Happy animals make good science Trevor Poole Universities Federation for Animal Welfare, 8 Hamilton

Close, Potters Bar, Hertfordshire

EN6 3QD, UK

Summary In this paper the question is posed whether it is not only better for the animal to be happy, but whether its state of mind may also have the potential to influence the scientific results derived from it. To ensure good science, the animal should have a normal physiology and behaviour, apart from specific adverse effects under investigation. There is a growing body of evidence from a wide variety of sources to show that animals whose well-being is compromised are often physiologically and immunologically abnormal and that experiments using them may reach unreliable conclusions. On scientific, as well as ethical grounds, therefore, the psychological well-being of laboratory animals should be an important concern for veterinarians, animal technicians and scientists. Well-being; laboratory animals; endocrine; immune response; handling; experimental method

Keywords

What are happy animals? Most people who have worked closely with animals or who keep them as pets are aware if they are suffering or unwell; the signs may be small but we become aware that all is not well. Colloquially, we will say to our colleagues 'that animal is not happy'. The signs which tell us that there is something wrong are changes in the behaviour which we have come to expect from the individual, for example, we may find it sitting huddled in a comer, failing to greet us or lacking interest in events taking place around it. If the behavioural change persists, we take action and may even call in veterinary advice. Happiness and unhappiness, or distress, refer to states of mind of the animal; they cannot be measured directly but the whole concept of animal welfare is based on the belief that higher animals, like us, are able to experience pain and pleasure. The best way to decide whether an animal is happy or distressed is by observing its behaviour. On Corespondence Accepted

to: Trevor Poole

6 September

1996

this basis, I will define a 'happy animal' as one which is alert and busy (shows a wide repertoire of behaviour), is able to rest in a relaxed manner, is confident (outward going and does not display fear towards trivial nonthreatening stimuli) and does not show abnormal behaviour. It is, of course, important to be familiar with the particular .animal's character to make these judgements. Some individuals are naturally extrovert and active, while others are quiet and lethargic. In the laboratory, those of us who care for animals like to think that our charges are happy and that any procedures which have to be performed on them cause them the absolute minimum of distress. In this article the question is posed whether it is not only better for the animal to be happy, but whether its state of mind may also have the potential to influence the scientific results which are derived from it. The most obvious cases of unhappiness will occur when animals are sick or injured and symptoms will vary from mild to severe. Generally, with modem laboratory practice this source of distress has virtually been eliminated. We take it for granted that LaboratoryAnimals

(1997) 31,116-124

Happy animals make good science

scientists do not work with animals which are ill or injured. Nor do laboratory animals lack essential physical needs, such as food, water or suitable climatic conditions. There remain, however, a variety of potential causes of distress, such as social problems with aggressive cage mates, overcrowding, or social isolation. There are also features of the physical environment, such as loud or sudden noises, including ultrasound which can be perceived by rodents, dogs and smaller primates, which might also be sources of distress. Finally, there are the attitudes and sometimes inexpert manipulations by staff. Mammals, particularly become distressed if they are badly handled, especially when they are restrained by personnel unfamiliar to them. This may be a common occurrence, as the experimenter is not usually the person in day-to-day care. A factor which is increasingly being recognized as a source of unhappiness, is the failure of the captive environment to meet the animal's behavioural needs and assure its psychological well-being. It is becoming apparent that captive mammals can be bored or resort to abnormal behflviour if their environment is not sufficiently complex and interesting to them (Wemelsfelder 1990, Poole 1988).

