425
23
Section ofComparative Medicine
adjusted according to the number of mice in each group. At 8 weeks the heaviest male and female were selected to enter the next generation. A cyclical mating programme was used to minimize in-breeding effects. Early results showed that breeding floor areas of 600 cm2 or more impaired infantile growth and survival and it was shown that survival was directly related to cage size. Female mice bred in cages of 150 cm2 lost approximately 20% of their litters whilst corresponding animals in cages of 1350 cm2 suffered an 80% offspring mortality. A reversal of cage conditions for the same breeding females for the next litters simply reversed the mortality figures so it was possible to obtain high or low infantile mortality by placing the female in a large or small cage. Assessments of maternal behaviour were also conducted and the results showed that females with smaller breeding areas produced the highest maternal scores. The offspring mortality in large cages was invariably due to a failure of the female to nurse the young and mammary development was inhibited. However, post-weaning growth was not affected by floor area. The results therefore suggested that the effect of floor area was maximal during the parturient and nursing periods. Furthermore, cross-fostering of young at birth showed that the viability of the young was unaffected by conditions in utero. Such data therefore emphasize the importance of space and illustrate the sensitivity of the dam during the post-partum period. It is important to note that the mouse is less productive when available space is increased which contradicts the opinion that restricted space leads to a reduction in performance. A possible interpretation of these results suggests there may be an optimum spatial requirement and increases or decreases lead to changes in reproductive efficiency. Although only spatial effects were studied, there are now indications that when group size is introduced as another variable, interaction effects begin to emerge. In the dairy industry the resultant effect of increasing the size of the herd and reducing floor area per cow has been a suppression of cestrus. Work in the UK and USA suggests that optimum group size is between 70 and 100 cows at present day floor area allowances of 80-100 sq ft (7.4-9.3 m2) per cow. Some herds exceeding these numbers during the housed winter period have reported a decline in behavioural aestrus, which means that artificial insemination is delayed and there is a subsequent loss of milk production. Other work at Bristol with battery housed hens shows that they too reflect the effects of group size and floor area allowances by altering not only egg output but also the composition of the egg.
This paper has attempted to give some indication of the attitudes of man towards stress in domestic animals. Man's attitude and ability is important because he has to decide whether an animal is stressed and we still rely heavily on subjective measures of assessment as knowledge of behaviour patterns under different management systems is very inadequate, although information is now coming forward. Much additional work is now in progress featuring the adrenal gland and corticosteroid levels although in terms of an onthe-spot appraisal of stress the behaviour and/or the reproductive performance of the animal is and probably will remain for a considerable time the most important means of assessment of stress in domestic animals. REFERENCES Arave C W & Albright J L (1974) Proceedings of the Indiana Academy of Science 83 Brittain P (1973) PhD Thesis, Bristol Departmental Committee on Experiments on Animals (1965) Report (Chairman: Sir Sidney Littlewood). Home Office, London; Cmnd 2641 Technical Committee to Enquire into the Welfare of Animals Kept under Intensive Livestock Husbandry Systems (1965) Report (Chairman: F W R Brambell). HMSO, London; Cmnd 2836
Dr R T Wilkinson (MRC Applied Psychology Unit Annexe, Cambridge, CB2 2B W)
Performance Under Stress Think about the idea of stress for a moment and one begins to feel that there are as many meanings to it as there are contexts to consider it in. In the context of performance it has proved useful to think in terms of high and low arousal stress: performance may be worse either because a man is too anxious, overloaded, or afraid, or because what he has to do is too little or too uninteresting to help him maintain alertness. The stress of under-arousal is reasonably well documented (e.g. Wilkinson 1965), that of over-arousal less so, mainly because it is a difficult topic to study experimentally. Nevertheless, researchers have indeed succeeded in studying the performance of laboratory-type tasks in field settings of high anxiety and considerable danger, for example before a maiden parachute jump (Hammerton & Tickner 1968) and during open sea diving (Baddeley 1972). Performance was impaired. due possibly to some distraction of attention from the test in face of more pressing needs of selfpreservation. What happens when the performance of the test is itself made relevant to avoiding danger? Only one experiment appears to have addressed the question successfully:
