(Accepted 31 March 19921
of performance in a six-point alley-T-maze was studied in s the error rate, the latency time for leaving the start box as well er of alleys entered per min were determined. From O&SO hrs on, a group of s was investigated every four hours. Each group received one trial per day. Home was continousty monitored over a M-day period. The results revealed the eunt of a circadian rhythm of maze performance in the course of training, which ificant as from the 4th day. The best maze performance was observed in the e dark phase, from 23.87 to 03.15 hrs. No indication of a direct relationship en home cage and ambulatory activity in the maze could be found.
Circadian rhythm; Spatial fearning; Mouse
Introduction Behavioral rhythms of motor activity in general, as well as food and water intake and (Rusak, 1981). However, have been extensively studied in rn,,ampals ,. ..
social behavior,
rhythms of learning behavior and memory have been less considered and investigations have often been confined to simple light-dark comparisons. The diversity of the applied performance tasks and the complexity of the involved sensory and motor corppone:lis yielded contradictory results.
Correspondence to: H.J. Hoffman, institute of Neurobiology and Brain Research, PF 1860, Magdeburg O-301 0, FRG.
78
Davies et ~1. (197.3 I, using a passive avoidance paradigm, found a petiormance the middle of the light phase. This was supported by data from Chields and Redfern who investigated 12 time-points. These results suggest a facilitation in the acquisitio passive
avoidance task during phases of rest or low motor activity. As regards
used active avoidance learning, some facts point to a better performance in the (Bialik et al., 1984; Cataia et al., 1985; Novakova et al., 3983). For this high activity may have a facilitating effect. It is to be supposed that t rhythm in motor activity contributed to the observed circadian alt performance. rhythmicity
Nevertheless, of motor
evaluating six time-points Learning
Ghiselli and Patton (1976,
activity
and performance
were abfe to disca
in a free operant
avoidance
daily.
paradigms with
positive reinforcement
may, to some extent,
circep
influences of motor activity but, on the other hand, raise new questions of d in motivation (Hunsicker and Mellgren, 1977). Hostetter (39693 reported on di performance during the light and the dark phase in an appetitively rei learning task. However, mice.
he found contradictory
results
for two dstierent inb
Therefore the aim of the present paper was to demonstrate a circadian rhythm of spatial learning with positive reinforcement and its relationship to activity or motivation. To minimize influences of stress, the animals were not deprived of food or water. Reinforcement consisted of returning the an’mal to the home cage at the goal of t Additionally, ambulatory activity in the maze as well as home cage activity was and assessed for correlations to maze learning. For reduction of intragroup variance we selected the C57Bi/6 inbred mouse strain, which has been found to be a well-performing strain in spatial learning tasks (Schwegler and Buselmaier, 1981).
Animals After weaning, male C5761/6 Ola mice were housed in groups of ten in a temperature controlled room (22 + 2°C) maintained at a 12/12 light/dark cycle (light from 06.00 to 18.00 hrs). During the dark phase the animal room was iit by dim red bulbs. Food and water were supplied ad libitum. At the time of training, animals were 113 + 10 days oid.
Home cage activity Home cage activity was determined for two groups (IO days per group) on an ‘Animex’ activity meter (Farad Electronics, I lagersten, Sweden) 14 days prior to the experiment.
Exploratory
activity
Latency time for leaving the start box of the maze and ambulation scores (expressed as number of alleys entered per min), both examined on the first day of experiment, were taken as a measure of exploratory activity in a novel environment.
lea
tey-T-maze4LhRRtL, Fig. 1) covered with a removable Plexiglass ransported to the test room in their home cage and transferred into a e expebiment started 5 min later. Reinforcement consisted of returning to e goal of the maze, thus ensuring a relatively low stress level. On their of their home cage the animals were left undisturbed for 1 min and were then returned to the cage. Animals were randomly assigned to six time groups (04.00, 08.00, 12.00, 16.00, 20.00 and 00.00 hrs, respectively). Each group was tested once per day for 9 days. Between individual trials the maze was wiped out and dried. To avoid olfactory cues, blind alleys and start-goal alleys were interchanged regularly. Training of a group lasted for less than 1 h. In the dark p ase, experiments were carried out under dim red light. Learning performance was expressed as number of errors (entrances in blind alleys or return to the start box).
