the brain (e.g., cytokines such as IL-1 and ??F?, Kelley et al., 2003), are generally non-specific, and include a reduction in feed intake and feeding behaviour, as well as a general reduction in activity and social behaviours (Dantzer, 2001).  In our study, cows with signs of infectious disease and a febrile response ate less during the first day after calving. A similar response has been found in other studies using cows with metritis (Huzzey et al., 2007) and mastitis (Fogsgaard et al., 2012). This difference disappeared in the next 2 d after calving, likely because 7 of the 9 ill cows were treated for their illnesses (5 on d 2 and 2 on d 3). Moreover, ill cows tended to spend more time lying down, which is also a common sickness behaviour. This result differs from previous work which showed that mastitic 63  cows spent less time lying compared to healthy cows, likely as a means to relieve pressure and pain from the udder (Siivonen et al., 2011; Fogsgaard et al., 2012; Cyples et al., 2012). The slight increase in lying time recorded here was likely driven by cows with metritis and pneumonia. A novel finding from this study was that ill cows spent more time in the secluded area compared to healthy cows. Although this study is the first to demonstrate isolation behaviour in ill dairy cows, reduced social exploration has been found in studies using laboratory animal models of sickness behaviour (e.g., Crestani et al., 1991; Bluth? et al., 1999). In these studies, sickness behaviour was caused by injecting the animals with either LPS (an endotoxin) or IL-1 to mimic an infectious disease. Animals infected with LPS or IL-1 reduced social exploration when presented with an unfamiliar conspecific (e.g., sniffing, grooming and chasing, Crestani et al., 1991; Bluth? et al., 1999). Animals may benefit from avoiding others when they have an infection, as avoiding other animals prevents them from being introduced to secondary infections (Loehle, 1995). In this study, we were limited to a small number of ill cows affected with 3 different diseases. Although sickness behaviours are normally generalized across infectious diseases when a fever is present (Hart, 1988), it would be useful to conduct a larger study to reduce some of the variation due to illness type and severity as well as day of diagnosis and treatment.  Further research is also encouraged to determine the mechanism that drives this behaviour during calving. The mechanism of sickness behaviour is becoming relatively well-known in the literature; during illness, markers of inflammation including pro-inflammatory cytokines in the brain produce changes in behaviour (Dantzer, 2001). It is not unlikely that hiding during parturition is driven by a similar mechanism, as there is evidence that pro-inflammatory cytokines are expressed during labour in dairy cows (Jonsson et al., 2013), and are thought to play a key role initiating labour in humans (Kamel, 2010). Perhaps cytokines expressed during parturition also mediate changes in behaviour, such as social isolation and reduced feed intake.   5.5. Conclusions When given the opportunity, cows housed in an individual maternity pen preferentially used a secluded area to calve. Cows housed in the covered pens began using the secluded area more than those in the uncovered pens in the hour before calving, and continued to use it in the hour after calving. Ill cows ate less, spent slightly more time lying, and more time using the secluded area compared to healthy cows.  64  6. Behavioural Indicators of Dystocia5  6.1. Introduction The second part of Chapter 5 provided evidence that cows with infectious disease change their behaviour near the time of diagnosis. The current chapter as well as the next chapter will continue on this theme, and describe studies that set out to measure behaviour well in advance of poor health, including dystocia and lameness.  It has been estimated that between 2 and 23% of cows in a herd experience difficult calvings that require farmer or veterinarian assistance (Mee, 2008). Calvings which are prolonged and where the extraction of the calf is especially difficult, requiring the assistance of farm staff, are referred to as dystocia (Mee, 2004). Dystocia can result from the failure of expulsive forces during parturition, birth canal inadequacy, fetal mal-positioning and disproportionate calf size to the dam?s pelvic size (Noakes, 2001).  Dystocia is painful (Huxley and Whay, 2006), is associated with reduced milk yield, increased risk of health disorders and reproductive complications for the cow (Oltenacu et al., 1988) and increased risk of death for the calf (Tenhagen et al., 2007). In response to these welfare and production costs, previous research has focused on identifying factors that increase the risk of dystocia. For instance, young and small cows are at high risk for dystocia, especially when the calf is large relative to the size of the dam (Meijering, 1984; Johanson and Berger, 2003). A mal-positioned calf (Meijering, 1984), abnormally short or long gestation length (Philipsson, 1976), as well as overfeeding and underfeeding in the last trimester (Grunert, 1979) can also increase the risk of dystocia.  Some research has described behavioural changes associated with dystocia, but have been largely based on anecdotal evidence and descriptive data (Mee, 2008). It remains unclear if detailed feeding and standing behaviour are altered when cows have dystocia, and how well these changes in behaviour can identify cows at-risk for dystocia. The objectives of the current study were to determine if feeding or standing behaviour differed for cows with and without dystocia, and to assess how well these measurements could identify cows at-risk for dystocia.                                                     5 A version of Chapter 6 has been published: Proudfoot, K. L., J. M. Huzzey, and M. A. G. von Keyserlingk. 2009. The effect of dystocia on dry matter intake and behaviour of Holstein cows. J. Dairy Sci. 92: 4937-4944. 65  6.2. Materials and methods 6.2.1 Animals, experimental design and diet The study was conducted between August 2005 and March 2006 at the University of British Columbia?s Dairy Education and Research Centre (Agassiz, British Columbia, Canada). Animals were cared for according to the guidelines of the Canadian Council on Animal Care (1993). Thirty-two primiparous and 69 multiparous (parity = 3.2 ? 1.3, mean ? STD) Holstein dairy cows were used. Cows were housed in 3 group pens depending on stage of lactation (i.e., pre-calving, maternity, and post-calving). Pre- and post-calving pens contained 20 free-stalls fitted with a mattress (Pasture Mat, Promat Inc., Woodstock, Ontario, Canada) covered with 5 cm of sand. Stocking density was maintained at 20 animals per pen. The maternity pen was the same dimension as the pre- and post-calving pens; the free-stall hardware in this pen was removed to create a sand-bedded pack. All pens contained vulcanized rubber floors in the alleys and intersections (Red Barn Dairy Mat, North West Rubber Mats, Ltd., Abbotsford, British Columbia, Canada). Each pen contained 12 Insentec electronic feed bins and 2 Insentec electronic water troughs that recorded individual intake and time spent at the feed bin per visit (Insentec, Marknesse, Holland). Chapinal et al. (2007) provide a description and validation of this system.  Cows entered the pre-calving pen 25 ? 2 d before their expected calving date and were moved to a maternity pen when they showed physical signs of imminent calving (i.e., udder enlargement, milk let-down, raised tail and/or relaxation of sacrosciatic ligament). The maternity pen consisted of a sand-bedded pack, 6 Insentec feed bins and 1 Insentec water trough. A maximum of 2 cows were kept in the maternity pen at any given time and cows were moved to the post-calving pen within 16 h after calving. Cows in the post-calving pen were milked twice daily at approximately 0700 and 1700 h.  Feed bins were refilled twice daily at approximately 0800 and 1600 h. Cows in the pre-calving and maternity pens were fed a TMR ad libitum consisting of 21.3% corn silage, 42.8% alfalfa hay, and 35.9% concentrate and mineral mix on a DM basis (DM: 50.8 ? 1.2%, CP: 14.4 ? 1.0% DM, ADF: 35.0 ? 2.7% DM, NDF: 45.6 ? 2.6% DM, and NEL: 1.4 ? 0.1 Mcal/kg). Cows in the post-calving pen were fed a TMR consisting of 21.3% grass silage, 14.7% corn silage, 12.3% alfalfa hay, and 51.7% concentrate and mineral mix on a DM basis (DM: 51.1 ? 1.8%, CP: 17.7 ? 1.0% DM, ADF: 23.7 ? 1.4% DM, NDF: 36.1 ? 1.8% DM, and NEL: 1.66 ? 0.02 Mcal/kg). 6.2.2. Dystocia classification and cow participation  Upon pelvic examination from the farm manager or one of 6 experienced farm staff, a cow was assisted with calving under the following criteria: 1) the calf was mal-positioned, 2) the 66  amnionic sac was ruptured and the front feet of the calf protruded for longer than 60 min, or 3) the cervix was fully dilated but the dam was unable to deliver the calf. When assistance was considered necessary, the standard operating procedure for calving assistance was used. Firstly, the farm manager and experienced staff restrained the cow and lubricated the vaginal opening. Obstetrical (pulling) chains were then attached to the front or back feet of the calf, depending on orientation, and feet were pulled in synchrony with the dam's abdominal contractions until the calf was fully expelled. Following calf delivery, cows were given one of three calving scores: 1 = unassisted delivery (i.e., ?eutocia?), 2 = easy assistance (i.e., one person required to pull the calf out) or 3 = difficult assistance (i.e., ?dystocia?, 2 or more people required to pull the calf). Cows that were classified as having dystocia (calving score of 3) were compared to cows with eutocia (calving score of 1).  Of the 101 cows in the study, 54 cows scored 1, 23 scored 2 and 24 scored 3. Of the 24 cows with dystocia, 13 were removed from the final analysis due to a condition diagnosed by the herd veterinarian or hoof trimmer that might severely affect their feeding or standing time (i.e., lameness, ketosis, hypocalcaemia or retained placenta), or if they were removed to another maternity pen to calve where we were unable to monitor feed or water intake, leaving 11 cows in the final dystocia group. The behaviour of these 11 cows was compared to 11 cows with eutocia; the groups were balanced for parity (5 primiparous and 6 multiparous cows in each group). Previous work in our laboratory showed that the occurrence of sub-clinical metritis post-calving can affect DMI during transition (Huzzey et al., 2007); thus, treatment groups were balanced for sub-clinical metritis incidence. Sub-clinical metritis was diagnosed using vaginal discharge scores recorded every 3 d post-calving until 21 d based on a scoring system described by Huzzey et al. (2007). Moreover, we ensured that the approximate period of calving (i.e., day or evening) and feed intake on d -4 prior to calving were not statistically different between groups to reduce variation due to individual feed intake and lying behaviour during the experimental period.  6.2.3. Intake and behavioural data collection  Calving time (i.e., the hour of delivery) was used as the reference point from which to analyze feed and water intake and feeding, drinking and standing behaviour pre- and post-calving. Calving time was determined from 2 video cameras (CCTV camera, model WV-BP330; Panasonic, Osaka, Japan) with F1.4/2.5-6 mm vari-focal lenses mounted above the feed alley and bedded pack of the maternity pen. These cameras were connected to a video multiplexer (Panasonic Video Multiplexer, WJ FS 416) and a time-lapse videocassette recorder (Panasonic Time-Lapse VCR, AG-6540), which recorded the video on VHS tapes for later use. Cows were 67  identified on the videos by a hair-dyed alphanumeric symbol on their back and sides. To facilitate video recording during the night hours, red lights (100 W) were hung adjacent to the cameras. Calving time was defined as the time in which the full calf was visibly outside of the dam. For cows that calved standing, calving time was the moment the calf hit the ground, and when cows were lying down, calving was considered the time when the back legs were completely visibly outside of the dam.  The Insentec feed and water intake system was used to calculate daily feed and water intake and feeding and drinking time. Dry matter intake measures were obtained by correcting weekly as-fed intakes for the DM content of the feed. Data from the Insentec system were summarized in 24-h intervals for the period from 48 h before calving time to 48 h after the recorded time of calving. The number of meals consumed per day was computed from the Insentec data by first establishing a meal criterion. A meal criterion was the minimum time interval away from the feeder such that the next visit was considered a new meal, and was identified using discontinuities in the distribution of intervals using the mixed distribution method described by Tolkamp and Kyriazkis (1999). Based on our data, a pooled meal criterion of 16.7 min was used to calculate the number of daily meals and the size and duration of each meal. Standing behaviour data was collected using modified dataloggers (Gemini Dataloggers Ltd., Chichester, UK), validated by O?Driscoll et al. (2008), which were fitted on the hind leg of each cow. These dataloggers were modified with a mercury switch that recorded leg orientation (horizontal or vertical) at 1 min intervals, and these data were used to quantify total daily standing and lying time as well as the number of times cows transitioned from standing to lying positions (i.e., ?standing bouts?). Dataloggers were attached to the hind leg of cows upon entry into the pre-calving pen and were removed 21 d post-calving; these data were offloaded weekly when cows were restrained in the milking parlor. 6.2.4. Statistical analysis Feeding and drinking variables were averaged into 24 h periods relative to calving for analysis (i.e., -48 h: 48 to 25 h before calving; -24 h: 24 to 1 h before calving; 24 h: 0 to 23 h after calving; 48 h: 24 to 47 h after calving). The cow was considered the experimental unit. All data were analyzed with SAS software (version 9.2; SAS institute Inc., Cary, NC). A repeated measures ANOVA (PROC MIXED) was used to determine differences in feeding and drinking behaviours (DMI, feeding time, number of meals, meal size, latency to feed before and after calving, water intake and drinking time) between cows with dystocia and eutocia. The models included the fixed effects of parity (primiparous or multiparous), 24 h period relative to calving, 68  level of assist (dystocia or eutocia), and the period ? level of assist interaction. When a period ? level of assist interaction (P < 0.05) was detected, the dependent variable was stratified by period.  The period with the most dramatic differences in feed intake occurred in the 24 h before calving (Figure 6.1). To determine the time in which cows with dystocia began to differ in dry matter intake, we calculated the cumulative hourly intake from 24 h before to the hour of calving, and determined the hour in which the rate of intake first differed based on the first hour in which the groups differed (P < 0.05) using a t-test.  It was hypothesized that the time between calving and eating would impact the size of the first meal of the post-calving diet after calving. To test this, regression (PROC REG) was used to calculate an R2 value for each level of assistance.  The same 24 h periods were used to analyze standing behaviour. Due to missing datalogger data from 1 cow in each group (1 primiparous and 1 multiparous), groups were rebalanced for parity, resulting in 9 animals in each group (i.e., 2 were removed from each group to maintain the same number of multiparous and primiparous cows in each group). The model used to test differences in standing variables was constructed and interpreted in the same manner as the models for the feeding and drinking variables. To determine how well changes in behaviour during the 24 h before calving could estimate the likelihood of dystocia, discriminant analysis (PROC DISCRIM) was used. Only variables that were found significantly different between the 2 levels of assistance during the 24 h before calving (P < 0.05) were included in the discriminant analysis, these included: DMI, feeding time, water intake and cumulative standing bouts. Thresholds for each variable were derived using the PROC DISCRIM procedure, and from these the sensitivity (likelihood that a cow with dystocia was correctly identified as higher or lower than the threshold depending on the variable), specificity (likelihood that a cow without dystocia was correctly identified as higher or lower than the threshold depending on the variable), and accuracy (likelihood that a cow was correctly classified as having dystocia or not by the threshold) were calculated.   