What is good science? The quality of experimental laboratory animal science depends on three essential conditions being satisfied. Firstly, there should be an important problem for which an answer is sought, secondly, the experiment should yield unambiguous results which provide an answer to the problem and, finally, variables which are not under investigation should be strictly controlled. I shall take for granted the assumption that the first two conditions have been met and only be concerned with the third which, can also have a direct bearing on the well-being of the animals. Good laboratory animal science is based on normal, healthy subjects, unless the illness is itself the subject of investigation. Scientific method assumes the absence of confounding factors or uncontrolled variables. Clearly, unhappiness might be a confounding variable unless, for example, its

117

alleviation was the subject of the study, as in the case of testing an anti-depressant. Whatever the subject under investigation, all unnecessary stress should be minimized during the experiment to reduce the variability of the results and thus the number of animals required. This requires, firstly, a thorough understanding of the animal and its biology and, secondly, experiments which are well designed, statistically valid and appropriate. Even in situations where the experiment itself creates unhappiness for the animal, such as premature removal of young, any effect may be diminished or even lost if the animal was not happy in the first instance. We can therefore conclude that, in all aspects, apart from unavoidable adverse effects of the experiment, the animal should be happy. Most scientists working with animals will make the assumptions that they will have normal blood pressure, heart rates, levels of stress hormones, immunological competence, digestion, appetite and behaviour. To avoid confounding variables, experimental animals should have both normal physiology and behaviour. Some might argue that behaviour is of less significance than physiology, but this is based on the erroneous supposition that mind and body are separate entities and that one will not influence the other. Recent scientific work has shown how the brain, behaviour, hormones and even the immune system, are all interdependent and that disturbances in one of these systems commonly influences one or all of the others (see review article by Martin 1989 and Bohus & Koolhaas 1991).In fact, behavioural changes are usually more sensitive indicators of distress than physiological ones. I shall now consider some of the main factors which may influence the psychological well-being of laboratory animals.

Factors influencing psychological well-being Social factors Laboratory mice are commonly kept in single sex groups in stock cages. While females generally tolerate such conditions, males

118

fight and establish a hierarchy, but the type of social structure depends on the number of individuals in the cage. Poole and Morgan (1973) found that, in small colonies of 3--4 male CFW mice, the dominant's aggression was directed to few rivals and these subordinates were highly intimidated and restricted in their movements about the cage. Five male mice however formed a linear hierarchy with each individual knowing its place and thus subordinates were able to develop strategies to avoid conflict. In larger colonies of nine or more individuals the dominant was unable to control such a large number of subordinates, so that the social structure broke down after 3-5 days and another dominant arose. This contrasted with the situation in smaller groups of up to five in number where, barring disturbance, aggression gradually decreased and was minimal after 12-15 days. Physiological data from laboratory mice have shown that, in standard housing, subordinates exhibit higher levels of stress and sex hormones than dominants (Hucklebridge et al. 1976, Benton et al. 1978). In addition to experiencing fear, suffering injury and high levels of stress hormones, Beden and Brain (1984, 1985) found that the immunological response to an antigen (sheep red blood cell) was reduced in subordinate or defeated mice. Brayton and Brain (1973) found that crowding mice lowered their resistance to a digenean parasite and Edwards and Dean (1977) also showed that crowding affects the mouse's immune response. Age is another important consideration when housing male micej litter mates, unless significantly disturbed will usually live amicably together, as will members of different litters who have been grouped together from an early age. Clearly, wherever possible, animals, should be kept in conditions where their social grouping leads to the minimum of aggression, and hence distress. There has been a tendency to believe that isolation in the form of single housing is distressing for other mammals, as it is for humans. However, it has been found that isolated male mice have hormonal profiles similar to dominants (Brain 1975, Hucklebridge et al. 1976, Brain & Benton 1977,