Proc. roy. Soc. Med. Volume 68 July 1975
24
Berkun et al. (1962) required an isolated soldier to make a radio call for help during an army exercise in order to extricate himself or (in another experiment) a comrade from simulated danger. The man finds his radio transmitter has 'failed'. Under this pressure he has to repair it, and this constitutes the performance test. Performance was again impaired, this time presumably due to the high arousalper se. In this situation the men were under observation as they attempted to cope. Two features of their behaviour were particularly noticeable: first, they often showed perseverative behaviour, concentrating too narrowly on one aspect of the problem and ignoring other peripheral information which could have made their task easier. Second, they showed an abnormal variability in their approach to the task when compared with controls who were in no danger. The first feature, perseverative behaviour, calls to mind work on the breadth of attention under stress, which has been carried out since the mid1950s by Cambridge psychologists, Bursill (1958) and later Hockey (1970, 1971). When subjects were put in an anxiety-provoking situation (high temperature or low frequency, gut-shaking noise at 100 dB) they tended to concentrate more on what was perceived to be the more important of two tasks in a double task situation; or the most important channel of information where a number of channels were to be dealt with. When sleep had been lost, which tends to lower arousal and reduce people's anxiety to do well, there were signs of the reverse happening. There was difficulty in concentrating on a central task and a greater tendency to be over-distracted by less important peripheral information. Thus as a person departs from optimal levels towards high or low arousal he may, respectively, tend to take in information too narrowly, or too diffusely, for optimal performance. This idea of the narrowing of attention with high arousal agrees well with the incidental observations of 'peripheral blindness' in the attempts of soldiers to mend their transmitters in the combat danger experiments. The second incidental observation from the 'danger' studies was an increased variability in performance. Bills (1931) observed that people may suffer 'blocks' in responding after working under stress for some time; for example an overanxious speaker may stop, at loss for a quite obvious word. Such involuntary pauses in information processing will cause unevenness in the rhythm of work, or errors where an automatic task calls for a steady stream of responses. The choice serial reaction time test is one of the best for revealing these phenomena. It was used by Bills in his original studies of 'blocks' and also during extensive work on environmental stress
by the Cambridge laboratory (Wilkinson 1969, Poulton 1970). The subject taps one of five metal discs with a stylus in response to the illumination of any one of five light bulbs. Responding to one bulb, whether on the right or wrong disc, puts it out and brings another on in random fashion, to which the subject again responds, and so on. The test may last from one minute, if the subject's time is at a premium, to 30 minutes if extreme sensitivity to stress is sought. We record each reaction time and examine how the speed, accuracy and variability of responding changes with time and in the presence of stressors of various kinds, such as noise, loss of sleep, alcohol, lack of incentive, and so on. To give just one example from this work, both loss of sleep for one night and working in loud noise at 100 dB will increase errors. Yet when men work with loss of sleep and in noise the errors are fewer than with either singly. We think that the errors are due to over-arousal with noise and under-arousal with loss of sleep. Combine the stresses and transactions take place nearer the centre of the arousal continuum where performance is better. Thus the choice serial reaction time test is a very sensitive indicator of stress due to either low or high arousal. Unfortunately the machine used in our laboratory studies has been an old and venerable one designed in the 1950s. It weighs over 60 lb (27 kg), occupies 10 cubic feet (0.28 i3), and requires mains electricity. Fig 1 illustrates a portable equivalent (Wilkinson & Houghton 1975) which we have built into a well-known portable, battery-powered, cassette tape recorder. The subject responds on the four, square, press buttons to the onset of any one of the four lights set in a square format just above the buttons.
