UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Facial expression, vocalizations and eye temperature as potential indicators of pain in harbour seals… MacRae, Amelia Mari 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2019_february_macrae_amelia.pdf [ 3.17MB ]
Metadata
JSON: 24-1.0375838.json
JSON-LD: 24-1.0375838-ld.json
RDF/XML (Pretty): 24-1.0375838-rdf.xml
RDF/JSON: 24-1.0375838-rdf.json
Turtle: 24-1.0375838-turtle.txt
N-Triples: 24-1.0375838-rdf-ntriples.txt
Original Record: 24-1.0375838-source.json
Full Text
24-1.0375838-fulltext.txt
Citation
24-1.0375838.ris

Full Text

FACIAL EXPRESSION, VOCALIZATIONS AND EYE TEMPERATURE AS POTENTIAL INDICATORS OF PAIN IN HARBOUR SEALS  (PHOCA VITULINA) by Amelia Mari MacRae  B.A., University of Victoria, 1998 B.Sc., University of British Columbia, 2005 M.Sc., University of British Columbia, 2009  A THESIS SUBMITTED IN PARTIAL FULFILLMENT 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)  December 31, 2018  © Amelia Mari MacRae, 2018  ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled: Facial Expression, Vocalizations and Eye Temperature as Potential Indicators of Pain in Harbour Seals (Phoca vitulina)  submitted by Amelia Mari MacRae in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Applied Animal Biology Examining Committee: Dr. David Fraser, Applied Animal Biology Supervisor  Dr. Marina von Keyserlingk, Applied Animal Biology Supervisory Committee Member  Dr. Kenneth Craig, Psychology Supervisory Committee Member Dr, Ronaldo Cerri, Animal Reproduction University Examiner Dr. Jeffrey Richards, Zoology University Examiner  iii Abstract Pain assessment in animals typically relies on a combination of physiological and behavioural measures.  Unfortunately, many of these measures require handling the animals or physical sampling, which are invasive and can result in a stress response. The aim of this thesis was to identify some possible non-invasive indicators of pain in harbour seals as no clear species-specific indicators had been established. I investigated whether or not seals showed changes in facial expression and other behaviours (Chapter 2), vocalizations (Chapter 3), or eye temperature (Chapter 4) in response to the routine procedures of flipper-tagging and microchipping. Seals showed changes in facial expression and in several other behaviours in response to the procedures. Most notably, orbital tightening increased from before to after tagging and microchipping (p < 0.001), whereas the behaviours of looking around (p < 0.01) and struggling (p < 0.05) decreased. Sham treatment produced no similar changes (Chapter 2). The number of vocalizations increased from before tagging to after (p < 0.001) and the peak frequency increased from 837.1 ± 75 Hz before to 1041 ± 75 Hz after (mean ± SEM; p < 0.01). Similarly, there were more vocalizations after chipping than before (p < 0.001), and peak frequencies of the calls increased from 848.8 ± 79 Hz before to 1111.2 ± 79 Hz after (mean ± SEM; p < 0.05). No similar changes in vocalizations were seen after sham treatments (Chapter 3). Lastly, seals’ eye temperature increased after tagging but not after sham-tagging (p < 0.05), suggesting that a rise in eye temperature may reflect pain. However, eye temperature also increased in response to handling and an injection of lidocaine, suggesting change in eye temperature is non-specific to pain. Lidocaine, at the dosage used, did not appear to have a mitigating effect on the pain  iv from tagging and chipping (Chapter 4). These results show promise for the use of facial expressions and other behaviours, including vocalizations, to assess potentially painful procedures in seals.  Similarly, the use of eye temperature has potential to indicate a stress response and to evaluate the potential aversiveness of routine procedures in this species.   v Lay Summary  Very little published work exists on pain in harbour seals and before this research there were no reliable indicators of their pain. The objective of this thesis was to provide an initial evaluation of the use of facial expression, vocalizations and eye temperature as three potential non-invasive measures of pain in this species. In response to the routine painful procedures of flipper-tagging and microchipping, seals showed at least one change in facial expression (increased orbital tightening). They also increased the number and peak frequency (Hz) of their vocalizations in response to the procedures. Eye temperature increased in response to the procedures but also to handling. All three measures show promise for the assessment of pain in seals, and changes in eye temperature may also be useful for evaluating the potential aversiveness of routine procedures. This work contributes to the welfare of seals by improving understanding of their pain responses.    vi Preface All of the work presented here was conducted at the University of British Columbia’s Animal Welfare Program and all data were collected at the Vancouver Aquarium Marine Mammal Rescue, Vancouver, Canada. All projects and associated methods were approved by the University of British Columbia’s Animal Care Committee (Protocol A12-0095), (Protocol A16-0175) and by the Vancouver Aquarium Animal Care Committee.   A version of Chapter 2 has been published: MacRae, A.M., Makowska, I. J. and Fraser, D. 2018. Initial evaluation of facial expressions and behaviours of harbour seal pups (Phoca vitulina) in response to tagging and microchipping. Applied Animal Behaviour Science, 205: 167-174. A.M. MacRae developed and researched the main ideas for this paper and managed all aspects of data collection. Dr. David Fraser supervised, helped develop ideas, interpret material and edited drafts. Dr. I. Joanna Makowska helped develop ideas, interpret materials, collect data and edited drafts.  A version of Chapter 3 has been submitted for publication: MacRae, A.M., Makowska, I. J., and Fraser, D. Vocal changes as indicators of pain in harbour seal pups (Phoca vitulina). A.M. MacRae developed and researched the main ideas for this paper and managed all aspects of data collection. Dr. David Fraser supervised, helped interpret material and edited drafts. Dr. I. Joanna Makowska helped collect data and edited drafts.   vii A version of Chapter 4 has been submitted for publication: MacRae, A.M., Daros, R.R., McGreevy, P., and Fraser, D. Can surface eye temperature be used to indicate a stress response in seals (Phoca vitulina)? A.M. MacRae developed and researched the main ideas for this paper and managed all aspects of data collection.  Dr. David Fraser supervised, helped develop ideas, interpret material and edited drafts. R.R Daros helped with statistical analysis. Dr. Paul McGreevy (University of Sidney, Australia) helped develop ideas and edited drafts.   viii Table of Contents  Abstract .......................................................................................................................... iii Lay Summary .................................................................................................................. v Preface ........................................................................................................................... vi Table of Contents ........................................................................................................ viii List of Tables ............................................................................................................... xiii List of Figures .............................................................................................................. xiv List of Abbreviations ................................................................................................... xvi Acknowledgements ................................................................................................... xviii Dedication ..................................................................................................................... xx 1.  Introduction ............................................................................................................... 1 1.1 Recognizing and assessing pain in animals ........................................................... 1 1.1.1 Evolving attitudes towards pain ....................................................................... 1 1.1.2 Importance and difficulties of pain assessment ............................................... 4 1.1.3 Terminology and mechanisms: pain, stress, distress ...................................... 7 1.1.4 Pain .................................................................................................................. 7 1.1.5 Stress and the stress response ..................................................................... 11 1.1.6 Distress .......................................................................................................... 12 1.1.7 Assessment of pain in animals ...................................................................... 13 1.1.7.1 Physiological responses to pain .............................................................. 13 1.1.7.1.1 Measures of SNS response .............................................................. 13 1.1.7.1.2 Measures of HPA response .............................................................. 16  ix 1.1.7.2 Behavioural measures of pain ................................................................. 20 1.1.7.2.1 Pain-related behaviours .................................................................... 21 1.1.7.2.2 Change in the frequency or magnitude of certain behaviours .......... 23 1.1.7.2.3 Choice or preference behaviours ...................................................... 24 1.1.7.2.4 Cognitive bias ................................................................................... 26 1.1.8 Administration of analgesia ............................................................................ 27 1.1.9 Pain-scoring systems ..................................................................................... 29 1.1.10 Limitations of behavioural measures ........................................................... 32 1.1.11 Possible factors affecting the pain response ............................................... 32 1.1.12 Conclusions ................................................................................................. 36 1.2 Facial expressions indicative of pain .................................................................... 36 1.2.1 Introduction .................................................................................................... 36 1.2.2 The human ‘pain face’ .................................................................................... 37 1.2.3 The primacy of the face ................................................................................. 38 1.2.4 Measuring facial changes .............................................................................. 39 1.2.4.1 The Facial Action Coding System (FACS) .............................................. 39 1.2.4.2 The Neonatal Facial Coding System (NFCS) .......................................... 40 1.2.4.3 Pain faces in non-human animals ........................................................... 41 1.2.4.4 Grimace scales ........................................................................................ 41 1.2.4.5 Measuring acute versus long-term pain .................................................. 43 1.2.5 Conclusions ................................................................................................... 45 1.3 Vocalizations as potential indicators of pain ......................................................... 45 1.3.1 Introduction .................................................................................................... 45  x 1.3.2 Honest signals of need .................................................................................. 47 1.3.3 Vocalizations as indicators of pain ................................................................. 49 1.3.4 Conclusions ................................................................................................... 52 1.4. Infrared thermography as a non-invasive measure of stress and pain in animals .................................................................................................................................... 52 1.4.1 Introduction .................................................................................................... 52 1.4.2 Applications of IRT ......................................................................................... 53 1.4.3 Peripheral vasoconstriction and affective state .............................................. 56 1.4.4 Changes in eye temperature .......................................................................... 57 1.4.5 Eye temperature and the stress response ..................................................... 58 1.4.6 Advantages and limitations ............................................................................ 62 1.4.7 Conclusions ................................................................................................... 63 1.5 Thesis aims........................................................................................................... 64 2. Initial evaluation of facial expressions and behaviours of harbour seal pups (Phoca vitulina) in response to tagging and microchipping .................................... 66 2.1 Introduction ........................................................................................................... 66 2.2 Materials and methods ......................................................................................... 69 2.2.1 Animals and housing ...................................................................................... 69 2.2.2 Tagging and chipping ..................................................................................... 70 2.2.3 Development of ethogram .............................................................................. 71 2.2.4 Experiment 1 – FAUs and behaviours before and after tagging and chipping ................................................................................................................................ 71 2.2.5 Experiment 2 – Cross-over experiment ......................................................... 72  xi 2.2.6 Experiment 3 – Analgesic pilot study ............................................................. 72 2.2.7 Video analysis ................................................................................................ 73 2.2.8 Statistical analysis .......................................................................................... 74 2.3. Results ................................................................................................................. 75 2.3.1 Ethogram ....................................................................................................... 75 2.3.2 Experiment 1 – FAUs and behaviours before and after tagging and chipping ................................................................................................................................ 78 2.3.3 Experiment 2 – Cross-over experiment ......................................................... 81 2.3.4 Experiment 3 – Analgesic pilot study ............................................................. 84 2.4. Discussion ........................................................................................................... 84 2.5. Conclusions ......................................................................................................... 90 3. Vocal changes as indicators of pain in harbour seal pups (Phoca vitulina) ...... 92 3.1 Introduction ........................................................................................................... 92 3.2 Materials and methods ......................................................................................... 95 3.2.1 Animals and housing ...................................................................................... 95 3.2.2 Tagging and chipping ..................................................................................... 96 3.2.3 Experiment 1 – Vocalizations before and after tagging and chipping ............ 97 3.2.4. Experiment 2 – Cross-over experiment ........................................................ 97 3.2.5. Vocal analysis ............................................................................................... 98 3.2.6. Statistical analysis ....................................................................................... 100 3.3 Results ................................................................................................................ 101 3.3.1 Experiment 1- Vocalizations before and after tagging and chipping ............ 101 3.3.2 Experiment 2 – Crossover experiment ......................................................... 102  xii 3.4 Discussion .......................................................................................................... 105 3.5 Conclusions ........................................................................................................ 108 4. Can surface eye temperature be used to indicate a stress response in seals (Phoca vitulina)? ........................................................................................................ 109 4.1 Introduction ......................................................................................................... 109 4.2 Materials and methods ....................................................................................... 112 4.2.1 Animals and facilities ................................................................................... 112 4.2.2 Injection and tagging protocol ...................................................................... 113 4.2.3 Eye temperature measurements .................................................................. 114 4.2.4 Treatments ................................................................................................... 116 4.3 Statistical analysis and results ............................................................................ 119 4.3.1 Eye temperature changes after handling ..................................................... 120 4.3.2 Eye temperature changes after injection ..................................................... 122 4.3.3 Eye temperature changes after tagging or sham-tagging ............................ 124 4.3.4 Eye temperature changes after attempted pain management ..................... 126 4.4 Discussion .......................................................................................................... 126 4.5 Animal welfare implications and conclusions ..................................................... 130 5. General Discussion and Conclusions .................................................................. 132 5.1 Thesis findings .................................................................................................... 132 5.2 Discussion and conclusions ............................................................................... 137 References .................................................................................................................. 144   xiii List of Tables  Table 2.1. Ethogram of all facial action units (FAUs) and behaviours observed before and after tagging and chipping of harbour seal pups……………………………….….….76  Table 2.2. Scores for facial action units and other behaviours (mean ± SEM) before and after sham or actual tagging and chipping for the 9 seals in Experiment 2………..……83    xiv List of Figures  Figure 2.1. Facial action units of harbour seals (Phoca vitulina) recorded in this study. [Blinking and vocalizing are not included. Detailed descriptions of each FAU are included in Table 2.1]………………………………………………………………………...77  Figure 2.2. Experiment 1 – Facial expression of one seal before (left) and after (right) tagging and chipping. [The two photos are stills taken from the pup’s 90-s video clips]…………………………………………………………………………………………….78 Figure 2.3. Scores for facial expression and other behaviour (mean ± SEM) before and after tagging and chipping for the 19 seals in Experiment 1. [*** p ≤ 0.001, ** p ≤ 0.01]…………………………………………………………………………………………….80 Figure 3.1. Example of spectrographic representation of one harbour seal pup’s vocalizations for a 5-s period created with Raven Pro Interactive Sound Analysis Software (Version 1.5)……………………………………………………………………….99  Figure 3.2.  Number of calls of harbour seal pups (n = 10) in the 30 s before and after tagging or sham-tagging. [A: Seals that received sham-tagging on Day 1 and tagging on Day 2; B: seals that received the reverse order of treatments]………………………..104  Figure 4.1. Example of a thermal image of a seal’s eye. [Maximum eye temperature is indicated by the dark red triangle]…………………………………………………………115   xv Figure 4.2. Pictorial representation of the four treatments (n= 13 seal pups per treatment) over the different periods from Baseline (before any handling began) to Post 60 (60 min after all handling had ended).  [The dashed lines indicate periods when pups were not handled and the solid lines indicate periods when pups were restrained]…………………………………………………………………………………… 117  Figure 4.3. Examples of the change in eye temperature ° C (mean ± SE) showing four seals before (Period A) and after (Period B) receiving an injection of lidocaine (Lidocaine treatment)………………………………………………………………………..122   Figure 4.4. Examples of change in eye temperature ° C showing five seals in the Tag Only treatment before (Period C) and after (Period D) the animals were received flipper tags…………………………………………………………………………………………….124   xvi List of Abbreviations ACTH – adrenocorticotropic hormone  ANS – autonomic nervous system APP – acute phase protein AU – action unit bCRH – bovine corticotropin-releasing hormone CBG – corticosteroid-binding globulin CRH – corticotropin-releasing hormone FACS – Facial Action Coding System FAU – Facial Action Unit HF – high frequency HGS – horse grimace scale HPA – hypothalamic-pituitary-adrenal HRV – heart rate variability IBI – inter-beat interval IRT – infrared thermography LF – low frequency LF:HF ratio – low frequency to high frequency ratio LMM – linear mixed model MGS –mouse grimace scale NFS – Neonatal Facial Coding System NRS – numerical rating scale NSAID – nonsteroidal anti-inflammatory  xvii PNS – parasympathetic nervous system RMSSD – root mean square of successive differences SDS – simple descriptive scale SNS – sympathetic nervous system TMD – temporomandibular disorder VAS – visual analogue scale   xviii Acknowledgements I would like to express my deepest gratitude to David Fraser, my supervisor.  His guidance, humour, seemingly endless knowledge and patience have been a constant source of inspiration to me. I feel extraordinarily privileged to have enjoyed his mentorship both through my MSc and PhD experiences. I am grateful, too, for the support and guidance of my supervisory committee members. I thank Ken Craig for his insightful feedback and encouragement, and Nina von Keyserlingk for her expertise, as well as her advice and emotional support throughout the years. I feel truly fortunate to have had the guidance of such a formidable group of people. I also appreciate the helpful input and feedback from Dan Weary and the energy and enthusiasm he brings to our program.  An additional thanks to Chris McGill for his good humour and assistance in all things program related.  My experience in the Animal Welfare Program was greatly enriched by the varied backgrounds and passions of my fellow students. I especially want to thank Joanna Makowska for her exceptional generosity with feedback, even when busy with her own projects, but mostly I thank her for her friendship. I could not have asked for a better collaborator. I am grateful too, to Kristen Walker for her support and for blazing a trail with her research on pinniped pain. Thank you to Liv Baker for her constant encouragement, friendship and always being up for adventure, to Cathy Schuppli for her support, friendship and for accepting many late-night calls about sick animals and to Gosia Zobel and Elisabeth Ormandy for their feedback and advice. I also want to thank Erin Ryan for the laughs and chats about all things, and for all her help and support in the final stages of my thesis. I have been extraordinarily lucky to have formed  xix friendships with these incredible people, and I am truly inspired by their passion and dedication to animal welfare. I owe a special thanks to the wonderful team at the Vancouver Aquarium Marine Mammal Rescue, especially Lindsaye Akhurst, Emily Johnson, Martin Haulena, Shanie Fradette, Belinda To, Jenelle Leedam, Sion Cahoon and Lisa B. who were central to the fruition of this research. I also thank the many other staff and volunteers who take such amazing care of the animals. My deep gratitude goes to Duncan MacRae, Lauren Cee, Janet Bauer, Kathryn Gibb, and Tiffany Lin for their work in data collection and data management, as well as Ruan Daros, who was always willing to assist with R. I acknowledge my receipt of doctoral funding from NSERC CGS-D and UBC’s four-year fellowship, which enabled me to pursue research in my areas of interest. I especially want to thank my amazing family for their unconditional love and encouragement. I am grateful to my Mom and Dad for always supporting me and helping in more ways than I can list, my O.G. for all the laughs and adventures, Loree and Brian for their constant support and care of Lilah, and to Lilah for showing me all that is truly important by being the most wonderful daughter. I thank my husband, Robin, for his years of love and patience, not to mention hours of editing. In many ways, this achievement also belongs to him. To Pinky, Cash and Brynn, thank you for your faithful companionship.  Last but not least, I am grateful to the seals for affording me the motivation to continue my work.   xx Dedication      To my family and to the seals  1 1.  Introduction  1.1 Recognizing and assessing pain in animals   1.1.1 Evolving attitudes towards pain Pain is a complex phenomenon with physiological, subjective and behavioural dimensions (Williams and Craig, 2016). The emotional aspect of pain is suggested to be fundamental to the pain experience (Le Bars and Cadden, 2005).  However, opposition to the idea that animals feel pain has existed for centuries with an intense controversy persisting to this day (e.g., Bermond, 2001; Rose et al., 2014). Beliefs about animals influence how we treat them (Fraser, 2008). It follows that improvements to their welfare largely depend on the societal beliefs and values related to issues such as our understanding of the nature of pain in animals. Central to the discussion, and perhaps informing the varying attitudes toward animals, is the question of whether or not animals have the capacity to consciously experience such negative states as hunger, fear and pain (Dawkins, 2017; Weary et al., 2006). While in recent years, the issue of animal pain has become a central consideration for those concerned with the welfare of animals (Bateson, 1991), animals in our care continue to be frequently subjected to potentially painful procedures as part of routine management.  It is difficult for many people to understand how there could be doubt about animals’ capacity to experience pain. However, when considering that not long ago, only privileged members of human society were considered deserving of pain management, it is unsurprising that there is still confusion and disagreement over attempts to assess pain in non-human species. For example, it is documented that as   2 recently as the 1860s in Pennsylvania more than 30% of human limb amputations were done without anesthesia.  While anaesthetic drugs were readily available and inexpensive, they were not employed for the poor, uneducated, alcoholics, people of colour or recent immigrants, as these groups were considered insensitive to pain (Pernick 1885 cited in Phillips, 1993). Although in the assessment of pain and delivery of treatment there is still bias related to socioeconomic status, as well as sexism, racism and ageism (Narayan, 2010), it would now be unthinkable to subject a person to major surgery without first administering anaesthetic.  Interestingly, views on pain in infants and young children are still evolving. While some recently argued that neonates have reduced pain sensitivity (Derbyshire, 2003), others conclude that the importance of early pain experience in an individual’s development demands that pain be rigorously addressed and managed (Anand and Craig, 1996). Unfortunately, there is much evidence that pain care in infants and young children remains inadequate (Allegaert and van den Anker, 2016). Attitudes toward animal pain are undergoing similar shifts. Examination of anaesthetic use in animal surgeries over the last century follows a similar pattern to its use in underprivileged humans. In the 19th century, British antivivisectionists actively campaigned to overcome the widely held belief that animals were insensitive to pain and therefore did not require anesthesia for surgery (Pernick 1885 cited in Phillips, 1993). Although the use of anaesthetics is now routine, the management of animal pain has not yet attained the same standard as it has in humans. There have been dramatic advances in our understanding and treatment of animal pain (Mogil et al., 2010). However, reports on the use of analgesics over the last several   3 decades suggest their administration is still infrequent and haphazard and seems to vary by species, by procedure and by clinician. Despite the acknowledgement of most veterinary surgeons that animals have the capacity to feel pain in many circumstances (e.g., after surgery) only 50% of dogs and cats in Canada were reported to receive analgesia after ovariohysterectomy in the 1990s (Dohoo and Dohoo, 1996a, 1996b). A study on the administration of analgesics in dogs and cats by French veterinarians showed rates ranging from as low as 17% for castration up to 84% for orthopedic surgeries (Hugonnard et al., 2004). Surveys regarding views on equine pain after castration revealed professional disagreement on the question of whether castration is in fact painful for horses, and that the identification of pain, as well as the application of analgesics, lacked uniformity (Price et al., 2005, 2002). The administration of analgesics in laboratory animals after painful procedures has been shown to be much less frequent in small rodents such as rats and mice than in larger mammals such as rabbits, sheep, dogs, pigs and non-human primates (Coulter et al., 2009; Stokes et al., 2009). In fact, only 3% of studies in peer-reviewed journals published between 1990 and 1992 reported the administration of analgesics to mice and rats post-surgery. Subsequent follow-up with the researchers who did not report the use of analgesics revealed that 71% of them had not, in fact, administered any analgesic drugs reportedly because there were no observable signs of pain (35%) and belief that analgesics were unnecessary (35%) (Richardson and Flecknell, 2005). Rates of analgesic use increased to 10% by 2000-2001 and 20% by 2005-2006 (Stokes et al., 2009). These rates were significantly lower, however, than the 50% in 2000-2001 and 63% in 2005-2006 reported for larger species (Coulter et al., 2009).   4 Reportedly, the main reasons for the discrepancies in administering pain-relief drugs were difficulties in recognizing animal pain and uncertainty as to the appropriate drug therapies (Hugonnard et al., 2004). In fact, it seems likely that the frequent failure to adequately address animal pain often stems from our poor ability to assess it (Flecknell, 2000). Clearly, there is a need to better understand animal pain. Therefore, the continued development of reliable methods for its identification and evaluation in a wide range of species is required.  1.1.2 Importance and difficulties of pain assessment Pain perception is essential to an animal’s fitness and survival, warning of potential harm and motivating individuals to protect themselves from injury. The negative experiential aspect of pain is likely adaptive in that it helps animals learn to avoid potentially dangerous circumstances, and arguably, cognitive engagement is required if long-term fitness benefits are to be conveyed (Brown, 2016; Sneddon et al., 2014). However, despite conferring protective benefits, pain is unpleasant and can have consequences if left untreated. It can impede biological function and lead to an animal’s inability to experience normal pleasures (anhedonia). Pain can affect quality of life, impair multiple cognitive functions (including attention, memory, motivation, and sleep), affect social interaction, and create negative affective states such as depression and anxiety (reviewed by Mogil, 2009). The impact of pain on an individual’s well-being, and the reliance of effective pain management on accurate assessment, make the detection and evaluation of pain fundamentally important for animal welfare. Further, the reliable assessment of animal pain has important implications for human medicine, as animal   5 models are widely used in studies on human pain and for the development and evaluation of analgesics (Mogil, 2009). Accurate identification and assessment of pain has posed a challenge in both human and non-human animals, but is particularly problematic in animals.  As the pain experience is subjective, it can only be inferred, as opposed to directly measured (Livingston and Chambers, 2000). Noxious insult triggers an extensive range of physiological and behavioural reactions and the challenge lies in distinguishing those that reflect non-conscious reaction from those that indicate an experience of pain. In humans, we largely rely on our ability to verbally communicate pain status (Anand and Craig, 1996), but as animals cannot self-report, other methods are needed. Further confounding our ability to recognize animal pain is that the affective aspect of pain is key to the pain experience.  Our difficulty in distinguishing unconscious detection of noxious stimuli from conscious pain is due to our inability to access animals’ internal mental states (Broom, 1998). Determining which measures actually reflect pain is additionally challenged by the fact that the pain response may vary according to several factors: (1) the type, location or severity of damage; (2) differences among individuals and between species; (3) differences between acute and chronic conditions or different disease stages; and (4) changes in the intensity and nature of pain over time (Livingston and Chambers, 2000).   Currently, there is no perfect method for the evaluation of pain. Accurate assessment requires that tools be valid (measuring what they purport to measure) and reliable (producing stable and consistent results). Weary et al. (2006) suggest that   6 measures of responses to painful stimuli can be validated by examining responses with (P) and without (p) the painful condition, and with (A) and without (a) analgesics known to be effective for mitigating the painful condition. A useful measure should effectively differentiate animals with the painful condition (Pa) from the other conditions (PA, pA, pa) (Weary et al., 2006).  In human medicine and psychometrics, assessment tools are also expected to meet criteria of sensitivity and specificity. The sensitivity of a test (also called the true positive rate) refers to the correct identification of the proportion of individuals with a condition (e.g., pain), and specificity (also called the true negative rate) refers to the proportion of healthy individuals who are correctly identified as not having the condition (Altman and Bland, 1994). An ideal indicator of pain would be specific to pain such that it would be expressed only when pain was present, exist across a range of pain modalities, distinguish between nociceptive reflexes and the affective experience of the pain, and inform about the type, intensity and duration of pain, all while being practical to apply. The challenge with many of the parameters used to assess pain is that, while they may be sensitive to a painful condition, they lack specificity. For example, a response being measured may be evoked by pain but may also occur in response to non-painful events or situations (e.g., heart rate will accelerate in response to noxious stimuli as well as to fear, exertion etc.) (Rietmann et al., 2004). This lack of specificity could lead to the assumption that pain is present when in fact it is not (false positive). Despite the challenges, developing reliable assessment schemes that enable the identification of the occurrence of pain, as well as the quantification of its intensity, is critically important to its appreciation, understanding and management.    7 1.1.3 Terminology and mechanisms: pain, stress, distress Before proceeding with a discussion of the various measures that have been used to assess animal pain, the following section will attempt to clarify briefly some of the terminology that will be used throughout this paper and, as pain and stress responses are physiologically complex, provide a simplified overview of the general processes that are involved.  1.1.4 Pain There is currently no single, universally accepted definition of animal pain. The International Association for the Study of Pain (IASP) defines pain as “an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage” (International Association for the Study of Pain, 1979).  Following this definition is a note explaining that the IASP purposely avoids tying pain to any specific stimulus.  Although it often has a proximate physical cause, pain is a subjective, psychological state, and it is the unpleasant sensory experience itself that is defined as pain (International Association for the Study of Pain, 1979). The IASP’s definition was devised to apply to human pain but is problematic when applied to animals (or nonverbal humans) as it has been tied to the ability to self-report (Anand and Craig, 1996). Williams and Craig (2016) propose an updated definition that reads: “Pain is a distressing experience associated with actual or potential tissue damage with sensory, emotional, cognitive and social components.” They argue that this revised definition reflects the advancements in our understanding of pain by acknowledging cognitive and social components, by recognizing that the word “unpleasant” is likely   8 inadequate for characterizing the experience of severe pain, and by no longer prioritizing self-report at the expense of other non-verbal behaviours (Williams and Craig, 2016). The revision by Williams and Craig (2016) better applies to non-human species than the original IASP definition but, for the purposes of this paper, I will use the working definition of pain proposed by Molony and Kent (1997), which also avoids the need for self-report and assumes the capacity of animals to affectively experience pain. They describe pain as “an aversive sensory and emotional experience representing an awareness by the animal of damage or threat to the integrity of its tissues; it changes the animal’s physiology and behavior to reduce or avoid damage, to reduce the likelihood of recurrence and to promote recovery; unnecessary pain occurs when the intensity or duration of the experience is inappropriate for the damage sustained or when the physiological and behavioral responses to it are unsuccessful at alleviating it” (Molony and Kent, 1997).  