6.3. Results 6.3.1. DMI and feeding time  A period ? level of assistance interaction was detected for DMI (P = 0.04; Figure 6.1A). During the 48 h period before calving, cows with dystocia consumed 12% less DM than cows with eutocia (P = 0.07), and consumed 24% less during the 24 h before calving (P = 0.02). There were no differences in DMI during 24 h and 48 h period after calving between the groups.  69  The daily feeding time of cows with dystocia was 34 min less than that cows without dystocia during the 24 h period before calving (P = 0.04; Figure 6.1B), but this difference was not observed during any other period. During the 24 h before calving, cows with dystocia began to decrease their intake compared to cows with eutocia by 11 h before calving (P = 0.04; Figure 6.2).                            ? P < 0.10, * P < 0.05   Figure 6.1. Feeding behaviour of cows with and without dystocia LS means ? SE of daily (A) DMI and (B) feeding time for cows with dystocia (n = 11) and eutocia (n = 11) during two 24 h periods before and after calving.   70      Figure 6.2. Cumulative DMI of cows with and without dystocia LS means ? SE of hourly cumulative DMI from 24 h before calving until the hour before calving for cows with dystocia (n = 11) and eutocia (n = 11). 71  6.3.2. Meals and time to feed after calving No differences were detected in the number of meals per period between cows with dystocia and eutocia. Yet, cows with dystocia ate smaller meals during the 48 h period before calving (1.4 vs. 1.7 ? 0.1 kg/meal; P = 0.04) and the 24 h period before calving (0.8 vs. 1.1 ? 0.1 kg/meal; P = 0.01) compared to cows with eutocia. No differences in meal size were found after calving.  No differences were found in the time between the last visit to the feed bins and time of calving (303 vs. 224 ?71 min, dystocia vs. eutocia; P = 0.44). But, cows with dystocia had a shorter latency to visit the feed bins after calving compared to cows with eutocia (141 vs. 326 ? 44 min; P = 0.01). Regression analysis revealed a positive relationship between latency to begin feeding after calving and the size of first meal for cows with dystocia (Figure 6.3). The longer a cow waited to access the feeder, the larger her first meal after fresh feed delivery (R2 = 0.43, P = 0.02); this relationship was not found for cows with eutocia (R2 = 0.01, P = 0.79).     Figure 6.3. Meal size after calving for cows with and without dystocia Relationship between the time to feed after calving and the size of the first meal after calving for cows with dystocia (n = 11) and eutocia (n = 11). 72  6.3.3. Drinking behaviour and water intake During the 24 h before calving, cows with dystocia consumed less water than those with eutocia (P = 0.04; Figure 6.4A). During the 24 h after calving, this relationship reversed, as cows that had dystocia tended to consume more water (P = 0.08). There were no differences in water intake between the groups during any other period. There were no differences in drinking time between cows with dystocia and cows with eutocia throughout the experimental period (Figure 6.4B).          ? P < 0.10, * P < 0.05  Figure 6.4. Water consumption of cows with and without dystocia LS means ? SE of (A) water intake and (B) drinking time for cows with dystocia (n = 11) and eutocia (n = 11) during two 24 h periods before and after calving. 73  6.3.3. Standing behaviour Cows with dystocia had higher cumulative standing bouts throughout the experimental period compared with cows with eutocia (P < 0.001; Figure 6.5A). No difference in standing time was found between cows with dystocia and cows with eutocia (Figure 6.5B).        Figure 6.5. Standing behaviour of cows with and without dystocia LS means ? SE of the number of cumulative (A) standing bouts and (B) standing time for cows with dystocia (n = 9) and eutocia (n = 9) during two 24 h periods before and after calving.  74  6.3.4. Discriminant analysis Table 6.1 shows the thresholds, specificity, sensitivity and accuracy of DMI, feeding time, water intake and cumulative standing bouts during the 24 h before calving at identifying cows with dystocia. Cumulative standing bouts were the most accurate at identifying cows with dystocia (77.8%); cows with more than 30 bouts over the 24 h before calving were more likely to experience dystocia. Dry matter intake was 77.3% accurate at detecting cows with dystocia; cows that ate less than 9.6 kg during the 24 h before calving were more likely to experience dystocia at calving. Feeding time and water intake were moderately accurate at detecting cows with and without dystocia.  Table 6.1. Discriminant analysis of behaviours used to predict dystocia Thresholds, specificity, sensitivity and accuracy of behaviours in differentiating cows with dystocia from cows with eutocia during the 24 h before calving   Threshold1 Specificity2 Sensitivity2 Accuracy2 DMI (kg/d) 9.6 72.7% 81.8% 77.3% Feeding time (min/d) 103.6 63.6% 54.6% 59.1% Water intake (kg/d) 31.7 81.8% 54.6% 68.2% Cumulative standing bouts (no./d) 33.8 77.8% 77.8% 77.8%  1 Thresholds for each variable were determined using discriminant analysis.  2 Specificity and sensitivity refer to the proportion of cows with eutocia and dystocia, respectively, were correctly identified as higher or lower than the threshold depending on the variable. Accuracy refers to the proportion of cows that were correctly classified as having dystocia or not by the threshold of each variable.  6.4. Discussion It is clear from previous research that cows will adjust their behaviour around the time of calving. For instance, Huzzey et al. (2005) showed that as calving approached, cows decreased their time spent at the feed bunk, and transitioned from standing to lying positions more often. This current study is the first to show that cows with difficult calvings changed their feeding and resting behaviour in a different way than cows with unassisted calvings. To our knowledge, only 2 studies measured the effect of dystocia on cow behaviour, but both monitored cows during the stages of parturition and not in the period before the onset of parturition. Wehrend et al. (2006) reported that a higher proportion of cows with dystocia rubbed against the wall, scraped the floor 75  and discharged urine compared to cows with eutocia, but detected no differences in feeding or standing behaviour.   A major outcome was that cows with dystocia reduced their feed and water intake, as well as feeding time, during the hours leading up to calving. These differences in behaviour could be due to a number of factors. For example, as dystocia is often caused by a disproportion of calf to dam size, larger calves could reduce the amount of space available in the rumen, and thus, reduce intake for cows with dystocia (Stanley et al., 1993). Alternatively, reductions in intake could be due to pain associated with a large or mal-positioned calf. According to Weary et al. (2006), one type of response to pain was a decline in the frequency or magnitude of certain behaviours that animals are highly motivated to perform, such as feeding and drinking behaviours. Although no study to date has attempted to assess the extent of tissue damage associated with dystocia prior to calving, decreases in feed intake and feeding time were identified as behavioural indicators of painful lameness conditions in dairy cows (Gonzalez et al., 2008). An increase in position changes before calving performed by cows with dystocia may be indicative of pain, as increased general restlessness is a validated behavioural indicator of pain associated with tissue damage in other species, such as lambs following castration and tail docking (Molony and Kent, 1997). It was not the intention of the current study to assess the level or duration of pain associated with dystocia; thus, future research should assess the physiological or psychological mechanisms before calving that drive these behavioural changes.  Cows with dystocia compensated for deficits in water intake before calving by tending to consume more water during the 24 h after calving compared to cows without dystocia. There was a greater increase in DMI for cows with dystocia between the period 24 h before and 24 h after calving compared to cows that calved without assistance, and a shorter latency to eat a large meal following calving at fresh feed delivery when cows are most motivated to access the feed bunk (DeVries et al., 2003). These results suggest that cows with difficult calvings were hungrier due to reduced intakes before calving, and may be at risk for complications associated with eating a high energy diet following a period of food deprivation such as acute ruminal acidosis (Radostits et al., 1994). But, as this study was not designed to test signs of ruminal acidosis, more research is needed in this area.   This was the first study to show that DMI and the number of standing bouts during the 24 h before calving may help correctly identify cows that experience dystocia or eutocia at calving. Although a sensitive measurement tool such as the Insentec system used may be impractical for commercial use at present, the advancement of technology may soon allow use of automatic 76  measurements of feeding and standing behaviour as predictors of dystocia in the 24 h period before calving. Special attention should be given to monitoring feed intake during the brief period before calving (i.e., 11 h), when cows are may be showing other signs of calving (e.g., restlessness, raised tail, an enlarged udder, and relaxation of the tail ligament).  Previous studies have used the finding that cows will alter their behaviour around calving as an indication that special attention and alternative management be given to transition cows (Cook and Nordlund, 2004; Huzzey et al., 2005). The current study found that cows with dystocia had exaggerated decreases in feed intake, and increases in position changes compared to cows without dystocia in the 24 h before calving. Thus, if a maternity area with easy access to feed and water and comfortable lying space is limited, priority should be given to cows most at-risk for dystocia.    6.5. Conclusion  Cows with dystocia had lower DMI and water intake, and higher standing bouts during the hours leading up to calving compared to cows that calved without assistance. These behavioural changes may be indicative of pain beginning up to 24 h before calving. In an attempt to compensate for these intake deficits before calving, cows with dystocia approached the feed bunk faster than cows with eutocia and consumed larger first meal sooner. The behavioural changes before calving may be useful tools to help identify cows at risk for dystocia, and may help focus future research into the design of maternity pens and management of cows at-risk for dystocia.  77  7. Behavioural Indicators of Lameness6  7.1. Introduction In Chapter 6, evidence was provided that behaviour could identify cows at-risk for dystocia about 24 h before calving. Here, we describe an experiment where behaviour during transition was used to predict lameness months after calving. Lameness during this post-calving period is caused primarily by the presence of claw horn lesions, including sole hemorrhages, white line hemorrhages and sole ulcers (Bicalho et al., 2009). Claw horn lesions typically become visible 8 to 12 wk after calving (Leach et al., 1998), but the transition period is considered a ?trigger factor? for when these lesions usually start (Cook and Nordlund, 2009). Physiological changes during transition that predispose cows to claw horn lesions include weakening of the connective tissue of the hoof suspensory apparatus (Tarlton et al., 2002) and a decrease in the thickness the digital cushion (Bicalho et al., 2009), both of which can impair the resilience of the feet to external stresses. However, the development of claw horn lesions during transition is likely multi-factorial, as physiological changes are compounded by external risks for lesions, such as a hard standing surface (Webster, 2002).  Cow behaviour during transition may also contribute to the development of claw horn lesions. In one study using transition cows, Chapinal et al. (2009) found that cows diagnosed with ulcers in mid-lactation had a faster increase in standing time during transition compared to cows that remained healthy, but reasons for this increase in standing time were not assessed. During other stages of lactation, social status within the pen (Galindo and Broom, 2000), and feeding behaviour (Nocek, 1997) have been linked to the risk of lameness. However, these behavioural predictors have not been assessed during the transition period. The objective of this study was to determine if cows that are diagnosed with sole hemorrhages or sole ulcers in mid-lactation differ in their standing, social and feeding behaviour during the transition period from those cows that do not develop lesions. A second objective was to determine which behaviours during transition that would best predict which cows would be later diagnosed with these claw horn lesions.                                                   6 A version of Chapter 7 has been published: Proudfoot, K. L., D. M. Weary, and M. A. G. von Keyserlingk. 2010. Behaviour during transition differ for cows diagnosed with claw horn lesions in mid-lactation. J. Dairy Sci. 93: 3970-3978. 78  7.2. Materials and methods 7.2.1. Animals, housing and diet  This study was conducted at the University of British Columbia?s Dairy Education and Research Centre (Agassiz, BC, Canada); all animals were cared for according to the guidelines of the Canadian Council on Animal Care (1993). We monitored 55 multiparous Holstein dairy cows (parity = 2.9 ? 1.1, mean ? STD) over a 9 month period. Animals were housed in one pre-calving and one post-calving pen, each containing 20 free stalls fitted with mattresses (Pasture Mat, Promat Inc., Ontario, Canada) covered with 5 cm of sand bedding. Pens had vulcanized rubber floors in the alleys and crossovers (Red Barn Dairy Mat, North West Rubber Mats, Ltd., Abbotsford, British Columbia, Canada) and an electronic feeding system that included 12 feed bins (Insentec, Marknesse, Holland, see Chapinal et al., 2007 for a validation of this system). All cows could access all feed bins, but only one cow could access a single bin at any given time. The feed bin recorded the time of day, duration and intake during each cow?s visit. Stocking density was maintained at 20 cows per pen (i.e., 20 cows: 12 feed bins).  Cows entered the pre-calving pen 25 ? 2 d before their expected calving date and were moved to the maternity pen when they showed physical signs of imminent calving (i.e., udder enlargement, milk let-down and relaxation of sacrosciatic ligament). The maternity pen contained a sand-bedded pack with 6 Insentec feed bins and 1 Insentec water trough; there was a maximum of 2 cows in this pen at any given time. Cows were moved to a post-calving pen within 24 h after calving where they remained for 21 d. Cows in the post-calving pen were milked twice daily at approximately 0700 h and 1700 h. Only behavioural data collected from the 2 wk before calving to 3 wk after calving were included in the analysis. For the remainder of the trial (i.e. 15 wk), cows were housed in one of two pens with vulcanized rubber floors in the alley adjacent to the feed bunk; the flooring in the remaining alleys and crossovers was grooved concrete.  For all pens, fresh feed was provided twice per day at approximately 0800 h and 1600 h.  Three samples of pre- and post-calving TMR were taken per week and pooled before analysis. Samples were dried at 600C for 2 d to determine dry matter (DM). Dried weekly samples were pooled into monthly samples and sent for nutrient analysis (Cumberland Valley Analytical Services INC., Maryland, USA) to determine the average CP, ADF, NDF and NEL content of the feed (Table 7.1).   79  Table 7.1. Feed composition of pre- and post-calving diets Ingredient and chemical composition of the pre- and post-calving total mixed ration  Composition Pre-calving Post-calving Ingredient, %DM   Corn silage 21.3 14.7 Grass silage -- 21.3 Alfalfa hay 42.8 12.3 Concentrate1 35.9 51.7 Chemical composition   DM, % 50.8 51.1 CP, % of DM 14.4 17.7 ADF, % of DM 35.0 23.7 NDF, % of DM 45.6 36.1 NEL (Mcal/kg) 1.40 1.66  160.0% ground beet pulp, 10.0% rye distillers, 8.0% canola meal, 7.0% Amipro (Unifeed Ltd., Chilliwack, BC, Canada), 6.0% flattened barley, 2.95% dairy trace mineral/vitamin premix, 1.33% Vitamin E, 1.0% dicalcium phosphate, 0.88 % niacin, 0.75 % magnesium oxide,  0.6% iodized salt, 0.59 % limestone, 0.5% calcium sulfate, 0.4 % cane molasses.    