- Email: [email protected]
Reorganization of behaviour in laboratory mice,Mus musculus, with varying cost of access to resources
Anim. Behav., 1996, 51, 1087–1093
Reorganization of behaviour in laboratory mice, Mus musculus, with varying cost of access to resources C. M. SHERWIN & C. J. NICOL Division of Animal Health and Husbandry, Department of Clinical Veterinary Science, University of Bristol, U.K. (Received 12 June 1995; initial acceptance 19 July 1995; final acceptance 11 September 1995; MS. number: 4950)
Abstract. Gaining access to resources can be measured in terms of the time spent with, or the number of visits to, a resource. For some resources, time spent with the resource might be more important, for others the number of visits might be more meaningful. By using traverses of shallow water, the costs of gaining access to food, shelter, a conspecific, increased space, a running wheel, deep sawdust, or enrichments (e.g. balls, a variety of small objects) were increased for laboratory mice. When 30 cm of water was present, the number of visits to each resource decreased to 39–64% of the number recorded when no water was present, but the proportion of time spent with each resource was defended. Increasing the width of water to 120 cm had no further effect on the number of visits or on the proportion of time spent with each resource: thus both the frequency of visits and the proportion of time with each resource were ultimately defended. Possible reasons for this change in behavioural organization, including the importance of patrolling, are discussed. The data support previous findings that laboratory mice are highly motivated to patrol areas made accessible to them, and suggest that care is needed when interpreting what animals perceive as reinforcement during visits to resources. Furthermore, it was shown that for laboratory mice under conditions that penalize initial access only, the time spent with a wide variety of resources was defended more diligently than the number of visits, but the number of visits was rarely allowed to decrease to zero. ?
At any given moment, animals might be motivated to gain access to several resources. If constraints are placed on visiting resources (e.g. limited time, limited energy, increased risk of injury), the animals must organize their behaviour to visit in a manner that accrues them the greatest benefit and/or the least cost. Understanding how animals organize such activities is necessary for both fundamental and applied studies of animal behaviour. In fundamental studies, how behaviour is organized and given priority has received much attention. However, despite many theoretical models (e.g. McFarland 1977; Hursh 1980; McNamara & Houston 1986; Stephens & Krebs 1986), behavioural responses have been examined largely in the presence of only one or two Correspondence: C. M. Sherwin, Division of Animal Health and Husbandry, Department of Clinical Veterinary Science, University of Bristol, Langford, Bristol BS18 7DU, U.K. (email: [email protected]). 0003–3472/96/051087+07 $18.00/0
?
1996 The Association for the Study of Animal Behaviour
resources (e.g. Bolles 1961; McFarland 1965; Roper 1975; Johnson & Cabanac 1982; Sherwin & Nicol 1995). In applied studies, understanding how animals organize visits to resources allows us to improve welfare by designing environments to include those resources that the animals rank as ‘important’ or essential. The importance of visiting areas of a complex environment, as perceived by an animal, can be assessed by determining the extent to which visits are made when the cost of these is increased. If visits are essential, they will prevail despite an increased cost, that is visits will be defended. If visits are unimportant, they will not be defended and will feature less prominently in the repertoire, or perhaps disappear (Wilson 1975; Houston & McFarland 1980; McFarland & Houston 1981; Dawkins 1988; McFarland 1993). When visiting some resources, the length of time spent with the resource will be important; but, if the cost incurred whilst making a visit deters the animal 1996 The Association for the Study of Animal Behaviour
1087
1088
Animal Behaviour, 51, 5
from gaining access to the resources, less time is then available to spend with the resources and time becomes an indirect constraint on behaviour. Therefore, the animal should defend the time spent with resources only if it perceives these as important. Under some circumstances, however, the frequency of visits might be more meaningful to the animal. If patrolling is the reason for visiting an area, for example monitoring for intruders, mates or new food sources, visits might occupy a small amount of time but should be conducted frequently if environmental changes are transitory or if rapid detection is important. Therefore, we can state three possible responses predicting how animals might organize behaviour when the cost of visiting resources is increased. Response 1: number of visits is important (defended)/time spent is not important (not defended and bout length is decreased). Response 2: number of visits is not important (not defended)/time spent is important (defended by compensatory increases in bout length). Response 3: number of visits is important/time spent is important (both are defended). We are aware of only one study (Collier et al. 1990) to report both the frequency and duration of visits in a multiple-resource environment where the cost of gaining access was variable. Collier et al. reported that as the costs of gaining access to food, water, a nest and a running wheel were increased, the frequency of visits by rats, Rattus norvegicus, to the resources decreased. For eating and drinking, this was compensated by increases in visit length, but the lengths of visits to the nest and wheel did not significantly increase. This study used a small number of animals and limited range of resources. It is appropriate to investigate the generality of these responses. It is widely presumed that the cost or benefit associated with gaining access to a resource is realized in terms of a decrease or increase (respectively) in the fitness of the animal. However, the cost as perceived by the animal is likely to be a variable with more proximate consequences such as energy depletion, increased risk of drowning or injury (Dawkins 1990; Sherwin & Nicol 1995). We have previously shown (Sherwin & Nicol 1995) that if mice had to traverse shallow water to gain access to food, consumption was maintained but the number of visits decreased, indicating the mice perceived water as a (potential) cost. In the present study, we used shallow water as a natural
Food Resource tunnel
One-way door
Mouse
Shelter
Central cage
Space Figure 1. Experimental apparatus for experiment 1 (see text for description of resources present in experiment 2). Each resource tunnel contained four 30-cm sections which could be independently filled with water to a depth of 2 cm.
cost. This cost was incurred directly whilst making the traverse, and then indirectly as a time constraint when the mice avoided incurring the cost by not visiting resources. The apparatus therefore allowed us to determine how mice organized both the frequency and duration of visits to various resources when the cost of gaining access was varied.
