ABSTRACT FEEDING PATTERN IN PEROMYSCUS MANICULATUS: THE RESPONSE TO TEMPORAL FOOD DEPRIVATION BY Michael M. Jaeger Temporal food deprivation can be imposed on Peromyscus by unfavorable meteorological conditions, particularly during periods of cold winter weather. The effects of light, temperature, humidity, wind, snow, etc. on the activity of deer mice are reviewed. Mechanisms of energy storage (caches, fat deposition, etc.) and conservation (torpor, nest building, etc.) are then considered. From this review it is hypothesized that 239. bairdi are temporally opportunistic in their nocturnal feeding and, if deprived, will concentrate their daily intake into those times when access to food is most available. 239, bairdi from the laboratory, however, showed little accomodation to a schedule of temporal food deprivation. When per- mitted only the initial six hours of their 12 hour dark activity period for feeding, five of the six animals died by Day 4. This intolerance to temporal food deprivation prompted an examination of the experi- mental conditions which affect the accomodation to food deprivation. Experiment I examined the effect of prior deprivation experi- ence on survivability under a more severe test deprivation (a 15 day period where food was present each day during the initial seven hours Michael M. Jaeger of the 12 hour dark period). It was hypothesized that laboratory Peromyscus, maintained under ad_1ibitum conditions, would show greater mortality than those with prior experience. Such was supported. Furthermore, the experienced mice consumed no more food than the con- trols. The hypothesis that greater activity (induced by access to a running wheel) reduces survivability under these conditions of temporal food deprivation was also tested. The results supported this hypoth- esis, suggesting that the nature of the deprivation experience lies, at least in part, in an inhibition of general activity or movement. Regardless of experience, 3393 bairdi demonstrated little capacity to adjust their food intake to the seven hours of availability per day. Experiment II was conducted to see if the food intake is influenced by the temporal properties of when food is accessible. Intake is greater and survival enhanced when the food is available during the final six hours of the nocturnal period as opposed to the initial six hours. This result implies a circadian regulation of food consumption. The strictly nocturnal activity of Peromyscus might necessitate an organization where late-night feeding can correct for deficit and prepare the mouse for the upcoming light phase. It is hypothesized that foraging and hoarding are most prominent early in the night and can become increased at that time in response to deprivation. FEEDING PATTERN IN PEROMYSCUS MANICULATUS: THE RESPONSE TO TEMPORAL FOOD DEPRIVATION BY 0‘» Michael M. Jaeger A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Zoology 1976 DEDICATION I would like to dedicate this to my fellow graduate students and to Jan Harper who helped make this a most enjoyable experi- ence. ii ACKNOWLEDGMENTS I am particularly grateful to Dr. Martin Balaban, my advisor, for his guidance and support of my graduate training. In addition, I very much appreciate the constructive comments of Dr. Rollin H. Baker, Dr. Glenn I. Hatton, and Dr. John H. King. Dr. King generously pro- vided space and facilities for this research. Margaret Jaeger trans- lated this work from the original language and aided in the research; for this I offer special thanks. The research was supported in part by NIH Animal Behavior Training Grant #5 TOl GMl751-05 BHS and by a teaching assistantship provided by the Department of Zoology. iii TABLE OF CONTENTS LIST OF TABLES . . . . LIST OF FIGURES. INTRODUCTION. . . . LITERATURE REVIEW . I. Influence of Weather on Activity . Light . Rain . . . Humidity . Snow . . Temperature . . Wind Summary . ommonw> II. Mechanisms of Food Storage and Energy Conservation . A. Storage Mechanisms. . B. Conservation Mechanisms . C. Summary and Conclusion EXPERIMENT I--EFFECT OF PRIOR DEPRIVATION EXPERIENCE ON SURVIVABILITY. A. Purpose . . B. Subjects . . . . C. Experimental Design D. Procedure. E. Analysis . . F. Results . . G. Discussion . . . . iv Page vi viii 10 ll 11 14 15 16 17 17 20 25 27 27 27 28 33 34 35 40 Page EXPERIMENT II--THE EFFECT OF THE TIME OF PRESENTATION ON FOOD INTAKE: EARLY-NIGHT XS: LATE-NIGHT . . . . . . . . 49 A. Purpose . . . . . . . . . . . . . . . . . 50 B. Subjects . . . . . . . . . . . . . . . . 50 C. Experimental Design . . . . . . . . . . . . . 51 D. Analysis . . . . . . . . . . . . . . . . . 53 E. Results . . . . . . . . . . . . . . . . . S7 1. Survivability . . . . . . . . . . . . . 57 2. Body Weight Loss . . . . . . . . . . . . 62 3. Food Consumption . . . . . . . . . . . . 69 F. Discussion . . . . . . . . . . . . . . . . 79 GENERAL DISCUSSION. . . . . . . . . . . . . . . . 84 FUTURE EXPERIMENTS. . . . . . . . . . . . . . . . 88 1. Organization of Feeding . . . . . . . . . . . . 88 2. Food Storage. . . . . . . . . . . . . . . 89 3. Experience with Feeding . . . . . . . . . . . . 90 APPENDIX THE EFFECT OF ACCESS TO A RUNNING WHEEL ON SURVIVABILITY . . . . . . . . . . . . . . 91 LITERATURE CITED . . . . . . . . . . . . . . . . 96 Table 10. LIST OF TABLES Design for Experiment I . . . . . . . Contingency table (2 x 2) of survivability for experienced vs. control P.m. bairdi on the test deprivation (food available during the first seven hours of the 12 hour dark period for 15 days) . . . . . . . Contingency table (2 x 2) of survivability for past- experience vs, long-term control P,m, bairdi on the test deprivation. . . . . . . . . . Analysis of variance (Model I single classification analysis of variance) for grams consumed per gram body weight in 2,2, bairdi on Day 3 of the test deprivation . Chi-square analyses (Run 1, Run 2, pooled, and homogeneity) of survivability for wheel vs. non-wheel P.m. bairdi on the test deprivation (foBd'available during the first seven hours of the 12 hour dark period for 15 days) Design for Experiment II. . . Fatalities of P,m, bairdi under the test deprivation and reverse test deprivation conditions . Chi-square analyses (Run 1, Run 2, pooled, and homogeneity) of survivability for late food test deprivation (food available during the final six hours of the 12 hour dark period for five days), vs, early food test deprivation (food available during the initial six hours), in P.m. bairdi . . . . . . . -_'_ Chi-square analyses (Run 1, Run 2, pooled, and homogeneity) of survivability for early food vs, late food P,m, bairdi on the reverse test deprivation Chi-square analyses (Run 1, Run 2, pooled, and homogeneity) of survivability of early food experience vs, early food control P,m, bairdi on the early food test deprivation . vi Page 29 36 39 39 45 52 58 59 6O 61 Table Page 11. Chi-square analyses (Run 1, Run 2, pooled, and homogeneity) of survivability for early food experience vs. late food experience P.m. bairdi on the early food test-deprivation and the earIy_food reverse test deprivation, respectively . . . . . . . . . . . . . . . 61 12. Analysis of variance table for body weight loss (grams) for experience vs, control P,m, bairdi after three days of early vs, late test (or reverse test) deprivation. . 63 13. Comparison of mean body weight loss in grams for experi— ence vs, control P,m, bairdi after three days of early vs, late test (or reverse test) deprivation. . . . . 65 14. Fmax tests for homogeneity of variance within error terms of body weight change analysis . . . . . . . . . 68 15. Analysis of variance table for total food consumed (grams) for experience vs. control E,m, bairdi for 5 days of early vs. late'EESt (or reverse test) deprivation. . . 70 16. Comparison of mean total food consumption in grams for experience vs, control 3,9, bairdi for five days of early vs, late test (or reverse test) deprivation. . . 72 17. Fmax tests for homogeneity of variance within error terms of 5-day totals of food consumed (grams). . . . 75 vii LIST OF FIGURES Figure l. The food intake pattern (taken from Esterline Angus recordings) of four 3,9, bairdi for five consecutive nights each . . . . . . . . . . . . 2. Daily presentation of food for Experiment I . 3. Daily mortality on the test deprivation of Experiment I for 2,9, bairdi. . . . . . . . . . . . 4. Food consumption of 2,9, bairdi on the test deprivation of Experiment I. . . . . . . . . . . . . S. Split-plot, report measure design with Latin squares in the subplots for Experiment II. . . . . . . . 6. Body weight loss (in grams) for experience (early or late) vs, control 3,9, bairdi from Day 0 to Day 3 of early vs, late test (or reverse test) deprivation . . . 7. Total food consumed (in grams) for experience (early or late) vs, control £,m. bairdi during five days of early vs, late test (or reverse test) deprivation 8. Food consumption (in grams) during Day 1 and Day 5 of the test (or reverse test) deprivation for late vs. early deprived 2,9, bairdi . . . . . . '—_ A. Running wheel apparatus with adjoining nest-box. . . viii Page 31 38 42 55 67 74 78 93 INTRODUCTION The natural occurrence of temporal food deprivation in rodents and its consequences to their feeding behavior have not been explored. Food deprivation generally connotes a shortage in quantity or density, which requires increased searching and optimization of effort (energetic efficiency in terms of type and size of food consumed). In contrast to the quantity of food, however, many species experience temporal restrictions on foraging, which may be the more prevalent form of food deprivation. The temporal world of the animal must therefore be accounted for when describing feeding strategies. If unfavorable weather conditions affect the opportunity for foraging in the strictly nocturnal Peromyscus, we would hypothesize that these mice are able to adjust their daily feeding to coincide with the more favorable variations in light intensity, temperature, precipitation, etc. The result would be to enhance thermoregulation by avoiding unnecessary energy and water loss and to maximize concealment from predators while the mouse is out of the nest and active. This dis- sertation focuses on the question of the capacity of Peromyscus maniculatus bairdi to compensate for daily periods of temporal food deprivation by increasing intake during the times of accessibility. In this laboratory prairie deer mice, P.m. bairdi, showed little accomodation to a deprivation schedule which allowed for six continuous hours of access to food per day (Jaeger, unpub. obs.). Adult male mice were maintained in separate cages on a L:D ratio of 12:12 with 99_libitum food and water and with access to a running wheel. When permitted only the initial six hours of their dark activity period for feeding, five of the six animals died by the fourth day. This result is particularly surprising when it is con- sidered that under 99_libitum laboratory conditions these mice spent on the average a total of only one and one half hours per day feeding, considerably less time than the six hours allowed here. Further, this feeding appeared to be distributed throughout the dark hours and not concentrated in the late night (see Figure 1). Rats, in contrast, can be made to concentrate their daily intake into fixed-periodic meals (Lawrence and Mason, 1955; Tepperman and Tepperman, 1958; Cohn 32.313’ 1962; Baldessari SE 91,, 1971). A maximal adjustment to a single, one-hour meal per day was reported by Dufort (1964). This observation of intolerance to deprivation in these mice suggests the following alternatives: (1) Either 2,9, bairdi have a more or less fixed pattern of- daily food intake which cannot be collapsed to fit a reduced period of access, or (2) These mice can adjust their feeding to accomodate temporal food restriction, but under a set of circumstances different from those of the original observation (i.e., adequate experience, optimal circadian timing, etc.). The first alternative is likely if 2,9, bairdi is not naturally confronted with temporal food deprivation, or if other mechanisms are |I|I|I l I .530 munwfic 33300980 3: .Hom vafimn .e.m psom mo Amwcfiupoumh msmc< ocflapoumm scum coxmuv :uouumm oxmucw poem oLF .H opsmflw JequmN lowguv available which exclude the need for an adjustable pattern of daily feeding. The literature review will examine these possibilities. Here, the assumption will be defended that fluctuating weather con- ditions (light, rain, humidity, air temperature, snow, and wind) can and do impose temporal restrictions on the activity of Peromyscus. Next, the mechanisms of food storage (hoarding, fat deposition, coprophagy) and energy conservation (torpor, seasonal acclimatization, nest building, huddling) will be evaluated as alternatives to feeding adjustments. Where possible the limitations of each mechanism will be pointed out as they are understood for Peromyscus. From this, it will be suggested that storage mechanisms fail after prolonged periods of unfavorable weather (i.e., late winter). It is concluded from the literature review that 2,9, bairdi must then rely on conserv- ing energy by remaining in a suitable microhabitat and foraging for only limited periods of time, during those weather fluctuations which most favor out-of-the-nest activity. The experimental portion of this dissertation will address itself to the second of the above alternatives. Conditions of the original observation (lab-reared animals with access to food only during the initial six hours of dark) suggest two variables of potential importance to the feeding biology of Peromyscus: (a) pvior experience with temporal food deprivation, which lab- reared animals would not have had, might enable these mice to respond to restricted temporal access to food by increasing their rates of intake; and/or (b) an early vs. late night circadian influence on food intake may exist, whereby the greater adjustment to deprivation is made with late-night access to food. Prior experience and circadian influence will, therefore, be examined for their effects on the survivability, weight loss, and food con- sumption of 2,9, bairdi. LITERATURE REVIEW 1. Influence of Weather on Activigy Both laboratory and field observations reveal that "elements" of weather influence activity in small mammals. Considered here will be light, rain, humidity, snow, temperature, and wind, as well as their combined long-term effects. The mouse's response to any of these weather factors is based either on its need for concealment from predators or its regulation of energy and water resources. Activitv, as it is used here, is a general term taken to include feeding, exploration, grooming, etc. It is assumed that the animal cannot be feeding or foraging if it is not moving about. In this laboratory the feeding and wheel-running of 2,9, bairdi show a close temporal association, such that a bout of feeding is imme- diately preceded or followed by running. Therefore, throughout this review "activity" will be taken as a measure of feeding. The reader should be cautioned, however, that different measures of activity (i.e., v99_wheel-running, movement back and forth across a treadle, or a delicately balanced cage which detects subtle disturbances) measure different sets of behaviors and can therefore appear to contradict one another. When this situation occurs, we will question the extent to which the measures include feeding. A. Light It has been well established that the daily light—dark cycle acts to entrain the circadian activity of the nocturnal Peromyscus (Johnson, 1926; Behney, 1936; Hatfield, 1940; Stinson, 1952; Falls, 1953; French, 1956; Getz, 1959; Layne, 1971). 