7.2.2. Feeding behaviour The time of calving for each cow was determined from video recordings. Two cameras (CCTV camera, model WV-BP330, Panasonic, Osaka, Japan) were mounted directly above the lying stalls of the pre-calving and maternity pens. Cameras were connected to a video multiplexer (Panasonic Video Multiplexer, WJ FS 416) and a time-lapse videocassette recorder (Panasonic Time-Lapse VCR, AG-6540). For all behaviours, the day after calving (24 h) was adjusted so that the first hour of this period was the hour of calving. All other days (2 wk before to 3 wk after calving) were recorded from midnight to midnight. Data collected from the feed bins were used to determine the time of day, duration and the amount of feed consumed at each visit to the bin. Dry matter intakes were calculated by correcting weekly intakes for DM content. Feeding rate was calculated as the ratio of DMI to time spent feeding for each visit to the feed bins; these feeding rates were then averaged over a day to create one value per cow per day.  80  The number of meals consumed per day was computed by first establishing a meal criterion. A meal criterion was the minimum time away from the feeder such that the next visit was considered a new meal; this criterion was identified using discontinuities in the distribution of intervals using the mixed distribution method described by Tolkamp and Kyriazkis (1999). Based on our data, a pooled meal criterion of 16.7 min for the 2 wk before calving and 20.1 min for the 3 wk after calving were used to calculate the number of meals and the size of each meal. 7.2.3. Standing behaviour Standing behaviour was collected using modified dataloggers (Gemini Dataloggers Ltd., Chichester, UK) validated by O?Driscoll et al. (2008). Dataloggers were modified with a mercury switch that recorded leg orientation (horizontal or vertical) at 1 min intervals. These data were used to calculate daily standing time as well as the frequency of transitions from standing to lying positions (i.e., ?standing bouts?). A standing bout began when a cow stood up and ended when the cow returned to lying. The average duration of standing bouts per day was calculated as a ratio of minutes standing to the number of standing bouts per day. Dataloggers were attached to the hind leg of cows upon entry into the pre-calving pen and were removed 21 d after calving; data were offloaded weekly when cows were in the milking parlor. The location of standing was recorded from video using 10 min scan sampling during 4 d in the 2 wk before calving. Scan sampling has been validated as a reliable measure of daily standing behaviour (validated by Mitlohner et al., 2001). Standing locations included: standing while feeding, standing in the feed alley (not feeding), standing in the alley adjacent to the lying stalls, standing with 2 front feet in the stall (i.e., ?perching?) and standing with 4 feet in the stall.  7.2.4. Social behaviour Displacements at the feed bins were recorded using video from 2 cameras mounted 6 m above the feed bins in each of the pre-calving pens. Displacements were recorded for 90 min following the twice-daily delivery of fresh feed for the same 4 d used to measure standing location 2 wk before calving. Data collected from the morning and afternoon feeding were summed to create one value per cow and day, and averaged to create one value per cow.  A displacement was recorded when a butt or a push from the ?Actor? resulted in the complete withdrawal of another cow?s head (the ?Reactor?) from the feed bin. An index of displacements was created, as described by Galindo and Broom (2000), calculating the proportion of displacements a cow instigates relative to her overall displacements in the group. This index was used to categorize animals as low ranking (index < 0.4), middle ranking (index > 0.4 and < 0.6) or high ranking (index > 0.6):  81      7.2.5. Clinical examination of hooves  Hooves were examined approximately 2 wk pre-calving, 3 wk post-calving and at monthly intervals thereafter (i.e. wk 7, 11 and 15). At each examination, hooves were pared by a trained hoof trimmer to reveal a clean horn surface. The presence of claw horn lesions (sole hemorrhages, sole ulcers, white line hemorrhages), digital dermatitis and interdigital hyperplasia were recorded for front and rear hooves using a foot map that divided each claw into 6 zones (Greenough and Vermunt, 1991). Claw horn lesions were scored for location and severity by a trained observer using a validated 6 point scoring system modified from Leach et al. (1998): 1 = diffuse yellow discoloration, 2 = diffuse red discoloration, 3 = deep, dense red lesion, 4 = port-colored lesion, 5 = red color, raw and bleeding lesion, and 6 = ulcer, corium exposed.  7.2.6. Cow participation in the study Cows were retrospectively assigned to a lesion category based on the most severe lesion on the sole (i.e., zone 4 or 5) diagnosed between wk 7 and wk 15 after calving. As a result of the association between disease and feeding behaviour during the transition period (e.g., see Huzzey et al., 2007; Goldhawk et al., 2009), animals diagnosed with any clinical diseases including ketosis, retained placenta or metritis were excluded from the study (n = 13). Moreover, standing behaviour is known to be directly affected by lameness (Chapinal et al., 2009), thus, cows diagnosed with a sole ulcer or interdigital hyperplasia during transition were removed from the final dataset (n = 6). Ten cows that were diagnosed with a moderate lesion (score 1, 2 or 3) between wk 7 and 15 were also removed from the analysis. Together this resulted in the exclusion of 29 cows. Of the remaining 26 cows, 13 had no visible lesions, 5 had sole ulcers, and 7 had ?severe? sole hemorrhages with a score of 4 or 5. 7.2.7. Statistical analysis Cow was considered the experimental unit (n = 26). In a preliminary analysis, an ANOVA using SAS software (PROC GLM, version 9.1; SAS institute, 2003) was used to determine if cows that were diagnosed with severe sole hemorrhages differed from cows that were diagnosed with sole ulcers. We found no differences for any behaviour except in the number and duration of standing bouts during the 2 wk before calving, thus, chose to group cows with severe sole hemorrhages and sole ulcers together as having lesions. These 13 lesion cows were compared to the 13 cows with no lesions.  Displacement Index     =                    Number of displacements a cow instigates (actor) Total number of displacements she was involved in (actor + reactor) 82  To detect differences in feeding and standing behaviour between cows with and without lesions, data were summarized into the following periods: 2 wk before calving (d -14 to d -1, ?wk -2?), the first 24 h after calving (?24 h?), the first wk after calving (d 2 to d 7, ?wk 1?) and the 2 to 3 wk after calving (d 8 to d 21, ?wk 2?). An ANOVA (PROC GLM) was used to detect differences between cows with an without lesions for daily DMI, feeding time, feeding rate, the number and duration of meals, standing time, number of standing bouts and average duration of standing bouts for each period. The model for each variable included parity as a covariate and lesion category (lesion or no lesion) as a fixed effect. The LSMEANS statement was used to calculate least-square means and pooled standard errors for presentation. Differences between groups in the location of standing were tested using the same model during wk -2.  It was clear from the above analysis that there were differences in feeding and standing behaviour in the week before calving. To detect any diurnal differences in behaviour during this pre-calving period, DMI, feeding rate and standing time data were summarized into hourly periods. A repeated measures ANOVA (PROC GLM) was used to assess differences in diurnal patterns between cows that were diagnosed with lesions later in lactation and those without lesions. The model included parity as a covariate, hour as a repeated measure, lesion category (lesion or no lesion) and a lesion category ? hour interaction. Autoregressive covariance structure was used to model the correlation between data points. Differences between cows with and without lesions in the total number of displacements each cow was engaged in and her displacement index during wk -2 before calving were analyzed using an ANOVA (PROC GLM). The model included parity as a covariate and lesion category (lesion or no lesion) as a fixed effect. Differences in the proportion of cows that became lame across the three displacement index categories were tested using a Fisher?s Exact test. Multivariate logistic regression (PROC LOGISTIC) was used to determine the predictive value of the variables that were significantly different between treatments. Univariate logistic models were first used to fit one variable at a time. Correlation between variables was assessed (PROC CORR) to avoid including highly related variables in the same model. Receiver operating characteristics (ROC) curves were plotted and the area under the curve (AUC) was used to compare the accuracy of the different models in discriminating cows that were diagnosed with lesions from those that were not. Multivariate models were constructed using a forward stepwise procedure based on a selection criterion of P < 0.10.  Differences between groups in production measurements (milk production, body weight and parity) were tested using ANOVA (PROC GLM). 83  7.3. Results 7.3.1. Milk production There were no differences between groups in BW before calving (no lesion vs. lesion: 742 ? 16 vs. 717 ? 15.3 kg, P = 0.27), BW after calving (no lesion vs. lesion: 694 ? 14 vs. 679 ? 14, P = 0.45), 21 d milk production (no lesion vs. lesion: 41.9 ? vs. 42.9 ? kg/d, P = 0.67), or parity (no lesion vs. lesion: 2.8 ? 0.3 vs. 3.0 ? 0.3 parities, P = 0.62). 7.3.2. Standing behaviour Cows diagnosed with lesions in mid-lactation stood for approximately 2 h (121 min) longer per day during the 2 wk before calving than cows without lesions (Figure 7.1; P = 0.007). This difference was also found for the 24 h after calving (P = 0.001), but not in wk 1 or wk 2 after calving. The frequency of standing bouts during the 24 h after calving was lower for cows that were diagnosed with lesions (P = 0.04), and the duration of each standing bout was longer for these cows (P = 0.005). No differences in standing bouts or bout duration were found for any other period.  Figure 7.2 shows the time per day that cows in each group stood in different locations of the pen. Perching time during the 2 wk before calving averaged 94 min/d higher for cows that were diagnosed with lesions than for cows without lesions (P = 0.007). There was a tendency for cows that were diagnosed with lesions to stand longer in the feed alley (not feeding) compared to cows without lesions (P = 0.06).     84   * P ? 0.05, ** P ? 0.01, *** P ? 0.001   Figure 7.1. Standing behaviour of cows with and without hoof lesions LS means ? SE of daily (A) standing time, (B) number of standing bouts and (C) standing bout duration of cows that were diagnosed with severe sole hemorrhages and sole ulcers (n = 13) and those without lesions (n = 13) from 2 wk before to 3 wk after calving.    85     Figure 7.2. Standing location of cows with and without hoof lesions Standing location of cows that were diagnosed with severe sole hemorrhages and sole ulcers (n = 13) and those without lesions (n = 13) during the 2 wk before calving. Locations included: standing while feeding (?feeding?), standing in the feed alley and not feeding (?feed alley?), standing in the alley adjacent to the lying stalls (?alley?), standing with 4 feet in the stall (?4 ft in stall?) standing with 2 front feet in the stall (?2 ft in stall?). 86  7.3.3. Feeding behaviour Cows that were diagnosed with lesions in mid-lactation ate at a faster rate and had more frequent but slightly smaller meals during the 2 wk before calving compared to cows that did not develop lesions (Table 7.2). In the 24 h after calving, cows with lesions consumed more feed in more frequent meals and tended to eat at a faster rate than did cows without lesions. During wk 1, cows with lesions continued to consume more feed in larger, rather than more frequent, meals compared to cows without lesions. By wk 2 after caving there were no detectable differences in feeding behaviour.  Table 7.2. Feeding behaviour of cows with and without hoof lesions LS means and SEM for the feeding behaviour of cows that were diagnosed with severe sole hemorrhages and sole ulcers (n = 13) and those without lesions (n = 13) from 2 wk before to 3 wk after calving.       No Lesion Lesion SEM P DMI (kg/d)      wk -2 16.8 17.0 0.6 0.82  24 h 12.3 17.9 0.9 <0.001  wk 1 15.8 18.5 0.6 0.01  wk 2 19.8 21.3 0.7 0.15 Feeding time (min/d)     wk -2 228 213 9 0.26  24 h 110 139 12 0.11  wk 1 158 172 8 0.20  wk 2 208 212 11 0.84 Feeding rate (g/min)     wk -2 77 86 3 0.05  24 h 121 137 7 0.09  wk 1 106 113 4 0.29  wk 2 102 107 5 0.50 Meals (no./d)      wk -2 6.3 7.1 0.2 0.02  24 h 6.0 7.6 0.5 0.05  wk 1 8.1 7.8 0.3 0.60  wk 2 7.4 7.2 0.2 0.60 Meal size (kg/meal)     wk -2 2.8 2.5 0.1 0.10  24 h 2.2 2.5 0.4 0.42  wk 1 2.1 2.6 0.1 0.01   wk 2 2.8 3.0 0.1 0.10 87  7.3.4. Diurnal pattern There was no difference in the diurnal pattern of DMI during the 2 wk before calving between cows with and without lesions (P = 0.79; Figure 7.3). There was an hour ? lesion interaction for feeding rate and standing time (feeding rate: P = 0.002; standing time: P = 0.001), whereby cows diagnosed with lesions spent more time standing in the early morning and between peak feeding times, and ate faster at peak feeding times compared to cows without lesions.   Figure 7.3. Pattern of daily feed intake for cows with and without hoof lesions Diurnal pattern of (A) DMI, (B) feeding rate and (C) standing time of cows that were diagnosed with severe sole hemorrhages and sole ulcers (n = 13) and those without lesions (n = 13) during the 2 wk before calving.  A C B 88   7.3.5. Social behaviour No differences were found during the 2 wk before calving between cows with lesions and without lesions in the total number of displacements at the feed bins or in the average displacement index. When cows were categorized as low-, middle- and high-ranking, we detected a difference in the proportion of cows with and without lesions that fell into each category (P = 0.02). Cows without lesions were more likely to be middle-ranking (displacement index > 0.40 and < 0.60) than cows with lesions. Of the 13 cows without lesions after calving, 10 were middle-ranking, 1 was low ranking (< 0.40) and 2 were high-ranking (> 0.60). In contrast, 3 cows that were diagnosed with lesions after calving fell into the middle-ranking category, 4 were low-ranking and 6 were high-ranking (Figure 7.4).    Figure 7.4. Social behaviour of cows with and without hoof lesions Number of cows with and without severe sole hemorrhages and sole ulcers (n = 13 per group) in the low (displacement index < 0.4), middle (> 0.4 and < 0.6) and high-ranking (> 0.6) categories. This index is calculated as a ratio of the number of displacement instigations to the total number of displacements (following Galindo and Broom, 2000) during the 2 wk before calving.  7.3.6. Logistic regression Table 7.3 shows the odds ratio (OR) with 95% confidence intervals (CI), area under the ROC curve (AUC), R2 and PWald of the following variables: wk -2 standing time, perching time, feeding rate and meals, 24 h standing time, DMI, standing bout duration and meals, and wk 1 89  DMI and meal size. Standing time during wk -2 was correlated with perching time (R2 = 0.75) and 24 h standing time (R2 = 0.64), 24 h standing time was correlated with 24 standing bout duration (R2 = 0.69), and 24 h DMI was correlated with wk 1 DMI (R2 = 0.54). These correlated variables were not combined in the same multivariate model. Moreover, meal variables were highly correlated with intake, so these variables were not included in complete logistic models. The variable with the highest AUC was 24 h DMI (0.86). The AUC increased from 0.86 to 0.87 after the addition of wk -2 daily standing time (PLikelihood Ratio = 0.0007) and to 0.91 after the addition of wk -2 feeding rate (PLikelihood Ratio = 0.0007). When wk -2 perching time was substituted for standing time, the AUC of the model was 0.95 (PLikelihood Ratio = 0.0003). When wk 1 DMI was substituted for 24 h DMI (wk 1 DMI, wk -2 standing time and wk -2 feeding rate), the AUC was 0.91 (PLikelihood Ratio = 0.0009); and when wk -2 perching time was substituted for standing time in this model the AUC was 0.90 (PLikelihood Ratio = 0.008).   Table 7.3. Behaviours predictive of hoof lesions Results of logistic regression for a cow diagnosed with a lesion (i.e. severe sole hemorrhage or sole ulcer; n = 13) or not (n = 13) using measures of feeding and standing behaviours as predictors   OR1 CI2 AUC3 R2 PWald wk -2 standing time4 1.47 1.00 ? 1.02 0.81 0.27 0.02 wk -2 perching time5 1.49 1.00 ? 1.03 0.81 0.24 0.02 wk -2 feeding rate 1.11 0.99 ? 1.24 0.71 0.18 0.07 wk -2 meals 3.17 1.03 ? 