METHODS Apparatus and Housing The apparatus comprised four resource cages each connected to a single central cage by two separate plastic tunnels (Fig. 1). All cages, including the floors, were made of transparent Perspex with wire mesh tops and contained a water bottle. All had the same floor dimensions of 19#19 cm (except the ‘space’ cage in experiment 1 which measured 38#38 cm) and were 12 cm high (except the ‘sawdust’ and ‘wheel’ cages in experiment 2 which were 25 cm high). The mice were housed in the same room maintained at 21)C and on a 12:12 h light:dark cycle. The food used throughout this study was standard laboratory mouse feed (Bantin & Kingman, U.K.). To move from the central cage to a resource cage, the mouse had to move through 136 cm of transparent yellow, plastic tunnel 5 cm in diameter (‘Habitrail’). Each tunnel contained four
Sherwin & Nicol: Organization of behaviour by mice 30-cm traverses, created by gluing semi-circles of Perspex to the bottom half of the tunnel. These allowed each section of each tunnel to be independently filled with water. Once in a resource cage, the mouse could return to the central cage by either retracing its route through the tunnel, or moving via a second tunnel incorporating a oneway exit door. This door comprised a circle of transparent Perspex, hinged half-way up its height so that the mouse could push under the door. Under normal circumstances, the door was locked by an electromagnet. Infra-red sensors were located at the ends of each tunnel and on the central-cage side of the one-way door to detect movement of the mouse through the tunnels. The sensors and the electromagnet were controlled by an Acorn computer linked to an Arachnid laboratory interface (Paul Fray Ltd, Oxford, U.K.) which automatically recorded the movements of the mouse about the apparatus. When the mouse moved into a resource cage, the appropriate oneway exit door was unlocked. Once the mouse had moved through the door towards the central cage, the electromagnet was energized thus locking the exit door. This prevented the mouse gaining access to the resource cage from the central cage without traversing the water tunnel, but also allowed the mouse to return to the central cage without traversing the water a second time. The apparatus therefore imposed a cost on the mouse as it gained access to a resource rather than during consumption of the resource, but imposed negligible cost when the animal wished to leave the resource. Training Prior to being introduced into the apparatus, each mouse was placed individually into a training system. This comprised two standard mouse cages joined by a length of tunnel incorporating an unlocked one-way door as described above. The mice rapidly learned to move through the door. This required little apparent effort or aversion, and within 2 days all the mice were voluntarily moving through it many times (more than 30 times) each 24 h. Procedure Each mouse was placed individually in the experimental apparatus for 9 consecutive days. During this time, zero, one, or four sections of
1089
each tunnel were filled with water to a depth of 2 cm. Each mouse was exposed to each cost (0, 30 or 120 cm of water) for 3 consecutive days. The order of presentation was either ascending or descending. On days when one section was filled, this was the section nearest each resource cage. At the beginning of each 24-h period we replenished the food and when necessary, poured water into the tunnels to compensate for evaporation and spillage by the mouse. After each mouse’s set of trials, we thoroughly scrubbed the entire apparatus with disinfectant, and rotated it 90). We repeated this procedure for 11 male T/O laboratory mice in experiment 1, and seven male mice of the same strain in experiment 2. In experiment 1, one resource cage contained powdered food, one contained an overturned, opaque plastic cup (‘shelter’) 11 cm high and 8 cm wide, and one allowed visual access to an unfamiliar male mouse of the same strain housed in a separate cage (barriers prevented visual access from all other areas of the apparatus). The fourth resource cage (38#38#12 cm) offered four times the floor space of the other cages, but contained no other additional resource. In experiment 2, the resource cages contained powdered food, a 15-cm-diameter running wheel, sawdust 6–7 cm deep or two to four objects (‘enrichments’) chosen at random from a pool of two table tennis balls painted various colours, a 7-#-7-cm piece of steel mesh, a Perspex disc 5 cm in diameter, a steel washer 3 cm in diameter, a piece of iron 8 cm long bent at 90), and a plastic funnel 4 cm long. The enrichments were washed and disinfected daily, and placed randomly in the ‘enrichments’ cage at the beginning of each 24-h period. For practical reasons, it was impossible to obtain complete 24-h records of use of the resources for all mice throughout the study. In addition, records of ‘resource-use’ would be difficult to interpret and perhaps misleading. For example, if the experimental mouse entered the cage allowing visual access to another mouse but gave no obvious indication that it had observed the second mouse, this could mean that it was not seen, or that it was seen with a glance imperceptible to human observers. Statistical Analyses For each mouse, we calculated the mean number of visits per 24 h and the mean proportion
Animal Behaviour, 51, 5
1090
Table I. The proportion of time spent in resource cages and the number of visits per 24 h with no water traverse or traverses of 30 or 120 cm Time (proportion of 24 h) Type of resource Experiment 1 Food Mouse Shelter Space Centre Overall Experiment 2 Food Wheel Enrichments Sawdust Centre Overall
Number of visits per 24 h
0 cm
30 cm
120 cm
0 cm
30 cm
120 cm
0.57&0.06a 0.06&0.01b 0.06&0.05 0.09&0.01 0.23&0.06b
0.65&0.01ab 0.02&0.01a 0.05&0.03 0.05&0.01 0.23&0.05b
0.76&0.05b 0.03&0.01ab 0.05&0.01 0.08&0.02 0.09&0.01a
30.1&3.6a 23.7&2.1a 16.2&2.5a 32.4&2.8a
16.8&2.1b 15.1&2.5b 8.7&2.3b 19.5&3.6b
12.9&1.7b 12.2&2.3b 7.8&1.6b 14.9&2.3b
23.9&11.3a
11.8&0.9b
9.4&0.6b
23.4&4.3a 29.2&5.3a 19.1&5.1a 19.1&3.9a
11.1&1.5b 11.3&2.1b 8.4&1.6b 9.4&1.2b
9.5&1.2b 9.0&1.8b 5.1&1.3b 8.9&1.6b
22.7&2.3a
10.1&0.8b
8.1&0.8b
0.13&0.01 0.13&0.0 0.02&0.01 0.67&0.03 0.07&0.01
0.10&0.01 0.09&0.01 0.01&0.01 0.70&0.02 0.09&0.01
0.13&0.02 0.08&0.02 0.01&0.01 0.69&0.05 0.08&0.01
Values represent X& calculated from mean values for 11 (experiment 1) or 7 (experiment 2) mice. Significant differences (ANOVA: P<0.05) between means within resources are indicated by dissimilar superscripts.
of time spent in each resource cage per 24 h for each cost (N=3 days). We used these mean values in subsequent analyses. The times spent in the resource cages were non-independent so raw data of proportions of time spent in each resource cage were subjected to arcsine, square root transformation prior to ANOVA (Sokal & Rohlf 1981) and PLSD tests.
