‘ .1 «$3.. .01...V \. "3%.. .0113: v xv , . . homage? . s. w. m L“. :h 5 hi“!- , a: . mm. (ital 4'0. x. . c: azurch... : ‘ . ‘53. v. a." pa. ,5 L. v3?” \r. ~0va: .4: :3. .2 3. . gag fitlt. L ff: :1. LE. ”mknwraxmmw THESIS (95100 1 LIBRARY Michigan State University This is to certify that the thesis entitled Behavioral and Physiological Responses of Horses to Initial Training: The Comparison Between Pastured Versus Stalled Horses presented by Elissette Rivera has been accepted towards fulfillment of the requirements for Animal Masters degree in Sc1ence Date §Z/}/ 3.) 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 1/98 chlRCIDabOmpfiS—pu Behavioral and Physiological Responses of Horses to Initial Training: The Comparison Between Pastured Versus Stalled Horses By Elissette Rivera A Thesis Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE 1999 Department of Animal Science Abstract Behavioral and Physiological Responses Of Horses To Initial Training: The Comparison Between Pastured Versus Stalled Horses By Elissette Rivera There are evidences that learning ability may be impaired in animals housed in social isolation or barren environments. Responses to initial training may be affected by housing conditions. Sixteen 2-yr-old Arabian horses were kept on pasture (P) (n=8) or in individual stalls (S) (n=8). Twelve horses (6P and 68) were subjected to a standardized training procedure, carried out by two trainers in a round pen, and 4 horses (2P and ZS) were used as control. On sample collection day 0, 7, 21 and 28, behaviors were recorded, blood samples were drawn and heart rates were monitored. Total training time for the stalled housed horses was significantly higher than total time for the pastured horses. The stalled group required more time to habituate to the activities occurring from the start of training to mounting. Frequency of undesirable behaviors such as bucks and jumps was higher in the stalled horses. On day 0 and day 21, basal plasma cortisol levels were lower when compared to basal plasma cortisol levels on day 7 and day 28. Pastured horses tended to have higher basal heart rates on day 0. While the physiological data failed to identify differences among housing groups, the behavioral data suggests that pasture-kept horses adapt easier to training than stalled horses. This thesis is dedicated to my parents Enrique Rivera and Delia Diaz and my son Pork Chop for their love, support and encouragement. iii Acknowledgments I would like to thank the many people who have supported me through this project. I have been fortunate to have two wonderful mentors Dr. John E. Shelle and Dr. Adoraldo J. Zanella and I would like to thank them for giving me the opportunity to conduct research under their supervision and for believing in me. My committee members, Dr. Christine Corn, Dr. Judy Marteniuk for all there support and Dr. Brian Nielsen for giving me the opportunity to conduct my study during his own research project. I would like to especially thank Scott Benjamin who’s effort and great attitude on collection days made this study possible. My honorary committee member, Carol Brown for being a friend and providing advice and humor when needed. My friends Renee Bell, Molly Nicodemus and Kristine Lang for their support and deep, philosophical lunchtime discussions. I am also very grateful to the Animal Behavior team and the Horse section for their valuable scientific input on this project. Dr. J arnes Jay and the Minorities and Women Graduate Assistanship program which financed this study. Finally, my dear friend Ramzi Mansoob, whom I have known for eight years, thank you for opening my eyes to the world and helping me become a better person. iv Table of Contents PAGE LIST OF TABLES ................................................................................................ vii LIST OF FIGURES ................................................................................................. x INTRODUCTION ....................................................................................... 1 LITERATURE REVIEW Housing ........................................................................................... 3 Training in horses ............................................................................ 4 Stress Response ............................................................................... 7 Behavioral ...................................................................................... 8 Physiological Indicators of Stress ................................................. 10 Hypothalamic- Pituitary- Adrenal Axis ................................ 11 Plasma cortisol ...................................................................... 14 Plasma cortisol and environment .......................................... 15 Plasma cortisol, training and exercise ................................... 16 Heart Rate .............................................................................. 19 Heart Rate and training ......................................................... 21 Summary ....................................................................................... 22 MATERIALS AND METHODS Animals ......................................................................................... 23 Housing ......................................................................................... 23 Animal Management ..................................................................... 24 Trainers .......................................................................................... 25 Training Arena .............................................................................. 25 Equipment Training tack ..................................................................... 26 Video observations ............................................................ 28 Blood collection preparation ............................................. 29 Heart rate monitors ............................................................ 29 Training Timeline ............................................................................. 30 Procedure ............................................................................ 3O Groundwork ....................................................................... 32 Riding ................................................................................. 34 Control group ..................................................................... 34 Data Collection Heart rate ............................................................................... 35 Blood sampling ..................................................................... 35 Preliminary data analysis Behavior data! Program configuration .................................. 35 Event recording ..................................................................... 36 Data Analysis Behavior data ......................................................................... 39 Reliability tests ...................................................................... 39 Cortisol assay ........................................................................ 39 Heart rate data ....................................................................... 40 Statistical Analysis ........................................................................ 40 RESULTS Training Times .............................................................................. 42 Behavior Results Horse behaviors during groundwork training ...................... 44 Trainer behaviors towards the horses during groundwork training ................................................................................. 46 Horse behaviors during riding portion of training ............... 49 Horse behaviors during groundwork and riding portion of training ................................................................................ 53 Physiological Data Cortisol ................................................................................. 55 Heart rate .............................................................................. 