9.74 0.73 0.18 0.04 24 h standing time4 1.25 1.00 ? 1.01 0.78 0.30 0.02 24 h DMI 1.51 1.12 ? 2.04 0.86 0.37 0.007 24 h bout duration5 1.22 1.05 ? 2.14 0.81 0.27 0.03 24 h meals 1.56 0.98 ? 2.56 0.73 0.16 0.06 wk 1 DMI 1.62 1.07 ? 2.46 0.78 0.24 0.02 wk 1 meal size 2.03 2.23 ? 537 0.82 0.31 0.01  1 Odds Ratio 2 95% Confidence Interval 3 Area under the receiver operating characteristics curve  4 OR adjusted to a 30-min increase 5 OR adjusted to a 10-min increase   90  7.4. Discussion Our results provide the first evidence that lesions diagnosed 7 to 15 wk after calving were associated with changes in behaviour during the transition period. What remains unclear is whether these behavioural differences contribute to the development of the lesions, or are a consequence of an existing ailment. Regardless, these differences in behaviour at transition can serve as early indicators of lesions that can only be diagnosed much later in lactation.  Cows that were diagnosed with lesions stood for longer than cows without lesions during the 2 wk before calving. Our results support those of Chapinal et al.(2009), in that cows diagnosed with sole ulcers in mid-lactation began to increase their standing time during the 2 wk before calving. High standing times during transition likely increase the risk for lesions, as there is evidence that prolonged standing on a hard surface can increase pressure on the third phalanx thereby compressing the horn-producing tissue in the corium and compounding parturition-related changes in the hoof (Lischer et al., 2002). A higher daily standing time during the weeks before calving for cows that develop lesion in mid-lactation was driven by an increase in time spent perching with 2 front feet in the stall. The literature on perching has considered the behaviour both a cause and an effect of lameness (Leonard et al., 1994; Galindo et al., 2000). An increase in this behaviour could also be driven by the design of the lying stall, as cows spend less time perching when the stalls are designed to allow cows to stand fully inside the stall (Fregonesi et al., 2009). Cows diagnosed with lesions in mid-lactation consumed more feed in more frequent meals during the 24 h and 1 wk after calving. Previous work has shown that lame cows reduce DMI in the days before diagnosis (Gonzalez et al., 2008), likely as a result of pain associated with the condition. Variation in feed intake in early lactation is not likely a consequence of the developing lesions; more likely this is a response to the introduction of a higher-concentrate diet and a new social group. The current data do not allow us to determine if the increased intake contributed to acidosis-related changes in the hoof, but cows with a high intake after calving may have an exaggerated drop in pH (Fairfield et al., 2007). Future work is encouraged to identify the possible mechanisms by which the development of lesions may be affected by feed consumption immediately after calving. Cows diagnosed with a lesion in mid-lactation consumed feed more rapidly during the 2 wk before and 24 h after calving, and this difference was most evident at peak feeding time. A high rate of feed intake has been associated with more extensive sorting of the feed for small particles (DeVries et al., 2007) as well as greater social pressure at the feed bunk (Nielsen, 1999; 91  Proudfoot et al., 2009). There is likely a combination of factors influencing feeding rate in transition cows and research addressing the causes and consequences of increased feeding rate during transition is encouraged. The displacement index was used as an indicator of social rank, as it reflects how often individual cows displace other cows versus being displaced. Cows diagnosed with a lesion in mid-lactation were more likely to be either high or low ranking compared to cows without lesions that were primarily categorized as middle ranking. High social rank has been associated with older age, a factor that is also associated with lameness (Espejo et al., 2006); however, with a small sample of cows in each age group we were unable to provide a powerful test of age effects in our study. Low social rank has been associated with lameness and high standing time (Galindo et al., 2000; Galindo and Broom, 2000).  Implications of this study follow the suggestions of Knott et al. (2007) and Cook and Nordlund (2009) that good husbandry during the transition period helps curb lameness in mid-lactation. Results from the multivariate analysis suggest that a combination of behaviours, particularly standing time before calving and consuming a large amount of feed the day after calving, can be used to predict which cows may develop lesions later in lactation. Webster (2002) suggested that providing a comfortable standing surface during transition can mitigate external influences on the hoof; results provided here show that transition cow behaviour, rather than the environment alone, can contribute to sole hemorrhages and sole ulcers diagnosed in mid-lactation. Further research is encouraged to determine if transition management practices aimed to limit excess standing and perching time (i.e., by providing a comfortable lying or standing surface), and better regulation of DMI directly after calving, can reduce the incidence of claw horn lesions in the weeks after calving.  7.5. Conclusions Compared to cows without lesions, cows diagnosed with lesions in mid-lactation spent more time standing, particularly perching with their 2 front feet in the lying stall, during the 2 wk before calving. Cows with lesions also ate faster than cows without lesions during the 2 wk before calving, and consumed more feed in more frequent meals during the 24 h after calving. Transition cow behaviour can serve as early indicators of sole hemorrhages and sole ulcers that become visible several weeks after calving. 92  8. General Conclusions and Discussion   8.1. Summary of thesis This thesis opened with a review of the existing literature concerning transition cow behaviour, health and management, as well as the relationship between social behaviour and disease in farm animals. Gaps in the literature were identified, and some of these gaps were addressed in the subsequent research chapters.  The studies presented in Chapters 3, 4 and 5 provided evidence that management practices can affect behaviour around calving, and that cows preferred to be in a secluded environment during labour. Chapter 3 reported that when cows were moved during a later stage of labour they had prolonged stage II labour and spent more time standing in the hour before calving compared to those moved earlier. Chapters 4 and 5 reported that, under certain conditions, cows preferred a secluded area to calve.  The studies presented in Chapters 5, 6 and 7 showed that behaviours can be useful indicators of three major welfare concerns in dairy cattle: infectious disease, dystocia and lameness. In Chapter 5, cows with signs of infectious disease immediately after calving ate less, spent more time lying and spent more time in a secluded area of a pen compared to healthy cows. In Chapter 6, cows with dystocia ate less before calving, and were more restless compared to those with eutocia. In Chapter 7, cows that became lame after calving spent more time standing with their two front feet in the stall (?perching?) during transition and consumed more feed at a faster rate on the first day after calving.  Collectively, these findings provide evidence that knowledge of behaviour can be useful for designing suitable environments and addressing welfare concerns of transition dairy cows. In the following sections, I will discuss the findings of my studies within the broader literature, will then highlight the strengths and limitations of the research, and will end with recommendations for future research.  8.2. Using knowledge of behaviour to improve housing and management  8.2.1. Discussion within the broader literature Results from Chapters 3, 4 and 5 provide evidence that knowledge of behaviour during the parturient period can help improve management and housing practices. One method to inform good practice is to determine the effect of management practices on cow welfare. To claim that a 93  management practice reduces welfare, it is necessary to include measurements related to her natural behaviour, health status or affective state (Fraser et al., 1997). In the case of Chapter 3, moving cows during a sensitive period just before stage II labour caused a disturbance in normal behaviour and prolonged labour. Together, these results provide some evidence that moving cows during this period may impact her welfare, depending on the aspect of welfare that is considered. Prolonged labour may be associated with increased pain, a negative affective state (Mainau and Manteca, 2011), as well as increased inflammation (as shown in Chapter 4). Moreover, this movement also limits her ability to perform a natural behaviour (i.e., recumbency in the hour before calving, Schuenemann et al., 2011).  A second way to use behaviour to help improve practice is by gaining a better understanding of environments that cows prefer and are motivated to access. Chapters 4 and 5 provided the first evidence that indoor-housed cows preferred a secluded environment to calve, particularly when there was high activity in the barn during the daytime. The choice tests used were long in duration, as it would be impractical (but not impossible) to do an instantaneous choice test during parturition. Our methodology was similar to those used in previous choice tests (e.g., Tucker et al., 2003; Fregonesi et al., 2009), which compared two housing options that a producer may realistically use on their farm. The shelter in Chapter 4 was large, but there are variations of this idea that may be more practical for use on commercial farms, such as the option presented in Chapter 5. Together, the results of Chapters 3, 4 and 5 have implications for both scientists and producers. For scientists, they provide the first steps in what will hopefully be a series of experiments used to determine suitable housing and management of cows during parturition (see below for suggestions for future research). For producers, these results suggest that disturbances, including movements and high human activity, during labour should be limited. Producers often move cows to individual pens and keep them in high traffic areas to facilitate observation, but these practices may be doing more harm than good. More research is needed, but the results presented in this thesis give some justification for providing a less disturbed environment during labour (e.g., moving cows well in advance of labour rather than ?just-in-time? when clear calving signs are present).  8.2.2. Strengths and limitations  The thesis chapters that focused on the relationship between behaviour, housing and management are novel, as very few studies have examined the period in the hours prior to and during calving. Although Chapter 3 is particularly applicable to Denmark where the research was 94  conducted, many Canadian dairy farms also use individual maternity pens. A main strength of the preference tests is the use of previous knowledge about cow behaviour in a more natural setting during parturition. There was evidence from both wild ungulates and domesticated dairy cows housed on pasture that given the opportunity, cows will leave the herd and find a place to hide as calving approaches (Lidfors et al., 1994; Barbknecht et al., 2011). Although the design of our shelter may also be a limitation (see below), this background research gave me more confidence that the secluded environment would be preferred. There are two main limitations to the research presented in this thesis regarding management and behaviour: 1) small sample size (Chapter 3), and 2) the accuracy and usefulness of preference testing as indicators of cow welfare (Chapters 4 and 5). In Chapter 3, cows were meant to be moved into individual pens at standard times relative to labour; however, there was a high variation in the time cows spent in the maternity pens. Although cows were categorized based on stage of labour, a larger sample size would have allowed for more power to test changes in behaviour as well as health outcomes. For instance, the effect of prolonged labour on uterine health was measured but could not be included due to the low occurrence of illness in the sample. The insights and associated limitations that arise from preference tests have been discussed in the literature (e.g., see reviews by Dawkins, 1983; Duncan, 1992; Fraser, 1993; Kirkden and Pajor, 2006). One common criticism of preference tests is that researchers control the alternatives offered to the animal, thus running the risk that neither option is suitable. In this thesis, cows were given two choices for a calving location: a secluded area and an open area exposed to other cows and human activity. It is possible that neither option was suitable ? the secluded area may have been chosen because this was less aversive than the open area. In other words, the secluded environment could have been the better of two bad options. I had predicted that the secluded environment would be preferable to cows based on evidence from wild ungulates and domesticated cows kept on pasture (Lidfors et al., 1994; Barbknecht et al., 2011). However, the design of our shelter was different from the tree and brush cover identified as preferred calving sites in those studies. Perhaps a shelter with the opportunity to escape on both sides would be more preferred, or a shelter with some visual access to the group (e.g., as was found with chickens, Newberry and Shackleton, 1997).  A second criticism of preference tests is that they may provide clues to the short-term preference of animals, but may not reflect long-term benefits. Although cows prefer a secluded area to an open area during calving, they may benefit, or not benefit, equally with either option. Lastly, the preference tests used in Chapters 4 and 5 provide no information about the strength of 95  the motivation to hide during calving. Determining the strength of the motivation to hide would provide more insight into the importance of this behaviour to a parturient cow.  8.2.3. Recommendations for future research The take-home message from Chapters 3, 4 and 5 is that cows prefer a secluded environment during labour, and disturbing them while in labour can reduce their ability to calve normally. Follow-up research is encouraged to determine the health benefits to both limited disturbance and seclusion during labour. Moreover, a motivation test is recommended to determine the value of a secluded environment for parturient cows. As a follow-up to Chapter 3, I encourage research to determine the potential long-term health effects of prolonged labour caused by management practices such as pen-movements. Moreover, research is needed to determine the effect of housing cows in group maternity pens on behaviour, labour progress and health. I would predict that cows housed in groups have little or no opportunity to seclude themselves and may experience disturbance from other cows (as was found with ewes, Arnould et al., 1991). This could result in similar changes in behaviour and labour progress as was found in Chapter 3 when cows were disturbed by being moved to a new pen, and may increase the risk of uterine disease after calving. For Chapters 4 and 5, I encourage caution when interpreting the results of the preference tests as being reflective of good cow welfare and best practice; however, this research provides a necessary first step. A preference test combined with other measurements of welfare (e.g., health outcomes), or done in a way that measured the strength of the motivation or preference (i.e., motivation test) would provide additional evidence in support of providing a more secluded area for calving. Research aimed at measuring the effect of this preferred, secluded environment during parturition on cow behaviour and health is encouraged. A good example of relating a preferred environment to health benefits comes from work using mid-lactation cows and free stall surface: cows prefer to lie on soft, deep-bedded surfaces compared to harder mattresses (Tucker and Weary, 2004), and these softer surfaces reduce the risk of hock lesions (Barrientos et al., 2013). I predict that a more disturbed environment (e.g., a group pen with cows entering and exiting frequently, or an individual pen in a very high traffic area) would cause disturbance to cows during labour, changing behaviour and potentially influencing labour progress and subsequent uterine health. I also encourage future work to determine the value of a secluded area to cows during parturition. A motivation test would provide insight into the value that parturient cows place on 96  this behaviour. Some motivation tests have been done using mid-lactation cows, but developing an appropriate ?price? for a parturient cow to pay for access to seclusion may take some careful thought. For instance, using a weighted gate may not be the best task, as labour causes visceral pain that might influence the motivation to push through a heavy gate (Mainau and Manteca, 2011). Pushing a lever has been used to determine motivation for nest-building and locomotion in farrowing sows (Hutson and Haskell, 1990; Haskell et al., 1997), so this may also be a good task for parturient cows. However, Haskell et al. (1997) discovered that aspects related to the operant task may have interfered with the motivation of sows to express locomotor behaviour on the day of farrowing. Authors speculated that this interference may have partially been due to the ?unnaturalness? of pressing a lever for farrowing sows, and that a more natural task (such as walking on a treadmill) may have been a preferred task. Walking a long distance on a treadmill, pasture or in a barn may also be a more natural task for parturient dairy cows. In an experiment using cows housed on pasture, Lidfors et al. (1994) found that cows will travel as much as 500 m away from the herd to find a birth-site. Thus, a suitable follow-up to Chapter 4 and 5 would be for cows to be trained to walk progressively longer distances to find a secluded environment. The motivation to hide on the day of calving (e.g., maximum price paid, or maximum distance walked) could then be compared to previous days to determine the relative strength of the motivation to perform this behaviour at calving.   