Poole

Benton et al. 1978, Brain 1990), so that singly-housed mice do not suffer from 'isolation stress'. This finding is compatible with the fact that mice are territorial in the wild and thus actively repel other males. However, providing a simulation of the wild in the laboratory may actually be deleterious because Bishop and Chevins (1988) found that territorial male mice placed in an arena had high levels of stress hormones, presumably associated with the defence of their territories. While access to social companions benefits many species of laboratory animals, some solitary species, may have to be kept alone because of their aggressive tendencies, for example, male rabbits and ferretsj this is particularly common in cases where males have bred or been exposed to members of the opposite sex, or females have infant young. Even individuals of social species must have adequate space to avoid one another and thus minimize any conflict. Bohus and Koolhaas (1991) reviewed the literature on psycho-immunology and came to the conclusion that social stress is only likely to impair immune function significantly when the animal is unable to exert some control over the situation by developing a coping strategy. Inability to escape from an attacker and chronic overcrowding are clear examples of situations which animals are unable to control. It is important to remember that a rat is not simply a scaled-up mouse. Rats are much more sociable than mice and consequently seem likely to suffer more when isolated, although Brain and Benton (1979) could find no evidence of isolation stress in this species. However, young rats show very active social play which involves both chasing and wrestling (Poole & Fish 1975, 1976), so they need adequate space to do this. During their developmental stage, rats acquire skills and experience; adult rats reared in deprived conditions are not only less intelligent, but also have smaller brains than those from rich and stimulating environments (Renner & Rosenzweig 1978). They are thus less normal, but this does not necessarily mean that they are less happy. Beynen (19921 reviewed literature which indicated that control rats in the same room

Happy animals make good science

as the experimentals showed raised levels of corticosteroids, as compared to controls in another room. This suggests that some rats (and therefore possibly other animals) may be able to communicate their feelings to other individuals either by vocalizations or pheromones and that situations which cause distress may also upset others within the range of these forms of communication. Mendoza et a1. (1991)found that squirrel monkeys showed differences in levels of corticosteroids which related to their sex and social grouping. Females showed higher levels when kept singly, with a single female companion, or with a male. Three females housed together appeared to be the minimum social group which would reduce the levels of corticosteroids to a normallevelj this number also showed much higher levels of reproductive cycling as compared to the smaller groups or females paired with a male. Likewise male squirrel monkeys showed lower levels of corticosteroids when housed with male companions. The practical implication for husbandry and breeding is that the minimum breeding group should consist of two males and three females. Both mammals and some birds have a period of life when they acquire knowledge of the world and are able to test its properties while protected by a vigilant mother or family group. The developmental environment determines, to a large extent the kind of situations with which they are able to cope when they reach adulthood. During childhood they practise skills and motor coordinations which will be of benefit in later life. Many mammals play fight when young and thus practise the strategies of attack and defence which they will need when faced with real rivals. They show extraordinary curiosity and inventiveness and thus learn the properties of objects and other organisms in their environment. Mammals enjoy play and experimentation in the sense that it is self-rewarding, so that they should be provided with a stimulating and complex developmental environment. Likewise, the presence of a mother is important because it allows the young to express their wide repertoire of play and curiosity without fear.

119

Early weaning is undoubtedly stressful for mammals because of the upset caused by losing their mothers. The problem is further exacerbated by the sudden loss of maternal antibody. This immune deficiency is temporary until the young independent animal develops its own fully functioning immune system. However, the temporary immunodeficiency may be enhanced by stress resulting from separation and immuno-suppression has been recorded from both adult pigs (Blecha & Kelley 1981, Blecha et a1. 1983)and primates (Reite 1987)which had been removed from their mothers unnaturally early. For example, Reite showed that the separation of young Macaca nemestrina for 2 weeks at the age of 6 months still had an effect on the immune system in the form of a reduction in T-cell proliferation in response to a mitogen 6 years later. Early separation from the mother, is commonly practised by the breeders of non-human primates, so that the young not only develop abnormal behaviour (Goosen 1989)but may also suffer from reduced immunological competence as adults, and thus, would be unsatisfactory as experimental animals. The phYSical environment There is increasing evidence that a number of physical environmental factors which influence an animal's psychology may also affect its immune system. Unpleasant events, such as inescapable electric shock, increase the incidence of tumours in rats (Keller et a1. 1981).However, not all stress is deleterious and some may even be beneficial. For example Marsh et a1. (1963)found that cynomolgus monkeys (Macaca fascicularis) trained over a 24-h period to avoid electric shocks increased resistance to polio virus infection. Similarly, mice which had been trained to avoid electric shocks showed greater resistance to malaria than controls not subjected to the training. What seems to be important in these animals is not the shock itself but the degree of control which the animal is able to exert to avoid it. Mice conditioned to drink saccharine solution followed by an immuno-suppressive drug (cyclophosphamide) were subsequently