2, leucopus, for instance, were reported to be 99.5% nocturnal in their wheel-running on lighting regimes which included artificial twilights (Kavanau, 1969). Seasonal changes in the length of the photoperiod influence the temporal limits of daily activity by extending them over a longer period during winter than during summer nights (Johnson, 1926; Falls, 1953). When the available time for nocturnal activity is very short, as it is during mid-July in the Yukon Territory, activity can occur in as few as five hours per day (Swade and Pittendrigh, 1967). The effect of the dark period duration on the total daily amount of activity is, however, less clear. Activity is not continuous through- out a 12 hour nocturnal period, but rather interspersed with varying periods of "rest" or inactivity (Figure l). Foraging might be con- stant during the extremely reduced night of the sub-arctic summer, without differing in total amount from that occurring during long dark periods. Habitat characteristics can also affect day length. The forest-dwelling species 2,9, gracilis remained active longer than did 2,9, bairdi on an adjoining beach (Falls, 1953). Similarly, 2, floridanus, found in a relatively more open habitat than is 2. gossvpinus, showed the greater decline in time Spent out of the nest and in movement across a treadle with increased photoperiod (Layne, 1971). Exceptions occur, however, when activity is greater on a long- day photoperiod (16:8 LD) than on a short-day photoperiod (9:15 LD) (Lynch and Lynch, 1973). In outdoor cages in northwest Canada, 2, maniculatus began increasing daytime movement in mid-March with the nocturnal peak of activity becoming bimodal. This trend continued through the spring and by mid-June, when the dark period was only about three hours, activity persisted for six hours after dawn (Stebbins, 1971). The presence of snow in the cages appeared to correlate negatively with the extent of movement across treadles; snow cover and air temperature might together mask an influence of day length. Beyond light as a Zeitgeber, changes in its intensity within either the light or dark phase of the day can influence outside movement through effects on both vision and concealment. Nocturnal, large-eyed Peromyscus will favor dark but visually perceptible surroundings for "outside” activity. Manipulation of light intensity in the laboratory bears this out. The greatest amount of activity occurs in a range of dim light, between darkness and that equivalent to a full moon, approx- imately 0.024 ft-c (see Fall's review, 1968). When exposed to a dim- dark light cycle, 2, leucopus were active during the dim phase, not the dark (Kavanau, 1968). He noted an optimum range for activity from 0.00018 to 0.015 ft-c. Illumination from the moon is, therefore, a potential deterent to unconcealed movement. It was found to be the most potent factor 10 affecting the activity of Peromyscus polionotus leucocephalus on the open beach at Santa Rosa Island, Florida (Blair, 1951). The trapping and tracking records indicated much less movement at light intensities from a half to a full moon than at those when the moon was slight or obscured by clouds. Similar findings have been reported for other species of Peromyscus, particularly those found in more open habitat such as old-field or desert (Gentry and Odum, 1957; Gentry, Golley, and McGinnis, 1966; Falls, 1968; Owings and Lockard, 1971). The forest dwelling 2,9, gracilis, on the other hand, appeared to be unaffected by the lunar cycle (Falls, 1968). This lunar inhibition of movement has been reported to reduce greatly trapping success in a population of known density for a period of up to six consecutive days (Caldwell and Connell, 1968). Avoiding visual exposure, there- fore, seems to be the important consequence of inactivity at "higher" light intensities. For Per09yscus it appears that those situations with the greatest potential for serious temporal food restriction due to light intensity are (1) extremely short summer nights such as found in the Canadian sub-arctic and (2) a period of successive nights which are clear and moonlit. B. Rain The effect of rain on the activity of small mammals is unclear. The accompanying overcast provides concealment, whereas the rain can be a masking noise, making a predator's detection of a mouse by sound more difficult. Direct contact with rain, however, can result in wet pelage and reduced insulation. In general, Peromyscus are more active on nights with a light to moderate rainfall (Burt, 1940; Blair, 11 1951; Gottschang, 1952, from Martin, 1973; Gentry and Odum, 1957; Hays, 1958; Mirth, 1959; Gentry s: 91., 1966; Falls, 1968), presumably because rainy nights are also the darkest nights. Heavy rainfall, however, restricts the movement of small mammals (Falls, 1968; O'Farrell, 1974). C. Humidity The influence of humidity on deer mouse activity is also diffi- cult to assess; its effects are confounded with cloud cover, air temperature, and wind. Similar to the other weather factors, there likely exists a favored range of humidity, with either extreme being avoided. Field studies do suggest a direct relationship between humidity and activity, except possibly near saturation (Orr, 1959; Falls, 1968; Martin, 1973). Under controlled lighting and temperature in the laboratory, however, the activity of 2,9, bairdi decreased as the vapor pressure increased (Stinson, 1952; from Falls, 1968). The high humidity associated with cloud cover and light rain evidently has no adverse influence on movement. If humidity does reduce the opportunity for foraging by Peromyscus, this may be at low vapor pressures, which threaten desiccation. The cactus mouse, 2, eremicus, for instance, becomes torpid in response to negative water balance and will spend up to several weeks in humid burrows during the desert summer (MacMillen, 1965). 0. Snow Snow cover can influence the activity of Peromyscus in a variety of ways. It offers concealment for subsurface movement, but 12 if densely packed it can also become an obstacle to tunneling. Con- siderable tunneling and sporadic daytime activity has been reported for 2.1. noveboracensis;when theirlcages contained several inches of newly fallen snow (Behney, 1936). The over-all movement of three individually housed 2, maniculatus was found to be reduced during the winter in the Northwest Territories (Stebbins, 1971). The animals had to tunnel to find their food or to exercise. The increase in activity in late March coincided with the loss of snow cover over the top of the cage. Deep or densely packed snow could, therefore, have an adverse effect on foraging. Snow, in fact, seems related to seasonal migrations of 2, boylei in California (Storer s£_sl,, 1944). They observed mice moving back into previously vacated areas after the snow had melted. In addition, the surface of the snow can provide a contrasting background making the movement of the mice across it very conspicuous. Peromyscus, nevertheless, are active on the snow, as is evidenced by their tracks. O'Farrell (1974) noted 2, maniculatus as the only desert rodent he studied that remained active when snow was present. The effect of the snow per ss_on the extent of surface activity is, however, not well understood. Reduced catches of 2,1, noveboracensis were observed in Michigan following heavy snowfalls (Brand, 1955; from Stickel, 1968). This result might have been due to the placement of his traps at regular intervals. Snow can possibly restrict the move- ment of Peromyscus to protected areas such as brush piles and under logs (Fitch, 1958; Beer, 1961). 13 Snow as an insulative agent can also be an important influ- ence on small mammal activity. Under the conditions of winter in northern Canada (i.e., low temperatures and deep snow) Peromyscus were reported as almost impossible to trap (Fuller s£_92., 1969). But with very little snow cover, a temperature of -19°F did not seem to curtail the activity of either 2, leucopus or Clethrionomys gappsri, although both populations suffered heavy losses during that period (Beer, 1961). Pruitt (1957) examined subnivean conditions on the taiga near Fairbanks, Alaska. He observed that in late October when the snow reached a critical depth of 15 to 20 cm that a change occurred in the behavior of forest floor mammals. The activity of shrews and red-backed voles switched from the snow surface to below it. Pre- sumably the animals went beneath the snow to take advantage of its insulative properties. Snow density can modify this insulative effect by altering the thermal conductivity (Fuller SE 21-» 1969). Thawing increases snow density, which lowers subnivean temperatures. A southerly ranging subspecies, 2,9, bairdi, can be exposed to wider seasonal extremes and more fluctuation in temperature than one found farther north as is 2,9, gracilis (Pruitt, 1959). This is attributed to the permanent winter snow cover on the northern study plots in contrast to inter- mittent snow cover, thaws, winter rains, and a greater amount of summer insolation (heat from the sun) in the south. Comprehensively, snow cover appears to reduce or restrict the surface movement of Peromyscus. The subnivean refuge makes the mice less conspicuous to predators and offers shelter from low temperature 14 and wind. The extent to which mice are active beneath the snow and are able to forage for food is, however, unknown. E. Temperature Activity away from the nest not only requires predator evasion, but also thermoregulation. For each homeotherm there exists a region of thermoneutrality in which the metabolism is independent of the ambient temperature. At low temperatures the rate of metabolism is proportional to the temperature differential between the body and the environment. Above this region the metabolism again increases until the lethal temperature is reached. McNab (1963) pointed out that a homeotherm, such as Peromyscus, expends most of its energy for thermo- regulation. At lower temperatures increased movement and shivering can compensate for heat losses; for example, oxygen consumption is 480% greater at 10°C than at 31° to 34°C in antelope ground squirrels, Ammospermophilis leucurus (Dawson, 1955). This suggests that outside activity at low temperatures would necessitate more foraging for greater amounts of food. It has been found that 2,9, bairdi consume almost 2.5 times the number of calories at 0°C than at 36°C (Dice, 1922). He observed that "most small mammals are able to survive one day or longer without food at ordinary laboratory temperatures, but it is doubtful that any trapped small rodent can exist overnight without food or nest material when the temperature drops much below 10°C." A similar observation has been made by Howard (1951) for both 2,9, bairdi and 2,1, noveboracensis regarding what he termed "cold weather starvation." 15 Mammals commonly select a suitable temperature zone when in an environment which is not thermally uniform (Herter, 1934; 1936; Kalabuchov, 1939; Bodenheimer, 1941; all from Stinson and Fisher, 1953). Individual 2,9, bairdi distributed themselves between 20°C and 30°C when placed in an eight foot aluminum tube in which a temperature gradient existed ranging from 6°C to 50°C (Stinson and Fisher, 1953). The perception of the thermal zone was apparently through foot contact with the substratum, rather than v99 the ambient temperature. Correspondingly, Peromyscus activity seems reduced at hot or cold temperatures (Howard, 1951; Hart, 1953; Falls, 1953; Orr, 1959). For instance, reduced surface movement has been observed in conjunction with low winter temperatures (Stebbins, 1971; O'Farrell, 1974); and the distance of capture of 2, leucopus from their trapping centers has been reported to diminish at high temperatures (Haresign, 1964). In a four year study, involving 2,666 rodents, Sidorowicz (1960) found that Clethrionomys glareolus and Apodemus flavicollis were caught most readily at certain optimum temperatures. The greatest nocturnal cap- tures were between 24.1°C and 27.0°C, and were reduced above and below this with the most reduction occurring at minimum temperatures. F. Wind Air movements can accentuate the physiological effects of temperature by increasing both evaporative water loss and the extent of cooling. Little attention, however, has been given to wind as an influence on activity. Trapping success of 2, maniculatus was recorded at regular distances from a windbreak (Vose and Dunlap, 1968). More mice were found in the area of greatest wind shelter and 16 snow depth. The subnivean temperatures, however, were not markedly different at any distance, probably because snow cover remained thin, not reaching 15 cm in depth. High wind speeds accompanied by blowing sand were observed to lower rodent activity in the desert (O'Farrell, 1974). It was reasoned that "fine wind-blown sand would be highly uncomfortable and a hindrance to vision; additionally, since many rodents locate seeds by olfaction, high wind speeds would greatly affect this ability negatively." It has also been suggested that the noise associated with wind may make detection of predators more difficult and, as a consequence, inhibit movement (Davis and Golley, 1963). G. Summary In summary, unfavorable weather conditions can have a pro- hibitive effect on the outside movement of Peromyscus. Favorable weather might include light rainfall or cloud cover, moderately high temperature, and high humidity. Alternatively, mice avoid situations which make them conspicuous and/or which make excessive energy or water demands: for example, a clear, cold night. Winter weather, particularly in the more northern latitudes, offers the greatest number of adversive conditions. Low temperature, wind, reduced foliage, clear moonlit nights, a contrasting background of snow, insufficient snow depth for insulation, or densely packed snow can all provide for periods of temporal restriction to foraging. 17 II. Mechanisms of Food Storage and Enevgy Conservation If "exposed" foraging can be temporally restricted by unfavor- able weather, can temporal food deprivation result? Must Peromvscus concentrate feeding during the "most favorable" times? The answer to these questions will be "yes," unless they depend upon other strategies, such as food storage or torpor. The role of such alternative mechanisms will now be examined. Those dealing with food storage will be con- sidered first; they include hoarding, body fat deposits, coprophagy, and cannibalism. Other mechanisms exist which function to conserve energy; those to be considered are both physiological (torpor, geo— graphic differences in metabolic rate, and seasonal acclimatization) and behavioral (burrowing, nest building, and huddling). A. Storsge Mechanisms (l) Deprivation from restricted foraging can be avoided by maintaining sufficient stores of food. The occurrence of hoarding behavior or the resulting food caches have been reported for many species of Per09yscus (Cogshall, 1928; Sumner and Karol, 1929; Criddle, 1950; Linduska, 1950; McCabe and Blanchard, 1950; Howard, 1951; Houtcooper, 1972; Rice and Terman, 1972; Barry, 1974). The caches of 2,9, bairdi in outdoor next boxes appeared to be most pro- minent in October and early November, and were found to contain non- perishable foodstuffs, such as seeds and acorns (Howard, 1949; Howard and Evans, 1961). The largest store occurred in a nest box in which 12 deer mice resided. In less than one month, 1050 cc of weed seeds, 565 acorns, and six deer pellets were cached. 