8.3. Using behaviour as an indicator of poor health 8.3.1. Discussion within the broader literature Results from Chapters 5, 6 and 7 provide additional evidence that behaviour can be a useful tool to identify both ill animals and those at-risk of becoming ill. Although sickness behaviours are becoming more common in the dairy cow literature (e.g., Siivonen et al., 2011; Fogsgaard et al., 2012; Cyples et al., 2012), to my knowledge, the research presented here was the first to show that dairy cows change their social behaviour when ill. This was not surprising, as a reduction in social exploration is a classical sign of infectious disease in other mammals, such as laboratory animals (e.g., Crestani et al., 1991; Bluth? et al., 1999). Additional research in this area is still needed, as I believe a better understanding of the relationship between illness and social behaviour in cattle will both contribute to the broader literature regarding sickness behaviours in mammals, and provide producers a more comprehensive picture of early stages of illness.  The work described in this thesis also provided evidence that behaviour during transition can predict poor health in the form of dystocia and lameness. Since Chapter 6 was published, 97  there have been a number of studies assessing the effect of dystocia on behaviour and labour progress (e.g., Miedema et al., 2011; Barrier et al., 2012), as well as pain behaviours after calving and behaviour of the calf (Barrier et al., 2012). Results of these studies suggest that cows with dystocia have longer stage II labour (Schuenemann et al., 2011; Barrier et al., 2012) and may be experiencing more pain after calving (i.e., spent more time self-grooming, Barrier et al., 2012). Calves from cows with dystocia show higher latencies to stand, walk and reach the udder than unassisted calves (Barrier et al., 2012). Together with the results of my thesis, these studies provide evidence that dystocia is an important welfare concern for dairy cows, and that more research identifying dystocia at early stages should continue. Two recent studies have been published using methodologies that were very similar to those in Chapter 6, although some of the outcomes differed (Miedema et al., 2011; Barrier et al., 2012). Similar to our results, Barrier et al. (2012) found that cows with dystocia were more restless compared to cows without dystocia before calving. In addition, these authors discovered that some cows with dystocia also raised their tail earlier than cows without dystocia (those with mal-positioned calves). In contrast, Miedema et al. (2011) did not find any effect of dystocia on feeding and standing behaviour before calving.  The difference in the outcomes of these studies may be attributed to differences in the definition of dystocia. In fact, all three studies used different definitions to categorize cows into the ?dystocia? or ?calving difficulty? level. Our definition was based entirely on producer records, and only included cases where the producer categorized the calving difficulty as severe (e.g., a mal-positioned calf and/or a very difficult pull was necessary to extract the calf). Barrier et al. (2012) further differentiated cows in this category by separating those with mal-positioned calves from those with very difficult calvings without mal-positions. Miedema et al. (2011) included cows that had been assisted using a calving jack (for > one minute). These authors suggest that the difference between their results and ours was likely due to the inclusion of cows that may not have needed assistance (or would have been an ?easy? assistance in our category), and thus may not have been experiencing the same severity of dystocia compared to the other two studies. A greater consensus about the definition of dystocia is still needed in the literature. This definition could be an integration of those outlined here, and include both assessments from the producer and/or veterinarian, as well as position of the calf and use of a jack.  As with the dystocia chapter, a couple of similar studies have taken place since Chapter 7 was published (e.g., Calderon and Cook, 2011; Dippel et al., 2011). One study with similar methodology (Dippel et al., 2011) found that cows at-risk for hoof lesions (including hemorrhages 98  and white line lesions) spent more time ?perching? with their two front feet in the stall during the 2 wk post-calving compared to those without lameness. Although differences in perching time were noted in the 2 wk pre-calving in our work, it was not measured post-calving, so it is likely that these at-risk cows are displaying this behaviour over both periods. This study also recorded feeding behaviour, but unlike Chapter 7, found no differences between groups. A likely explanation for this difference is the time period when the data were collected, as well as the measurements used. Chapter 7 reports the biggest differences in feeding behaviour during the 24 h and 1 wk after calving, whereas Dippel et al. (2011) failed to differentiate between weeks, summarizing the 3 wk period after calving as a whole. In addition, Chapter 7 only reported differences in feed intake and feeding rate, but not in feeding time (i.e., cows spent the same amount of time eating, but ate more during that period), whereas Dippel et al. (2011) were only able to measure feeding time.  Together, Chapters 5, 6 and 7 were able to relate behaviour to illness, but did so in different ways. In Chapter 5, cows showed signs of illness, including fever, concurrent with the behavioural changes, which likely meant that these changes were an effect of the illness. In the studies of dystocia and lameness, this relationship was less clear. In the case of dystocia, I argued that the behavioural changes were likely an early sign of a difficult calving caused by something else related to dystocia (i.e., cows were eating less and were more restless because they already felt discomfort due to a mal-positioned calf or prolonged labour). Lastly, Chapter 7 reported that behaviours predictive of lameness occurred far in advance of diagnosis, and were thus more likely a cause, rather than an effect of the disease (although this could be debated, as discussed in the next section).  In all three cases, behaviour could be a useful tool to help indicate illness. However, it is still important to distinguish between those behaviours that were caused by, rather than an effect of the illness, to develop the best strategy for how to treat those animals. For instance, cows showing behavioural changes alongside other signs of illness would likely be treated, or examined by a veterinarian. In the case where a behaviour is likely tied to an illness, but clear, clinical signs of the illness are not yet present (as was the case of dystocia), these cows may require closer observation or perhaps a more comfortable environment rather than treatment. Lastly, for those cases where behavioural changes likely cause illness (such could have been the case with lameness), there is much more room for prevention. For instance, if high perching time during transition causes lameness, then producers could attempt to limit this behaviour by changing management, such as removing the neck-rail in pre- and post- calving pens (Bernardi et al., 2009).  99  8.3.2. Strengths and limitations of the thesis research Research presented in Chapters 5, 6 and 7 provides some of the first evidence that behaviour around the time of parturition can be useful as an early indicator of infectious disease, dystocia and lameness. Strengths of both of Chapters 6 and 7 were the careful selection of animals for our retrospective analyses. A usual dilemma with transition cow research is the high number of cows that have multiple afflictions. For instance, many cows have both metabolic and infectious disease, in addition to dystocia and lameness; these diseases have been shown to alter behaviour, potentially confounding results (e.g., Huzzey et al., 2007; Goldhawk et al., 2009). To avoid this problem, a broad spectrum of health measures were collected and used these to eliminate cows that had multiple conditions. By removing these animals, or balancing the groups for certain conditions (i.e., sub-clinical metritis in the Chapter 6), I were more confident that the behavioural changes discovered were specifically related to dystocia and lameness and not artifacts of other transition issues. There are three main limitations to this research on behavioural indicators of health: 1) sample size (Chapter 5), 2) determining the cause and effect relationship between behavioural changes and poor health (Chapter 6 and 7), and 3) neglecting individual differences and social behaviours as early indicator of poor health (Chapter 6 and 7). To address my hypothesis in the second part of Chapter 5, cows were included in the study if they had a variety of infectious illnesses that were diagnosed between 2 and 3 d after calving. Although this provided important information about the behaviour of cows with infectious illnesses, I was unable to determine the exact time-course of these changes as cows were diagnosed on different days relative to calving. A greater sample size would have allowed for more details about the day of diagnoses, the disease type and the severity of the disease. The studies described in Chapters 6 and 7 measured the behaviour of cows with and without dystocia and lameness. A limitation with this approach is the inability to determine the cause and effect relationship between the behaviours and the health conditions.  For example, cows with lameness spent more time perching in the stalls beginning months before they were diagnosed with claw horn lesions. However, in previous studies this behaviour has been shown to be both a risk-factor for lameness, as well as an effect of lameness (Leonard et al., 1994; Galindo et al., 2000). Results from Chapter 7 provide strong evidence that perching was a risk-factor for claw horn lesions because behaviour was measured months before cows were diagnosed as lame. However, the possibility that cows may have already been in a very early stage of the lesion 100  (which takes months to become visible, Leach et al., 1998) and perching for this reason instead cannot be ruled out. A third limitation is the lack of emphasis on social behaviours and individual differences in behaviour as early indicators of lameness and dystocia. Chapter 2 describes the potential relationship between social behaviour and illness, and how individual differences in social status or coping strategy may be useful as early predictors of poor health. The study in Chapter 7 measured the social status of cows at-risk for lameness later in lactation and found that cows that became lame were either more dominant or more subordinate compared to healthy cows. Social status is one way of measuring individual differences in behaviour and risk of poor health, but can be altered based on the composition of the pen. For instance, a young cow may be of high status in a pen of other young cows, but low status in a pen of older cows. Arguably, more stable measurements of individual differences in behaviour like coping style or personality may be better indicators of those animals at-risk for poor health (section 2.2.2.). A greater understanding of the influence of social behaviour as a risk factor for dystocia and lameness is still needed. 8.3.3. Recommendations for future research The take-home message from Chapters 5, 6 and 7 is that behaviour can be a used to detect cows with infectious disease, and those at-risk for dystocia and lameness. Further work is still needed to determine the promise of social behaviour as an early predictor of at-risk animals. Lastly, research is needed to take these concepts and apply them on-farm using technology that allows for the automatic collection and interpretation of behavioural changes. A main recommendation for future research is a greater emphasis on individual differences in behaviour as early indicators of illness. For instance, animals with different coping styles and personality traits are thought to be at-risk for different ailments (section 2.2.2.). Determining attributes that can identify cows with different traits is encouraged. For instance, coping style could be determined by presenting cows with a stressful situation and measuring their behavioural and neuroendocrine response. These measures could then be used to categorize cows as active or passive copers (i.e., to start, using behaviours and physiological responses outlined in Figure 2.2). Moreover, a personality trait that has been linked to infectious illness in laboratory animals is low sociability (Capitanio, 1999). This trait could also be measured in cows by determining the motivation of cows to access a familiar group (Gibbons et al., 2010). Once cows can be distinguished based on these traits, measurements about their influence on the risk of poor health after calving can be made. 101  Secondly, in order to transfer these results onto commercial farms, it is necessary to develop automatic methods of measuring behaviour. In the experiments described in this thesis I used methods of automatically recording feeding behaviour (e.g., Insentec, Marknesse, Holland) and standing behaviour (e.g., Ice Tag, IceRobotics, Edinburgh, UK), but these methods were designed for research and are likely not cost-effective for commercial farms. I highly encourage future research to validate more cost-effective tools to monitor behaviour. Future work is also encouraged to develop algorithms that identify when behaviours are deviating from normal to help producers identify at-risk animals. Once the animal is identified, researchers in collaboration with producers, veterinarians and hoof-trimmers would then need to decide the best course of action for the animal (e.g., early treatment if other signs of disease are present, or preventative measures if no signs of disease are present).  8.4. Final conclusions I have argued that research during the transition period holds great potential for improving the lives of dairy cattle. Cows face a number of challenges during this short period, and many succumb to one or multiple diseases and other painful conditions. These cows may also face environments that may be very different from their preferred surroundings. A major gap in the transition cow literature is a lack of understanding about cow behaviour. Research presented in this thesis provides examples of how behaviour can be used to identify a cow at-risk for poor health, and provide insight into environments that she prefers.  Based on this research, I have recommendations for both scientists and dairy producers. For scientists, I recommend further research addressing the optimal environment for parturient cows, the development of commercially-available tools to measure behaviour, and identification of at-risk cows drawing from concepts already used for other species. For producers, I recommend a visit to their calving pen and careful thought about how much disturbance the cows are experiencing and practical ways to minimize these disturbances. For both scientists and producers, I also recommend more conversations regarding creative alternatives to the traditional management practices during parturition, as I believe this is essential to develop practices that are beneficial to transition cow welfare, as well as sustainable and practical to implement on-farm.  102  Bibliography  Almeida, P. E., P. S. D. Weber, J. L. Burton and A. J. Zanella. 