The term nociception is sometimes used interchangeably with the term pain. However, nociception refers to the activation of the peripheral and central nervous system in response to noxious stimuli but does not imply the conscious, subjective aspect of the pain experience (Gregory, 2004). Nociceptors are the sensory receptors that respond to noxious stimuli (mechanical, chemical, thermal). They are free nerve endings, which vary in amount and location but exist in most tissue, both on the skin and internally. Noxious stimuli detected at the periphery will excite the nociceptors, causing them to depolarize, which generates nerve impulses (action potentials). The nerve impulses are transferred along the afferent nerve fibers where they first synapse at the dorsal horn of the spinal cord. The input is then relayed via the spinal cord to   9 multiple sites of the brain where it is processed, resulting in various bodily responses (Livingston and Chambers, 2000). Two main types of afferent nerve fiber facilitate different pain sensations and often work in conjunction: myelinated A-delta fibers associated with sharp, mechanical stimuli, and unmyelinated C-fibers, which evoke dull, burning, longer-lasting pain (Livingston and Chambers, 2000). Reflex withdrawal which is controlled at the spinal cord, and autonomic responses which are controlled in the brain, result from the stimulation of nociceptive fibers; however, their activation does not necessarily result in the affective experience of pain (Clark, 1994). The local release of chemical mediators and cytokines (proteins released from immune cells that promote inflammation) can sensitize nociceptors so that their threshold for activation is lowered (Zhang and Jianxiong, 2007). The lowered threshold to noxious stimuli is called hyperalgesia. Primary hyperalgesia is the increased sensitivity at the site of injury and secondary hyperalgesia occurs when sensitivity is increased in the surrounding uninjured area. Allodynia is a type of hyperalgesia in which sensation is altered so that innocuous stimuli are registered as noxious and provoke a pain response (e.g., gentle stroking of the skin is interpreted as painful) (Gregory, 2004).  Pain differs depending on its type, site of origin, and duration. Somatic pain originates in the periphery as either ‘superficial’ at skin level, or as ‘deep’ pain in muscles, bones or joints, and is usually well localized (Molony and Kent, 1997). Visceral pain, on the other hand, originates in the viscera, is more diffuse – often described as more unpleasant than somatic pain – and can be referred to other areas of the body (felt remotely from the site of tissue damage) (Paine et al., 2009). Neuropathic pain   10 signifies pain originating from a pathology of the nervous system (Campbell and Meyer, 2006). Acute pain typically relates to specific tissue damage, occurs at the time of injury, is of short duration and resolves once the damage is healed. It is usually accompanied by autonomic changes (e.g., increased heart-rate, respiration rate, blood pressure), may evoke escape or avoidance behaviours and responds well to analgesics (Dobromylskyj et al., 2000b). In most cases, acute pain is considered to have adaptive value as it functions to warn of potential injury, evoking withdrawal from and avoidance of future stimuli. Acute pain is followed by sub-acute pain, which is the phase that results in protective behaviour that aids in recovery; it is this phase that may develop into chronic pain (Bateson, 1991). Chronic pain is more difficult to define. The term ‘chronic’ can be ambiguous in how it is applied to describe pain. As opposed to the term reflecting the actual nature of the pain, it may refer to the prolonged time period during which an individual experiences the pain or refer to the difficulty of managing the pain. In general, chronic pain is considered to persist after the time of expected healing, may exist in the absence of obvious tissue damage and be of an intensity that does not parallel the severity of the pathology (Dobromylskyj et al., 2000a). Chronic pain may be more difficult to recognize than acute pain as indicators may be more subtle and less specific (e.g., sleep disturbance, appetite suppression, depression or long-term changes in behaviour). Unlike acute pain, chronic pain is thought to be without adaptive value (Brearley and Brearley, 2000).    11 1.1.5 Stress and the stress response Stress is a commonly used term that nonetheless has always been fraught with ambiguity. In 1936, Hans Seyle first defined the term stress as the non-specific bodily response to any demand for change.  However, as Seyle’s theories garnered interest, it became clear that the term ‘stress’ was problematic as it had become a catch-all for any type of unpleasant threat. Throughout his career Seyle tried to clarify this definition to more accurately reflect the phenomena he was describing, but it remains a term that can have different meanings in different contexts (Fraser, 2008). To avoid the confusion, in this paper I will not use the term stress but instead will refer to stress responses. Under normal circumstances an individual’s body rests in a state of physiological balance, or homeostasis. Bateson (1991) proposed a stressor is anything that upsets that balance, and the stress response is the set of bodily processes that work to reestablish the homeostatic state. The stress response is dominated by the activation of the hypothalamic-pituitary adrenal axis (HPA) and sympathetic nervous system (SNS) and can be initiated by a multitude of physiological or psychological stressors. The SNS is the first to respond to a stressor with the release of catecholamines: norepinephrine from the nerves of the SNS and epinephrine from the adrenal medulla.  The activation of the SNS involves a number of physiological responses that prepare the body for a ‘fight or flight’ response, such as pupil dilation, changes in blood pressure, an increases in heart rate, respiration rate and temperature, the increase of blood flow to skeletal muscles, the mobilization of stored energy (glycogen from liver, fatty acids from adipose tissue) and a decrease or cessation of digestive processes (Morton and Griffiths, 1985; Sapolsky, 2002).    12 The activation of the SNS is followed by the cascading HPA response. When the HPA axis is triggered, the hypothalamus releases corticotropin-releasing hormone (CRH), which stimulates (within ~15 seconds) the anterior pituitary gland to synthesize and release adrenocorticotropic hormone (ACTH) into the blood, which within minutes results in the cortex of the adrenal gland releasing glucocorticoid hormones (cortisol and corticosterone) (Sheriff et al., 2011). Corticosteroids affect metabolism and mediate the body’s inflammatory response. At a basal level, the secretion of glucocorticoids supports essential life processes such as the conversion of sugar, fat and protein stores into usable energy (i.e., basic energy regulation). However, when the system is stimulated in response to a challenge, the secretion of glucocorticoids increases well above basal levels (Landys et al., 2006; Sapolsky et al., 2000).   1.1.6 Distress The term distress, which is often used in conjunction with the term pain, can have different meanings in different contexts but in general is used to encompass various unpleasant experiences such as anxiety or boredom (Brown et al., 2006).  A panel of welfare experts came together to create The Report of Working Group on Animal Distress in the Laboratory published in 2006, in which they addressed the difficulties and confusion in defining the term distress.  They noted that it is important to recognize that distress is not necessarily related to (or a consequence of) pain, as the alleviation of one will not necessarily affect the other. They determined “distress might therefore be considered as one or more negative psychological states indicative of poor well-being and diminishing to an animal’s well-being and quality of life” and that “distress is   13 therefore multi-modal, and its extent can vary depending on both the duration and intensity of the stimulus causing distress, as well as the context in which it occurs” (Brown et al., 2006).   1.1.7 Assessment of pain in animals  Although some expressions of pain likely may be conserved among species  (e.g., facial expression) (Williams, 2002), the assessment of pain is species-specific and often relies on the interpretation of behavioural and physiological indicators. Subjective and objective measures of pain have been developed for various species and usually focus on general body function (such as food and water intake), physiological responses (such as heart rate, blood pressure, biochemical changes), and behaviour (such as body postures, vocalizations) (Weary et al., 2006). Problematically, many of the indicators used to assess pain are not necessarily pain-specific, as they may present in other contexts that do not involve painful conditions.  1.1.7.1 Physiological responses to pain  1.1.7.1.1 Measures of SNS response Many of the physiological measures of pain reference the stress response system (SNS and HPA). The activation of the sympathetic nervous system, subsequent release of catecholamines and the resulting autonomic changes in heart rate, blood pressure, pupillary diameter, respiration rate and body temperature have all been used as indicators of pain (Morton and Griffiths, 1985). The quick response of the SNS to   14 stressors mean that changes in this system can be useful for the determination of short-lived acute pain, such as at the start of painful procedures (Mellor et al., 2000). For example, changes in catecholamine concentrations in blood have been recorded in response to castration and tail-docking in lambs and dehorning in cattle (Mellor et al., 2002). However, because catecholamines activate quickly, have a short half-life (1-2 minutes; Hjemdahl, 1993) and are rapidly removed from blood through enzymatic degradation, they must be sampled almost immediately. This brief sampling window can create practical challenges as well as difficulties ensuring these changes in concentration are not, at least in part, due to the sampling procedures themselves (Rutherford, 2002).  Heart-rate is also commonly used as a physiological parameter of ANS activity, and there is evidence of changes in heart rate in response to painful stimuli such as branding (Lay et al., 1992a; Walker et al., 2011b), castration, and tail-docking (Peers et al., 2002).  However, the relationship between heart rate and pain has not been clearly established. Many factors may influence heart rate, including fear, excitement, physical exertion, hydration status and endotoxemia (Rietmann et al., 2004). Despite its lack of validation as a metric, change in heart-rate is cited by veterinarians as a primary parameter in pain assessment (Price et al., 2002), likely because it is easy to measure and non-invasive.  Many researchers argue that, because heart rate varies between beats and reflects the net interactions of both the parasympathetic and sympathetic divisions of the autonomic nervous system (ANS), it is of limited use for the assessment of sympatho-vagal regulation. Heart rate variability (HRV) is suggested as a better   15 measure of the autonomic modulation of cardiac function, as it represents the balance between sympathetic nervous system (SNS) and parasympathetic nervous system (PNS) activity. HRV analysis uses the heart’s inter-beat interval  (IBI or R-R interval), which is the time interval between two successive beats or R waves of an electro-cardiograph (von Borell et al., 2007). Different types of stressors can shift the ANS to either sympathetic or parasympathetic dominance, and decreases in HRV measures typically indicate a shift to sympathetic dominance (Sgoifo et al., 1999, 1997).  The most widely used methods of HRV analysis are of two main types: time-domain analysis, which is based on inter-beat interval, and frequency-domain analysis, which counts the number of inter-beat intervals corresponding to assigned bands of frequency.  High-frequency (HF) bands are thought to represent parasympathetic activity whereas low-frequency (LF) bands integrate both parasympathetic (vagal) and sympathetic influences.  The LF/HF ratio has been demonstrated to be an accurate indicator of sympathetic activity during a number of physiological and physical stressors (von Borell et al., 2007).  For example, HRV indices have been used to assess the stress responses in horses exposed to novel objects (Visser et al., 2002) and in non-ambulatory dairy cows in response to flotation therapy (Stojkov et al., 2016). HRV has also been used to assess pain responses in horses suffering from laminitis (Rietmann et al., 2004) and in calves after surgical castration (Stewart et al., 2010).  von Borell et al. (2007) provide a review of HRV as a measure for assessing stress in farm animals. Blood pressure (Keating et al., 2012; Peers et al., 2002), body temperature  and respiratory rate (Hellebrekers et al., 1994) have also been used as possible indices, but there is debate about their reliability to determine the presence of pain (Conzemius et   16 al., 1997). Holton et al. (1998) suggested there is a relationship between pupil dilation and pain but concluded it was not a viable clinical indicator due to the difficulty in measuring diameter and because pupillary changes have other causes. Because autonomic responses are influenced by so many factors, it is unlikely any of these measures would be accurate as a sole indicator of the presence of pain.  Nonetheless, all of the above listed measures may have value when integrated into a multi-factorial pain scoring system (Livingston and Chambers, 2000).  There has been a rapid increase in the use of infra-red thermography (IRT) to detect changes in eye temperature as an indication of physiological stress responses in animals (Fenner et al., 2016). Proposed as a proxy for activity of the SNS and HPA axis, IRT has the appeal of allowing the estimation of the stress response from a distance, thus reducing operator effects, such as flight responses. It also seems possible that both SNS and HPA activation due to pain could successfully be measured using IRT. For example, Stewart et al. (2008b) recorded a significant drop in calves’ eye temperature in response to being disbudded without anaesthetic, and this was thought to be the result of sympathetically-mediated vasoconstriction. Following the initial temperature decrease, eye temperatures then increased above baseline for the remainder of the experimental sampling period, likely because of parasympathetic and HPA activity. Schaefer et al. (2007) also found temperature changes around the eyes of cattle in response to painful stimuli such as an electric prod. Measuring eye temperature in response to pain will be discussed in more detail in Section 1.4.  1.1.7.1.2 Measures of HPA response   17 Activation of the HPA response by painful stimuli is most commonly monitored via corticosteroid production. The initial release of glucocorticoids into the blood can be detected within a few minutes of exposure to a perceived stressor, and high levels may last for several hours (e.g., levels are elevated for 6 to 7 hours in calves after dehorning; Petrie et al., 1996). Mellor and Murray (1989) were the first to measure changes in plasma cortisol in lambs in response to castration and tail-docking, finding that levels were elevated for 60 minutes in animals that had undergone both procedures, as opposed to control animals that had been subjected only to restraint and blood sampling.   Subsequently, corticosteroids have been used in multiple species to assess various painful stimuli: celiotomy for colic in horses (Pritchett et al., 2003); branding of cattle (Lay et al., 1992a, 1992b); disbudding and dehorning of cattle (Stafford and Mellor, 2011); ischemia in rats (Kalliokoski et al., 2010) and tattooing of rabbits (Keating et al., 2012). Changes in cortisol concentration have been shown to closely correlate with changes in certain types of behaviour after painful procedures in some species (e.g., abnormal standing and lying positions, restlessness and rolling in lambs after tail docking and castration) (Kent et al., 1993) suggesting cortisol may be useful for signifying the presence of pain, as well as for helping to validate behaviours indicative of pain.    The measurement of glucocorticoids provides only an indirect measure of pain, but it may nonetheless reflect the noxiousness of the stimuli. For example, Lester et al. (1996) measured cortisol levels and behaviour after tail-docking of lambs via rubber ring versus surgical removal. Both methods resulted in a spike in cortisol concentration in the hour after the procedure.  However, whereas the levels of those animals docked by   18 rubber ring returned to normal soon after the procedure, the levels of those surgically docked remained high over the 4-hour observation period. Lester et al. (1996) suggest that changes in cortisol can therefore be useful for determining the degree of distress associated with different management procedures.    As with the other physiological indicators discussed, there are limitations to the use of corticosteroids to indicate pain. Cortisol is most effective when used to measure acute, moderate pain and may be inadequate to reflect severe pain due to a ‘ceiling effect’ where moderate and high levels of pain produce similar corticosteroid responses (Molony et al., 1997). Additionally, changes in cortisol levels are inconsistent in response to chronic pain (Ley et al., 1994).  Many factors affect the amount of circulating hormones. Their release is pulsatile; hence, the concentrations will vary depending on when sampling occurs. They may fluctuate seasonally or diurnally, be affected by diet, age, reproductive, social or health status as well as be different between species and individuals.  Elevated levels are associated with negative stimuli such as the presence of a predator, as well as positive stimuli such as mating (Moberg, 2000). It is therefore necessary to ensure that a baseline sample is collected, followed by multiple samples taken at appropriate intervals, to determine if changes in concentration are due to the stimulus the researcher is trying to measure or to some other influence (Molony and Kent, 1997). Unfortunately, obtaining blood samples in animals is invasive and requires restraint, which in itself can act as a severe stressor (Weary et al., 2006). Repeat venipuncture can cause acute pain and be a potential source of infection and inflammation at the injection site. A catheter can be placed for repeat samplings but is not always practical,   19 depending on species and test conditions. Corticosteroids can also be collected via saliva or feces, both of which methods are much less invasive. Within seconds of being released into the blood, the majority of cortisol becomes bound to proteins called corticosteroid-binding globulins (CBG). The remaining free cortisol (~5-15%) has low molecular weight and high lipid solubility so is able to enter bodily cells and fluids such as saliva (Kirschbaum and Hellhammer, 1999). Cortisol in blood plasma is usually measured as total cortisol (both bound and free fractions), and the unbound, free cortisol is what is measured in saliva. Salivary cortisol has been correlated to cortisol in the blood in several species, including swine, sheep, goats, horses, guinea pigs and dogs (reviewed in Cook, 2012). Although the method still needs validation for many species, sampling cortisol from saliva has the advantage of being non-invasive. While glucocorticoids in feces have also been successfully used as biomarkers of HPA activation, the inevitable delay from the onset of the stressor to the time of defecation may make such a measure impractical for many situations (Keay et al., 2006). Various attempts have been made to find biomarkers, other than corticosteroids, that may be indicative of pain, including glucose, free fatty acids and lactic acid (Peers et al., 2002), ACTH (Benson et al., 2000) and neuropeptide substance P (Coetzee et al., 2008). McCarthy (1993) attempted to use plasma B-endorphin to assess pain and stress in horses but considered it to be too non-specific to be useful. Some acute phase proteins (APP), which are primarily produced by the liver in response to infection or tissue damage, have been shown to increase in response to some painful conditions such as castration in horses (Kent and Goodall, 1991) and clinical mastitis in dairy cows   20 (Eckersall, 2000), but which particular APPs have relevance likely varies with species and with the specific injury or disease (Eckersall, 2000). Overall, none of these substances are considered to provide a practical measure of the pain response.  Because of the difficulties in interpreting physiological indices as measures of pain, behavioural indicators are increasingly used for both humans and non-human animals (Price et al., 2003).   1.1.7.2 Behavioural measures of pain Multiple behavioural indicators and concomitant changes in physical appearance have been used in attempts to identify and assess pain in animals (Morton and Griffiths, 1985). Physical appearance may vary due to changes in food and water intake or reduction of maintenance behaviour necessary for general health, such as grooming. Animals may show changes in their social behaviour such as increased aggression or withdrawal from social interactions, reduced exploratory behaviour, changes in locomotion, latency to feed, or other general deviations from their normal time budgets. Changes to normal or baseline behaviour, or the development of abnormal behaviours, have been used as indicators.  Examples include defensive aggression in cats and dogs (Hart et al., 1985), the change in demeanour and posture in cats after surgery (Stanway et al., 2002), lethargy in multiple species (Morton and Griffiths, 1985), the reaction to handling in lambs after castration (Thornton and Waterman-Pearson, 1999), changes in general activity level or type of activity in rats (Roughan and Flecknell, 2000) and lameness-related changes in gait in cattle (Sprecher et al., 1997).   21   Weary et al. (2006) suggest that behaviour used to assess pain can be classified into three types: 1) pain-specific behaviours, 2) a change in the frequency or magnitude of certain behaviours and 3) choice or preference behaviours.  Most behaviours used to assess pain are not technically pain-specific as they may displayed in contexts that do not involve painful conditions; therefore, they will be referred to as pain-related rather than pain-specific.   1.1.7.2.1 Pain-related behaviours Pain-related behaviours may present at the time of, or in the period after, painful stimuli. Several types of behaviour, in multiple species, have been observed as an immediate response to painful procedures. Examples include escape or avoidance behaviours, tail flicking, kicking, and falling during cattle branding (Lay et al., 1992a; Schwartzkopf-Genswein et al., 1998, 1997), escape behaviours and rearing by lambs and beef calves during castration (Rault et al., 2011), and the vocalization of piglets during castration (Taylor and Weary, 2000; White et al., 1995). Other behaviours that can still be considered pain-related may not immediately manifest and may persist for hours or days depending on the type and intensity of the stimuli. For example, both rats (Roughan and Flecknell, 2003) and cats (Waran et al., 2007) were noted to arch their backs, writhe, twitch and crouch in the hours following abdominal surgery; mice were observed to writhe and flinch after vasectomy surgery (Wright-Williams et al., 2007); ear flicking was recorded for up to 24 hours after hot-iron disbudding of calves (Faulkner and Weary, 2000); attention to the damaged area and abnormal standing posture were recorded in calves for up to 42 days following castration (Molony et al., 1995); increased   22 wound-directed behaviour was observed in sea lions (Eumetopias jubatus) after hot-iron branding (Walker et al., 2011b) and tail wagging after docking has been shown by both lambs and piglets (Dobromylskyj et al., 2000a).  With longer-lasting pain, many species show guarding behaviour, in which they appear to protect the injured area from further damage or stimulation (Zimmermann, 1986). For example, chickens that have had a joint injected with sodium urate crystals spend more time standing on their untreated leg, presumably to protect the one made painful by the treatment (Gentle and Corr, 1995).  Animals will frequently adopt specific postures apparently to minimize the pain.  In response to abdominal pain dogs often display a distinctive ‘bowing’ or ‘prayer’ pose in which their front legs are stretched in front of them on the ground and their hind end is raised (Hansen et al., 1997). Horses with severe laminitis will adopt a ‘stiff board-like’ position, may be reluctant to move and will limp or repeatedly shift weight, seemingly in order to avoid pressure on affected hooves (Taylor et al., 2002). Comprehensive discussion of behaviourally based pain indices are provided by Carstens and Moberg (2000), Rutherford (2002) and Sneddon et al. (2014). Humans and many other species show specific changes in facial expression in response to pain. The challenge of assessing pain in subjects without the ability to verbally report was addressed in human neonates by the development of the Neonatal Facial Coding System (NFCS) (Grunau and Craig, 1987). NFCS is adapted from the Facial Action Coding System (FACS) which is anatomically based and identifies minimal units of facial movement called Facial Action Units (FAUs) with detailed descriptions of the resulting changes in facial appearance (Ekman and Friesen, 1978). This approach,   23 now reliably used by doctors and nurses for pain assessment in infants (Stevens et al., 2010), has provided the basis of a framework for pain assessment in animals and has resulted in grimace scales being developed for mice (Langford et al., 2010), rats (Sotocinal et al., 2011), rabbits (Keating et al., 2012), cats (Holden et al., 2014), sheep (McLennan et al., 2016), horses (Dalla Costa et al., 2014), ferrets (Reijgwart et al., 2017) and piglets (Di Giminiani et al., 2016). Despite the species differences, these expressions in response to pain are well conserved (Chambers and Mogil, 2015), suggesting the strong likelihood that other mammals also exhibit similar expressions in response to painful stimuli. Facial expressions have proven to be a repeatable, accurate and valid way to identify pain, and there is promise for the development of a grimace scale in other species as a new area of research. Facial expressions in response to pain will be discussed in greater detail in Section 1.2.  1.1.7.2.2 Change in the frequency or magnitude of certain behaviours  Decreases in the frequency and magnitude of behaviour associated with pain can include a reduction in general activity, locomotion, and maintenance behaviours such as feeding and drinking. For example, after castration, piglets locomote less, avoid certain sitting postures and either spend more time with the sow or more time socially isolated (Llamas Moya et al., 2008). Sea lions showed increased latency to eat and drink, decreased locomotor activity, increased lying times and were frequently noted back-arching after abdominal surgery (Walker et al., 2011b). Following the castration of lambs, acute signs of pain abated within 8 hours, but the animals reduced feeding for 24 hours and reduced play behaviour for approximately 3 days (Dobromylskyj et al.,   24 2000a). After beak trimming, chickens showed guarding behaviour of the damaged area by minimizing feeding, drinking and preening. The birds also demonstrated more generalized changes in activity levels and social behaviour which persisted for weeks, suggesting the procedure results in chronic pain (Duncan et al., 1989). However, not all behaviours have the same weight or importance as indicators of a pain response. Weary et al. (2006) suggest that changes in behaviours that could have a direct effect on the animal’s fitness, or those that the animal would normally be highly motivated to perform, may be particularly indicative of pain.   1.1.7.2.3 Choice or preference behaviours The challenge with many of the aforementioned measures is that they cannot with certainty demonstrate that the animal is actually experiencing pain as opposed to responding unconsciously to stimuli.  As mentioned, many physiological changes happen unconsciously and not only in response to negative stimuli. Behavioural changes might be reflexive actions that protect against further injury and are not necessarily indicative of an animal’s experience of the unpleasant or cognitive aspect of pain. There are, however, several test methods that suggest animals do, in fact, experience pain.  Preference or choice experiments can indicate how animals value or perceive the aversiveness of a stimulus or treatment by measuring their choice in a given situation or their motivation to escape or avoid it (Chapman et al., 1985).  Aversion learning, for example, has been used to assess the degree of fear or pain an animal associates with a particular treatment. In being repeatedly exposed to a stimulus, an animal can learn   25 what to expect and thus what to avoid (Weary et al., 2006). For example, sheep became comparatively more difficult to move through a chute where they had been restrained by electro-immobilization versus a chute where they had been manually restrained, suggesting they found electro-immobilization more aversive than manual restraint (Rushen and Congdon, 1986).  Rats demonstrate conditioned place aversion, avoiding the places they have received noxious stimuli (Gao et al., 2004) and will attempt to cover electric probes with bedding (Pinel and Mana, 1989).  Another compelling form of choice experiment is the self-selection of analgesic medication by animals presumed to be in pain. Colpaert et al. (1982, 1980) were the first to use this method, demonstrating that rats with chronically inflamed joints would consume more water containing pain medication (both an NSAID and an opioid) versus a sweetened solution, compared to rats without the presumably painful condition. Similarly, lame broiler chickens were shown to consume more food containing an anti-inflammatory analgesic than healthy controls. Interestingly, the amount of the drug consumed increased proportionally with the severity of the birds’ lameness (Danbury et al., 2000). An animal’s choice to actively seek pain medication, as well as the level of self-medication it chooses to consume, may be indicative of the severity of the pain it is experiencing. Additionally, as many species attempt to mask their pain so as to not appear vulnerable, this type of experimentation may be useful for identifying pain when observable pain behaviours may be absent (Fraser, 2008). A study with zebra fish (Danio rerio) examined whether or not the fish were willing to pay a cost in order to gain pain relief.  When given the option of an enriched chamber or one that was barren and brightly lit, zebra fish consistently chose the   26 enriched one. This strong preference remained for groups of fish that were injected with acetic acid (noxious) and for fish injected with saline (control).  However, once the barren chamber was infused with an analgesic, the fish injected with acid no longer demonstrated the strong environmental preference for the enriched chamber and would spend more than fifty percent of their time in the barren but medicated chamber.  The authors interpreted this behaviour as indicating a fish’s willingness to pay a cost in the form of an undesirable environment (brightly lit barren tank) in order to access analgesia and presumably mitigate pain (Sneddon et al., 2003).   Preference and choice experiments demonstrate some of the more complex behavioural responses associated with pain. This type of testing shows animals can use their pain experience to make decisions and alter their behaviour, seemingly in order to avoid the recurrence of the noxious stimuli (Sneddon et al., 2014).  1.1.7.2.4 Cognitive bias  Testing cognitive bias has potential to directly assess the affective state of animals and thus help determine the significance of their pain experience. When experiencing pain, which is a negative emotional state, humans will demonstrate a judgment bias in which ambiguous stimuli are interpreted as negative (Pincus et al., 1994). This type of cognitive or judgment bias has been used to assess the impact that environmental and husbandry practices may have on animals’ affective states (Sneddon et al., 2014). Primarily, however, research on negative judgment bias resulting from pain has focused on humans (Pincus and Morley, 2001). The first evidence of a judgment bias in response to pain in a non-human species was seen in newly dehorned dairy calves   27 (Neave et al., 2013). Hot-iron dehorning is a common management procedure known to be painful (Morisse et al., 1995). Prior to dehorning, calves were trained to respond differentially to red and white video screens and were then tested with unreinforced ambiguous colours. Calves were more likely to judge ambiguously coloured screens as negative in the 22 hours after being dehorned, compared to before dehorning, suggesting a negative (‘pessimistic’) bias in their judgment of the ambiguous stimuli (Neave et al., 2013). This work provides support for the affective component of pain in these animals and demonstrates promise for future research of this kind.  1.1.8 Administration of analgesia The administration of analgesic (or in some cases, local anaesthetic) is useful for identifying pain and has been used to validate multiple pain measures. When pain causes certain behaviours to occur more or less frequently, the removal of that pain should return behaviour to normal (pain-free baseline). In other words, analgesics have the dual purpose of, on the one hand, decreasing those behaviours that increase if pain-specific (e.g., wound-directed licking) and, on the other hand, increasing those behaviours that decrease in response to pain (e.g., activity). If an animal is exhibiting behaviour indicative of pain, and if the signs are reduced or disappear after analgesic treatment, it can be assumed the condition or injury is in fact painful. For example, nerve blocks are frequently used to identify lameness in horses (Rietmann et al., 2004). If the animal’s mobility is improved, it could be concluded that the impaired gait is due to pain rather than a mechanical problem of the limb (McGeown et al., 1999).  Many pain-related behaviours and physiological changes have been observed to   28 lessen or disappear when the animal is treated with appropriate pain-relieving medication (Rutherford, 2002). Calves showed a lower frequency of pain-related behaviours (ear flicks) after dehorning when provided an NSAID, compared to animals without an NSAID (Faulkner and Weary, 2000). After castration (by crushing the blood vessels that supply the testes), lambs had reduced plasma cortisol, trembled less and had fewer abnormal postures when administered an NSAID (Molony et al., 1997). Plasma cortisol responses were also reduced by administration of anti-inflammatory drugs to cattle after surgical castration (Stafford et al., 2002) and scoop dehorning (Sutherland et al., 2002). Dogs treated with analgesics prior to ovariohysterectomy had reduced physiological stress responses (i.e., no increase in catecholamine release and delayed ACTH and cortisol responses) when compared to dogs without pain medication (Benson et al., 2000). In a study by Petrie et al. (1996), calves were either dehorned without anaesthetic, dehorned with a local anaesthetic, or not dehorned but handled and restrained as if they were to be dehorned (sham). Immediately after the procedure, cortisol increased in the group dehorned without anesthesia but not in either the anaesthetic group or the sham group. Interestingly, 2 to 3 hours later, when the effect of the local anaesthetic was expected to wear off, cortisol levels of the group that had received the anaesthetic began to increase to levels higher than those that had not received it. Presumably, as the anaesthetic wore off, these animals began to experience pain from the injured area, which activated their HPA response. The animals that did not receive anaesthetic had an initial release of cortisol in response to the procedure, which has natural anti-inflammatory properties. The comparatively elevated cortisol levels in   29 the group without pain management, and the delayed onset of cortisol elevation once the anaesthetic wore off, is consistent with procedural pain (Petrie et al., 1996). As noted above, validation of pain measures requires responses to be observed with and without the painful condition and with and without effective pain medications (Weary et al., 2006). There are, of course, ethical issues in conducting experiments that cause pain without providing pain management.  Dobromylskyj et al. (2000a) suggest that if an experiment includes animals undergoing painful procedures without analgesics, an interventional pain therapy protocol should be applied in those cases where pain thresholds are considered to have been exceeded. It is also suggested it may be possible to undertake the approach used in some post-operative analgesia studies on children in which administration of analgesia is delayed until the patient has made a full recovery from the anaesthetic (Dobromylskyj et al., 2000a). Additionally, because some analgesic drugs may have inhibitory or excitatory behavioural effects unrelated to pain or nociception, caution must be used to ensure that any change is due to the alleviation of pain as opposed to a general response to the drug (Wall, 1992).  1.1.9 Pain-scoring systems Various pain-scoring techniques used in human medicine have been adapted for use in animals. Particularly applicable are methods used to assess acute pain in pre-verbal children that are based on vocalizations (crying), facial expressions, posture and behaviour (McGrath, 1987). Most of the scoring systems used for animals have been simple descriptive scales, visual analogue scales, numerical rating scales, or composite   30 scales and rely on observers’ assigning scores based on their subjective evaluation of the signs of pain they observe (Dobromylskyj et al., 2000a). The simple descriptive scale (SDS) typically uses verbal descriptors such as ‘mild’, ‘moderate’, or ‘severe’ to represent a range of no pain to severe pain.  Often each expression has an assigned index value that becomes the pain score.  Verbal descriptors have been useful for allowing humans to describe their pain experience but evaluation of their use by veterinary staff demonstrated an inconsistent relationship between descriptors and the severity of pain being scored (Holton et al., 2001).  Visual analogue scales (VAS) use a 100mm line with ‘no pain’ at 0mm and ‘worst imaginable pain’ at 100mm.  The observer assessing the animal places a mark on the line where she believes the level of an animal’s pain to be. The pain score is the distance in mm from ‘no pain’ to the observer’s mark. This method has been used for scoring different types of pain in several species including post-operative pain in dogs (Firth and Haldane, 1999), castration pain in lambs (Thornton and Waterman-Pearson, 1997) and degree of lameness and wound severity in sheep (Welsh et al., 1993).  Numerical rating scales (NRS) are similar to VAS scales but, as opposed to a mark on a line, the observer assigns a numerical score (typically 0 to 5 or 10) to reflect the pain intensity. Examples of painful conditions that have been assessed using numerical rating scales include degree of pain from lameness in sheep (Welsh et al., 1993), as well as a range of husbandry procedures and diseases in sheep and cattle (Fitzpatrick et al., 2006, 2002).  In general, both VAS and NRS systems have been demonstrated to be more sensitive than SDS. When scoring lameness in sheep and post-operative pain in dogs, both NRS and VAS had strong inter-observer agreement   31 but VAS was slightly more sensitive in both cases (reviewed by Dobromylskyj et al., 2000a).  Composite scales that combine multiple relevant behaviours have been used to assess pain in several species. For example, a combined index of the time spent lying ventrally and the incidence of the active behaviours (e.g., restlessness, rolling, stamping, kicking) has been proposed as a practical method for measuring pain from rubber ring castration and tail docking in lambs (Molony and Kent, 1997).  Roughan and Flecknell (2003b) developed a composite behaviour score designed to assess post-operative pain in rats.  