Statistical evaluation Since the Gauss distribution was found to disappear in the course of the experiment the median and three nonparametric statistical methods were used in parallel (Median, H-test by Kruskal-Wallis, multiple comparison of Nemeny, and U-test by Mann-Whitney). Cosine functions with periods of 24 h, 12 h and 8 h, respectively, were fitted to the original data by non-linear regression to determine the dominant spectral component and the corresponding rhythmic parameters mesor (24-h mean), amplitude and acrophase,
Results The number of errors obtained from the first three trials did not display any significant differences between the six groups. However, between trials two and four a remarkably improved learning could be observed in the dark phase, improving further up to tnal 9 (Fig. 2). On the contrary, the light phase performance remained on the level of day three. This resulted in a significant maze learning rhythm from day 4 which was accompanied by an increase in amplitude from 2.1 errors on day 4 to 2.7 on day 6 (Table I). The gradual
80
Fig. 2. Mean
learning performance, expressed as rrumber of errors, during the Dight and dark p Each value represents the mean of three groups. * P < 8.01 (U-test)
appearance
of this rhythm
was supported
by the calculated
correlation
coefficient,
continu-
ously rising from 0.17 on the first day to 0.61 at the end of the experi With the exception of day 7, acrophases (maximum number of er to occur around noon (e.g. between 11 .Oi’ and 15.15 hrs on day 9). performance could be observed 12 h (half a period) later in the mid from 23.07 to 03.15 hrs. On day 7, the rhythm was flatter and the no two smail maxima at 08.00 and 16.00 hrs, respectively. Consequent1 al 24 h-cosine function but the 12 h-cosine wave yielded the best correlation, reduced amplitude (2.0). On days 8 and 9, the 24 h-component returned to be powerful one and the amplitude increased again, reaching maximum levels on day 8. The 24 h-mean declined from 9.1 errors on day 1 to 3.8 on day 9. The computed acrophases showed a tendency towards phase delay during the last training days. In addition to t results of the nonlinear regression, rhythm in maze performance was confirmed by nonparametric tests. These tests revealed significant group differences beginning on day 4 W-test) and day 6 (H-test and Nemeny comparison), respectively (Fig. 2), up to the end of the experiment. The development of the circadian rhythm of maze learning is depicted in
TABLE 1 Farameters
of the rhythm
Day
Mesor
1 2
9.13 11.20
3 4
6.11 5.62kl.22
5 6
4.92kO.94 4.99kO.87
7 8 9
5.28kO.96 4.82+0.83 3.79kO.83
of maze learning errors Amplitude
Period
Acrophase
n.s. 2.09k1.73 2.38+X35
24 24
11.05+3.10 11.46+2.07
2.73k1.23 1.68f1.35
24 12
11.32k1.43 10.06+3.04
3.20f1.17
24
2.17+1.17
24
12.13f1.23 13.11f2.04
n.s. n.s.
er
the circadian rhythm of maze performance (number of errors) during nine ys. Results are plotted two-fold. Shaded areas represent dark phase.
can be seen, the high error rate during the dark phase around midnight on day I ly over the next few days and a clear-cut minimum of errors, i.e. maximum rmance develops during the dark phase. We used two parameters to characterize test related activity: latency for leaving the start box and ambulatory activity in the maze. No significant differences in latency time for leaving the start box could be found between experimental groups. Scores for ambulatory activity, expressed as number of alleys entered per minute, were slightly higher in the dark phase but this difference could be verified statistically only for the 08.00 hrs group (10.0 alleys/min) in comparison to the 2&O@hrs and the OO.OO-hrs group (17.8 alleys/min for both; P < 0.01). From the first to the second trial, latency time dropped down from 11 s to about 1 s in all groups. Correspondingly, in trial 2 a doubling of running speed, which remained constant afterwards, could be observed at all time-points. Home cage activity showed the usual circadian periodic&y with maxima around dawn and dusk and a preponderance of activity in the dark phase after lights-off (Fig. 4). Thus the group investigated at 20.00-hrs was tested in a relatively high activity phase which was found to be lower (but still above 100%) in the group tested at 00.00 hrs, the time of the best maze performance. The good maze performance at 04.00 hrs matched the 100%
time (hrl
Fig. 4. Rhythmicity of home cage activity (counts per h; mean + S.E.M).
82
activity level. The 12.00-hrs and 16.00-hrs groups were in a phase of increasing activity but clearly below the 100% level. The 08.00-hrs group matched low activiv.