58 Correlation among horse behaviors and plasma cortisol ..... 59 DISCUSSION Training Times ............................................................................... 62 Horse and trainer behaviors during groundwork training 64 Horse behaviors during the riding portion of training .......... 66 Trainers .......................................................................................... 69 Cortisol .......................................................................................... 69 Heart rate ....................................................................................... 71 Correlation among horse behaviors and plasma cortisol ............. 72 CONCLUSION ........................................................................................... 74 APPENDIX A Glossary .......................................................................................... 75 APPENDIX B Behavior categories and codes ....................................................... 77 vi APPENDIX C Supplemental horse and trainer behavior categories ...................... 80 LIST OF REFERENCES ...................................................................................... 94 vii List of Tables Table 1. Calculated nutrient content of total ration as-fed .................................... 24 Table 2. Horse and trainer behavior categories used in study for the groundwork training period ..................................................................................................................... 37 Table 2b. Horse behavior categories and definitions that occurred during the riding portion of training .................................................................................................. 38 Table 2c. Behavior categories and definitions that occurred during both the groundwork and riding periods of training ................................................................................ 38 Table 2d. Description of training periods .............................................................. 38 Table 3. Duration (mean :1: SEM) of time spent standing still, during the groundwork period, by the treatment groups according to the experience of their trainers ...... 46 Table 4. Frequency (mean 1' SEM) of sacking with the blanket by the experienced and novice trainers ....................................................................................................... 48 Table 5. Frequency (mean 1 SEM) of looking towards the wall, during the riding period, for training days .................................................................................................... 49 Table 6. Frequency (mean :i: SEM) of looking towards the wall, during riding portion of training, for treatment groups according to the experience of their trainer ........... 49 Table 7. Frequency (mean 3: SEM) of head and neck extension upward behavior performed by treatment groups, during the riding period, according to experience of their trainer .................................................................................................................... 51 Table 8. Frequency (mean :I: SEM) of the head and neck extension straight behavior performed by the treatment groups, during the riding period, according to the experience of their trainer ........................................................................................................ 52 Table 9. Frequency (mean :i: SEM) of the head and neck extension down behavior performed by the treatment groups, during the riding period, according to the experience of their trainer ........................................................................................................ 52 Table 10. Frequency (mean :1: SEM) of the bucking and jumping, performed by the treatment groups, during the groundwork and riding period, according to the experience of their trainer ........................................................................................................ 54 Table 11. Duration (mean :1: SEM) of time spent carrying a normal tail setting, by treatment groups, during the groundwork and riding periods, according to the experience of their trainer ........................................................................................................ 55 viii Table 12. Correlation analysis for horse behaviors, performed throughout the study and cortisol ................................................................................................................... 61 Table 13. Correlation analysis of categories performed by the trainers and cortisol ............................................................................................................................... 61 Table 14. Frequency (mean :1: SEM) of applying pressure on the lead rope/reins by the trainers, for training days ...................................................................................... 81 Table 15. Frequency (mean :1: SEM) of pulling away from the pressure of the reins by the treatment groups, per training day ......................................................................... 81 Table 16. Duration (mean :t SEM) of chasing (min) by trainers on treatment groups per training days .......................................................................................................... 82 Table 17. Frequency (mean :1: SEM) of swinging rope/arms performed by the trainers, on the treatment groups, per training day ................................................................... 82 Table 18. Frequency (mean :1: SEM) of swinging rope/arms performed by the trainers according to the experience of the trainer ............................................................. 82 Table 19. Frequency (mean 1 SEM) of hitting the horses with the rope, per training day ............................................................................................................................... 83 Table 20. Frequency (mean :1: SEM) of head tossing performed by the treatment groups, during the groundwork and riding portion for training days ................................. 83 Table 21. Frequency (mean :1: SEM) of head tossing, performed by the treatment groups, during the groundwork and riding portion according to the experience of their trainer ............................................................................................................................... 83 Table 22. Duration (mean :1: SEM) of walking (min) by treatment groups for training days ............................................................................................. g ........................... 84 Table 23. Duration (mean :1: SEM) of walking (min) by treatment groups according to the experience of their trainer ..................................................................................... 84 Table 24. Duration (mean i SEM) of trotting (min) by treatment groups per training day ............................................................................................................................... 85 Table 25. Duration (mean :1: SEM) of walking (min) by treatment groups according to the experience of their trainer ..................................................................................... 85 Table 26. Duration (mean :1: SEM) of cantering (min) by treatment groups per training day ......................................................................................................................... 86 ix Table 27. Frequency (mean :1: SEM) of stopping, speeding and slowing down behaviors performed by the treatment groups for training days ............................................ 86 Table 28. Frequency (mean 1 SEM) of stopping behavior performed by treatment groups per training day ...................................................................................................... 87 Table 29. Frequency (mean :1: SEM) of speeding performed by treatment groups for training days .......................................................................................................... 