RESULTS Within minutes of being placed in the apparatus, the mice began moving about the tunnels in an apparently relaxed manner and were largely unresponsive to the occasional presence and noise of humans in the room. Use of the resource cages was consistent within and between individuals, as indicated by the small standard errors (Table I). The number of visits to a resource cage decreased to zero on only 14 days; this was limited to three mice for the ‘shelter’ in experiment 1 (10 days), and one mouse for visits to the ‘enrichments’ in experiment 2 (4 days). There was no obvious difference in the behaviour of the mice presented with ascending or descending orders of traverse length. When water was present in the tunnels, the mice often appeared reluctant to move through. Some individuals arched their backs and waded, others attempted to straddle the water by pushing
against the sides of the tunnel. They frequently groomed after completing the crossing. General movement about the apparatus was reduced by the presence of water in the resource tunnels. The overall number of daily visits was halved by the presence of 30 cm of water: increasing this length to 120 cm further reduced the mean number of visits, but this difference was not significant. The overall number of visits made at each length of water was similar in experiments 1 and 2, despite the different resources offered in the two experiments. The number of visits to a resource usually changed substantially only on the first day on which the length of the water traverse was changed; thereafter the number of visits was remarkably consistent. The proportion of time spent in each resource cage was not significantly decreased by 30 or 120 cm of water and was therefore completely resilient to the imposed cost. The greatest proportion of time was spent in the resource cage used for sleeping, the least in the ‘enrichments’. When 30 cm of water was present in the tunnels, the frequency of visits to each resource cage was significantly lower (ANOVA: P<0.05) in each case than the number of visits when no water was present. When the traverse of water was lengthened to 120 cm, there was no further reduction in the number of visits; rather, the frequency observed at 30 cm was defended. When no
Sherwin & Nicol: Organization of behaviour by mice water was present, the ‘space’ was visited most frequently and the ‘shelter’ the least. In experiment 1, the proportion of time spent in the food cage significantly increased as the length of water was increased. The mice frequently slept in this cage, often in food spilt or dug from the containers. This was prevented in experiment 2 by placing the food bowls on a wire mesh floor. The number of visits and the proportion of time spent in a resource cage were largely unrelated. For example, in experiment 1, the ‘space’ was visited most frequently but the amount of time spent there was only 8.4–15.0% of the time spent in the food cage. Similarly, in experiment 2, the number of visits to the ‘enrichments’ and the ‘sawdust’ when no water was present were identical; however, the mean proportion of time spent in the latter was 30 times greater. The mean time spent in the ‘centre’ ranged between 1.68 to 5.52 h/day and therefore represented a considerable proportion of the available time. In experiment 1, the proportion of time in the ‘centre’ was generally higher than in experiment 2. In addition, 120 cm water significantly reduced this proportion in experiment 1 but not in experiment 2.
DISCUSSION Access to resources can be measured in terms of the number of visits, or the length of time with a resource, and can be predicted from responses 1–3 given in the Introduction. Our results show that a 30-cm traverse of water reduced the number of visits to each of the resources compared with the number recorded when no water was present, but, the proportion of time spent in the resource cages was defended. This is consistent with response 2 and supports previous findings in which the number of visits to food decreased as the cost of visiting was increased, but the time spent feeding was maintained (presumably) by a compensatory strategy of increasing bout length (Johnson & Cabanac 1982; Collier et al. 1990; Sherwin & Nicol 1995). Our results show laboratory mice responded in this way in two experiments both providing a wide diversity of resources, several of them unrelated to maintenance activities. When the traverse of water was lengthened to 120 cm, the mice defended both the number of visits and the time spent with each resource. This
1091
is consistent with response 3, in contrast with the response to a traverse of 30 cm discussed above. There are three possible explanations for this change in behavioural organization in the present study. First, the presence of water per se might have reduced access to the resources to the absolute minimum acceptable to the mice: gaining access to the resources at 30 or 120 cm of water was unconditionally resilient and defended as the minimum required. Second, the reason for traversing the water may have become less related to the resource when the cost of visits increased; if a constant and irreducible number of monitoring or patrolling visits to each resource cage was required regardless of what the cage contained, this constant would represent a greater proportion of the visits to the resource cage when the total number of visits decreased. Ultimately, the mice might have defended patrolling visits. Third, the nature of the cost might have been ‘all-ornothing’; for example, if the cost was assessed as the extent to which the fur became wet, this is unlikely to have been influenced by increasing the length of water from 30 to 120 cm. However, subsequent work directly comparing lever pressing and a traverse of water (C. Nicol, C. Sherwin & S. Pope, unpublished data) does not support this suggestion. Although it is not possible to state unequivocally the reasons for this reorganization in the defence of visiting resources, the results show that both the number of visits and time spent in various areas of a complex environment were ultimately important to the mice. Our data support the idea that laboratory mice are highly motivated to patrol areas made accessible to them. But, it seems unlikely that monitoring activity was the sole reason for defending access to the resource cages in the present study. The data in Table I show consistent differences in the number of visits made to, and the proportion of time spent in, the various resource cages (e.g. ‘wheel’ versus ‘enrichments’ in experiment 2). These show that the mice differentiated between the resource cages, and defended use of each of them in a manner evidently related to the resource. Therefore, gaining access to the resources themselves was perceived as important by the mice. The apparent high importance of monitoring behaviour, and the disparity between number of visits to and time spent in each resource cage, have important implications for studies determining the