120

found to have a suppressed immune system when drinking saccharine alone. This indicates that the immune response may actually be susceptible to Pavlovian conditioning (Kelley et a1. 1984, 1985). Differences in levels of stress hormones also relate to the animal's ability to control events. Rats were subjected to an electric shock from the grid on which they walked. Where the rat could prevent the grid from becoming electrified by pressing a lever, corticosteroid levels were much lower than in controls which were shocked but had no way of switching it off. There are also marked species differences in responsiveness to stressors, even when they are closely related, for example rhesus, bonnet and cynomolgus macaques show marked differences in hormonal responses to routine procedures (Clark et a1. 1988). Mendoza and Mason (1984) compared the physiology of titi (Callicebusl and squirrel monkeys (Saimiri). The former are notoriously difficult to keep in the laboratory whereas squirrel monkeys readily adapt to captivity. The two species show very different endocrinological responses to stress. Interestingly, Callicebus shows very limited corticosteroid response to stress compared to Saimiri and the authors suggest that the capacity to show a high level of corticosteroids relates to the lifestyles of the two species. Saimiri is an active, mobile and highly exploratory species.whereas Callicebus has a small home range, is monogamous and leads a rather quiet life. Thus in keeping species in captivity it is important to consider their ability to adapt to an artificial environment which is far removed from their natural way of life. The titi monkey is clearly unhappy in captivity and has only a limited ability to cope with a profound change from its natural environment. Mild unpredictable stressors (such as water deprivation, continuous illumination, cage tilt, living in a soiled cage, or loud noises) have been shown to influence the appetite of rats and mice for sweet substances (Willner et a1. 1987, Monleon et a1. 1995). These uncontrolled variables, which may be associated with poor husbandry, could seriously compromise experiments using a food re-

Poole

ward. Exposure to bright light, often resulting from cages being high on a rack, can also be aversive to nocturnal rodents and can induce retinal degeneration, prenatal mortality and decreased growth rates in various strains of rats and mice (Clough 19841.Schlingmann et a1. (1993) showed that rats show avoidance behaviour to light intensities as low as 20-25 lux, which is well below the threshold for retinal degeneration (60 lux). In many instances experiments on rats and mice are carried out during the daytime and in bright light, when the animals would normally be asleep. This would seem certain to cause them some distress but, as far as I am aware, no one has investigated whether such conditions may add to any stress caused by the experiment. The environment may also contain stressors of which we are unaware. An obvious example is noise which is in the ultrasonic range and can be perceived by and is known to influence the behaviour of rodents. I recall entering an experimental laboratory where drugs were tested on rodents. It contained five computers with visual display units. Such equipment emits a high pitched ultrasonic scream which is almost indistinguishable from the fear cry of a rat (Sales et a1. 1988). Three computer visual display units

(VDUs) were switched on when I entered the laboratory and, when I asked whether the computers were switched off during experiments, the scientists were obviously surprised by my question and unaware of the possibility that VDUs might affect experimental results! Some of the most stressful events in every day husbandry result from changes in environment. Animals may be placed in unfamiliar/ clean cages and, for mammals which scent mark their home range, this may be highly disturbing. For example, rodents often fight when moved to another cage. Even more stressful is the situation where an animal is moved from its home cage to an unfamiliar one and then subjected to an experiment. It is good practice to allow plenty of time for the animal to acclimatize to its new situation, not only on welfare grounds but also because the experimental results may be influenced. For example