87 Table 30. Frequency (mean a: SEM) of slowing down performed by treatment groups per training days .......................................................................................................... 88 Table 31. Frequency (mean :1: SEM) of stopping, speeding and slowing down performed by the treatment groups, according to the experience of their trainer ................... 88 Table 32. Frequency (mean i SEM) of side stepping performed by the treatment groups, for training days .................................................................................................... 89 Table 33. Frequency (mean :1: SEM) of side stepping by the horses according to the experience of their trainer ..................................................................................... 89 Table 34. Frequency (mean :1: SEM) of kicking by the trainers on the treatment groups on training days ..................................................................................................... 90 Table 35. Frequency (mean :t SEM) of kicking by the trainers on the treatment groups ............................................................................................................................... 90 Table 36. Frequency (mean :1: SEM) of smelling the environment performed by the treatment groups on training days ......................................................................... 91 Table 37. Frequency (mean :1: SEM) of smelling the environment performed by the treatment groups, according to the experience of their trainer .............................. 91 Table 38. Frequency (mean a: SEM) of defecation by the treatment groups, for training days ........................................................................................................................ 92 Table 39. Frequency (mean :1: SEM) of cow kicking performed by the treatment groups on training days ..................................................................................................... 92 Table 40. Frequency (mean :t SEM) of cow kicking performed by treatment groups, according to the experience of their trainer ........................................................... 93 List of Figures Figure 1. Hypothalamus-Pituitary-Axis system diagram ..................................... 12 Figure 2. Schematic diagram of roundpen used in the study ............................... 26 Figure 3. Photograph of equipment used during training ..................................... 27 Figure 4. Picture of bridle and snaffle bit, similar to the one used in the study... 28 Figure 5. Photograph of heart rate monitor .......................................................... 30 Figure 6. Timeline (days) for study ...................................................................... 30 Figure 7. Photographs of the hot walker used for the stalled horses during the study ............................................................................................................................... 3 1 Figure 8. Differences (mean :1: SEM) in total training time between treatment groups on training days. Time began once horses entered the roundpen until the trainer dismounted ............................................................................................................................... 43 Figure 9. Latency (mean :t SEM) of time between the horses entering roundpen until the trainer mounted the horses for the first time ......................................................... 43 Figure 10. Latency (mean :1: SEM) between the time the trainer first mounting the horses until the trainer dismount, on training days ........................................................... 45 Figure 11. Duration (mean :1: SEM) of time spent standing still on training days ............................................................................................................................... 45 Figure 12. Frequency (mean :1: SEM) of circling behavior performed by the horses during the groundwork period, over training days ........................................................... 47 Figure 13. Frequency (mean :1: SEM) of sacking the horses with a blanket, during the groundwork period, for training days .................................................................... 47 Figure 14. Frequency (mean :1: SEM) of head and neck extension up behavior performed by the horses, during the riding period, over training days ................................... 50 Figure 15. Frequency (mean i SEM) of head and neck extension straight performed by the horses, during the riding portion of training, over training days ..................... 51 Figure 16. Frequency (mean 1 SEM) of bucking and jumping by the treatment groups, over training days .................................................................................................. 53 xi Figure 17. Duration (mean i SEM) of normal tail setting during groundwork and riding periods of training, over training day .................................................................... 54 Figure 18. Plasma cortisol concentrations on training days, for treatment groups .. ............................................................................................................................... 57 Figure 19. Plasma cortisol concentrations at time sampling period for training days ........................................................................................................................ 57 Figure 20. Basal plasma cortisol concentrations on training days ....................... 58 Figure 21. Heart rate values at sample collection times ....................................... 59 xii Introduction Horses kept in stalls are deprived of opportunities for social bonding and the performance of natural behaviors (Hogan et al., 1988). Housing horses on pasture or in box stalls are accepted management techniques in the horse industry. Pasture sizes will vary according to the owners’ available land, while box stalls are typically 9 m2 or 13 m2 in size, possibly with a small window and/or dividing bars between stalls (Evans et al., 1990). Humans are accustomed to seeing horses in box stalls and may not perceive their housing condition as negatively as they might a calf in a veal crate. Housing horses in stalls is different from their natural, pasture-like environment. In the wild, horses have a structured social environment. Harem bands, averaging five to seven horses, are typically comprised of fillies, a few yearlings and foals and one stallion (Kirkpatrick and Francis, 1994). These cohesive bands may roam over vast areas of land and stay together even in the absence of the stallion. For the convenience of feeding and accessibility, humans stable horses inhibiting some behavior patterns and social interactions. Cunningham (1991) suggests that environmental conditions contribute to 65% of a horse’s performance potential while 35% of their potential is genetically inherited. Trainability affects the monetary value of horses since the horse must be handled for racing, showing and pleasure riding. If the environment negatively affects the animal’s learning ability, trainability along with the value of the horse may decrease. Environmental variables include, but are not lirrrited to, nutrition, trainer experience, degree of handling, temperature, lighting and housing. An animal’s ability to learn, solve problems or survive stressful situations is influenced by the environment they are raised in. The environmental condition a horse is exposed to does not only affect its learning ability, but may affect its welfare. If the environment causes the horse to make significant adjustments in behavior and physiology, the housing situation may be deemed as stressful (Fraser, 1992). Animals may respond to environmental challenges or stressors imposed on them through a variety of physiological, biochemical, anatorrrical, immunological and behavioral adaptation mechanisms (Ewbank, 1985). Behavioral and physiological measures have become widely used to determine stress levels and