- Email: [email protected]
Simultaneous colour discrimination in chicks is improved by brief exposure to light
Anim. Behav ., 1978,26, 1 -5
SIMULTANEOUS COLOUR DISCRIMINATION IN CHICKS IS IMPROVED BY BRIEF EXPOSURE TO LIGHT By JEREMY JON CHERFAS Sub-Department of Animal Behaviour, Madingley, Cambridge CB3 8AA
Abstract . Dark-reared chicks were exposed to white light for 1 h at age 24 h. At 48 h these chicks, and others kept in the dark throughout, were given the opportunity to obtain a heat reward by choosing the correct arm in a Y maze . The discriminanda were red and blue illuminated circles, each being the positive stimulus for half the chicks . Light-exposure resulted in faster learning, especially when the positive stimulus was blue . This is taken as further support for the view that brief exposure to light activates the visual system non-specifically, affecting many kinds of visually guided behaviour . Brief exposure to light is known to have an effect on visually mediated behaviour in the chick . For example, imprinting takes place more readily (Bateson & Seaburne-May 1973) and results in a stronger preference (Bateson & Wainwright 1972) if chicks are first exposed to light for 30 min . Avoidance learning, too, is enhanced by prior exposure to light (Cherfas 1977) . These, and other, results have been interpreted as evidence for a non-specific activation of the visual system by prior exposure (Bateson 1976) . In particular, Bateson & Wainwright (1972) suggested that the light-exposed chick is able to learn faster because its visual system is already, as it were, primed to process incoming information. Dimond & Adam (1972) said much the same thing, that light `acts on the learning mechanism itself and brings it into a state of readiness' (p . 419) . Further evidence of the non-specific activating effects of light exposure was provided in an earlier experiment (Cherfas 1977) . Light-exposed and dark-reared chicks were allowed to peck at a coloured target coated with methyl anthranilate . Subsequent retention tests, consisting of successive presentations of the training target colour and a neutral target, revealed that lightexposed chicks increase their pecks to the neutral colour and decrease their pecks to the training colour, but dark-reared chicks do not alter their behaviour as a result of their experience with methyl anthranilate . This last study, and those concerned with aspects of imprinting, suggest that light-exposed birds pay more attention to visual stimuli and thereby learn more about those stimuli . If this is the case one would expect enhanced performance on any task where the chick is required to learn about visual stimuli . The present study had two aims ; to extend the `enhanced learning'
observations to a simultaneous discrimination based on conventional positive reward, and to look at performance in a task involving learning over a number of trials . A positive reward to motivate the behaviour of young chicks is not easy to find . They need neither food nor water, so the performance of a group of chicks working for these rewards is very variable . Working for an imprinting object involves complex perceptual rewards, and it would be difficult to establish a discriminative stimulus independent of the imprinting object . Young chicks, however, are not very good at regulating their body temperature (Wekstein & Zolman 1967) and heat has proved to be a potent reward for young chicks (Zolman & Martin 1967 ; Zolman and Becker 1968 ; Zolman 1968) . Dark-reared chicks were exposed to light, or held in darkness, for 1, h . Twenty-four hours later, when the chicks were 2 days old, they were given 20 trials in which to learn to approach a particular coloured stimulus in a Y maze for a heat reward . Methods Rearing and exposure techniques have been described elsewhere (Cherfas 1977). Briefly, chicks were removed from a dark hatching incubator about 6 h after hatching and marked to enable subsequent recognition . They were transferred to individual compartments in a second light-tight incubator until they were 24 h old . Half the birds were then moved to similar compartments in an incubator with a transparent lid. A 100-W tungsten bulb behind an opal screen 50 cm above the birds was switched on for 1 h, after which the chicks were returned to their compartments in the light-tight incubator . The other half of each batch was moved between compartments in light-tight incubators . There 1
2
ANIMAL BEHAVIOUR, 26,