2008. Depressed DHEA and increased sickness response behaviors in lame dairy cows with inflammatory foot lesions. Domest. Anim. Endocrinol. 34:89-99.  Arakawa, H., K. Arakawa and T. Deak. 2010. Sickness-related odor communication signals as determinants of social behavior in rat: A role for inflammatory processes. Horm. Behav. 57:330-341.  Arnould, C., V. Piketty and F. L?vy. 1991. Behaviour of ewes at parturition toward amniotic fluids from sheep, cows and goats. Appl. Anim. Behav. Sci. 32:191-196.  Aubert, A., G. Goodall, R. Dantzer and G. Gheusi. 1997. Differential effects of lipopolysaccharide on pup retrieving and nest building in lactating mice. Brain Behav. Immun. 11:107-118.  Avitsur, R., J. Hunzeker and J. F. Sheridan. 2006. Role of early stress in the individual differences in host response to viral infection. Brain Behav. Immun. 20:339-348.  Barbknecht, A. E., W. S. Fairbanks, J. D. Rogerson, E. J. Maichak, B. M. Scurlock and L. L. Meadows. 2011. Elk parturition site selection at local and landscape scales. J. Wildl. Manage. 75:646-654.  Barrientos, A. K., N. Chapinal, D. M. Weary, E. Galo and M. A. G. von Keyserlingk. 2013. Herd-level risk factors for hock injuries in freestall-housed dairy cows in the northeastern united states and california. J. Dairy Sci. 96:3758-3765.  Barrier, A. C., M. J. Haskell, A. I. Macrae and C. M. Dwyer. 2012a. Parturition progress and behaviours in dairy cows with calving difficulty. Appl. Anim. Behav. Sci. 139:209-217.  Barrier, A. C., E. Ruelle, M. J. Haskell and C. M. Dwyer. 2012b. Effect of a difficult calving on the vigour of the calf, the onset of maternal behaviour, and some behavioural indicators of pain in the dam. Prev. Vet. Med. 103:248-256.  Barrier, A. C. and M. J. Haskell. 2011. Calving difficulty in dairy cows has a longer effect on saleable milk yield than on estimated milk production. J. Dairy Sci. 94:1804.  Bartolomucci, A., P. Palanzaa, L. Gaspanic, E. Limirolic, A. E. Paneraic, G. Ceresinid, M. D. Polib and S. Parmigiania. 2001. Social status in mice: Behavioral, endocrine and immune changes are context dependent. Physiol. Behav. 73:401-410.  103  Bartolomucci, A., P. Palanzaa, P. Sacerdotec, G. Ceresinid, A. Chirieleisona, A. E. Paneraic and S. Parmigiania. 2003. Individual housing induces altered immuno-endocrine responses to psychological stress in male mice. Psychoneuroendocrinolo. 28:540-558.  Bartolomucci, A. 2007. Social stress, immune functions and disease in rodents. Front. Neuroendocrinol. 28:28-49.  Bernardi, F., J. Fregonesi, C. Winckler, D. M. Veira, M. A. G. von Keyserlingk and D. M. Weary. 2009. The stall-design paradox: Neck rails increase lameness but improve udder and stall hygiene. J. Dairy Sci. 92:3074-3080.  Bicalho, R. C., V. S. Machado and L. S. Caixeta. 2009. Lameness in dairy cattle: A debilitating disease or a disease of debilitated cattle? A cross-sectional study of lameness prevalence and thickness of the digital cushion. J. Dairy Sci. 92:3175-3184.  Black, P. H. and L. D. Garbutt. 2002. Stress, inflammation and cardiovascular disease. J. Psychosom. Res. 52:1-23.  Blanchard, R. J., C. R. McKittrick and D. C. Blanchard. 2001. Animal models of social stress: Effects on behavior and brain neurochemical systems. Physiol. Behav. 73:261-271.  Bluth?, R., N. Castanon, F. Pousset, A. Bristow, C. Ball, J. Lestage, B. Michaud, K. W. Kelley and R. Dantzer. 1999. Central injection of IL-10 antagonizes the behavioural effects of lipopolysaccharide in rats. Psychoneuroendocrinolo. 24:301-311.  Bolhuis, J. E., H. K. Parmentier, W. G. P. Schouten, J. W. Schrama and V. M. Wiegant. 2003. Effects of housing and individual coping characteristics on immune responses of pigs. Physiol. Behav. 79:289-296.  Bontekoe, E. H. M., J. F. Blacquiere, C. Naaktgeboren, S. J. Dieleman and P. P. M. Willems. 1977. Influence of environmental disturbances on uterine motility during pregnancy and parturition in rabbit and sheep. Behav. Processes. 2:41-73. Borderas, T. F., A. M. de Passill? and J. Rushen. 2008. Behavior of dairy calves after a low dose of bacterial endotoxin. J. Anim. Sci. 86:2920-2927.  Brambell, F. W. R. 1965. Report of the Technical Committee to Enquire into the Welfare of Animals Kept Under Intensive Livestock Husbandry Systems. Her Majesty's Stationary Office, London, UK. Breivik, T., F. Sluyter, M. Hof and A. Cools. 2000. Differential susceptibility to periodontitis in genetically selected wistar rat lines that differ in their behavioral and endocrinological response to stressors. Behav. Genet. 30:123-130.  104  Brooks-Pollock, E. and M. Keeling. 2009. Herd size and bovine tuberculosis persistence in cattle farms in great britain. Prev. Vet. Med. 92:360-365.  Calderon, D. F. and N. B. Cook. 2011. The effect of lameness on the resting behavior and metabolic status of dairy cattle during the transition period in a freestall-housed dairy herd. J. Dairy Sci. 94:2883-2894.  Canadian Council on Animal Care. 2009. CCAC Guidelines on: The Care and use of Farm Animals in Research, Teaching and Testing. CCAC, Ottawa. Capitanio, J. P. 1999. Personality dimensions in adult male rhesus macaques: Prediction of behaviors across time and situation. Am. J. Primatol. 47:299-320.  Capitanio, J. P., K. Abel, S. P. Mendoza, S. A. Blozis, M. B. McChesney, S. W. Cole and W. A. Mason. 2008. Personality and serotonin transporter genotype interact with social context to affect immunity and viral set-point in simian immunodeficiency virus disease. Brain Behav. Immun. 22:676-689.  Carrier, J., S. Godden, J. Fetrow, S. Stewart and P. Rapnicki. 2006. Predictors of stillbirth for cows moved to calving pens when calving is imminent. Proc. 39th Ann. Amer. Assn. Bov. Pract. 158-159.  Chapinal, N., A. M. de Passill?, D. M. Weary, M. A. G. von Keyserlingk and J. Rushen. 2009. Using gait score, walking speed, and lying behavior to detect hoof lesions in dairy cows. J. Dairy Sci. 92:4365-4374.  Chapinal, N., D. M. Veira, D. M. Weary and M. A. G. von Keyserlingk. 2007. Technical note: Validation of a system for monitoring individual feeding and drinking behavior and intake in group-housed cattle. J. Dairy Sci. 90:5732-5736.  Charlton, G. L., S. M. Rutter, M. East and L. A. Sinclair. 2013. The motivation of dairy cows for access to pasture. J. Dairy Sci. 96:4387-4396.  Chase, I. D., C. Tovey, D. Spangler-Martin and M. Manfredonia. 2002. Individual differences versus social dynamics in the formation of animal dominance hierarchies. Proc. Natl. Acad. Sci. U. S. A. 99:5744.  Chida, Y., N. Sudo and C. Kubo. 2005. Social isolation stress exacerbates autoimmune disease in MRL/lpr mice. J. Neuroimmunol. 158:138-144.  Clericia, M. and G. M. Shearer. 1993. A TH1 TH2 switch is a critical step in the etiology of HIV infection. Immunol. Today. 14:107-111.  Coe, C. L. and G. R. Lubach. 2003. Critical periods of special health relevance for psychoneuroimmunology. Brain Behav. Immun. 17:3-12.  105  Coe, C. L., G. R. Lubach, M. L. Schneider, D. J. Dierschke and W. B. Ershler. 1992. Early rearing conditions alter immune responses in the developing infant primate. Pediatrics. 90:505-509.  Cohen, S., S. Line, S. B. Manuck, B. S. Rabin, E. R. Heise and J. R. Kaplan. 1997. Chronic social stress, social status, and susceptibility to upper respiratory infections in nonhuman primates. Psychosom. Med. 59:213.  Cohen, S., W. J. Doyle, R. Turner, C. M. Alper and D. P. Skoner. 2003. Sociability and susceptibility to the common cold. Psyhcol. Sci. 14:389-395.  Cohen, S., D. Janicki-Deverts and G. E. Miller. 2007. Psychological stress and disease. JAMA-J. Am. Med. Assoc. 298:1685-1687.  Cohn, D. W. H. and L. C. de Sa-Rocha. 2006. Differential effects of lipopolysaccharide in the social behavior of dominant and submissive mice. Physiol.  Behav. 87:932-937.  Cole, S. W., M. E. Kemeny, S. E. Taylor and B. R. Visscher. 1996. Elevated physical health risk among gay men who conceal their homosexual identity. Health Psychol. 15:243-251.  Cole, S. W., M. E. Kemeny, J. L. Fahey, J. A. Zack and B. D. Naliboff. 2003. Psychological risk factors for HIV pathogenesis: Mediation by the autonomic nervous system. Biol. Psychiatry. 54:1444-1456.  Cole, S. W., S. P. Mendoza and J. P. Capitanio. 2009. Social stress desensitizes lymphocytes to regulation by endogenous glucocorticoids: Insights from in vivo cell trafficking dynamics in rhesus macaques. Psychosom. Med. 71:591-597.  Cook, N. B. and K. V. Nordlund. 2009. The influence of the environment on dairy cow behavior, claw health and herd lameness dynamics. Vet. J. 179:360-369.  Cook, N. B. and K. V. Nordlund. 2004. Behavioral needs of the transition cow and considerations for special needs facility design. Vet. Clin. N. Am-Food A. 20:495-520.  Council for Agricultural Science and Technology (CAST). 2012. The Direct Relationship between Animal Health and Food Safety Outcomes. CAST Commentary QTA2012-1 ed. CAST, Ames, Iowa. Couret, D., W. Otten, B. Puppe, A. Prunier and E. Merlot. 2008. Behavioural, endocrine and immune responses to repeated social stress in pregnant gilts. Animal. 3:118-127.  Couret, D., A. Jamin, G. Kuntz-Simon, A. Prunier and E. Merlot. 2009. Maternal stress during late gestation has moderate but long-lasting effects on the immune system of the piglets. Vet. Immunol. Immunopathol. 131:17-24.  Creel, S. R. and J. L. Albright. 1988. The effects of neonatal social isolation on the behavior and endocrine function of holstein calves. Appl. Anim. Behav. Sci. 21:293-306.  106  Crestani, F., F. Seguy and R. Dantzer. 1991. Behavioural effects of peripherally injected interleukin-1: Role of prostaglandins. Brain Res. 542:330-335.  Croney, C. C. and S. T. Millman. 2007. Board-invited review: The ethical and behavioral bases for farm animal welfare legislation. J. Anim. Sci. 85:556-565.  Cyples, J. A., C. E. Fitzpatrick, K. E. Leslie, T. J. DeVries, D. B. Haley and N. Chapinal. 2012. Short communication: The effects of experimentally induced escherichia coli clinical mastitis on lying behavior of dairy cows. J. Dairy Sci. 95:2571-2575.  Dantzer, R. 2001. Cytokine-induced sickness behavior: Where do we stand? Brain Behav. Immun. 15:7-24.  Dantzer, R. and K. W. Kelley. 2007. Twenty years of research on cytokine-induced sickness behavior. Brain Behav. Immun. 21:153-160.  Dantzer, R. and K. W. Kelley. 1989. Stress and immunity: An integrated view of relationships between the brain and the immune system. Life Sci. 44:1995-2008.  Dawkins, M. S. 1990. From an animal's point of view: Motivation, fitness, and animal welfare. Behav. Brain Sci. 13:1-9.  Dawkins, M. S. 1983. The current status of preference tests in the assessment of animal welfare. Pages 20-26 S. H. Baxter, M. R. Baxter and J. A. C. MacCormack eds. Martinus Nijhoff, Boston. de Groot, J., M. A. W. Ruis, J. W. Scholten, J. M. Koolhaas and W. J. A. Boersma. 2001. Long-term effects of social stress on antiviral immunity in pigs. Physiol.  Behav. 73:145-158.  de Paula Vieira, A., M. A. G. von Keyserlingk and D. M. Weary. 2010. Effects of pair versus single housing on performance and behavior of dairy calves before and after weaning from milk. J. Dairy Sci. 93:3079-3085.  Devoino, L., E. Alperina and T. Pavina. 2003. Immunological consequences of the reversal of social status in C57BL/6J mice. Brain Behav. Immun. 17:28-34.  DeVries, T. J., K. A. Beauchemin and M. A. G. von Keyserlingk. 2007. Dietary forage concentration affects the feed sorting behavior of lactating dairy cows. J. Dairy Sci. 90:5572-5579.  DeVries, T. J., M. A. G. von Keyserlingk and K. A. Beauchemin. 2003. Short communication: Diurnal feeding pattern of lactating dairy cows. J. Dairy Sci. 86:4079-4082.  Dhabhar, F. S. and B. S. McEwen. 1997. Acute stress enhances while chronic stress suppresses cell-mediated immunity in vivo: A potential role for leukocyte trafficking. Brain Behav. Immun. 11:286-306.  107  Dhabhar, F. S., A. H. Miller, B. S. McEwen and R. L. Spencer. 1995. Effects of stress on immune cell distribution. dynamics and hormonal mechanisms. J. Immunol. 154:5511-5527.  Dickerson, S. S. and M. E. Kemeny. 2004. Acute stressors and cortisol responses: A theoretical integration and synthesis of laboratory research. Psychol. Bull. 130:355-391.  Dippel, S., C. B. Tucker, C. Winckler and D. M. Weary. 2011. Effects of behaviour on the development of claw lesions in early lactation dairy cows. Appl. Anim. Behav. Sci. 134:16.  Drackley, J. K. 1999. Biology of dairy cows during the transition period: The final frontier? J. Dairy Sci. 82:2259-2273.  Duffield, T. F., K. E. Leslie, K. D. Lissemore and S. T. Millman. 2009. Research and teaching of dairy cattle well being: Finding synergy between ethology and epidemiology. J. Appl. Anim. Welf. Sci. 12:132-142.  Duncan, I. J. H. 1992. Measuring preferences and the strength of preferences. Poult. Sci. 71:658-663.  Duve, L. R. and M. B. Jensen. 2012. Social behavior of young dairy calves housed with limited or full social contact with a peer. J. Dairy Sci. 95:5936-5945.  Dwyer, C. M. 2008. Individual variation in the expression of maternal behaviour: A review of the neuroendocrine mechanisms in the sheep. J. Neuroendocrinol. 20:526-534.  Dzialak, M. R., S. M. Harju, R. G. Osborn, J. J. Wondzell, L. D. Hayden-Wing, J. B. Winstead and S. L. Webb. 2011. Prioritizing conservation of ungulate calving resources in multiple-use landscapes. PLoS One. 6:e14597.  Ebbesen, P., J. A. Villadsen, H. D. Villadsen and K. E. Heller. 1991. Effect of subordinance, lack of social hierarchy, and restricted feeding on murine survival and virus leukemia. Exp. Gerontol. 26:479-486.  Edwards, S. A. 1983. The behaviour of dairy cows and their newborn calves in individual or group housing. Appl. Anim. Ethol. 10:191-198.  Edwards, S. A. and D. M. Broom. 1982. Behavioural interactions of dairy cows with their newborn calves and the effects of parity. Anim. Behav. 30:525.  Ehrchen, J. M., J. Roth, K. Roebrock, G. Varga, W. Domschke, R. Newberry, C. Sorg, C. Muller-Tidow, C. Sunderkotter, T. Kucharzik and T. W. Spahn. 2008. The absence of cutaneous lymph nodes results in a Th2 response and increased susceptibility to leishmania major infection in mice. Infect. Immun. 76:4241-4250.  108  Elenkov, I. J. and G. P. Chrousos. 1999. Stress hormones, Th1/Th2 patterns, pro/anti-inflammatory cytokines and susceptibility to disease. Trends Endocrinol. Metabol. 10:359-368.  Elenkov, I. J. and G. P. Chrousos. 2002. Stress hormones, proinflammatory and antiinflammatory cytokines, and autoimmunity. Ann. N. Y. Acad. Sci. 966:290-303.  Ely, D. L. and J. P. Henry. 1978. Neuroendocrine response patterns in dominant subordinate mice. Horm. Behav. 10:156-169.  Engler, H., M. T. Bailey, A. Engler and J. F. Sheridan. 2004. Effects of repeated social stress on leukocyte distribution in bone marrow, peripheral blood and spleen. J. Neuroimmunol. 148:106-115.  Enting, H., D. Kooij, A. A. Dijkhuizen, R. B. M. Huirne and E. N. Noordhuizen-Stassen. 1997. Economic losses due to clinical lameness in dairy cattle. Livest. Prod. Sci. 49:259-267.  Espejo, L. A., M. I. Endres and J. A. Salfer. 2006. Prevalence of lameness in high-producing holstein cows housed in freestall barns in minnesota. J. Dairy Sci. 89:3052-3058.  Fahey, A. G. and H. W. Cheng. 2008. Group size and density effects on physical indices and cell-mediated immunity in two genetic lines of white leghorn layers. Poult. Sci. 87:2500-2504.  Fairfield, A. M., J. C. Plaizier, T. F. Duffield, M. I. Lindinger, R. Bagg, P. Dick and B. W. McBride. 2007. Effects of prepartum administration of a monensin controlled release capsule on rumen pH, feed intake, and milk production of transition dairy cows. J. Dairy Sci. 90:937-945.  Fisher, A. D., M. A. Crowe, P. O'Kiely and W. J. Enright. 1997. Growth, behaviour, adrenal and immune responses of finishing beef heifers housed on slatted floors at 1.5, 2.0, 2.5 or 3.0 m2 space allowance. Livest. Prod. Sci. 51:245-254.  Fogsgaard, K. K., C. M. R?ntved, P. S?rensen and M. S. Herskin. 2012a. Sickness behavior in dairy cows during escherichia coli mastitis. J. Dairy Sci. 95:630-638.  Fogsgaard, K. K., M. S. Herskin, P. J. Gorden, A. M. O'Connor, L. L. Timms and S. T. Millman. 2012b. Survey of management and design of hospital pens at Iowa dairy farms. Dairy Cattle Welfare Symposium, Guelph, Canada.  Fraser, D. and D. M. Weary. 2009. Social and reproductive behaviour. Pages 73-84 in The Ethology of Domestic Animals. 2nd ed. P. Jensen ed. CABI, UK. Fraser, D., D. M. Weary, E. A. Pajor and B. N. Milligan. 1997. A scientific conception of animal welfare that reflects ethical concerns. Anim. Welfare. 6:187-205.  109  Fraser, D. 2008. Toward a global perspective on farm animal welfare. Appl. Anim. Behav. Sci. 113:330-339.  