welfare in animals (Broom and Johnson, 1993). Abnormal behaviors that are not commonly observed in pastured horses have been attributed to stabling of horses. Weaving, cribbing and stall walking are a few of the behavior abnormalities linked to stalling horses (Kiley-Worthington, 1990). Physiological measures such as plasma cortisol (Boulton et al., 1997) and heart rate (Baldock and Sibling, 1990) have been associated with activation of the hypothalamus- pituitary-adrenal axis system, which is triggered during stressful situations in animals. Based on this information, the hypothesis of this study is that pastured yearlings will acclimatize more readily to initial training than individually stalled horses. The objective of this study is to determine if housing conditions may have an effect on behavioral and physiological measures in horses that are subjected to a standardized training procedure. In this study, stress is a response to stressors, the environment in which the animals are housed and the training regimen. The long-term goal of the study is to improve the welfare and housing conditions of horses. Literature Review Housing Housing environments can create behavioral problems as well as interfere with naturally occurring behaviors. Stalling weaned foals altered their behavior qualitatively and quantitatively when compared to foals weaned in a paddock (Heleski et al., 1999). In this study, paddock-housed foals spent the majority of their time standing in close contact with the other horses, interacting with the horses or eating. The stalled horses, could not perform some of the behaviors observed in the paddock housed foals (i.e. grazing, social interactions). In addition, the stalled foals performed more uncommon behaviors such as licking, chewing and kicking the stall walls. Rearing, bucking, bolting and head throwing in horses have been associated with some environmental variables such as a barren environment with insufficient stimuli, restriction in movement and isolation from conspecific or social partners (Kiley-Worthington, 1990). In addition to altering behaviors, barren environments can affect learning. Heird et a1. (1986) demonstrated that young horses in a richer environment perform better during complex tasks. Similarly, group housed rats raised with various toys learned a radial maze more quickly and accurately than rats raised in an isolated environment (Juraska et al., 1983). Moore and Spear (1995) reported that memory retention was decreased in rats housed with anesthetized dams and siblings as well as those housed in social isolation. However, when housed with foster dams, sires and littermates, these rats showed no alteration in their olfactory memory. Isolated male rats during an active avoidance test could not learn to avoid an electric shock when warned by an indicator light (Viveros et al., 1990). Forgay and Read (1962) demonstrated rats provided with wooden and metal objects to enrich their environment made fewer errors in a maze than rats housed in a barren environment. Housing conditions reflected through behavior and various trial experiments have influenced an animals’ learning ability which may affect trainability. Training in Horses The value of a horse is greatly increased by its trainability. Training relies on manipulating the horses’ environment into stimuli and reinforcements in order to attain a desired response (Yeates, 1994). Learning processes such as habituation, classical and operant conditioning, shaping, positive and negative reinforcements are used for training horses (Fraser, 1992). Habituation is defined as the decrease or cessation of a physiological or behavioral response as a result of repeated exposure to the stimulus (V oith, 1986). Horses refusing to load into a trailer will gradually habituate, with practice over time, and easily load into the trailer (Houpt, 1986). Classical conditioning is using unconditioned stimuli and unconditioned responses and pairing them with conditioned stimuli (Schmajuk, 1997). With time, the unconditioned response is associated with the conditioned stimuli and the unconditioned response becomes a conditioned response. Leading a stallion towards a customary service area (conditioned stimulus) can act as a sexual stimulus (unconditioned response) (Fraser, 1992). Sexual stimulation (unconditioned response) in the stallion becomes associated with the service area and the stimulation becomes the conditioned response. Operant conditioning is the process of modifying the frequency of a behavior as a result of an association between an environmental stimuli, an unconditioned stimuli and a response (Schmajuk, 1997). With horses, the most common use of operant conditioning is avoidance learning which rewards the horse by removing something unpleasant (Houpt, 1986). In order to cue a horse for forward motion, riders will press with their legs (stimulus), if the horse has learned to avoid the pressure from the legs (unconditioned stimulus), it should move forward on cue (operant response). Shaping, a form of operant conditioning, is teaching the horse small parts of a large maneuver over time (Houpt, 1986). Shaping begins by initially rewarding the animal when it performs a behavior that is similar to the desired response, then rewarding when successive approximations of the desired behavior are achieved until only the precise behavior is rewarded (Fraser, 1992). Reinforcements during training can be primary or secondary and positive or negative. Primary reinforcements are natural situations for the horse, such as food or returning to the herd. Secondary reinforcements are learned such as a pat on the neck or vocal praises (Waring, 1983). Positive and negative reinforcements develop a stronger link between a stimulus and a desired response in horses, increasing the chance of repeating its correct response (Tarpy, 1982). Positive reinforcements can be primary or secondary and are referred to as reward training. Dougherty and Lewis (1991) during discrimination training, demonstrated that horses learned quickly to press a lever once the correct picture was displayed resulting in a food reward. Haag et a1. (1980) demonstrated ponies that promptly learned to avoid a negative reinforcement, an electrical shock also, promptly learned a maze test that used positive reinforcement. Avoidance and punishment are methods of negative reinforcements which use aversive stimuli (Yeates, 1994). Using the reins to apply light contact on the mouth for the horse to back up is an avoidance responses since there is a possibility for a stronger application of the reins if the response is not reached (Yeates, 1994). However, this stimulus can also be used as a cue. Reinforcements increase the chance the response will occur again while punishment works to suppress or eliminate a response (T arpy, 1982). The purpose of punishment is to eliminate or weaken a response (Yeates, 1994). Both punishment and negative reinforcement use aversive stimuli. Punishment as well as any positive or negative reinforcement, must be paired closely to the response in order to be effective (Fraser, 1992; Yeates, 1994). Kratzer et al. (1977) observed fewer errors in Quarter Horses after presentation of an aversive stimulus; however a greater amount of time was spent on deciding which side of the maze to go through. Learning, a change in the brain in response to an event “outside of the brain”, will occur with or without human intervention, throughout the life of an animal (Broom and Johnson, 1993). Every encounter with a stimulus and/or response has the chance to become memorized and influence future behaviors and physiological functions (Broom and Johnson, 1993). “Learning from learned” means a horse is capable of incorporating what they have learned and apply it to the task at hand, suggesting future tasks will be easier to complete (Fiske and Potter, 1979). In the studies conducted by McCall et al., (1981) and Baer et a1. (1983) using discrimination learning and Hebb-Williams Closed Field Maze animals which had the opportunity