1092
Animal Behaviour, 51, 5
motivation to gain access to resources. In such studies (e.g. Johnson & Cabanac 1982; Collier et al. 1990; Matthews & Ladewig 1994; Sherwin & Nicol 1995) ‘motivation’ is sometimes quantified as the number of reinforcements obtained. In experiment 2 of the present study, only 1.1% of 24 h was spent in the ‘enrichments’ when 120 cm of water was present, but it was visited a similar number of times to the other resources, including the food (an undeniably essential resource). The present results indicate that a considerable understanding of what animals perceive as reinforcement is needed when motivation to gain access to resources is examined. Our results show that the mice preferred to rest in deep sawdust than in scattered food or the tunnels. In experiment 1, when 0 or 30 cm of water was present, that is when portions of the tunnels were dry and no sawdust was present, the mice spent a considerable proportion of time in the ‘centre’ which included the dry portion of the tunnels. When 120 cm of water was present, the time spent in the ‘centre’ decreased, and time spent in the ‘food’ cage (with scattered food available) increased. In experiment 2, the mice spent considerable proportions of time in the ‘sawdust’ and a relatively small proportion of time in the ‘centre’, even when no water was present in the tunnels. The proportion of time spent in areas other than the four resource cages (‘centre’) was significantly reduced by increasing the length of water from 30 to 120 cm in experiment 1. This indicates that the time spent moving between the resource cages was not defended as the cost of performing this activity was increased, but was presumably sacrificed to maintain or increase time spent with the resources. This response was not evident in experiment 2. We suggest that the probable dual function of the ‘food’ resource cage in experiment 1 (as an energy source and sleeping site) permitted the mice to be more liberal in the time they spent moving about the tunnels when the cost was low. When 120-cm water was present, the values for ‘centre’ were very similar in experiments 1 and 2. This possibly represents the minimum time necessary to allow adequate visits and patrolling activity. Alternatively, it is possible that this is ‘buffer’ time (i.e. time spent in no activity) which has been reported as a resilient behaviour in chicks, Gallus gallus domesticus (Hill et al. 1986).
Previous studies (e.g. Johnson & Cabanac 1982; Collier et al. 1990; Sherwin & Nicol 1995) have shown that several species respond to increased cost of visits to resources by reducing the frequency but defending the duration of visits. However, the wide variety of resources for which this response can occur has not been shown before. Using a variable length of water traverse, we have shown that laboratory mice were highly motivated to visit all accessible areas of their environment to gain access to resources, or to perform patrolling behaviour. These findings have important implications for both fundamental and applied studies of animal behaviour.
ACKNOWLEDGMENTS We thank Dr C. Lindberg and two anonymous referees for helpful comments on an early version of the manuscript. The study was funded by the Home Office (U.K.).
REFERENCES Bolles, R. C. 1961. The interaction of hunger and thirst in the rat. J. comp. physiol. Psychol., 54, 580– 584. Collier, G. H., Johnson, D. F., CyBulski, K. A. & McHale, C. A. 1990. Activity patterns in rats (Rattus norvegicus) as a function of the cost of access to four resources. J. comp. Psychol., 1, 53–65. Dawkins, M. S. 1988. Behavioural deprivation: a central problem in animal welfare. Appl. Anim. Behav. Sci., 20, 209–225. Dawkins, M. S. 1990. From an animal’s point of view: motivation, fitness and animal welfare. Behav. Brain. Sci., 13, 1–61. Hill, W. L., Rovee-Collier, C., Collier, G. & Wasserloos, L. 1986. Time budgets in growing chicks. Physiol. Behav., 37, 353–360. Houston, A. I. & McFarland, D. 1980. Behavioural resilience and its relation to demand functions. In: Limits to Action: The Allocation of Individual Behaviour (Ed. by J. E. R. Staddon), pp. 177–203. New York: Academic Press. Hursh, S. 1980. Economic concepts for the analysis of behavior. J. exp. Analysis Behav., 34, 219–238. Johnson, K. G., & Cabanac, M. 1982. Homeostatic competition in rats fed at varying distances from a thermoneutral refuge. Physiol. Behav., 29, 715– 720. McFarland, D. J. 1965. Hunger, thirst and displacement pecking in the Barbary dove. Anim. Behav., 13, 293– 300. McFarland, D. J. 1977. Decision making in animals. Nature, Lond., 269, 15–21.
Sherwin & Nicol: Organization of behaviour by mice McFarland, D. J. 1993. Animal Behaviour. 2nd edn. Harlow, Essex: Longman Scientific and Technical. McFarland, D. J. & Houston, A. I. 1981. Quantitative Ethology: The State Space Approach. London: Pitman Publishing. McNamara, J. M. & Houston, A. I. 1986. The common currency for behavioral decisions. Am. Nat., 127, 359–378. Matthews, L. R. & Ladewig, J. 1994. Environmental requirements of pigs measured by behavioural demand functions. Anim. Behav., 47, 713–719.
1093
Roper, T. J. 1975. Nest material and food as reinforcers for fixed-ratio responding in mice. Learn. Motiv., 6, 327–343. Sherwin, C. M. & Nicol, C. J. 1995. Changes in meal patterning by mice measure the cost imposed by natural obstacles. Appl. Anim. Behav. Sci., 43, 291–300. Sokal, R. R. & Rohlf, F. J. 1981. Biometry. 2nd edn. New York: W. H. Freeman. Stephens, D. W. & Krebs, J. R. 1986. Foraging Theory. Princeton, New Jersey: Princeton University Press. Wilson, E. O. 1975. Sociobiology. Cambridge, Massachusetts: Harvard University Press.