426
rc,k\ .oj
Fig 1 Portable four-choice reaction time recorder. A, on-offswitch; B, tape transport lever; c, button which puts a marker on tape to indicate start and end of recorded test data
427
25
Section of Comparative Medicine
'Correct' and 'error' responses are recorded on the tape cassette, which can be returned to the laboratory for analysis. Portable devices of this kind are an essential tool in our future attack on the problem of stress in its many forms. Our studies so far have underlined the importance of time in relation to the effects of stress. Both for high arousal stresses like noise and low arousal ones like loss of sleep some time has to be spent at work if significant effects are to appear. Unfortunately the available time scale of the laboratory must be a short one due to the availability of subjects. This, coupled with the difficulty of imposing a genuinely high arousal stress in the laboratory, argues the need to study stress in natural situations of life and work. Thus portable performance test devices are needed which can be used for brief test runs day in, day out by subjects in their domestic or job environments, where the stresses of every day life and work can have free and natural play. Such devices must be designed both to administer the test and store a permanent record of the results in the absence of the experimenter and with a minimum of preliminary instruction to the subject on how to use it. Similar methods need to be developed for taking portable physiological and biochemical measures in the home or workplace, for correlation with the performance measures. On the performance side we at Cambridge are developing a range of such portable devices covering a number of human abilities, for example, reaction time, short-term memory, vigilance, tracking, and calculation. On the physiological side similar moves towards portability of measurement have been under way for some time (Cashman & Stott 1974, Wolff et al. 1967). We hope that field research using such tools may ultimately enable us to provide more objective methods of contributing to the diagnosis of undue stress in people, thus relieving the general practitioner of the burden of making difficult decisions based on information which is inevitably limited by the time he can devote to each patient.
Wilkinson R T (1965) In: The Physiology of Human Survival. Ed. 0 G Edholm & A L Bacharach. Academic Press, New York (1969) Psychological Bulletin 72, 260-272 Wilkinson R T & Houghton D (1975) Behaviour Research Methods and Instrumentation (in press) Wolff H S, Baker J A & Humphrey S J E (1967) Journal ofPhysiology 188, 4P
REFERENCES
Baddeley A D (1972) British Journal ofPsychology 63, 537-546 Berkun M M, Bialek H M, Kern R P & Yagi K (1962) Psychological Monographs 76, No. 15 Bills A G (1931) American Journal ofPsychology 43, 230-245 Bursill A E (1958) Quarterly Journal ofExperimental Psychology 10, 1 13-129 Cashman P M M & Stott F D (1974) Biomedical Engineering 9, 54-57 Hammerton M & Tickner A H (1968) Ergonomics 12, 851-855 Hockey G R J (1970) Quarterly Journal of Experimental Psychology 22, 28-36 (1971) British Journal ofPsychology 61, 473-480 Poulton E C (1970) Environment and Human Efficiency. Thomas, Springfield, I11.
Dr B M Freeman (Houghton Poultry Research Station, Houghton, Huntingdon, PEJ7 2DA)
Physiological Basis of Stress It is essential that the cells of an organism are maintained in a relatively constant environment if life is to be sustained. This notion has been admirably expressed by Claude Bernard in his famous aphorism 'la fixite du milieu interieur est la condition de la vie libre'. It follows that mechanisms must exist for the maintenance of this required constancy of the cellular environment. However, when conditions become particularly hostile these mechanisms are likely to fail and death will result unless the organism has available a mechanism of 'last resort' such as movement, migration, encystment or hibernation. In the past fifty years or so, largely as a result of the work of Cannon (1929) and Selye (1950), the physiological responses of the whole animal to hostile environments - that is conditions likely to upset or break down cellular homeostasis - have been elucidated. It was Cannon who characterized the immediate response, the fight or flight syndrome, while Selye extended our understanding of the phenomenon by showing that if the adverse stimulation persists then further responses can be distinguished. These are firstly directed towards adapting the organism to the new conditions - the resistance stage ofthe general adaptation syndrome - but if this proves impossible then the organism becomes physiologically exhausted and death supervenes. Selye has further shown that, no matter the type of stimulus, the physiological responses of the animal are largely stereotyped. The adverse stimuli are now conveniently termed 'stressors'; it follows that the response of the animal should therefore be termed the 'stress response'. It is now generally accepted that the adrenal gland plays a key role in regulating the response of an animal to stressors. The adrenal medulla is largely, but not exclusively, concerned with the immediate response while the adrenal cortex is concerned with the longer term. In an abnormal situation, therefore, it is characteristic for the adrenal medulla to release adrenaline or noradrenaline, or both, into the circulation. The responses to these catecholamines include in-