to observe the correct response were more efficient in solving the problems themselves. Time between training sessions can also influence learning. Ponies that were trained once a week, in fewer sessions reached the highest level of performance when compared to ponies trained daily and biweekly (Rubin et al., 1980). Sm Response In order to survive, animals maintain homeostasis, which is defined by Broom and Johnson (1993) as the “relatively steady state of a body variable, which is maintained by means of physiological or behavioral regulation”. The physical and chemical properties of body fluids and tissues will remain constant despite the changes that are occurring. Any stimulus or adverse condition that can alter or disturb this balance can be considered a stressor. When confronted with a stressor, the human and animal body will undergo adaptation to maintain or regain homeostasis (Broom and Johnson, 1993). Stressors may be physical, psychological or psychosocial in nature (Mason, 1975). Physical stressors include, but are not limited to, injuries (Boyce et al., 1998), changes in body temperature (Geor et al., 1998; Friend et al., 1998), electric shock (Pynoos et al., 1996; Brundige, 1998) and changes in the environment such as confinement (Boyce et al., 1998). Psychological stressors are considered emotional or mental, such as fear or anxiety (Boissy et al., 1998). Psychosocial stress, as the term describes, refers to social interactions or lack of social interactions, for example isolation (Asterita, 1985). Although there are different forms of stressors, stress responses can activate three major pathways; autonomic nervous system, neuroendocrine system and behavior (Moberg, 1985). Behavior Behavioral observations can determine the response of an individual to a stressor. Variations in intensity, frequency and duration of behavioral responses from the normal level can indicate a disturbance (Broom and Johnson, 1993). Behavioral observations are inexpensive, relatively easy and do not subject the animals to additional stressors such as blood sampling aVIoberg, 1985). Behavioral measurements can provide an indication of short- term and long-term problems and if an animal is having difficulty in coping with a problem (Broom and Johnson, 1993). Coping is defined as “having control of mental and bodily stability” (Fraser and Broom, 1990). Coping with problems may not necessarily be harmful to the animal. Coping can be divided into three levels. In the first level, problems are tolerable and coping is easy (i.e. moderate exposure to cold or heat exposure). In the second level, the animal copes but with difficulty (i.e. stomach ulcers) and in the third level, the animal fails to cope and may collapse or die (Broom and Johnson, 1993). Short-term behavioral responses include situations of human intervention, handling, certain training techniques, transport, and pre-slaughter procedures. In horses, changes in ear, nostril, tail positions and posture, pressing their heads against the wall or staring at the abdomen can be evidence of short-term pain or discomfort (Fraser, 1969). In other species pawing, rolling, staring at flanks, groaning, frequent lying down and getting up, sitting in a dog position are indicators of pain (Waring, 1983). During a freeze and hot-iron branding study, cattle experiencing the hot-iron branding exhibited greater avoidance reactions (jumping away and kicking) toward the procedure (Lay et al., 1992). Long-term behavior responses deal with situations where the animal lacks control of their interactions with their environment (Broom and Johnson, 1993). Stereotypies are repetitive, invariant sequences of movements that have no obvious function (Mason, 1991). They are amongst the most important indicators of long-term welfare problems (Fraser and Broom, 1990). Inadequate or improper environmental conditions, isolation, and overcrowding are a few examples of causative factors leading to stereotypies (Mason, 1991). Pawing, in horses, was interpreted as a “response to frustration, a displacement activity that originated from the activity of uncovering food buried under snow” (Odberg, 1973). Pawing has been noted to occur in different situations, such as movement limitations, eating grain, anticipation of feed and trying to get a lying foal to stand (Houpt, 1986). Cribbing is when the horse grabs an object with its incisor teeth and force swallows gulps of air (Evans et al., 1990). Cribbing may lead to gastric upsets or colic. Weaving is when the horse weaves its head back and forth while rocking side to side (Evans et al., 1990). In 1995, McGreevy et al., using questionnaires, studied 1750 horses in different equestrian disciplines and revealed a higher percentage of cribbing and weaving in dressage and event horses than in horses used in endurance riding. There was also a positive correlation, both within and between the dressage and eventing horses, in the amount of time spent in a stall to the occurrence of stereotypic behavior. Although wood chewing occurs in horses housed both in restricted environments and pasture, Sarnbraus (1985) determined the behavior is aggravated in close confinement. Piglets which were reared in a “poor environment” (farrowing crate) compared to piglets reared in a “rich environment” (outdoor pasture with half-open crates) displayed more aggressive behaviors (DeJonge et al., 1996). In addition, the subordinates in the poor environment showed symptoms of chronic social stress exposure by delayed onset of puberty, smaller weight gain and elevated cortisol levels. Environmental stress and welfare of animals can be determined using long and short-term behavioral observations. Physiological Indicators of Stress Hans Seyle was the first to report the neuroendocrine response to stress in 1956. Seyle (1956) reported that animals who undergo challenges to their homeostasis such as heat, cold, and infection will elicit morphological reactions in the body, such as decrease in size of the thymus and lymphatic structures, occurrence of ulcers in the gastrointestinal tract and adrenocortical stimulation. Seyle concluded this triad of reactions represents a non-specific reaction to almost all-harmful stimuli and refers to this syndrome as the “general adaptation syndrome”. Selye hypothesized the general adaptation syndrome, through a series of nervous and endocrine (neuroendocrine system) pathways can help the body to adapt to the stress. The general adaptation syndrome is divided into three stages (Seyle, 1956). Alarm reaction, the initial response, is the body’s first defense mechanism. In this stage, the adrenal cortex will release glucocorticoids into the bloodstream, depleting its stores. If the animal does not die during the alarm phase, the stage of resistance follows. During the resistance stage the body counteracts the alarm reaction by coping with the stressor and returning the body to normal conditions. If the stressor persists, the third phase of exhaustion is reached. At this stage, the animal or human is no longer able to adapt, leading to death. Three main criticisms have been argued against Selyes’ concept of stress (Broom and Johnson, 1993; Mason, 1975). First, the adrenal cortex response does not always 10 occur or may decrease during stressful situations. Second, there are other physiological measures that may occur during a stressful situation. Finally, Seyle has been criticized for his inconsistent usage of the word stress. Nonetheless, Selyes’ work has stimulated decades of research on stress and the neuroendocrine system (Mason, 1975). Broom and Johnson (1993) define stress as an “environmental effect on an individual which overtaxes its control systems and reduces its fitness or appears likely to do so”. The term stress means many things to different people. It is a term that has been applied to a stimulus, response and an interaction between stimulus and response (Barnett and Hemsworth, 1990). Lazarus (1971) suggested that scientists avoid using the term loosely by clearly defining the manner