Fraser, D. 1993. Assessing animal well-being: Common sense, uncommon science. Page 37-51 in The proceedings of the conference on food animal well-being, West Lafayette, IN. Fregonesi, J. A., C. B. Tucker and D. M. Weary. 2007a. Overstocking reduces lying time in dairy cows. J. Dairy Sci. 90:3349-3354.  Fregonesi, J. A., D. M. Veira, M. A. G. von Keyserlingk and D. M. Weary. 2007b. Effects of bedding quality on lying behavior of dairy cows. J. Dairy Sci. 90:5468.  Fregonesi, J. A., M. A. G. von Keyserlingk and D. M. Weary. 2009. Cow preference and usage of free stalls compared with an open pack area. J. Dairy Sci. 92:5497-5502.  Friedman, H. S. 2008. The multiple linkages of personality and disease. Brain Behav. Immun. 22:668-675.  Galindo, F. and D. M. Broom. 2000. The relationships between social behaviour of dairy cows and the occurrence of lameness in three herds. Res. Vet. Sci. 69:75-79.  Galindo, F., D. M. Broom and P. G. G. Jackson. 2000. A note on possible link between behaviour and the occurrence of lameness in dairy cows. Appl. Anim. Behav. Sci. 67:335-341.  Galindo, F. and D. M. Broom. 2002. The effects of lameness on social and individual behavior of dairy cows. J. Appl. Anim. Welf. Sci. 5:193-201.  Gay, E. and J. Barnouin. 2009. A nation-wide epidemiological study of acute bovine respiratory disease in france. Prev. Vet. Med. 89:265-271.  Gibbons, J. M., A. B. Lawrence and M. J. Haskell. 2010. Measuring sociability in dairy cows. Appl. Anim. Behav. Sci. 122:84-91.  Godson, D. L., M. Campos, S. K. Attah-Poku, M. J. Redmond, D. M. Cordeiro, M. S. Sethi, R. J. Harland and L. A. Babiuk. 1996. Serum haptoglobin as an indicator of the acute phase response in bovine respiratory disease. Vet. Immunol. Immunopathol. 51:277-292.  Goldhawk, C., N. Chapinal, D. M. Veira, D. M. Weary and M. A. G. von Keyserlingk. 2009. Prepartum feeding behavior is an early indicator of subclinical ketosis. J. Dairy Sci. 92:4971-4977.  Gonyou, H. W. and J. M. Stookey. 1985. Behavior of parturient ewes in group-lambing pens with and without cubicles. Appl. Anim. Behav. Sci. 14:163-171.  Gonz?lez, L. A., B. J. Tolkamp, M. P. Coffey, A. Ferret and I. Kyriazakis. 2008. Changes in feeding behavior as possible indicators for the automatic monitoring of health disorders in dairy cows. J. Dairy Sci. 91:1017-1028.  110  Gosling, S., P. J. Rentfrow and W. B. Swann. 2003. A very brief measure of the big-five personality domains. J. Res. Pers. 37:504-528.  Grant, R. and J. Albright. 1995. Feeding behavior and management factors during the transition period in dairy cattle. J. Anim. Sci. 73:2791-2803.  Greenough, P. R. and J. J. Vermunt. 1991. Evaluation of subclinical laminitis in a dairy herd and observations on associated nutritional and management factors. Vet. Rec. 128:11-17.  Gregory, N. G., S. R. Payne, C. D. Devine and C. J. Cook. 2009. Effect of lipopolysaccharide on sickness behaviour in hens kept in cage and free range environments. Res. Vet. Sci. 87:167-170.  Griffith, F. L. I. 1898. The Petrie Papyri?Hieratic Papyri from Kahun and Gurob. Bernard Quartich, London, UK. Gross, W. B. 1984. Effect of a range of social stress severity on escherichia coli challenge infection. Am. J. Vet. Res. 45:2074-2076.  Grunert, E. 1979. Clinical aspects of the nutritional status of the dam and parturition. Pages 468-477 in Calving Problems and Early Viability of the Calf. Current Topics in Veterinary Medicine and Animal Science. 4th ed. B. Hoffmann, I. L. Mason and J. Schmidt eds. Springer, New York, NY. Gulliksen, S. M., E. Jor, K. I. Lie, I. S. Hamnes, T. L?ken, J. ?kerstedt and O. ?ster?s. 2009a. Enteropathogens and risk factors for diarrhea in norwegian dairy calves. J. Dairy Sci. 92:5057-5066.  Gulliksen, S. M., E. Jor, K. I. Lie, T. L?ken, J. ?kerstedt and O. ?ster?s. 2009b. Respiratory infections in norwegian dairy calves. J. Dairy Sci. 92:5139-5146.  Gundelach, Y., K. Essmeyer, M. K. Teltscher and M. Hoedemaker. 2009. Risk factors for perinatal mortality in dairy cattle: Cow and foetal factors, calving process. Theriogenology. 71:901-909.  Gupta, S., B. Earley and M. A. Crowe. 2007. Pituitary, adrenal, immune and performance responses of mature Holstein ? Friesian bulls housed on slatted floors at various space allowances. Vet. J. 173:594-604.  Haller, J., E. Fuchs, J. Hal?sz and G. B. Makara. 1999. Defeat is a major stressor in males while social instability is stressful mainly in females: Towards the development of a social stress model in female rats. Brain Res. Bull. 50:33-39.  111  Hammon, D. S., I. M. Evjen, T. R. Dhiman, J. P. Goff and J. L. Walters. 2006. Neutrophil function and energy status in holstein cows with uterine health disorders. Vet. Immunol. Immunopathol. 113:21-29.  Hart, B. L. 1988. Biological basis of the behavior of sick animals. Neurosci.  Biobehav.  Rev. 12:123-137.  Haskell, M. J., G. D. Hutson, L. G. Dickenson and S. Palmer. 1997. The pre-farrowing behaviour of sows with operant access to space for locomotion. Appl. Anim. Behav. Sci. 51:51-58.  Heegaard, P. M. H., D. L. Godson, M. J. M. Toussaint, K. Tj?rneh?j, L. E. Larsen, B. Viuff and L. R?nsholt. 2000. The acute phase response of haptoglobin and serum amyloid A (SAA) in cattle undergoing experimental infection with bovine respiratory syncytial virus. Vet. Immunol. Immunopathol. 77:151-159.  Hessing, M. J. C., G. J. Coenen, M. Vaiman and C. Renard. 1995. Individual differences in cell-mediated and humoral immunity in pigs. Vet. Immunol. Immunopathol. 45:97-113.  Hessing, M. J. C., C. J. M. Scheepens, W. G. P. Schouten, M. J. M. Tielen and P. R. Wiepkema. 1994. Social rank and disease susceptibility in pigs. Vet. Immunol. Immunopathol. 43:373-387.  Hill, A. E., A. L. Green, B. A. Wagner and D. A. Dargatz. 2009. Relationship between herd size and annual prevalence of and primary antimicrobial treatments for common diseases on dairy operations in the united states. Prev. Vet. Med. 88:264-277.  Holm, L., M. B. Jensen and L. L. Jeppesen. 2002. Calves? motivation for access to two different types of social contact measured by operant conditioning. Appl. Anim. Behav. Sci. 79:175-194.  Hopster, H., J. T. N. van der Werf and H. J. Blokhuis. 1998. Stress enhanced reduction in peripheral blood lymphocyte numbers in dairy cows during endotoxin-induced mastitis. Vet. Immunol. Immunopathol. 66:83-97.  Hudson, S. J. and M. M. Mullord. 1977. Investigations of maternal bonding in dairy cattle. Appl. Anim. Ethol. 3:271.  Hultgren, J. and C. Svensson. 2009. Heifer rearing conditions affect length of productive life in swedish dairy cows. Prev. Vet. Med. 89:255-264.  Hutson, G. D. and M. J. Haskell. 1990. The behaviour of farrowing sows with free and operant access to an earth floor. Appl. Anim. Behav. Sci. 26:363-372.  Huxley, J. N. and H. R. Whay. 2006. Current attitudes of cattle practitioners to pain and the use of analgesics in cattle. Vet. Rec. 159:662-668.  112  Huzzey, J. M., T. F. Duffield, S. J. LeBlanc, D. M. Veira, D. M. Weary and M. A. G. von Keyserlingk. 2009. Short communication: Haptoglobin as an early indicator of metritis. J. Dairy Sci. 92:621-625.  Huzzey, J. M., D. M. Veira, D. M. Weary and M. A. G. von Keyserlingk. 2007. Prepartum behavior and dry matter intake identify dairy cows at risk for metritis. J. Dairy Sci. 90:3220-3233.  Huzzey, J. M., M. A. G. von Keyserlingk and D. M. Weary. 2005. Changes in feeding, drinking, and standing behavior of dairy cows during the transition period. J. Dairy Sci. 88:2454-2461.  Hybschmann, G. K., A. K. Ersb?ll, H. Vigre, N. P. Baadsgaard and H. Houe. 2011. Herd-level risk factors for antimicrobial demanding gastrointestinal diseases in danish herds with finisher pigs: A register-based study. Prev. Vet. Med. 98:190-197.  Illmann, G. and M. ?pinka. 1993. Maternal behaviour of dairy heifers and sucking of their newborn calves in group housing. Appl. Anim. Behav. Sci. 36:91.  Ingvartsen, K. L., R. J. Dewhurst and N. C. Friggens. 2003. On the relationship between lactational performance and health: Is it yield or metabolic imbalance that cause production diseases in dairy cattle? A position paper. Livest. Prod. Sci. 83:277-308.  Ingvartsen, K. L. 2006. Feeding- and management-related diseases in the transition cow: Physiological adaptations around calving and strategies to reduce feeding-related diseases. Anim. Feed Sci. Technol. 126:175-213.  Iwersen, M., U. Falkenberg, R. Voigtsberger, D. Forderung and W. Heuwieser. 2009. Evaluation of an electronic cowside test to detect subclinical ketosis in dairy cows. J. Dairy Sci. 92:2618-2624.  Jackson, P. G. G. 2004. Normal birth. Pages 1-12 in Handbook of Veterinary Obstetrics. 2nd ed. Normal birth. Saunders, Philadelphia, PA. Jacobsen, S., P. H. Andersen, T. Toelboell and P. M. H. Heegaard. 2004. Dose dependency and individual variability of the lipopolysaccharide- induced bovine acute phase protein response. J. Dairy Sci. 87:3330-3339.  Jarvis, S., B. J. van der Vegt, A. B. Lawrence, K. A. McLean, L. A. Deans, J. Chirnside and S. K. Calvert. 2001. The effect of parity and environmental restriction on behavioural and physiological responses of pre-parturient pigs. Appl. Anim. Behav. Sci. 71:203-216.  Jawor, P. E., J. M. Huzzey, S. J. LeBlanc and M. A. G. von Keyserlingk. 2012. Associations of subclinical hypocalcemia at calving with milk yield, and feeding, drinking, and standing behaviors around parturition in holstein cows. J. Dairy Sci. 95:1240-1248.  113  Jenny, B. F., G. E. Gramling and T. M. Glaze. 1981. Management factors associated with calf mortality in south carolina dairy herds. J. Dairy Sci. 64:2284-2289.  Jensen, M. B. 2012. Behaviour around the time of calving in dairy cows. Appl. Anim. Behav. Sci. 139:195-202.  Jensen, M. B. and L. J. Pedersen. 2008. Using motivation tests to assess ethological needs and preferences. Appl. Anim. Behav. Sci. 113:340-356.  Jensen, M. B., L. J. Pedersen and L. Munksgaard. 2005. The effect of reward duration on demand functions for rest in dairy heifers and lying requirements as measured by demand functions. Appl. Anim. Behav. Sci. 90:207-217.  Johanson, J. M. and P. J. Berger. 2003. Birth weight as a predictor of calving ease and perinatal mortality in holstein cattle. J. Dairy Sci. 86:3745-3755.  John, O. P., R. W. Robins and L. A. Pervin. 2008. Handbook of Personality: Theory and Research. The Guilford Press, USA. Johnson, R. C. and P. Slade. 2003. Obstetric complications and anxiety during pregnancy: Is there a relationship? J. Psychosom. Obst. Gyn. 24:1-14. Johnson, R. R., R. Storts, T. H. Welsh, C. J. R. Welsh and M. W. Meagher. 2004. Social stress alters the severity of acute theiler's virus infection. J. Neuroimmunol. 148:74-85.  Jonsson, N. N., M. R. S. Fortes, E. K. Piper, D. M. Vankan, J. de Cisneros, J. Prada and T. Wittek. 2013. Comparison of metabolic, hematological, and peripheral blood leukocyte cytokine profiles of dairy cows and heifers during the periparturient period. J. Dairy Sci. 96:2283-2292.  Kamel, R. M. 2010. The onset of human parturition. Arch. Gynecol. Obstet. 281:975-982.  Kanitz, E., M. Tuchscherer, B. Puppea, A. Tuchscherer and B. Stabenowa. 2004. Consequences of repeated early isolation in domestic piglets (sus scrofa) on their behavioural, neuroendocrine, and immunological responses. Brain Behav. Immun. 18:35-45.  Kavelaars, A., C. J. Heijnen, B. Ellenbroek, H. van Loveren and A. Cools. 1997. Apomorphine-susceptible and apomorphine-unsusceptible wistar rats differ in their susceptibility to inflammatory and infectious diseases: A study on rats with group-specific differences in structure and reactivity of hypothalamic-pituitary-adrenal axis. J.  Neurosci. 17:2580-2584.  Kavelaars, A., C. J. Heijnen, R. Tennekes, J. E. Bruggink and J. M. Koolhaas. 1999. Individual behavioral characteristics of wild-type rats predict susceptibility to experimental autoimmune encephalomyelitis. Brain Behav. Immun. 13:279-286.  114  Kelley, K. W., R. Bluth?, R. Dantzer, J. Zhou, W. Shen, R. W. Johnson and S. R. Broussard. 2003. Cytokine-induced sickness behavior. Brain Behav. Immun. 17:112-118.  Kent, S., R. Bluth?, K. W. Kelley and R. Dantzer. 1992. Sickness behavior as a new target for drug development. Trends Pharmacol. Sci. 13:24-28.  Kidd, P. 2003. Th1/Th2 balance: The hypothesis, its limitations, and implications for health and disease. Altern. Med. Rev. 8:223-246.  Kimura, K., J. P. Goff, M. E. Kehrli and T. A. Reinhardt. 2002. Decreased neutrophil function as a cause of retained placenta in dairy cattle. J. Dairy Sci. 85:544-550.  Kirkden, R. D. and E. A. Pajor. 2006. Using preference, motivation and aversion tests to ask scientific questions about animals' feelings. Appl. Anim. Behav. Sci. 100:29.  Knott, L., J. F. Tarlton, H. Craft and A. J. F. Webster. 2007. Effects of housing, parturition and diet change on the biochemistry and biomechanics of the support structures of the hoof of dairy heifers. Vet. J. 174:277-287.  Koolhaas, J. M., A. Bartolomucci, B. Buwalda, S. F. De Boer, G. Flugge, S. M. Korte, P. Meerlo, R. Murison, B. Olivier, P. Palanza, G. Richter-Levin, A. Sgoifo, T. Steimer, O. Stiedl, G. van Dijk, M. Wohr and E. Fuchs. 2011. Stress revisited: A critical evaluation of the stress concept. Neurosci.  Biobehav.  Rev. 35:1291-1301.  Koolhaas, J. M. 2008. Coping style and immunity in animals: Making sense of individual variation. Brain Behav. Immun. 22:662-667.  Koolhaas, J. M., S. M. Korte, S. F. De Boer, B. J. Van Der Vegt, C. G. Van Reenen, H. Hopster, I. C. De Jong, M. A. W. Ruis and H. J. Blokhuis. 1999. Coping styles in animals: Current status in behavior and stress-physiology. Neurosci. Biobehav. Rev. 23:925-935.  Korte, S. M., J. M. Koolhaas, J. C. Wingfield and B. S. McEwen. 2005. The darwinian concept of stress: Benefits of allostasis and costs of allostatic load and the trade-offs in health and disease. Neurosci. Biobehav. Rev. 29:3-38.  Kossaibati, M. A. and R. J. Esslemont. 1997. The costs of production diseases in dairy herds in england. Vet. J. 154:41-51.  Leach, K. A., D. N. Logue, J. M. Randall and S. A. Kempson. 1998. Claw lesions in dairy cattle: Methods for assessment of sole and white line lesions. Vet. J. 155:91-102.  LeBlanc, S. J., K. D. Lissemore, D. F. Kelton, T. F. Duffield and K. E. Leslie. 2006. Major advances in disease prevention in dairy cattle. J. Dairy Sci. 89:1267-1279.  Leonard, F. C., J. O'Connell and K. O'Farrell. 1994. Effect of different housing conditions on behaviour and foot lesions in friesian heifers. Vet. Rec. 134:490-494.  115  Leuthold, W. 1977. African Ungulates. Springer, Berlin. Lidfors, L. and P. Jensen. 1988. Behaviour of free-ranging beef cows and calves. Appl. Anim. Behav. Sci. 20:237-247.  Lidfors, L. M., D. Moran, J. Jung, P. Jensen and H. Castren. 1994. Behaviour at calving and choice of calving place in cattle kept in different environments. Appl. Anim. Behav. Sci. 42:11-28.  Lindberg, C. 2001. Group life. Pages 406 in Social Behaviour in Farm Animals. L. J. Keeling and H. W. Gonyou eds. CABI, NY. Lischer, C. J., P. Ossent, M. R?ber and H. Geyer. 2002. Suspensory structures and supporting tissues of the third phalanx of cows and their relevance to the development of typical sole ulcers (rusterholz ulcers). Vet. Rec. 151:694-698.  Lo Fo Wong, D. M. A., J. Dahl, H. Stege, P. J. Van Der Wolf, L. Leontides, A. Von Altrock and B. M. Thorberg. 2004. Herd-level risk factors for subclinical salmonella infection in european finishing-pig herds. Prev. Vet. Med. 62:253-266.  Loehle, C. 1995. Social barriers to pathogen transmission in wild animal populations. Ecology. 76:326.  Lott, D. F. and J. C. Galland. 1985. Parturition in american bison: Precocity and systematic variation in cow isolation. Z.  Tierpsychol. 69:66-71.  MacDonald, J. M. and W. D. McBride. 2009. The Transformation of U.S. Livestock Agriculture: Scale, Efficiency, and Risks. United States Department of Agriculture, USA. Mainau, E. and X. Manteca. 2011. Pain and discomfort caused by parturition in cows and sows. Appl. Anim. Behav. Sci. 135:241-251.  Mee, J. F. 2008. Prevalence and risk factors for dystocia in dairy cattle: A review. Vet. J. 176:93-101.  Mee, J. F. 2004. Managing the dairy cow at calving time. Vet. Clin. N. Am-Food A. 20:521-546.  Meijering, A. 1984. Dystocia and stillbirth in cattle - A review of causes, relations and implications. Livest. Prod. Sci. 11:143-177.  Mendl, M., A. J. Zanella and D. M. Broom. 1992. Physiological and reproductive correlates of behavioural strategies in female domestic pigs. Anim. Behav. 44:1107-1121.  Merlot, E., M. C. Meunier-Sala?n and A. Prunier. 2004. Behavioural, endocrine and immune consequences of mixing in weaned piglets. Appl. Anim. Behav. Sci. 85:247-257.  116  Miedema, H. M., M. S. Cockram, C. M. Dwyer and A. I. Macrae. 2011a. Behavioural predictors of the start of normal and dystocic calving in dairy cows and heifers. Appl. Anim. Behav. Sci. 132:14-19.  Miedema, H. M., M. S. Cockram, C. M. Dwyer and A. I. Macrae. 