are thus two groups of birds, both of which received the same non-visual stimulation, but only one of which received the hour of exposure to light ; the two groups will be called the dark and light groups . Chicks were assigned to the groups at random . Training began about 24 h after the end of priming. The apparatus was a discrimination maze divided into a start box, a choice area, and two goal boxes . The maze was constructed of hardboard with a wire-mesh floor . Except in the goal boxes there was a false floor of hardboard beneath the wire mesh . The start box was 15 . 2 x 15 . 2 cm and opened onto a choice area 30 . 5 x 58 . 4 cm . The goal boxes were each 15 . 2 x 17 . 8 cm, and the whole apparatus was 38 cm deep . The internal walls and floors were painted matt black . On the rear wall of each goal box was a circular hole, 10 . 2 cm in diameter, with its bottom edge 3 . 8 cm above the wire mesh floor . Behind the hole was an opal diffusing screen which could be illuminated from behind by a 21-W tungsten bulb . Coloured acetate filters could be inserted between the diffusing screen and the hole, so that the rear wall of each box contained an illuminated red or blue circle . The colour of the stimulus in the goal box was easily changed by changing the filters . The room in which the maze was situated was in complete darkness, and the only light in the maze was provided by the stimuli. Ambient temperature in the room was about 15 C . Heat was provided by a fan heater mounted 30 cm below the goal box floor . For each correct choice the heater was switched on for 15 s . This raised the ambient temperature in the goal box from 15 to 25 C in about 2 s . Temperature returned to baseline within 15 s of the end of heat presentation. (All measurements were made with a Grant continuous recording thermometer connected to a probe 2 cm above the goal box floor.) Training involved pre-training and acquisition of a simultaneous colour discrimination . Pretraining consisted of four trials in which the chick was shaped to approach the goal box to receive heat . The subject was removed from the dark incubator and placed in a box cooled to 10 C . After 1 min it was removed and placed in one of the goal boxes . The light, without filter, was switched on and the chick was given 15 s of heat. After an inter-trial interval of 30 s, spent in the dark in the cooled box, the subject was placed in the other goal box and given another heat reward. On the remaining two trials the
1
chick was placed directly in front of the goal box and rewarded when it entered the box . Acquisition trials began 30 s after the end of the last pre-training trial . Subjects were required to approach and enter the goal box containing a particular coloured stimulus. For half the subjects the positive stimulus was the red circle and the negative the blue (referred to as red chicks) . For the other half (referred to as blue chicks) the significance of the colours was reversed . The position of the stimuli was varied for each acquisition trial according to a predetermined schedule (Fellows 1967) . On each training trial the chick was placed in the start box facing the choice area . The stimuli were illuminated and the latency before the chick left the start box was recorded. Time to run down the alley and initial choice were also noted . If the initial choice was correct the chick immediately received 15 s of heat reward. A correction method was used . If the first choice was incorrect the chick was allowed to retrace or, after 60 s, was guided into the other goal box where it received a reward . After each reward had been delivered the stimuli were switched off and the chick was moved into the cooled box where it remained in the dark for the intertrial interval of 30 s . All chicks were given 20 acquisition trials with a 15-s heat reward on every trial. Light and dark chicks were trained alternately with red or blue stimulus positive, according to a predetermined schedule . A total of 32 birds from four batches were trained, eight in each group (light/red, light/blue, dark/red, dark/blue) . Results Data were analysed using an analysis of variance, with repeated measures where appropriate . The r Llght
Dark Bed BIue
1-5
6-10
11-15
16201rials
Mean time to leave Start-box
1-5
6-10
11-15
16dOtrials
Mean time to run down Alley
Fig . 1 . Mean values for (left) time to leave start box and (right) time to run down alley . Scores on blocks of five consecutive trials have been averaged to give four values for each subject . For both measures only the main effect of blocks of trials is significant .