They selected behaviours that occurred frequently enough to be easily scored and that would occur after different types of surgery (e.g., back arch, abnormal gait, writhing and staggering/falling). The cumulative frequencies of the behaviours were used to assess the severity of pain rats experienced after abdominal surgery, as well as the efficacy of analgesia. Recently, a composite pain score was developed for horses that combined multiple pain behaviours seen in horses suffering from either abdominal or musculoskeletal pain. The authors reported the scale to have high inter-observer reliability, to be more reliable than NRS scores, and to be predictive of a horse’s survival after emergency gastrointestinal surgeries (van Loon et al., 2014). All of these scoring systems rely on subjective interpretation of behavioural signs, but their correlations to other physiological measures have not been well established. Composite scales seem to provide a more complete picture of the pain being scored, particularly as individuals may respond differently to the same pain source.    32 1.1.10 Limitations of behavioural measures Using behavioural indicators to identify pain has become a common strategy. Scoring behaviour can be easily employed under a variety of conditions and is minimally invasive as it does not require restraint or physical sampling. Because behaviour can be observed in the present, it has the advantage of enabling immediate assessment. However, behaviours may vary between individuals, between species and between different contexts (e.g., an animal may exhibit different behavioural responses in its familiar home environment compared to a veterinary clinic where it may be anxious or afraid). Therefore, the successful scoring of pain-related behaviour requires a detailed understanding of species-specific behaviour as well as a baseline of normal pain-free behaviour for the particular species and for the individual being observed. Observers may vary in their interpretations, making it difficult to ensure inter- and even intra-observer reliability. Additionally, behavioural indices may identify that an animal is in pain, but may provide little evidence of how specific behaviours relate to pain intensity. Although there is a growing body of literature on animal pain behaviour, still relatively few painful procedures have been assessed in only a small number of species. The process of identifying and validating pain-specific behaviours and developing a scoring system is time-consuming and, once established, may be difficult to apply in a practical setting (Dobromylskyj et al., 2000a).  1.1.11 Possible factors affecting the pain response Numerous environmental and psychological factors may influence the pain response. In humans, the pain experience can be affected by present surroundings, past   33 experiences of pain, or fear of incurring future pain (Craig, 1994).  In addition, social context has been shown to affect risk factors and outcomes of chronic pain (Mogil, 2015). In animals, both increases and decreases in pain responses have been observed as a result of social conditions, environmental surroundings, shifts in attention and affective states such as fear. Social factors such as communication and social organization may influence pain responses.  Depending on social organization and life histories, individuals of different species may overtly display signs of pain if these are helpful in incurring social support or warning con-specifics of danger, or individuals may suppress visible signs of pain in situations where they may be vulnerable to predation or aggression from conspecifics (Flecknell, 2000).  Social communication of pain may benefit the animal signaling pain if it leads to support, and may benefit the animal receiving the signal in the form of warning or the chance for observational learning (Mogil, 2015). Social communication was demonstrated to affect the pain response in mice when latency to withdraw the tail from heat was tested. If the first mouse was placed back in the home cage before the others were tested, subsequent mice had a faster response rate.  This effect of test order was removed if the first mouse was not returned to the cage before the others were tested, suggesting that some type of social communication took place (Chesler et al., 2002). Social modulation of distress specifically related to pain has been shown in several species, including mice, rats and pigeons (Langford et al., 2006). When mice in pain were housed together, they displayed more pain behaviours than those housed alone or in the presence of strange mice (Langford et al., 2006). Mogil (2015) provides   34 multiple study examples of social modulation of pain in rodents including examples of both social buffering (reduction of pain or pain-induced fear due to the presence of a social partner) and social contagion (spread of pain or pain-induced fear between animals). Social status may also play a role in pain sensitivity although the actual effect is unclear. For example, one study observed that dominant mice had decreased sensitivity in the few minutes after a formalin injection, followed by a period of increased sensitivity (Aghajani et al., 2013). However, a separate study using the same model found that it was the subordinate, as opposed to the dominant mice, that had decreased sensitivity in the minutes after the injection followed by increased sensitivity (Gioiosa et al., 2009). Another phenomenon that may suppress the pain response and confound pain assessment is stress-induced analgesia. Many species have endogenous systems that modulate pain. Descending pathways from the brain can modulate spinal transmission of pain signals, affecting both pain perception and inflammation (Carstens and Moberg, 2000).  Stress-induced analgesia is likely a defensive strategy that enables animals to hide signs of vulnerability and to attend to immediate threats such as a predator. In particular, prey species will cease overtly observable pain-behaviour if threatened. For example, the presence of a human observer, or even horizontally placed eyes suggestive of a predator, can inhibit pain-related behaviours in mice  (MacIntyre 2007 in Mogil, 2009). There are multiple examples of stress-induced analgesia both in laboratory and field studies (Sneddon et al., 2014). In humans, pain can be mitigated by a redirection of attention using cognitive behavioural therapies and relaxation techniques such as meditation or hypnosis. Gentle   35 (2001) investigated the effect of attentional shifts on pain perception in chickens. He hypothesized that if birds’ pain response was more than a simple, unconscious reaction, then directing their attention elsewhere should suppress the pain response in a way that is similar to what has been recorded in humans.  Sodium urate was injected into the leg joints of chickens to simulate arthritic joint pain and produce behaviours consistent with inflammatory pain for approximately 3 hours. For several hours after injection, chickens would favour the injured leg by standing on the unaffected leg or walk with a limp.  However, when chickens were provided with a distraction in the form of an enclosure that had litter of wood shavings (novel) or an unfamiliar conspecific, the limb-guarding behaviour disappeared. This behavioural change was also demonstrated immediately before egg-laying, during which time chickens are highly motivated to build a nest.  Once the egg was laid and no longer demanded their attention, the pain-related behaviours returned.  No behaviour indicative of fear was observed during the study, so the authors ruled out the possibility that behavioural changes were due to stress-induced analgesia. Instead, they concluded that shifts in attention reduced pain and reduced signs of peripheral inflammation of chickens being tested. The ability to have attention distracted away from pain suggests that the pain response in chickens must be in part mediated by an awareness of the pain, otherwise the behaviour would persist regardless of their attention being drawn elsewhere (Gentle, 2001).  In summary, pain assessment is complicated by the fact that an animal’s pain experience can be affected by a variety of factors that are not necessarily related to the pain itself. It is therefore important to consider the context in which animals are being   36 tested or observed and to be aware that impinging stressors may have a confounding effect. 1.1.12 Conclusions Reducing or eliminating the pain of the animals in our care is critical to animal welfare. Our ability to manage pain is dependent on its accurate identification and reliable measurement. In recent years many advances have been made in understanding and quantifying pain but due to its complicated, multi-dimensional nature, accurate assessment is still a challenge. A variety of physiological and behavioural indicators have been identified as useful in pain assessment although no single measure is likely valid in isolation. Instead, multi-factorial pain scoring systems that integrate several measures may be more effective at capturing the different dimensions of the pain experience (Livingston and Chambers, 2000). Continued work is needed to develop more non-invasive measures and scoring methods that reliably detect the presence of pain, assess its severity and duration, and are practical to employ.  1.2 Facial expressions indicative of pain  1.2.1 Introduction  As noted above, clinical assessment of pain in humans has relied heavily on the ability to verbally communicate one’s pain status (Anand and Craig, 1996), but accurate pain assessment in animals is difficult and can be controversial (Ranger et al., 2007). Facial expressions have been recognized as important for quantifying pain in patients unable to self-report, as in the case of human infants or adults with cognitive impairments   37 (Grunau and Craig, 1987; Hadjistavropoulos et al., 2014).  Facial expression can also be an important non-verbal indicator of pain in people who are able to communicate verbally, as facial expressions are more reflexive in nature than self-report and therefore are less vulnerable to intentional modulation (Craig et al., 1991). In recent years, specific expressions associated with pain have been identified in multiple species of non-human animals, and are proposed as promising for the assessment of animal pain (Flecknell, 2010).   1.2.2 The human ‘pain face’ Within the repertoire of facial expressions that have evolved in humans, there appears to be a distinctive expression of pain (Craig et al., 2001). Facial expressions of pain appear to be evolutionary stable and may be both unlearned and adaptive.  The social communication of pain by facial expression can function to warn of potential threat and to signal a need for assistance or support (Williams, 2002). In humans, it has been demonstrated that even when an individual attempts to conceal pain, the ‘pain face’ is difficult to wholly suppress (Prkachin and Mercer, 1989). Additionally, children who are congenitally blind have a full range of spontaneous expressions, including pain, suggesting that these expressions develop without learning (Freedman, 1964). The facial expression of pain is consistent across developmental life stages (Kunz et al., 2008) and across different pain modalities (Prkachin, 1992). In fact, the ‘pain face’ may be one of the most salient, non-verbal behaviours indicative of the pain experience (Simon et al., 2006).   38 Facial response to pain appears to be primarily reflexive and spontaneous, involving only a modicum of conscious control (Chambers and Mogil, 2015; Craig et al., 2010).  In neonates, facial change resulting from pain is displayed before their expressions are influenced by experiential factors. In adults, however, the pain response can be affected by contextual events and learned coping strategies (Grunau and Craig, 1987). It has been demonstrated that children as young as eight are able to manipulate their facial response to noxious stimuli (the immersion of hands into hot (30ᵒ C) or cold (10ᵒ C) water) (Larochette et al., 2006). However, it is possible that if the stimulus had been of a more severe or enduring nature, the children’s attempts to control their expressions would have been less successful. For example, adults experiencing chronic lower-back pain were unable to entirely mask their facial displays (Craig et al., 1991), and evidence of pain in other chronically pained patients is conveyed despite their efforts to conceal it (Poole and Craig, 1992).   1.2.3 The primacy of the face Facial expressions may be inherently easier for human observers to decode than other behavioural manifestations (Flecknell, 2010). Facial expressions are a prime means of conveying social information between humans, and we have a specialized neural apparatus that enables us to detect and recognize facial cues immediately (Kadosh and Johnson, 2007). Perhaps in part because of this natural skill, non-verbal cues are considered by many observers to be more credible than verbal indicators such as self-report (Craig, 1992; Poole and Craig, 1992). Moreover, the tendency to look for cues from facial expression seems to transfer to our assessment of other species.  For   39 example, when asked to gauge pain in rabbits without specific instructions, observers would focus on the rabbits’ faces first, more frequently and for longer durations than on other bodily regions (Leach et al., 2011).  Seemingly, the use of grimace scales may take advantage of humans’ natural inclination to attend to and process facial cues.  1.2.4 Measuring facial changes   1.2.4.1 The Facial Action Coding System (FACS) Ekman and colleagues did pivotal work in the area of facial expression, asserting that most ‘basic’ emotions (e.g., happiness, fear, anger, disgust, sadness) have distinct, corresponding facial expressions (Williams, 2002). In 1978, Ekman and Friesan published the Facial Action Coding System (FACS), which comprehensively classified human facial muscle movement into 46 discrete, anatomically-based action units (AUs). These facial action units are said to be those necessary and sufficient to describe objectively every possible facial expression (Ekman and Friesen, 1978). Specific constellations of AUs have been recognized as unique to pain, with multiple studies reporting the occurrence and increased intensity of these facial changes when people are in pain (see review by Prkachin, 1992). Prkachin (1992) suggests that the majority of the information about the pain experienced by an individual is conveyed by only 4 of the reported AUs: lowered brow, orbital tightening and closing, nose wrinkling and raising of the upper lip, and that these may comprise a generalized ‘basic signal’ across different types of pain. Since its development, the FACS has been used successfully for the evaluation of experimental and clinical pain in humans across life stages ranging from infancy   40 (Grunau and Craig, 1987) to old age (Kunz et al., 2008) and in populations of non-verbal adults such as those with cognitive impairment such as dementia (Hadjistavropoulos et al., 2014). The FACS framework has also been applied to multiple species: chimpanzees (Pan troglodytes) (Vick et al., 2007), orangutans (Pongo pygmaeus) (Caeiro et al., 2013), rhesus macaques (Macaca mulatta) (Parr et al., 2010), gibbons and siamangs (Hylobatids ) (Waller et al., 2012), and domestic dogs (Waller et al., 2013), cats (Caeiro et al., 2013) and horses (Wathan et al., 2015).   1.2.4.2 The Neonatal Facial Coding System (NFCS) The Neonatal Facial Coding System (NFCS) was developed specifically to address the issue of pain in human neonates.  Adapted from the FACS, the NFCS focused only on those facial movements associated with noxious stimuli and included some modifications to account for the subtle differences between infant and adult facial structure (Grunau and Craig, 1987). Changes in 5 facial movements (brow lowered, eyes squeezed shut, naso-labial furrow deepened, lips and mouth opened, tongue taut and cupped) are consistently found in healthy, full-term infants, and when used in combination with crying behaviour, can reliably differentiate between invasive and non-invasive tactile stimuli (Grunau et al., 1990). This pattern of infants’ facial changes in response to noxious stimuli closely resembles that of adults experiencing pain (Craig, 1992). Facial expressions are now considered a reliable measure of pain in both pre- and full-term infants and can be used to evaluate pain in both acute procedures (Grunau et al., 1990; Grunau and Craig, 1987) and the post-operative period following abdominal and thoracic surgeries (Peters et al., 2003). The development of the NFCS   41 not only provided an objective measure for scoring infant pain, but also provided evidence of infants’ ability to perceive pain, a view which was still met with doubt at the time (Chambers and Mogil, 2015).  1.2.4.3 Pain faces in non-human animals Darwin proposed that facial expressions are evolutionarily conserved, such that humans and non-human animals display facial expressions of specific emotional states in similar ways (Darwin, 1872). In fact, continuities and parallels between the facial expression of pain in human and non-human animals makes its use as a measure of pain translatable for multiple species. Using methods analogous to the NFCS, species-specific grimace scales have now been developed for mice (Langford et al., 2010), rats (Sotocinal et al., 2011), rabbits (Keating et al., 2012), horses (Dalla Costa et al., 2014; Gleerup et al., 2015), cats (Holden et al., 2014), sheep (McLennan et al., 2016), ferrets (Reijgwart et al., 2017) and piglets (Di Giminiani et al., 2016).   1.2.4.4 Grimace scales The first systematic assessment of a ‘pain face’ in a non-human species was completed in mice. By quantifying the facial expressions of mice in response to moderately noxious stimuli, Langford et al. (2010) developed the behavioural scoring system called the mouse grimace scale (MGS). The MGS consists of 5 pain-related action units: orbital tightening, nose bulge, cheek bulge, and ear and whisker position. Each AU is scored on a 3-point scale and then the sum of the 5 scores is totaled to assign a grimace score.   42 This scale has been shown to discriminate accurately between mice with and without pain, to quantify reliably pain of moderate duration (10 minutes to 4 hours), and to be sensitive to changes in facial expression when pain is attenuated by analgesics (Langford et al., 2010).  Since its development, the MGS has been applied clinically for the evaluation of analgesia efficacy (Matsumiya et al., 2012) and post-operative pain in mice (Leach et al., 2012). Interestingly, the MGS might represent a measure of the affective component of mouse pain. The rostral anterior insula is a part of the brain associated in humans with the affective component of the pain experience (Schweinhardt et al., 2006). When this area was chemically destroyed in mice, their facial expressions of pain were prevented whereas the expression of other pain-specific behaviour (abdominal constrictions) remained unaffected.  This suggests that facial display may reflect the negative affect of the pain experience (Langford et al., 2010). Subsequent grimace scales that have been developed demonstrate that, despite some species-specific variations, the facial change in response to pain is remarkably similar between species. For example, 3 of the 4 AUs in the rat grimace scale are the same as those observed in mice and rabbits (Keating et al., 2012; Sotocinal et al., 2011). Similarly, the distinguishing features of the cat pain face are orbital tightening, nose/cheek flattening and ear and whisker changes (Holden et al., 2014). Even horses, a species with arguably few physical features in common with mice or rats, have AUs largely similar to those described in other grimace scales (Dalla Costa et al., 2014).  In fact, all of these animals, including humans, share characteristics of a ‘pain face’ typified   43 by orbital tightening and a change in the nose/cheek area supporting the view that key expressions are phylogenetically conserved (Williams, 2002).  For grimace scales to be useful in veterinary medicine, they must be applicable in clinical settings and agree with other validated pain measures. The horse grimace scale (HGS) has been shown to be a valid measure of pain from surgical castration and to be highly correlated to other validated measures of horse pain, such as the Composite Pain Scale (Dalla Costa et al., 2016). Recently, the HGS was applied to evaluate pain from laminitis, a common equine disease in which the laminae of the hoof are inflamed causing acute and chronic pain.  Typically, the pain severity of this condition is scored using the Obel grading system, which requires the afflicted horse to move through several gaits thus exacerbating the pain (Viñuela-Fernández et al., 2011). When used to assess laminitis, the HGS scores were found to be consistent with those of the Obel system and had the advantage that horses could be assessed at rest (Dalla Costa et al., 2016).   1.2.4.5 Measuring acute versus long-term pain Facial expression has been demonstrated as a reliable indicator of acutely painful stimuli in multiple species, but its use as a measure of longer-lasting or chronic pain is less clear. In general, chronic pain may be more difficult to recognize and evaluate than acute pain, as indicators may be more subtle and less specific, and there may be periodic fluctuations of acute exacerbation (Brearley and Brearley, 2000). Craig and Patrick (1985) found that facial responses to pain were strongest at the beginning of noxious stimuli and would rapidly habituate without continued stimulation. There is,   44 however, evidence of measurable facial cues in humans experiencing chronic pain. For example, LeResche et al. (1992) observed that the frequency of pain-related AUs was higher at baseline for patients experiencing chronic pain from temporomandibular disorder (TMD) and became more frequent over time if the condition persisted.  While AU frequency increased, the verbal reports of pain from TMD patients remained consistent. The authors concluded that the facial expression of pain is likely socially reinforced in humans, whereas verbal reporting may be interpreted as complaining and may actually be socially punished (LeResche et al., 1992). One challenge in using facial change to assess longer-lasting pain is the difficulty of establishing a pain-free baseline (Williams, 2002). A study using the MGS for cage-side clinical assessment demonstrated that pain-free mice did not have grimace scores of zero, and that there were significant differences in the pain-free baseline between the sexes and the different mouse strains (Miller and Leach, 2015). It is possible the differences seen in the pain-free baselines were a consequence of the non-specificity of any given AU, as it is the concatenation of AUs that best signals pain to the observer. The establishment of baselines of facial expressions for different species would improve the accuracy of grimace scales when used for clinical assessment and possibly for the assessment of pain beyond the acute phase.   To date, grimace scales developed for non-human species have been focused on spontaneous, evoked pain. In rats, a pain grimace persisted for less than 48 hours after several inflammatory assays and laparotomy. However, in this case, it is possible that the disappearance of the painful expression reflects the adaptive strategy of prey   45 species of inhibiting overt pain signals as soon as possible, as opposed to the actual disappearance of pain (Sotocinal et al., 2011).   1.2.5 Conclusions Facial expressions of pain have been established as a valuable indicator of pain across species.  Largely reflexive and difficult to suppress (Craig et al., 2010), the ‘pain face’ seems to provide a sensitive and specific measure of an individual’s pain. Grimace scales are non-invasive in application, use the natural tendency of humans to focus on the face when gauging pain (Williams, 2002), can be used to assess pain ranging from mild to severe, and can be applied in real-time.  There is also the compelling suggestion that facial expression could reflect the affective component of pain (Langford et al., 2010). Although, continued research is needed to establish the presence of facial grimaces in other species, for all types of pain and over prolonged durations of pain, grimace scales are a promising tool for enhancing our ability to assess animal pain.  1.3 Vocalizations as potential indicators of pain  1.3.1 Introduction Vocalizations are a critical component of communication for many species (Manteuffel et al., 2004). An individual’s calls can convey meaningful information about its biological condition (Watts and Stookey, 2000), signal attraction or warning among conspecifics, or reflect various affective states (Manteuffel et al., 2004). Vocalizations in many species are distinctive enough to convey specific meaning (Seyfarth and Cheney,   46 2003); thus, if interpreted correctly, they could be a useful indicator of particular affective states, including pain. Some species are thought to have specific types of vocal utterances that convey information about their affective experience. If a piglet becomes separated from its mother, it will emit calls that will attract her (Weary et al., 1996) in a pattern distinctive enough that experienced pig keepers are said to easily associate it with a lost piglet (Fraser, 2008). Hens emit a specific ‘gakel-call’ when thwarted from performing important behaviours such as dust-bathing, eating and drinking, such that the rate of this specific type of call is considered to accurately reflect their degree of frustration (Zimmerman et al., 2000). However, occasionally signals are deceptive; for example, roosters have been shown to falsely use ‘food calls’ to attract hens (Gyger and Marler, 1988). Therefore, the correct interpretation of different species’ vocalizations requires some knowledge about each species' natural history and the contexts in which certain signals are likely to occur (Weary and Fraser, 1995). Evolution should favor signals that convey a selective advantage. Individuals that are effective at signaling their needs may be expected to leave more copies of their genes if there is a recipient that perceives the signal and responds in a way that provides assistance (Johnstone and Grafen, 1993; Maynard Smith, 1994; Weary and Fraser, 1995).  For example, calls that indicate ‘separation distress’ are given by dependent young of many species where parents come to the aid of their young (Weary et al., 1997). Similarly, ‘alarm’ calls that are thought to provide a fitness advantage in contexts where the signaler is genetically related to at least some the animals that hear   47 the warning and may thus have better opportunity to escape the threat (Fraser et al., 1995). The begging calls of nutritionally dependent young provide an interesting example of offspring communicating an affective state, hunger, to their parents (Godfray and Johnstone, 2000).  There is obviously a fitness advantage for the parents in responding to these calls, thus ensuring survival of viable young. Theoretically, however, an individual could amplify or exaggerate its signals to indicate a greater need than it actually has and thus receive more parental resources. In fact, it has been demonstrated in some bird species (e.g., pied fly catcher, Ficedula hypoleuca) that increased begging does increase parental provisioning of food. In such cases, there may be a fitness conflict between offspring and parents if additional provisioning decreases the fitness of the parents’ future clutches (Ottosson et al., 1997). In this context, there must therefore be an equilibrium to ensure the fitness of both parties, which requires a level of ‘honesty,’ or the accurate representation of individual need implicit in these type of signals (Godfray, 1991).   1.3.2 Honest signals of need ‘Honest signals’ can be considered traits or behaviours that reliably convey accurate and otherwise unobservable information to the receiver. Theories of ‘honest’ signaling stipulate conditions under which signals that accurately indicate an individual’s fitness needs would be selectively favoured. These stipulations are that: 1) the signaler and receiver are genetically related so that the receiver obtains a fitness benefit by detecting and responding to the signal; 2) signals vary in a way that reflects the degree of need   48 for a response, and; 3) signaling incurs some ‘cost’ to individual fitness (Johnstone and Grafen, 1993; Maynard Smith, 1994; Weary and Fraser, 1995). Begging calls of nestling birds can, in fact, be viewed as examples of  ‘honest’ signaling of need because the calls increase and decrease in response to food intake (e.g., Redondo and Castro, 1992). Additionally, in producing these calls, nestlings incur a cost through a greater risk of being detected by a predator. Therefore, if a nestling is truly in need of feeding, the potential gain from begging outweighs the risk of predation; in contrast, if begging does not truly indicate need, then the risk of predation will outweigh the potential gain of begging (Godfray, 1991).  Weary and Fraser (1995) suggest that ‘honest’ signals that reliably reflect an individual’s degree of need are particularly useful for assessing animal welfare. A series of studies on piglets showed that certain piglet vocalizations do meet criteria for honest signaling and thus can convey information about the animals’ welfare (Taylor and Weary, 2000; Weary and Fraser, 1995; Weary et al., 1996; Weary and Fraser, 1997). When unweaned piglets are separated from their dam, they become very active and produce a characteristic series of vocalizations that increase progressively in loudness and frequency. These calls are rarely exhibited when piglets are with the sow and hence the calls are considered to signal distress from separation (Fraser, 1975). Presumably, the calls of piglets that need maternal attention, such as those that are hungry or cold, should increase with the extent of their need. In fact, when non-thriving piglets (those with low body weight and poor weight gain) and piglets that had not recently been fed were isolated from their dam, they used more high-frequency calls, longer calls, and calls that rose more in frequency than their thriving and recently fed   49 litter-mates (Weary and Fraser, 1995) . Similarly, sows showed stronger responses to playback calls from piglets in need than to those from piglets not in need; specifically, the sows vocalized more and approached and spent more time near the playback speaker (Weary et al., 1996). As piglets become older, and hence less dependent on their dam, they show fewer of the calls and behaviours associated with separation distress (Fraser, 1975). Variations in vocal responses to varying degrees of need have also been demonstrated in other species. For example, rat pups isolated from their dam and kept in a cold environment produce more ultrasonic calls than those isolated in a warm environment, presumably reflecting their need for parental warmth (Olivier et al., 1994).  1.3.3 Vocalizations as indicators of pain Vocalizations can be a behavioural sign of distress and pain (Morton and Griffiths, 1985) and have been used to assess potentially painful procedures in many species. Such procedures include the circumcision of human infants (Porter et al., 1988, 1986); the tail-docking of puppies (Noonan et al., 1996); the tail-docking, teeth-clipping and ear-notching of piglets (Di Giminiani et al., 2017; Noonan et al., 1994); the hot-iron branding of calves (Watts and Stookey, 1999); and the castration of lambs, goat kids, calves and piglets (Mellor et al., 1991; Weary et al., 1998). Different acoustic parameters have been used in the assessment of vocal responses to pain, including the presence or absence of vocalizations (Mellor et al., 1991), the proportion of animals that vocalize (Lay et al., 1992b), the total rate of calls (Noonan et al., 1994) and the rate of different call types (Noonan et al., 1996). Other measures such as the duration,   50 amplitude and frequency (Hz) of calls have identified some distinctive acoustic characteristics associated with pain responses in the species examined. Because the affective state of an individual causes changes in muscular tension and action of the vocal apparatus, different acoustic features of the vocalizations are influenced (Briefer, 2012). For example, acoustic properties that distinguish pain cries from non-pain cries in human infants are shorter latency to cry, longer cry duration, higher frequency (Hz) and greater intensity of cries (Grunau et al., 1990). In rats, long-pulse, ultrasonic calls of frequencies between 22–28 kHz are associated with pain (Dinh et al., 1999), and piglets emit more high-frequency calls (>1000hz) after castration than those sham treated (Taylor and Weary, 2000). The variation in the acoustic properties of vocalizations is important for distinguishing the presence of pain and potentially for discriminating intensities of pain (Porter et al., 1986). Porter (1986) showed that in human infants, frequency or pitch of cries was associated with the most invasive aspects of circumcision. Similarly, during castration, piglets had the highest rate and peak frequency of calls during the pulling and severing of the spermatic cord, indicating this to be the most painful component of the procedure (Taylor and Weary, 2000). Thus, vocal analysis seems to have the potential to indicate the presence of pain and to provide important information about its intensity. To serve as effective signals, vocal responses to pain in young need to be distinct from other types of vocal utterances and sufficiently arousing to elicit a response from a caregiver (Grunau et al., 1990). Human mothers can accurately differentiate their own infant's pain cries from other types of cries (Wiesenfeld et al., 1981). To distinguish   51 pain-related vocalizations in other species, it may be possible to translate our ability to identify specific calls in our own species to those of others through the use of cross-species acoustic universals.  In fact, humans have been shown to correctly classify the emotional content of vocalizations produced by chimpanzees (Pan troglodytes), dogs (Scheumann et al., 2014), pigs (Tallet et al., 2010) and cats (Mccomb et al., 2009). Vocalizations indicating an aroused state do have acoustic properties in common across different species (Filippi et al., 2017). When listening to play-back recordings, human participants could accurately identify levels of arousal in the vocalizations not only of other mammals but also in species of birds, reptiles and amphibians. The acoustic attributes that signified the degree of arousal differed among species, but in general, listeners relied on the fundamental frequencies of calls (Filippi et al., 2017). Despite species differences, a heightened state of arousal was reflected in the acoustic modulation of vocalizations, suggesting the potential for their use to identify pain in many species. However,  despite some similarities in vocal parameters, vocal responses to pain are often species-typical (Morton and Griffiths, 1985). Not all species will vocalize in response to pain, not all individuals will have the same degree of vocal response (Cooper and Vierck, 1986), and some types of vocalizations are outside the range of human hearing (Jourdan et al., 2002). Therefore, the absence of a vocal response to potential pain does not necessarily mean the absence of pain. Additionally, there are contexts in which an overt expression of pain would confer a selective disadvantage. For example, in an adult prey species (e.g., cattle), vocally signaling injury would not likely garner help from conspecifics but would increase the risk of attracting predators   52 (Watts and Stookey, 2000). In general, the use of vocal indicators of pain may be most effective for animals whose signaling would be expected to initiate a response in a conspecific, as in the case of dependent young.  1.3.4 Conclusions In conclusion, vocalizations are quantifiable, can reflect an animal's affective state and, under certain circumstances, may reliably indicate an individual’s degree of need (Weary and Fraser, 1995). Many species vocalize as a reaction to pain, and there is evidence that these calls can reliably be used to quantify pain (Weary et al., 1998). Although the relationship between vocalizations and pain requires careful validation, vocalizations are potentially a non-invasive indicator of pain in many species, especially in the case of dependent young or other animals that receive assistance from conspecifics.  1.4. Infrared thermography as a non-invasive measure of stress and pain in animals  1.4.1 Introduction Infrared thermography (IRT) is a technology that detects the infrared radiation emitted from an object and displays it as a visual map of temperature gradients (McCafferty, 2007). This technology has many applications and has been used successfully in both veterinary and human medicine for the diagnosis of disease and injury, and for the identification of metabolic and thermoregulatory changes (Marches et al., 2013). IRT   53 also is increasingly recognized as a reliable, non-invasive method to assess both acute and chronic stress in animals, as it is influenced by both sympathetic and hypothalamic-pituitary-adrenocortical responses (Stewart et al., 2010). Peripheral vasoconstriction is dependent on the affective and physiological state of an animal (Ludwig et al., 2007). Hence, by measuring variations in skin temperature due to changes in local blood perfusion, IRT can provide a dynamic, real-time tool that can be used in the evaluation of an animal’s welfare. All objects with a temperature greater than absolute zero (-273.16 ᵒC) emit infrared radiation that can be detected and measured via thermography. Thermal energy is the portion of the electromagnetic spectrum perceived as heat.  Broad-range infrared radiation wavelengths (3-12µm) are longer than those of visible light and cannot be detected by the human eye (Stewart et al., 2005). It is estimated that 40-60% of an animal’s heat loss is within this range (Kleiber, 1975). Thermal-imaging cameras detect the electromagnetic radiation from a surface and translate it into a pictorial image of temperature distribution.  Small changes in radiated energy can be accurately detected using IRT (0.1 ᵒC precision) and then represented as readable variations in surface body temperature (Edgar et al., 2013). Inasmuch as these changes in radiated energy reflect physiological stress responses, they may provide a non-invasive tool for animal welfare assessment.  1.4.2 Applications of IRT Within veterinary medicine, IRT has been used largely as a tool to diagnose injury and disease. It has had particular application in equine medicine (Redaelli et al., 2014),   54 having been used successfully for assessing the subluxation of vertebrae, abscesses, periostitis and laminitis, as well as a range of inflammatory and lameness conditions (Eddy et al., 2001). Eddy et al. (2001) found that flexor tendon injuries in horses could be identified with IRT before any clinical signs of lameness developed.  