Happy animals make good science

Damon et a1. (1986) compared the nephrotoxic response of rats to implanted, refined uranium ore. They implanted group I rats and then moved them to metabolism cages; group II was allowed 21 days to acclimatize to the metabolism cage before implantation, while group III were housed in polycarbonate cages and implanted after 21 days, four being retained without implantation, as controls. The surprising result was that, for group I rats 3-8 mg/kg proved toxic, while for the acclimatized experimental rats in groups II and III, the toxic dose was 220-650 mg/kg. This example provides concrete data to support the practice of carrying out experiments in a familiar environment. Handling and training Another important consideration which can affect their well-being, is the way in which animals are handled or restrained. Barclay et a1. (1988) found significant values of their 'Disturbance Index' resulted from longer periods of restraint lover 20 s) while Blecha et a1. (1982) showed that mice restrained in a wire cone for 2 h, had increased levels of corticosteroids and that their immune response was suppressed. Barclay et a1. (loc cit) found that mice showed significant Disturbance Indexes when the handler had previously contacted cat urine. Mice exposed to a cat (Hamilton 19741, or those which had been fighting, both had lower levels of resistance to tapeworms indicating a stress-induced immunosuppresion. It is well known that restraint can be highly stressful to mammals (Cronin 1985, Lawrence 1991) and it may also lead to suppression of the immune system (Rasmussen et a1. 1957, Levine et a1. 1962). In spite of this, in many laboratories monkeys are manhandled, restrained in crush cages or even anaesthetized to carry out even trivial procedures such as injections or blood sampling. Monkeys, like most mammals, can easily be trained to cooperate in procedures. For example, rhesus monkeys trained to extend an arm for an injection showed a much lower incidence of diarrhoea as compared with animals who were physically

121

restrained IReinhardt 1990, 1992). Monkeys in zoos have been trained to perform a variety of tasks, for example, female drills were encouraged to present for artificial insemination and a male to masturbate in a special area to facilitate the collection of sperm. Reinhardt (1991, 1992) also found that the home cage is the best place to carry out venipuncture on Macaca nemestrina and that cortisol levels were lower in animals trained to accept this procedure when compared with individuals who were restrained. Monkeys can also be trained to open their mouths for dental examination. Most, if not all, laboratory mammals and birds recognize humans as individuals and are nervous of strangers. When an experimental procedure is to be carried out it is therefore preferable that the handler should be a person familiar to the animal, in whom it has confidence. An unfamiliar handler will undoubtedly cause the animal fear and stress. Wherever possible, laboratories should ensure that the animal is familiar with those in direct contact with it during the experiment. Thus, the role of the animal technician is of vital importance and should be appreciated by the scientist (Biological Council 19921. Kind and gentle handling make all the difference to the animal. This is referred to as 'stockmanship' in farming circles, where it has been shown that a friendly stockman can increase the milk yield of cows as compared with persons who treat them humanely but do not form any relationship with their animals (Seabrook 1984). Atherosclerosis is reduced in rabbits handled in a consistent and friendly way as opposed to the more usual laboratory procedure of simply picking them up and restraining them (Nerem et a1. 1980). Good handling and friendly approaches have been shown to lead to greater growth rates and reproductive success in pigs (Hemsworth & Barnett 1987). Insome instances experimental pigs with catheters did better than controls, probably because they received more attention from staff (Wiepkema 1990). Good handling and training animals to cooperate, not only improves the quality of the relationship between carer and animal, but also allows the animal to exercise its intelligence. One of the major

Poole

122

problems in experiments with conscious animals is the fear or anxiety which the subject experiences. If the experimental animal has been trained to cooperate and has confidence and trust in the handler it will be much less stressed and the experiment will be much improved by the removal of this unwanted variable. Thus a positive, caring attitude by staff not only improves the wellbeing of the animal but also makes it more willing to cooperate in any procedures to which it is subjected. There are, of course sources of potential stress even in the best run animal houses. Husbandry procedures themselves can disturb animals and, for example, the routines which occur regularly may affect animals in ways which are not apparent. Line et al. (1989) found that during cage cleaning the heart rate of rhesus monkeys increased and remained significantly above normal for 2 h afterwards. Thus, there could be considerable effects on, for example, drug metabolism, depending on the time of day in which the animal was given the dose. Barclay et al. (1988) showed that the behaviour of rats was significantly disturbed when they were restrained by an inexperienced handler as compared to an experienced one. One of the referees rightly pointed out that this paper seems to have a bias toward rats, mice and non-human primates; this is correct, but I do believe that it reflects the coverage in the scientific literature. While it would be surprising if other species of mammal, and probably birds, were not affected similarly, there is no doubt that additional information on other laboratory animals would be valuable.