18 Concerning seasonal food stores, two important questions arise: (1) how long can they last, and (2) how easily can they be replenished? The gathering and maintaining of a food supply capable of sustaining a group of mice through five months of subarctic winter has not been studied. Howard and Evans (1961) did note that much of the stored food, in nest boxes on the George Reserve in southern Michigan, was gone by the end of December, and all of it by March. Their work, however, considered but one year. It is unfortunate that neither they nor Stebbins (1971) give winter information about the amount of food consumed per unit time, the addition of new food to a cache, or the establishment of new caches. (2) Seasonal food storage in deer mice can also be accomplished physiologically in the form of elevated fat deposits (Hadaway, 1972; Morris, 1975). Hayward (1965a) examined the gross body composition of six geographic races of Peromyscus and found that the average fat content in relation to the fat-free body weight was nearly twice in winter what it was in summer for all races except p7m, sonoriensis where the reverse was true. He attributed the higher winter fat values to the functions of increased tissue insulation and energy reserve. The P.m. sonoriensis, taken from the high altitude desert of Nevada, showed the highest summer fat levels (Bodenheimer, 1952; Wright, 1954). Desert homeotherms, including Peromyscus, tend to store fat for use during the hot, dry summers at a time when food intake is reduced because of reduced water intake (McNab, 1968). How long can this energy supply from fat last? Is it readily mobilized? Fat reserves of 2, polionotus could account for 1.2 to 1.7 19 days of energy reserves when foraging was limited (Caldwell and Connell, 1968). For a lactating female, however, with three young of five days of age, the estimate was lowered to between 0.7 and 1.2 days. With the stomach contents included, the mice were provisioned for 2.8 days at the basal rate of metabolism. The nursing female had a supply for 1.1 days. It was concluded that ”some individuals must at times forage for food under bright moonlight since our esti- mates of fat and stomach energy reserves would not otherwise permit survival through an extended period of clear weather. However, as the trap data suggest, these forages are most certainly restricted to a short radius from the nest” (Caldwell and Connell, 1968). They also found captures in live traps down approximately 42% from overcast to moonlit nights, even for four to six day periods under clear skies. (3) Coprophagy, too, might be considered a form of food storage. This refers to the recycling of food through the consumption of feces. The laboratory rat recycles 30 to 50 percent of its total fecal output on a nutritionally adequate diet, and up to 100 per cent on an inadequate diet (Barnes, 1962). When the experimental con- ditions permitted access to feces, 2,9, bairdi endured the removal of all dietary protein for 100 days (Karch, 1974). Coprophagy, however, is generally held as a way to compensate for a particular metabolic deficiency, and is probably of limited value as an energy source. (4) In addition, each animal itself can be viewed as a potential nutrient store. Both 2,1, noveboracensis and 2,9, bairdi showed a 20 tendency toward cannibalism when deprived of food and water at low temperature (Sealander, 1952). B. Conservation Mechanisms (1) Many birds and mammals respond to energy stress with intermittent periods of torpor. Here a reduction in body temperature is accompanied by behavioral lethargy and a decrease in energy expenditure. This condition has been observed in many species of Peromyscus under both controlled (Sealander, 1953; Morrison and Ryser, 1959; McNab and Morrison, 1963; MacMillen, 1965; Morhardt and Hudson, 1966; Fink, 1973) and field conditions (Howard, 1951; Mullen, 1971; Stebbins, 1971). Food availability has the greatest effect on whether or not torpor will occur (Hudson and Bartholomew, 1964; Morhardt and Hudson, 1966). In the laboratory, 2, eremicus undergoes a typical cycle of daily torpor when food is limited. Well-fed animals maintain their body temperatures between 33° and 38°C, even when exposed to ambient temperatures as low as 5°C (MacMillen, 1965). Similarly, Morhardt (1970) found five species of Peromyscus to become torpid in response to starvation at ambient temperatures from 0° to 23°C. Torpor was not observed in animals provided with excess food. It has been reported, however, that Peromyscus will enter torpor in the presence of food (Stebbins, 1971). This "spontaneous" torpor was observed during winter in 2, leucopus that had been acclimatized to cold weather the preceding fall (Gaertner et al., 1973). 21 A torpid animal, by definition, rewarms without the aid of external heat. Species with relatively labile body temperatures have been designated heterotherms. When the phenomenon occurs in Peromyscus it is a daily event taking place during the normal period of inactivity (light phase) and terminating before the time the mouse would usually begin to seek food. For one species the duration of daily torpor was usually less than eight hours and never as long as 12 hours (Morhardt, 1970). (2) As another physiological adaptation, it has been suggested that species and races of Peromyscus can be characterized by different metabolic rates, each appropriate to a general climatic situation (i.e., desert, alpine, taiga). These adaptations are presumed to be genetically fixed. Desert Peromyscus, as an example, have lower basal metabolic rates and body temperatures than mesic forms, pre- sumably to prevent overheating (McNab and Morrison, 1963). Similar findings have been put forth regarding both desert and alpine mice (Cook and Hannon, 1954; Murie, 1961). Other investigators, however, maintain that the metabolic rates of races they examined, including a desert form, are not different from a predictable, weight-dependent relationship and are, therefore, not adapted to climate (Scholander 23.31:» 1950; Hayward, 1965b). (3) Seasonal acclimatization is also believed to have a sig- nificant effect on the mouse's response to temperature. Cold- acclimatized, winter Peromyscus can withstand lower ambient tempera— tures than warm-acclimatized, summer individuals (Sealander, 1951; Hart and Heroux, 1953). These mice also show a lower existence 22 metabolism than the summer-acclimatized mice when compared at the same temperature (Morris, 1975). This is attributed to the greater insulation afforded by longer hair, greater pelage weight, and increased lipid reserves. The fact, however, that no change in the critical temperature (lower limit of the thermoneutral zone) accom- panied seasonal acclimatization to cold in Peromyscus suggested to Hart (1957) that insulative changes are insignificant. He also demonstrated that the metabolic rate at any temperature below thermo- neutrality is similar in summer and winter mice, with the exception that in winter the low temperature limit for survival is extended by approximately 20°C. This extension, he concluded, was due to an enhanced "capacity" for more intensive and sustained metabolic activity. Hayward (1965b) pointed out that "this phenomenon of seasonal metabolic acclimatization would not allow continuous exposure to sub- freezing temperatures. Its function is likely to permit short periods of intensive energy production associated with occasional exposure to very low winter temperatures. Such exposure may occur when it is necessary for animals to forage for food." This exposure can be metabolically costly. It was calculated that a 20 g Peromyscus eats approximately 1.95 g of food per day at 21° to 23°C (around 5.7 calories), and that it would have to more than quadruple its food intake to meet energy requirements at -15°C (Hayward, 1965b). He concluded "it is safe to say that a mouse in its natural winter envi- ronment is forced, in terms of energy balance and long-term weight stasis, to seek a moderate microclimate.” 23 The above mechanisms can be considered physiological controls of metabolic rate. What follows are behavioral controls whereby Peromyscus seek out and/or construct a suitable microhabitat. (4) The role of snow in providing a microhabitat of relatively constant temperature has been discussed. Burrowing, in general, appears to be a climatic adaptation to reduce the body-to—air tempera- ture gradient. The digging and gnawing movements employed by Peromyscus for burrowing or enlarging cavities in tree stumps have been described (Eisenberg, 1968). Deer mice from widely divergent habitats have been observed to use these excavations (Sumner and Karol, 1929; Hayne, 1936; MacMillen, 1965; Pengelley and Fisher, 1968; Houtcooper, 1972; Davenport, 1974; Jaeger, pers. obs.). Hayward (1965c) attached thermistors to representatives of five geographic races of 2, mani- culatus and monitored temperatures in the burrows and other shelters they entered. His results indicated that the microhabitat provides an environment of moderate temperature where seasonal extremes are avoided. Winter microhabitat temperatures were never below freezing and were very similar in all habitats despite the considerable dif- ferences in gross climates. (5) Nest building is a means of further modifying the micro- environment. The implications of this behavioral insulation for deer mice have been discussed by King (1963, 1968). It appears that winter acclimatization and a more northern geographic distribution both correlate positively with the extent of nest construction. For 292. noveboracensis the nest temperature in cold weather has been reported to be as much as 24°C higher than that of the surroundings (Johnson, 24 1926). Nest temperatures of Peromyscus were also recorded that ranged from 7° to 21°C higher than air temperatures (—30° to -35°C) with a tendency for the gradient to diminish as the thermogenic fatigue of the mouse set in (Sealander, 1952). Sealander (1951) compared survival times and total percentage of initial weight lost by winter animals having nest protection with data for unprotected winter animals at the same temperature. He found that winter animals thus protected were able to increase their sur- vival to the equivalent of a rise in the temperature of their imme- diate surroundings of as much as 25°C. Summer animals having nest protection were similarly able to prolong their survival times. Nests have been observed which contain aggregations of deer mice. Larger aggregations are formed in winter, and these groups occasionally consist of unrelated mice of the same or of different species (Howard, 1949). One of these contained seven 2,9, bairdi and three 2,1, noveboracensis. Stebbins (1971) counted 28 2, maniculatus in one nest discovered in late March. (6) This huddling is yet another important mechanism of food conservation and winter survival; its function is to reduce the amount of surface area across which heat is lost. Food consumption per gram of body weight by laboratory mice varies inversely with the number of animals in the group (Prychodko, 1958). As the ambient temperature is lowered there is an increasing divergence in the amount of food consumed by the different groups. The effect of huddling on survival times 0f 2,1, noveboracensis and 2,9, bairdi exposed to temperatures of -23°C was investigated by Sealander (1952). In this study, the 25 mean survival time of each individual in a group of two was nearly double that of individuals exposed singly. Survival time increased in much smaller increments as the group size was increased beyond two. Similarly, the oxygen consumption for a group of albino mice at 8°C was determined to be roughly the same as for a single mouse at 15°C (Mount and Wilmott, 1967). C. Summary and Conclusion It is difficult to summarize the combined effects of these mechanisms of energy storage and conservation on the feeding biology of Peromyscus. Nevertheless, in light of the wide range of available alternatives just discussed, it would appear that isolated, short-term restrictions (as might occur with the lunar cycle) would not result in serious food shortage and energy depletion. Moreover, food caches and fat reserves together with coprophagy would likely assure energy for longer periods of restriction, imposed for instance by winter weather. The building of these energy stores, however, coincides with favorable weather and optimal food availability (i.e., in autumn). In a temperate-zone winter (e.g., Michigan), these stores can become depleted by late winter or early spring, at a time of poor food availability when the caches are difficult to replace. This is also a time of continued unfavorable weather, including deteriorating snow conditions, which reduce cover and increase thermal conductivity v19 thawing. Late winter through early spring can, therefore, be an energetically critical period for deer mice. This is the time of greatest mortality, lowest reproduction, and lowest population levels for many species of Peromyscus (Howard, 1949; McCabe and Blanchard, 26 1950; Beer, 1961; Brant, 1962; Sadleir, 1965; Fuller_s£_s1., 1966; Martin, 1973; Myton, 1974). The mechanisms of energy conservation become particularly important when food stores are depleted. When the mouse conserves energy while huddled in its nest without stored food, it is temporally food-deprived. It then has two alternative strategies, either to remain sheltered, or to venture out for food, even at a relatively great energy expense. A temporally flexible feeding pattern would determine which alternative the mouse takes. The question thus remains viable, to what extent are these mice able to adjust their daily feeding to accomodate the more favor- able fluctuations in light intensity, low temperature, precipitation, etc? What is the capacity of Peromyscus to compensate for daily periods of temporal food deprivation by increasing intake during the times of accessibility? EXPERIMENT I--EFFECT OF PRIOR DEPRIVATION EXPERIENCE ON SURVIVABILITY A. Purpose There is reason to expect that Peromyscus must compensate for daily periods of temporal food deprivation by increasing their intake when the food is accessible. An initial observation indicated that 2,9, bairdi have very little tolerance to six hours 0f food avail- ability per day. Since these were laboratory-reared mice, they might have been particularly vulnerable because they had no prior experience with food deprivation. The purpose of this experiment is, therefore, to examine the effect of prior experience on the deer mouse's survival under temporal food deprivation. More specifically, does experience with periods of progressively more restricted daily exposure to food enable Peromyscus to survive longer in a later and more severe test deprivation? Gradual experience with progressively greater food deprivation is intended to be analogous to the experience wild mice might incur as fall progresses into winter and the weather becomes more severe . B. Subjects For this study 38 male 2,9, bairdi were used. They represent the males of 18 litters, from 12 distinct matings. At the initiation 27 28 of the experiment these mice averaged 109.4 :_2.4 days of age (:_one standard error), with an average body weight of 15.82 :_0.28 g. All animals were the first laboratory generation from parents caught in the vicinity of East Lansing, Michigan. At weaning (21 days of age) the mice were housed with the male littermates in plastic laboratory cages (28.6 x 12.7 x 15.2 cm). They were provided wood shavings and a cotton Nestlet (Anicare). Food (Wayne Mouse Breeder Blox) was provided 99_libitum by means of a wire mesh hopper in the top of each cage. Water was also available 99_libitum. The laboratory 'environment was maintained at approximately 21°C under a 15:9 hour light-dark cycle. C. Experimental Design The deprivation experience consists of successive blocks of time (approximately 21 days each) during the first of which the food is available for nine hours each day and during the second for eight hours. The "experienced" mice are then compared with controls for their survivability under a test deprivation where food is accessible for only seven hours per day. The experimental design is presented in Table 1, while the types of experience and the test deprivation are illustrated in Figure 2. Initially, both the group to receive the deprivation experience and the controls undergo a baseline period (24 days) where their food is available each day during the 12 hours of dark (see Row 1 of both Table l and Figure 2). (Under all experi- mental conditions a 12:12 hour light-dark cycle prevailed.) .xmv you voom op mmooom mo muse; NH ow mafi:93uou noumm mxmv conga .mo xma :o vofiv maonm was» mo Hmawcm ecclea .mcficofiuflucoo ocfiaommn mo w xma :o vofiu msouw was» mo Hmswcm macaw .N unawam oom¥ 29 n A n n ma umoh .m 2 w .2 S om 8523a .4 w m S om 8:33qu .m m S 2 R 859883 .N NH NH NH vN ocflaommm .H «seam u fiv mm H CV «.kmNH u CV mm H fig mocofipomxm mocofiuoaxm Honucou Monacou amen oumflonEH Euop-uno:m Ehohumcoq oswmom comm no :ofiumNflcmeo mxwo mo Honesz Hmucoefiuomxm mocowpomxm Honucou maneuoEwh Hancosfiuomxm 1:uoom on mmoou< mo >ma\mnsoz .H unosfihomxm How cmflmoonu.fl oanmb 30 .H ucosfiuomxm How voow mo cofiumucomoum xaflma .N ouzwflm 31 Sum E6: “ [ book 0: ODE book 0: E& v [llllllll Sam 0 [1 Eu: E6: M n 03 o: QOQK boo.