IRT has also been used to identify and monitor infectious diseases. These include rabies in raccoons (Procyon lotor) (Dunbar and MacCarthy, 2006), foot-and-mouth disease in cattle (Rainwater-Lovett et al., 2009) and mule deer (Odocoileus hemionus) (Dunbar et al., 2009), bluetongue virus in sheep (Perez de Diego et al., 2013), and bovine viral diarrhea and bovine respiratory disease  in cattle (Schaefer et al., 2012, 2007). Schaefer et al. (2012) found that IRT could detect and predict bovine respiratory disease at an earlier stage than other conventional methods. Using an automated IRT system that scans the eyes of cattle at the water station, they found that animals with bovine respiratory disease had higher eye temperatures, indicating infection prior to the presentation of any other clinical signs. In these examples, because IRT can identify temperature variations in otherwise clinically healthy animals, it can be used to detect early warning signs for the development of both injury and disease, enabling earlier interventions for management and treatment (Schaefer et al., 2012). IRT has also been used for other purposes in animal production. These include to identify the predisposition of animals to producing poor quality meat, such as Dark-Firm-Dry beef (Tong et al., 1995) and Pale-Soft-Exudative pork (Schaefer et al., 1989), infertility in bulls using scrotal surface temperatures (Kastelic et al., 1996), and estrus in dairy cattle (Hurnik et al., 1985). In the dairy industry, IRT has also been used to assess   55 lameness (Nikkhah et al., 2005) and for the early detection of mastitis (Polat et al., 2010). IRT has also been used to monitor physiological and welfare outcomes from a wide range of routine management and husbandry procedures. For example, it has been employed for assessing and monitoring the stress in cattle from transport (Schaefer et al., 1988), the pain and stress in Wapiti (Cervus elaphus canadensis) and reindeer (Rangifer tarandus) from velvet-antler removal (Cook et al., 2005), the stress response in dogs from veterinary examination (Travain et al., 2015), the stress in chickens from being handled (Edgar et al., 2013), the changes in the skin temperature of weaned piglets after vaccination (Cook et al., 2015), and the pain and stress responses in cattle after castration and disbudding (Stewart et al., 2010, 2009, 2008b). Thermography has also been employed to assess the extent of tissue damage and duration of inflammation after different methods of cattle branding (Schwartzkopf-Genswein et al., 1998, 1997) and to assess the pain responses of heifers after tail docking (Eicher et al., 2006), and of piglets after clipping and grinding of teeth (Moya et al., 2006). It is likely that IRT would be useful for determining the degree of injury and inflammation, as well as the pain and distress, associated with many other invasive procedures such as ear cropping and de-beaking (Stewart, 2008). Comprehensive reviews of the use of IRT have been published for farm animals (Stewart et al., 2005) and animal research (McCafferty, 2007).     56 1.4.3 Peripheral vasoconstriction and affective state Short-term temperature responses due to peripheral blood flow are associated with the activation of the autonomic nervous system (ANS) and may be a physiological correlate of an animal’s affective state in response to painful, stressful or arousing stimuli (Stewart, 2008). The amount of blood-flow through peripheral vessels influences the temperature of the skin and extremities; thus changes in blood flow as a result of a stress response will alter the amount of radiated heat from such sites. In humans, changes in blood flow, measured as small changes in skin temperature, have been shown to correspond to both negative and positive alterations in affective state (Khan et al., 2009). For example, eye temperature increased when people told lies (Pavlidis et al., 2002; Pavlidis and Levine, 2002) and in response to fear, with a simultaneous decrease in cheek temperature (Levine et al., 2001). When infants were considered to experience a joyful emotion, their facial temperature briefly decreased (Nakanishi and Imai-Matsumura, 2008). Similar responses have also been recorded in non-human animals. For example, changes have been recorded in the temperature of the skin and extremities of rats (Vianna and Carrive, 2005), mice (Lecorps et al., 2016), horses (Dai et al., 2015) and Rhesus macaques (Macaca mulatta) (Nakayama et al., 2005) due to fear; of domestic rabbits in response to alerting stimuli (Blessing, 2003); of sheep (Lowe et al., 2005), cattle (Stewart et al., 2008b, 2007b) and horses (McGreevy et al., 2012) due to aversive procedures; and of chickens in response to both positive (food reward) and negative (handling) situations (Edgar et al., 2013; Moe et al., 2012).     57 1.4.4 Changes in eye temperature Temperature variation at multiple sites on the body has been measured and used for the evaluation of the stress response. Examples of such body sites are the nasal cavity of rhesus macaques (Macaca mulatta) (Nakayama et al., 2005), the pinna of sheep and rabbits (Blessing, 2003; Lowe et al., 2005), the dorsal area of Wapiti (Cervus elaphus canadensis) (Cook and Schaefer, 2002), the tail and paw of rats (Vianna and Carrive, 2005) and the head and comb of chickens (Edgar et al., 2013). Ludwig et al. (2007) reported the eye region as one of the best sites for measuring the stress response, and recent studies have demonstrated that multiple mammalian species undergo a change in eye temperature in response to stressful and painful procedures (Cook and Schaefer, 2002; Stewart et al., 2010, 2009, 2008b). Changes in eye temperature are reflective of the activity of the sympathetic nervous system (SNS) and the hypothalamic-pituitary-adrenal axis (HPA) and thus have been proposed as a useful index of the stress response (Stewart et al., 2010).  Animals have multiple thermal windows (e.g., eyes, ears, nasal cavities, flippers) from which heat is lost to the environment. The eye area is a particularly practical thermal window for IRT measurements because it is present in (virtually) all species; it is easily accessible; is free from insulative pelage and plumage, and is a region with a multitude of capillary beds innervated by the SNS (McCafferty, 2007). The areas of the lacrimal caruncle and palpebral border of the lower eyelid have been recognized as being particularly sensitive to changes in blood flow from the stress response (Pavlidis and Levine, 2002; Stewart et al., 2008b). In addition to being an easy location on most animals to measure, the eye seems to show temperature changes more consistently   58 than other anatomical areas.  For example, when monitoring eye temperature in cattle, bovine viral diarrhea could be detected as early as day 1 of infection compared to day 5-6 when monitoring other anatomical areas such as the nose, ears, body or hooves (Schaefer et al., 2004).  It has been suggested that the change recorded in eye temperature could be due to tears in the eye and subsequent evaporative heat loss. However, seeing that the eye is a small area and that water has an emissivity of 0.96, which is close to that of bare skin, it is unlikely that any moisture on the eye surface would have much effect on the amount of heat radiated from the area (Stewart, 2008).  1.4.5 Eye temperature and the stress response A rapid drop in eye temperature followed by a prolonged increase has been demonstrated in multiple species in response to a range of stressors. This response suggests that changes in eye temperature are mediated by the activation of both the SNS and the HPA axis. Cook et al. (2001) were the first to investigate the relationship between infrared radiation and HPA activity by measuring cortisol levels in blood and saliva and matching them to IRT images of horses’ eyes before and after an adrenocorticotropic hormone (ACTH) challenge. Their study showed a significant correlation between maximum eye temperature and cortisol concentrations signifying that the HPA axis is likely associated with increases in eye temperature (Cook et al., 2001). Cortisol concentrations and IRT images of the eye have also been correlated to assess pain and fear in cattle (Stewart et al., 2010, 2009, 2007a), stress in rabbits   59 (Ludwig et al., 2007) and the stress experienced by horses during competition (Valera et al., 2012).  To further investigate the mechanism driving temperature change in the eye, Stewart et al. (2007b) subjected cows to a series of exogenous stress challenges, including the administration of ACTH and bovine corticotropin-releasing hormone (bCRH). Eye temperature remained unaffected by both ACTH and bCRH, leading the authors to conclude that although the HPA axis was activated, it was not driving the change.  In the same trial, on two separate occasions, calves underwent jugular catheterization (Stewart et al., 2007b), which is known to be aversive for cattle (Alam and Dobson, 1986). Calves did not demonstrate a significant change in eye temperature or cortisol concentration after initial catheterization, but a significant increase in both measures was recorded following the second catheterization one week later.  The authors suggest the possibility that a cognitive awareness of the stressor might be required for an eye temperature response to occur. Specifically, the calves may have demonstrated a response the second time they were catheterized because they likely anticipated what they had learned would be an aversive experience (Stewart et al., 2007b).  In a subsequent study, Stewart et al. (2008) examined the possibility that pain from disbudding via hot-iron could be assessed with eye-temperature change and heart-rate variability (HRV). Calves disbudded without a local anaesthetic had a rapid drop in eye temperature in the 2 to 5 minutes after the procedure (-0.27 ᵒC) whereas animals disbudded with local anaesthetic showed only a small non-significant decrease. Eye   60 temperature then increased above baseline in animals disbudded with and without the local anaesthetic (+0.6 and +0.66 ᵒC respectively). There was no significant difference in eye temperature of control animals. When the disbudded calves were subjected to an ACTH challenge one week later, there was no change in eye temperature, confirming that eye-temperature change following disbudding was not due to HPA activity alone. A change in HRV (increased low-frequency to high-frequency ratio (LF:HF)) was seen only in the animals disbudded without local anaesthetic, suggesting an acute, sympathetically mediated response to the procedural pain.  The resulting vasoconstriction from the SNS response then may have caused the decrease in eye temperature (Stewart et al., 2008b).   In a similar trial, Stewart et al. (2009) extended the observation time to include the period when the local anaesthetic would wear off (2-3 hours after administration). In this 2-3 hour period, animals that had been disbudded with a local anaesthetic showed a drop in eye temperature (-0.6 ᵒC) similar to those animals disbudded without a local (Stewart et al., 2008b). This study indicated that eye temperature drops in response to the immediate pain of the procedure and also when pain commences as the mitigating effects of the anaesthetic wear off (Stewart et al., 2009). After surgical castration, a similar trend was seen in which calves experienced a rapid, but in this case non-significant, decrease in eye temperature followed by a significant increase that was maintained throughout the 20-minute post-procedure observation period. Eye temperature increased in calves that were castrated with and without local anaesthetic, but the change was greater for those castrated without (+0.28 ± 0.05 ᵒC and +0.47 ± 0.05 ᵒC, respectively). Calves castrated without the local   61 anaesthetic also demonstrated a rapid increase in heart rate, a decrease in LF:HF ratio, and an increase in RMSSD (root mean square of successive beat-to-beat differences) and HF power indicating an increase in parasympathetic activity (Stewart et al., 2009; von Borell et al., 2007). The ANS is activated in response to stressors. When the SNS is activated, there is a rapid diversion of blood from cutaneous areas to areas of more urgent requirement (the muscles and brain), for an immediate fight-or-flight response. This change in perfusion also has a protective function in minimizing the risk of blood loss from injury to vulnerable areas (the skin and extremities) (Blessing, 2003) (see section 1.1). This sympathetically mediated vasoconstriction is believed to drive the initial, rapid drop in eye temperature. The rapidity of the drop in eye temperature when an animal is shocked with an electric prod, or startled, provides further evidence that this is the mechanism in effect (Stewart et al., 2007a). Interestingly, although animals’ eye temperature decreased immediately after being administered a painful electric prod or frightened by shouting or exposure to the flapping of a bag, the animals that had experienced acute pain took much longer to return to baseline eye temperature than those that had experienced fear.  This reaction suggests that IRT has the potential to discriminate between stressors such as fear and pain (Stewart et al., 2007a). The mechanism is less clear for the increase in eye temperature that follows the initial sympathetically mediated decrease. Stewart et al. (2008) confirmed that the increase was not a result of heat from the hot iron, physical activity or local anti-inflammatory processes, and suggested it could be due to the parasympathetic nervous system lowering cardiac output and blood pressure resulting in vasodilation. They also   62 suggested the possibility that the release of vasodilators such as nitric oxide or those known to be associated with pain (adrenaline, prostaglandins, bradykinins and histamine) (Mellor et al., 2000) may contribute.  When calves were castrated without local anaesthetic, the changes in HRV revealed an increase in parasympathetic tone that corresponded in time to the increase in eye temperature (Stewart et al., 2009). When the HPA axis is activated by a stressor, the release of cortisol can last from minutes to hours, resulting in multiple thermo-genic reactions in tissue metabolism, which may also contribute to noted temperature increases (Valera et al., 2012).  It is, therefore, probable that the increase in eye temperature is a result of both vasodilation due to parasympathetic activity and activity of the HPA axis (Cook et al., 2001; Valera et al., 2012).  1.4.6 Advantages and limitations The main advantage of IRT is that it is non-invasive, enabling an assessment  of stress and animals’ stress responses without physical contact. IRT can precisely measure dynamic, cutaneous changes in temperature due to changes in metabolic activity in real-time. Measurements can be taken from various distances. Additionally, thermography is more sensitive than manual palpation in detecting skin temperature differences of <0.1 ᵒC (McCafferty, 2007). Nonetheless, while recent technological advancements in IRT cameras have made the technology more accessible and easier to use than in the past, it is not without limitations. One of the main disadvantages of IRT is that the emissivity of an object can be greatly affected by solar radiation, meaning that all images must be collected out of direct   63 sunlight. This poses logistical challenges in many research settings. Additionally, emissivity can be affected by foreign material such as dirt or moisture. Water has an emissivity of 0.96, bare skin of 0.98, and the emissivity of dry fur is consistently between 0.98 and 1.0 across mammals (McCafferty, 2007). If an animal’s coat is wet, the greater thermal conductivity of water may affect the conduction of body heat so that local heat loss is increased to drier regions of the animal’s coat and to the environment (Palmer 1981). Weather conditions, wind drafts, lighting and radiative surfaces in the surrounding area may also interfere with accurate measurements. Therefore, when using thermography efforts must be taken to account for factors that could affect emissivity and conductivity and result in extraneous variation. Importantly, although IRT has been successful in helping detect injury or disease, it can identify thermal anomalies but not specific pathologies (Redaelli et al., 2014).   1.4.7 Conclusions For many years, IRT has been used successfully for the identification of injury and disease in veterinary medicine (McCafferty, 2007) and, increasingly, to measure stress responses in animals by changes in eye temperature. It has the advantage of enabling a non-invasive estimation of the stress response (Stewart et al., 2005). There is evidence that both physiological and psychological stress are indicated by changes in eye temperature (Cook et al., 2001; Pavlidis and Levine, 2002) and that changes in eye temperature reflect changes in both the ANS and HPA axis.  In representing more than one component of the stress response, the measurement of eye temperature via IRT has great potential for monitoring and evaluating many aspects of animal health and   64 welfare, especially those husbandry and handling procedures known to cause pain and result in a stress response.  1.5 Thesis aims In writing this introduction, I was particularly struck by the paucity of reliable, non-invasive methods for the assessment of pain in animals.  Both grimace scales and IRT technology are relatively new ways of quantifying animal pain and although the use of vocalizations to indicate pain has been established in multiple species (e.g., pigs; Weary et al., 1998), vocal behavior in pinnipeds has mainly focused on its function and phylogeny (Ravignani et al. 2016). Changes in facial expression, eye temperature and vocalizations to indicate pain are particularly interesting to me as their application does not require physical contact or invasive sampling.  Grimace scales and vocalizations have been shown to be reliable measures of spontaneous, acute pain in several species but have yet to be developed for any species of pinniped. Similarly, the use of IRT to measure pain and stress responses has been focused primarily on domestic species and used to assess a limited number of painful procedures.   For the last decade, I have worked as a marine mammal rehabilitator at the Vancouver Aquarium Marine Mammal Rescue Centre.  The majority of our patients are neonatal harbour seal pups (Phoca vitulina) that are admitted to our centre following maternal separation.  During their time in our care, the primary focus is on their health and functioning, with much less attention given to other welfare concerns such as pain and distress. Typically, pain is managed only in animals that have severe injuries or are reported by staff to be in pain.  This approach is ad hoc and problematic. No systematic   65 method exists to determine which animals are in pain, and to date, no clear species-specific indicators of harbour seal pain have been established.  Additionally, seals in rehabilitation are subjected to a multitude of stressors (e.g., gavage feeding, restraint, blood sampling) as well as routine procedures that likely cause pain, such as tagging and chipping.   The general objective of this thesis was to identify some possible indicators of pain in harbour seals. In each experimental chapter I used the routine procedures of flipper tagging and microchipping as my potentially painful stimuli and aimed to determine whether pups showed a change in facial expressions and/or behaviours (Chapter 2), a change in vocalizations (Chapter 3) and a change in eye temperature (Chapter 4) in response to the procedures. I also tried to determine whether the observed responses were the result of tagging and chipping, as opposed to possible effects of restraint (Chapters 2,3,4) and lastly, whether the observed responses would be mitigated by administration of an analgesic (Chapters 2 and 4). Because the three chapters were written as scientific papers, there is some repetition of background and relevant literature from Chapter 1.     66 2. Initial evaluation of facial expressions and behaviours of harbour seal pups (Phoca vitulina) in response to tagging and microchipping  2.1 Introduction  Behavioural responses including postural changes are commonly used for assessing animal pain (Rutherford, 2002). More recently, specific facial expressions associated with pain have been identified in multiple species of non-human animals, and are proposed as promising for the assessment of animal pain (Flecknell, 2010).  Facial expressions in humans have been extensively researched as indicators of affective states (Ekman, 1993), and are recognized as a valuable tool for assessing pain (Craig et al., 2011), including with subjects who are unable to self-report (Anand and Craig, 1996). For example, pain in human neonates is assessed by the Neonatal Facial Coding System (NFCS) (Grunau and Craig, 1987). NFCS is adapted from the Facial Action Coding System, which is an objective, anatomically based behavioural coding scale that identifies minimal units of facial movement called Action Units and yields detailed descriptions of the resulting changes in facial appearance (Ekman and Friesen, 1978).  This approach has provided the basis for a framework for pain assessment in animals, and has resulted in grimace scales being developed for mice (Langford et al., 2010), rats (Sotocinal et al., 2011), rabbits (Keating et al., 2012), horses (Dalla Costa et al., 2014), cats (Holden et al., 2014), sheep (McLennan et al., 2016), pigs (Di Giminiani et al., 2016) and ferrets (Reijgwart et al., 2017). Facial expressions provide a non-invasive means of pain assessment that exploits the natural tendency of humans to   67 focus on the face when gauging pain (Williams, 2002), and can be used to assess pain ranging from mild to severe (Leach et al., 2012). Moreover, facial expressions in response to pain are well conserved between species, although some species-specific actions have been identified (Chambers and Mogil, 2015).   Pinnipeds are commonly kept in captivity, and harbour seals (Phoca vitulina) are a common species admitted to rehabilitation facilities (MacRae et al., 2011). In the Southern North Sea (Denmark, Germany, Netherlands and the UK) more than 2000 harbour seals were admitted to rehabilitation centres between 2000 and 2005 (Reijnders et al., 2009) and since 2014, the Marine Mammal Center in California, USA has rehabilitated and released more than 1900 pinnipeds, many of which were harbour seals (The Marine Mammal Center, 2018). In Vancouver, Canada, close to 200 harbour seal pups are admitted to the Vancouver Aquarium Marine Mammal Rescue and Rehabilitation Centre (MMR) each year.   Pinnipeds are routinely marked for identification by methods that include external tags attached to flippers or more invasive procedures such as scarring (toe clipping), hot-iron branding, or surgical implantation of telemetry devices (Walker et al., 2012). While these procedures are likely painful (Walker et al., 2009), there is little work on pain in any species of pinniped. Consequently, pain therapy for these species may be inadequate or inconsistent (Flecknell, 2000). In a rehabilitation environment, the welfare of large numbers of animals may depend on the ability to rapidly assess pain ‘cage-side’. Facial expressions have potential for relatively easy, real-time, non-invasive identification of pain (Leach et al., 2012; Leung et al., 2016), and could therefore provide a practical method for identification of pain in pinnipeds.   68  Seal pups admitted to MMR are hand-raised and, once deemed healthy and capable of independent survival, are released. Before release, the Canadian Department of Fisheries and Oceans (DFO) requires that each pup have an identification tag placed in a hind flipper (‘tagging’) and a microchip placed at the base of the tail (‘chipping’).  Adverse, unintended consequences are possible.  In rodents, inflammation and neoplasia have resulted from both ear tagging (Waalkes et al., 1987) and chipping (Blanchard et al., 1999). After flipper tagging, gray seals (Halichoerus grypus) had swelling, exudate and partially open wounds at the tag site for up to 24 days (Paterson et al., 2011). Chipping is known to cause immediate sensitivity and visible inflammation at the chip site for up to three days in horses (Gerber et al., 2012) and swelling around the site for up to a week in guinea pigs, rabbits and woodchucks (Marmota monax) (Mrozek et al., 1995). Both procedures cause tissue damage by piercing soft tissues. Similar procedures in other species are associated with pain; examples include ear piercing in humans (Spafford et al., 2002), ear tagging and microchipping in rodents (Dahlborn et al., 2013) and ear tagging (Leslie et al., 2010) and ear notching in piglets (Torrey et al., 2009). Additionally, when tagged and chipped, seals have anecdotally been observed to flinch, vocalize, exhibit escape behaviours, and curl flippers after tagging, which are consistent with behavioural responses to pain in multiple other species (National Research Council, 2009; Rutherford, 2002).   In harbour seals, little to no research exists on sensitive and specific behavioural indicators of pain, including those that may be species-specific. Therefore, in this study we looked for consistent facial expressions (characterized by specific facial action units) and behavioural changes in response to the routine, potentially painful procedures of   69 tagging and chipping. The aims of the study were to determine: 1) whether pups showed a change in facial expressions and/or behaviours after tagging and chipping, 2) whether the observed responses were the result of tagging and chipping, as opposed to possible effects of restraint, and 3) whether the observed responses would be mitigated by administration of an analgesic.  2.2 Materials and methods The research was approved by the University of British Columbia Animal Care Committee (Protocol A12-0095) and by the Vancouver Aquarium Animal Care Committee. Some of the seals included in this protocol were simultaneously used in a study of vocal responses (Chapter 3).   2.2.1 Animals and housing Forty-seven harbour seal pups (24 males, 23 females) had been recovered along the coastline of British Columbia, Canada, by MMR staff, or were brought to the facility by members of the public, between June and September of 2012. Data were collected between August and October 2012, once pups were weaned, estimated to be more than 60 days old, and considered to be free of disease or injury. Mean (±SEM) body weight at the time of testing was 14.75 ± 0.5 kg (range 9.4 kg–24.6 kg).   Animals were kept at MMR following the standard operating procedures of the facility. Animals were housed in groups of up to 8 in fiberglass pre-release pools of approximately 23,000 L (4.87 m diameter X 1.21 m depth, with a 2.34 m X 1.21 m haul-out). Pups were scatter-fed whole herring three times per day at approximately 12% of   70 body weight. 2.2.2 Tagging and chipping Between 07:00 and 09:00 and before their first daily feed, pups were taken from their pool to a quiet outdoor area that was out of visual and auditory range of other staff and animals. The area was sheltered from direct sunlight to ensure there were no shadows on pups’ faces or any glare that could affect facial appearance or expressions.  Individuals were placed on a foam mat where they were restrained by a person who held the seal in ventral recumbency, with the animal’s body between the restrainer’s knees, the head between the restrainer’s feet and hind flippers secured in the restrainer’s hands.  The animals were then tagged and chipped, in that order. For tagging, a 5 cm plastic tag with an animal identification number and DFO contact information was attached by piercing the webbing of the 2nd and 3rd digits of the hind flipper. Males were tagged on the right hind flipper and females on the left. For chipping, a 2.1 x 11.5 mm microchip which comes preloaded in a 2.6 x 40 mm sterile single use cannula (TrovanÒ ID-162B, Germany), was implanted subcutaneously on the dorsal aspect of the base of the tail. Following standard practice, no analgesic was provided for either of these procedures. The whole procedure lasted approximately 6 min: 2 min of baseline restraint before the procedure, about 2 min for tagging and chipping, and 2 min of restraint after the procedure. During this 6-min period, each animal was recorded using a Canon HG10 HDD camcorder (Canon USA Inc., New York, USA), from a distance of approximately 1.2 m. The camera was placed approximately 15 cm above the ground and directly in front of, and level with, the animals’ faces. The camera was stabilized   71 with a tripod (JobyÒ GorillaPod, San Francisco, USA) and maneuvered by hand when required. Only the head and neck of the pups were visible in the camera frame. The same three people restrained the seal, performed tagging and chipping, and filmed the procedure for all animals.   2.2.3 Development of ethogram  To identify the facial action units (FAUs) and behaviours seen during the procedure, eight seals undergoing tagging and chipping were video-recorded. These eight pups were a convenience sample chosen by MMR staff from animals that were weaned, estimated to be more than 60 days old, and considered to be free from disease or injury. Two observers watched each of the videos multiple times independently, and recorded all FAUs and other behaviours observed. Observers then watched the videos together to ensure agreement on recognition and definitions of all elements. After refining the definitions, both observers again scored the videos from all eight seals to ensure that the described behaviours were scored consistently. These eight seals are not included in the experiments below.  2.2.4 Experiment 1 – FAUs and behaviours before and after tagging and chipping The aim of Experiment 1 was to determine if pups show a change in FAUs and other behaviours as a result of the procedures. To achieve this, 19 seal pups undergoing tagging and chipping were recorded following the procedure outlined above. There were three test days: six pups were tested on the first day, six on the second day and seven   72 on the third day. FAUs and behaviours were compared before versus after the procedure.  2.2.5 Experiment 2 – Cross-over experiment The aim of Experiment 2 was to determine whether the FAUs and behaviours observed to change in Experiment 1 were the result of tagging and chipping or simply effects of the associated restraint.  A cross-over experiment with 10 pups was conducted over 2 days, spaced 5 days apart. On day 1, all animals were restrained as outlined above, but every second pup was actually tagged and chipped, while the others were ‘sham’ tagged and chipped. On day 2, the five animals that had been sham tagged and chipped on day 1 were tagged and chipped, and the five pups that had been tagged and chipped underwent the sham procedure.  Everything was identical for the sham procedure: seals were restrained and their tagging and chipping sites were cleaned, but instead of tagging, the tagging unit was touched to the flipper without piercing it (the sound from depressing and releasing the device was still audible), and instead of chipping, a syringe without a needle was briefly placed in contact with the chipping site.  2.2.6 Experiment 3 – Analgesic pilot study Experiment 3 was intended as a pilot study to test whether the responses observed in Experiments 1 and 2 could be reduced by pain mitigation. To achieve this, 10 animals were tagged and chipped on the same day using the procedure outlined above. One hour before the procedure, five of the pups received a single intramuscular dose (0.01mg/kg) of buprenorphine (Buprenorphine 0.3mg/ml, VetergesicÒ, Sogeval UK   73 Limited), and five received a saline injection.  Injections were delivered into the femoral biceps using a 22-G x ¾” (0.71 mm x 1.9 cm) needle (Kendall, Mansfield, Massachusetts 02048, USA) on a 1-ml syringe (Terumo Medical Corporation, Somerset, New Jersey 08870, USA). As there is no validated analgesic protocol for acute pain in seals, the drug and dosage were suggested by the staff veterinarian as an empirically safe dose, based on the low end of the recommended dosage for analgesia in dogs (0.01 – 0.03 mg/kg), but it was not known whether it would provide pain control.   2.2.7 Video analysis For each 6-min video recording in Experiments 1, 2 and 3, two 90-s video clips (segments) were extracted using Windows Movie Maker (Windows 7). The first 90-s clip started when seals were restrained before tagging and chipping (‘before’) and ended exactly 5 s before the procedure began. The second 90-s clip (‘after’), began exactly 5 s after the procedure ended. The resulting 90-s video clips were randomized and renamed to conceal their origin (seal, date, before or after procedure). As only the head and neck region of seals was filmed, there were no visible cues that could indicate to the observer if clips were before or after the procedures. The 2-min period during which pups were tagged and chipped was not included in the analysis. This was because pups often vocalized and struggled during the procedures, thus making the possible change in FAUs difficult to assess, and because there were slight variations in timing of the tagging and chipping. The 38 clips from Experiment 1 (19 ‘before’ and 19 ‘after’, renamed and randomized) were each watched twice in real time by two observers. The first time,   74 observers made a subjective overall judgement of whether they thought, from the pup’s responses, that pain was present or absent in the featured pup. The second time breathing rate was recorded by counting movement of nares and audible inspiration/exhalation. Next, the clips were watched in VLC media player (Version 2.2.1) at 25% - 50% speed between eight and 25 times, depending on the activity level of the individual in the clip. Each FAU and behaviour from the ethogram was scored separately and marked as either present or absent for each second of the 90-s clip, thus yielding a ‘one-zero score’ ranging from zero to 90 for each behaviour (Martin and Bateson, 2007). Observers did not know whether they were scoring animals before or after the procedure. Both observers had several years of experience with scoring animal behaviour; however, one observer was considered to be ‘experienced’, having worked with harbour seals as a rehabilitator and in a research capacity for eight years, and the other ‘inexperienced’, having had no direct experience with this species.   Only the experienced observer scored the 40 clips from Experiment 2 (10 ‘before’ and 10 ‘after’ for day 1 and 10 ‘before’ and 10 ‘after’ for day 2, renamed and randomized), and the 20 clips from Experiment 3 (10 ‘before’ and 10 ‘after’, renamed and randomized). In Experiment 2, video data were corrupted for one animal, leaving only nine animals in the analysis.   2.2.8 Statistical analysis For each 90-s clip, the total one-zero score (the number of seconds when the behaviour was seen) was tabulated for each FAU and behaviour. Mean ± standard error of the   75 mean (SEM) was used for all results. A paired-samples t-test was used to compare the score for each FAU and behaviour before versus after the tagging and chipping procedure. Breathing rate was counted as breaths per minute and also compared before and after the procedure. Wilcoxon signed-rank test was used to compare the nose bulge behaviour, as it was seen in only some seals and did not meet the criteria for parametric analysis. Changes were considered significant at p ≤ 0.05.  In Experiment 1, inter-observer reliability was calculated for each FAU and each behaviour using Pearson correlations. All analyses were conducted in R version 3.2.2 (R Core Team, 2015).  2.3. Results   2.3.1 Ethogram The resulting ethogram included 15 FAUs and five other behaviours (Table 2.1). Photographs illustrating some of the FAUs are in Figure 2.  76 Table 2.1. Ethogram of all facial action units (FAUs) and behaviours observed before and after tagging and chipping of harbour seal pups..  FAU/behaviour Description Eye Change  Eyes closed  Eyes tightly closed. Narrowing of orbital area with tightly closed eyelid or eye squeeze (often denoted by wrinkle around or under the eye) Orbital tightening Narrowing of orbital area, partial or complete closure of eye; eye takes on an almond shape, may have skin wrinkle around or under the eye; nictating membrane may be visible Blinking Eyes briefly close (partially or fully) with no pause before returning to full open Eyes open Eyes wide open, eye has rounded shape Nose Change   Nose bulge Visible tightening and rounded, convex bulge in the area under the eyes and above the whiskers; muzzle may appear shorter and wider Nares open Nares wide open and oval shaped; nares are oriented almost horizontally creating a wide "U" shape; air being inhaled  Nares partially open Nares open but narrow; nares are oriented more vertically creating a "V" shape; this occurs briefly as nares are going from closed to open; air being inhaled Nares closed Nares pressed shut, nares are oriented almost vertically creating a narrow 'V' shape; air being held in  Mouth Change   Closed  Mouth closed with lips touching, teeth and tongue not visible Open mouth without vocalization Mouth opens so that bottom and top lips separate, tongue and/or bottom incisors and canines visible; no audible vocalizations  Partially open mouth with vocalization As seal is vocalizing, mouth opens so that bottom and top lips separate, tongue and/or bottom incisors and canines visible Open mouth with  vocalization As seal is vocalizing, mouth wide open so that upper and lower jaws are stretched, tongue and full set of bottom teeth visible, top teeth and upper palate may be visible Closed mouth with  vocalization Mouth closed with lips touching, teeth and tongue not visible; making audible vocalizations Whisker Change   Whiskers forward Whiskers are lifted and away from face, often positioned horizontally or pushed forward, become erect and separate from each other; gives a stiff 'standing on end' impression Whiskers backward  Whiskers have a gentle downward curve, slightly oriented towards the face, have a 'relaxed' appearance Other Behaviours   Struggling Struggling against restraint, often moving head rapidly back  and forth and/or left or right or pulling head into neck Biting Bites or attempts to bite handler (body, boot, pants etc.) Trembling Small rapid vibrating/shaking movement of entire head and/or body Looking around  Head turns to follow movement in surrounding environment; eyes are focused and track environmental stimuli; often whites of eyes are visible   Open-mouth breathing Rhythmic opening and closing of mouth as animal breathes; nares open simultaneously as mouth opens; seal is not vocalizing but intake of air may be audible 	 	   77 Figure 2.1. Facial action units of harbour seals (Phoca vitulina) recorded in this study. Blinking and vocalizing are not included. Detailed descriptions of each FAU are included in Table 2.1.    78 2.3.2 Experiment 1 – FAUs and behaviours before and after tagging and chipping For both observers, the subjective judgement of whether pain was present or absent corresponded closely to whether pups had or had not undergone tagging and chipping. Representative still images taken from one seal’s 90-s videos before and after tagging and chipping are shown in Figure 2.2.     Figure 2.2. Experiment 1 – Facial expression of one seal before (left) and after (right) tagging and chipping. The two photos are stills taken from the pup’s 90-s video clips.     79 The experienced observer had 95% correspondence, with one false positive (pain scored as present for a pup that had not been tagged and chipped) and one false negative (pain scored as absent for a tagged/chipped pup) out of 38 clips. The inexperienced observer had 89% correspondence with two false positives and two false negatives.  Breathing rate was lower in the period after tagging and chipping (t= -2.45, df = 18, p < 0.03). The mean ± SEM rate was 29 ± 1 breaths min–1 (range 19-39 breaths min–1) before and 24 ± 1 breaths min–1 (range 13-35 breaths min–1) after the procedure.  Pups displayed an increase in orbital tightening after the procedure (t = 10.80, df = 18, p < 0.001). This FAU occurred in 27 ± 4 of the 90 1-s intervals before compared to 72 ± 4 intervals after tagging and chipping. From before to after the procedure, scores for blinking decreased from 1 ± 0 to 0 (t = -3.52, df = 18, p < 0.01), looking around from 21 ± 4 to 6 ± 2 (t = -5.92, df = 18, p < 0.001), trembling from 2 ± 1 to 0 (t= -2.91, df = 18, p < 0.01) and struggling from 14 ± 3 to 5 ± 1 (t= -3.43, df = 18, p < 0.01) (Figure 2.3)  80     Figure 2.3. Scores for facial expression and other behaviour (mean ± SEM) before and after tagging and chipping for the 19 seals in Experiment 1. *** p ≤ 0.001, ** p ≤ 0.010102030405060708090OrbitaltightenBlink Look Tremble StruggleMean ±SEM score BehaviourBeforeAfter*** ** *** ** **    81 Nose bulge was seen in only seven of the 19 animals, but was displayed only after tagging and chipping. For this subset of seven animals, nose bulge was different from before the procedures (score of 0) to after (16 ± 7; range 2 to 51; p < 0.01).  There were no significant differences between before and after the procedure for any other FAUs or behaviours in the sample of 19 seals.   