Reorganization of behaviour in laboratory mice, Mus musculus, with varying cost of access to resources C. M. SHERWIN & C. J. NICOL Division of Animal Health and Husbandry, Department of Clinical Veterinary Science, University of Bristol, U.K. (Received 12 June 1995; initial acceptance 19 July 1995; final acceptance 11 September 1995; MS. number: 4950)
Abstract. Gaining access to resources can be measured in terms of the time spent with, or the number of visits to, a resource. For some resources, time spent with the resource might be more important, for others the number of visits might be more meaningful. By using traverses of shallow water, the costs of gaining access to food, shelter, a conspecific, increased space, a running wheel, deep sawdust, or enrichments (e.g. balls, a variety of small objects) were increased for laboratory mice. When 30 cm of water was present, the number of visits to each resource decreased to 39–64% of the number recorded when no water was present, but the proportion of time spent with each resource was defended. Increasing the width of water to 120 cm had no further effect on the number of visits or on the proportion of time spent with each resource: thus both the frequency of visits and the proportion of time with each resource were ultimately defended. Possible reasons for this change in behavioural organization, including the importance of patrolling, are discussed. The data support previous findings that laboratory mice are highly motivated to patrol areas made accessible to them, and suggest that care is needed when interpreting what animals perceive as reinforcement during visits to resources. Furthermore, it was shown that for laboratory mice under conditions that penalize initial access only, the time spent with a wide variety of resources was defended more diligently than the number of visits, but the number of visits was rarely allowed to decrease to zero. ?
At any given moment, animals might be motivated to gain access to several resources. If constraints are placed on visiting resources (e.g. limited time, limited energy, increased risk of injury), the animals must organize their behaviour to visit in a manner that accrues them the greatest benefit and/or the least cost. Understanding how animals organize such activities is necessary for both fundamental and applied studies of animal behaviour. In fundamental studies, how behaviour is organized and given priority has received much attention. However, despite many theoretical models (e.g. McFarland 1977; Hursh 1980; McNamara & Houston 1986; Stephens & Krebs 1986), behavioural responses have been examined largely in the presence of only one or two Correspondence: C. M. Sherwin, Division of Animal Health and Husbandry, Department of Clinical Veterinary Science, University of Bristol, Langford, Bristol BS18 7DU, U.K. (email: [email protected]). 0003–3472/96/051087+07 $18.00/0
?
1996 The Association for the Study of Animal Behaviour
resources (e.g. Bolles 1961; McFarland 1965; Roper 1975; Johnson & Cabanac 1982; Sherwin & Nicol 1995). In applied studies, understanding how animals organize visits to resources allows us to improve welfare by designing environments to include those resources that the animals rank as ‘important’ or essential. The importance of visiting areas of a complex environment, as perceived by an animal, can be assessed by determining the extent to which visits are made when the cost of these is increased. If visits are essential, they will prevail despite an increased cost, that is visits will be defended. If visits are unimportant, they will not be defended and will feature less prominently in the repertoire, or perhaps disappear (Wilson 1975; Houston & McFarland 1980; McFarland & Houston 1981; Dawkins 1988; McFarland 1993). When visiting some resources, the length of time spent with the resource will be important; but, if the cost incurred whilst making a visit deters the animal 1996 The Association for the Study of Animal Behaviour
1087
1088
Animal Behaviour, 51, 5
from gaining access to the resources, less time is then available to spend with the resources and time becomes an indirect constraint on behaviour. Therefore, the animal should defend the time spent with resources only if it perceives these as important. Under some circumstances, however, the frequency of visits might be more meaningful to the animal. If patrolling is the reason for visiting an area, for example monitoring for intruders, mates or new food sources, visits might occupy a small amount of time but should be conducted frequently if environmental changes are transitory or if rapid detection is important. Therefore, we can state three possible responses predicting how animals might organize behaviour when the cost of visiting resources is increased. Response 1: number of visits is important (defended)/time spent is not important (not defended and bout length is decreased). Response 2: number of visits is not important (not defended)/time spent is important (defended by compensatory increases in bout length). Response 3: number of visits is important/time spent is important (both are defended). We are aware of only one study (Collier et al. 1990) to report both the frequency and duration of visits in a multiple-resource environment where the cost of gaining access was variable. Collier et al. reported that as the costs of gaining access to food, water, a nest and a running wheel were increased, the frequency of visits by rats, Rattus norvegicus, to the resources decreased. For eating and drinking, this was compensated by increases in visit length, but the lengths of visits to the nest and wheel did not significantly increase. This study used a small number of animals and limited range of resources. It is appropriate to investigate the generality of these responses. It is widely presumed that the cost or benefit associated with gaining access to a resource is realized in terms of a decrease or increase (respectively) in the fitness of the animal. However, the cost as perceived by the animal is likely to be a variable with more proximate consequences such as energy depletion, increased risk of drowning or injury (Dawkins 1990; Sherwin & Nicol 1995). We have previously shown (Sherwin & Nicol 1995) that if mice had to traverse shallow water to gain access to food, consumption was maintained but the number of visits decreased, indicating the mice perceived water as a (potential) cost. In the present study, we used shallow water as a natural
Food Resource tunnel
One-way door
Mouse
Shelter
Central cage
Space Figure 1. Experimental apparatus for experiment 1 (see text for description of resources present in experiment 2). Each resource tunnel contained four 30-cm sections which could be independently filled with water to a depth of 2 cm.
cost. This cost was incurred directly whilst making the traverse, and then indirectly as a time constraint when the mice avoided incurring the cost by not visiting resources. The apparatus therefore allowed us to determine how mice organized both the frequency and duration of visits to various resources when the cost of gaining access was varied.