“ o: — c193 _ xm P>0.025). The experience, therefore, either immediate or three weeks past, has been effective in reducing mortality. It was hypothesized that experience would manifest itself in a greater daily food consumption under the test deprivation. Through the stepwise presentation of the deprivation experience 2,9, bairdi would learn to increase their rates of intake in re5ponse to food availability. To examine this, the food consumed per gram of body weight (that at the initiation of the test deprivation) is compared between groups for Day 3 of the test. Day 3 was selected because it appeared to be a critical time in the animal's reSponse to the deprivation (mode for time of death was Day 4). Results of the analysis of variance are shown in Table 4. No differences exist among 37 Figure 3. Daily mortality on the test deprivation of Experiment I for 2,9, bairdi. Accumulative % Mortality 80 70 m O 00 O b O 01 O N O I IO 38 !l IMMEDIATE EXPERIENCE -_-_ .g-_l , PAST EXPERIENCE 12 3 4 5 6 7 a 9l0 ll l2l3 Day of Test Deprivation 39 Table 3.--Contingency table (2 x 2) of survivability for past experience vs, long-term control 2,9, bairdi on the test deprivation. Past-Experience Long-Term Control 2 Observed Expected Observed Expected Died 0 1.88 4 2.12 4 Survived 8 6.12 5 6.88 13 E 8 9 17 x2 = 4 64° x2 = 3 841° 0 os>p>o 025 ' ’ 0.05(l) ' ’ ' ' Table 4.--Analysis of variance (Model I single classification analysis of variance) for grams consumed per gram body weight in 2,9, bairdi on Day 3 of the test deprivation. Source of Variance d.f. 88 MS F Ratio Treatments 3 1,299.68 433.23 0.19 Control vs. Experience (1) (940.12) 940.12 0.42 Within 33 73,902.05 2,239.46 Total 36 75,201.73 40 the four groups (P>0.75) or between the grouped-controls and the grouped-experience animals (0.50P>0.2), supporting the contention that food intake per ss_was not the critical variable. 45 Table 5.--Chi-square analyses (Run 1, Run 2, pooled, and homogeneity) of survivability for wheel vs, non-wheel 2,9, bairdi on the test deprivation (food available during the first seven hours of the 12 hour dark period for 15 days). Wheel No Wheel Survived Died Survived Died X2 d.f. p Run 1 4 9 9 4 3.85 1 0.05>p Run 2 0 10 8 2 13.34 1 0.005>p Sum of l G 2 17.19 2 Pooled 4 l9 l7 6 14.81 1 0.005>p Homogeneity 2.38 1 0.5>p>0.l Woolf, 1968. The data suggest, rather, that it is the level of general activity that is important for survival under these conditions. In fact, it was observed that wheel running, as visually monitored through an Esterline-Angus event recorder, intensified in response to the removal of food and often continued well into or through the light period. Deprivation has been shown to increase activity in running wheels much more than in stationary cages when the two devices are directly compared (Treichler and Hall, 1962; Weasner EE.El:’ 1960). This is probably because locomotion on a running wheel produces greater visual, auditory, and kinesthetic stimulation than does locomotion in a stationary cage (Gross, 1968). This running seems to be "self- reinforcing," as rats will bar-press to obtain access to a wheel (Kagun and Berkun, 1954). In addition, the confinement to a small 46 nest-box, such as those employed here, would encourage use of the adjoining wheel. This outlet for increased activity (wheel running) has also been observed to have a profound detrimental effect on the survival of food-deprived rats (Weasner s9fls1., 1960; Routtenberg and Kuznesof, 1967). This outcome is presumed to be due to the disparity between energy input (food), which remains static, and energy output (v1s_the running wheel), which greatly increases. The greater this disparity, the higher is the likelihood of mortality. Symons (1973) found that the level of wheel-running for each of four mouse strains correlated negatively with their survival time on the food deprivation. In addition, Routtenberg and Kuznesof (1967) reported that administration of the drug chloropromazine lowered the wheel activity of food-deprived rats and resulted in a significantly reduced mortality. It should be made clear that these findings only suggest an association between deprivation experience and activity level. Very little is understood concerning the behavioral effects resulting from prolonged or repeated experience with "mild" food deprivation. The experience may affect activity by altering the animal's perception of the deprivation situation. Considerable evidence indi- cates that deprivation ps9_ss does not increase activity, but rather increases the reactivity to external or novel stimuli; this is manifested as more locomotor activity (Baumeister ss_s1,, 1964). This is consistent with the result here that the food removal seemed to stimulate an immediate increase in activity. Fehrer (1956) found that hungry animals left a familiar environment to explore a novel one faster than did satiated ones. Halliday (1968) suggested that food 47 deprivation increases exploratory activity only until the animal finds that there is no food in the situation; thereafter exploratory behavior decreases with deprivation. No data, however, have been put forward to support this hypothesis. Experience might also partially extinguish the learned association between eating and movement activities associated with the procurement of food. Richter (1972) first reported that periods of locomotor activity (v19 the running wheel) are closely associated in time with the feeding bouts of the rat. A similar observation has been made for 2,9, bairdi (Jaeger, unpub. obs.). Feeding and wheel- running have, in addition, been shown to be mutually reinforcing, such that an animal will not only increase wheel-running to obtain food but also increase food intake to obtain access to a running wheel (Premack and Premack, 1958). Finger 33.31: (1960) found a differential rein- forcement of activity on mild deprivation depending on the length of the interval between the period of access to a running wheel and the food presentation. Significantly less activity occurred when food availability followed access to the wheel by one hour than when the food was available immediately upon removal from the wheel. These examples, therefore, suggest that the experience might provide suffici- ent time for the feeding reinforcement to become dissociated from a burst of running activity, and that this results in a diminished running reaction to the food removal. The results of Premack and Premack (1958) support this interpretation. They found that when deprived of a wheel the food intake of rats increased to a maximum on about the third post-wheel deprivation day, and that intake did not return to baseline for about 14 days. It is interesting to note here 48 that the mode for day of death on the test deprivation was Day 4, similar to the time of their maximum response. It appears, therefore, that activities associated with foraging might increase or decrease depending on the reinforcement properties of the feeding situation. Experience might, therefore, play a very important role in whether or not an ”apprOpriate" response is made in a deprivation situation. For circumstances of low food density that response might be to increase search activity, while for unfavorable weather, reduced movement (i.e., remaining in a suitable microhabitat) might be the more appr0priate. It is also conceivable that foraging for a food, which is evenly scattered but in very low density, can be energetically more costly than the food return. Here again conserving energy output might be the most effective approach, particularly if the deprivation situation is temporary, for example a temporal gap between the maturity or availability of different food sources. Deprivation experience and energy conservation notwithstanding, inactivity must be complemented with periods of foraging, especially when food stores are unavailable. It is hypothesized that deer mice will concentrate this foraging into periods of restricted access (i.e., due to conditions where exposure to either low temperature or predators must be minimized). EXPERIMENT II-—THE EFFECT OF THE TIME OF PRESENTATION ON FOOD INTAKE: EARLY—NIGHT vs. LATE-NIGHT Regardless of the deprivation experience, 2,9, 991991 demonstrated little capability in adjusting their daily intake to seven hours of food availability. Both control and experience animals ate the same reduced amounts of food per gram of body weight; and both groups lost appreciable weight. This outcome suggests that food intake might be influenced by the temporal preperties of the presenta- tion and removal. It will be recalled that, during the test depri- vation, food was available early, from the onset of the dark period. Some evidence shows that feeding can be affected by the circadian timing of food availability (Welch, 1968; Baldessari ss_s1,, 1971). Nelson s£_s1, (1973) reported that the survival of young BALB/c mice, after abrupt restriction to a single four hour span of daily food accessibility, depended on the temporal placement of this feeding span in relation to the lighting regime. These mice allowed access to food during the initial four hours of their twelve hour diurnal period began dying within three days with 50 percent mortality by the fifth day; while mice allowed food in either the initial or final four hours of their nocturnal period began dying on the fifth 49 50 or sixth day with 35 percent mortality by Day 13. That there can be a circadian influence on the nocturnal foraging in Peromyscus is suggested by the work of Hirth (1959). He observed a differential reaction to rainfall within the nocturnal period of 2,1, noveboracensis, with a tendency for catches to be low if rain fell before midnight and higher if it fell afterward. A. Puvpose Are 2,9, bairdi able to concentrate their daily food intake into a particular time period during the diel cycle in order to com- pensate for a period of deprivation? Experiment II addresses itself to this question. The Specific hypothesis to be tested is that food intake is greater and survival is enhanced when the food is available during the final six hours of the dark period as opposed to the initial six hours. This experiment is further intended as a replication of Experiment I, but with a new experience schedule. Here the effects of both early deprivation and late deprivation experience will be examined under early and late test deprivation conditions. B. Subjects Successive runs of this experiment employed a total of 60 male 2,9, bairdi. The 26 mice used in the inital run (from 31 October, 1974 to 20 December, 1974) weighed an average of 18.25 :_0.36 g and were 118 :_2.8 days of age at the beginning of the experiment. These animals represented nine litters from five mated pairs. The second run (from 18 January, 1975 to 5 March, 1975) was experimentally identical to the first; it used 34 deer mice at an average age of 134.8 :_3.6 days and an average body weight of 18.00 :_0.31. These animals Sl represented nine more recent litters from the same five mated pairs. The genetic history, rearing conditions, and test room conditions were the same as those described for Experiment I. C. Experimental Design The experimental design shown in Table 6 is similar to that of Experiment I. The animals are divided into four equal groups of 15 mice each, forming two control groups and two experience groups. Initially all of the animals undergo a baseline period of ten days where the food is available each day during the 12 hours of dark (Row 1, Table 6). Both control groups remain on this schedule for an additional 25 days (Rows 2-7), after which time they undergo five days of test deprivation (Row 8). At this time the Early—food controls have access to food for the initial six hours of the dark period and are deprived for the final six hours. The Late~food controls, on the other hand, are deprived during the initial six hours of dark and have access to food during the last six hours. Following the test deprivation the survivors of both control groups are given one recovery day of 12 hours of access to food (Row 9). The reverse test deprivation is then imposed for five days (Row 10). Here, the Early-food controls receive early deprivation and late access and vice versa for the Late—food controls. Following the baseline period, the experience groups undergo 25 days of their respective Early-food or Late-food experience. The Late-food experience begins with ten days where the access to food is in the final nine hours of the dark period (Row 2). This is followed by successive five-day blocks (Rows 4 and 6) allowing first the final 52 .poHHem ecHHommn moptceu ecu wcHnsp poHp moon» some :H mHmEHco mace xHHem o ouwH o xHHom p epmH o amok omho>om m umeh .oH NH NH NH NH Npo>ooom H pooh .m oome o NHoom o oooe o NHpom o oooe m oooe .m NH NH NH NH xpo>ooom m oocoHHomxm .N oumH N xHHmm N NH NH eocoHHomxm m oecowuomxm .0 NH NH NH NH Ano>eeom H ooceHHemxm .m ean w xHHmm w NH NH eecoHHomxm m eoceHHomxm .p NH NH NH NH xno>oeom H eoceHHemxm .m eumH m xHHom m NH NH eecoHuemxm oH eocoHHomxm .N NH NH NH NH ecHHomom oH ezHHommm .H .HmH u av .HmH u av HmH u av HmH u so osHmom msoo :oHooNHoomoo poem oHEH poem xHHmm poem oumH poem xHumm HmucoEHHomxm ooceHHomxm Heupcou eHnmuoEHH HouceaHhomxm poem ou mmeeu< we xmo\musoz .HH useEHHemxm Hem :meoouu.p oHnmh 53 eight hours and then the final seven hours for access to food. A recovery day with 12 hours of access separates each experience block (Rows 3 and 5), with three such days after the seven hour block (Row 7). The Early-food schedule of experience is the same as above with the exception that the nine, eight, and seven hours blocks refer to the initial hours of food availability per dark period (Rows 2, 4, and 6). The respective test and reverse test deprivations are the same as those described for the controls (Rows 8 and 10). The experi- mental procedures throughout are the same as those described for Experiment I. D. Analysis The dependent variables considered in this experiment are survival, body weight change, and daily food consumption. Survival vs, death is again tested v1s_a chi—square (Model II design) analysis of a two-way contingency table. Because the experi- ment was conducted in two separate, but equal, runs, it is necessary to check the subsamples for homogeneity before legitimately pooling the results. The chi-square values and the degrees of freedom for the two subsamples are added. A probability value (i.e., whether the sub- samples were taken from the same population) results from the differ— ence between the