23
Section ofComparative Medicine
adjusted according to the number of mice in each group. At 8 weeks the heaviest male and female were selected to enter the next generation. A cyclical mating programme was used to minimize in-breeding effects. Early results showed that breeding floor areas of 600 cm2 or more impaired infantile growth and survival and it was shown that survival was directly related to cage size. Female mice bred in cages of 150 cm2 lost approximately 20% of their litters whilst corresponding animals in cages of 1350 cm2 suffered an 80% offspring mortality. A reversal of cage conditions for the same breeding females for the next litters simply reversed the mortality figures so it was possible to obtain high or low infantile mortality by placing the female in a large or small cage. Assessments of maternal behaviour were also conducted and the results showed that females with smaller breeding areas produced the highest maternal scores. The offspring mortality in large cages was invariably due to a failure of the female to nurse the young and mammary development was inhibited. However, post-weaning growth was not affected by floor area. The results therefore suggested that the effect of floor area was maximal during the parturient and nursing periods. Furthermore, cross-fostering of young at birth showed that the viability of the young was unaffected by conditions in utero. Such data therefore emphasize the importance of space and illustrate the sensitivity of the dam during the post-partum period. It is important to note that the mouse is less productive when available space is increased which contradicts the opinion that restricted space leads to a reduction in performance. A possible interpretation of these results suggests there may be an optimum spatial requirement and increases or decreases lead to changes in reproductive efficiency. Although only spatial effects were studied, there are now indications that when group size is introduced as another variable, interaction effects begin to emerge. In the dairy industry the resultant effect of increasing the size of the herd and reducing floor area per cow has been a suppression of cestrus. Work in the UK and USA suggests that optimum group size is between 70 and 100 cows at present day floor area allowances of 80-100 sq ft (7.4-9.3 m2) per cow. Some herds exceeding these numbers during the housed winter period have reported a decline in behavioural aestrus, which means that artificial insemination is delayed and there is a subsequent loss of milk production. Other work at Bristol with battery housed hens shows that they too reflect the effects of group size and floor area allowances by altering not only egg output but also the composition of the egg.
This paper has attempted to give some indication of the attitudes of man towards stress in domestic animals. Man's attitude and ability is important because he has to decide whether an animal is stressed and we still rely heavily on subjective measures of assessment as knowledge of behaviour patterns under different management systems is very inadequate, although information is now coming forward. Much additional work is now in progress featuring the adrenal gland and corticosteroid levels although in terms of an onthe-spot appraisal of stress the behaviour and/or the reproductive performance of the animal is and probably will remain for a considerable time the most important means of assessment of stress in domestic animals. REFERENCES Arave C W & Albright J L (1974) Proceedings of the Indiana Academy of Science 83 Brittain P (1973) PhD Thesis, Bristol Departmental Committee on Experiments on Animals (1965) Report (Chairman: Sir Sidney Littlewood). Home Office, London; Cmnd 2641 Technical Committee to Enquire into the Welfare of Animals Kept under Intensive Livestock Husbandry Systems (1965) Report (Chairman: F W R Brambell). HMSO, London; Cmnd 2836
Dr R T Wilkinson (MRC Applied Psychology Unit Annexe, Cambridge, CB2 2B W)
Performance Under Stress Think about the idea of stress for a moment and one begins to feel that there are as many meanings to it as there are contexts to consider it in. In the context of performance it has proved useful to think in terms of high and low arousal stress: performance may be worse either because a man is too anxious, overloaded, or afraid, or because what he has to do is too little or too uninteresting to help him maintain alertness. The stress of under-arousal is reasonably well documented (e.g. Wilkinson 1965), that of over-arousal less so, mainly because it is a difficult topic to study experimentally. Nevertheless, researchers have indeed succeeded in studying the performance of laboratory-type tasks in field settings of high anxiety and considerable danger, for example before a maiden parachute jump (Hammerton & Tickner 1968) and during open sea diving (Baddeley 1972). Performance was impaired. due possibly to some distraction of attention from the test in face of more pressing needs of selfpreservation. What happens when the performance of the test is itself made relevant to avoiding danger? Only one experiment appears to have addressed the question successfully:
Proc. roy. Soc. Med. Volume 68 July 1975
24
Berkun et al. (1962) required an isolated soldier to make a radio call for help during an army exercise in order to extricate himself or (in another experiment) a comrade from simulated danger. The man finds his radio transmitter has 'failed'. Under this pressure he has to repair it, and this constitutes the performance test. Performance was again impaired, this time presumably due to the high arousalper se. In this situation the men were under observation as they attempted to cope. Two features of their behaviour were particularly noticeable: first, they often showed perseverative behaviour, concentrating too narrowly on one aspect of the problem and ignoring other peripheral information which could have made their task easier. Second, they showed an abnormal variability in their approach to the task when compared with controls who were in no danger. The first feature, perseverative behaviour, calls to mind work on the breadth of attention under stress, which has been carried out since the mid1950s by Cambridge psychologists, Bursill (1958) and later Hockey (1970, 1971). When subjects were put in an anxiety-provoking situation (high temperature or low frequency, gut-shaking noise at 100 dB) they tended to concentrate more on what was perceived to be the more important of two tasks in a double task situation; or the most important channel of information where a number of channels were to be dealt with. When sleep had been lost, which tends to lower arousal and reduce people's anxiety to do well, there were signs of the reverse happening. There was difficulty in concentrating on a central task and a greater tendency to be over-distracted by less important peripheral information. Thus as a person departs from optimal levels towards high or low arousal he may, respectively, tend to take in information too narrowly, or too diffusely, for optimal performance. This idea of the narrowing of attention with high arousal agrees well with the incidental observations of 'peripheral blindness' in the attempts of soldiers to mend their transmitters in the combat danger experiments. The second incidental observation from the 'danger' studies was an increased variability in performance. Bills (1931) observed that people may suffer 'blocks' in responding after working under stress for some time; for example an overanxious speaker may stop, at loss for a quite obvious word. Such involuntary pauses in information processing will cause unevenness in the rhythm of work, or errors where an automatic task calls for a steady stream of responses. The choice serial reaction time test is one of the best for revealing these phenomena. It was used by Bills in his original studies of 'blocks' and also during extensive work on environmental stress
by the Cambridge laboratory (Wilkinson 1969, Poulton 1970). The subject taps one of five metal discs with a stylus in response to the illumination of any one of five light bulbs. Responding to one bulb, whether on the right or wrong disc, puts it out and brings another on in random fashion, to which the subject again responds, and so on. The test may last from one minute, if the subject's time is at a premium, to 30 minutes if extreme sensitivity to stress is sought. We record each reaction time and examine how the speed, accuracy and variability of responding changes with time and in the presence of stressors of various kinds, such as noise, loss of sleep, alcohol, lack of incentive, and so on. To give just one example from this work, both loss of sleep for one night and working in loud noise at 100 dB will increase errors. Yet when men work with loss of sleep and in noise the errors are fewer than with either singly. We think that the errors are due to over-arousal with noise and under-arousal with loss of sleep. Combine the stresses and transactions take place nearer the centre of the arousal continuum where performance is better. Thus the choice serial reaction time test is a very sensitive indicator of stress due to either low or high arousal. Unfortunately the machine used in our laboratory studies has been an old and venerable one designed in the 1950s. It weighs over 60 lb (27 kg), occupies 10 cubic feet (0.28 i3), and requires mains electricity. Fig 1 illustrates a portable equivalent (Wilkinson & Houghton 1975) which we have built into a well-known portable, battery-powered, cassette tape recorder. The subject responds on the four, square, press buttons to the onset of any one of the four lights set in a square format just above the buttons.