2011b. Changes in the behaviour of dairy cows during the 24 h before normal calving compared with behaviour during late pregnancy. Appl. Anim. Behav. Sci. 131:8-14.  Miller, G. E., S. Cohen and A. K. Ritchey. 2002. Chronic psychological stress and the regulation of pro-inflammatory cytokines: A glucocorticoid-resistance model. Health Psychol. 21:531-541.  Miller, G. E., E. Chen and S. W. Cole. 2009. Health psychology: Developing biologically plausible models linking the social world and physical health. Annu. Rev. Psychol. 60:501-524.  Miller, G. E., E. Chen and E. S. Zhou. 2007. If it goes up, must it come down? Chronic stress and the hypothalamic-pituitary-adrenocortical axis in humans. Psychol. Bull. 133:25-45.  Miller, K. and D. G. M. Wood-Gush. 1991. Some effects of housing on the social behaviour of dairy cows. Anim. Prod. 53:271-278.  Mitlohner, F. M., J. L. Morrow-Tesch, S. C. Wilson, J. W. Dailey and J. J. McGlone. 2001. Behavioral sampling techniques for feedlot cattle. J. Anim. Sci. 79:1189-1193.  Moberg, G. P. 2000. Biological response to stress: Implications for animal welfare. Pages 377 in The Biology of Animal Stress: Basic Principles and Implications for Animal Welfare. G. P. Moberg and J. A. Mench eds. CABI, UK. Molony, V. and J. E. Kent. 1997. Assessment of acute pain in farm animals using behavioral and physiological measurements. J. Anim. Sci. 75:266-272.  Moore, A. S., H. W. Gonyou, J. M. Stookey and D. G. McLaren. 1994. Effect of group composition and pen size on behavior, productivity and immune response of growing pigs. Appl. Anim. Behav. Sci. 40:13-30.  Morrow-Tesch, J. L., J. J. McGlone and J. L. Salak-Johnson. 1994. Heat and social stress effects on pig immune measures. J. Anim. Sci. 72:2599-2609.  Mulligan, F. J., L. O?Grady, D. A. Rice and M. L. Doherty. 2006. A herd health approach to dairy cow nutrition and production diseases of the transition cow. Anim. Reprod. Sci. 96:331-353.  Newberry, R. C. and D. M. Shackleton. 1997. Use of visual cover by domestic fowl: A venetian blind effect? Anim. Behav. 54:387.  117  Nielsen, B. L. 1999. On the interpretation of feeding behaviour measures and the use of feeding rate as an indicator of social constraint. Appl. Anim. Behav. Sci. 63:79-91.  Noakes, D. E. 2001. Dystocia and other disorders associated with parturition. Pages 205-217 in Arthur's Veterinary Reproduction and Obstetrics. 8th ed. D. E. Noakes, T. J. Parkinson and G. C. W. England eds. Saunders, Philadelphia, PA. Noakes, D. E., T. J. Parkinson, G. C. W. England and G. H. Arthur. 2001. Parturition and the care of parturient animals. Pages 155-187 in Arthur's Veterinary Reproduction and Obstetrics. 8th ed. D. E. Noakes, T. J. Parkinson and G. C. W. England eds. Saunders, Philadelphia, PA. Nocek, J. E. 1997. Bovine acidosis: Implications on laminitis. J. Dairy Sci. 80:1005-1028.  O'Driscoll, K., L. Boyle and A. Hanlon. 2008. A brief note on the validation of a system for recording lying behaviour in dairy cows. Appl. Anim. Behav. Sci. 111:195-200.  Ohl, F. and F. J. van der Staay. 2012. Animal welfare: At the interface between science and society. Vet. J. 192:13-19.  Olsson, I. A. S. and L. J. Keeling. 2002. The push-door for measuring motivation in hens: Laying hens are motivated to perch at night. Anim. Welfare. 11:11-19.  Oltenacu, P. A., A. Frick and B. Lindhe. 1988. Use of statistical modelling and decision analysis to estimate financial losses due to dystocia and other diseases in swedish cattle. Acta. Vet. Scand. Suppl. 84:353-355.  Olynk, N. J. 2012. Assessing changing consumer preferences for livestock production processes. Animal Frontiers. 2:32-38.  Overton, T. R. and M. R. Waldron. 2004. Nutritional management of transition dairy cows: Strategies to optimize metabolic health. J. Dairy Sci. 87:E105-E119.  Pajor, E. A., J. Rushen and A. M. de Passill?. 2003. Dairy cattle?s choice of handling treatments in a Y-maze. Appl. Anim. Behav. Sci. 80:93-107.  Palanza, P., L. Gioiosa and S. Parmigiani. 2001. Social stress in mice gender differences and effects of estrous cycle and social dominance. Physiol. Behav. 73:411-420.  Pedersen, L. J. and T. Jensen. 2008. Effects of late introduction of sows to two farrowing environments on the progress of farrowing and maternal behavior. J. Anim. Sci. 86:2730-2737.  Pervin, L. A. and O. P. John. 1997. Personality: Theory and Research. 7th ed. John Wiley and Sons, Oxford. Philipsson, J. 1976. Studies on calving difficulty, stillbirth and associated factors in swedish cattle breeds. III. genetic parameters. Acta Agric. Scand. 26:211-220.  118  Pinheiro Machado, L. C., J. F. Hurnik and G. J. King. 1997. Timing of the attraction towards the placenta and amniotic fluid by the parturient cow. Appl. Anim. Behav. Sci. 53:183-192.  Prescott, N. B., T. T. Mottram and A. J. F. Webster. 1998. Relative motivations of dairy cows to be milked or fed in a Y-maze and an automatic milking system. Appl. Anim. Behav. Sci. 57:23-33.  Proudfoot, K. L., D. M. Veira, D. M. Weary and M. A. G. von Keyserlingk. 2009. Competition at the feed bunk changes the feeding, standing, and social behavior of transition dairy cows. J. Dairy Sci. 92:3116-3123.  Radostits, O. W., D. C. Blood and C. C. Gay. 1994. Acute carbohydrate engorgement of ruminants (rumen overload). Pages 262-269 in Veterinary Medicine. O. M. Radostits, D. C. Blood and C. C. Gay eds. W.B. Saunders, Philadelphia, PA. Ribble, C. S., A. H. Meek, P. E. Shewen, P. T. Guichon and G. K. Jim. 1995. Effect of pretransit mixing on fatal fibrinous pneumonia in calves. JAMA-J. Am. Vet. Med. Assoc. 207:616-619.  Rohleder, N., N. C. Schommer, D. H. Hellhammer, R. Engel and C. Kirschbaum. 2001. Sex differences in glucocorticoid sensitivity of proinflammatory cytokine production after psychosocial stress. Psychosom. Med. 63:966.  Rudine, A. C., M. A. Sutherland, L. Hulbert, J. L. Morrow and J. J. McGlone. 2007. Diverse production system and social status effects on pig immunity and behavior. Livest. Sci. 111:86-95.  Ruis, M. A. W., J. de Groot, J. H. A. te Brake, E. D. Ekkel, J. A. van de Burgwal, J. H. F. Erkens, B. Engel, W. G. Buist, H. J. Blokhuis and J. M. Koolhaas. 2001a. Behavioural and physiological consequences of acute social defeat in growing gilts: Effects of the social environment. Appl. Anim. Behav. Sci. 70:201-225.  Ruis, M. A. W., J. H. A. te Brake, B. Engel, W. G. Buist, H. J. Blokhuis and J. M. Koolhaas. 2001b. Adaptation to social isolation: Acute and long-term stress responses of growing gilts with different coping characteristics. Physiol.  Behav. 73:541-551.  Rushen, J., A. Boissy, E. M. C. Terlouw and A. M. de Passill?. 1999. Opioid peptides and behavioral and physiological responses of dairy cows to social isolation in unfamiliar surroundings. J. Anim. Sci. 77:2918-2924.  Sanderson, M. W., D. A. Dargatz and B. A. Wagner. 2008. Risk factors for initial respiratory disease in united states? feedlots based on producer-collected daily morbidity counts. Can. Vet. J. 49:373-378.  Sapolsky, R. M. 2005. The influence of social hierarchy on primate health. Science. 308:648-652.  119  Sapolsky, R. M., L. M. Romero and A. U. Munck. 2000. How do glucocorticoids influence stress responses? integrating permissive, suppressive, stimulatory, and preparative actions. Endocr. Rev. 21:55-89.  Sapolsky, R. M. 1990. Adrenocortical function, social rank, and personality among wild baboons. Biol. Psychiatry. 28:862-878.  Sa-Rocha, V. M., L. C. Sa-Rocha and J. Palermo-Neto. 2006. Variations in behavior, innate immunity and host resistance to B16F10 melanoma growth in mice that present social stable hierarchical ranks. Physiol. Behav. 88:108-115.  Schirmann, K., N. Chapinal, D. M. Weary, W. Heuwieser and M. A. G. von Keyserlingk. 2011. Short-term effects of regrouping on behavior of prepartum dairy cows. J. Dairy Sci. 94:2312.  Schmidt, D., S. O. Reber, C. Botteron, T. Barth, D. Peterlik, N. Uschold, D. N. M?nnel and A. Lechner. 2010. Chronic psychosocial stress promotes systemic immune activation and the development of inflammatory th cell responses. Brain Behav. Immun. 24:1097-1104.  Schrama, J. W., J. M. Schouten, J. W. Swinkels, J. L. Gentry, G. de Vries Reilingh and H. K. Parmentier. 1997. Effect of hemoglobin status on humoral immune response of weanling pigs differing in coping styles. J. Anim. Sci. 75:2588-2596.  Schuenemann, G. M., I. Nieto, S. Bas, K. N. Galv?o and J. Workman. 2011. Assessment of calving progress and reference times for obstetric intervention during dystocia in holstein dairy cows. J. Dairy Sci. 94:5494-5501.  Schutz, K. E., D. Davison and L. R. Matthews. 2006. Do different levels of moderate feed deprivation in dairy cows affect feeding motivation? Appl. Anim. Behav. Sci. 101:253-263.  Schutz, K. E., N. R. Cox and L. R. Matthews. 2008. How important is shade to dairy cattle? choice between shade or lying following different levels of lying deprivation. Appl. Anim. Behav. Sci. 114:307-318.  Seegers, H., C. Fourichon and F. Beaudeau. 2003. Production effects related to mastitis and mastitis economics in dairy cattle herds. Vet. Res. 34:475-491.  Segerstrom, S. C. 2000. Personality and the immune system: Models, methods, and mechanisms. Ann.  Behav.  Med. 22:180-190.  Selye, H. 1956. The Stress of Life. McGraw-Hill, New York. Sep?lveda-Varas, P., J. M. Huzzey, D. M. Weary and M. A. G. von Keyserlingk. 2013. Behaviour, illness, and management during the periparturient period in dairy cows. Anim. Prod. Sci. Online. Available: http://dx.doi.org/10.1071/AN12286. 120  Siivonen, J., S. Taponen, M. Hovinen, M. Pastell, B. Lensink, S. Py?r?l? and L. H?nninen. 2011. Impact of acute clinical mastitis on cow behaviour. Appl. Anim. Behav. Sci. 132:101-106.  Snowder, G. D., L. D. Van Vleck, L. V. Cundiff and G. L. Bennett. 2006. Bovine respiratory disease in feedlot cattle: Environmental, genetic, and economic factors. J. Anim. Sci. 84:1999-2008.  Stanley, T. A., R. C. Cochran, E. S. Vanzant, D. L. Harmon and L. R. Corah. 1993. Periparturient changes in intake, ruminal capacity, and digestive characteristics in beef cows consuming alfalfa hay. J. Anim. Sci. 71:788-795.  Stark, K. D. C. 2000. Epidemiological investigation of the influence of environmental risk factors on respiratory diseases in swine?A literature review. Vet. J. 159:37-56.  Stefanski, V. and H. Engler. 1998. Effects of acute and chronic social stress on blood cellular immunity in rats. Physiol. Behav. 64:733-741.  St?hulov?, I., M. ?pinka, R. ??rov?, L. M?chov?, R. Kn?z and P. Firla. 2013. Maternal behaviour in beef cows is individually consistent and sensitive to cow body condition, calf sex and weight. Appl. Anim. Behav. Sci. 144:89-97.  Stricklin, W. R. and J. A. Mench. 1987. Social organization. Vet. Clin. N. Am-Food A. 3:307-322.  Suls, J. and C. K. Wan. 1993. The relationship between trait hostility and cardiovascular reactivity: A quantitative review and analysis. Psychophysiology. 30:615-626.  Sutherland, M. A., S. R. Niekamp, R. W. Johnson, W. G. Van Alstine and J. L. Salak-Johnson. 2007. Heat and social rank impact behavior and physiology of PRRS-virus-infected pigs. Physiol. Behav. 90:73-81.  Sutherland, M. A., S. R. Niekamp, S. L. Rodriguez-Zas and J. L. Salak-Johnson. 2006. Impacts of chronic stress and social status on various physiological and performance measures in pigs of different breeds. J. Anim. Sci. 84:588-596.  Svensson, C., A. Linder and S. O. Olsson. 2006. Mortality in swedish dairy calves and replacement heifers. J. Dairy Sci. 89:4769-4777.  Tarlton, J. F., D. E. Holah, K. M. Evans, S. Jones, G. R. Pearson and A. J. F. Webster. 2002. Biomechanical and histopathological changes in the support structures of bovine hooves around the time of first calving. Vet. J. 163:196-204.  Taylor, S. E., L. C. Klein, B. P. Lewis, T. L. Gruenewald, R. A. R. Gurung and J. A. Updegraff. 2000. Biobehavioral responses to stress in females: Tend-and-befriend, not fight-or-flight. Psychol. Rev. 107:411-429.  121  Tenhagen, B. A., A. Helmbold and W. Heuwieser. 2007. Effect of various degrees of dystocia in dairy cattle on calf viability, milk production, fertility and culling. J. Vet. Med. A. 54:98-102.  Tolkamp, B. J. and I. Kyriazakis. 1999. To split behaviour into bouts, log-transform the intervals. Anim. Behav. 57:807-817.  Tuchscherer, M., E. Kanitz, B. Puppe and A. Tuchscherer. 2010. Altered immunomodulation by glucocorticoids in neonatal pigs exposed to a psychosocial stressor. Pediatr. Res. 68:473-478.  Tuchscherer, M., B. Puppe, A. Tuchscherer and E. Kanitz. 1998. Effects of social status after mixing on immune, metabolic, and endocrine responses in pigs. Physiol.  Behav. 64:353-360.  Tuchscherer, M., E. Kanitz, B. Puppe, A. Tuchscherer and B. Stabenow. 2004. Effects of postnatal social isolation on hormonal and immune responses of pigs to an acute endotoxin challenge. Physiol.  Behav. 82:503-511.  Tuchscherer, M., E. Kanitz, B. Puppe, A. Tuchscherer and T. Viergutz. 2009. Changes in endocrine and immune responses of neonatal pigs exposed to a psychosocial stressor. Res. Vet. Sci. 87:380-388.  Tucker, C. B. and D. M. Weary. 2004. Bedding on geotextile mattresses: How much is needed to improve cow comfort? J. Dairy Sci. 87:2889-2895.  Tucker, C. B., D. M. Weary and D. Fraser. 2003. Effects of three types of free-stall surfaces on preferences and stall usage by dairy cows. J. Dairy Sci. 86:521-529.  Turner, S. P., M. Ewen, J. A. Rooke and S. A. Edwards. 2000. The effect of space allowance on performance, aggression and immune competence of growing pigs housed on straw deep-litter at different group sizes. Livest. Prod. Sci. 66:47-55.  van Reenen, C. G., M. H. Mars, I. E. Leushuis, F. A. M. Rijsewijk, J. T. van Oirschot and H. J. Blokhuis. 2000. Social isolation may influence responsiveness to infection with bovine herpesvirus 1 in veal calves. Vet. Microbiol. 75:135-143.  van Reenen, C. G., N. E. O'Connell, J. T. N. van der Werf, S. M. Korte, H. Hopster, R. B. Jones and H. J. Blokhuis. 2005. Responses of calves to acute stress: Individual consistency and relations between behavioral and physiological measures. Physiol. Behav. 85:557-570.  VandeHaar, M. J. and N. St-Pierre. 2006. Major advances in nutrition: Relevance to the sustainability of the dairy industry. J. Dairy Sci. 89:1280-1291.  VandeHaar, M. J., G. Yousif, B. K. Sharma, T. H. Herdt, R. S. Emery, M. S. Allen and J. S. Liesman. 1999. Effect of energy and protein density of prepartum diets on fat and protein metabolism of dairy cattle in the periparturient period. J. Dairy Sci. 82:1282-1295.  122  Veenema, A. H., O. C. Meijer, E. R. D. Kloet, J. M. Koolhaas and B. G. Bohus. 2003. Differences in basal and stress-induced HPA regulation of wild house mice selected for high and low aggression. Horm. Behav. 43:197-204.  Vitale, A. F., M. Tenucci, M. Papini and S. Lovari. 1986. Social behaviour of the calves of semi-wild maremma cattle, bos primigenius taurus. Appl. Anim. Behav. Sci. 16:217-231.  von Keyserlingk, M. A. G. and D. M. Weary. 2007a. Maternal behavior in cattle. Horm. Behav. 52:106-113.  von Keyserlingk, M. A. G., A. Barrientos, K. Ito, E. Galo and D. M. Weary. 2012. Benchmarking cow comfort on north american freestall dairies: Lameness, leg injuries, lying time, facility design, and management for high-producing holstein dairy cows. J.  Dairy Sci. 95:7399.  von Keyserlingk, M. A. G., D. Olenick and D. M. Weary. 2008. Acute behavioral effects of regrouping dairy cows. J. Dairy Sci. 91:1011-1016.  von Keyserlingk, M. A. G. and D. M. Weary. 2007b. Maternal behavior in cattle. Horm. Behav. 52:106.  Waldron, M. R., A. E. Kulick, A. W. Bell and T. R. Overton. 2006. Acute experimental mastitis is not causal toward the development of energy-related metabolic disorders in early postpartum dairy cows. J. Dairy Sci. 89:596-610.  Weary, D. M., J. M. Huzzey and M. A. G. von Keyserlingk. 2009. Board-invited review: Using behavior to predict and identify ill health in animals. J. Anim. Sci. 87:770-777.  Weary, D. M., L. Niel, F. C. Flower and D. Fraser. 2006. Identifying and preventing pain in animals. Appl. Anim. Behav. Sci. 100:64-76.  Webster, A. J. F. 2002. Effects of housing practices on the development of foot lesions in dairy heifers in early lactation. Vet. Rec. 151:9-12.  Wehrend, A., E. Hofmann, K. Failing and H. Bostedt. 2006. Behaviour during the first stage of labour in cattle: Influence of parity and dystocia. Appl. Anim. Behav. Sci. 100:164-170.  Wittrock, J. M., K. L. Proudfoot, D. M. Weary and M. A. G. Von Keyserlingk. 2011. Short communication: Metritis affects milk production and cull rate of holstein multiparous and primiparous dairy cows differently. J. Dairy Sci. 94:2408-2412.  Wood-Gush, D. G. M., I. J. H. Duncan and D. Fraser. 1975. Social stress and welfare problems in agricultural animals. Pages 182-200 in The Behaviour of Domestic Animals. E. S. E. Hafez ed. Bailli?re Tindall, London.   

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0103338/manifest

Comment

Related Items