The results demonstrate
a clear circadian rhythm in maze ge
error rate approximately 5-7 h after lights-off. AS indicated coefficrent this rhythm gradually appears in the course of the statistically significant on day 4. Taking account of the chosen 4~%I h seems to be the dominant spectral component in the rhyth performance. However, on day 7 the maze performance displayed a the I 2-h cosine wave was computed to yield the best correlation. SUC function could not be expected to provide an adequate fit o indicates that beyond the circadian range (24 +_4 h) rhythms periods might contribute, at least temporarily, to the generatio maze performance. Our results confirm and extend the findings of Hostetter (1966) who-after a simple reported a better maze performance of the C57B1/6 strain during light : dark comparisonthe light phase. Food reward hc cf been used as reinforcement, however, without any attempt to compare the motivational strength between the two groups. By restricting to only two time-points, the probzhility of finding the peaks and troughs of rhythmic behavior is very low. Therefore, phase differences of a few hours may falsely show an inverted rhythm, as reported for the DBA/2 strain in his paper. In our experiment the latency time for leaving the start box and the number of alleys entered per min at the first trial were used to characterize exploratory behavior in a novel environment. Whiie the latency time did not vary significantly between the individuai titne-points, there was a slight tendency towards higher ambulatory activitv during the dark phase. Rats and mice exhibited a higher exploration in novel environment during the dark phase (Catala et al., 1985; Kuribara and Tadokord, 1982). However, the weak light/dar difference observed in our experiments cannot explain the strong variation in learning performance. During the following days an equally rapid decline in latency time and an increase in running speed were o)jserved for all groups (data not shown). Thus, the analysis of motor activity gave no indication of any intergroup differences in motivational strength. Whereas no rhythms of activity could be established in the maze, home cage activity displayed the expecte: time course with higher activity in the dark phase and peaks around dawn and dusk (Meyer-Lohmann, 1355; Weinert and Schuh, 1984). The comparison between the rhythms of maze learning behavior and home cage activity revealed two main differences: (1) while the maze learning rhythm was unirnodal, the home cage activity displayed a bimodal pattern; and (2) the phase of the best maze performance did not match the peaks of home cage activity. Since a relationship between learning performance and home cage activity or ambulatory activity could not be observed in the maze, we are led to conclude that the described rhythm in maze performance represents a true rhythm of learning and/or memory processes.
83
References D.C.S., 1884. Defidt~ IEImndita-n& avoidance rqxnding cenFd norepinephrine depiction are dependent the diurnal cycle. Eehav. Neurosci., 98: 847-557. wtanueva, m Orbavoidi%Ke
P. and v-w
Giner, j.M.,
on p&surgical 19385. Effect of
havior in the ma&erat. Phy;ioL Behav., 34:
I. A ekadian rhythm in passive avoidance behaviour: The effect of europharmacology, P.H.,
1973.
20: ?365-T 366.
A 24-hour
rhythm in passive-avoidance
maI variation in performance of
free-operant avoidance
cts on teaming and open field activity. Psychon. Sci., 5: 77. Muftipfe deficits in the retention of aE appetitivety avior across a 24-h period in rats. Anim. Learn. Behav., 5: 14-16. Tadokoro, S., 7 982. Circadian variation in methamphetamine- and apomorphineinduced increase in am&uiaFory activity in mice. Pharmacol. Biochem. and Behav., 17: 12511256. Meyer-Lohmann, J., 1955. ibber den Einfluss taglicher Futtergaben auf die 24-Stunden-Periodik det lokomotorischen AktivitZt weisser MZuse. Pfliigers Arch. ges. Physiol., 260: 292-305. Novakova, V., Sterc, j. and Knez, R., 1983. The active avoidance reaction of laboratory rats: differences between experiments carried out in the phase of motor acitivity and inactivity. Physiol. Bohemoslov., 32: 38-44. Rusak, B., 1981. Vertebrate
Behavioral Rhythms. in: J. Aschoff (Editor), Handbook
of Behavioral
Neurobiology, Vol. 4, Biological Rhythms, Plenum Press, N.Y., pp. 183-213. Schwegler, H. and Buselmaier, W., 1981, Behavior gene&‘r analysis of water T-maze learning in inbred strains of mice, their hybrids, and seleted second generation crosses. Psychol. Res., 43: 335-345. Weiner-t, D. and Schuh, J. 1984. Untersuchungen zur Entwicklung der Zeitstruktur ausgewzhlter Parameter im Verlaufe der postnatalen Ontogenese im Tagesmuster der lokomotorischen Aktivit& der Futter- und Wasseraufnahme.