in which the concept is used. Physiological indicators of stress, such as the hormones released by the hypothalamic-pituitary-adrenal axis (Fig. 1) and heart rate measurements can be used to assess animal welfare. Hypothalamic-Pituitary—Adrenal Axis The hypothalamus is responsible for regulating body functions (i.e. thirst, hunger, pleasure and pain) and releasing neurohormones, which affect endocrine function in the body (Asterita, 1985). Stimulation of body functions and neurohormones can occur by visual, auditory, tactile (pain), or physiological disturbances leading to a change in homeostasis. Once stimulation has occurred, the hypothalamus releases corticotrophin- releasing factor (CRF). The paraventricular nucleus (PVN) and the anterior paraventricular nuclei (PVA) are the primary locations for the synthesis of CRF (Brown, 11 paraventricular nuclei (PVA) are the primary locations for the synthesis of CRF (Brown, 1994). Corticotropin releasing factor is the dominant regulator of adrenocorticotropic hormone (ACTH) release from the pituitary gland. VP¢CRF ¢ ACTH l Cortisol Fig. l. Hypothalamus — Pituitary — Adrenal Axis system diagram The pituitary gland connects to the hypothalamus at the base of the brain via the hypophyseal stalk (Brown, l994). The gland’s primary function is to regulate neuroendocrine function. Adrenocorticotropic hormone is produced in the corticotroph cells of the anterior lobe and melanotroph cells in the intermediate lobe (adenohypophysis) of the pituitary gland (Brown, 1994). Adrenocorticotropic hormone is derived from a prohormone. proopiomelanocortin (Eipper and Manis. 1980) and is 1’) l— hormones. Proopiomelanocortin is. a large molecule that, once enzymatically cleaved, releases the sequence for seven pituitary peptides, including ACTH (Brown, 1994). The distal lobe of the pituitary gland produces oxytocin and vasopressin. The intermediate lobe synthesizes melanocyte-stimulating hormone, while the anterior lobe produces growth hormone, thyroid stimulating hormone, follicle stimulating hormone, luteinizing hormone, prolactin, and ACTH. Simultaneously and independently, under stressful situations, both CRF and VP can trigger the release of adrenocorticotropic hormone (ACTH) from the pituitary gland in horses (Redekopp et al., 1986; Livesey et al., 1988). Vasopressin is an anti-diuretic hormone that promotes water absorption in the kidneys and increases blood pressure. During exercise, horses displayed a significant increase in vasopressin and CRF secretions remained unchanged (Alexander et al., 1991). In all farm animal species studied, CRF and VP combined magnified the secretion of ACTH (Minton and Parssons, 1993; Familari et al., 1989). Once the pituitary gland is activated, the adrenal cortex is stimulated to release steroid hormones; mineralcorticoids (aldosterone), sex steroids (androgens) and glucocorticoids. Glucocorticoid levels increase during many short-term situations. Glucocorticoid measurement allows for valuable information on the welfare of the animals (Broom and Johnson, 1993). Glucocorticoids influence blood glucose levels, carbohydrate and protein metabolism. In addition, glucocorticoids have been found to inhibit formation of antibodies, lower lymphocyte counts and delay the growth of new tissue (Vining and McGinely, 1986). Glucocorticoids can maintain or inhibit the production of ACTH and CRF (Broom and Johnson, 1993). Antoni et al. (1990) showed 13 l' increases in CRF and VP in adrenalectomized rats and that the increases could be blocked by replacement glucocorticoids. Furthermore, ACTH and glucocorticoid responses to stress can be blocked by prior adrrrinistration of glucocorticoids (Keller-Wood and Dallman, 1984). McFarlane (1995) demonstrated in sheep that low concentrations of cortisol inhibited ACTH secretion responses to exogenous CRF and AVP. However, a combination of CRF and AVP was able to override the feedback system, stimulating ACTH release. In rats, Castro and Moreira (1996) observed that CRH secretion was inhibited by ACTH administration in a negative, dose-dependent manner, demonstrating a short-loop feedback mechanism for CRH. Plasma Cortisol Cortisol is the primary glucocorticoid produced by the adrenal cortex in horses (Zolovick et al., 1966). Most of the circulating cortisol is bound to plasma proteins. Cortisol binding globulins bind to 80-90% of circulating cortisol and 5-10% is loosely bound to alburrrin (Brody et al., 1994). Cortisol levels participate in the stress response by potentiating the activities of the sympathetic nervous system, increasing hepatic gluconeogenesis and glycogenolysis, increasing protein catabolism, inhibiting ACTH secretion (negative feedback mechanism), maintaining blood pressure by sensitizing arterioles to the action of noradrenaline and increasing renal excretion (Brody et al., 1994). In horses, circadian rhythms have been observed in cortisol (Lebelt et al., 1996). The age and gender of a horse does not have an affect on the circadian rhythm (Hoffis et al., 1970). Discrepancies in the literature have been reported on the occurrence of 14 circadian rhythms in horses. Some studies have reported changes (Hoffis et al., 1970; Evans et al., 1977) in the circadian rhythms of horses while others did not record any changes (Elier et al., 1979). Typically, circadian cortisol activity peaks in the morning (0600-0900) and nadir at night (0700-2100) (Lebelt et al., 1996; Larsson et al., 1979). Irvine et al. (1994) concluded that a circadian cortisol rhythm in horses occurs when the animals are left undisturbed. However, any minor disruption, such as environment change and training, can alter the rhythm. Although the circadian rhythm could be lost, Irvines’ et al. (1994) study also showed horses could adapt to their environment, reestablishing the circadian rhythm. Slone et al. (1983) examined the disappearance rate of cortisol in bilateral adrenalectomized horses. Cortisol was observed having a biphasic disappearance. Phase one, the redistribution phase, is the rapid (0 to 30 min) liner decrease of free and bound cortisol from the blood. Phase two, elimination phase, is a slow (2 to 12 h), progressive linear decrease to non-detectable levels. During this study, the half—life of cortisol was measured at 2.1 i 0.6 h. Plasma cortisol and environment Plasma cortisol has been examined under various, and possibly stressful, environmental conditions such as housing isolation, barren or complex environments and temperature variability. Pseudopregnant gilts, injected with prostaglandin F20t, in which nest building behavior was induced, showed a greater level of cortisol when housed in a farrowing crate versus a pen (Boulton et al., 1997). In addition, the amount of time spent inactive linearly correlated with the cortisol concentrations in both penned and crated 15 gilts. Siberian dwarf hamsters displayed high level of cortisol when separated from their mates (Castro and Matt, 1997). Contrary to the gilts examined by Boulton et al. (1997) the Siberian dwarf hamsters did not display a correlation with cortisol and inactivity time. Alexander et al. (1988) showed mean cortisol levels were not significantly altered during a ten-min isolation period, however the horses displayed agitation, vocalization, hyperventilation and sweating. In squirrel monkeys, a 32% increase in cortisol from basal was observed in monkeys housed without their companions (Lyons et al., 1995). Al- Gahtani et al. (1991) determined that sheep isolated from their herdmates had higher levels of cortisol. These studies show separation, barren environments and isolation accompanied by an increase in hypothalamic-pituitary-adrenal axis function can be perceived as stressful in several species. Plasma cortisol, training and exercise The role of glucocorticoids, such as cortisol, is to stimulate gluconeogenesis thereby mobilizing amino acids and fats. Gluconeogenesis increases liver conversion of amino acids into glucose, which increases liver glycogen and blood glucose (Tharp, 1975). Mobilization of amino