426
rc,k\ .oj
Fig 1 Portable four-choice reaction time recorder. A, on-offswitch; B, tape transport lever; c, button which puts a marker on tape to indicate start and end of recorded test data
427
25
Section of Comparative Medicine
'Correct' and 'error' responses are recorded on the tape cassette, which can be returned to the laboratory for analysis. Portable devices of this kind are an essential tool in our future attack on the problem of stress in its many forms. Our studies so far have underlined the importance of time in relation to the effects of stress. Both for high arousal stresses like noise and low arousal ones like loss of sleep some time has to be spent at work if significant effects are to appear. Unfortunately the available time scale of the laboratory must be a short one due to the availability of subjects. This, coupled with the difficulty of imposing a genuinely high arousal stress in the laboratory, argues the need to study stress in natural situations of life and work. Thus portable performance test devices are needed which can be used for brief test runs day in, day out by subjects in their domestic or job environments, where the stresses of every day life and work can have free and natural play. Such devices must be designed both to administer the test and store a permanent record of the results in the absence of the experimenter and with a minimum of preliminary instruction to the subject on how to use it. Similar methods need to be developed for taking portable physiological and biochemical measures in the home or workplace, for correlation with the performance measures. On the performance side we at Cambridge are developing a range of such portable devices covering a number of human abilities, for example, reaction time, short-term memory, vigilance, tracking, and calculation. On the physiological side similar moves towards portability of measurement have been under way for some time (Cashman & Stott 1974, Wolff et al. 1967). We hope that field research using such tools may ultimately enable us to provide more objective methods of contributing to the diagnosis of undue stress in people, thus relieving the general practitioner of the burden of making difficult decisions based on information which is inevitably limited by the time he can devote to each patient.
Wilkinson R T (1965) In: The Physiology of Human Survival. Ed. 0 G Edholm & A L Bacharach. Academic Press, New York (1969) Psychological Bulletin 72, 260-272 Wilkinson R T & Houghton D (1975) Behaviour Research Methods and Instrumentation (in press) Wolff H S, Baker J A & Humphrey S J E (1967) Journal ofPhysiology 188, 4P
REFERENCES
Baddeley A D (1972) British Journal ofPsychology 63, 537-546 Berkun M M, Bialek H M, Kern R P & Yagi K (1962) Psychological Monographs 76, No. 15 Bills A G (1931) American Journal ofPsychology 43, 230-245 Bursill A E (1958) Quarterly Journal ofExperimental Psychology 10, 1 13-129 Cashman P M M & Stott F D (1974) Biomedical Engineering 9, 54-57 Hammerton M & Tickner A H (1968) Ergonomics 12, 851-855 Hockey G R J (1970) Quarterly Journal of Experimental Psychology 22, 28-36 (1971) British Journal ofPsychology 61, 473-480 Poulton E C (1970) Environment and Human Efficiency. Thomas, Springfield, I11.
Dr B M Freeman (Houghton Poultry Research Station, Houghton, Huntingdon, PEJ7 2DA)
Physiological Basis of Stress It is essential that the cells of an organism are maintained in a relatively constant environment if life is to be sustained. This notion has been admirably expressed by Claude Bernard in his famous aphorism 'la fixite du milieu interieur est la condition de la vie libre'. It follows that mechanisms must exist for the maintenance of this required constancy of the cellular environment. However, when conditions become particularly hostile these mechanisms are likely to fail and death will result unless the organism has available a mechanism of 'last resort' such as movement, migration, encystment or hibernation. In the past fifty years or so, largely as a result of the work of Cannon (1929) and Selye (1950), the physiological responses of the whole animal to hostile environments - that is conditions likely to upset or break down cellular homeostasis - have been elucidated. It was Cannon who characterized the immediate response, the fight or flight syndrome, while Selye extended our understanding of the phenomenon by showing that if the adverse stimulation persists then further responses can be distinguished. These are firstly directed towards adapting the organism to the new conditions - the resistance stage ofthe general adaptation syndrome - but if this proves impossible then the organism becomes physiologically exhausted and death supervenes. Selye has further shown that, no matter the type of stimulus, the physiological responses of the animal are largely stereotyped. The adverse stimuli are now conveniently termed 'stressors'; it follows that the response of the animal should therefore be termed the 'stress response'. It is now generally accepted that the adrenal gland plays a key role in regulating the response of an animal to stressors. The adrenal medulla is largely, but not exclusively, concerned with the immediate response while the adrenal cortex is concerned with the longer term. In an abnormal situation, therefore, it is characteristic for the adrenal medulla to release adrenaline or noradrenaline, or both, into the circulation. The responses to these catecholamines include in-