Zool. jb. Anat., 111: 133-l
46.
of performance in a six-point alley-T-maze was studied in s the error rate, the latency time for leaving the start box as well er of alleys entered per min were determined. From O&SO hrs on, a group of s was investigated every four hours. Each group received one trial per day. Home was continousty monitored over a M-day period. The results revealed the eunt of a circadian rhythm of maze performance in the course of training, which ificant as from the 4th day. The best maze performance was observed in the e dark phase, from 23.87 to 03.15 hrs. No indication of a direct relationship en home cage and ambulatory activity in the maze could be found.
Circadian rhythm; Spatial fearning; Mouse
Introduction Behavioral rhythms of motor activity in general, as well as food and water intake and (Rusak, 1981). However, have been extensively studied in rn,,ampals ,. ..
social behavior,
rhythms of learning behavior and memory have been less considered and investigations have often been confined to simple light-dark comparisons. The diversity of the applied performance tasks and the complexity of the involved sensory and motor corppone:lis yielded contradictory results.
Correspondence to: H.J. Hoffman, institute of Neurobiology and Brain Research, PF 1860, Magdeburg O-301 0, FRG.
78
Davies et ~1. (197.3 I, using a passive avoidance paradigm, found a petiormance the middle of the light phase. This was supported by data from Chields and Redfern who investigated 12 time-points. These results suggest a facilitation in the acquisitio passive
avoidance task during phases of rest or low motor activity. As regards
used active avoidance learning, some facts point to a better performance in the (Bialik et al., 1984; Cataia et al., 1985; Novakova et al., 3983). For this high activity may have a facilitating effect. It is to be supposed that t rhythm in motor activity contributed to the observed circadian alt performance. rhythmicity
Nevertheless, of motor
evaluating six time-points Learning
Ghiselli and Patton (1976,
activity
and performance
were abfe to disca
in a free operant
avoidance
daily.
paradigms with
positive reinforcement
may, to some extent,
circep
influences of motor activity but, on the other hand, raise new questions of d in motivation (Hunsicker and Mellgren, 1977). Hostetter (39693 reported on di performance during the light and the dark phase in an appetitively rei learning task. However, mice.
he found contradictory
results
for two dstierent inb
Therefore the aim of the present paper was to demonstrate a circadian rhythm of spatial learning with positive reinforcement and its relationship to activity or motivation. To minimize influences of stress, the animals were not deprived of food or water. Reinforcement consisted of returning the an’mal to the home cage at the goal of t Additionally, ambulatory activity in the maze as well as home cage activity was and assessed for correlations to maze learning. For reduction of intragroup variance we selected the C57Bi/6 inbred mouse strain, which has been found to be a well-performing strain in spatial learning tasks (Schwegler and Buselmaier, 1981).
Animals After weaning, male C5761/6 Ola mice were housed in groups of ten in a temperature controlled room (22 + 2°C) maintained at a 12/12 light/dark cycle (light from 06.00 to 18.00 hrs). During the dark phase the animal room was iit by dim red bulbs. Food and water were supplied ad libitum. At the time of training, animals were 113 + 10 days oid.
Home cage activity Home cage activity was determined for two groups (IO days per group) on an ‘Animex’ activity meter (Farad Electronics, I lagersten, Sweden) 14 days prior to the experiment.
Exploratory
activity
Latency time for leaving the start box of the maze and ambulation scores (expressed as number of alleys entered per min), both examined on the first day of experiment, were taken as a measure of exploratory activity in a novel environment.
lea
tey-T-maze4LhRRtL, Fig. 1) covered with a removable Plexiglass ransported to the test room in their home cage and transferred into a e expebiment started 5 min later. Reinforcement consisted of returning to e goal of the maze, thus ensuring a relatively low stress level. On their of their home cage the animals were left undisturbed for 1 min and were then returned to the cage. Animals were randomly assigned to six time groups (04.00, 08.00, 12.00, 16.00, 20.00 and 00.00 hrs, respectively). Each group was tested once per day for 9 days. Between individual trials the maze was wiped out and dried. To avoid olfactory cues, blind alleys and start-goal alleys were interchanged regularly. Training of a group lasted for less than 1 h. In the dark p ase, experiments were carried out under dim red light. Learning performance was expressed as number of errors (entrances in blind alleys or return to the start box).