total chi-square value and the pooled chi-square value and is distributed with degrees of freedom equal to the difference between those of the total and pooled degrees of freedom (Woolf, 1968). The data on body weight change and daily food consumption are analyzed using a split-plot, repeat measure design with Latin Squares in the sub-plots. This is illustrated in Figure 5. Here each subject 54 .HH ucoEHHemxm How mueHmnsm 0:» CH monmscm :HumH :sz :meop enammoe umemeu .peHmnuHHmm .m oHSNHm 55 836 Hols Bis Bra. 0% a m eeem eoem eeom eo em am 3&3 E3 3m 3. E3 5 I! IHW Hers Blaze Hows: BIS m w... eoem eeem eoeu eeE Mp3 E3 3&3 E3 3m 3 axe SQ axe axm E3 3&3 oz oz , me SEES 393R ZQQ 56 receives both the late and early test (or reverse test) deprivation, but with either deprivation experience or no deprivation experience (control group). The main effects of the experience vs, control treatment are said to be completely confounded with other differences among these blocks. This confounding, however, does not affect the interpretability of the treatment effects, only the precision of the estimate. On the other hand, the main effects of early vs, late food presentation and the interaction between the experience and the time of food presentation are free from such confounding and are generally the more powerful tests (Kirk, 1968). This split-plot design has two error terms. One corresponds to the pooled variation among subjects within the experience vs, the control groups; while the other source is the pooled interaction of the early vs, late test treatment within each block. For the F ratios associated with each error term to follow an F distribution, the sources of variation within each term must be homogeneous. These can be tested by means of the Fmax test (Kirk, 1968). The repeated measures aspect of this design offers the advantage of controlling subject heterogeneity. Differences among subjects are often such as to obscure the treatment effects. An important limitation to the use of these repeated measures, however, lies in the likelihood of the error components from the reverse test deprivation not being independent from those of the preceding test deprivation. When this is the case, the variance-covariance matrix departs from the required form. The procedures found in Kirk (1968) for testing the equality of covariance matrices are those employed here. When the hypothesis of equality of the variance-covariance S7 matrices is accepted (the obtained chi-square value is less than the tabled value), independence is assumed. In addition to information from the above tests, the means of individual elements within the design will be compared. If the inter- action between experience vs, no experience and the time of food presentation (early vs, late) is significant, it will be of interest to determine in what way the early vs, late differences are not of the same magnitude for controls as for the experienced mice. It will be of further interest to compare the means of the early-late deprivation test sequence controls with those of the early-late sequence experience mice, as well as the means of the late-early controls with those of the late-early experience animals. E. Results The results are organized under three headings: survivability, bodyweight loss, and food consumption. The effects of both treatments (control vs, experience and early vs, late food presentation) will be considered for each of these dependent variables. In addition, each section will Open with a summary of the findings as they are inter- preted for that particular variable. 1. Survivability As was hypothesized, there is a circadian effect (within the nocturnal period) on survivability under the present conditions of temporal food deprivation. All of the fatalities of 2,9, bairdi under the four conditions resulting from late or early food presentation with or without deprivation experience are presented in Table 7. All of the fatalities of 2,9, bairdi (11 dead out of 58 animals) occurred when 58 Table 7.--Fatalities of 2,9, bairdi under the test deprivation and reverse test deprivation conditions. Test Deprivation (TD) Reverse Test Deprivation (RTD) Group Initial Number Number of Initial Number Number of of Animals Fatalities of Animals Fatalities 1. Early food control 15 5 * 10 0 2. Early food experience 14 0 l4 0 3. Late food control 15 0 15 3 * 4. Late food experience 14 0 l4 3 *Brackets represent the condition where food is available for only the initial six hours of dark; no brackets therefore represent the condition where food is available during the final six hours. mice were subjected to the early food presentation (initial six hours of the 12 hour dark period). Refer to TD(1), RTD(3), and RTD(4) in Table 7. This result occurred regardless of experience with late food presentation (RTD(3)) or the sequence of testing (early-late or late- early). No mice, however, with early food access experience died on the early food access test (or reverse test) deprivation (TD(2)). The chi-square analysis (Table 8) indicates that the mortality of 2,9, 991991_on the test deprivation depends on the time of the food presentation (pooled chi-square = 0.05; 0.025>p>0.01). The results from the two runs do appear homogeneous (chi-square = 0.05; 0.9>p> 0.5). During the five days of the early-food test deprivation, five of 29 mice died (17.24%), while not one of the 29 mice subjected to 59 Table 8.--Chi-square analyses (Run 1, Run 2, pooled, and homogeneity) of survivability for late food test deprivation (food available during the final six hours of the 12 hour dark period for five days), vs, early food test deprivation (food available during the initial six hours), in 2,9, bairdi. Late Food Test Early Food Test Deprivation Deprivation Survived Died Survived Died X2 d.f. p Run 1 13 0 10 3 3.39 1 0.1>p>0?05 Run 2 l6 0 14 2 2.13 l O.S>p>0?l Sum of l G 2 5.52 2 Pooled 29 0 24 5 5.47 l 0.025>p>0.01 Homogeneity 0.05 1 0.9>p>0.5 Woolf, 1968. the late-food deprivation test died. Two of the animals were dead by the morning of Day 3, and one each by Days 4, 5, and 6. Similar findings are presented in Table 9 for early food vs, late food 2,9, 991991_on the reverse test deprivation (pooled chi— square = 5.53; 0.025>p>0.01; homogeneity chi-square = 0.45; 0.9>p>0.5). Here six of 29 mice died (20.68%) during five days of the early food reverse test deprivation. Two were dead by Day 3, three by Day 4, and one by the morning of Day 5. Again no mice with access to food during the final six hours of the 12 hour dark period died. Those 11 animals that died in the test or reverse test deprivations did so at an average of 76.92 :_l.57 (S.E.) percent of their initial body weights. The finding of Experiment 1, regarding the survival advantage associated with early food experience, was confirmed here. The 60 Table 9.--Chi-square analyses (Run 1, Run 2, pooled, and homogeneity) of survivability for early food vs, late food 2,9, bairdi on the reverse test deprivation. , - -4.-~_.-’ --- Late Food Early Food Reverse Test Reverse Test Deprivation Deprivation Survived Died Survived Died x2 d.f. p Run 1 10 0 ll 2 1.80 1 0.5>p>0.1 Run 2 14 0 12 4 4.18 l 0.05>p>0.025 Sum of 1 G 2 5.98 2 Pooled 24 0 23 6 5.53 1 0.025>p>0.01 Homogeneity 0.45 l 0.9>p>0.5 Woolf, 1968. chi-square analysis (Table 10) indicates that survivability on the early food test deprivation is dependent upon experience with early food deprivation (pooled chi-square = 5.59; 0.025>p>0.01; homogeneity chi-square = 0.11; 0.9>p>0.5). Here five of the fifteen control mice died (33.33%) while all of the early food experience animals survived the five days of the test deprivation. The effect of the late food experience on survivability is less clear, although it appears to be minimal. Under the reverse test deprivation, with early food, the survivability of late food experi- ence mice (Table 7, RTD(4)) does not differ significantly from that of the early food experience animals on the early food test deprivation (Table 7, TD(2)) (pooled chi-square = 3.36; 0.l>p>0.05; homogeneity chi-square = 0.02; 0.9>p>0.5). These data are presented in Table 11. 61 Table 10.--Chi-square analyses (Run 1, Run 2, pooled, and homogeneity) of survivability of early food experience vs, early food control 2,9, bairdi on the early food test deprivation. Early Food Early Food Experience Control Survived Died Survived Died X2 d.f. p Run 1 6 0 4 3 3.41 l 0.l>p>0.05 Rm12 8 (1 6 2 129 l. asnpql Sum of l 6 2 5.70 2 Pooled l4 0 10 5 5.59 l 0.025>p>0.01 Homogeneity 0.11 1 0.9>p>0.5 Woolf, 1968. Table ll.—-Chi-square analyses (Run 1, Run 2, pooled, and homogeneity) of survivability for early food experience vs. late food experience 2,9, bairdi on the early food teEE deprivation and the early food reverse test deprivation, respectively. Early Food Late Food Experience Experience Survived Died Survived Died X2 d.f. p Run 1 6 0 5 l 1.09 1 0.5>p>0.1 Run 2 8 0 6 2 2.29 1 0.5>p>0.l Sum of 1 G 2 3.38 2 Pooled 14 0 ll 3 3.36 1 0.1>p>0.05 Homogeneity 0.02 1 0.9>p>0.5 Woolf, 1968. 62 The trend, however, suggests that the early experience mice are more successful under these conditions. In this case three of the late experience animals died while, again, there was no mortality among the early experience mice. 2. Body Weight Loss The results here confirm the early vs, late effect shown on survivability; significantly more weight was lost under early food availability. There is, however, no effect of experience on body weight loss. The early food experience mice, therefore, lose weight to the same degree as do the controls or the late food experience mice when on the early food deprivation. In addition, there appears to be no interaction between experience and the time of food presentation. These results do not support the hypothesis that experience affects a reduction in activity during deprivation. The analysis of variance for body weight loss for experience vs, control 2,9, bairdi is presented in Table 12. This is based on the difference between the initial weight measured at the onset of both the test and reverse test deprivations and the weight after three days (taken on the morning of Day 4) of early or late test or reverse test deprivations. The analysis is based on a sample size of 80 (40 mice in each of two test conditions). The nature of this design (repeat measures) dictates that the sample size be the same for each of the four groups (combination of early vs, late with experience vs, control), and that those animals considered in the reverse test he the same as those in the test deprivation (see Figure 5). Because five of the 15 early food controls died during the test deprivation (Table 7, 63 Table 12.--Analysis of variance table for body weight loss (grams) for experience vs, control 2,9, bairdi after three days of early vs, late test (or reverse test) deprivation. Source 58 d.f. MS F 1. Between subjects 22.3705 39 2. Experience vs. control 0.3919 1 0.3919 0.6391 3. Subjects within groups 23.3036 38 0.6132 4. Within subjects 83.8950 40 5. Test vs. reverse test deprivation 0.1619 1 0.1619 0.2752 6. (2) x (5) 0.6130 1 0.6130 1.0423 7. Early vs. late 60.5519 1 60.5519 102.9619* 8. (2) x (7) 0.0607 1 0.0607 0.1032 9. Error 21.1725 36 0.5881 10. Total 106.2555 79 *p<0.001. 64 Row 1), it was necessary to select randomly ten animals from each of the remaining three groups for this analysis. Table 12 indicates that early vs, late test (or reverse test) conditions had significantly different effects on body weight loss (F = 102.9619; p<0.001). Mice with access to food during the initial six hours of the dark period lost an average of 1.84 :_0.10 (S.E.) g compared with an average loss of 0.10 :_0.10 g for those with access during the final six hours. These Peromyscus with previous deprivation experience (early or late) lost an average of 0.90 :_0.17 g while the controls dropped 1.01 :_0.17 g. This difference is not significant (F = 0.6391; 0.50> p>0.25). The effects of early and late experience are compared separately in Table 13. Neither type of experience differs from its control for either the test or reverse test deprivation or for the two tests taken in sequence. As was suggested from Experiment 1, early food experience does not manifest itself through reduced weight loss. Figure 6 illustrates these effects of time of food presentation, experience, and the degree of interaction between them (F = 0.1032; p>0.50). It should also be noted that no difference is found between the test and reverse test deprivations (F = 0.2752; p>0.5 (Table 12(5))), nor does this variable appear to interact with the experience vs, control factor (F = 1.0423; p>0.25 (Table 12(6))). Results of the Fmax tests, for homogeneity of variance within error terms, are shown in Table 14. Homogeneity is accepted for both. In addition, the hypothesis of the equality of the experience vs, 65 Table l3.--Comparison of mean body weight loss in grams for experience vs, control 2,9, bairdi after three days of early vs, late test (or reverse test) deprivation. Control Experience Test Early food Late food Early food Late food Dsprivation (l) (3) (5) (7) (Y :_S.E.) -l.52 :_0.17 -0.29 :_0.26 -l.59 :_0.24 -0.29 :_0.24 Reverse Late food Early food Late food Early food Test (2) (4) (6) (8) Deprivation -0.10 :_0.14 -2.24 :_0.31 +0.29 :_0.10 -2.00 :_0.22 (Y :_S.E.) Comparison of Means F Late food experience (3) x (7) 1.000 (4) x (8) 0.4978 Early food experience (1) x (5) 0.0423 (2) x (6) 0.2800 Late vs, early sequence (3)(4) x (7)(8) 0.2495 Early vs, late sequence (l)(2) x (5)(6) 0.4429 F (0.05) 1,38 = 4.08 66 Figure 6. Body weight loss (in grams) for experience (early or late) vs, control 2,9, bairdi from Day 0 to Day 3 of early vs, late test (or reverse test) deprivation. Body Weight Loss (qms) tl S.E. 67 o . o.oo t.l4 -O.201:.I4 £ \ t io I - I.O -I.2- V - L4 -I.6- -l.80 1 .IT 4.88: .20 EARLY LA'TE Test Deprivation 68 Table l4.--Fmax tests for homogeneity of variance within error terms of body weight change analysis. ,, --. a. Partition of subjects within groups sum of squares (no. 3, Table 12) Source SS d.f. MS Fmax Subjects within groups 23.3036 38 Within control group 11.4938 19 0.6490 1.0442 Within experience group 11.8098 19 0.6215 Fmax (0.05) 2,20 = 2.46 b. Partition of early vs. late subjects within groups sum of squares (no. 9, Table 11) Source 85 d.f. MS Fmax Early vs. late subjects within groups 21.1725 36 Within early group 13.4480 18 0.7471 1.7409 Within late group 7.7245 18 0.4291 Fmax (0.05) 2,20 = 2.46 69 control variance-covariance matrices is tenable (chi-square = 0.9552; p>0.5, with three degrees of freedom). 3. Food Consumption These results are consistent with those found for body weight loss. That is, on the late food regime significantly more is consumed than with early food availability. Here the total five day food consumption, for the test or reverse test deprivation, is the dependent variable. Again, there appears to be no effect of either prior experi- ence or an interaction between experience and the time of presentation. This re-confirms the finding that experience is not being manifested in an increased food consumption. Furthermore, evidence indicates that this late food advantage is due to an adjusted, or concentrated, six-hour uptake which excedes that for the same period under sg_libitum conditions. It is noted that a greater adjustment is made during the first day on late food than over the five days on early food. . The analysis of variance for the total amount of food consumed during the five day test or reverse test deprivation periods is found in Table 15. Here the early vs, late test (or reverse test) conditions had significantly different effects on the total food intake (F = 100.9693; p<0.001). The animals with immediate access to food con- sumed an average of 9.13 :_0.47 (S.E.) g in five days compared to 12.24 1 0.67 g for the mice with the food available during the final six hours. Previous deprivation experience (early or late) again effected no significant difference (F = 0.9146; p>0.25 (Table 15(9))). The 7 0 Table 15.