Conclusions Happy animals are busy, confident and behave normally. They will not be in pain, or distressed and should resist disease and reproduce successfully. I have given examples which show that both the endocrine condition and immunology of laboratory animals, which experimenters may assume to be normal, can be compromised by social conditions, developmental history and stressors in the animal unit or experimental

laboratory. One of the most important aspects of the life of laboratory animals are their relations with human handlers and care givers on whom they are totally dependent. Good, kindly treatment and simple humane training are beneficial both in reducing stress and in producing animals which are confident, cooperative and easily handled; they will also be the best subjects for scientific investigation. While it has to be accepted that no animal can live an entirely stress-free life, what I have termed a happpy animal is readily able to cope with the stressors to which it is subjected. Unhappy animals have to put up with distressing conditions beyond their control which result in behavioural and physiological disabilities such as permanently raised levels of stress hormones or reduced concentrations of sex hormones and a compromised immune system; these uncontrolled variables make them unsuitable subjects for scientific studies. These findings make it obligatory for scientists to do everything practicable to ensure the happiness of laboratory animals if the quality of their research is to be beyond reproach. Acknowledgment I am extremely grateful to Dr E. D. Williamson for her very helpful comments on this manuscript, however, I alone take responsibility for the opinions expressed therein.

References Barclay RJ, Herbert WJ, Poole TB (1988) The Dis· turbance Index: a behavioural method of assessing the severity of common laboratory procedures on rodents. UFA W Animal Welfare Research Report 2, Potters Bar: Universities Federation for Animal Welfare Beden SN, Brain PF (1984) Effects of attack-related stress on the primary immune responses to sheep red blood cells in castrated mice. IRCS Medical Science 12, 675 Beden SN, Brain PF (1985) The primary immune responses to sheep red blood cells in mice of differing social rank or from individual housing. IRCS Medical Science 13, 364-5 Benton D, Goldsmith JF, Gamal-el Din L, Brain PF, Hucklebridge FH 11978) Adrenal activity in isolated mice and mice of different social status. Physiology and Behaviour 20, 459-64 Beynen AC (19921 Communication between rats of experiment-induced stress and its impact on experimental results. Animal Welfare 1, 153-9