METHODS Apparatus and Housing The apparatus comprised four resource cages each connected to a single central cage by two separate plastic tunnels (Fig. 1). All cages, including the floors, were made of transparent Perspex with wire mesh tops and contained a water bottle. All had the same floor dimensions of 19#19 cm (except the ‘space’ cage in experiment 1 which measured 38#38 cm) and were 12 cm high (except the ‘sawdust’ and ‘wheel’ cages in experiment 2 which were 25 cm high). The mice were housed in the same room maintained at 21)C and on a 12:12 h light:dark cycle. The food used throughout this study was standard laboratory mouse feed (Bantin & Kingman, U.K.). To move from the central cage to a resource cage, the mouse had to move through 136 cm of transparent yellow, plastic tunnel 5 cm in diameter (‘Habitrail’). Each tunnel contained four
Sherwin & Nicol: Organization of behaviour by mice 30-cm traverses, created by gluing semi-circles of Perspex to the bottom half of the tunnel. These allowed each section of each tunnel to be independently filled with water. Once in a resource cage, the mouse could return to the central cage by either retracing its route through the tunnel, or moving via a second tunnel incorporating a oneway exit door. This door comprised a circle of transparent Perspex, hinged half-way up its height so that the mouse could push under the door. Under normal circumstances, the door was locked by an electromagnet. Infra-red sensors were located at the ends of each tunnel and on the central-cage side of the one-way door to detect movement of the mouse through the tunnels. The sensors and the electromagnet were controlled by an Acorn computer linked to an Arachnid laboratory interface (Paul Fray Ltd, Oxford, U.K.) which automatically recorded the movements of the mouse about the apparatus. When the mouse moved into a resource cage, the appropriate oneway exit door was unlocked. Once the mouse had moved through the door towards the central cage, the electromagnet was energized thus locking the exit door. This prevented the mouse gaining access to the resource cage from the central cage without traversing the water tunnel, but also allowed the mouse to return to the central cage without traversing the water a second time. The apparatus therefore imposed a cost on the mouse as it gained access to a resource rather than during consumption of the resource, but imposed negligible cost when the animal wished to leave the resource. Training Prior to being introduced into the apparatus, each mouse was placed individually into a training system. This comprised two standard mouse cages joined by a length of tunnel incorporating an unlocked one-way door as described above. The mice rapidly learned to move through the door. This required little apparent effort or aversion, and within 2 days all the mice were voluntarily moving through it many times (more than 30 times) each 24 h. Procedure Each mouse was placed individually in the experimental apparatus for 9 consecutive days. During this time, zero, one, or four sections of
1089
each tunnel were filled with water to a depth of 2 cm. Each mouse was exposed to each cost (0, 30 or 120 cm of water) for 3 consecutive days. The order of presentation was either ascending or descending. On days when one section was filled, this was the section nearest each resource cage. At the beginning of each 24-h period we replenished the food and when necessary, poured water into the tunnels to compensate for evaporation and spillage by the mouse. After each mouse’s set of trials, we thoroughly scrubbed the entire apparatus with disinfectant, and rotated it 90). We repeated this procedure for 11 male T/O laboratory mice in experiment 1, and seven male mice of the same strain in experiment 2. In experiment 1, one resource cage contained powdered food, one contained an overturned, opaque plastic cup (‘shelter’) 11 cm high and 8 cm wide, and one allowed visual access to an unfamiliar male mouse of the same strain housed in a separate cage (barriers prevented visual access from all other areas of the apparatus). The fourth resource cage (38#38#12 cm) offered four times the floor space of the other cages, but contained no other additional resource. In experiment 2, the resource cages contained powdered food, a 15-cm-diameter running wheel, sawdust 6–7 cm deep or two to four objects (‘enrichments’) chosen at random from a pool of two table tennis balls painted various colours, a 7-#-7-cm piece of steel mesh, a Perspex disc 5 cm in diameter, a steel washer 3 cm in diameter, a piece of iron 8 cm long bent at 90), and a plastic funnel 4 cm long. The enrichments were washed and disinfected daily, and placed randomly in the ‘enrichments’ cage at the beginning of each 24-h period. For practical reasons, it was impossible to obtain complete 24-h records of use of the resources for all mice throughout the study. In addition, records of ‘resource-use’ would be difficult to interpret and perhaps misleading. For example, if the experimental mouse entered the cage allowing visual access to another mouse but gave no obvious indication that it had observed the second mouse, this could mean that it was not seen, or that it was seen with a glance imperceptible to human observers. Statistical Analyses For each mouse, we calculated the mean number of visits per 24 h and the mean proportion
Animal Behaviour, 51, 5
1090
Table I. The proportion of time spent in resource cages and the number of visits per 24 h with no water traverse or traverses of 30 or 120 cm Time (proportion of 24 h) Type of resource Experiment 1 Food Mouse Shelter Space Centre Overall Experiment 2 Food Wheel Enrichments Sawdust Centre Overall
Number of visits per 24 h
0 cm
30 cm
120 cm
0 cm
30 cm
120 cm
0.57&0.06a 0.06&0.01b 0.06&0.05 0.09&0.01 0.23&0.06b
0.65&0.01ab 0.02&0.01a 0.05&0.03 0.05&0.01 0.23&0.05b
0.76&0.05b 0.03&0.01ab 0.05&0.01 0.08&0.02 0.09&0.01a
30.1&3.6a 23.7&2.1a 16.2&2.5a 32.4&2.8a
16.8&2.1b 15.1&2.5b 8.7&2.3b 19.5&3.6b
12.9&1.7b 12.2&2.3b 7.8&1.6b 14.9&2.3b
23.9&11.3a
11.8&0.9b
9.4&0.6b
23.4&4.3a 29.2&5.3a 19.1&5.1a 19.1&3.9a
11.1&1.5b 11.3&2.1b 8.4&1.6b 9.4&1.2b
9.5&1.2b 9.0&1.8b 5.1&1.3b 8.9&1.6b
22.7&2.3a
10.1&0.8b
8.1&0.8b
0.13&0.01 0.13&0.0 0.02&0.01 0.67&0.03 0.07&0.01
0.10&0.01 0.09&0.01 0.01&0.01 0.70&0.02 0.09&0.01
0.13&0.02 0.08&0.02 0.01&0.01 0.69&0.05 0.08&0.01
Values represent X& calculated from mean values for 11 (experiment 1) or 7 (experiment 2) mice. Significant differences (ANOVA: P<0.05) between means within resources are indicated by dissimilar superscripts.
of time spent in each resource cage per 24 h for each cost (N=3 days). We used these mean values in subsequent analyses. The times spent in the resource cages were non-independent so raw data of proportions of time spent in each resource cage were subjected to arcsine, square root transformation prior to ANOVA (Sokal & Rohlf 1981) and PLSD tests.