-—Analysis of variance table for total food consumed (grams) for experience vs, control 2,9, bairdi for 5 days of early vs, late test (or reverse test) deprivation. Source SS d.f. MS F 1. Between subjects 1,034.3349 39 2. Experience vs, control 24.3101 1 24.3101 0.9146 3. Subjects within groups 1,010.0248 38 26.5796 4. Within subjects 268.7250 40 5. Test vs, reverse test deprivation 3.0811 1 3.0811 1.6004 6. (2) x (5) 0.0102 1 0.0102 0.0052 7. Early vs. late 194.3761 1 194.3761 100.9693* 8. (2) x (7) 1.9532 1 1.9532 1.0145 9. Error 69.3044 36 1.9251 10. Total 1.303.0599 79 *p<0.001. 71 experienced animals did, however, consume more food than did the controls: 11.24 :_0.58 (S.E.) g and 10.13 :_0.66 g respectively. Table 16 breaks experience into its early and late components and compares each with its control for either the test or reverse test situation and for these two situations combined. No significant differences appear, however. Further, there is no interaction between the experience with deprivation and the early vs, late test or reverse test conditions (F = 1.0145; p>0.25 (Table 15(8))). The relationship is shown graphically in Figure 7. Both error terms test homogeneous (Table 17). In addition, the variance-covariance matrices (experience vs, control) are accepted as equal (chi-square = 6.7063; 0.l>p>0.05). Having established the importance of the time of food avail- ability within the nocturnal activity period, it is of interest to know whether the late food mice increase their six-hour food intake over the baseline amount for the same period, or whether they normally consume the greater percentage of their lZ-hour sg_libitum intake during the final six hours. When the amount of food consumed on Day 1 of the early food test or reverse test deprivation is taken as a per- centage of that consumed the previous ss_libitum day, the result is an estimate of the percentage of total intake consumed during the initial six hours of ss_libitum day (one with access to food during the entire 12 hours of dark). In this experiment that percentage was 49.41 :_ 1.96 (S.E.)%. (This and subsequent percentages are all derived from the same 47 survivors.) It follows, then, that 50.59 percent of the total intake occurred during the final six hours. Now, the amount of 72 Table l6.--Comparison of mean total food consumption in grams for experience vs, control 2,9, bairdi for five days of early vs, late test (or reverse test) deprivation. Control Experience Test Early food Late food Early food Late food Dsprivation (l) (3) (5) (7) (Y :_S.E.) 7.92 :_0.87 11.94 :_1.68 10.27 :_0.78 11.84 :_0.96 Reverse Late food Early food Late food Early food Test (2) (4) (6) (8) Deprivation 11.14 :_l.42 9.55 :_1.12 14.07 :_1.35 8.78 :_l.05 (Y :_S.E.) Comparison of Means F Late food experience (3) x (7) 0.0019 (4) x (8) 0.1115 Early food experience (1) x (5) 1.0389 (2) x (6) 1.6149 Late vs, early sequence (3)(4) x (7)(8) 0.0696 Early vs, late sequence (l)(2) x (5)(6) 2.6221 F (0.05) 1,38 = 4.08 73 Figure 7. Total food consumed (in grams) for experience (early or late) E: control 2.9. bairdi during five days of early vs. late test (or reverse test) deprivation. I4.0 l2.0 I0.0 8.0 Food Consumed (qmsitl S.E. 6.0 4.0 EARLY l2.95 t .84 74 EXPERIENCE CONTROL 9.52 $.68 8.73 1: .72 LA'TE Test Deprivation 75 Table 17.--Fmax tests for homogeneity of variance within error terms of 5-day totals of food consumed (grams). a. Partition of subjects within groups sum of squares (no. 3, Table 15) Source 85 d.f. MS Fmax Subjects within groups 1010.0248 38 Within control group 595.5288 19 31.3436 1.4367 Within experience group 414.4960 19 21.8155 Fmax (0.05) 2,20 = 2.46 b. Partition of early vs, late subjects within groups sum of squares (no. 9, Table 14) Source SS d.f. MS Fmax Early vs, late subjects within groups 69.3044 36 Within early group 45.3047 18 2.5169 1.8877 Within late group 23.9997 18 1.3333 Fmax (0.05) 2,20 = 2.46 76 food consumed on Day 1 of the late food test or reverse test deprivation is found to be 69.43 :_2.07% of that consumed the previous sg_libitum day. The late food mice ate an average of 2.27 :_0.13 g on Day 1 of the test or reverse test deprivation compared with 1.55 :_0.08 g for the early food animals. This implies that the late food animals did increase food intake per unit time. Of further interest is whether the daily food intake of deprived mice increases through the course of the deprivation period. The difference in food intake between Day 1 and Day 5 is an average increase of 0.43 :_0.06 (S.E.) g for the early food mice and 0.14 :_ 0.06 g for the late food mice. These differences are illustrated in Figure 8. On Day 5 the early food mice ate 64.45 i 2.77% (1.98 i 0.09 g) of the Day 0 sg_libitum intake of 3.23 :_0.17 g; this is an increase over the 49.41 :_1.96% (1.55 :_0.08 g) observed on Day 1. The late food animals increased their Day 5 intake (2.41 :_0.12 g) slightly to 75.21 :_2.9% of their baseline intake (3.28 :_0.16 g). It appears, therefore, that 2,9, 991991_can make a greater accommoda- tion in one day with late food (access for the final six hours of dark) than through five days with early food (access for the initial six hours). Such is the case within both the test deprivation (T = 1.752, p<0.05, one-tailed) and the reverse test deprivation (T = 1.834, p<0.05, one-tailed). These values are less than they might be because only the survivors were considered; those animals that died did so during the late deprivation. 77 .HpHHmn .HM po>HHmop xHHmo .umum oumH How :eHumSHHmop Hume» omHo>oH Hem umou exp we m moo paw H xma wcHuzp Hmemuw :HV coHumaomcoo poem .m oestm 78 ZOE<>EQMQ «kmmst>m& m0 mkmMm. k0 xpd IN 65 i v.8 \l T... 8 v.9 \t 9 0. m Q to QOOK wqu QOOK \St QM Mop/Stub) NOIldW/iSNOO 000:1 79 F. Discussion These results show that 2,9, bairdi can indeed compensate temporal food deprivation by increasing their intake per unit of time but only when food is available in the late food presentation. The experiment however did not explore the limits of this capacity. Can Peromyscus for example adjust their daily intake to a single one-hour meal per day as Dufort (1964) reported for the rat? This concentration of feeding into fewer but larger bouts can have important implications for over-wintering mice. "Meal—eating," in contrast to a more diffuse "nibbling" pattern, when food consumption per day is kept constant in each, can bring about increased lipid metabolism and fat deposition (Cohn and Joseph, 1960; Hollifield and Parson, 1962; Fabry and Braun, 1967; Leveille, 1967) together with increased gastric hypertrophy which results in greater absorptive efficiency of the gut (Holéckova and Fabry, 1959). In addition, questions arise regarding the possible temporal interdependency of feeding with other activities. How are these activities affected by an extreme feeding pattern? What are the ecological consequences? These ramifications of temporal food depri- vation have not been explored. That Peromvscus will compensate temporal food deprivation is not surprising in itself. What is of particular interest, however, is that this capability is under circadian restraints beyond those imposed by the light-dark cycle. This appears to limit the opportunity for compensation v1s_increased food intake. Consider, also, that early morning, when the compensation seems to be greatest, is often the coldest and therefore the most unfavorable time of the day for food 80 procurement. How is this paradoxical "late-night" effect on feeding to be explained? The term feeding has, throughout this dissertation, implied that both foraging and food consumption occur simultaneously. This is to say that the food is consumed when it is located; and this might often be the case, particularly with live, mobile prey such as insects or lumbricids. It remains very likely, however, that other foods such as seeds, fungi, herbaceous plants, or associated plant parts, are transported back to a concealed cache to be eaten at a later time. Might, therefore, Per09vscus be more opportunistic as to when they will gather and store food than as to when they will consume it? Does Experiment 11 imply an early-night tendency to forage for and heard food, with some consumption, and a late-night tendency to depend on a cache to make up any deficit in food intake? This question can be tested in a manner similar to that employed here, but including the opportunity for, and a measure of, food hoarding. The daily activity records of 2,9, bairdi support such a temporal organization of feeding (Jaeger, unpub. obs.). Under sg_libitum conditions, the feeding bouts of longest duration were often observed late in the nocturnal period, at a time when wheel-running was greatly reduced. (See Figure I, particularly Numbers 32 and 37.) A circadian regulation of food consumption is suggested by the result that 2,9, bairdi could not correct their daily intake over a five-day period of early-night access and late-night deprivation, but could readily do so in a single day when these six-hour periods of deprivation and access were reversed. This result may, in part, provide a clue to the understanding of the regulation of food intake 81 for the strongly nocturnal Per09vscus. A great deal of effort has focused on this problem in the laboratory rat (see the review by LeMagnen, 1971). These studies have used the underlying assumption that food intake is ultimately regulated at the level of the individual meal, whose size is determined in part by both that meal which preceded it and that which will follow. Daily consumption, then, is a consequence of the regulation of individual meals. This can be appreciated when it is considered that rats eat a more or less constant Inumber of discrete meals per day, and that these meals are spaced throughout both the day and night (LeMagnen and Tallon, 1966). The present evidence for Peromyscus, however, points to a daily regulation of food intake where late-night feeding can make up for a deficit incurred during the early-night period. The relatively strict nocturnal existence of Peromyscus offers an added dimension to the problem; that is, the nocturnal intake must be adjusted to account for the diurnal half of the day when food is not normally consumed. In light of this, the making up of the nightly deficit prior to the onset of the day is understandable. It assures a nutrient supply until feeding is again resumed the following night. It has been reported however that 2, maniculatus are frequently caught with their stomachs filled to less than one-third of their capacity (Harling and Sadleir, 1974). However, only two mice of their sample were captured seven or more hours after sunset (suggesting reduced foraging after this time), and one of those had a full stomach. The fact that actively foraging deer mice had relatively empty stomachs supports the hypothesis that daily intake might be regulated in the late night at a food cache. 82 The nocturnal organization of feeding that has been prOposed here also offers an explanation for the seasonal formation of caches and the conservative use of them. It may be that Peromyscus hoard storable food throughout the year in a manner described above; that is, foraging, consuming, and boarding during the early-night and feeding from the cache prior to the onset of day, during the late-night. If food is not plentiful, if it is not in a storable form, or if meteorological conditions are unfavorable, then daily caches do not accumulate. If, however, the above factors are favorable, as might be expected in the autumn, then caches become conspicuous, giving the appearance that hoarding takes place only at this specific time. Along the same line of reasoning, a conservative approach to the use of the cache would result if foraging and hoarding were ongoing behaviors. The cache, then, becomes depleted only when foraging success draps, as might occur in winter. This is opposed to the mouse depending exclusively on an available cache until it disappears and then having to resort solely to foraging, possibly at an inopportune time. We have discussed the functional ramifications of this late vs, early night adjustment in food intake. Can we now account for this phenomenon in terms of an underlying control mechanism? What physiological process(es) allow 2,9, bairdi to concentrate intake readily if food is present "late"? The answer to this question lies, in part, in an understanding of the conditions which initiate and maintain feeding. It has been mentioned previously that food intake and locomotor activity (i.e., in a running wheel) are very closely associated in time. This is very apparent in the rat where the relatively few (eight to ten), but discrete, daily bouts of feeding 83 are preceded by bouts of physical activity (Wyrwika, 1967). One consequence of this activity would be an increase in metabolic rate. Does this elevated metabolic state, with its resulting energy con- sumption, affect food intake such that the greater the metabolic rate the more food consumed per unit time? This is to hypothesize that for maximal feeding, activity must precede the feeding period. Now for the early-food 2,9, bairdi the food is presented at the point of transition from light to dark where activity, followed by metabolic rate, are only beginning to build and gather momentum. The late-food mice, on the other hand, receive food after a six hour period of activity where we would expect the metabolic rate to be higher and the daily energy deficit to be greater. Not only does the level of activity differ for the periods preceding early vs, late food availability but it differs in the reverse for the periods which follow. Early food is succeeded by a period of activity whereas late food is followed by rest. It is known that inactivity subsequent to feeding will facilitate or promote energy storage (fat deposition, etc.) (Nisbett, 1972). Rats that are returned to food availability from deprivation will eat large meals which are followed by relatively long periods of rest (Finger, 1951; Evans, 1971). This was also observed in Peromyscus (Jaeger, unpub. obs.). This "recovery response" agrees with the hypothesis that for maximum energy storage, feeding has to be followed by inactivity. If this applies in the present situation, the late-food mice would have greater energy reserves per gram of food consumed than would early-food mice. This could be an important advantage in dealing with 18 hours of deprivation per day. GENERAL DISCUSSION In his review on feeding strategies Schoener (1971) passed over a consideration of "optimal foraging period" except to state that, "No formal theory yet covers optimal placement of feeding periods over the activity cycle. Obviously basic components are metabolic costs of activity under different climatic conditions and time distributions of climatic factors, food, and predator abundance. A host of observations indicate that climate will be overwhelmingly important in models of optimal feeding period." The results presented here suggest two features important in understanding the feeding biology of 2,9, bairdi: (1) that there exists a temporal organization of nocturnal feeding behavior; and (2) that previous experience with temporal food deprivation cannot effect an increased consumption during periods of restricted access to food, but that this experience may result in a reduced locomotor response to the deprivation and thereby play a role in energy conservation. These features will now be considered in terms of a