Inter-observer reliability ranged from r = 0.82 to 0.92 (p < 0.01) for the five responses that changed significantly after tagging and chipping: orbital tightening (r = 0.89), blinking (r = 0.82), looking around (r = 0.9), struggling (r = 0.9), and trembling (r = 0.92), and ranged from r = 0.84 to 1 (p < 0.01) for all other facial expressions and behaviours.  2.3.3 Experiment 2 – Cross-over experiment When pups received the sham treatment, there were no differences in FAUs or behaviours from before to after the treatment. In contrast, when these pups underwent actual tagging and chipping, their responses were similar to those in Experiment 1.  From before to after the procedures, scores for orbital tightening increased whereas blinking, looking around and struggling declined after the procedures (Table 2.2). Trembling was not seen. Nose bulge was not seen before the procedure nor after the sham procedure, but occurred in five of the nine animals after tagging and chipping (4 ± 1 after versus 0 before; p < 0.05). Breathing rate was lower in the period after both sham (t= -3.88, df = 8, p < 0.005) and actual tagging and chipping (t= -2.85, df = 8, p < 0.05). For the sham procedures, the mean ± SEM rate was 28 ± 3 breaths min–1 (range 15-47 breaths min–1)   82 before and 22 ± 2 breaths min–1 (range 10-32 breaths min–1) after. For the actual procedures, the rate was 29 ± 2 breaths min–1 (range 21 to 37 breaths min–1) before and 24 ± 2 breaths min–1 (range 19-36 breaths min–1) after.   83 Table 2.2 Scores for facial action units and other behaviours (mean ± SEM) before and after sham or actual tagging and chipping for the 9 seals in Experiment 2.   FAU/Behaviour  Sham   Tagging/chipping  Before  After  t  df  P   Before  After  t  df  P  Orbital tightening  21 ± 3  36 ± 4  2.17  8  p = 0.06   25 ± 2  73 ± 3  14.32  8  p < 0.001  Blinking  1 ± 1  1 ± 1  - 0.19  8  p = 0.9   1 ± 1  0 ± 0  - 2.68  8  p < 0.05  Looking around  30 ± 4  22 ± 3  - 1.66  8  p = 0.1   26 ± 3  6 ± 1  - 3.81  8  p < 0.01  Struggling  9 ± 1  9 ± 1  - 0.10  8  p = 0.9   13 ± 2  5 ± 1  - 5.01  8  p < 0.05   84 2.3.4 Experiment 3 – Analgesic pilot study There was no difference between the five pups given analgesic compared to the five pups without analgesic.  When the treatment groups were combined (n = 10), behavioural responses before and after tagging and chipping followed a similar pattern to those seen in Experiments 1 and 2: orbital tightening increased from 21 ± 5 to 67 ± 5 (t = 13.14, df = 9, p < 0.001), looking around decreased from 37 ± 7 to 8 ± 2 (t= -5.33, df = 9, p < 0.001) and struggling decreased from 9 ± 3 to 3 ± 1 (t = -2.23, df = 9, p < 0.05). None of the other behaviours were different before versus after tagging and chipping. Nose bulge was displayed by two of 10 pups, in both cases after being tagged and chipped. Breathing rate was lower after tagging and chipping (t= -3.38, df = 9, p < 0.01). The mean ± SEM rate was 31 ± 2 breaths min–1 (range 20-44 breaths min–1) before and 25 ± 2 breaths min–1 (range 17-35 breaths min–1) after the procedure.   2.4. Discussion Collectively the findings of this study showed that harbour seals consistently react to a presumably painful stimulus with observable behavioural changes, including quantitative changes in orbital tightening and declines in looking around and struggling. Nose bulge is a potential indicator of pain, but because it appeared in only some of the seals, it needs further investigation before it can be considered a reliable response to the procedures.    Seals possess the facial musculature to generate facial change and are known to have complex facial expressions in various social and non-social contexts (Miller, 1975).    85 Since facial expressions are well conserved between species (Chambers and Mogil, 2015), it is unsurprising that harbour seals demonstrate changes in facial expressions in response to potentially painful procedures. The orbital tightening and nose bulge seen in the seals after tagging and chipping are similar to changes described in other species in pain. In fact, all the animals for which grimace scales have been developed, including humans, share characteristics of a ‘pain face’ typified by orbital tightening and for many, a change in the nose/cheek area (Chambers and Mogil 2015).  Humans tend to focus on the face and, in particular, the eyes of other people when attempting to gauge affective states such as pain (Deyo et al., 2004). This attention to facial change seems to transfer to and assist our assessment of pain in other species (Leach et al., 2011). Of the FAUs and behaviours observed in seals, orbital tightening showed the strongest response, with an almost three-fold increase from before to after tagging and chipping in all three experiments. The eyes of harbour seals are prominent, show a marked change in response to the procedures, and because they are objects of our inherent observational bias, they may be a particularly useful indicator when assessing seal pain in a ‘cage-side’ situation, for example when monitoring recovery from injury.   The FAU we called nose bulge was shown only after tagging and chipping and was seen in only some pups in each experiment. Nose bulge was seen in seven of the 19 pups in Experiment 1, in five of the nine pups in Experiment 2, and in two of the 10 pups in Experiment 3. Pain responses may vary among individuals (Livingston and Chambers, 2000), and it is possible that the variation in nose bulge reflects such individual differences. As this FAU was never seen before tagging and chipping, nor   86 after the sham procedure, its occurrence seems a potential indicator of pain but one that is not seen in all animals. However, due to its low occurrence, this FAU needs further investigation before being considered a valid indicator of pain in this species. Despite commonalities with other species in facial expressions in response to pain, harbour seals have some unique adaptations that make some of the FAUs in existing grimace scales inapplicable. For example, ear position, which is controlled by the facial muscles, is included in all currently published grimace scales for other species but was not useful for harbour seals because of their lack of external ear pinna.  In rabbits, the shape of the nostrils can be a sign of pain (Keating et al., 2012). However, seals can fully open and close their nares as they breathe (Figure 2.1), likely as an adaptation to diving. Hence, changes in nostril position or shape often occur during simple breathing. Finally, changes in whisker position are indicative of pain in mice (Langford et al., 2010), rats (Sotocinal et al., 2011) and rabbits (Keating et al., 2012) but we did not find any consistent pattern in seal pups.  Our results also showed that looking around and struggling decreased consistently after tagging and chipping, whereas blinking decreased in Experiments 1 and 2 and trembling decreased only in Experiment 1. The behaviour we term ‘looking around’ refers to animals’ visible tracking of objects, activities, and noises in the surrounding environment by focusing the direction of their gaze or orienting their heads towards such stimuli. Pups performed this behaviour more often before being tagged and chipped. We interpreted looking around as inquisitive or exploratory behaviour indicating attention to environmental stimuli. Pain can disrupt the allocation of attention (Eccleston and Crombez, 1999). The pain associated with tagging and chipping may   87 have temporarily captured the seals’ attention and distracted them from their surroundings.  After tagging and chipping, there was also a decrease in both struggling (Experiments 1, 2, 3) and trembling (Experiment 1). Struggling is an escape behaviour that has been observed in response to pain (e.g., tattooing the ears of rabbits: Keating et al., 2012) and restraint (e.g., rats: Grissom et al., 2008). Although pups struggled and vocalized at the moment of tagging and chipping, the amount of struggling was higher before the procedure than after. Trembling was brief and intermittent and was also seen more often before the procedure. Trembling can occur in response to pain (e.g., castration and tail docking of lambs: Molony and Kent, 1997) but is more often associated with excitement, fear and anxiety (Jankovic and Fahn, 1980).  For example, dogs have been shown to tremble in response to fear (e.g., sudden loud noise) and to both actual and anticipated restraint (Beerda et al., 1998). Thus, both struggling and trembling recorded in the seals may have been a response to capture and restraint rather than pain.  In general, the seals appeared alert and active before tagging and chipping but more subdued afterward, with reduced overall body movement (less looking around, struggling and trembling).  These differences did not exist in the sham procedure, suggesting that these changes are not a result of exhaustion or restraint.  This change in demeanour is consistent with changes found in cats that were reported to be less reactive to environmental change when in pain, and in dogs that were reported to appear less alert and more quiet when in pain (National Research Council, 1992).   88 In humans, blink rate is affected by multiple factors and has been shown to increase with anxiety (Harrigan and O’Connell, 1996) and in situations perceived to be stressful (Haak et al., 2009). In horses, however, a preliminary study showed that blink rate declined after exposure to several known stressors (e.g., separation from herd, exposure to a novel object) (Garnett and Merkies, 2017). In our study, the decrease in blinking is difficult to interpret but may be an artefact of seals having an increase in orbital tightening and a decrease in the amount of looking around at their surroundings after the procedures.  Breathing rate decreased after tagging and chipping, but overall was unusually high in these experiments.  Normal breathing rate for harbour seals is between 13 and 16 breaths min–1 for pups, and between 3 and 4 breaths min–1 for adults (Gulland et al., 2001) whereas the rates seen in these experiments were much higher (e.g., 29 ± 1 breaths min–1 before and 24 ± 1 breaths min–1 after the procedures in Experiment 1). Breathing may become more rapid and shallow in response to the stress of handling (Gulland et al., 2001) as evidenced in one study in which pre-weaned pups had rates of 36 ± 6 breaths min–1 after handling (Lapierre et al., 2007). The increased breathing rate in pups suggests either a stress response to handling and restraint or an effect of the restraint technique, in which the restrainer placed weight on the thoracic region of the animal. Because in Experiment 2 the seals that received the sham procedures showed the same decline as the tagged and chipped animals, the decline does not seem to be caused by pain. Perhaps the very high rate when first restrained was a hyperventilation response that waned gradually, possibly in response to a resulting alkalosis.  While the reason for lower breathing rate post-procedure is not clear, a similar effect was found in   89 harbour porpoises (Phocoena phocoena) over the period of being tagged (Eskesen et al., 2009).  Our results suggest that observers can accurately judge the presence or absence of pain based on the animals’ facial expressions and behaviours alone.  The observers’ subjective judgement of whether pain was present or absent corresponded closely to whether the seal had been tagged and chipped, regardless of the observers’ experience with the species. Assuming that tagging/chipping was painful, the accuracy of 95% (for the experienced observer) and 89% (for the inexperienced observer) is high compared to rates reported in other grimace work. These include 67% accuracy in sheep (McLennan et al., 2016), 73% in horses (Dalla Costa et al., 2014), 82% in rats (Sotocinal et al., 2011), 84% in rabbits (Keating et al., 2012), 89% in ferrets (Reijgwart et al., 2017) and 97% in mice (Langford et al., 2010). This result suggests that simple subjective use of facial expressions may be an effective observational tool for distinguishing the presence of pain in seals, but this would need to be confirmed with a larger number of observers.   The use of analgesics is a standard method of testing whether behavioural responses are due to pain (Weary et al., 2006) but this method requires an analgesic of known effectiveness. There is little information on the use of analgesics in pinnipeds. Existing research is limited to a few preliminary pharmacokinetic studies, for example with tramadol in California sea lions (Zalophus californianus) (Boonstra et al., 2013) and butorphanol and buprenorphine in elephant seals (Mirounga angustirostris) (Molter et al., 2015; Nutter et al., 1998). Buprenorphine is known to provide relatively effective, long-lasting analgesia in domestic dogs and cats (Lamont and Mathews, 2007) and is   90 suggested to be a reasonable choice for use in seals (Molter et al., 2015). Since there is no drug or dosage validated for controlling acute pain in harbour seals, our small pilot experiment used buprenorphine, known to be empirically safe for seals, at the low end of the recommended dosage for analgesia in dogs in order to avoid a sedative effect and respiratory depression. In this study, the drug had no observable effect on orbital tightening or other changes, but this may simply be because the drug and/or the low dosage rate were inadequate for this type of acute pain in this species.  There are several limitations to this study. Due to the time-consuming nature of scoring each second of video, this study used only two observers for Experiment 1 to establish inter-observer agreement, and a single observer once the high level of agreement had been established. Future studies with more observers could test how reliably different people score the key measures, especially orbital tightening, in real time rather than from video recordings. Similarly, this study examined the combined effect of tagging and chipping, while other studies (e.g., mouse grimace scale: Langford et al., 2010) tested various durations and types of noxious stimuli. In order to validate the key FAUs and behaviours indicative of pain in seals, further work examining their responses to other types of routine, potentially painful procedures is required.  Lastly, only a single drug and dosage was tried; future work would benefit from including an analgesic of known efficacy, should one become available.   2.5. Conclusions  This study was a preliminary evaluation of the facial expressions and behaviours in harbour seals in response to the presumably painful, routine procedures of flipper   91 tagging and microchipping. Seals demonstrated an observable change in orbital tightening and other behaviours after being tagged and chipped.  Both observers showed considerable accuracy in distinguishing between animals that had undergone the procedures and those that had not, suggesting that the use of these measures has promise as a tool for acute pain assessment in seals.  Orbital tightening may be particularly useful for identifying potential pain in seals as it changed in all seals after the procedure in all three experiments, was easily recognized by observers and has been identified as a key FAU in all animal grimace scales developed thus far.  92 3. Vocal changes as indicators of pain in harbour seal pups (Phoca vitulina)  3.1 Introduction Vocalizations are a critical aspect of the behaviour of most bird and mammalian species, conveying important information among conspecifics (e.g., attraction or warnings) (Manteuffel et al., 2004) and about an individual’s biological condition (e.g., fitness, dominance, reproductive state) (Watts and Stookey, 2000). Additionally, vocalizations may reflect an individual’s internal, affective state (Manteuffel et al., 2004). Given that affective states are recognized as integral to an individual’s welfare (Fraser et al., 1997), the ability to relate vocalizations to particular affective states could be advantageous for evaluating many aspects of animal welfare, including pain. Because pain is a subjective state, its accurate assessment in both human and non-human animals is challenging (Sneddon et al., 2014). However, because many species vocalize in response to presumably painful events (Anil et al., 2002) analysis of vocalizations has been used to study pain-inducing procedures in a variety of species. Such procedures include circumcision (Porter et al., 1988, 1986) and heel lancing and injections in human infants (Craig et al., 1993; Grunau et al., 1990); tail-docking in puppies (Noonan et al., 1996); tail docking, teeth clipping and ear notching in piglets (Noonan et al., 1994); hot-iron branding in calves (Watts and Stookey, 1999); and castration in lambs, goat kids, calves and piglets (Mellor et al., 1991; Weary et al., 1998).    93 Vocal measures used for pain assessment include the mere presence or absence of vocalizations (Mellor et al., 1991), the total rate of calls (Noonan et al., 1994) and the rate of different call types (Noonan et al., 1996). In addition, detailed acoustic parameters (e.g., duration, amplitude and frequency (Hz) of calls) have helped identify some distinctive characteristics associated with pain responses in the species examined. For example, shortened latency to cry, longer cry duration, and higher frequency and intensity typify human infants’ responses to invasive procedures such as injections (Grunau et al., 1990). Similarly, during castration, piglets emit more high-frequency calls (>1000hz) than those sham treated (Taylor and Weary, 2000).  However, because vocal responses to pain are often species typical (Morton and Griffiths, 1985), the usefulness of different parameters will likely vary among species.  Harbour seals (Phoca vitulina) are a marine mammal species commonly admitted to rehabilitation facilities as orphaned or injured pups (MacRae et al., 2011). In British Columbia, Canada, nearly 200 harbour seal pups are admitted annually to the Vancouver Aquarium Marine Mammal Rescue and Rehabilitation Centre (MMR), where they are hand-raised and then released once deemed healthy and capable of independent survival. Many of the pups are admitted to the centre with potentially painful conditions or injuries and they are subjected there to many routine interventions that may cause pain (e.g., blood draws, injections, wound cleaning, marking for identification).   The Canadian Department of Fisheries and Oceans (DFO) requires pre-release identification of each rehabilitated pup by affixing a tag in a hind flipper (‘tagging’) and implanting a microchip at the base of the tail (‘chipping’). In other species, similar   94 identification procedures have been associated with pain; these include ear tagging and microchipping in rodents (Dahlborn et al., 2013), microchipping in horses (Gerber et al., 2012) and ear tagging (Leslie et al., 2010) and ear notching (Torrey et al., 2009) in piglets. Seals demonstrate increases in orbital tightening after being tagged and chipped (a facial change indicative of pain in many species) (MacRae et al., 2018), an increase in eye temperature after being tagged (a physiological change associated with pain) (Chapter 4) and have been observed to exhibit behaviours that are consistent with pain responses in multiple other species (e.g., vocalizing, flinching, attempting escape, limb guarding) (National Research Council, 2009; Rutherford, 2002).  Notably, it is not standard practice to provide analgesic for these procedures.  Harbour seal pups vocalize frequently, beginning within hours of birth (Lawson and Renouf, 1985). These vocalizations, termed ‘mother-attraction calls’, are mother directed (Khan et al., 2006) and cease at weaning (Khan et al., 2006; Perry and Renouf, 1988), which usually occurs 26 - 40 days after birth (Cottrell et al., 2002). These calls likely play a role in maintaining mother–pup contact during lactation (Renouf, 1984). Unlike most phocids, lactating harbour seals leave their pups for brief periods to forage, and this necessitates effective mother-pup recognition for reunification (Boness et al., 1994). In fact, harbour seal pup vocalizations are individually distinct (Khan et al., 2006), and mothers appear to recognize their pups' calls within hours after parturition (Sauvé et al., 2015a). Early studies suggest that fundamental frequency may transmit individual pup identity, but increased calling bouts, call rate and harmonics may signal states of distress (Perry and Renouf, 1988; Renouf, 1984).   95 Very little research has been done on pain-related behaviour in any species of pinniped, and to our knowledge, no published work exists on the vocal responses of harbour seals to painful stimuli. Without valid indicators of pain for this species, it is difficult to design appropriate pain therapies (Flecknell, 2000). In this study, we looked for vocal changes in response to the potentially painful routine procedures of tagging and chipping. The aims of the study were to determine: 1) whether the parameters of seal pup vocalizations changed after tagging and chipping; and 2) whether the observed vocal responses were the result of the invasive procedures themselves, as opposed to possible effects of the restraint that they involve.  3.2 Materials and methods The research was approved by the University of British Columbia Animal Care Committee (Protocol A12-0095) and by the Vancouver Aquarium Animal Care Committee. Some of the seals included in this protocol were simultaneously used in a study on facial expression (Chapter 2: MacRae et al., 2018).   3.2.1 Animals and housing Thirty-one orphaned harbour seal pups (15 males, 16 females) were recovered along the coastline of British Columbia, Canada, by MMR staff or were brought to the facility by members of the public between June and September of 2012. Data were collected between August and October 2012, once pups had been weaned, estimated to be more than 60 days old, and considered to be free of disease or injury. Mean (± SEM) body weight at the time of testing was 14.7 ± 0.6 kg (range 11.4 kg–24.6 kg).    96  Animals were kept at MMR following the standard operating procedures of the facility. Animals were housed in groups of up to eight in fiberglass pre-release pools of approximately 23,000 L (4.87 m diameter X 1.21 m depth, with a 2.34 m X 1.21 m haul-out). Pups were scatter-fed whole herring three times per day at approximately 12% of body weight.  3.2.2 Tagging and chipping Between 07:00 and 09:00, before their first daily feed, individual pups were taken from their pools to a quiet outdoor area that was out of visual and auditory range of other staff and animals. Individuals were placed on a foam mat and restrained by a staff member holding the seal in ventral recumbency with the animal’s body between the restrainer’s knees, its head between the restrainer’s feet and its hind flippers secured in the restrainer’s hands.  The animals were then tagged and chipped, in that order, by another trained individual with many years of experience performing these procedures.  Tagging consisted of piercing the webbing of the 2nd and 3rd digits of the hind flipper and attaching a 5-cm plastic tag bearing an animal identification number and DFO contact information. Males were tagged on the right hind flipper and females on the left. Chipping consisted of implanting a 2.1 x 11.5 mm microchip preloaded in a 2.6 x 40 mm sterile single-use cannula (TrovanÒ ID-162B, Germany) subcutaneously on the dorsal aspect of the base of the tail. It is not standard practice to provide analgesic for either procedure.   The entire procedure together lasted approximately six minutes: two minutes of baseline restraint before tagging and chipping, about two minutes for tagging and   97 chipping, and two minutes of restraint afterward. Throughout the 6 min, the vocalizations of each animal were recorded using a Zoom 4HNPro digital audio recorder (Zoom Corporation, Japan) and a Beyer Dynamic MCE86 N(C) microphone (Beyerdynamic, Germany) suspended approximately 0.9 m over the seals’ heads. Audio was recorded as .wav files at a sampling rate of 44.1 kHz, 24 bit depth. Also, for the purposes of another study, the faces of the pups were filmed using a Canon HG10 HDD camcorder (Canon USA Inc., USA), from a distance of approximately 1.2 m (MacRae et al., 2018). The same three staff members restrained the seals, performed the tagging and chipping, and filmed the procedures for all animals.   3.2.3 Experiment 1 – Vocalizations before and after tagging and chipping The aim of Experiment 1 was to determine if pups show a change in vocalizations from before to after the procedures. To do this, the vocalizations of 21 seal pups undergoing tagging and chipping were recorded as outlined above. There were three test days: eight pups were tested on the first day, eight on the second and five on the third. Vocalizations before and after the procedures were then compared.   3.2.4. Experiment 2 – Cross-over experiment The aim of Experiment 2 was to determine whether the changes in vocalizations observed in Experiment 1 were the result of tagging and chipping or simply effects of the associated restraint.  A cross-over experiment with 10 pups was conducted over two test days. On Day 1, all animals were restrained as outlined above, but every second pup was actually tagged and chipped, while the others were ‘sham’ tagged and chipped.   98 On Day 2, which occurred five days later, the five animals that had been sham tagged and chipped on Day 1 were tagged and chipped, and the five pups that had been tagged and chipped on Day 1 underwent the sham procedure.  The sham procedures were nearly identical to the actual ones: seals were restrained and their tagging and chipping sites were cleaned, but instead of tagging, the tagging unit was touched to the flipper without piercing it (the sound from depressing and releasing the device was still audible), and instead of chipping, a syringe without a needle was briefly placed in contact with the chipping site.  3.2.5. Vocal analysis Sound files were analyzed spectrographically using Raven Pro Interactive Sound Analysis Software, Version 1.5 (The Cornell Lab of Ornithology, 2014). Spectrographs were made with a 512-point Hann window (3 dB bandwidth 124 Hz), 50% overlap and colour palette ‘jet’ (Figure 3.1). Measurements included for analysis were number of individual calls within a specified time block (30s or 10s), the duration of each call calculated from the waveform (osillogram) to a specificity of 1 ms, and the peak frequency defined as the frequency (Hz) at which peak power (dB) occurs within the selected vocalization.  Preliminary analysis of each seal’s 6-min clip showed that the majority of vocal activity occurred in the brief period immediately after the procedures.  We therefore chose the 30 s immediately before and the 30 s immediately after tagging to capture the peak period of vocalizations together with a comparison period before the procedure.  Because chipping always followed tagging and there was slight variation in the time   99 between the procedures, some seals did not have a full 60 s between tagging and chipping. Therefore, we used a shorter interval (10 s) before and after the chipping procedure.    Figure 3.1. Example of spectrographic representation of one harbour seal pup’s vocalizations for a 5-s period created with Raven Pro Interactive Sound Analysis Software (Version 1.5)  																									1																																				2																																			3																																			4																																				5	12	3	4	5	6	7	8	9	10	11	12	13	14	15	KHz	Seconds	  100  3.2.6. Statistical analysis Data analysis was done using R version 3.2.2 (R Core Team, 2015). Linear mixed-effects models (LMM) were calculated with the R package lme4 (Bates et al., 2015) and lmerTest (Kuznetsova et al., 2017). For each seal, the total number of calls, duration of calls and peak frequency (Hz) of each call were tabulated for 30s before and after tagging and for 10s before and after chipping. Adjusted mean ± standard error of the mean (SEM) was used for all results. For both experiments the response variables were the total number of calls, average duration of calls, average peak frequency (Hz) of calls, and maximum peak frequency (the peak frequency of the single call with the highest peak frequency) in the 30s before and after tagging and in the 10s before and after chipping. The analysis tested the effects of sex, body weight and treatment as fixed effect variables. In Experiment 2 the same models were used, except ‘treatment order’ (whether pups were tagged or sham tagged first) was included as a fixed effect variable, as well as interactions between treatment and treatment order. All interactions were tested but were kept in the models and reported only if significant. Seal was included as a random effect. Seals that did not vocalize at all were removed from the models but are included where appropriate in descriptive statistics. Changes were considered significant at p ≤ 0.05.      101 3.3 Results   3.3.1 Experiment 1- Vocalizations before and after tagging and chipping The mean peak frequency of calls was higher in the 30s after tagging (1041 ± 75 Hz) compared to before (837.1 ± 75 Hz; p < 0.01; n = 14). All 21 seals vocalized immediately after tagging, and all but one of the animals had a higher maximum peak frequency in the 30s after tagging (1476 ± 93 Hz; range 1119.7 - 2153.3 Hz), compared to before (996.4 ± 117 Hz; range 430.7- 2067 Hz; p < 0.001).  Similarly for chipping, seals emitted calls of higher peak frequency in the 10s after the procedure (1111.2 ± 79.2 Hz), compared to before (848.8 ± 79.2 Hz; p < 0.05; n = 16). Twenty of the seals vocalized in the 10s immediately after chipping, and all but one’s calls had a higher maximum peak frequency after chipping (1434.1 ± 74.8 Hz; range 1033.6 - 2584 Hz) than before (942.1 ± 101 Hz; range 344.5 – 1550.4 Hz) (p < 0.01). The duration of individual calls was not affected by tagging or chipping; seals called for 0.7 ± 0.05s before and 0.7 ± 0.04s after tagging and 0.8 ± 0.1s before and 0.8 ± 0.06 after chipping. The seven seals that did not vocalize in the 30s before tagging and the five seals that did not vocalize in the 10s before chipping were removed from this portion of the analysis.  The number of individual calls was greater in the 30s after tagging (9 ± 1) compared to the 30s before (5 ± 1; p < 0.001). Similarly, there were twice as many calls in the 10s after chipping (4.3 ± 0.4) than before (2 ± 0.4; p < 0.001; n = 21). Of the 21 pups, seven did not vocalize before tagging, and five did not vocalize before chipping;   102 however, all pups vocalized in the 30s after tagging, and all but one vocalized in the 10s after chipping.   There was no effect of sex or body weight on the number of calls or call duration immediately before and after tagging and chipping. However, there was a tendency for the calls of males to have overall lower average peak frequency than those of females before and after tagging (813 ± 111.2 Hz for males compared to 1065.9 ± 82.8 for females; p = 0.1; n = 14) and before and after chipping (847 ± 98.6 for males compared to 1113  ± 86.7; p < 0.07; n = 16). Means of peak frequency before and after tagging and before and after chipping were pooled for this comparison.  3.3.2 Experiment 2 – Crossover experiment In Experiment 2, all 10 pups vocalized after tagging and after chipping, but several did not vocalize before the procedures or after the sham procedures. Because of the small number of animals, comparisons of call duration and peak frequency were done by t tests. As in Experiment 1, pups had higher average peak frequency in the 30s after tagging (1062.3 ± 155 Hz) than before tagging (872.8 ± 155 Hz) (p < 0.001).  There was no difference in average peak frequency before and after sham tagging (p = 0.4). There was also trend for pups to have higher average peak frequency in the 10 s after chipping (1236  ± 147 Hz) compared to before (924.8  ±  89 Hz; p = 0.15). Because only three animals vocalized both before and after sham chipping the difference in peak frequency could not be analyzed.    103 There was no difference in call duration in the 30s before versus after tagging (p = 0.5) or sham tagging (p = 0.8). However, call duration in the 10 s before chipping was somewhat shorter (0.7 ± 0.09 s) than after (0.9 ± 0.1 s) (p < 0.05). Call duration could not be analyzed for sham chipping as only three animals vocalized both before and after.  We also compared call duration after tagging and chipping versus after the sham procedures. In the 30s after tagging pups tended to have longer call duration than after sham tagging (p = 0.1); however, this is based on only five of 10 pups because five did not vocalize in the 30s after the sham procedure.   All 10 pups vocalized after chipping, whereas only five vocalized after sham chipping. For these five pups, the duration of vocalizations was on average 0.3s longer after chipping compared to the sham procedure (p = 0.01).  In Experiment 2, seals also produced more calls in the 30s after tagging (8.0 ± 1.5) than before (3.6 ± 1.5) but there was no similar change after sham tagging (5.9 ± 1.5 before compared to 3.3 ± 1.5 after). This is reflected by the highly significant interaction between ‘treatment’ (tag versus sham) and ‘phase’ (before versus after the procedure) (p < 0.001; n = 10). There was also a less pronounced but still significant interaction between ‘treatment’ (tag versus sham) and ‘treatment order’ (p < 0.05; Figure 3.2), evidently because the five seals in the tag-first group tended to give more calls than the five in the sham-first group on the day of actual tagging. As in Experiment 1, pups also gave a greater number of calls after chipping (4 ± 0.3) compared to before   104 (2 ± 0.5; p < 0.001), but there was no difference in the number of calls before and after sham chipping (p = 0.4).      Figure 3.2. Number of calls of harbour seal pups (n = 10) in the 30 s before and after tagging or sham-tagging. A: Seals that received sham-tagging on Day 1 and tagging on Day 2; B: seals that received the reverse order of treatments.   															sham-tag																	tag																			sham-tag																	tag																																A																																																				B	30	s	before	30	s	a2er	Number	of	vocaliza:ons	10	86420  105 3.4 Discussion Our results show that seals had a strong vocal response to both tagging and chipping. Specifically, the number and peak frequency of calls consistently increased after both procedures. Similar changes were not observed with sham-tagging and sham-chipping. An increased number and frequency of calls has been associated with pain in many species (Manteuffel et al., 2004). Previous studies demonstrate that seals show other potential signs of pain after both tagging and chipping.  After the procedures, for example, seals had increased orbital tightening (Chapter 2) which is recognized as a common response to pain in many species (Chambers and Mogil, 2015). Seals also had increased eye temperature after tagging, a physiological change potentially associated with pain (Chapter 4). Therefore, the findings from this study – together with previous findings of increased orbital tightening and increased eye temperature – suggest that these vocal changes are due to the immediate pain caused by tagging and chipping. The increased number of vocalizations after both tagging and chipping is consistent with the greater number of calls per calling bout and faster rate of call emission reported in seal pups when distressed (Perry and Renouf, 1988; Renouf, 1984). An increased number  and rate of vocalizations has been reported also in human infants when in pain (Porter et al., 1986), and higher-frequency calls produced at a faster rate have been reported in piglets in response to castration (Weary et al., 1998).  Typical harbour seal mother-attraction calls are described as tonal, having an inverted ‘u’ or ‘v’ shaped spectrograph with harmonics, with fundamental frequencies averaging 200-600 Hz and peak frequencies of ~800 Hz, and with durations of ~0.6s   106 (Khan et al., 2006; Sauvé et al., 2015b). The vocalizations of the seals in our study were consistent with this description (Figure 3.1), with peak frequencies averaging ~870 Hz and durations ~0.7s in the periods before tagging and chipping. In both Experiments 1 and 2, the average peak frequency increased after both tagging and chipping to surpass peak frequencies of typical mother-attraction calls.  Additionally, in Experiment 1, many pups did not vocalize at all before tagging and chipping but all 21 pups vocalized in the 30s after tagging, and all but one vocalized in the 10s after chipping. In contrast, when pups received the sham treatments, there were no differences in either the number or peak frequency of vocalizations from before to after the procedures, suggesting that the changes in vocal response are a reaction to the procedure itself rather than to restraint. Change in call duration is often linked to affective state (Briefer, 2012) and has been used to assess pain in piglets (e.g., Taylor and Weary, 2000). Within the confines of our study (e.g., short time intervals and brief punctate procedures), call duration did not change from before to after tagging or chipping. However, in Experiment 2, calls tended to have a longer duration in the 30-s period after tagging than in the same period after sham-tagging. Similarly, in the 10-s period after chipping, the duration of vocalizations was on average 0.3s longer than after sham-chipping. This suggests that call duration may sometimes be affected by painful procedures in seals, but more animals are needed to determine how consistently this occurs. There was no effect of sex or body weight on any of the vocal response variables, except a tendency for males to have lower average peak frequencies.  In general, vocalizations of both male and female harbour seal pups have similar peak frequencies, but as animals mature, the peak frequencies of the females rise while   107 those of the males fall (Khan et al., 2006). The trend toward lower peak frequency for males found in this study may reflect early sex differences in vocalizations as seals develop. As well as indicating the presence of pain, vocalizations may also reflect the nature or intensity of pain and may thus be useful for identifying particularly aversive aspects of routine procedures. For example, Taylor and Weary (2000) used detailed vocal analysis to assess piglet responses to different components of the castration procedure. Using rate and frequency of calls, they concluded that the pulling and severing of the spermatic cord is likely the most painful part of the castration procedure (Taylor and Weary, 2000). In our study, the increase in the number and peak frequency of calls was similar for both tagging and chipping, suggesting that the degree of pain evoked by each procedure was equivalent. However, this would need to be verified as tagging was always done before chipping in our study.      Our results suggest that harbour seal vocalizations—in particular, changes in the number of calls and their peak frequency—can potentially indicate pain in harbour seal pups. It is unclear, however, if these measures could be extrapolated to older animals. Generally, adult harbour seals are considered vocally reticent compared to other pinnipeds (Schusterman 2008). Adults of this species associate with each other in using shared haul-out sites, but do not appear to have strong affiliative relationships (Godsell 1988). Hence there is unlikely to be any advantage for mature animals to vocally signal pain to con-specifics. Since mother-attraction calls have the primary purpose of soliciting maternal attention, it is logical that the behaviour would cease once the maternal bond is severed by weaning, and in fact this type of call is not produced past   108 weaning in mother-raised pups (Renouf, 1984). Pups in rehabilitation become noticeably less vocal once weaned, but many continue to emit mother-attraction type calls after weaning (pers. obs.), possibly because some hand-raised pups may transfer mother-attraction calls to their caregivers while they remain dependent on humans for food. Pups are still physically able to emit calls with the acoustic properties of mother-attraction calls past the natural age of weaning (26-40 days; Cottrell et al., 2002) as evidenced by the calls of the seals in this trial, all of which were > 6 weeks of age. However, as pups develop, their vocal parameters change (Khan et al., 2006; Sauvé et al., 2015b), and at a certain point, their ability to produce these calls is likely diminished. Therefore, the use of vocalizations to indicate pain in older seals needs further investigation.  