RESULTS Within minutes of being placed in the apparatus, the mice began moving about the tunnels in an apparently relaxed manner and were largely unresponsive to the occasional presence and noise of humans in the room. Use of the resource cages was consistent within and between individuals, as indicated by the small standard errors (Table I). The number of visits to a resource cage decreased to zero on only 14 days; this was limited to three mice for the ‘shelter’ in experiment 1 (10 days), and one mouse for visits to the ‘enrichments’ in experiment 2 (4 days). There was no obvious difference in the behaviour of the mice presented with ascending or descending orders of traverse length. When water was present in the tunnels, the mice often appeared reluctant to move through. Some individuals arched their backs and waded, others attempted to straddle the water by pushing
against the sides of the tunnel. They frequently groomed after completing the crossing. General movement about the apparatus was reduced by the presence of water in the resource tunnels. The overall number of daily visits was halved by the presence of 30 cm of water: increasing this length to 120 cm further reduced the mean number of visits, but this difference was not significant. The overall number of visits made at each length of water was similar in experiments 1 and 2, despite the different resources offered in the two experiments. The number of visits to a resource usually changed substantially only on the first day on which the length of the water traverse was changed; thereafter the number of visits was remarkably consistent. The proportion of time spent in each resource cage was not significantly decreased by 30 or 120 cm of water and was therefore completely resilient to the imposed cost. The greatest proportion of time was spent in the resource cage used for sleeping, the least in the ‘enrichments’. When 30 cm of water was present in the tunnels, the frequency of visits to each resource cage was significantly lower (ANOVA: P<0.05) in each case than the number of visits when no water was present. When the traverse of water was lengthened to 120 cm, there was no further reduction in the number of visits; rather, the frequency observed at 30 cm was defended. When no
Sherwin & Nicol: Organization of behaviour by mice water was present, the ‘space’ was visited most frequently and the ‘shelter’ the least. In experiment 1, the proportion of time spent in the food cage significantly increased as the length of water was increased. The mice frequently slept in this cage, often in food spilt or dug from the containers. This was prevented in experiment 2 by placing the food bowls on a wire mesh floor. The number of visits and the proportion of time spent in a resource cage were largely unrelated. For example, in experiment 1, the ‘space’ was visited most frequently but the amount of time spent there was only 8.4–15.0% of the time spent in the food cage. Similarly, in experiment 2, the number of visits to the ‘enrichments’ and the ‘sawdust’ when no water was present were identical; however, the mean proportion of time spent in the latter was 30 times greater. The mean time spent in the ‘centre’ ranged between 1.68 to 5.52 h/day and therefore represented a considerable proportion of the available time. In experiment 1, the proportion of time in the ‘centre’ was generally higher than in experiment 2. In addition, 120 cm water significantly reduced this proportion in experiment 1 but not in experiment 2.
DISCUSSION Access to resources can be measured in terms of the number of visits, or the length of time with a resource, and can be predicted from responses 1–3 given in the Introduction. Our results show that a 30-cm traverse of water reduced the number of visits to each of the resources compared with the number recorded when no water was present, but, the proportion of time spent in the resource cages was defended. This is consistent with response 2 and supports previous findings in which the number of visits to food decreased as the cost of visiting was increased, but the time spent feeding was maintained (presumably) by a compensatory strategy of increasing bout length (Johnson & Cabanac 1982; Collier et al. 1990; Sherwin & Nicol 1995). Our results show laboratory mice responded in this way in two experiments both providing a wide diversity of resources, several of them unrelated to maintenance activities. When the traverse of water was lengthened to 120 cm, the mice defended both the number of visits and the time spent with each resource. This
1091
is consistent with response 3, in contrast with the response to a traverse of 30 cm discussed above. There are three possible explanations for this change in behavioural organization in the present study. First, the presence of water per se might have reduced access to the resources to the absolute minimum acceptable to the mice: gaining access to the resources at 30 or 120 cm of water was unconditionally resilient and defended as the minimum required. Second, the reason for traversing the water may have become less related to the resource when the cost of visits increased; if a constant and irreducible number of monitoring or patrolling visits to each resource cage was required regardless of what the cage contained, this constant would represent a greater proportion of the visits to the resource cage when the total number of visits decreased. Ultimately, the mice might have defended patrolling visits. Third, the nature of the cost might have been ‘all-ornothing’; for example, if the cost was assessed as the extent to which the fur became wet, this is unlikely to have been influenced by increasing the length of water from 30 to 120 cm. However, subsequent work directly comparing lever pressing and a traverse of water (C. Nicol, C. Sherwin & S. Pope, unpublished data) does not support this suggestion. Although it is not possible to state unequivocally the reasons for this reorganization in the defence of visiting resources, the results show that both the number of visits and time spent in various areas of a complex environment were ultimately important to the mice. Our data support the idea that laboratory mice are highly motivated to patrol areas made accessible to them. But, it seems unlikely that monitoring activity was the sole reason for defending access to the resource cages in the present study. The data in Table I show consistent differences in the number of visits made to, and the proportion of time spent in, the various resource cages (e.g. ‘wheel’ versus ‘enrichments’ in experiment 2). These show that the mice differentiated between the resource cages, and defended use of each of them in a manner evidently related to the resource. Therefore, gaining access to the resources themselves was perceived as important by the mice. The apparent high importance of monitoring behaviour, and the disparity between number of visits to and time spent in each resource cage, have important implications for studies determining the