general feeding strategy. In the temporal organization suggested here, feeding behavior is reduced to three components: Foraging (locating), hoarding 84 8S (gathering), and consumption. Under sg_libitum laboratory conditions these behaviors are probably distributed throughout the nocturnal period with only a tendency for the late-night bouts of food intake to be of longer duration than the early-night bouts. When stressed with temporal food deprivation, these mice will adjust by concentrating food intake on the late-night but not in the early-night. It is hypothesized that foraging and hoarding are most prominent early in the night and can become increased at that time in response to deprivation. What might such an arrangement mean in terms of feeding strategy? As has been previously suggested, the "nocturnalism" of Peromyscus may be an important determinant of its short-term feeding pattern. One can speculate that food storage in the stomach (or in caches) must be at the daily maximum just prior to the onset of the light or inactive period (where presumably little or no foraging occurs). The temporal organization of feeding put forth here could assure this. If the stomach is gradually filled once, or even twice, during the nocturnal period, a late—night feeding could correct for any deficit leaving the animal "full" for the upcoming inactive period. In addition the early-night foraging and hoarding would assure that a food supply is available for this late-night feeding. This pattern differs from that in which the animal forages for and eats a series of discrete meals, each of which leads the stomach (such as was described for the rat). In a meal-type pattern, the day- night transition might occur between meals, when there exists the likelihood of an empty or partially filled stomach. Also, foraging from meal to meal might leave an animal with no late-night food, if 86 the weather is unfavorable at that time. Either situation enhances the possibility of a 2,9, bairdi facing the light period with less than a full stomach. Moreover, the nocturnal organization described here is, in a sense, loose; as such it can provide a margin of error which allows for unfavorable weather, time-consuming social interactions, etc. Models of feeding strategy are too often inflexible in this regard. They place animals on a very strict time budget. This circadian organization, however, relegates the various aspects of feeding to relatively wide time slots (i.e., early or late). The early-night proclivity to foraging and hoarding, regardless of the not-yet-full status of the stomach, will ensure these activities whenever no con- flict exists with the weather or social situation. The food store which results will then allow for the late-night feeding. As was pointed out earlier, feeding in Peromyscus appears to be temporally associated with other activities. This being the case, the temporal organization of feeding would have important consequences for these other behaviors. For instance, synchronized early-night foraging might facilitate social contact, and, by so doing, increase the time spent "in contact" or reduce the time spent in pursuit of it. Organization, therefore, on the circadian level might lead to greater temporal efficiency than that on a smaller scale (i.e., meal to meal). Again, this can be of particular importance to the strongly nocturnal Peromyscus. Experience with temporal food deprivation did not affect this organization of feeding. It appears, rather, that experience may play a role in the mouse's recognition of temporal food deprivation or 87 unfavorable weather such that it seeks a more suitable microenvironment (i.e., becomes less active). Experience, very likely, influences the effectiveness of other of the mechanisms of food storage and energy conservation. FUTURE EXPERIMENTS An important outcome of the interpretations put forth in this dissertation is the generation of ideas for further experimentation. Those that are of particular interest to me will now be outlined; these are (1) organization of feeding, (2) food storage, and (3) experi- ence with feeding. 1. Organization of Feeding The main impetus of the work presented here is toward testing the hypothesis that feeding behaviors are organized within the nocturnal period such that foraging and hoarding are predominant early while food consumption can be concentrated late. This can be tested in a manner similar to that employed in Experiment II. Here we will also require measures of hoarding and foraging. These can be made by presenting discrete food pellets which can be transported from the bin in which they are offered back through a tunnel (equipped with treadles) to the plastic cage housing the mouse (via Barry, 1974). The food bin and tunnel form a unit that can be attached to and removed from the home cage. Hoarding is simply measured as the number of pellets found in the home cage and foraging by the number of treadle crossings in the tunnel. It is, therefore, hypothesized that the early-food mice will show more foraging and hoarding than the late-food 88 89 mice, but again that the late-food animals will consume the greater amount of food. All hoarded food is removed following the six hours of availability and the amount of food consumed during that period is recorded. It is also of interest to test the hypothesis that food gradually accumulates in the animal throughout the nocturnal period and that the stomach is filled just prior to the onset of daylight. This can be examined by sacrificing 2,9, bairdi (on sg_libitum food) at regular intervals throughtout the night and measuring their stomach contents. In addition, is the bi-modal running activity character- istic of longer nights reflected in a bi-modal stomach filling? 2. Food Storags_ In regard to food caches the following questions have been generated: (a) Is food hoarded by 2,9, 991991 throughout the year, but found in accumulated caches only during the times of abundance (i.e., autumn)? (b) 00 food caches become exhausted by late winter or early spring? (c) Are 2,9, bairdi able to replenish their food stores during the winter? These and related questions can be approached through the use of outdoor nest boxes (v1s_Howard, 1949; Stebbins, 1971) which allow stored food to be examined, weighed, and marked for turnover by the investigator. If, for example, food is continuously hoarded we would expect to find, with the addition of excess "supplemental food," accumulated caches at those times when they are not usually found (i.e., during summer). 9O 3. Experience with Feeding Only one basic regime of deprivation experience was considered here. Another might be more effective; for instance, one where the deprivation is more severe but presented intermittently or for fewer days. In addition, the possible effect of experience with temporal food deprivation on reduced locomotor activity has yet to be resolved. Furthermore, food hoarding and foraging under conditions of a test deprivation might be much more responsive to the appropriate experience than was food intake. These alternatives also merit further investi- gation. APPENDIX THE EFFECT OF ACCESS TO A RUNNING WHEEL ON SURVIVABILITY APPENDIX THE EFFECT OF ACCESS TO A RUNNING WHEEL ON SURVIVABILITY This appendix contains a description of the experimental pro— cedures which were referred to in the discussion of Experiment I. The purpose of this experiment was to examine the effect of general activity on the survival of 2,9, bairdi subjected to temporal food deprivation. In the experiment 46 male 2,9, bairdi were considered. These animals were of similar age, body weight, and genetic history as those described in Experiment I. Colony and test room conditions were also as before with the exception that now a second test room was employed as well. This housed ten nest boxes with their attached running wheels (five on each of two tiers). Above each tier were two 60 watt and two 7 1/2 watt light bulbs. Again the 7 1/2 watt bulbs remained lit during the 12 hours of dark. The door separating the test rooms was kept open throughout the experiment to facilitate ventilation and to insure like conditions of temperature and humidity. This two-room unit was rela- tively soundproof with respect to activity in the adjacent colony room. The running wheel apparatus is pictured in Figure A. The wheel itself is eight inches in diameter with a three inch wide running 91 92 Figure A.--Running wheel apparatus with adjoining nest-box. 93 94 surface of perforated sheet metal. The wheel rotates around a bicycle hub fixed to the side of the nest box. This plywood box has a floor area 3 1/4 inches by 4 inches and is 3 1/4 inches high. 99_1ibitum access to water is available through a 100 ml graduated cylinder whose spout enters the back side of the nest box. The Purina food pellets are presented at the front in a removable plexiglas face-plate. The pellets are lodged in a 1 1/4 inch diameter plexiglas cylinder fixed horizontally into the center of the face-plate. Access to the food requires the animal to move a pendular aluminum door to one side of the aperture of the cylinder. To remove the food for deprivation, the entire unit is dislodged and replaced by a flat plexiglas face- plate. Two pellets are presented per day, as in Experiment I. The procedures involved with the rearing and with the adjust- ment to the test conditions (12:12 hour light-dark cycle, social isolation, and the switch to Purina food pellets) were the same as those previously employed. This experiment was conducted in two separate runs. In the first, 13 animals were randomly selected from Experiment I (animals tested without running wheels) and compared with an equal number of mice which had been tested concurrently in running wheels. Previous deprivation experience, although varied within a group, was matched across groups. The experimental timetable, including the test depri- vation (food available for the initial seven hours of the 12 hour dark period), was, therefore, that described in Table 1. Run 2 was made immediately following Run 1. It consisted of ten no-wheel mice and ten wheel mice. Again the deprivation experience 95 was matched across groups. The test deprivation was the same one used in Run 1. Survival vs. mortality is tested for independence from wheel vs, no wheel v19 the chi-square analysis of a 2 x 2 contingency table. Wheeled animals from both runs are pooled (n = 23) as are the no-wheel animals (n = 23). Homogeneity between runs is analyzed to satisfy the legitimacy of this procedure (Woolf, 1968). The results are pre- sented in the discussion portion of Experiment I. LITERATURE C ITED LITERATURE CITED Baldessari, B., Carlucci, F., Ticca, M., and Mariani, A. 1971, Analysis of the modifications of body composition and spontaneous activity in the meal-eating rat. Nutr. Rep. Inter. 3:149-158. Barnes, R. H. 1962. Nutritional implications of coprophagy. Nutr. Revs. 20(10):289-291. Barry, W. J. 1974. The effect of photoperiod and temperature on food hoarding in deer mice (Peromyscus). Unpub. Ph.D. dissertation, Michigan State University. 165 pp. Bauer, E. R., and Stebbins, M. M. 1972. Refeeding after differential severity of deprivation. Physiol. Behav. 9(3):489-49l. Baumeister, A., Hawkins, W. F., and Cromwell, R. T. 1964. Need states and activity level. Psychol. Bull. 61:438-453. Beer, J. R. 1961. Winter home ranges of the red-backed mouse and white-footed mouse. Jour. Mamm. 42:174—180. Behney, W. H. 1936. Nocturnal explorations of the forest deer-mouse. Jour. Mamm. 17:225-230. Blair, W. F. 1951. Population structure, social behavior and environ— mental relations in a natural population of the beach mouse, Peromyscus polionotus leucocephalus. Contr. Lab. Vert. Biol. Univ. Mich. 48:1—47. Bodenheimer, F. S. 1941. Observations on rodents in Herter's temperature gradient. Physiol. 2001. 14:186-192. Bodenheimer, F. S. 1952. Problems of animal ecology and physiology in deserts. In Desert Research. Res. Council Israel, Spec. Publ. No. 2, pp. 205-229. Bolles, R. C. 1963. Food deprivation and home cage behavior in the rat. Jour. Comp. Physiol. Psychol. 56:456-460. Bolles, R. C. 1965. Effects of deprivation conditions upon the rat's home-cage behavior. Jour. Comp. Physiol. Psychol. 60:244-248. 96 97 Brand, R. H. 1955. Abundance and activity of the wood mouse (Pero9yscus leucopus noveboracensis) in relation to the character of its habitat. Ph.D. Thesis, University of Michigan. 168 pp. Brant, D. H. 1962. Measures of the movement and population densities of small rodents. Univ. Calif. Publ. Zool. 622105-184 (from Sadleir, 1965). Burt, W. H. 1936. Nocturnal explorations of the forest deer mouse. Jour. Mamm. 17:225-230. Caldwell, L. D., and Connell, C. E. 1968. A précis on energetics of the old-field mouse. Ecology 49:542-548. Campbell, B. A., Teghtsoonian, R., and Williams, R. A. 1961. Activity, weight loss, and survival time of food-deprived rats as a function of age. Journ. Comp. Physiol. Psychol. 54:216-219. Cogshall, A. S. 1928. Food habits of the deer mice of the genus Peromyscus in captivity. Jour. Mamm. 9:217-221. Cohn, C., and Joseph, D. 1960. Role of rate of ingestion of diet on regulation of intermediary metabolism ("meal-eating" vs. "nibbling"). Metabolism 9:492-500. Cohn, C., Joseph, D., and Allweiss, M. D. 1962. Nutritional effects of feeding frequency. Am. J. Clin. Nutr. 11:356-361. Cook, S. F., and Hannon, J. P. 1954. Metabolic differences between three strains of 2, maniculatus. Jour. Mamm. 27:63-78. Cornish, E. R., and Mrosovsky, N. 1965. Activity during food deprivation and satiation in six species of rodent. Anim. Behav. 13:242-248. Criddle, S. 1950. The 2, 9, bairdi complex in Manitoba. Canad. Field-Nat. 64:169-177. Davenport, L. B., Jr. 1964. Structure of two Pero9yscus p91ionotus populations in old-field ecosystems at the AEC Savannah River Plant. Jour. Mamm. 45:95-113. Davis, D. E., and Golley, F. B. 1963. Principles 19_Mammalogv. Reinhold Publ. Corp., New York. 335 pp. Dawson, W. R. 1955. The relation of oxygen consumption to tempera- ture in desert rodents. Jour. Mamm. 36(4):543-552. 98 Denenberg, V. H., and Karas, G. G. 1959. Effects of differential infantile handling upon weight gain and mortality in the rat and mouse. Science 130:629-630. Denenberg, V. H., and Karas, G. G. 1961. Interactive effects of infantile and adult experience upon weight gain and mortality in the rat. Jour. Comp. Physiol. Psychol. 54:685-689. Dice, L. R. 1922. Some factors affecting the distribution of the Prairie Vole, Forest Deer Mouse, and Prairie Deer Mouse. Ecology 3:29-47. Dufort, R. H. 1964. The rat's adjustment to 23-, 47-, and 71-hour food deprivation schedules. Psychol. Rept. 17:663-669. Eisenberg, J. F. 1968. Behavior patterns. In: J. A. King (ed.), Biology s£_Peromyscus. American Society of Mammalogists. Pp. 451-495. Evans, H. L. 1971. Rat's activity: Influence of light-dark cycle, food presentation and deprivation. Physiol. Behav. 7:455-460. Fabry, P., and Braun, T. 1967. Adaptation to the pattern of food intake: Some mechanisms and consequences. Proc. Nutr. Soc. 26:144-152. Falls, J. B. 1953. Activity and local distribution of deer mice in relation to certain environmental factors. Unpubl. Ph.D. dissertation, University of Toronto (from Falls, 1968). Falls, J. B. 1968. Activity. In: J. A. King (ed.), Biologv s£_ Peromyscus. American Society of Mammalogists. Pp. 543-570. Fehrer, E. 1956. The effects of hunter and familiarity of locale on exploration. Jour. Comp. Physiol. Psychol. 49:549-552. Finger, F. W. 1951. The effect of food deprivation and subsequent satiation upon general activity in the rat. Jour. Comp. Physiol. Psychol. 44:557-564. Finger, F. W., Reid, L. S., and Weasner, M. H. 1960. Activity changes as a function of reinforcement under low drive. Jour. Comp. Physiol. Psychol. 