3.5 Conclusions Acoustic analysis of the vocalizations of harbour seal pups has promise as an objective, non-invasive measure of the animals’ response to painful stimuli.  The vocal indicators recorded in this study, combined with other indicators of pain previously found in response to tagging and chipping (Chapters 2 and 4), give strong support to the view that these routine procedures are painful. Future studies are needed to identify pain management therapies for this species.   109 4. Can surface eye temperature be used to indicate a stress response in seals (Phoca vitulina)?   4.1 Introduction  The assessment of potentially painful or aversive husbandry procedures typically relies on a combination of physiological and behavioural measures (Rutherford, 2002).  Unfortunately, many of these measures are time-consuming and may require handling or sampling which can themselves result in a stress response (Stewart et al., 2005). As an alternative, infrared thermography (IRT) is increasingly recognized as a reliable, non-invasive method to detect changes in heat emission, especially from around the eye, as an indication of physiological stress responses in animals (Stewart et al., 2010).  For many years, IRT has been used successfully to identify injury and disease in veterinary medicine (McCafferty, 2007) and, increasingly, to measure stress responses in animals via changes in local blood perfusion. There is evidence that both physiological and psychological stress are indicated by changes in eye temperature (Cook et al., 2001; Pavlidis and Levine, 2002) and that eye temperature may reflect the activity of both the autonomic nervous system (ANS) and the hypothalamic-pituitary-adrenal axis (HPA) (Cook et al., 2001; Stewart et al., 2008). Short-term temperature responses due to changes in peripheral blood flow are associated with the ANS and may be a physiological correlate of an animal’s affective state, for example in response to painful, stressful or arousing stimuli (Stewart et al., 2008a). In such cases a change in the amount of blood-flow through peripheral vessels influences the temperature of the   110 skin and extremities leading to a change in radiated heat which can be measured with IRT.  The eye, and in particular the areas of the lacrimal caruncle and palpebral border of the lower eyelid, has been recognized as being particularly sensitive to changes in blood flow from the stress response (Pavlidis and Levine, 2002; Stewart et al., 2008b). For example, changes in eye temperature have been recorded in horses (Dai et al., 2015) and cattle (Stewart et al., 2008a) in response to fear-producing stimuli, in horses after aversive procedures such as coat-clipping (Yarnell et al., 2013) and the application of restrictive nosebands (Fenner et al., 2016; McGreevy et al., 2012), in elk (Cervus elaphus canadensis) and reindeer (Rangifer tarandus) from velvet-antler removal (Cook et al., 2005), in dogs from veterinary examination (Travain et al., 2015), and in cattle after castration and disbudding (Stewart et al., 2008b; Stewart et al., 2009, 2010). Hence, variations in eye temperature with IRT can provide a dynamic, real-time tool that can potentially be used for the evaluation of an animal’s stress response to a range of handling and husbandry interventions.  Harbour seal pups (Phoca vitulina) are commonly admitted to rehabilitation facilities (MacRae et al., 2011) where they are hand-raised and, once deemed healthy and capable of independent survival, released. These animals are frequently admitted into rehabilitation with conditions or injuries that are potentially painful, and while in care they are subjected to many potential stressors (e.g., handling, restraint, veterinary interventions) some of which may cause pain. However, little research has assessed the stress responses of pinnipeds to such procedures. Moreover, there is limited information on identifying pain in any species of pinniped. Poor identification of pain in   111 these species may thus result in inadequate or inconsistent pain therapies (Flecknell, 2000). Additionally, exposure to prolonged or severe stress can have a range of negative health consequences, including impaired growth and immune competence (Moberg, 2000). Because these pups are already vulnerable, it is important to identify and possibly moderate husbandry routines and procedures that may contribute to animals’ stress responses.  In Canada, the Canadian Department of Fisheries and Oceans (DFO) requires each rehabilitated pup to be marked before release with a tag placed in a hind flipper and a microchip at the base of the tail. Similar identification procedures in other species have been associated with pain; examples include ear tagging and microchipping in rodents (Dahlborn et al., 2013) and ear tagging (Leslie et al., 2010) and ear notching (Torrey et al., 2009) in piglets. While being tagged and microchipped, seals have been observed to flinch, vocalize, exhibit escape behaviours, and curl their flippers after tagging, behaviours consistent with responses to pain in multiple other species (National Research Council, 2009; Rutherford, 2002). Seals also demonstrate significant increases in orbital tightening after being tagged and chipped (a facial change indicative of pain in many species), as well as higher than normal breathing rates during these procedures, thought to be in response to their concomitant handling and restraint (MacRae et al., 2018). Notably, it is not standard practice to provide analgesic for these procedures. Measuring changes in eye temperature may be particularly applicable in contexts that necessitate the temporary captivity of wild species. For example, in wildlife rehabilitation, handling and human proximity may be especially stressful to the animals,   112 potentially dangerous to the handler, and may increase the risk of habituation to humans and thus potentially jeopardize the animals’ success once released. IRT may therefore provide a practical method for identifying stress responses caused by fear and pain from routine procedures and husbandry practices in pinnipeds undergoing rehabilitation. The aims of the study were to determine: 1) whether harbour seal pups showed a change in eye temperature after the potentially painful procedure of flipper tagging; 2) whether the observed responses were the result of tagging, as opposed to possible effects of handling; and 3) whether the observed responses would be mitigated by the administration of a local anaesthetic (lidocaine).  4.2 Materials and methods The research was approved by the University of British Columbia Animal Care Committee (Protocol A16-0175) and by the Vancouver Aquarium Animal Care Committee.   4.2.1 Animals and facilities  The 52 harbour seal pups (30 males, 22 females) had been recovered along the coastline of British Columbia, Canada, by Vancouver Aquarium Marine Mammal Rescue (MMR) staff, or had been brought to the facility by members of the public, between June and August of 2016. Data were collected between August and December of 2016, once pups were deemed ready for release (weaned, estimated to be more than 90 days old, and free of disease or injury). Mean (± SEM) body weight at the time of testing was 23.6 ± 0.3 kg (range 18 kg – 30 kg).    113  Animals were kept at MMR following the standard operating procedures of the facility. Animals were housed in groups of up to 8 in fiberglass pre-release pools of approximately 4,500 L (2.4 m diameter x 0.8 m depth) with a 3.7 m x 4.0 m haul-out area.  Testing was done in two locations: the home pen and the tagging area. Pups remained in their home pen (described above) during periods that required no restraint. The roof and sides of the home pen were covered with canvas during the experiment to eliminate sunlight and minimize cross drafts. The tagging area was a 3 m x 3 m pavilion tent canopy with four sides where all testing that required handling was completed.  The two locations were approximately 9 m apart. Pups were captured by hand and carried between locations in plastic totes (81 x 51.4 x 44.5 cm, RubbermaidÒ).  The ambient temperature and humidity of the relevant testing location were recorded at the start of each sampling Period (Taylor Wireless Indoor/Outdoor ThermometerÒ, Oak Brook, IL, USA 60523).  4.2.2 Injection and tagging protocol After being carried to the tagging area, individuals were placed on a foam mat where they were restrained in ventral recumbency, with the animal’s body between the restrainer’s knees, its head between the restrainer’s feet and hind flippers secured in the restrainer’s hands. For animals that received injections, 1 ml of lidocaine hydrochloride 2% (20mg/ml, Zoetis Canada Inc., Kirkland, QC, H9H 4M7) or sterile saline solution was injected into the planned area of insertion of the identification tag (webbing of the 2nd and 3rd digits of the hind flipper) with a 22-G x ¾ "(0.711 mm x 1.9   114 cm) needle (Kendall, Mansfield, Massachusetts 02048, USA) on a 3-ml syringe (Terumo Medical Corporation, Somerset, New Jersey 08870, USA). The site had been cleaned with chlorhexidine and alcohol.  All injections were performed by the same person. The 1 ml dose of lidocaine was established by manually pinching the skin and flipper webbing in the area of the lidocaine injection and observing the animals’ reactions. The pups that were pinch tested were not included in this study. For tagging, animals were restrained and tag sites cleaned as described. A 5-cm plastic tag bearing an animal identification number and DFO contact information was attached by piercing the webbing of the 2nd and 3rd digits of the hind flipper. Males were tagged on the right hind flipper and females on the left. The same two people alternated between restraining and tagging the seals.   4.2.3 Eye temperature measurements Infrared images of each pup’s eyes were collected with an IRT camera (FLIR T300, FLIR Systems AB, Danderyd, Sweden). The camera has a thermal sensitivity of < 0.05°C and can detect temperatures between -20°C and 650°C. Consecutive thermograms (1 approximately every 10 s) were taken from approximately 1 m from seals’ eyes for the duration of each sampling period (Baseline, Periods A, B, C, D, Post-30 and Post-60). All thermal images were taken by the same person. Once thermal images had been collected, they were uploaded to FLIR Tools+analysis software. Thermal images were analyzed to determine maximum eye temperature (°C) within the area of the medial posterior palpebral border of the lower eyelid and the lacrimal caruncle (Figure 4.1). No temperature difference was found between the left and right   115 eyes so these images were pooled. Animals were dry at the time of recording.   116      Figure 4.1. Example of a thermal image of a seal’s eye. Maximum eye temperature is indicated by the dark red triangle.  4.2.4 Treatments Twelve test days, approximately one per week for twelve weeks, were coordinated with the times when seals were ready for release. The seals tested each day were first selected based on their readiness for release and were then assigned to one of four treatments using a random number generator. Four pups were tested per day (1 pup   117 per treatment) except on the first day, when eight animals were tested (2 per treatment). Testing was done between 08:00 and 14:00; daily feeding was delayed until testing was completed. The Baseline eye temperatures of all pups were first recorded for 3 min in their home pen before any handling began. The four treatments were (Figure 4.2):  Lidocaine (n = 13, restrained twice): after the 3-min Baseline recording in the home pen, pups were captured and moved to the tagging area where they were immediately restrained and eye-temperature recording was begun. After 3 min of restraint (Period A), they received 1 ml lidocaine (2%) injected subcutaneously into the tagging site. Restraint continued for 10 min after the injection (Period B). Pups were then placed in the plastic carrying tote and left undisturbed in the tagging area for a 10-min rest when eye temperature was not recorded. Pups were then recaptured and restrained a second time. After 3 min of restraint (Period C), pups were tagged and restraint continued for 10 min (Period D). Then pups were returned to the home pen.  Saline (n = 13, restrained twice): the procedure was identical to the Lidocaine treatment except that 1 ml of saline solution was injected instead of lidocaine. Tag Only (n = 13, restrained once): after the 3-min Baseline recording in the home pen, pups had eye temperatures recorded for a further 13 min while still unhandled in their home pen (Periods A and B). Ten min later, pups were captured and moved to the tagging area. After 3 min of restraint (Period C), pups were tagged and restraint continued for 10 min (Period D). Then pups were returned to their home pen.  Sham Tag (n = 13, restrained once): the procedure was identical to the Tag Only treatment but instead of the animals being tagged, they were sham tagged by touching   118 the tagging unit to the flipper without piercing it. The sound from depressing and releasing the tagging unit was the same for both actual tagging and sham tagging. Once returned to their home pen, eye temperatures were recorded for all pups for 3 min starting at each of 30 min and 60 min after treatment.     Figure 4.2. Pictorial representation of the four treatments (n= 13 seal pups per treatment) over the different periods from Baseline (before any handling began) to Post 60 (60 min after all handling had ended).  The dashed lines indicate periods when pups were not handled and the solid lines indicate periods when pups were restrained.   	1																				Baseline					A														B																																	C															D																										Post	30							Post	60																						2																																														Inj.																																																			Tag							3	Lidocaine						3	min								3	min								10	min																																3	min									10	min																															3	min															3	min	4																										5		6																																														Inj.																																															Tag	7	Saline											3	min								3	min								10	min																																3	min									10	min																															3	min															3	min	8		 	9																																																																																																		Tag	10		11	Tag	Only							3	min								3	min								10	min																																3	min									10	min																															3	min															3	min	12		 	13		14																																																																																											Sham	Tag	15	Sham	Tag					3	min								3	min								10	min																																3	min									10	min																															3	min															3	min	16		 	17		18		19	  119 4.3 Statistical analysis and results Thermograms were binned by minute to obtain a mean of the maximum eye temperature in each minute of each period for each seal. For Periods A through D each minute was treated as a separate data point. Three minutes of data were collected for each seal in the Baseline period and in the two periods after all handing was completed (Post-30 and Post-60 minutes); as there was no systematic difference in eye temperature between the respective minutes of these periods, they were pooled to make a single maximum eye temperature for each seal for each of these three periods. Subsets of the data were analyzed separately to determine the effects of treatment, tagging, handling and of receiving an injection. Mean ± standard error of the mean (SEM) was used for all results.  For each model, normality and homoscedascity of residuals were assessed graphically.  Data analysis was done using R version 3.2.1 (R Core Team, 2015). Linear mixed-effects models (LMM) were calculated with the R package lme4 (Bates et al., 2015) and lmerTest (Kuznetsova et al., 2017). The models included the seals’ maximum eye temperature (as the response variable), and tested the effects of sex, body weight, treatment, treatment period (Baseline, Periods A, B, C, D, Post-30 and Post-60) within the treatment, minute within treatment period, and handling (i.e., capture, restraint, and/or injection, tagging), as fixed effect variables, as well as interactions between treatment and treatment period and between treatment period and minute. Seal was included as a random effect. Results were considered significant at p ≤ 0.05.   120 The pups showed large and consistent individual differences in eye temperature throughout the experiment. For example, one of the Sham Tag pups (‘Bubbles’) had eye temperature values from 30 to 34.4 ° C throughout the study whereas another (‘Peter’) remained in the range of 27-29 ° C.  Eye temperature values were correlated weakly with ambient temperature in some periods and more strongly in others. For example, the correlation of eye temperature and ambient temperature was 0.38 in the Baseline period and 0.80 in the final readings taken 60 min after pups had been returned to their home pens. Because of the correlation with ambient temperature, all comparisons of eye temperature values were made after including ambient temperature in the model, except when comparing Periods A to B and C to D because the same ambient temperature applied during those comparisons. Handling (i.e., capturing, restraining, injecting) varied across treatments. This allowed us to test the effect of different handling events separately. Treatments were grouped when possible to increase power as described below. To assess the effect of handling events on eye temperature change, the interaction between period and treatments was included. When not significant, the interaction term was removed and only main effects are reported.  4.3.1 Eye temperature changes after handling  To examine the effect of handling (capture and restraint) versus no handling, we compared the change in eye temperature from Baseline to Period A for the 26 pups that were moved to the tagging area (Lidocaine and Saline treatments) versus the 26 pups   121 that remained in the home pen (Tag Only and Sham Tag treatments). To explore the effect on seals of being handled a second time, eye temperatures for the Lidocaine and Saline treatments were compared for Period A (first handling) and Period C (second handling).  Eye temperature increased after pups were captured and restrained. Mean eye temperature in Period A (when half the pups had been carried to the tagging area) was 0.6 ± 0.09 higher than Baseline levels. The increase was seen both in pups that were moved to the test pen (Lidocaine and Saline treatments) and those that remained in the home pen while their pen-mates were being removed (Sham Tag and Tag Only treatments), but the increase was greater in those that were moved. Regression analysis showed a higher eye temperature in Period A compared to Baseline (p < 0.05) and an interaction of period and treatment (p < 0.05) because those pups that were moved increased more than those that were left in the home pen.  An apparent effect of capture and restraint was also seen when pups in the Lidocaine and Saline treatments were restrained a second time in Periods C and D. These animals, which had already been restrained and injected in the tagging area during Period B, showed a further increase of 0.7 ± 0.08° C in eye temperature from Period B to Period C (p < 0.001) with no difference between the Lidocaine and Saline groups.  Eye temperature also increased after the pups were returned to the home pen. The first reading, taken 30 min after the pups were returned and reunited, was 1.1 ± 0.2 ° C higher than the mean for Baseline (P < 0.001) regardless of treatment. Eye temperature declined 0.5 ± 0.15 ° C (p < 0.001) at the final reading taken 60 min after   122 the pups were returned but still remained higher than Baseline (0.6 ± 0.26 ° C, p < 0.05).  Treatment did not affect the change in eye temperature from Baseline at either 30 min or 60 min after treatment.  Because of the apparent effect of capture and restraint, treatments were compared only for Lidocaine versus Saline and for Tag Only versus Sham Tag because pups in these pairs of treatments had been handled in the same way.   4.3.2 Eye temperature changes after injection  After lidocaine injection, some pups showed an upward trend in eye temperature over the 10 min of Period B (Figure 4.3). Of the 13 seals, 10 had a higher mean eye temperature after lidocaine injection (Period B) than before (Period A), the mean increase of all 13 seals being 0.3 ± 0.16 degrees (p < 0.07). No similar change was seen in the pups injected with saline.  123        Figure 4.3. Examples of the change in eye temperature ° C (mean ± SE) showing four seals before (Period A) and after (Period B) receiving an injection of lidocaine (Lidocaine treatment).2526272829303132331 2 3 inj. 1 2 3 4 5 6 7 8 9 10Period	A	(3	min) Period	B	(10	min)KingMaverickGuinevereRosie33	32	31	30	29	28	27	0	1  2							3										inj.					1							2							3							4							5								6							7							8							9						10	Period	A	(3	min.)	 Period	A	(3	min.)	King	Maverick	Guinevere	Rosie	Eye		temperature	°C	Period B (3 in.)   124  4.3.3 Eye temperature changes after tagging or sham-tagging To examine whether there was a change from before to after tagging or sham-tagging, the difference between the 3 min before tagging (Period C) and the 10 min after tagging (Period D) was compared for each of the Tag Only and Sham Tag treatments separately, using a paired t-test. This approach was used because the assumption of homogeneity of variance was not met and because it was not necessary to adjust for ambient temperature when comparing Periods C and D because the same ambient temperature applied in both Periods. After tagging, pups showed large individual differences in eye temperature response but no uniform pattern (Figure 4.4). For example, of the pups that were tagged with no previous intervention (Tag Only), two (‘Oscar’ and ‘Satchi’) showed a sharp 1-degree increase in eye temperature in minutes 2-4 after tagging, whereas most others did not. However, mean eye temperature increased gradually during the 10 min, with mean temperature being 0.3 ± 0.11 ° C (mean ± SE) degrees higher after tagging (Period D) compared to before (Period C) (t = 2.58, df = 12, p = 0.02), and increased 0.4 ± 0.17 ° C (mean ± SE) from minute 1 min to minute 10 after tagging (Period D) (t = 2.43, df = 12, p < 0.05). Pups in the Sham Tag treatment showed no similar increase (mean change of 0.0 ± 0.19).  125      Figure 4.4. Examples of change in eye temperature ° C showing five seals in the Tag Only treatment before (Period C) and after (Period D) the animals were received flipper tags.25.026.027.028.029.030.031.032.01 2 3 TAG 1 2 3 4 5 6 7 8 9 10FinnLouieSatchiOscarMaid	M.		1								2								3							TAG					1								2								3								4								5								6								7									8								9							10	Finn	Louie	Satchi	Oscar	Maid	M.	0	27	28	29	30	31	32	Period	C	(3	min.)	 Period	D	(10	min.)	Eye	temperature	°	C	  126  4.3.4 Eye temperature changes after attempted pain management As noted above, pups that had been restrained and injected in Period B (Lidocaine and Saline treatments) showed a further increase in eye temperature when restrained a second time for Period C. These animals then showed a decrease in eye temperature in the minutes immediately after tagging (decrease of 0.5  ±  0.11 from Period C to the first 3 min of Period D, p < 0.001), possibly as their initial response to handling diminished. Thereafter, mean eye temperature showed no clear changes or differences between treatments. Specifically, eye temperature tended to increase during the remainder of Period D for the Lidocaine treatment (increase of 0.3 ± 0.17 C from min 1-3 to min 7-10 of Period D, p = 0.1) but not for the Saline treatment (decline of 0.1 ± 0.2, p = 0.7).  4.4 Discussion These results suggest that changes in eye temperature may be useful to detect responses to potentially stressful and painful procedures such as handling and flipper tagging in harbour seals.  When captured, moved and restrained for the first time, pups had higher eye temperatures than non-handled pups, suggesting that these handling events caused a physiological stress response detectable via IRT. Restraint, in particular full body immobilization, is known to be aversive and result in a stress response in several species (Buynitsky and Mostofsky, 2009). Full body restraint is a common technique for seals. Increased breathing rate can be associated with a stress response (Gulland et   127 al., 2001) and has been recorded in grey seals (Halichoerus grypus) (Lapierre et al., 2007) and harbour seals (MacRae et al., 2018) in response to handling.  Interestingly, pups that were not handled also had an increase in eye temperature from Baseline to Period A, although the change was not as great as that seen in the handled group. Social transmission of fear has been demonstrated when animals observe pen-mates in distress (Kim et al., 2010) and can cause behavioural and physiological changes indicative of a stress response in the observing animal (Olsson and Phelps, 2007). Also, because the pups were group-housed and handlers were required to enter the enclosure both to capture pups and to record IRT images, it is possible that the proximity of the handlers and/or observing the capture and removal of pen-mates resulted in eye temperature increase in pups that were not themselves handled in this period.   The further increase in eye temperature when pups were handled a second time (Lidocaine and Saline treatments) suggests that the first handling, which was presumably aversive, resulted in an anticipatory stress response to subsequent handling. A similar response was seen in calves that underwent two separate jugular catheterizations, known to be aversive for cattle (Alam and Dobson, 1986). In that case, calves showed no change in eye temperature or cortisol concentration after initial catheterization, but both measures increased after the second catheterization one week later possibly because the calves were anticipating an event that they had learned would be aversive (Stewart et al., 2007b). Together, the increase in eye temperature after the first handling, and the further increase after the second, suggest the seals’ experience of the first handling was aversive. Also, seals that were handled twice   128 received an injection of lidocaine or saline during their first handling which may have further contributed to a negative association with handling.  Seals’ reaction to repeat handling suggest it would be advisable to combine necessary procedures and interventions into single handling sessions whenever possible.  We had anticipated that seal eye temperatures would quickly return to Baseline upon their return to their home pens for recovery. For example, the rise in eye temperature recorded in dogs in response to a 4-5 min veterinary examination decreased to pre-exam values within 5 min  post-examination (Travain et al., 2015) and decreased in horses within 10 min after wearing a tight noseband for 10 min (Fenner et al., 2016). However, regardless of treatment, the eye temperature of seals when returned to their home pens remained 1.1 ± 0.23 ° C higher than Baseline 30 min after handling, and although eye temperature had started to decrease at 60 min after handling, it still remained higher than Baseline. This change in eye temperature was also seen in the sham tagged animals, ruling out a prolonged response to tagging. Possibly, seals’ re-unification with pen-mates furthered the effect of their physiological disturbance; alternatively, the sampling times (30 min and 60 min after treatment) may have been insufficient for a complete recovery in this species. After surgical castration of calves, eye temperature had not started to decrease by the end of a 20-min observation period (Stewart et al., 2010). The slower recovery of the seals may reflect a species difference in response, a difference in the type of stressor, or the more prolonged exposure to stressful handling and proximity to humans. Eye temperature has been shown to increase in response to potentially painful procedures in multiple species, for example, in elk and reindeer after velvet-antler   129 removal (Cook et al., 2005) and in cattle after castration and disbudding (Stewart et al., 2008b; Stewart et al., 2009, 2010). In our study, eye temperature was higher in the 10 min after tagging but not after sham-tagging. Tagging for identification causes tissue damage and is associated with pain in other species. Harbour seal pups show several behaviours indicative of pain after being tagged, including orbital tightening (MacRae et al., 2018). The increase in eye temperature after tagging in this study is consistent with the view that flipper tagging is painful for seal pups. Cattle show a rapid, transient drop in eye temperature within the first few seconds of presentation of a stressor, likely because of an initial sympathetic response (peripheral vasoconstriction) (Stewart et al., 2008; Stewart et al., 2007a). It is possible seals had similar initial decreases in eye temperature; however, because sampling did not start until several minutes after seals were captured, any immediate thermal responses would have been missed. As noted above, pups that had been restrained and injected in Period B (Lidocaine and Saline treatments) showed an increase in eye temperature when restrained and recorded for Period C. These animals then showed a decrease in eye temperature in the minutes immediately after tagging. However, as the decrease in eye temperature was seen only in the pups that already had elevated eye temperatures in their second handling period, it is possible this decrease indicates their initial response to handling diminished as opposed to reflecting an acute response to the pain from tagging.  Lidocaine is known to provide effective pain control in many species for a range of procedures (Valverde and Gunkel, 2005), but in the present study its administration had little effect on eye temperature after tagging. While lidocaine is successfully used in   130 pinnipeds (Gutiérrez et al., 2016), an effective dosage and route of administration have not been established for flipper tagging. Moreover, lidocaine injection is known to produce pain due to its acidic pH (Cepeda et al., 2012). For example, although lidocaine reportedly provided analgesia for velvet antler removal in elk and reindeer, its application appeared to initiate an additional, attendant stress response in control animals that were not subjected to antler removal (Cook et al., 2005). Similarly in calves, lidocaine reduced changes in eye temperature after disbudding and surgical castration (Stewart et al., 2008b; Stewart et al., 2010), but its initial injection caused an increase in eye temperature (Stewart et al., 2008) and heart rate (Stewart et al., 2010), suggesting lidocaine administration itself is stressful. In this study, there was an upward trend in eye temperature in the 10 min after injection of lidocaine but not of saline (Period B) and there was no clear difference between lidocaine-treated and saline-treated pups after tagging (Period D). Our injection of lidocaine appears not to have effectively blocked the nerves at the tag site, may have caused pain, and required a second handling. For these reasons lidocaine cannot be recommended for the flipper tagging of seals. Further investigation to establish an effective pain therapy for this procedure is warranted.  4.5 Animal welfare implications and conclusions In summary, this study indicates that eye temperature measured by IRT can be used to detect responses to routine handling procedures in harbour seals. Handling had an important effect on eye temperature in the seals; hence, future use of eye temperature to monitor states like pain will need to control carefully for handling as the eye   131 temperature response is not pain-specific. The fact that eye temperature generally increased during the minutes after tagging but not sham-tagging, and after lidocaine injection but not after saline injection, suggests that such increases may reflect pain. If this is true, then the results provide no evidence that the lidocaine treatment used in this case diminished pain from tagging. Measuring eye temperature in seals may provide an objective, immediate and non-invasive method for evaluating their responses to husbandry and handling practices, thus enabling subsequent refinement of such practices to improve seal welfare.  132 5. General Discussion and Conclusions  5.1 Thesis findings To my knowledge, prior to this thesis there was no published research on species-specific indicators of pain in harbour seals. The only published work specifically on pinniped pain addressed the responses of Stellar sea lions to hot iron branding (Walker et al., 2011a) and the intra-abdominal surgical implantation of tracking devices (Walker et al., 2011b). Moreover, very little work has been done to establish pain management protocols for any species of pinniped (Molter et al., 2015). Difficulties in recognizing animal pain and uncertainty about appropriate drug therapies are reported as the primary barriers to effective pain management (Hugonnard et al., 2004). In fact, our poor ability to assess animal pain may factor foremost in our frequent failures to manage it adequately (Flecknell, 2000). Harbour seals are a marine mammal species frequently under human care (MacRae et al., 2011), and individuals admitted to rehabilitation often present with potentially painful conditions. Hence, there is an obvious need to develop reliable methods for the identification and evaluation of pain in this species. My experimental research summarized in this thesis has focused on three potential, non-invasive methods for identifying pain in harbour seals: 1) changes in facial expression and other behaviour, 2) changes in vocalizations and, 3) changes in eye temperature.  To summarize briefly, in Chapter 2, I investigated whether harbour seals’ facial expressions and several other behaviours changed in response to the potentially painful procedures of flipper tagging and microchipping. The analysis of facial expressions has   133 proven to be a repeatable, accurate and valid method of identifying pain in multiple species (Chambers and Mogil, 2015), but facial expressions had not been examined in any species of pinniped. These results showed that harbour seals do consistently react to a presumably painful stimulus by quantitative changes in at least one facial action unit, orbital tightening, which increased almost three-fold from before to after tagging and chipping in all seals but did not do so after sham treatments.  Nose bulge was another facial action unit that we identified as a possible indicator of pain. Nose bulge was observed in only a small portion of the seals; however, it occurred only after the procedures (not before) and never after the sham procedures. Changes in the nose/cheek region have been demonstrated as an important facial action in response to pain in many species (Chambers and Mogil, 2015). Nose bulge, specifically, has been identified as a valid pain indicator in at least one other species, the mouse (Langford et al., 2010).  Although my work showed that nose-bulge occured in only a small number of individuals, it should not be dismissed as a sign of pain because not all individuals express pain in the same way. For example, a study on the facial expressions and cry responses of human infants reported that not all infants cried in response to the noxious stimulus heel-lance (Grunau and Craig, 1987), yet infant cry reactions are still considered an important behavioural response to pain (Grunau et al., 1990). Therefore, nose bulge may be significant in seals but it requires more investigation before being considered a reliable indicator of pain.    In general, the seals appeared alert and active before tagging and chipping but more subdued afterward, with reduced overall body movement (less looking around, struggling and trembling).  These differences did not exist after the sham procedure,   134 suggesting that these observed changes were not the result of exhaustion or restraint.  Appearing quieter and less reactive to environmental stimuli is consistent with what is reported in domestic cats and dogs in pain (National Research Council, 1992).   Importantly, both observers responsible for scoring facial and behavioural changes in this study could, with considerable accuracy, distinguish between animals that had undergone the procedures and those that had not, pointing to the promise of these measures for acute pain assessment in seals.  Nonetheless, one of the limitations of this study was the low number of observers. We had only two observers for the first experiment and, once a high level of agreement had been established between observers, only a single observer for the second two experiments. Future studies with more observers could test how reliably different people score the key measures, especially orbital tightening. Additionally, if facial changes are to be useful for ‘cage-side’ assessment in seals, work would need to show that observers could apply the measures in real time rather than from video recordings. Vocalizations are another potential indicator of pain in animals, and the analysis of vocalizations has been used to study pain-inducing procedures in a variety of species (e.g., Watts and Stookey, 1999; Weary et al., 1998). In Chapter 3, I examined seal vocalizations in response to tagging and chipping and showed that both the number and peak frequency (Hz) of seal pups' calls consistently increased following both procedures.  Both the increased call rate and increased peak frequency are consistent with vocal changes seen in other species in response to pain (e.g., pigs; Taylor and Weary, 2000). Seal pups showed no similar changes in their vocalizations after being sham treated, suggesting the observed vocal changes reflect pain caused by the   135 procedure rather than simple restraint. As a means of monitoring pain in seals, vocal analysis thus appears promising, but its usefulness may be chiefly for young animals.  The vocalizations I recorded were consistent with ‘mother attraction calls,’ which typically cease once pups are weaned (Khan et al., 2006). Generally, adult harbour seals are considered vocally reticent compared to other pinnipeds. Adult females are relatively non-vocal and do not return their pups’ calls, as most other pinniped species do (Renouf, 1984). Interestingly, the males become vocal once again when they reach sexual maturity and begin to produce underwater ‘songs’ during the breeding season (Schusterman and Van Parijs, 2003).  As well as these notable changes in their vocal behaviour as they develop, vocal parameters also change; for example, they begin to have lower fundamental and minimum frequencies (Khan et al., 2006; Sauvé et al., 2015b). Therefore, the use of vocalizations to indicate pain in seals over the age of natural weaning needs further investigation as the ability to produce these calls likely diminishes with age and calls of the type that I studied eventually drop out of the seals’ vocal repertoire.  