Happy animals make good science

Biological Council (19921 Guidelines on the Handling and Training of Laboratory Animajs. Potters Bar: Universities Federation for Animal Welfare Bishop ML, Chevins PFD (1988) Territory formation by mice under laboratory conditions: welfare considerations. In Laboratory Animal Welfare Research Vol. 1 Rodents (UFAW, edl. Potters Bar: Universities Federation for Animal Welfare, pp 25-48 Blecha F, Kelley KW (1981) Effects of cold and weaning stressors on the antibody-mediated immune response of pigs. Journal of Animal Science 53,439 Blecha F, Kelley KW, Satterlee DG (1982) Adrenal involvement in the expression of delayed-type hypersensitivity to SRBC and contact sensitivity to DNFB in stressed mice. Proceedings of the Society for Experimental Biology and Medicine 169, 257 Blecha F, Pollman DS, Nichols DA (1983) Weaning pigs at an early age decreases cellular immunity. Journal of Animal Science 56, 396 Bohus B, Koolhaas JM 119911Psychoimmunology of social factors in rodents and other sub primate vertebrates. In Psychoneuroimmunology (Ader R, Felten DL, Cohen N, eds). San Diego: Academic Press, pp 807-30 Brain PF 119751What does individual housing mean to a mouse? Life Sciences 16, 187-200 Brain PF (1990) Stress in agonistic contexts in rodents. In: Stress in Domestic Animals (Dantzer R, Zayan R, eds). Dortrecht: Kluwer Academic, pp 74-85 Brain PF, Benton D 119771What does individual housing mean to a research worker? IRCS Journal of Medical Science 5, 459-63 Brain PF, Benton D 119791The interpretation of physiological correlates of differential housing in laboratory rats. Life Sciences 24, 99-116 Brayton AR, Brain PF 119731Effects of 'crowding' on endocrine function and retention of the digenean parasite Microphallus pygmaeus in male and female mice. Journal of Helminthology 48, 99-106 Clark AS, Mason W, Moberg G 119881Differential and behavioral and adrenocortical responses to stress among three macaque species. American Journal of Primatology 14, 37-52 Clough G 119841Environmental factors in relation to the comfort and well-being of laboratory rats and mice. In: Standards in Laboratory Animal Management (UFAW, ed). Potters Bar: Universities Federation for Animal Welfare, pp 7-24 Cronin GM (1985) The Development and Significance of Abnormal Stereotyped Behaviour in Tethered Sows (PhD Thesis). Netherlands: Agricultural University of Wageningen Damon EG, EidsonAF, Hobbs CH, Hahn F 119861 Effect of acclimation on nephrotoxic response of rats to uranium. Laboratory Animal Science 36,24-7 Edwards EA, Dean LM (1977) Effects of crowding mice on humeral antibody formation and protection to

123

lethal antigenic challenge. Psychosomatic Medicine 39, 19-24 Friedman SB, Ader R, Grota LJ (1973) Protective effect of noxious stimulation in mice infected with rodent malaria. Psychosomatic Medicine 35, 535-7 Goosen C (1989) Influence of the age of weaning on the behaviour and well-being of rhesus monkeys. In: Laboratory Animal Welfare Research: Primates (UFAW, edl. Potters Bar: Universities Federation for Animal Welfare, pp 17-22 Hamilton DR (19741 Immunosuppressive effects of predator-induced stress in mice with acquired immunity to Hymenolepsis nana. Journal of Psychosomatic Research 18, 143-53 Hemsworth PH, Barnett JL (1987) Human-animal interactions. The Veterinary Clinician of North America 3, 339-56 Hucklebridge FH, Reid A, Benton D, Brain PF (1976) A comparison of the levels of the adrenal catecholamines and aggession in isolated, dominant and subordinate mice. IRCS Medical Science 4, 154 Keller SE, Weiss JM, Miller NE, Stein M (1981) Suppression of immunity by stress: effect of graded scries of stressors on lymphocyte stimulation in the rat. Science 213, 1397 Kelley KW, Dantzer R, Mormede P, Salmon H, Aynaud J-M (19841 Conditioned taste aversion induces immunological suppression in the absence of an immunosuppressive drug. Comptes Rendus des Seances de l'Academie des Sciences Serie III 299, 123 Kelley KW, Dantzer R, Mormede P, Salmon H, Aynaud J-M (19851 Conditioned taste aversion suppresses induction of delayed-type hypersensitivity immune reactions. Physiological Behaviour 34, 189

Lawrence A (1991) Introduction: the biological basis of animals responses to handling. In: Practical Animal Handling (Anderson RS, Edney A, eds). Oxford: Pergamon Press, pp 1-13 Levine S, Strebel R, Wenk E1, Harman PJ 119621 Suppression of experimental allergic encephalomyelitis by stress. Proceedings of the Society for Experimental Biology and Medicine 109, 294 Line SW, Morgan KN, Markowitz H, Strong S 119891 Heart rate and activity of rhesus monkeys in response to routine events. Laboratory Primate Newsletter 28, 9-12 Marsh JT, Lavender JF, Chang S5, Rasmussen AF (1963) Poliomyelitis in monkeys: decreased susceptibility after avoidance stress. Science 140, 1414-15 Martin P (1989) Psychoimmunology: relations between brain, behaviour and immune function. In: Perspectives in Ethology 8: Whither Ethology! (Bateson PPG, Klopfer PH, edsl. New York: Plenum Press, pp 173-214 Mendoza SP, Mason WA (1984) Rambuctious Saimiri and reluctant Callicebus: systemic contrasts in