1092
Animal Behaviour, 51, 5
motivation to gain access to resources. In such studies (e.g. Johnson & Cabanac 1982; Collier et al. 1990; Matthews & Ladewig 1994; Sherwin & Nicol 1995) ‘motivation’ is sometimes quantified as the number of reinforcements obtained. In experiment 2 of the present study, only 1.1% of 24 h was spent in the ‘enrichments’ when 120 cm of water was present, but it was visited a similar number of times to the other resources, including the food (an undeniably essential resource). The present results indicate that a considerable understanding of what animals perceive as reinforcement is needed when motivation to gain access to resources is examined. Our results show that the mice preferred to rest in deep sawdust than in scattered food or the tunnels. In experiment 1, when 0 or 30 cm of water was present, that is when portions of the tunnels were dry and no sawdust was present, the mice spent a considerable proportion of time in the ‘centre’ which included the dry portion of the tunnels. When 120 cm of water was present, the time spent in the ‘centre’ decreased, and time spent in the ‘food’ cage (with scattered food available) increased. In experiment 2, the mice spent considerable proportions of time in the ‘sawdust’ and a relatively small proportion of time in the ‘centre’, even when no water was present in the tunnels. The proportion of time spent in areas other than the four resource cages (‘centre’) was significantly reduced by increasing the length of water from 30 to 120 cm in experiment 1. This indicates that the time spent moving between the resource cages was not defended as the cost of performing this activity was increased, but was presumably sacrificed to maintain or increase time spent with the resources. This response was not evident in experiment 2. We suggest that the probable dual function of the ‘food’ resource cage in experiment 1 (as an energy source and sleeping site) permitted the mice to be more liberal in the time they spent moving about the tunnels when the cost was low. When 120-cm water was present, the values for ‘centre’ were very similar in experiments 1 and 2. This possibly represents the minimum time necessary to allow adequate visits and patrolling activity. Alternatively, it is possible that this is ‘buffer’ time (i.e. time spent in no activity) which has been reported as a resilient behaviour in chicks, Gallus gallus domesticus (Hill et al. 1986).
Previous studies (e.g. Johnson & Cabanac 1982; Collier et al. 1990; Sherwin & Nicol 1995) have shown that several species respond to increased cost of visits to resources by reducing the frequency but defending the duration of visits. However, the wide variety of resources for which this response can occur has not been shown before. Using a variable length of water traverse, we have shown that laboratory mice were highly motivated to visit all accessible areas of their environment to gain access to resources, or to perform patrolling behaviour. These findings have important implications for both fundamental and applied studies of animal behaviour.
ACKNOWLEDGMENTS We thank Dr C. Lindberg and two anonymous referees for helpful comments on an early version of the manuscript. The study was funded by the Home Office (U.K.).
REFERENCES Bolles, R. C. 1961. The interaction of hunger and thirst in the rat. J. comp. physiol. Psychol., 54, 580– 584. Collier, G. H., Johnson, D. F., CyBulski, K. A. & McHale, C. A. 1990. Activity patterns in rats (Rattus norvegicus) as a function of the cost of access to four resources. J. comp. Psychol., 1, 53–65. Dawkins, M. S. 1988. Behavioural deprivation: a central problem in animal welfare. Appl. Anim. Behav. Sci., 20, 209–225. Dawkins, M. S. 1990. From an animal’s point of view: motivation, fitness and animal welfare. Behav. Brain. Sci., 13, 1–61. Hill, W. L., Rovee-Collier, C., Collier, G. & Wasserloos, L. 1986. Time budgets in growing chicks. Physiol. Behav., 37, 353–360. Houston, A. I. & McFarland, D. 1980. Behavioural resilience and its relation to demand functions. In: Limits to Action: The Allocation of Individual Behaviour (Ed. by J. E. R. Staddon), pp. 177–203. New York: Academic Press. Hursh, S. 1980. Economic concepts for the analysis of behavior. J. exp. Analysis Behav., 34, 219–238. Johnson, K. G., & Cabanac, M. 1982. Homeostatic competition in rats fed at varying distances from a thermoneutral refuge. Physiol. Behav., 29, 715– 720. McFarland, D. J. 1965. Hunger, thirst and displacement pecking in the Barbary dove. Anim. Behav., 13, 293– 300. McFarland, D. J. 1977. Decision making in animals. Nature, Lond., 269, 15–21.
Sherwin & Nicol: Organization of behaviour by mice McFarland, D. J. 1993. Animal Behaviour. 2nd edn. Harlow, Essex: Longman Scientific and Technical. McFarland, D. J. & Houston, A. I. 1981. Quantitative Ethology: The State Space Approach. London: Pitman Publishing. McNamara, J. M. & Houston, A. I. 1986. The common currency for behavioral decisions. Am. Nat., 127, 359–378. Matthews, L. R. & Ladewig, J. 1994. Environmental requirements of pigs measured by behavioural demand functions. Anim. Behav., 47, 713–719.
1093
Roper, T. J. 1975. Nest material and food as reinforcers for fixed-ratio responding in mice. Learn. Motiv., 6, 327–343. Sherwin, C. M. & Nicol, C. J. 1995. Changes in meal patterning by mice measure the cost imposed by natural obstacles. Appl. Anim. Behav. Sci., 43, 291–300. Sokal, R. R. & Rohlf, F. J. 1981. Biometry. 2nd edn. New York: W. H. Freeman. Stephens, D. W. & Krebs, J. R. 1986. Foraging Theory. Princeton, New Jersey: Princeton University Press. Wilson, E. O. 1975. Sociobiology. Cambridge, Massachusetts: Harvard University Press.