acids from tissues results in a protein catabolic effect in muscle and leaves the animal in negative nitrogen balance. Mobilization of fatty acids from adipose tissues increases free fatty acids. From these roles one can hypothesize that these changes during exercise can be helpful by providing greater blood supplies of compounds used for energy and synthesis of needed cellular compounds (Tharp, 1975). Plasma cortisol increases can be used to predict exercise-induced stress in horses (Linden et al., 1990). Linden et al. (1990) measured plasma cortisol after a physiological l6 and pharmacological stressor occurred. The physiological stressor was a cross-country competition, while the pharmacological stressor was the administration of ACTH one- week after the competition. After the occurrence of the pharmacological and physiological stressors a relative increase of plasma cortisol was observed, suggesting each individual has a specific adrenocortical response independent of the type of stress (Linden et al., 1990). Submaximal exercise studies in horses have varied in results. Glucocorticoid 11- hydroxycorticosteriod was found to increase progressively during and after a ten week training study conducted by Snow and Mackenzie in 1977. During a short standardized exercise test, Church et al. (1987) determined exercise produced a highly significant increase in plasma ACTH and cortisol followed by a decrease post exercise. In contrast, during light exercise, cortisol levels were decreased throughout the study and during the recovery period (Davies and Few, 1973). Horses swimming over 15 min had higher cortisol levels immediately after swimming than horses swimming less than 15 min (Garcia and Beech, 1986). During maximal exercise, as the workload increases, studies have clearly demonstrated elevated cortisol levels. Snow and Rose (1981) demonstrated high levels of plasma cortisol (440 :l: 15 nmol / 1) during a long distance exercise which did not decrease (343 i 11 nmol / 1) until thirty min post exercise. Lucke et al. (1980) examined endurance riding in Arabian horses competing in a two day, long distance ride (80 km followed by a 40 km). They demonstrated that mean cortisol levels were higher in both the 80 km ride (pre-ride: 202 i 1 nmol l; post-ride: 574 i nmol 1) and 40 km ride (pre-ride: 236 :1: 19.8; post-ride: 401.7 i 43.6). Lucke et al. (1980) examined Arabian horses during a marathon 17 race (42 km) and found a significant rise in blood glucose, plasma free fatty acid and glucagon associated with a rise in plasma cortisol (pre-ride: 151 i 14.8 nmol / 1; post- ride: 445 i 23). Distance has an effect on glucocorticoid levels. Additionally, speed, recovery rates and level of fitness in the animal can be correlated with glucocorticoids. Rose et al. (1983) found that the fast group (234 m rrrin) of horses which completed a 160 km endurance ride had significantly higher cortisol values post-ride (540 i 50 vs. 284 i 19 nmol 1) when compared to the slower group (144 m rrrin). Trotting horses at 5m/sec on a treadmill for 55 min resulted in increases of plasma cortisol levels while post-exercise cortisol levels continued to increase (V alberg et al., 1989). In addition, at increasing speeds, cortisol levels continuously rose (Valberg et al., 1989). Dybal et al. (1980) conducted a two-year study on plasma corticosteroids in horses at pre—ride, mid-ride and post-ride during a 160-km endurance ride. In the first year of this study, post-ride levels returned to pre-ride levels, however, in the second year, using different horses, post-ride levels remained significantly elevated. A study comparing cortisol concentrations in trained and untrained Arabian horses demonstrated that both groups exhibited elevated cortisol levels post exercise. However, the cortisol levels in the trained group returned to basal levels quicker during a standardized test (Hower and Wicker, 1989). Overall, there are many variables that may affect levels of cortisol in animals. Increased speed, intensity, duration, environmental temperature and prior training experience have lead to variable cortisol values. Researchers have hypothesized that the reason for such variation in results can be due to psychological stress and/or physiological 18 stress (Tharp, 1975). However, it is important to emphasize that increase of cortisol and ACTH is not synonymous with stress. Heart rate Sympathetic and para-sympathetic neural pathways are included in the autonomic nervous system. The interaction of these two pathways is commonly referred to as the “fight or flight” response (Cannon, 1929). This response provides the body with the necessary physiological measures needed to handle a dangerous situation when detected. Release of catecholarrrines from the adrenal medulla during sympathetic stimulation intensifies the sympathetic activity (Swenson and Vogel, 1983). Sympathetic activity during a stressful situation increases body activities needed, such as heart rate, circulation and oxygen uptake while inhibiting activities, i.e. gut mobility, that are not necessary for survival at the time. Heart rate can be a useful indicator for recognizing acute stress in animals (Marchant et al., 1995). Cardiac measurements using radiotelemetry have been widely used in animal behavior studies (Stohr, 1988). Radiotelemetry provides a non-invasive procedure to measure heart rate and heart rate variability in horses (Evans and Rose, 1986). In general, during stressful situations, the parasympathetic nervous system decreases (bradycardia) and the sympathetic nervous system increases (tachycardia) heart rate. Although the heart has a greater innervation of sympathetic nerves compared to parasympathetic nerves, there are cases where bradycardia occurs more readily (Hindell and Lea, 1998). Tachycardia was exhibited in sheep during visual isolation, introduction 19 to a new flock, human and canine interaction (herding the flock) and transportation (Baldock and Sibly, 1990). However, this study also showed no increase in heart rate when the sheep were isolated with a conspecific or confined in a stationary trailer. Morton and Griffith (1985) reported that elevated heart rates are associated with greater pain sensation in animals. Horse trailer transportation studies revealed travelling heart rates are higher when compared to non-travelling heart rates (Clark et al., 1993; Smith et al., 1994). Vibration, noise and handling were stressors used during a transportation study in pigs (Stephens and Rader, 1983). Throughout the study, heart rates remained increased but handling and restraint caused the maximal disturbance. Brundige (1998) demonstrated that the use of an electrical prod when moving pigs for loading increased heart rates 1.7 times that of basal levels. Increased heart rates were observed in cattle when restrained in a novel breeding box compared to restraining the cattle in a breeding chute to which they were typically exposed (Lay et al., 1992). Pollard et al. (1993) showed that deer stags confined with unfamiliar animals had higher heart rates than when confined with familiar deer stags. During a study examining the effects of twitching horses, heart rates were monitored (Lagerweij et al., 1984). In this study, painful stimuli were applied with and without twitching the horse. While twitching, the release of endogenous opiods helped to decrease the pain, lowering heart rates by 8% whereas, without twitching the stimuli caused a 22% increase in heart rate. In addition, the percent changes in heart rate were related positively to the reactive behaviors to the painful stimuli. Overall, handling, management techniques and environments can be stressful as demonstrated by the increases in heart rate in various species. 