Poole

124

stress physiology. American Journal of Primatology 6,415 Mendoza SP 11991) Behavioural and physiological indices of social relationships: comparative studies of New World monkeys. In: Primate Responses to Environmental Change (Box HO, ed). London: Chapman & Hall, pp 311-36 Monleon S, D'Aquila P, Parra A, Simon VM, Brain PF, Willner P 11995) Attenuation of sucrose consumption in l11ice by chronic mild stress and its restoration by imipramine. Psychopharmacology 117,453-7 Nerem RM, Levesque MJ, Comhill JF (1980) Social environment as a factor in diet induced atherosclerosis. Science 208, 1475-6 Poole TB (1988) Normal and abnormal behaviour in . captive primates. Primate Report 22, 3-12 Poole TB, Fish J (1975) An investigation of playful. behaviour in Rattus norvegicus and Mus musculus (Mammalia). Journal of the Zoological Society of London 175, 61-71 Poole TB, Fish J 11976) An investigation of individual, age and sexual differences in the play of Rattus norvegicus IMammalia: Rodential. Journal of the Zoological Society of London 179, 249-60 Poole TB, Morgan HDR 11973) Difierences in aggressive behaviour between male mice (Mus musculus) in colonies of different sizes. Animal Behaviour 21, 788-95 Rasmussen AF, Marsh JT, Brill NQ 11957) Increased susceptibility to herpes simplex in mice subjected to avoidance-learning stress of Il~straint. Proceedings of the Society for Experimental Biology and Medicine 96, 183 Reinhardt V (1990) Avoiding undue stress: catching individual animals in groups of rhesus monkeys. Lab Anima119, 52-3 Reinhardt V (1991) Training rhesus monkeys to actively cooperate during in-horne cage venipuncture. Animal Technology 42, 11-17

Reinhardt V (1992) Improved handling of experimental rhesus monkeys. In: The Inevitable Bond: Examining Scientist-Animal Interactions IDavis H, Balfour AD, eds). Cambridge: Cambridge University Press Reite M (19871 Infant abuse and neglect: lessons from the primate laboratory. Child Abuse and Neglect 11,347-55 Renner MJ, Rosenzweig MR (19781 Enriched and Impoverished Environments. Effects on Brain and Behaviour. New York: Springer-Verlag Sales GD, Wilson Environmental animal houses: welfare and use 369-75

KJ, Spencer KEV, Milligan SR (1988) ultrasound in laboratories and a possible cause for concern in the of animals. Laboratory Animals 22,

Schlingmann F, Derijk SHLM, Pereboom WJ, Remie R (1993) 'Avoidance' as a behavioural parameter in the determination of distress amongst albino and pigmented rats at various light intensities. Animal Technology 44, 87-96 Seabrook MF (1984) The psychological interaction between the stockman and his animals and its influence on performance of pigs and dairy cows. Veterinary Record 115, 84-7 Wemelsfelder F (19901 Boredom and laboratory animal welfare. In: The Experimental Animal in Biomedical Research. (Rollin BE, Kessel ML, eds). Boca Raton: CRC Press Wiepkema PR 119901Stress: ethological implications. In: Psychobiology of Stress (Puglisi-Allegra S, Oliverio A, eds). Netherlands: Kluwer, pp 1-13 Willner P, Towell A, Sampson D, Muscat R, Sopholeous S (19871 Reduction of sucrose preference by chronic mild stress and its restoration by a tricyclic antidepressant. Psychopharmacology 93,

358-64