Statistical evaluation Since the Gauss distribution was found to disappear in the course of the experiment the median and three nonparametric statistical methods were used in parallel (Median, H-test by Kruskal-Wallis, multiple comparison of Nemeny, and U-test by Mann-Whitney). Cosine functions with periods of 24 h, 12 h and 8 h, respectively, were fitted to the original data by non-linear regression to determine the dominant spectral component and the corresponding rhythmic parameters mesor (24-h mean), amplitude and acrophase,
Results The number of errors obtained from the first three trials did not display any significant differences between the six groups. However, between trials two and four a remarkably improved learning could be observed in the dark phase, improving further up to tnal 9 (Fig. 2). On the contrary, the light phase performance remained on the level of day three. This resulted in a significant maze learning rhythm from day 4 which was accompanied by an increase in amplitude from 2.1 errors on day 4 to 2.7 on day 6 (Table I). The gradual
80
Fig. 2. Mean
learning performance, expressed as rrumber of errors, during the Dight and dark p Each value represents the mean of three groups. * P < 8.01 (U-test)
appearance
of this rhythm
was supported
by the calculated
correlation
coefficient,
continu-
ously rising from 0.17 on the first day to 0.61 at the end of the experi With the exception of day 7, acrophases (maximum number of er to occur around noon (e.g. between 11 .Oi’ and 15.15 hrs on day 9). performance could be observed 12 h (half a period) later in the mid from 23.07 to 03.15 hrs. On day 7, the rhythm was flatter and the no two smail maxima at 08.00 and 16.00 hrs, respectively. Consequent1 al 24 h-cosine function but the 12 h-cosine wave yielded the best correlation, reduced amplitude (2.0). On days 8 and 9, the 24 h-component returned to be powerful one and the amplitude increased again, reaching maximum levels on day 8. The 24 h-mean declined from 9.1 errors on day 1 to 3.8 on day 9. The computed acrophases showed a tendency towards phase delay during the last training days. In addition to t results of the nonlinear regression, rhythm in maze performance was confirmed by nonparametric tests. These tests revealed significant group differences beginning on day 4 W-test) and day 6 (H-test and Nemeny comparison), respectively (Fig. 2), up to the end of the experiment. The development of the circadian rhythm of maze learning is depicted in
TABLE 1 Farameters
of the rhythm
Day
Mesor
1 2
9.13 11.20
3 4
6.11 5.62kl.22
5 6
4.92kO.94 4.99kO.87
7 8 9
5.28kO.96 4.82+0.83 3.79kO.83
of maze learning errors Amplitude
Period
Acrophase
n.s. 2.09k1.73 2.38+X35
24 24
11.05+3.10 11.46+2.07
2.73k1.23 1.68f1.35
24 12
11.32k1.43 10.06+3.04
3.20f1.17
24
2.17+1.17
24
12.13f1.23 13.11f2.04
n.s. n.s.
er
the circadian rhythm of maze performance (number of errors) during nine ys. Results are plotted two-fold. Shaded areas represent dark phase.
can be seen, the high error rate during the dark phase around midnight on day I ly over the next few days and a clear-cut minimum of errors, i.e. maximum rmance develops during the dark phase. We used two parameters to characterize test related activity: latency for leaving the start box and ambulatory activity in the maze. No significant differences in latency time for leaving the start box could be found between experimental groups. Scores for ambulatory activity, expressed as number of alleys entered per minute, were slightly higher in the dark phase but this difference could be verified statistically only for the 08.00 hrs group (10.0 alleys/min) in comparison to the 2&O@hrs and the OO.OO-hrs group (17.8 alleys/min for both; P < 0.01). From the first to the second trial, latency time dropped down from 11 s to about 1 s in all groups. Correspondingly, in trial 2 a doubling of running speed, which remained constant afterwards, could be observed at all time-points. Home cage activity showed the usual circadian periodic&y with maxima around dawn and dusk and a preponderance of activity in the dark phase after lights-off (Fig. 4). Thus the group investigated at 20.00-hrs was tested in a relatively high activity phase which was found to be lower (but still above 100%) in the group tested at 00.00 hrs, the time of the best maze performance. The good maze performance at 04.00 hrs matched the 100%
time (hrl
Fig. 4. Rhythmicity of home cage activity (counts per h; mean + S.E.M).
82
activity level. The 12.00-hrs and 16.00-hrs groups were in a phase of increasing activity but clearly below the 100% level. The 08.00-hrs group matched low activiv.