Whereas the focus of both Chapter 2 and 3 was on behavioural indicators of pain, in Chapter 4 I examined a potential physiological indicator – changes in eye temperature. Short-term temperature responses due to peripheral blood flow are associated with the activation of the autonomic nervous system and may be a physiological correlate of an animal’s affective state in response to painful, stressful or arousing stimuli (Stewart, 2008). Therefore, the aim in this study was to determine whether the eye temperature of harbour seals changes in response to routine handling   136 and the potentially painful procedure of flipper tagging, and if responses to tagging would be mitigated by a local anaesthetic (lidocaine). Results from this study showed that the eye temperature of tagged pups did, in fact, increase after the procedure, compared to pups that were sham tagged, which suggests that a rise in eye temperature can reflect pain. However, when comparing pups that were handled or not handled eye temperature also increased when pups were first restrained. Additionally, eye temperature increased further in pups that underwent a second restraint. The higher eye temperature of handled versus non-handled pups suggests that handling and restraint cause a physiological stress response that is detectable via IRT. The increased temperature seen in the second handling suggests the first handling was aversive, resulting in pups' anticipatory response to their second handling. Increased eye temperature in response to an event previously learned to be aversive has also been demonstrated in calves which showed increased eye temperature the second time they experienced jugular catheterization but not the first (Stewart et al., 2008a). Surprisingly, eye temperature also increased after pups received an injection of lidocaine but not saline. There is evidence that lidocaine causes vasodilation, which could increase local skin temperatures; however, this vasodilatory response is localized to the area immediately surrounding the injection site (Newton et al., 2007). Seals received only a 1-ml injection in their hind flipper and this could not likely account for an increased temperature in the eye.  Similarly, if lidocaine were causing a vasodilatory reaction affecting eye temperature, the increase should have been seen in the 10-min period following its administration. However, in that period, the eye temperature of   137 lidocaine-treated pups was no different than that of animals administered saline. Both saline and lidocaine injections used the same injection method, gauge of needle and volume of injectable fluid. Therefore, it seems more likely that the increase in eye temperature following the injection of lidocaine but not saline resulted from the painfulness of the lidocaine application itself, as is known from other species (Cepeda et al., 2012).  5.2 Discussion and conclusions Problematically, many of the indicators used to assess pain are not necessarily pain-specific as they may present in other contexts that do not involve pain (Bateson, 1991). Of the measures I examined, this may be particularly true of eye temperature. Influenced by both sympathetic and hypothalamic-pituitary-adrenocortical responses (Stewart et al., 2010), changes in eye temperature may be triggered by a variety of stimuli. The fact that eye temperature generally increased during the minutes after tagging but not sham-tagging, and after lidocaine injection but not after saline injection, suggests that such increases can reflect pain. However, increases after handling as well as after painful procedures suggest sensitivity rather than specificity to detecting pain. That said, there may be scope for using IRT to discriminate between stressors such as fear and pain. For example, cattle had an immediate, transient decrease in eye temperature both after being startled by a flapping bag or loud shout and after being administered a painful electric prod. However, it took much longer for a return to baseline temperatures after the acute pain from the electric prod than it did after being frightened (Stewart et al., 2007a).  Further research might determine whether such   138 differences in the magnitude of the response reflect differences in the severity of the stressor.  I wanted to examine procedures that were already routine in the management of the seals. Therefore, I used tagging and chipping as the potentially painful stimuli in all my studies as these procedures are required by the federal government department of Fisheries and Oceans Canada for every animal prior to release. However, there was some experimental variation between studies in how the procedures were approached.  The study on facial expression (Chapter 2) examined the combined effect of tagging followed by chipping which, together took approximately two minutes. Video observation occurred until five seconds before the onset of tagging and recommenced five seconds after chipping but did not include the two minutes during the procedures. This choice of sampling interval was largely a practical one as the amount of struggling and vocalizing during the two minutes of the procedures made facial change difficult to assess. However, an approach for future work could be to use a single stimulus and focus on the seconds immediately after application. In a study on human infants, 99% tested demonstrated the facial changes of brow bulge, eye squeeze, naso-labial furrow and open lips within six seconds of a heel-lance procedure (Grunau and Craig, 1987).  The majority of seal vocalizations (Chapter 3) were clustered in the periods immediately after each procedure; this allowed the responses to tagging and chipping to be studied separately. In Chapter 4, when examining changes in eye temperature, I decided only to tag and not chip the seals in order to minimize their handling and the number of sites to pre-inject with lidocaine. In consultation with the staff veterinarian, we assumed that a single lidocaine injection was more likely to effectively block the pain   139 from tagging than from chipping, due to the location of the sites on the seals’ bodies. Acute, cutaneous pain likely results from both tagging and chipping. Future studies examining seal responses to other types of routine, potentially painful procedures are required to further validate the measures.   Throughout this thesis, I attempt to acknowledge limitations specific to each of my studies; however, the significant challenge that prevailed across studies was the lack of validated pain management for this species. As discussed in Chapter 1, the validation of pain measures requires responses to be observed with and without the painful condition and with and without effective pain medication. The use of analgesics is a standard method of testing whether behavioural responses are due to pain (Weary et al., 2006), but this method requires an analgesic of known effectiveness. Problematically, pharmacological data for most species of marine mammal do not currently exist and, to date, research on the use of analgesics in pinnipeds is limited to a few preliminary pharmacokinetic studies in only a few species, for example with tramadol in California sea lions (Zalophus californianus) (Boonstra et al., 2013) and butorphanol and buprenorphine in elephant seals (Mirounga angustirostris) (Molter et al., 2015; Nutter et al., 1998). Because of the lack of data on effective drug therapies, it is often necessary to extrapolate doses and pharmacokinetic parameters across species (Molter et al., 2015) without certainty of how effective such translation may be. In Chapter 2, I piloted a single drug (buprenorphine) at a dosage recommended for analgesia in dogs. Buprenorphine had no observable effect on any of the seals' facial action units or behaviours; this may simply be because the drug and/or the low dosage rate were inadequate for this type of acute pain in this species. The drug's apparent   140 ineffectiveness also could have been due to a lack of statistical power, as buprenorphine was piloted in only five animals.  In Chapter 4, seals were administered a single injection of lidocaine at the tag site. Lidocaine is reported to be an effective pain reliever in pinnipeds (Gutiérrez et al., 2016), but there was no pre-existing evidence of its efficacy for mitigating the pain of tagging. The 1 ml dose of lidocaine used in Chapter 4 was established by manually pinching the skin and flipper webbing in the area of the lidocaine injection and observing the animals’ reactions. The pups that were used to pilot the lidocaine appeared to have reduced reactivity to pinching (e.g., limb withdrawal) after the injection. However, in this study the 1 ml dose did not attenuate the change in eye temperate in response to tagging likely because tagging was more invasive than pinching.   If lidocaine did have some mitigating effect on the pain from tagging, the change in eye temperature would be expected to be greater in the animals that had been tagged without lidocaine. However, there was no clear difference in mean eye temperature for pups that were injected with lidocaine compared to saline in the period after tagging. In actuality, eye temperature tended to increase for the Lidocaine treatment but not for the Saline treatment. Lidocaine, at the one dosage tested, did not appear to reduce the eye-temperature response to tagging; in fact, it may have exacerbated it. In all three studies, seals showed changes in all measures from before to after the procedures consistent with what has been seen in other species in response to pain.  However, a treatment group using an effective analgesic would have provided stronger evidence that the changes observed were, in fact, due to pain. There are both advantages and disadvantages to conducting research on animals   141 in a rehabilitation environment. Although pinnipeds are widespread in the wild and common in captivity compared to other marine mammal species (Brando et al., 2017), the number of individuals of each species kept at a single zoo or aquarium tends to be small (Higgins and Hendrickson, 2013). Thus, the Vancouver Aquarium Marine Mammal Rescue Centre provided an unusual opportunity to access a relatively large group of the same species, all of similar age and reared under the same conditions.  Even so, depending on the type of research, available sample sizes may still be quite limited by the number of available individuals.  For example, my work required pups to have been weaned, deemed healthy, and to be over six weeks old at the time of testing.  That meant that the size of treatment groups for each project was constrained by the number of individuals meeting those criteria as opposed to being established by an a priori analysis of statistical power.  Furthermore, doing research in a rehabilitation context requires constant awareness of the main rehabilitation objective, which is to return healthy animals to the wild.  Therefore, any experimental manipulations (e.g., extra handling that could result in habituation) had to be kept to a minimum lest they lower seals' chances of success once released.  There is also scope for more detailed analysis of harbour seal vocalizations. Some species have complex vocal repertoires with multiple types of call. For example, vervet monkeys (Cercopithecus aethiops) give acoustically distinct alarm calls for different kinds of predators. Their calls for ‘eagle’, ‘leopard’, or ‘snake’ each elicit different responses in the listener that are appropriate for the different types of danger. Such vocalizations contain information about the signaler’s affective state (alarm) as well as information that categorizes predator types for the listener (Seyfarth et al.,   142 1980). In pigs, 14 different call types have been distinguished by multiple acoustic parameters including amplitude, frequency (Hz), duration and grunt pattern (Kiley, 1972). For example, low-frequency grunts maintain social contact (Kiley, 1972) and a relatively fixed pattern of grunting rate indicates let-down of milk (Ellendorf et al., 1982). In harbour seal pups fundamental frequency may transmit individual pup identity, and increased calling bouts, call rate and harmonics may signal states of distress (Perry and Renouf, 1988; Renouf, 1984) but in general their vocalizations do not seem to have a lot of variation.   A study on vocal development in harbour seals analyzed close to 5000 vocalizations from 15 unweaned pups and found that the large majority of calls had very similar acoustic properties (e.g., relatively tonal, having an inverted ‘v’-or ‘u’-shaped spectrogram with the fundamental frequency around 200–600 Hz and harmonics). Only 19 of the total recorded calls were different, being described as harsh, broadband, staccato calls used in an aggressive context toward another seal pup or human (Khan et al., 2006). In my study on seal vocalizations (Chapter 3), I included only call number, duration and peak frequency (Hz); hence, future work could examine other acoustic components of the calls given in response to pain. That said, the calls of harbour seal pups may be a relatively simple signaling system considering that the primary function of these vocalizations is maternal contact and that the calls cease once individuals are independent (Khan et al., 2006; Sauvé et al., 2015a). As discussed in Chapter 1, no single perfect method exists for the assessment of pain in any species. Nevertheless, with this thesis I have tried to provide evidence-based research on both behavioural and physiological indicators of pain in young   143 harbour seals with the aim of contributing to the understanding of some of their modes of pain expression. I suggest that orbital tightening may be particularly useful for identifying pain in seals as this facial change occured in all seals after the procedure (and not during the sham procedure), was easily recognized by observers and has been identified as a key facial action unit in all animal grimace scales developed thus far. As well, vocalization — especially changes in the number and peak frequency (Hz) of calls — also shows promise for the assessment of pain in seal pups. Lastly, I demonstrate the use of eye temperature as a stress-response indicator and suggest its potential for evaluating the possible averseness of routine procedures in seals, including those that cause pain. As each method has limitations, a multi-factorial pain scoring system that integrates several measures may be more effective at capturing the different dimensions of the seals’ pain experience. Taken together the facial, vocal and eye temperature responses of the seals to tagging and chipping strongly suggest that these procedures do, in fact, cause pain. Therefore, I would urge that future research focus on establishing effective pain management for this species and, once established, that appropriate analgesic therapy be administered during both tagging and chipping procedures.     144 References Aghajani, M., Mahdavi, M.R.V., Najafabadi, M.K., Ghazanfari, T., Azimi, A., Soleymani, S.A., Dust, S.M., 2013. Effects of dominant/subordinate social status on formalin-induced pain and changes in serum proinflammatory cytokine concentrations in mice. PLoS One 8. https://doi.org/e80650 Alam, M.G.S., Dobson, H., 1986. The effect of various veterinary procedures on plasma concentrations of cortisol, luteinising hormone and prostogalndin F2 alpha metabolite in the cow. Vet. J. 118, 7–10. Allegaert, K., van den Anker, J.N., 2016. Neonatal pain management: still in search for the Holy Grail. Int. J. Clin. Pharmacol. Ther. 54, 514–523. https://doi.org/10.5414/CP202561 Altman, D.G., Bland, J.M., 1994. Diagnostic tests 1: sensitivity and specificity. Br. Med. J. 308, 1552. Anand, K.J.S., Craig, K.D., 1996. New perspectives on the definition of pain. Pain 67, 3–6. https://doi.org/10.1016/0304-3959(96)03135-1 Anil, S.S., Anil, L., Deen, J., 2002. Challenges of pain assessment in domestic animals. J. Am. Vet. Med. Assoc. 220, 313–319. https://doi.org/10.2460/javma.2002.220.313   145 Bates, D., Mächler, M., Bolker, B., Walker, S., 2015. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48. https://doi.org/10.18637/jss.v067.i01 Bateson, P., 1991. Assessment of pain in animals. Anim. Behav. 42, 827–839. Beerda, B., Schilder, M.B.H., van Hooff, J.A.R.A.M., de Vries, H.W., Mol, J.A., 1998. Behavioural, saliva cortisol and heart rate responses to different types of stimuli in dogs. Appl. Anim. Behav. Sci. 58, 365–381. https://doi.org/10.1016/S0168-1591(97)00145-7 Benson, G.J., Grubb, T.L., Neff-Davis, C., Olson, W.A., Thurmon, J.C., Lindner, D.L., Tranquilli, W.J., Vanio, O., 2000. Perioperative stress response in the dog: effect of pre-emptive administration of medetomidine. Vet. Surg. 29, 85–91. Bermond, B., 2001. A neuropsychological and evolutionary approach to animal consciousness and animal suffering. Anim. Welf. 10, S47–S62. Blanchard, K.T., Barthel, C., French, J.E., Holden, H.E., Moretz, R., Pack, F.D., Tennant, R.W., Stoll, R.E., 1999. Transponder-induced sarcoma in the heterozygous p53+/- mouse. Toxicol. Pathol. 27, 519–527. Blessing, W.W., 2003. Lower brainstem pathways regulating sympathetically mediated changes in cutaneous blood flow. Cell. Mol. Biol. 23, 527–538.   146 Boness, D.J., Bowen, W.D., Oftedal, O.T., 1994. Evidence of a maternal foraging cycle resembling that of otariid seals in a small phocid, the harbor seal. Behav. Ecol. Sociobiol. 34, 95–104. Boonstra, J.L., Barbosa, L., Van Bonn, W.G., Gulland, F.M.D., 2013. Pharmacokinetic study of orally administered tramadol in California sea lions (Zalophus californianus), in: Proceedings of the American Association of Zoo Veterinarians. p. 39. Brando, A.S., Broom, D.M., Clark, F., 2017. Optimal marine mammal welfare under human care: Current efforts and future directions. Behav. Processes. https://doi.org/10.1016/j.beproc.2017.09.011 Brearley, J.C., Brearley, M.J., 2000. Chronic pain in animals, in: Flecknell, P.A., Waterman-Pearson, A. (Eds.), Pain Management in Animals. W.B. Saunders, London, UK, pp. 147–160. https://doi.org/10.1016/B978-0-7020-1767-4.50008-4 Briefer, E.F., 2012. Vocal expression of emotions in mammals: mechanisms of production and evidence. J. Zool. 288, 1–20. https://doi.org/10.1111/j.1469-7998.2012.00920.x Broom, D.M., 1998. Welfare, stress, and the evolution of feelings, in: Moller, A.P., Miliniski, M., Slater, P.J.B. (Eds.), Advances in the Study of Behavior: Stress and Behavior. Academic Press, San Diego, USA, pp. 371–404.   147 Brown, C., 2016. Comparative evolutionary approach to pain perception in fishes. Anim. Sentience 1, 1–7. Brown, M., Carbone, L., Conlee, K., Dawkins, M., Duncan, I.J., Fraser, D., Griffin, G., Hampshire, V.A., Lamber, L.A., Mench, J.A., Morton, D., Richmond, J., Rollin, B.E., Rowan, A.N., Stephens, M.L., Würbel, H., 2006. Report of the Working Group on Animal Distress in the Laboratory. Lab. Anim. 35, 26–30. Buynitsky, T., Mostofsky, D.I., 2009. Restraint stress in biobehavioral research: Recent developments. Neurosci. Biobehav. Rev. 33, 1089–1098. https://doi.org/10.1016/j.neubiorev.2009.05.004 Caeiro, C.C., Burrows, A.M., Waller, B.M., 2013a. CatFACS: The cat facial action coding system. University of Portsmouth, U.K. www.CatFACS.com Caeiro, C.C., Waller, B.M., Zimmermann, E., Burrows, A.M., Davila-Ross, M., 2013b. OrangFACS: A muscle-based facial movement coding system for orangutans (Pongo spp.). Int. J. Primatol. 34, 115–129. https://doi.org/10.1007/s10764-012-9652-x Campbell, J.N., Meyer, R.A., 2006. Mechanisms of neuropathic pain. Neuron 52, 77–92. https://doi.org/10.1016/j.neuron.2006.09.021 Carstens, E., Moberg, G.P., 2000. Recognizing pain and distress in laboratory animals.   148 ILAR J. 41, 62–71. https://doi.org/10.1093/ilar.41.2.62 Cepeda, M.S., Tzortzopoulou, A., Thackrey, M., Hudcova, J., Arora Gandhi, P., Schumann, R., 2012. Adjusting the pH of lidocaine for reducing pain on injection. Evidence-Based Child Heal. 7, 149–215. https://doi.org/10.1002/ebch.1811 Chambers, C.T., Mogil, J.S., 2015. Ontogeny and phylogeny of facial expression of pain. Pain 156, 798–799. https://doi.org/10.1097/j.pain.0000000000000133 Chapman, C.R., Casey, K.L., Dubner, R., Foley, K.M., Gracely, R.H., Reading, A.E., 1985. Pain measurement: an overview. Pain 22, 1–31. Chesler, E.J., Wilson, S.G., Lariviere, W.R., Rodriguez-Zas, S.L., Mogil, J.S., 2002. Influences of laboratory environment on behavior. Nat. Neurosci. 51, 1101–1102. Clark, R.W., 1994. Central control of pain, in: Rothwell, N.J., Berkenbosch, F. (Eds.), Brain Control of Responses to Trauma. Cambridge University Press, Cambridge, UK, pp. 295–331. Coetzee, J.F., Lubbers, B.L., Toerber, S.E., Gehring, R., Thomson, D.U., White, B.J., Apley, M.D., 2008. Plasma concentrations of substance P and cortisol in beef calves after castration or simulated castration. Am. J. Vet. Res. 69, 751–762. Colpaert, F.C., DeWitte, P., Maroli, A.N., Awouters, F., Niemegeers, C.J.E., Janssen,   149 P.A.J., 1980. Self-administration of the analgesic suprofen in arthritic rats: evidence of Mycobacterium butyricum induced arthritis as an experimantal model of chronic pain. Life Sci. 27, 921–928. Colpaert, F.C., Meert, T., DeWitte, P., Schmitt, P., 1982. Further evidence validating adjuvant arthritis as an experimental model of chronic pain in the rat. Life Sci. 31, 67–75. Conzemius, M.G., Hill, C.M., Sammarco, J.L., Perkowski, S.Z., 1997. Correlation between subjective and objective measures used to determine severity of post-operative pain in dogs. J. Am. Vet. Med. Assoc. 210, 1619–1622. Cook, N.J., 2012. Review: Minimally invasive sampling media and the measurement of corticosteroids as biomarkers of stress in animals. Can. J. Anim. Sci. 92, 227–259. https://doi.org/10.4141/cjas2012-045 Cook, N.J., Chabot, B., Lui, T., Bench, C.J., Schaefer, A.L., 2015. Infrared thermography detects febrile and behavioural responses to vaccination of weaned piglets. Animal 9, 339–346. https://doi.org/10.1017/S1751731114002481 Cook, N.J., Church, J.S., Schaefer, A.L., Webster, J.R., Matthews, L.R., Suttie, J.M., 2005. Stress and pain assessment of velvet antler removal from Elk (Cervus elaphus canadensis) and Reindeer (Rangifer tarandus). Online J. Vet. Res. 9, 24–36.   150 Cook, N.J., Schaefer, A.L., 2002. Stress responses of wapiti (Cervus elaphus canadensis) to removal of velvet antler. Can. J. Anim. Sci. 82, 11–17. Cook, N.J., Schaefer, A.L., Warren, L., Burwash, L., Anderson, M., Baron, V., 2001. Adrenocortical and metabolic responses to ACTH injection in horses: An assessment by salivary cortisol and infrared thermography of the eye. Can. J. Anim. Sci. 81, 621. Cooper, B.Y., Vierck, C.J., 1986. Vocalizations as measures of pain in monkeys. Pain 26, 393–407. https://doi.org/10.1016/0304-3959(86)90065-5 Cottrell, P.E., Jeffries, S., Beck, B., Ross, P.S., 2002. Growth and development in free-ranging harbor seal Phoca vitulina pups from southern British Columbia, Canada. Mar. Mammal Sci. 18, 721–733. Coulter, C.A., Flecknell, P.A., Richardson, C.A., 2009. Reported analgesic administration to rabbits, pigs, sheep, dogs and non-human primates undergoing experimental surgical procedures. Lab. Anim. 43, 232–238. https://doi.org/10.1186/1746-6148-7-12 Craig, K.D., 1994. Emotional aspects of pain, in: Wall, P.D., Melzack, R. (Eds.), Textbook of Pain. Edinburgh: Churchill Livingstone, pp. 261–274. Craig, K.D., 1992. The facial expression of pain: Better than a thousand words? Phys.   151 Rev. Focus 1, 153–162. https://doi.org/10.1016/1058-9139(92)90001-S Craig, K.D., Hyde, S.A., Patrick, C.J., 1991. Genuine, suppressed and faked facial behavior during exacerbation of chronic low back pain. Pain 46, 161–171. https://doi.org/10.1016/0304-3959(91)90071-5 Craig, K.D., Patrick, C.J., 1985. Facial expression during induced pain. J. Pers. Soc. Psychol. 48, 1080–1091. Craig, K.D., Prkachin, K.M., Grunau, R.V.E., 2001. The facial expression of pain, in: Turk, D.C., Melzack, R. (Eds.), Handbook of Pain Assessment. New York: Guilford, pp. 153–169. Craig, K.D., Versloot, J., Goubert, L., Vervoort, T., Crombez, G., 2010. Perceiving others in pain: automatic and controlled mechanisms. J. Pain 11, 101–108. Craig, K.D., Whitfield, M.F., Grunau, R.V.E., Linton, J., Hadjistavropoulos, H.D., 1993. Pain in the preterm neonate: behavioural and physiological indices. Pain 52, 287–299. https://doi.org/10.1016/0304-3959(93)90162-I Dahlborn, K., Bugnon, P., Nevalainen, T., Raspa, M., Verbost, P., Spangenberg, E., 2013. Report of the Federation of European Laboratory Animal Science Associations Working Group on Animal Identification. Lab. Anim. 47, 2–11. https://doi.org/10.1177/002367712473290   152 Dai, F., Cogi, N.H., Heinzl, E.U.L., Dalla Costa, E., Canali, E., Minero, M., 2015. Validation of a fear test in sport horses using infrared thermography. J. Vet. Behav. Clin. Appl. Res. 10, 128–136. https://doi.org/10.1016/j.jveb.2014.12.001 Dalla Costa, E., Minero, M., Lebelt, D., Stucke, D., Canali, E., Leach, M.C., 2014. Development of the horse grimace scale (HGS) as a pain assessment tool in horses undergoing routine castration. PLoS One 9, e92281. https://doi.org/10.1371/journal.pone.0092281 Dalla Costa, E., Stucke, D., Dai, F., Minero, M., Leach, M., Lebelt, D., 2016. Using the horse grimace scale (HGS) to assess pain associated with acute laminitis in horses (Equus caballus). Animals 6, 47. https://doi.org/10.3390/ani6080047 Danbury, T.C., Weeks, C.A., Chambers, J.P., Waterman-Pearson, A.E., Kestin, S.C., 2000. Self-selection of the analgesic drug carprofen by lame broiler chickens. Vet. Rec. 146, 307–311. https://doi.org/10.1136/vr.146.11.307 Darwin, C., 1872. The expression of the emotions in man and animals. Albemarle, London, UK. Dawkins, M.S., 2017. Animal welfare with and without consciousness. J. Zool. 301, 1–10. https://doi.org/10.1111/jzo.12434 Derbyshire, S.W.G., 2003. Fetal ‘pain’ – a look at the evidence. Am. Pain Soc. Bull. 13,   153 1–4. Deyo, K.S., Prkachin, K.M., Mercer, S.R., 2004. Development of sensitivity to facial expression of pain. Pain 107, 16–21. https://doi.org/10.1016/S0304-3959(03)00263-X Di Giminiani, P., Brierley, V.L.M.H., Scollo, A., Gottardo, F., Malcolm, E.M., Edwards, S.A., Leach, M.C., 2016. The assessment of facial expressions in piglets undergoing tail docking and castration: toward the development of the piglet grimace scale. Front. Vet. Sci. 3, 1–10. https://doi.org/10.3389/fvets.2016.00100 Di Giminiani, P., Nasirahmadib, A., Malcolm, E.M., Leach, M.C., Edwards, S.A., 2017. Docking piglet tails: How much does it hurt and for how long? Physiol. Behav. 182, 69–76. Dinh, H.K., Larkin, A., Gatlin, L., Piepmeier, E., 1999. Rat ultrasound model for measuring pain resulting from intramuscularly injected antimicrobials. J. Pharm. Sci. Technol. 53, 40–43. Dobromylskyj, P., Flecknell, P.A., Lascelles, B.D., Livingston, A., Taylor, P., Waterman-Pearson, A., 2000a. Pain assessment, in: Flecknell, P.A., Waterman-Pearson, A.E. (Eds.), Pain Management in Animals. W.B. Saunders, London, UK, pp. 53–79. https://doi.org/10.1016/B978-0-7020-1767-4.50007-2   154 Dobromylskyj, P., Flecknell, P.A., Lascelles, B.D., Pascoe, P.J., Taylor, P., Waterman-Pearson, A., 2000b. Management of postoperative and other acute pain, in: Flecknell, P.A., Waterman-Pearson, A. (Eds.), Pain Management in Animals. W.B. Saunders, London, UK, pp. 81–145. https://doi.org/10.1016/B978-0-7020-1767-4.50008-4 Dohoo, S.E., Dohoo, I.R., 1996a. Factors influencing the postoperative use of analgesics in dogs and cats by Canadian veterinarians. Can. Vet. J. 37, 552–556. Dohoo, S.E., Dohoo, I.R., 1996b. Postoperative use of analgesics in dogs and cats by Canadian veterinarians. Can. Vet. J. 37, 546–551. Dunbar, M.R., MacCarthy, K., 2006. Use of infrared thermography to detect signs of rabies infection in raccoons (Procyon lotor). J. Zoo Wildl. Med. 37, 518–523. Dunbar, M.R., Ryan, J.C., Mccollum, M., 2009. Use of infrared thermography to detect thermographic changes in Mule deer (Odocoileus hemionus) experimentally infected with foot-and-mouth disease. J. Zoo Wildl. Med. 40, 296–301. Duncan, I.J.H., Slee, G.S., Seawright, E., Breward, J., 1989. Behavioural consequences of partial beak amputation (beak trimming) in poultry. Br. Poult. Sci. 30, 479–488. Eccleston, C., Crombez, G., 1999. Pain demands attention: A cognitive-affective model of the interuptive function of pain. Psychol. Bull. 125, 356–366.   155 Eckersall, P., 2000. Recent advances and future prospects for the use of acute phase proteins as markers of disease in animals. Rev. Med. Vet. (Toulouse). 151, 577–584. Eddy, A.L., Van Hoogmoed, L.M., Snyder, J.R., 2001. The role of thermography in the management of equine lameness. Vet. J. 162, 172–181. Edgar, J.L., Nicol, C.J., Pugh, C.A., Paul, E.S., 2013. Surface temperature changes in response to handling in domestic chickens. Physiol. Behav. 119, 195–200. https://doi.org/10.1016/j.physbeh.2013.06.020 Eicher, S.D., Cheng, H.W., Sorrells, A.D., Schutz, M.M., 2006. Behavioral and physiological indicators of sensitivity or chronic pain following tail docking. J. Dairy Sci. 89, 3047–3051. Ekman, P., 1993. Facial expression and emotion. Am. Psychol. 48, 384–392. https://doi.org/http://dx.doi.org/10.1037/0003-066X.48.4.384 Ekman, P., Friesen, W., 1978. Facial Action Coding System: A technique for the measurement of facial movement, Consulting Psychologists Press. Palo Alto. Ellendorf, F., Forsling, M.L., Poulain, D.A., 1982. The milk ejection reflex in the pig. J. Physiol. 333, 577–594.   156 Eskesen, I.G., Teilmann, J., Geertsen, B.M., Desportes, G., Riget, F., Dietz, R., Larsen, F., Siebert, U., 2009. Stress level in wild harbour porpoises (Phocoena phocoena) during satellite tagging measured by respiration, heart rate and cortisol. J. Mar. Biol. Assoc. United Kingdom 89, 885–892. https://doi.org/10.1017/S0025315408003159 Faulkner, P.M., Weary, D.M., 2000. Reducing pain after dehorning in dairy calves. J. Dairy Sci. 83, 2037–2041. https://doi.org/10.3168/jds.S0022-0302(00)75084-3 Fenner, K., Yoon, S., White, P., Starling, M., Mcgreevy, P., 2016. The effect of noseband tightening on horses ’ behavior, eye temperature, and cardiac responses. PLoS One 1–20. https://doi.org/10.1371/journal.pone.0154179 Filippi, P., Congdon, J. V., Hoang, J., Bowling, D.L., Reber, S.A., Pašukonis, A., Hoeschele, M., Ocklenburg, S., de Boer, B., Sturdy, C.B., Newen, A., Güntürkün, O., 2017. Humans recognize emotional arousal in vocalizations across all classes of terrestrial vertebrates: Evidence for acoustic universals. Proc. R. Soc. B Biol. Sci. 284, 1–9. https://doi.org/10.1098/rspb.2017.0990 Firth, A.M., Haldane, S.L., 1999. Development of a scale to evaluate postoperative pain in dogs. J. Am. Med. Assoc. 214, 651–659. Fitzpatrick, J., Scott, M., Nolan, A., 2006. Assessment of pain and welfare in sheep. Small Rumin. Res. 62, 55–61. https://doi.org/10.1016/j.smallrumres.2005.07.028   157 Fitzpatrick, J.L., Nolan, A.M., Scott, E.M., Harkins, L., Barrett, D.C., 2002. Observers perception of pain in cattle. Cattle Pract. 10, 209–212. Flecknell, P.A., 2010. Do mice have a pain face? Nat. Methods 7, 437–438. https://doi.org/10.1038/nmeth0610-437 Flecknell, P.A., 2000. Animal pain - an introduction, in: Flecknell., P.A., Waterman-Pearson, A. (Eds.), Pain Management in Animals. W.B. Saunders, London, UK, pp. 1–7. https://doi.org/10.1016/B978-0-7020-1767-4.50004-7 Fraser, D., 2008. Understanding Animal Welfare: The Science in its Cultural Context. Wiley-Blackwell, Oxford, UK. Fraser, D., 1975. Vocalizations of isolated piglets. 1. Sources of variation and relationships among measures. Appl. Anim. Ethol. 1, 387–394. Fraser, D., Kramer, D.L., Pajor, E.A., Weary, D.M., 1995. Conflict and cooperation: Sociobiological principles and the behavior of pigs. Appl. Anim. Behav. Sci. 44, 139–157. https://doi.org/10.1016/0168-1591(95)00610-5 Fraser, D., Weary, D.M., Pajor, E.A., Milligan, B.N., 1997. A scientific conception of animal welfare that reflects ethical concerns. Anim. Welf. 6, 187–205. Freedman, D.G., 1964. Smiling in blind infants and the issue of innate vs. acquired. J.   158 Child Psychol. Psychiatry 5, 171–184. Gao, Y.J., Ren, W.H., Zhang, Y.Q., Zhao, Z.Q., 2004. Contributions of the anterior cingulate cortex and amygdala to pain-and fear-conditioned place avoidance in rats. Pain 110, 343–353. Garnett, A., Merkies, K., 2017. Decreased eye-blink rate as a non-invasive measure of stress in the domestic horse, in: Randle, H., Waran, N., Kent, L. (Eds.), 13th International Equitation Science Conference. Wagga Wagga, Australia, p. 37. Gentle, M.J., 2001. Attentional shifts alter pain perception in the chicken. Anim. Welf. 10, 187–194. Gentle, M.J., Corr, S.A., 1995. Endogenous analgesia in the chicken. Neurosci. Lett. 201, 211–214. Gerber, M.I., Swinker, A.M., Staniar, W.B., Werner, J.R., Jedrzejewski, E.A., Macrina, A.L., 2012. Health factors associated with microchip insertion in horses. J. Equine Vet. Sci. 32, 177–182. https://doi.org/10.1016/j.jevs.2011.08.016 Gioiosa, L., Chiarotti, F., Alleva, E., Laviola, G., 2009. A trouble shared is a trouble halved: social context and status affect pain in mouse dyads. PLoS One 4, e4143. Gleerup, K., Forkman, B., Lindegaard, C., Andersen, P., 2015. An equine pain face.   159 Vet. Anaesth. Analg. 42, 103–114. https://doi.org/10.1111/vaa.12212 Godfray, H.C.J., 1991. Signalling of need by offspring to their parents. Nature 352, 328–330. Godfray, H.C.J., Johnstone, R.A., 2000. Begging and bleating: The evolution of parent-offspring signalling. Philos. Trans. R. Soc. B Biol. Sci. 355, 1581–1591. https://doi.org/10.1098/rstb.2000.0719 Gregory, N.G., 2004. Physiology and Behaviour of Animal Suffering. Blackwell Publishing, Oxford, UK. https://doi.org/10.1002/9780470752494 Grissom, N., Kerr, W., Bhatnagar, S., 2008. Struggling behavior during restraint is regulated by stress experience. Behav. Brain Res. 191, 219–226. https://doi.org/10.1016/j.bbr.2008.03.030 Grunau, R. V., Craig, K.D., 1987. Pain expression in neonates: facial action and cry. Pain 28, 395–410. https://doi.org/0304-3959(87)90073-X [pii] Grunau, R. V., Johnston, C.C., Craig, K.D., 1990. Neonatal facial and cry responses to invasive and non-invasive procedures. Pain 42, 295–305. https://doi.org/http://dx.doi.org/10.1016/0304-3959(90)91142-6 Gulland, F.M.D., Haulena, M., Dierauf, L.A., 2001. Seals and sea lions, in: Dierauf, L.A.,   160 Gulland, F.M.D. (Eds.), CRC Handbook of Marine Mammal Medicine: Health, Disease, and Rehabilitation. Taylor & Francis, Boca Raton, USA, pp. 907–915. Gutiérrez, J., Simeone, C., Gulland, F., Johnson, S., 2016. Development of retrobulbar and auriculopalpebral nerve blocks in California sea lions (Zalophus Californianus). J. Zoo Wildl. Med. 47, 236–243. https://doi.org/https://doi.org/10.1638/2015-0035.1 Gyger, M., Marler, P., 1988. Food calling in the domestic fowl, Gallus gallus: the role of external referents and deception. Anim. Behav. 36, 358–365. Haak, M., Bos, S., Panic, S., Rothkrantz, L.J.M., 2009. Detecting stress using eye blinks during game playing, in: 10th International Conference on Intelligent Games and Simulation, GAME-ON 2009. pp. 75–82. Hadjistavropoulos, T., Herr, K., Prkachin, K.M., Craig, K.D., Gibson, S.J., Lukas, A., Smith, J.H., 2014. Pain assessment in elderly adults with dementia. Lancet Neurol. 13, 1216–1227. https://doi.org/10.1016/S1474-4422(14)70103-6 Hansen, B.D., Hardie, E.M., Carroll, G.S., 1997. Physiological measurements after ovariohysterectomy dogs: what’s normal? Appl. Anim. Behav. Sci. 5, 101–109. Harrigan, J.A., O’Connell, D.M., 1996. How do you look when feeling anxious? Facial displays of anxiety. Pers. Individ. Dif. 21, 205–212. https://doi.org/10.1016/0191-8869(96)00050-5   161 Hart, B.L., Hart, L.A., Bain, M.J., 1985. Canine and Feline Behavioural Therapy. Lea & Febiger, Philadelphia, USA. Hellebrekers, L.J., Kemme, R.M.F.J., van Wandelen, R.W., 1994. Nalbuphine as a post-operative analgesic in the dog-a comparison with buprenorphine. J. Vet. Anaesth. 21, 4–5. Higgins, J.L., Hendrickson, D.A., 2013. Surgical procedures in pinniped and cetacean species. J. Zoo Wildl. Med. 44, 817–836. https://doi.org/10.1638/2012-0286R1.1 Hjemdahl, P., 1993. Plasma catecholamines—analytical challenges and physiological limitations. Baillieres. Clin. Endocrinol. Metab. 7, 307–353. https://doi.org/10.1016/S0950-351X(05)80179-X Holden, E., Calvo, G., Collins, M., Bell, A., Reid, J., Scott, E.M., Nolan, A.M., 2014. Evaluation of facial expression in acute pain in cats. J. Small Anim. Pract. 55, 615–21. https://doi.org/10.1111/jsap.12283 Holton, L., Reid, J., Scott, E.M., Pawson, P., Nolan, A., 2001. Development of a behaviour-based scale to measure acute pain in dogs. Vet. Rec. 148, 525–531. Holton, L., Scott, E.M., Nolan, A.M., Reid, J., Welsh, E., Flaherty, D., 1998. Comparison of three methods used for assessment of pain in dogs. J. Am. Vet. Med. Assoc. 212, 61–66.   162 Hugonnard, M., Leblond, A., Keroack, S., Cadore, J., Troney, E., 2004. Attitudes and concerns of French veterinarians towards pain and analgesia in dogs and cats. Vet. Anaesth. Analg. 31, 154–163. Hurnik, J.F., Webster, A.B., DeBoer, S., 1985. An investigation of skin temperature differentials in relation to estrus in dairy cattle using a thermal infrared scanning technique. J. Anim. Sci. 61, 1095–1102. International Association for the Study of Pain, 1979. Pain terms: a list with definitions and notes on usage. Pain 6, 249–252. Jankovic, J., Fahn, S., 1980. Physiologic and pathologic tremors: diagnosis, mechanisms and management. Ann. Intern. Med. 93, 460–465. Johnstone, R.A., Grafen, A., 1993. Dishonesty and the handicap principle. Anim. Behav. 46, 759–764. Jourdan, D., Ardid, D., Eschalier, A., 2002. Analysis of ultrasonic vocalisation does not allow chronic pain to be evaluated in rats. Pain 95, 165–173. Kadosh, K., Johnson, M., 2007. Developing a cortex specialized for face perception. Trends Cogn. Sci. 11, 367–369. https://doi.org/10.1016/j.tics.2007.06.007 Kalliokoski, O., Abelson, K.S., Koch, J., Boschian, A., Thormose, S.F., Fauerby, N.,   163 Rasmussen, R.S., Johansen, F.F., Hau, J., 2010. The effect of voluntarily ingested buprenorphine on rats subjected to surgically induced global cerebral ischaemia. In Vivo (Brooklyn). 24, 641–646. Kastelic, J.P., Cook, R.B., Coulter, G.H., Saacke, R.G., 1996. Insulating the scrotal neck affects semen quality and scrotal/testicular temperatures in the bull. Theriogenology 45, 935–942. Keating, S.C.J., Thomas, A.A., Flecknell, P.A., Leach, M.C., 2012. Evaluation of EMLA cream for preventing pain during tattooing of rabbits: changes in physiological, behavioural and facial expression responses. PLoS One 7, e44437. https://doi.org/10.1371/journal.pone.0044437 Keay, J.M., Singh, J., Gaunt, M.C., Kaur, T., 2006. Fecal glucocorticoids and their metabolites as indicators of stress in various mammalian species: a literature review. J. Zoo Wildl. Med. 37, 234–244. https://doi.org/10.1638/05-050.1 Kent, J.E., Goodall, J., 1991. Assessment of an immunoturbidimetric method for measuring equine serum haptoglobin concentration. Equine Vet. J. 23, 59–66. Kent, J.E., Molony, V., Robertson, I.S., 1993. Changes in plasma cortisol concentration in lambs of three ages after three methods of castration and tail docking. Res. Vet. Sci. 55, 246–251. https://doi.org/10.1016/0034-5288(93)90088-W   164 Khan, C.B., Markowitz, H., McCowan, B., 2006. Vocal development in captive harbor seal pups (Phoca vitulina richardii): Age, sex, and individual differences. J. Acoust. Soc. Am. 120, 1684–1694. https://doi.org/10.1121/1.2226530 Khan, M.M., Ward, R.D., Ingleby, M., 2009. Classifying pretended and evoked facial expressions of positive and negative affective states using infrared measurement of skin temperature. ACM Trans. Appl. Percept. 6, 1–22. https://doi.org/10.1145/1462055.1462061 Kiley, M., 1972. The vocalizations of ungulates, their causation and function. Z. Tierpsychol. 31, 171–222. Kim, E.J., Kim, E.S., Covey, E., Kim, J.J., 2010. Social transmission of fear in rats: the role of 22-kHz ultrasonic distress vocalization. PLoS One 5, e15077. Kirschbaum, C., Hellhammer, D.H., 1999. Noise and stress - salivary cortisol as a non-invasive measure of allostatic load. Noise Heal. 4, 57–65. Kleiber, M., 1975. The Fire of Life. R.E. Krieger Publishing Company, Huntington, NY. Kunz, M., Mylius, V., Schepelmann, K., Lautenbacher, S., 2008. Impact of age on the facial expression of pain. J. Psychosom. Res. 64, 311–318. https://doi.org/10.1016/j.jpsychores.2007.09.010   165 Kuznetsova, A., Brockhoff, P.B., Christensen, R.H.B., 2017. LmerTest: tests in linear mixed effects models. J. Stat. Softw. 82. https://doi.org/10.18637/jss.v082.i13 Lamont, L.A., Mathews, K.A., 2007. Opioids, nonsteroidal anti-inflammatories, and analgesic adjuvants, in: Tranquilli, W.J., Thurmon, J.C., Grimm, K.A. (Eds.), Lumb and Jones’ Veterinary Anesthesia and Analgesia. Blackwell Publishing, Ames, Iowa, pp. 241–273. Landys, M.M.M., Ramenofsky, M., Wingfield, J.C.C., 2006. Actions of glucocorticoids at a seasonal baseline as compared to stress-related levels in the regulation of periodic life processes. Gen. Comp. Endocrinol. 