53:385-387. Fink, A. J. 1973. Induction of torpor in the old-field mouse (Peromyscus polionotus). ASB Bull. 20(2):52 (Abs. 25012). Fitch, H. S. 1958. Home ranges, territories, and seasonal movements of vertebrates of the Natural History Reservation. Univ. Kans. Publ. Mus. Nat. Hist. 11:63-326. 99 French, R. L. 1956. Eating, drinking, and activity patterns in Peromyscus maniculatus sonoriensis. Jour. Mamm. 37:74-79. Fuller, W. A., Stebbins, L. L., and Dyke, G. R. 1969. Overwintering of small mammals near Great Slave Lake, northern Canada, in 1965-1966, and 1966-1967. Arctic 22:34-55. Gaertner, R. A., Hart, J. S., and Roy, 0. Z. 1973. Seasonal spon- taneous torpor in the white-footed mouse, Peromyscus leucopus. Comp. Biochem. Physiol. 45(14):169-181. Gentry, J. B., and Odum, E. P. 1957. The effect of weather on the winter activity of old-field rodents. Jour. Mamm. 38:72-77. Gentry, J. B., Golley, F. B., and McGinnis, J. T. 1966. Effect of weather on captures of small mammals. Am. Midl. Nat. 75: 526-530. Getz, L. L. 1959. Activity of Peromyscus leucopus. Jour. Mamm. 40:449-450. Gottschang, J. L. 1952. A life history study of the deer-mouse Peromyscus leucopus noveboracensis (Fisher) in the Ithaca, New York, region. Unpubl. Ph.D. thesis, Cornell University. Gross, C. G. 1968. General activity. In: Analvsis 92 Behavioral Change (Ed. L. Weiskranz), New York: Harper and Row. Pp. 91-106. Hadaway, H. C. 1972. Seasonal variation in the weight and composition of brown adipose tissue of Peromyscus maniculatus and Microtus californicus. Diss. Abstr. Int. B. Sci. Eng. 32(7): 4298. Hall, J. F. 1956. The relationship between external stimulation, food deprivation, and activity. Jour. Comp. Physiol. Psychol. 49:339-341. Halliday, M. 1968. Exploratory behavior. In: Analvsis 92 Behavioral Change (Ed. by L. Weiskranz) New York: Harper and Row. Pp. 107-126. Haresign, T. W. 1964. A study of the activity of Pero9vscus leucspus as influenced by certain environmental factors. Unpubl. Ph.D. thesis, University of Massachusetts. 86 pp. Harling, J., and Sadleir, R. M. F. S. 1974. The ecology of the deer mouse, Peromyscus maniculatus, in a coastal coniferous forest. III. Stomach-weight variation. Can. J. 2001. 52:1311-1315. 100 Hart, J. S. 1953. Energy metabolism of the white-footed mouse, Peromyscus leucopus noveboracensis, after acclimation at various environmental temperatures. Can. J. Zool. 31:99-105. Hart, J. S. 1957. Climate and temperature induced changes in the energetics of homeotherms. Rev. Can. Biol. 16:133-174. Hart, J. S., and Heroux, O. 1953. A comparison of some seasonal and temperature-induced changes in Peromyscus: cold resistance, metabolism, and pelage insulation. Can. J. Zool. 31:528-534. Hatfield, D. M. 1940. Activity and food consumption in Microtus and Peromyscus. Jour. Mamm. 21:29-36. Hays, H. A. 1958. The effect of microclimate on the distribution of small mammals in a tall grass prairie plot. Trans. Kans. Acad. Sci. 61:40-63 (from Falls, 1968). Hayward, J. S. 1965a. The gross body composition of six geographic races of Peromyscus. Can. J. Zool. 43:297-308. Hayward, J. S. 1965b. Metabolic rate and its temperature-adaptive significance in six geographic races of Peromyscus. Can. J. Zool. 43:309-323. Hayward, J. S. 1965c. Microclimate temperature and its adaptive significance in six geographic races of Peromyscus. Can. J. Zool. 43:341-350. Herter, K. 1934. Eine verbesserte Temperaturor-gel und ibre Anwendung auf Insekten and Sangtiere. Biol. Zentr. 54:487-507. Herter, K. 1936. Das thermotaktische Optimum bei Nagetieren, ein mendel des Art und Rassenmerkmal. Z. vergleich. Physiol. 23: 605-649. Hirth, H. F. 1959. Small mammals in old field succession. Ecology 40:417-425. Holéckova, E., and Fabry, P. 1959. Hyperphagia and gastric hyper- trophy in rats adapted to intermittent starvation. Brit. J. Nutr. 13:260-266. Hollifield, G., and Parson, W. 1962. Metabolic adaptations to a "stuff and starve" feeding program. I. Studies of adipose tissue and liver glycogen in rats limited to a short daily feeding period. J. Clin. Invest. 41:245-249. Houtc00per, W. C. 1972. Rodent seed supply and burrows of Peromyscus in cultivated fields. Proc. Ind. Acad. Sci. 1971. 101 Howard, W. E. 1949. Dispersal, amount of inbreeding, and longevity in a local population of prairie deermice on the George Reserve, southern Michigan. Contr. Lab. Vert. Biol. 43:52 pp. Howard, W. E. 1951. Relation between low temperature and available food to survival of small rodents. Jour. Mamm. 32:300-312. Howard, W. E., and Evans, F. C. 1961. Seeds stored by prairie deermice. Journ. Mamm. 42(2):260-263. Hudson, J. W., and Bartholomew, G. A. 1964. Terrestrial animals in dry heat: Estivators. In: Handbook of Physiology (Ed. by D. B. Dill) Sect. 4: Adaptation to the env1ronment. Amer. Physiol. Soc., Washington, D.C. Jakubezak, L. F. 1973. Frequency, duration, and speed of wheel running of rats as a function of age and starvation. Anim. Learning and Behav. 1:13-16. Johnson, M. S. 1926. Activity and distribution of certain wild mice in relation to biotic communities. Jour. Mamm. 7:245-277. Kagan, J., and Berkun, M. 1954. The reward value of running activity. Jour. Comp. Physiol. Psychol. 47:108. Kalabuchov, N. 1939. On ecological characters of closely related species of rodents. Zool. Zhur. 18:915-923. Karch, P. 1974. The role of experience in the behavioral-physiological adaptation of Peromyscus maniculatus bairdi to extended periods of a protein-free diet. Unpubl. Ph.D. dissertation, Michigan State University. 140 pp. Kavanau, J. L. 1968. Activity and orientational responses of white- footed mice to light. Nature 218:245-252. Kavanau, J. L. 1969. Influences of light on activity of the small mammals, Peromyscus spp., Tamias striatus, and Mustela rixosa. Experentia 25:208. Kendrick, D. C. 1973. The effects of infantile stimulation and inter- mittent fasting and feeding on life span in the black-hooded rat. Develop. Psychobiology 6:225-234. King, J. A. 1963. Maternal behavior in Peromyscus. In: H. E. Rheingold (ed.), Maternal Behavior 19_Mammals, New York: Wiley. Pp. 58-93. King, J. A. 1968. Psychology. In: J. A. King (ed.) Biology 92 Peromyscus. American Society of Mammalogists. Pp. 496-542. 102 Kirk, R. E. 1968. Experimental Design: Procedures for the Behavioral Sciences. Brooks/Cole Publ. Co., Belmont, California. 577 pp. Lat, J., and HolEckova, E. 1971. The effect of intermittent feeding and fasting on the nonspecific excitability level of the central nervous system in the rat. Physiol. Bohemoslovaca 20:441-445. Lawrence, D. H., and Mason, W. A. 1955. Intake and weight adjustments in rats to changes in feeding schedules. Jour. Comp. Physiol. Psychol. 48:267-271. Layne, J. N. 1971. Activity responses of two species of Peromyscus (Rodentia, Muridae) to varying light cycles. Oecologia (Berl.) 7:223-241. LeMagnen, J. 1971. Advances in studies of the physiological control and regulation of food intake. In: Progress in Physiological Psychology (Ed. E. Stellar and J. M. Sprague)-Vel. 4:203:261. Academic Press. LeMagnen, J., and Tallon, S. 1966. La périodicité spontanée de la prise d'aliments ad libitum du rat blanc. J. Physiol. Paris 58:323-349. Leveille, G. A. 1967. 19_vivo fatty acid synthesis in adipose tissue and liver of meal-fed rats. Proc. Soc. Exp. Biol. Med. 125:85-88. Levine, S. 1957. Infantile experience and consummatory behavior in adulthood. Jour. Comp. Physiol. Psychol. 50:609-612. Levine, S., and Otis, L. S. 1958. The effects of handling before and after weaning on the resistance of albino rats to later deprivation. Canad. J. Psychol. 12:103-108. Levitsky, D. A., and Barnes, R. H. 1972. Nutritional and environ- mental interactions in the behavioral development of the rat: long-term effects. Science 176:68. Linduska, J. P. 1950. Ecology and land-use relationships of small mammals on a Michigan farm. Mich. Dept. Cons., Fed. Aid Project 2R, ix + 144 pp. Lynch, G. R., and Lynch, C. B. 1973. PhotOperiodism and adaptive behavior in a small mammal. Nature 244:46-47. MacMillen, R. E. 1965. Aestivation in the cactus mouse, Peromyscqs eremicus. Comp. Biochem. Physiol. 16:227-248. 103 Marten, G. G. 1973. Time patterns of Peromyscus activity and their correlations with weather. Jour. Mamm. 54(1):l69-188. McCabe, T. T., and Blanchard, B. D. 1950. Three species of Peromyscus. Rood Associates, Publishers, Santa Barbara, California. 136 pp. McNab, B. K. 1963. A model of the energy budget of a wild mouse. Ecology 44(3):521-532. McNab, B. K. 1968. The influence of fat deposition on the basal rate of metabolism in desert homeotherms. Comp. Biochem. Physiol. 20:337-343. McNab, B. K., and Morrison, P. 1963. Body temperature and metabolism in subspecies of Peromyscus from arid and mesic environments. Ecol. Mono. 33:63-82. Morhardt, J. E. 1970. Body temperature of white-footed mice (Peromyscus spp.) during daily torpor. Comp. Biochem. Physiol. 33:423-439. Morhardt, J. E., and Hudson, J. W. 1966. Daily torpor induced in the white-footed mouse (Peromyscus spp.) by starvation. Nature 212:1046-1047. Morris, J. G. 1975. Population energetics of the prairie deer mouse Peromyscus maniculatus bairdi. Disr. Abst. Intr. (Jan.), 3672-B. Morrison, P., and Ryser, F. A. 1959. Body temperature in the white- footed mouse, Peromyscus leucopus noveboracensis. Physiol. Zool. 32:90-103. Mount, L. E., and Willmott, J. V. 1967. The relation between spontaneous activity, metabolic rate and the 24 hour cycle in mice at different environmental temperatures. J. Physiol. (London) 190:371-380. Mullen, R. K. 1971. Energy metabolism of Peromyscus crinitus in its natural environment. Jour. Mamm. 52(3):633-635. Murie, M. 1961. Metabolic characteristics of mountain, desert, and coastal populations of Peromyscus. Ecology 42:723-740. Myton, B. 1974. Utilization of space by Peromyscus leucopus and other small mammals. Ecology 55:277-290. Nelson, W., Cadotte, L., and Halberg, F. 1973. Circadian timing of single daily "meal" affects survival of mice. Proc. Soc. Exp. Biol. Med. 144:766-769. 104 Nisbett, R. E. 1972. Eating behavior and obesity in men and animals. Adv. Psychosom. Med. 7:173-193. O'Farrell, M. J. 1974. Seasonal activity patterns of rodents in a sagebrush community. Jour. Mamm. 55(4):809-823. Orr, H. D. 1959. Activity of white-footed mice in relation to environment. Jour. Mamm. 40:213-221. Owings, D. H., and Lockard, R. B. 1971. Different nocturnal activity patterns of Peromyscus californicus and Peromyscus eremicus in lunar lighting. Psychon. Sci. 22:63-64. Pengelley, E. T., and Fisher, K. C. 1968. Food habits and burrow associates of Peromyscus polionotus. Jour. Mamm. 49(3): 562-565. Premack, A. J., and Premack, D. 1958. Increments in ad libitum eating as a function of deprivation for wheel-turning_Eehavior. Paper read at Midwestern Psychological Association, Detroit, 1958. From Premack, 1959. Toward empirical behavior laws. 1. Positive reinforcement. Psychol. Rev. 66:219-233. Pruitt, W. 0., Jr. 1957. Observations on the bioclimate of some taiga mammals. Arctic 10:131-138. Pruitt, W. 0., Jr. 1959. Microclimates and local distribution of small mammals on the George Reserve, Michigan. Misc. Publ. Mus. 2001., University of Michigan 109:1-27. Prychodko, W. 1958. Effect of aggregation of laboratory mice (Mus musculus) on food intake at different temperatures. Ecology 39(3):500-503. Rice, L. Z., and Terman, C. R. 1972. Food hoarding in freely growing laboratory populations of prairie deermice (Peromyscus maniculatus bairdi). Virginia J. Sci. 23(3):115. Richter, C. P. 1927. Animal behavior and internal drives. Quart. Rev. Biol. 2:307-343. Routtenberg, A., and Kuznesof, A. W. 1967. Self-starvation of rats living in activity wheels on a restricted feeding schedule. Jour. Comp. Physiol. Psychol. 64:414-421. Sadleir, R. M. F. S. 1965. The relationship between agonistic behavior and population changes in the deer mouse, Peromyscus maniculatus (Wagner). J. Anim. Ecol. 34:331-352. Schoener, T. W. 1971. Theory of feeding strategies. In: Annual Review 92 Ecologv and Systematics. Vol. 2, Ann. Revs. Inc. Palo Alto, California. Pp. 369—404. 105 Scholander, P. F., Hock, R., Walters, V., and Irving, L. 1950. Adaptation to cold in arctic and tropical mammals and birds in relation to body temperature, insulation, and basal metabolic rate. Biol. Bull. 99:259-271. Sealander, J. A. 1951. Survival of Peromyscus in relation to environ- mental temperature and acclimation at high and low tempera- tures. Am. Midland Nat. 46:257-311. Sealander, J. A. 1952. Food consumption in Peromyscus in relation to air temperature and previous thermal experience. Jour. Mamm. 33:206-218. Sealander, J. A. 1953. Body temperature of white-footed mice in relation to environmental temperature and heat and cold stress. Biol. Bull. 104:87-99. Sidorowicz, J. 1960. Influence of the weather on capture of micro- mammalia. I. Rodents. Acta Theriologica 4:139-158. Sokal, R. R., and Rohlf, F. J. 1969. Biometry. The Principle and Practice s£_Statistics 19_Biolog1cal Research. W. H. Freeman and Co., San Francisco. 776 pp. Stebbins, L. L. 1971. Seasonal variations in circadian rhythms of deer mice in northwestern Canada. Arctic (Quebec) 24(2): 124-131. Stickel, L. F. 1968. Home range and travels. In: J. A. King (ed.), Biology 92_Peromyscus. American Society of Mammalogists. Pp. 373-411. Stinson, R. H. 1952. Effects of some environmental factors on the behavior of Peromyscus. Unpubl. Ph.D. dissertation, Univer- sity of Toronto (from Falls, 1968). Stinson, F. H., and Fisher, K. C. 1953. Temperature selection in deer mice. Can. J. Zool. 31:404-416. Storer, T. 1., Evans, F. C., and Palmer, F. G. 1944. Some rodent populations in the Sierra Nevada of California. Ecol. Monogr. 14:165-192. Strong, P. N. 1957. Activity in the white rat as a function of apparatus and hunger. Jour. Comp. Physiol. Psychol. 50: 596-600. Sumner, F. B., and Karol, J. J. 1929. Notes on the burrowing habits of Peromyscus polionotus. Jour. Mamm. 10:213-215. Swade, R. S., and C. S. Pittendrigh. 1967. Circadian locomotor rhythms of rodents of the arctic. Amer. Naturalist 101:431-466. 106 Symons, J. P. 1973. Wheel-running activity during ad-lib and food- deprivation conditions in four inbred mouse strains. Bull. Psychon. Sci. 1:78-80. Teghtsoonian, R., and Campbell, B. A. 1960. Random activity of the rat during food deprivation as a function of environmental conditions. Jour. Comp. Physiol. Psychol. 53:242-244. Tepperman, J., and Tepperman, H. M. 1958. Effects of antecedent food intake pattern on hepatic lipogenesis. Amer. J. Physiol. 193:55-64. Treichler, F. R., and Hall, J. F. 1962. The relationship between deprivation weight loss and several measures of activity. Jour. Comp. Physiol. Psychol. 55:346-349. Vose, R. N., and Dunlap, D. G. 1968. Wind as a factor in the local distribution of small mammals. Ecology 49(3):381-386. Weasner, M. H., Finger, F. W., and Reid, L. S. 1960. Activity changes under food deprivation as a function of recording device. Jour. Comp. Physiol. Psychol. 53:470-474. Weininger, 0. 1956. The effects of early experience on behavior and growth characteristics. Jour. Comp. Physiol. Psychol. 49:1-9. Welch, J. G. 1968. Effect of intermittent feeding on body com- position of mice. Lab. Anim. Care 18:596-601. Woolf, C. M. 1968. Principles g£_Biometry. D. Van Nostrand Co., Inc., Princeton, New Jersey. 359 pp. Wright, N. C. 1954. Domesticated animals inhabiting desert areas. In: Biology o£_Deserts, J. L. Cloudsley-Thompson (ed.), Inst. of 3101., Tavistock Square, London. Wyrwika, W. 1967. Conditioned behavioral analysis of feeding mechanisms. In: Handbook g£_Physiology. Sect. 6. Alimentary canal. Vol. 1. Control of food and water intake. Am. Physiol. Soc., Washington, D.C. Pp. 63-78. mug‘mulmwm Nlllfllll 083 30 i .1 3