The results demonstrate
a clear circadian rhythm in maze ge
error rate approximately 5-7 h after lights-off. AS indicated coefficrent this rhythm gradually appears in the course of the statistically significant on day 4. Taking account of the chosen 4~%I h seems to be the dominant spectral component in the rhyth performance. However, on day 7 the maze performance displayed a the I 2-h cosine wave was computed to yield the best correlation. SUC function could not be expected to provide an adequate fit o indicates that beyond the circadian range (24 +_4 h) rhythms periods might contribute, at least temporarily, to the generatio maze performance. Our results confirm and extend the findings of Hostetter (1966) who-after a simple reported a better maze performance of the C57B1/6 strain during light : dark comparisonthe light phase. Food reward hc cf been used as reinforcement, however, without any attempt to compare the motivational strength between the two groups. By restricting to only two time-points, the probzhility of finding the peaks and troughs of rhythmic behavior is very low. Therefore, phase differences of a few hours may falsely show an inverted rhythm, as reported for the DBA/2 strain in his paper. In our experiment the latency time for leaving the start box and the number of alleys entered per min at the first trial were used to characterize exploratory behavior in a novel environment. Whiie the latency time did not vary significantly between the individuai titne-points, there was a slight tendency towards higher ambulatory activitv during the dark phase. Rats and mice exhibited a higher exploration in novel environment during the dark phase (Catala et al., 1985; Kuribara and Tadokord, 1982). However, the weak light/dar difference observed in our experiments cannot explain the strong variation in learning performance. During the following days an equally rapid decline in latency time and an increase in running speed were o)jserved for all groups (data not shown). Thus, the analysis of motor activity gave no indication of any intergroup differences in motivational strength. Whereas no rhythms of activity could be established in the maze, home cage activity displayed the expecte: time course with higher activity in the dark phase and peaks around dawn and dusk (Meyer-Lohmann, 1355; Weinert and Schuh, 1984). The comparison between the rhythms of maze learning behavior and home cage activity revealed two main differences: (1) while the maze learning rhythm was unirnodal, the home cage activity displayed a bimodal pattern; and (2) the phase of the best maze performance did not match the peaks of home cage activity. Since a relationship between learning performance and home cage activity or ambulatory activity could not be observed in the maze, we are led to conclude that the described rhythm in maze performance represents a true rhythm of learning and/or memory processes.
83
References D.C.S., 1884. Defidt~ IEImndita-n& avoidance rqxnding cenFd norepinephrine depiction are dependent the diurnal cycle. Eehav. Neurosci., 98: 847-557. wtanueva, m Orbavoidi%Ke
P. and v-w
Giner, j.M.,
on p&surgical 19385. Effect of
havior in the ma&erat. Phy;ioL Behav., 34:
I. A ekadian rhythm in passive avoidance behaviour: The effect of europharmacology, P.H.,
1973.
20: ?365-T 366.
A 24-hour
rhythm in passive-avoidance
maI variation in performance of
free-operant avoidance
cts on teaming and open field activity. Psychon. Sci., 5: 77. Muftipfe deficits in the retention of aE appetitivety avior across a 24-h period in rats. Anim. Learn. Behav., 5: 14-16. Tadokoro, S., 7 982. Circadian variation in methamphetamine- and apomorphineinduced increase in am&uiaFory activity in mice. Pharmacol. Biochem. and Behav., 17: 12511256. Meyer-Lohmann, J., 1955. ibber den Einfluss taglicher Futtergaben auf die 24-Stunden-Periodik det lokomotorischen AktivitZt weisser MZuse. Pfliigers Arch. ges. Physiol., 260: 292-305. Novakova, V., Sterc, j. and Knez, R., 1983. The active avoidance reaction of laboratory rats: differences between experiments carried out in the phase of motor acitivity and inactivity. Physiol. Bohemoslov., 32: 38-44. Rusak, B., 1981. Vertebrate
Behavioral Rhythms. in: J. Aschoff (Editor), Handbook
of Behavioral
Neurobiology, Vol. 4, Biological Rhythms, Plenum Press, N.Y., pp. 183-213. Schwegler, H. and Buselmaier, W., 1981, Behavior gene&‘r analysis of water T-maze learning in inbred strains of mice, their hybrids, and seleted second generation crosses. Psychol. Res., 43: 335-345. Weiner-t, D. and Schuh, J. 1984. Untersuchungen zur Entwicklung der Zeitstruktur ausgewzhlter Parameter im Verlaufe der postnatalen Ontogenese im Tagesmuster der lokomotorischen Aktivit& der Futter- und Wasseraufnahme.
Zool. jb. Anat., 111: 133-l
46.