20 Heart rate and training Pre-exercise studies have demonstrated an anticipatory affect by heart rates in horses. Hall et al. (1976) revealed an increase in mean heart rates in Thoroughbred horses while moving between their home stall to a preparation shed and from the preparation shed to the race track, prior to exercise. Cardiovascular measurements can provide information on determining if training is capable of inducing adaptations and greater tolerance to the exercise (Gottlieb et al., 1995). Heart rates during sub-maximal exercise have been found to be lower after training (Bayle et al., 1983). Untrained Standardbred horses in poor physical fitness demonstrated higher resting heart rates, higher heart rates during exercise and a slower recovery or resting heart rate when compared to highly trained horses (Marsland, 1968). Foreman et al. (1990) examined Thoroughbred horses with previous training experience in a standardized test on a racetrack. In this study, recovery heart rates decreased indicating an increase in cardiovascular fitness. Heart rate monitoring can help prevent a trainer from excessively exercising subjects that cannot otherwise communicate fatigue (Ivers, 1982). 21 Summary This review examined the effects of housing environments on behavior and physiological responses and welfare of horses. Barren environments can hinder the acquisition of learning concepts and may induce the development of behavior abnormalities. Raising animals in a challenging and complex environment can help decrease the occurrence of abnormal behavior, increase learning abilities and improve the level of welfare for many species. Changes in behavior, as well as activation of the autonomic and neuroendocrine systems, occur under physical, psychological or psychosocial stress. These changes when monitored can indicate if an animal is having difficulty coping with a situation thereby, classifying the situation as stressful. Physiologically, during a stressful situation activation of the hypothalamic-pituitary—adrenal axis (Fig. 1) will occur. Once activated, the hypothalamic-pituitary-adrenal axis influences the release of glucocorticoids. Glucocorticoid levels have increased during many short-term situations allowing for assessment on the welfare of animals. Isolation and barren housing conditions have induced higher cortisol levels when compared to animals housed in a richer environment. During sub-maximal exercise studies, increases and decreases in plasma cortisol concentrations have been revealed. While undergoing maximal exercise, most studies have demonstrated elevated cortisol levels. Heart rate measurements using radiotelemetry is a non-invasive method in recognizing acute stress in animals. The sympathetic nervous system will increase in the presence of a stressor such as transportation, handling, restrainment and electric prodding. 22 Materials and Methods Animals Sixteen Arabian yearlings, owned by Michigan State University, were used in this experiment. The experimental group was comprised of castrated males (n=11) and intact females (n=5), with an average age of 18.6 month (range 16.6 —19.9). Average weight for the stalled and pastured horses at the start of the project was 349 :I: 10 kg and 342 i 10 kg, respectively. Housing All animals were housed at the Michigan State University Horse Teaching and Research facility. The horses were randomly assigned to three treatment groups according to housing, training and age: pasture (P) with training (n=6) or stalled (S) with training (n=6) and a control (C) group with no training (n=4). The control group was assigned into the pasture (n=2) or stalled (n=2) treatment group. Two weeks into the experiment, due to an accidental injury, one of the pastured horses was removed from the project. Therefore, only five horses were in the pasture group for the remainder of the study. Pasture horses were housed on approximately 29,166 m2 (seven acres) of land with a water dispenser, hay trough and grain trough in the field. Maximum temperatures on training days were 3 °C, 9 °C, 5 °C and 7 °C, respectively. The 10 m2 box stalls used for housing had dividing walls constructed with wooden panels on the bottom and steel guards on the top, allowing limited social 23 interaction between the horses. Each stall had a window along with a hay trough, grain bucket and water bucket. Stalls were bedded with wood shavings or straw. Natural sunlight and barn lights provided lighting during the day. At night (19:00 h) illumination in the corridor were turned off and individual stall lights were turned on. Animal Management The study was conducted during the winter when minimal nutrients were available from pasture grazing. All treatment groups were fed a diet recommended by NRC (1989) for long yearlings and two-year-olds (Table 1). Each horse was fed 1.8 kg, per day, of a commercial concentrate (Strategym, Purina Mills, Lansing, MI) divided into two equal feedings, at 06:00 and 18:00 h. In addition, horses had unlimited access to mixed alfalfa- grass hay. Stalls were cleaned daily at 16:00 h and all horses had their hooves trimmed and were dewormed routinely. Vaccinations against rhinopneumonitis, tetanus, influenza and equine encephalomyelitis were provided to all horses. Table 1 Calculated nutrient content of total ration as-fed Nutrient Units Ration DE Meal/kg 2. 169 CP % 14.7 Ca % .75 P % .29 Hoekstra, 1998 24 Trainers Two trainers were involved in the training procedure. The novice trainer (female, 50.0 kg) was a graduate student with no prior experience in the initial training of horses. The second trainer had ten years of horse experience (male, 62.3 kg). In the interest of human safety, horses that were subjectively determined as "easier" to train were assigned to the novice trainer. On day 0 of training, the experienced trainer trained eight horses (4 pasture and 4 stalled) while the novice trainer trained four (2 pasture and 2 stalled). For the remainder of the study, including the loss of the pastured horse, the experienced trainer trained five horses (2 pasture and 3 stalled) while the novice trainer trained six (3 pasture and 3 stalled). Training Arena All training sessions were conducted at the MSU Horse Teaching and Research facility. Training took place in an indoor round pen (Fig. 2). The round pen area was approximately 15 m in diameter. The flooring material throughout the arena was a mixture of sand and clay. The round pen had two sliding doors that were used. The north sliding door was the entrance where the experimental horses were lead from their respective housing situation. The second door, which faced west, lead to a breeding room where the video recording equipment was stored. Adjacent to the breeding room was another room used as a laboratory, for immediate processing of samples. The animals had no contact with the adjacent rooms. 25 Fig. 2. Schematic diagram of round pen used in the study. C 2 cameras Equipment Training tack Equipment used for initial training of the horses included a lead rope, halter, saddle pad, bridle with a jointed snaffle bit and reins (Fig. 3). The lead rope was 1.8 m long and was attached to a ring on the halter with a clip. In this study, the lead rope was used for a variety of purposes including, controlling the horses while walking and standing, encouraging forward movement and as reins for the first day of training. The halter is designed to fit on the horses’ head so the trainers can control the horses as they lead them. The saddle pad is placed on the horses” back before the saddle. 26 Fig. 3. Photograph of equipment used during training. The pad is placed near the withers and extends down towards the center of the shoulder (C.H.A., 1996'). The purpose of the saddle pad is to prevent chaffing or discomfort from the saddle while riding. Both trainers used a western style saddle. Bridles are used for steering or directing the horses. The snaffle bit allows the trainer to communicate with and control the horses through pressure (C.H.A., 1996). A drawing of a bridle with a snaffle bit can be found in Fig. 4. The snaffle bit consists of two rings joined by a smooth mouthpiece, which is jointed in the center. The snaffle bit applies pressure on the mouth of the horses. Reins are run through the rings of the bit, which are controlled by the trainer. 27 THROAT IA'I'CH CHEEK PIECE \v.