148, 132–149. Langford, D.J., Bailey, A.L., Chanda, M.L., Clarke, S.E., Drummond, T.E., Echols, S., Glick, S., Ingrao, J., Klassen-Ross, T., LaCroix-Fralish, M.L., Matsumiya, L., Sorge, R.E., Sotocinal, S.G., Tabaka, J.M., Wong, D., van den Maagdenberg, A.M., Ferrari, M.D., Craig, K.D., Mogil, J.S., 2010. Coding of facial expressions of pain in the laboratory mouse. Nat. Methods 7, 447–449. https://doi.org/10.1038/nmeth.1455 Langford, D.J.L., Crager, S.E., Shehzad, Z., Smith, S.B., Sotocinal, S.G., Levenstadt, J.S., Chanda, M.L., Levitin, D.J., Mogil, J.S., 2006. Social modulation of pain as evidence for empathy in mice. Science (80-. ). 312, 1967–1970. https://doi.org/10.1126/science.1128322   166 Lapierre, J.L., Schreer, J.F., Burns, J.M., Hammill, M.O., 2007. Effect of diazepam on heart and respiratory rates of harbor seal pups following intravenous injection. Mar. Mammal Sci. 23, 209–217. https://doi.org/10.1111/j.1748-7692.2006.00097.x Larochette, A.-C., Chambers, C.T., Craig, K.D., 2006. Genuine, suppressed and faked facial expressions of pain in children. Pain 126, 64–71. https://doi.org/10.1016/j.pain.2006.06.013 Lawson, J.W., Renouf, D., 1985. Parturition in the Atlantic harbor seal Phoca vitulina concolor. J. Mammal. 66, 395–398. Lay, D.C., Friend, T.H., Bowers, C.L., Grissom, K.K., Jenkins, O.C., 1992a. A comparative physiological and behavioral study of freeze and hot-iron branding using dairy cows. J. Anim. Sci. 70, 1121–1125. Lay, D.C., Friend, T.H., Grissom, K.K., Bowers, C.L., Mal, M.E., 1992b. Effects of freeze or hot-iron branding of Angus calves on some physiological and behavioral indicators of stress. Appl. Anim. Behav. Sci. 33, 137–147. https://doi.org/10.1016/S0168-1591(05)80003-6 Le Bars, D., Cadden, S.W., 2005. Pain, in: Whishaw, I.Q., Kolb, B. (Eds.), The Behavior of the Laboratory Rat. Oxford University Press, New York, pp. 69–80. Leach, M.C., Coulter, C.A., Richardson, C.A., Flecknell, P.A., 2011. Are we looking in   167 the wrong place? Implications for behavioural-based pain assessment in rabbits (Oryctolagus cuniculi) and beyond? PLoS One 6, e13347. https://doi.org/10.1371/journal.pone.0013347 Leach, M.C., Klaus, K., Miller, A.L., Scotto di Perrotolo, M., Sotocinal, S.G., Flecknell, P.A., 2012. The assessment of post-vasectomy pain in mice using behaviour and the mouse grimace scale. PLoS One 7, e35656. https://doi.org/10.1371/journal.pone.0035656 Lecorps, B., Rödel, H.G., Féron, C., 2016. Assessment of anxiety in open field and elevated plus maze using infrared thermography. Physiol. Behav. 157, 209–216. https://doi.org/10.1016/j.physbeh.2016.02.014 LeResche, L., Dworkin, S.F., Wilson, L., Ehrlich, K.J., 1992. Effect of temporomandibular disorder pain duration on facial expressions and verbal report of pain. Pain 51, 289–295. https://doi.org/10.1016/0304-3959(92)90212-T Leslie, E., Hernández-Jover, M., Newman, R., Holyoake, P., 2010. Assessment of acute pain experienced by piglets from ear tagging, ear notching and intraperitoneal injectable transponders. Appl. Anim. Behav. Sci. 127, 86–95. https://doi.org/10.1016/j.applanim.2010.09.006 Lester, S.J., Mellor, D.J., Holmes, R.J., Ward, R.N., Stafford, K.J., 1996. Behavioural and cortisol responses of lambs to castration and tailing using different methods. N.   168 Z. Vet. J. 44, 45–54. Leung, V., Zhang, E., Pang, D.S.J., 2016. Real-time application of the Rat Grimace Scale as a welfare refinement in laboratory rats. Sci. Rep. 6, 1–12. https://doi.org/10.1038/srep31667 Levine, J.A., Pavlidis, I., Cooper, M., 2001. The face of fear. Lancet 357, 1757. Ley, S.J., Waterman, A.E., Livingston, A., Parkinson, T.J., 1994. Effect of chronic pain associated with lameness on plasma cortisol concentrations in sheep: a field study. Res. Vet. Sci. 57, 332–335. https://doi.org/10.1016/0034-5288(94)90126-0 Livingston, A., Chambers, P., 2000. The physiology of pain, in: Flecknell, P.A., Waterman-Pearson, A. (Eds.), Pain Management in Animals. W.B. Saunders, London, UK, pp. 9–19. https://doi.org/https://doi.org/10.1016/B978-0-7020-1767-4.50005-9 Llamas Moya, S., Boyle, L.A., Lynch, P.B., Arkins, S., 2008. Effect of surgical castration on the behavioural and acute phase responses of 5-day-old piglets. Appl. Anim. Behav. Sci. 111, 133–145. https://doi.org/10.1016/j.applanim.2007.05.019 Lowe, T.E., Cook, C.J., Ingram, J.R., Harris, P.J., 2005. Changes in ear-pinna temperature as a useful measure of stress in sheep (Ovis aries). Anim. Welf. 14, 35–42.   169 Ludwig, N., Gargano, M., Luzi, F., Carenzi, C., Verga, M., 2007. Technical note: applicability of infrared thermography as a non-invasive measurement of stress in rabbits. World Rabbit Sci. 15, 199–205. MacRae, A.M., Haulena, M., Fraser, D., 2011. The effect of diet and feeding level on survival and weight gain of hand-raised harbor seal pups (Phoca vitulina). Zoo Biol. 30, 532–541. https://doi.org/10.1002/zoo.20356 MacRae, A.M., Makowska, I.J., Fraser, D., 2018. Initial evaluation of facial expressions and behaviours of harbour seal pups (Phoca vitulina) in response to tagging and microchipping. Appl. Anim. Behav. Sci. 205, 167–174. https://doi.org/10.1016/j.applanim.2018.05.001 Manteuffel, G., Puppe, B., Schön, P.C., 2004. Vocalization of farm animals as a measure of welfare. Appl. Anim. Behav. Sci. 88, 163–182. https://doi.org/10.1016/j.applanim.2004.02.012 Marches, S., Moioli, M., Di Giancamillo, M., Luzi, F., 2013. Veterinary diagnostic imaging, in: Luzi, F., Mitchell, M., Nanni Costa, L., Redaelli, V. (Eds.), Thermography: Current Status and Advances in Livestock Animals and in Veterinary Medicine. The Foundation for the Promotion of Zooprophylatics and Zootechnology of Brescia, Milan, pp. 127–146. Martin, P., Bateson, P., 2007. Measuring Behaviour: An Introductory Guide, 3rd ed.   170 Cambridge University Press, New York, USA. Matsumiya, L.C., Sorge, R.E., Sotocinal, S.G., Tabaka, J.M., Wieskopf, J.S., Zaloum, A., King, O.D., Mogil, J.S., 2012. Using the mouse grimace scale to reevaluate the efficacy of postoperative analgesics in laboratory mice. J. Am. Assoc. Lab. Anim. Sci. 51, 42–49. Maynard Smith, J., 1994. Must reliable signals always be costly? Anim. Behav. 47, 1115–1120. McCafferty, D.J., 2007. The value of infrared thermography for research on mammals: previous applications and future directions. Mamm. Rev. 37, 207–223. https://doi.org/10.1111/j.1365-2907.2007.00111.x McCarthy, R.N., 1993. Preliminary studies on the use of plasma beta endorphin in horses as an indicator of stress and pain. J. Equine Vet. Sci. 13, 216–219. Mccomb, K., Taylor, A.M., Wilson, C., Charlton, B.D., 2009. The cry embedded within the purr. Curr. Biol. 19, R507–R508. https://doi.org/doi:10.1016/j.cub.2009.05.033 McGeown, D., Danbury, T.C., Waterman-Pearson, A.E., Kestin, S.C., 1999. Effect of carprofen on lamesness of broiler chickens. Vet. Rec. 144, 668–671. McGrath, P.A., 1987. An assessment of children’s pain: a review of behavioral,   171 physiological and direct scaling techniques. Pain 31, 147–176. McGreevy, P., Warren-Smith, A., Guisard, Y., 2012. The effect of double bridles and jaw-clamping crank nosebands on temperature of eyes and facial skin of horses. J. Vet. Behav. Clin. Appl. Res. 7, 142–148. https://doi.org/10.1016/j.jveb.2011.08.001 McLennan, K.M., Rebelo, C.J.B., Corke, M.J., Holmes, M.A., Leach, M.C., Constantino-Casas, F., 2016. Development of a facial expression scale using footrot and mastitis as models of pain in sheep. Appl. Anim. Behav. Sci. 176, 19–26. https://doi.org/10.1016/j.applanim.2016.01.007 Mellor, D.J., Cook, C.J., Stafford, K.J., 2000. Quantifying some responses to pain as a stressor, in: Moberg, G.P., Mench, A.J. (Eds.), The Biology of Animal Stress: Basic Principles and Implications for Animal Welfare. CABI International, Wallingford, UK, pp. 171–198. Mellor, D.J., Molony, V., Robertson, I.S., 1991. Effects of castration on behavior and plasma cortisol concentrations in young lambs, kids and calves. Res. Vet. Sci. 51, 49–154. Mellor, D.J., Murray, L., 1989. Effects of tail docking and castration on behaviour and plasma cortisol concentrations in young lambs. Res. Vet. Sci. 46, 387–391. Mellor, D.J., Stafford, K.J., Todd, S.E., Lowe, T.E., Gregory, N.G., Bruce, R.A., Ward,   172 R.N., 2002. A comparison of catecholamine and cortisol responses of young lambs and calves to painful husbandry procedures. Aust. Vet. J. 80, 228–233. https://doi.org/10.1111/j.1751-0813.2002.tb10820.x Miller, A.L., Leach, M.C., 2015. The mouse grimace scale: A clinically useful tool? PLoS One 10, e0136000. https://doi.org/10.1371/journal.pone.0136000 Miller, E.H., 1975. A comparative study of facial expressions of two species of pinnipeds. Behaviour 53, 268–284. Moberg, G.P., 2000. Biological responses to stress: implications for animal welfare, in: Moberg, G.P., Mench, J.A. (Eds.), The Biology of Animal Stress: Basic Principles and Implications for Animal Welfare. CABI Publishing, Wallingford, UK, pp. 1–21. Moe, R.O., Stubsjøen, S.M., Bohlin, J., Flø, A., Bakken, M., 2012. Peripheral temperature drop in response to anticipation and consumption of a signaled palatable reward in laying hens (Gallus domesticus). Physiol. Behav. 106, 527–533. https://doi.org/10.1016/j.physbeh.2012.03.032 Mogil, J.S., 2015. Social modulation of and by pain in humans and rodents. Pain 156, S33–S41. https://doi.org/10.1097/01.j.pain.0000460341.62094.77 Mogil, J.S., 2009. Animal models of pain: progress and challenges. Nat. Rev. Neurosci. 10, 283–294.   173 Mogil, J.S., Davis, K.D., Derbyshire, S.W., 2010. The necessity of animal models in pain research. Pain 151, 12–17. https://doi.org/10.1016/j.pain.2010.07.015 Molony, V., Kent, J.E., 1997. Assessment of acute pain in farm animals using behavioral and physiological measurements. J. Anim. Sci. 75, 266–272. Molony, V., Kent, J.E., Hosie, B., Graham, M.J., 1997. Reduction in pain suffered by lambs at castration. Br. Vet. J. 153, 205–213. Molony, V., Kent, J.E., Robertson, I.S., 1995. Assessment of acute and chronic pain after different methods of castration of calves. Appl. Anim. Behav. Sci. 46, 33–48. Molter, C.M., Barbosa, L., Johnson, S., Knych, H.K., Chinnadurai, S.K., Wack, R.F., 2015. Pharmacokinetics of a single subcutaneous dose of sustained release buprenorphine in Northern elephant seals (Mirounga angustirostris). J. Zoo Wildl. Med. 46, 52–61. https://doi.org/10.1638/2014-0115R.1 Morisse, J.P., Cotte, J.P., Huonnic, D., 1995. Effect of dehorning on behaviour and plasma cortisol responses in young calves. Appl. Anim. Behav. Sci. 43, 239–247. https://doi.org/10.1016/0168-1591(95)00569-E Morton, D.B., Griffiths, P.H.M., 1985. Guidelines on the recognition of pain, distress and discomfort in experimental animals and hypothesis for assessment. Vet. Rec. 116, 431–436.   174 Moya, S.L., Boyle, L.A., Lynch, P.B., Arkins, S., 2006. Influence of teeth resection on the skin temperature and acute phase response in newborn piglets. Anim. Welf. 15, 291–297. Mrozek, M., Fischer, R., Trendelenburg, M., Zillmann, U., 1995. Microchip implant system used for animal identification in laboratory rabbits, guinea pigs, woodchucks and in amphibians. Lab. Anim. 29, 339–344. Nakanishi, R., Imai-Matsumura, K., 2008. Facial skin temperature decreases in infants with joyful expression. Infant Behav. Dev. 31, 137–144. https://doi.org/10.1016/j.infbeh.2007.09.001 Nakayama, K., Goto, S., Kuraoka, K., Nakamura, K., 2005. Decrease in nasal temperature of rhesus monkeys (Macaca mulatta) in negative emotional state. Physiol. Behav. 847, 83–790. Narayan, M.C., 2010. Culture’s effects on pain assessment and management. AJN Am. J. Nurs. 110, 38–47. National Research Council, 2009. Recognition and Alleviation of Pain in Laboratory Animals. The National Academies Press, Washington, DC. https://doi.org/https://doi.org/10.17226/12526 National Research Council, 1992. Recognition and assessment of pain, stress, and   175 distress, in: Recognition and Alleviation of Pain and Distress in Laboratory Animals. The National Academies Press, Washington, DC, pp. 32–52. https://doi.org/10.17226/1542 Neave, H.W., Daros, R.R., Costa, J.H.C., von Keyserlingk, M.A.G., Weary, D.M., 2013. Pain and pessimism: Dairy calves exhibit negative judgement bias following hot-iron disbudding. PLoS One 9, e96135. https://doi.org/10.1371/journal.pone.0096135 Newton, D.J., McLeod, G.A., Khan, F., Belch, J.J.F., 2007. Mechanisms influencing the vasoactive effects of lidocaine in human skin. Anaesthesia 62, 146–150. https://doi.org/10.1111/j.1365-2044.2006.04901.x Nikkhah, A., Plaizier, J.C., Einarson, M.S., Berry, R.J., Scott, S.L., Kennedy, A.D., 2005. Infrared thermography and visual examination of hooves of dairy cows in two stages of lactation. J. Dairy Sci. 2749–2753. Noonan, G.J., Rand, J.S., Blackshaw, J.K., Priest, J., 1996. Behavioural observations of puppies undergoing tail docking. Appl. Anim. Behav. Sci. 49, 335–342. https://doi.org/10.1016/0168-1591(96)01062-3 Noonan, G.J., Rand, J.S., Priest, J., Ainscow, J., Blackshaw, J.K., 1994. Behavioural observations of piglets undergoing tail docking, teeth clipping and ear notching. Appl. Anim. Behav. Sci. 39, 203–213. https://doi.org/10.1016/0168-1591(94)90156-  176 2 Nutter, F.B., Haulena, M., Bai, S.A., 1998. Preliminary pharmacokinetics of single-dose intramuscular butorphanol in elephant seals (Mirounga angustirostris), in: Proceedings of the American Association of Zoo Veterinarians. pp. 372–373. Olivier, B., Milewijk, E., van Oorschot, R., van der Poel, G., Zethof, T., van der Heyden, J., Mos, J., 1994. New animal models of anxiety. Eur. Neuropsychopharmacol. 4, 93–102. Olsson, A., Phelps, E.A., 2007. Social learning of fear. Nat. Neurosci. 10, 1095–1102. https://doi.org/10.1038/nn1968 Ottosson, U., Bäckman, J., Smith, H.G., 1997. Begging affects parental effort in the pied flycatcher, Ficedula hypoleuca. Behav. Ecol. Sociobiol. 41, 381–384. Paine, P., Kishor, J., Worthen, S.F., Gregory, L.J., Aziz, Q., 2009. Exploring relationships for visceral and somatic pain with autonomic control and personality. Pain 144, 236–244. https://doi.org/10.1016/j.pain.2009.02.022 Parr, L.A., Waller, B.M., Burrows, A.M., Gothard, K.M., Vick, S.J., 2010. MaqFACS: A muscle-based facial movement coding system for the rhesus macaque. Am. J. Phys. Anthropol. 143, 625–630. https://doi.org/10.1002/ajpa.21401   177 Paterson, W., Pomeroy, P.P., Sparling, C.E., Moss, S., Thompson, D., Currie, J.I., Mccafferty, D.J., 2011. Assessment of flipper tag site healing in gray seal pups using thermography. Mar. Mammal Sci. 27, 295–305. https://doi.org/10.1111/j.1748-7692.2010.00400.x Pavlidis, I., Eberhardt, N.L., Levine, J.A., 2002. Human behaviour: Seeing through the face of deception. Nature 415, 35. Pavlidis, I., Levine, J.A., 2002. Thermal image analysis for polygraph testing. IEEE Eng. Med. Biol. 21, 56–64. Peers, A., Mellor, D.J., Wintour, E.M., Dodic, M., 2002. Blood pressure, heart rate, hormonal and other acute responses to rubber-ring castration and tail docking in lambs. N. Z. Vet. J. 50, 56–62. Perez de Diego, A.C., Sanchez-Cordon, P.J., Pedrera, M., Martinez-Lopez, B., Gomez-Villamanlos, J.C., Sanchez-Vozcaino, J.M., 2013. The use of infrared thermography as a non-invasive method for fever detection in sheep infected with bluetongue virus. Vet. J. 198, 182–186. Perry, E.A., Renouf, D., 1988. Further studies of the role of harbour seal (Phoca vitulina) pup vocalizations in preventing separation of mother-pup pairs. Can. J. Zool. 66, 934–938.   178 Peters, J.W.B., Koot, H.M., Grunau, R.E., de Boer, J., van Druenen, M.J., Tibboel, D., Duivenvoorden, H.J., 2003. Neonatal Facial Coding System for assessing postoperative pain in infants: item reduction is valid and feasible. Clin. J. Pain 19, 353–363. https://doi.org/10.1097/00002508-200311000-00003 Petrie, N.J., Mellor, D.J., Stafford, K.J., Bruce, R.A., Ward, R.N., 1996. Cortisol response of calves to two methods of disbudding used with or without local anaesthetic. N. Z. Vet. J. 44, 9–14. Phillips, M.T., 1993. Savages, drunks, and lab animals: The researcher’s perception of pain. Soc. Anim. 1, 61–81. https://doi.org/doi:10.1163/156853093X00154 Pincus, T., Morley, S., 2001. Cognitive-processing bias in chronic pain: A review and integration. Psychol. Bull. 127, 599–617. https://doi.org/10.1037//0033-2909.127.5.599 Pincus, T., Pearce, S., McClelland, A., Farley, S., Vogel, S., 1994. Interpretation bias in responses to ambiguous cues in pain patients. J. Psychosom. Res. 38, 343–353. Pinel, J.P., Mana, M.J., 1989. Adaptive interactions of rats with dangerous inanimate objects: Support for a cognitive theory of defensive behavior, in: Blanchard, R.J., Brain, P.F., Blanchard, D.C., Parmigiani, S. (Eds.), Ethoexperimental Approaches to the Study of Behavior. Kluwer Academic/Plenum Publishers, New York, USA, pp. 137–150.   179 Polat, B., Colak, A., Cengiz, M., Yanmaz, L.E., Oral, H., Bastan, A., Kaya, S., Hayirli, A., 2010. Sensitivity and specificity of infrared thermography in detection of subclinical mastitis in dairy cows. J. Dairy Sci. 93, 3525–3532. https://doi.org/10.3168/jds.2009-2807 Poole, G., Craig, K.D., 1992. Judgments of genuine, suppressed, and faked facial expressions of pain. J. Pers. Soc. Psychol. 63, 797–805. Porter, F.L., Miller, R.H., Marshall, R.E., 1986. Neonatal pain cries: effect of circumcision on acoustic features and perceived urgency. Child Dev. 57, 790–802. Porter, F.L., Porges, S.W., Marshall, R.E., 1988. Newborn pain cries and vagal tone: parallel changes in response to circumcision. Child Dev. 59, 495–505. Price, J., Catriona, S., Welsh, E.M., Waran, N.K., 2003. Preliminary evaluation of a behaviour-based system for assessment of post-operative pain in horses following arthroscopic surgery. Vet. Anaesth. Analg. 30, 124–137. Price, J., Eager, R.A., Welsh, E.M., Waran, N.K., 2005. Current practice relating to equine castration in the UK. Res. Vet. Sci. 78, 277–280. https://doi.org/10.1016/j.rvsc.2004.09.009 Price, J., Marques, J.M.S., Welsh, E.M., Waran, N.K., 2002. Attitudes towards pain in horses – a pilot epidemiological survey. Vet. Rec. 151, 570–575.   180 Pritchett, L.C., Ulibarri, C., Roberts, M.C., Schneider, R.K., Sellon, D.C., 2003. Identification of potential physiological and behavioral indicators of postoperative pain in horses after exploratory celiotomy for colic. Appl. Anim. Behav. Sci. 80, 31–43. Prkachin, K.M., 1992. The consistency of facial expressions of pain: A comparison across modalities. Pain 51, 297–306. https://doi.org/10.1016/0304-3959(92)90213-U Prkachin, K.M., Mercer, S.R., 1989. Pain expression in patients with shoulder pathology: validity, properties and relationship to sickness impact. Pain 39, 257–265. https://doi.org/10.1016/0304-3959(89)90038-9 R Core Team, 2015. R: a language and environment for statistical computing. Rainwater-Lovett, K., Pacheco, J.M., Packer, C., Rodriguez, L.L., 2009. Detection of foot-and-mouth disease virus infected cattle using infrared thermography. Vet. J. 180, 317–324. https://doi.org/10.1016/j.tvjl.2008.01.003 Ranger, M., Johnston, C.C., Anand, K.J.S., 2007. Current controversies regarding pain assessment in neonates. Semin. Perinatol. 31, 283–288. Rault, J.L., Lay, D.C., Marchant-Forde, J.N., 2011. Castration induced pain in pigs and other livestock. Appl. Anim. Behav. Sci. 135, 214–225.   181 Redaelli, V., Bergero, D., Zucca, E., Ferrucci, F., Nanni, L., Das, C., Crosta, L., Luzi, F., 2014. Use of thermography techniques in equines: Principles and applications. J. Equine Vet. Sci. 34, 345–350. https://doi.org/10.1016/j.jevs.2013.07.007 Redondo, T., Castro, F., 1992. Signalling of nutritional need by magpie nestlings. Ethology 92, 193–204. Reijgwart, M.L., Schoemaker, N.J., Pascuzzo, R., Leach, M.C., Stodel, M., De Nies, L., Hendriksen, C.F.M., Van Der Meer, M., Vinke, C.M., Van Zeeland, Y.R.A., 2017. The composition and initial evaluation of a grimace scale in ferrets after surgical implantation of a telemetry probe. PLoS One 12, e0187986. https://doi.org/10.1371/journal.pone.0187986 Reijnders, P.J.H., Brasseur, S.M.J.M., Borchardt, T., Camphuysen, K., Czeck, R., Gilles, A., Jensen, L.F., Leopold, M., Lucke, K., Ramdohr, S., Scheidat, M., Siebert, U., Teilmann, J., 2009. Marine mammals. Thematic Report No. 20, in: Marencic, H., Vlas, J. de (Eds.), Quality Status Report 2009. Wadden Sea Ecosystem No. 25. Common Wadden Sea Secretariat, Trilateral Monitoring and Assessment Group, Wilhelmshaven, Germany. Renouf, D., 1984. The vocalization of the harbour seal pup (Phoca vitulina) and its role in the maintenance of contact with the mother. J. Zool. 202, 583–590. Richardson, C.A., Flecknell, P.A., 2005. Anaesthesia and post-operative analgesia   182 following experimental surgery in laboratory rodents: are we making progress? ATLA 33, 119–127. Rietmann, T.R., Stauffacher, M., Bernasconi, P., Auer, J.. A., Weishaupt, M.A., 2004. The association between heart rate, heart rate variability, endocrine and behavioural pain measures in horses suffering from laminitis. J. Vet. Med. A. Physiol. Pathol. Clin. Med. 51, 218–225. https://doi.org/10.1111/j.1439-0442.2004.00627.x Rose, J.D., Arlinghaus, R., Cooke, S.J., Diggles, B.K., Sawynok, W., Stevens, E.D., Wynne, C.D.L., 2014. Can fish really feel pain? Fish Fish. 15, 97–133. Roughan, J. V., Flecknell, P.A., 2003. Evaluation of a short duration behaviour-based post-operative pain scoring system in rats. Eur. J. Pain 7, 397–406. https://doi.org/10.1016/S1090-3801(02)00140-4 Roughan, J. V., Flecknell, P.A., 2000. Effects of surgery and analgesic administration on spontaneous behaviour in singly housed rats. Res. Vet. Sci. 69, 283–288. Rushen, J., Congdon, P., 1986. Relative aversion of sheep to simulated shearing with and without electro-immobilisation. Aust. J. Exp. Agric. 26, 535–537. Rutherford, K.M.D., 2002. Assessing pain in animals. Anim. Welf. 11, 31–53. https://doi.org/10.1016/B978-012370880-9.00209-7   183 Sapolsky, R.M., 2002. Endocrinology of the stress-response, in: Becker, J.B., Breedlove, S.M., Crews, D., and McCarthy, M.M. (Ed.), Behavioural Endocrinology. MIT Press, London, UK, pp. 409–450. Sapolsky, R.M., Romero, L.M., Munck, A.U., 2000. How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocr. Rev. 21, 55–89. https://doi.org/10.1210/er.21.1.55 Sauvé, C.C., Beauplet, G., Hammill, M.O., Charrier, I., 2015a. Mother-pup vocal recognition in harbour seals: Influence of maternal behaviour, pup voice and habitat sound properties. Anim. Behav. 105, 109–120. https://doi.org/10.1016/j.anbehav.2015.04.011 Sauvé, C.C., Beauplet, G., Hammill, M.O., Charrier, I., 2015b. Acoustic analysis of airborne, underwater, and amphibious mother attraction calls by wild harbor seal pups (Phoca vitulina). J. Mammal. 96, 591–602. https://doi.org/10.1093/jmammal/gyv064 Schaefer, A.L., Cook, N., Tessaro, S.V., Deregt, D., Desroches, G., Dubeski, P.L., Tong, A.K.W., Godson, D.L., 2004. Early detection and prediction of infection using infrared thermography. Can. J. Anim. Sci. 84, 73–80. Schaefer, A.L., Cook, N.J., Bench, C., Chabot, J.B., Colyn, J., Liu, T., Okine, E.K., Stewart, M., Webster, J.R., 2012. The non-invasive and automated detection of   184 bovine respiratory disease onset in receiver calves using infrared thermography. Res. Vet. Sci. 93, 928–935. https://doi.org/10.1016/j.rvsc.2011.09.021 Schaefer, A.L., Jones, S.D.M., Murray, A.C., Sather, A.P., Tong, A.K.W., 1989. Infrared thermography of pigs with known genotypes for stress susceptibility in relation to pork quality. Can. J. Anim. Sci. 69, 491–495. Schaefer, A.L., Jones, S.D.M., Tong, A.K.W., Vincent, B.C., 1988. The effects of fasting and transportation on beef cattle. 1. Acid-base-electrolyte balance and infrared heat loss of beef cattle. Livest. Prod. Sci. 20, 15–24. Schaefer, A.L., Cook, N.J., Church, J.S., Basarab, J., Perry, B.J., Miller, C., Tong, A.K.W., 2007. The use of infrared thermography as an early indicator of bovine respiratory disease complex in calves. Res. Vet. Sci. 83, 376–384. Scheumann, M., Hasting, A.S., Kotz, S.A., Zimmermann, E., 2014. The voice of emotion across species: how do human listeners recognize animals’ affective states? PLoS One 9, e91192. https://doi.org/10.1371/journal.pone. 0091192 Schusterman, R.J., Van Parijs, S.M., 2003. Pinniped vocal communication: an introduction. Aquat. Mamm. 29, 177–180. https://doi.org/10.1578/016754203101024103 Schwartzkopf-Genswein, K.S., Stookey, J.M., Crowe, T.G., Genswein, B.M.A., 1998.   185 Comparison of image analysis, exertion force, and behavior measurements for use in the assessment of beef cattle responses to hot-iron and freeze branding. J. Anim. Sci. 76, 972–979. Schwartzkopf-Genswein, K.S., Stookey, J.M., Welford, R., 1997. Behavior of cattle during hot-iron and freeze branding and the effects on subsequent handling ease. J. Anim. Sci. 75, 2064–2072. Schweinhardt, P., Glynn, C., Brooks, J., McQuay, H., Jack, T., Chessell, I., Bountra, C., Tracey, I., 2006. An fMRI study of cerebral processing of brush-evoked allodynia in neuropathic pain patients. Neuroimage 32, 256–265. Seyfarth, R.M., Cheney, D.L., 2003. Meaning and emotion in animal vocalizations. Ann. N. Y. Acad. Sci. 1000, 32–55. https://doi.org/10.1196/annals.1280.004 Seyfarth, R.M., Cheney, D.L., Marler, P., 1980. Monkey responses to three different alarm calls: Evidence of predator classification and semantic communication. Science  210, 801–803. https://doi.org/10.1126/science.7433999 Sgoifo, A., de Boer, S.F., Westenbroek, C., Maes, F.W., Beldhuis, H., Suzuki, T., Koolhaas, J.M., 1997. Incidence of arrhythmias and heart rate variability in wild-type rats exposed to social stress. Am. J. Physiol. 273, H1754–1760. Sgoifo, A., Koolhaas, J.M., Musso, E., De Boer, S.F., 1999. Different sympatho-vagal   186 modulation of heart rate during social and nonsocial stress episodes in wild-type rats. Physiol. Behav. 67, 733–738. Sheriff, M.J., Dantzer, B., Delehanty, B., Palme, R., Boonstra, R., 2011. Measuring stress in wildlife: techniques for quantifying glucocorticoids. Oecologia 166, 869–887. Simon, D., Craig, K.D., Miltner, W.H., Rainville, P., 2006. Brain responses to dynamic facial expressions of pain. Pain 126, 309–318. Sneddon, L.U., Braithwaite, V.A., Gentle, M.J., 2003. Novel object test: Examining nociception and fear in the rainbow trout. J. Pain 4, 431–440. https://doi.org/10.1067/S1526-5900(03)00717-X Sneddon, L.U., Elwood, R.W., Adamo, S.A., Leach, M.C., 2014. Defining and assessing animal pain. Anim. Behav. 97–133, 201–212. https://doi.org/10.1016/j.anbehav.2014.09.007 Sotocinal, S.G., Sorge, R.E., Zaloum, A., Tuttle, A.H., Martin, L.J., Wieskopf, J.S., Mapplebeck, J.C.S., Wei, P., Zhan, S., Zhang, S., McDougall, J.J., King, O.D., Mogil, J.S., 2011. The rat grimace scale: A partially automated method for quantifying pain in the laboratory rat via facial expressions. Mol. Pain 7, 55-65. https://doi.org/10.1186/1744-8069-7-55   187 Spafford, P.A., Von Baeyer, C.L., Hicks, C.L., 2002. Expected and reported pain in children undergoing ear piercing: A randomized trial of preparation by parents. Behav. Res. Ther. 40, 253–266. https://doi.org/10.1016/S0005-7967(01)00008-0 Sprecher, D.J., Hostetler, D.E., Kaneene, J.B., 1997. A lameness scoring system that uses posture and gait to predict dairy cattle reproductive performance. Theriogenology 47, 1179–1187. https://doi.org/10.1016/S0093-691X(97)00098-8 Stafford, K.J., Mellor, D.J., 2011. Addressing the pain associated with disbudding and dehorning in cattle. Appl. Anim. Behav. Sci. 135, 226–231. https://doi.org/10.1016/j.applanim.2011.10.018 Stafford, K.J., Mellor, D.J., Todd, S.E., Bruce, R.A., Ward, R.N., 2002. Effects of local anaesthesia or local anaesthesia plus a non-steroidal anti-inflammatory drug on the acute cortisol response of calves to five different methods of castration. Res. Vet. Sci. 73, 61–70. https://doi.org/10.1016/S0034-5288(02)00045-0 Stanway, G.W., Taylor, P.M., Broadbelt, D.C., 2002. A preliminary investigation comparing pre-operative morphine and buprenorphine for postoperative analgesia and sedation in cats. Vet. Anaesth. Analg. 29, 29–35. https://doi.org/10.1046/j.1467-2987.2001.00062.x Stevens, B., Johnston, C., Taddio, A., Gibbins, S., Yamada, J., 2010. The premature infant pain profile: evaluation 13 years after development. Clin. J. Pain 26, 813–  188 830. https://doi.org/10.1097/AJP.0b013e3181ed1070 Stewart, M., 2008. Non-invasive measurement of stress and pain in cattle using infrared thermography. Massey University, NZ. Stewart, M., Schaefer, A.L., Haley, D.B., Colyn, J., Cook, N.J., Stafford, K.J., Webster, J.R., 2008a. Infrared thermography as a non-invasive method for detecting fear-related responses of cattle to handling procedures. Anim. Welf. 17, 387–393. Stewart, M., Schaefer, A.L., Haley, D.B., J.J., C., Cook, N.J., Stafford, K.J., Webster, J.R., 2007a. Infrared thermography as a non-invasive method for detecting fear-related responses of cattle to different handling procedures. Anim. Welf. 17, 387–393. Stewart, M., Stafford, K.J., Dowling, S.K., Schaefer, A.L., Webster, J.R., 2008b. Eye temperature and heart rate variability of calves disbudded with or without local anaesthetic. Physiol. Behav. 93, 789–797. https://doi.org/10.1016/j.physbeh.2007.11.044 Stewart, M., Stookey, J.M., Stafford, K.J., Tucker, C.B., Rogers, A.R., Dowling, S.K., Verkerk, G.A., Schaefer, A.L., Webster, J.R., 2009. Effects of local anesthetic and a nonsteroidal antiinflammatory drug on pain responses of dairy calves to hot-iron dehorning. J. Dairy Sci. 92, 1512–1519. https://doi.org/10.3168/jds.2008-1578   189 Stewart, M., Verkerk, G.A., Stafford, K.J., Schaefer, A.L., Webster, J.R., 2010. Noninvasive assessment of autonomic activity for evaluation of pain in calves, using surgical castration as a model. J. Dairy Sci. 93, 3602–3609. https://doi.org/10.3168/jds.2010-3114 Stewart, M., Webster, J.R., Schaefer, A.L., Cook, N.J., Scott, S.L., 2005. Infrared thermography as a non-invasive tool to study animal welfare. Anim. Welf. 319–325. Stewart, M., Webster, J.R., Verkerk, G.A., Schaefer, A.L., Colyn, J.J., Stafford, K.J., 2007b. Non-invasive measurement of stress in dairy cows using infrared thermography. Physiol. Behav. 92, 520–525. https://doi.org/10.1016/j.physbeh.2007.04.034 Stojkov, J., Weary, D.M., Keyserlingk, M.A.G. Von, 2016. Nonambulatory cows : Duration of recumbency and quality of nursing care affect outcome of flotation therapy. J. Dairy Sci. 99, 2076–2085. https://doi.org/10.3168/jds.2015-10448 Stokes, E.L., Flecknell, P.A., Richardson, C.A., 2009. Reported analgesic and anaesthetic administration to rodents undergoing experimental surgical procedures. Lab. Anim. 43, 149–154. https://doi.org/10.1258/la.2008.008020 Sutherland, M., Mellor, D., Stafford, K.J., Gregory, N., Bruce, R., Ward, R., 2002. Cortisol responses to dehorning of calves given a 5-h local anaesthetic regimen plus phenylbutazone, ketoprofen, or adrenocorticotropic hormone prior to   190 dehorning. Res. Vet. Sci. 73, 115–123. https://doi.org/10.1016/S0034-5288(02)00005-X Tallet, C., Špinka, M., Maruščáková, I., Šimeček, P., 2010. Human Perception of Vocalizations of Domestic Piglets and Modulation by Experience With Domestic Pigs (Sus scrofa). J. Comp. Psychol. 124, 81–91. https://doi.org/10.1037/a0017354 Taylor, A.A., Weary, D.M., 2000. Vocal responses of piglets to castration: identifying procedural sources of pain. Appl. Anim. Behav. Sci. 70, 17–26. https://doi.org/10.1016/S0168-1591(00)00143-X Taylor, P.M., Pascoe, P.J., Mama, K.R., 2002. Diagnosing and treating pain in the horse. Where are we today? Vet. Clin. North Am. Equine Pract. 18, 1–19. The Cornell Lab of Ornithology, 2014. Raven Pro: Interactive Sound Analysis Software (Version 1.5). The Marine Mammal Center, 2018. Released or Deceased Patients [WWW Document]. URL http://www.marinemammalcenter.org/patients/released-deceased-patients/ Thornton, P.D., Waterman-Pearson, A.E., 1999. Quantification of the pain and distress responses to castration in young lambs. Res. Vet. Sci. 66, 107–118. https://doi.org/10.1053/rvsc.1998.0252   191 Thornton, P.D., Waterman-Pearson, A.E., 1997. Castration in young lambs produces changes in mechanical nociceptive threshold responses and behaviour as assessed by a dynamic and interactive visual analogue scale. J. Vet. Anaesth. 24, 41. Tong, A.K.W., Scheafer, A.L., Jones, S.D.., 1995. Detection of poor quality beef using infrared thermography. Meat Focus Int. 4, 443–445. Torrey, S., Devillers, N., Lessard, M., Farmer, C., Widowski, T., 2009. Effect of age on the behavioral and physiological responses of piglets to tail docking and ear notching. J. Anim. Sci. 87, 1778–1786. https://doi.org/10.2527/jas.2008-1354 Travain, T., Colombo, E.S., Heinzl, E., Bellucci, D., Prato Previde, E., Valsecchi, P., 2015. Hot dogs: Thermography in the assessment of stress in dogs (Canis familiaris)—A pilot study. J. Vet. Behav. Clin. Appl. Res. 10, 17–23. https://doi.org/10.1016/j.jveb.2014.11.003 Valera, M., Bartolomé, E., Sánchez, M.J., Molina, A., Cook, N., Schaefer, A., 2012. Changes in eye temperature and stress assessment in horses during show jumping competitions. J. Equine Vet. Sci. 32, 827–830. Valverde, A., Gunkel, C.I., 2005. Pain management in horses and farm animals. J. Vet. Emerg. Crit. Care 15, 295–307. https://doi.org/10.1111/j.1476-4431.2005.00168.x   192 van Loon, J.P., Jonckheer-Sheehy, V.S., Back, W., van Weeren, P.R., Hellebrekers, L.J., 2014. Monitoring equine visceral pain with a composite pain scale score and correlation with survival after emergency gastrointestinal surgery. Vet. J. 200, 109–115. Vianna, D.M.L., Carrive, P., 2005. Changes in cutaneous and body temperature during and after conditioned fear to context in the rat. Eur. J. Neurosci. 21, 2505–2512. https://doi.org/10.1111/j.1460-9568.2005.04073.x Vick, S., Waller, B.M., Parr, L.A., Smith Pasqualini, M.C., Bard, K.A., 2007. A cross-species comparison of facial morphology and movement in humans and chimpanzees using the Facial Action Coding System (FACS). J. Nonverbal Behav. 31, 1–20. https://doi.org/10.1007/s10919-006-0017-z Viñuela-Fernández, I., Jones, E., Chase-Topping, M.E., Price, J., 2011. Comparison of subjective scoring systems used to evaluate equine laminitis. Vet. J. 188, 171–177. Visser, E.K., Van Reenen, C.G., Van der Werf, J.T.N., Schilder, M.B.H., Knaap, J.H., Barneveld, A., Blokhuis, H.J., 2002. Heart rate and heart rate variability during a novel object test and a handling test in young horses. Physiol. Behav. 76, 289–296. https://doi.org/10.1016/S0031-9384(02)00698-4 von Borell, E., Langbein, J., Després, G., Hansen, S., Leterrier, C., Marchant-Forde, J., Marchant-Forde, R., Minero, M., Mohr, E., Prunier, A., Valance, D., Veissier, I.,   193 2007. Heart rate variability as a measure of autonomic regulation of cardiac activity for assessing stress and welfare in farm animals--a review. Physiol. Behav. 92, 293–316. https://doi.org/10.1016/j.physbeh.2007.01.007 Waalkes, M.P., Rehm, S., Kasprzak, K.S., Issaq, H.J., 1987. Inflammatory, proliferative, and neoplastic lesions at the site of metallic identification ear tags in Wistar [Crl:(WI)BR] rats. Cancer Res. 47, 2445–2450. Walker, K.A., Duffield, T.F., Weary, D.M., 2011a. Identifying and preventing pain during and after surgery in farm animals. Appl. Anim. Behav. Sci. 135, 259–265. https://doi.org/10.1016/j.applanim.2011.10.021 Walker, K.A., Mellish, J.E., Weary, D.M., 2011b. Effects of hot-iron branding on heart rate, breathing rate and behaviour of anaesthetised Steller sea lions. Vet. Rec. 169, 363–363. https://doi.org/10.1136/vr.d4911 Walker, K.A., Trites, A.B., Haulena, M.C., Weary, D.M., 2012. A review of the effects of different marking and tagging techniques on marine mammals. Wildl. Res. 39, 15–30. https://doi.org/http://dx.doi.org/10.1071/WR10177 Wall, P.D., 1992. Defining “pain” in animals, in: Short, C.E., Van Poznak, A. (Eds.), Animal Pain. Churchill Livingstone, New York, pp. 63–69. Waller, B.M., Lembeck, M., Kuchenbuch, P., Burrows, A.M., Liebal, K., 2012.   194 GibbonFACS: A muscle-based facial movement coding system for Hylobatids. Int. J. Primatol. 33, 809–821. https://doi.org/10.1007/s10764-012-9611-6 Waller, B.M., Peirce, K., Caeiro, C.C. Scheider, L. Burrows, A.M.,  et al., 2013. Paedomorphic facial expressions give dogs a selective advantage. PLoS One 8, e82686. https://doi.org/10.1371/journal.pone.0082686 Waran, N., Best, L., Williams, V., Salinsky, J., Dale, A., Clarke, N., 2007. A preliminary study of behaviour-based indicators of pain in cats. Anim. Welf. 16, 105–108. Wathan, J., Burrows, A.M., Waller, B.M., McComb, K., 2015. EquiFACS: The equine facial action coding system. PLoS One 10, e0131738. https://doi.org/10.1371/journal.pone.0131738 Watts, J.M., Stookey, J.M., 2000. Vocal behaviour in cattle: the animal’s commentary on its biological processes and welfare. Appl. Anim. Behav. Sci. 67, 15–33. https://doi.org/10.1016/S0168-1591(99)00108-2 Watts, J.M., Stookey, J.M., 1999. Effects of restraint and branding on rates and acoustic parameters of vocalization in beef cattle. Appl. Anim. Behav. Sci. 62, 125–135. https://doi.org/10.1016/S0168-1591(98)00222-6 Weary, D., Fraser, D., 1995. Calling by domestic piglets : reliable signals of need ? Anim. Behav. 50, 1047–1055.   195 Weary, D.M., Braithwaite, L.A., Fraser, D., 1998. Vocal response to pain in piglets. Appl. Anim. Behav. Sci. 56, 161–172. Weary, D.M., Fraser, D., 1997. Vocal response of piglets to weaning: effect of piglet age. Appl. Anim. Behav. Sci. 54, 153–160. Weary, D.M., Fraser, D., 1995. Signalling need : costly signals and animal welfare assessment. Appl. Anim. Behav. Sci. 44, 159–169. Weary, D.M., Lawson, G.L., Thompson, B.K., 1996. Sows show stronger responses to isolation calls of piglets associated with greater levels of piglet need. Anim. Behav. 1247–1253. Weary, D.M., Niel, L., Flower, F.C., Fraser, D., 2006. Identifying and preventing pain in animals. Appl. Anim. Behav. Sci. 100, 64–76. https://doi.org/10.1016/j.applanim.2006.04.013 Weary, D.M., Ross, S., Fraser, D., 1997. Vocalizations by isolated piglets : a reliable indicator of piglet need directed towards the sow. Appl. Anim. Behav. Sci. 53, 249–257. Welsh, E.M., Gettinby, G., Nolan, A.M., 1993. Comparison of a visual analogue scale and a numerical rating scale for assessment of lameness, using sheep as a model. Am. J. Vet. Res. 54, 976–983.   196 White, R.G., DeShazer, J.A., Tressler, C.J., Borcher, G.M., Davey, S., Waninge, A., Parkhurst, A.M., Milanuk, M.J., Clemens, E.T., 1995. Vocalization and physiological response of pigs during castration with or without a local anesthetic. J. Anim. Sci. 73, 381–386. Wiesenfeld, A.R., Zander Malatesta, C., Deloach, L.L., 1981. Differential parental response to familiar and unfamiliar infant distress signals. Infant Behav. Dev. 4, 281–295. Williams, A.C., 2002. Facial expression of pain: An evolutionary account. Behav. Brain Sci. 25, 439–455. https://doi.org/10.1017/S0140525X02000080 Williams, A.C., Craig, K.D., 2016. Updating the definition of pain. Pain 157, 2420–2423. https://doi.org/10.1097/j.pain.0000000000000613 Wright-Williams, S.L., Courade, J.P., Richardson, C.A., Roughan, J. V, Flecknell, P.A., 2007. Effects of vasectomy surgery and meloxicam treatment on faecal corticosterone levels and behaviour in two strains of laboratory mouse. Pain 130, 108–118. https://doi.org/10.1016/j.pain.2006.11.003 Yarnell, K., Hall, C., Billett, E., 2013. An assessment of the aversive nature of an animal management procedure (clipping) using behavioral and physiological measures. Physiol. Behav. 118, 32–39. https://doi.org/10.1016/j.physbeh.2013.05.013   197 Zhang, J., Jianxiong, A., 2007. Cytokines, inflammation and pain. Int. Anesthesiol. Clin. 45, 27–37. https://doi.org/10.1097/AIA.0b013e318034194e.Cytokines Zimmerman, P.H., Koene, P., van Hooff, J., 2000. The vocal expression of feeding motivation and frustration in the domestic laying hen, Gallus gallus domesticus. Appl. Anim. Behav. Sci. 69, 265–273. Zimmermann, M., 1986. Behavioural investigations of pain in animals, in: Duncan, I.J.H., Molony, V. (Eds.), Assessing Pain in Farm Animals. Commission of the European Communities, Luxembourg, pp. 16–27.    

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:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0375838/manifest

Comment

Related Items