CHERFAS : LIGHT EXPOSURE AND DISCRIMINATION LEARNING
factors included were exposure (light or dark), positive colour (red or blue) and training trials (scores were averaged over four blocks of five successive trials) . Transformations (square root or arcsin) were performed prior to the analyses but the figures all show untransformed data points . Figure 1 shows the means, across blocks of five trials, of time to leave the start box and time to run down the alley for the four groups . Scores were subjected to a square root transform and analysed with a 2 x 2 x 4 repeated measures analysis of variance (Winer 1971, p . 559) . All chicks left the start box more quickly as training progressed (F = 9 .908 ; df = 3, 84 ; P<0.001) and also ran from the start box to the goal box faster (F = 11 .574; df = 3, 84 ; P<0 . 001) . There were no other significant effects . StudentNewman-Keuls tests were used to locate the effect due to trials . For both measures the score for the first block of trials was different from the other three blocks, which did not differ among themselves (Leave start box ; block 1 versus block 2 ; q = 3 . 744, P = 2, P<0 . 05, block 2 versus block 4 ; q = 0 . 99, NS . Run down alley ; block 1 versus block 2 ; q = 3. 707, P = 2, ,,o _ ./. Mean percent Corrcct Choices
P<0.05, block 2 versus block 4 ; q = 1 .68, P = 3, NS .) Figure 2 shows the means across blocks of five trials of the per cent correct choices . Only correct initial choices were included, and the per cent correct in each block of trials was subjected to the arcsin transform before analysis . There is a clear main effect of positive colour (F = 58 .638, df = 1, 28 ; P<0.001) with chicks approaching red making more correct choices . There is also a significant effect of blocks of trials (F = 13 .401 ; df = 3, 84 ; P<0 . 001) with all groups making more correct choices as training progresses. As before, most of this difference is between the first and second blocks of trials (block 2 versus block 1 ; q = 3 . 043, P = 2, P<0.05, block 4 versus block 2; q = 3 . 06, P = 3, NS) . The interaction of exposure and trials is also significant (F = 4 . 535 ; df = 3, 84 ; P<0 . 02) with dark birds making fewer correct choices than light birds in the course of training ; dark birds learn more slowly than light birds . One other measure of learning (trials to criterion) was examined. The criterion was set at five correct choices in succession, and a bird that did not reach criterion was given a score of 20 . A two-factor analysis of variance revealed that both of the main effects, and the interaction between them, are significant . These results are most easily interpreted using Fig . 3, which shows 20 -trials Blue
15
10
5
0
0
I Light
Red
I Dark
Mean Trials to Criterion Fig. 2 . Mean correct choices as a percentage of trials in five-trial blocks . The effects of trials and positive colour are both significant, as is the interaction between preexposure and blocks of trials.
Fig . 3 . Mean trials to criterion (five successive correct choices) for each of the four groups . Both of the main effects and their interaction are significant .
4
ANIMAL BEHAVIOUR, 26,
the mean trials to criterion for the four groups . Birds being trained to approach red required fewer trials than birds approaching blue (F = 19 . 996 ; df = 1, 28 ; P<0.001). In addition birds kept in the dark also took more trials than those exposed to light (F = 7-198 ; df = 1, 28 ; P<0 . 05) but this is due largely to the strong interaction between exposure and positive colour. When the positive colour was blue, darkreared birds required considerably more trials to reach criterion than when the positive colour was red (F= 8 . 751 ; df = 1, 28 ; P<0 . 02) . In summary, there is no difference between groups in latency to leave the start box, nor in time to run down the alley . As training progresses these measures both show a decline for all groups, especially from the first five trials to the second five trials . Birds approaching red make fewer errors than birds approaching blue, and there is an interaction between prior exposure to light and correct choices in each block of trials ; dark-reared birds show a slower increase in correct choices than light birds . These findings are echoed by the measure of trials to criterion. Red-trained birds and light-exposed birds take fewer trials to reach criterion and there is an interaction between positive colour and exposure such that dark, blue-trained birds perform very badly. Discussion The results obtained fall into three categories ; the better performance of birds being trained to approach red, the lack of differences between groups in measures independent of discrimination, and the faster learning of the light-exposed birds . These will be discussed in turn . It is well known that, even in the absence of any contingent reward, chicks are differentially attracted to different colours, and most studies indicate a greater preference for red than for blue (Sluckin 1972) . The question of why birds approach the stimuli at all is discussed further below ; at the moment it is sufficient to note that one would expect naive birds to approach the red stimulus preferentially . In fact all birds do tend to choose red in the initial trials ; those that are rewarded for approaching red receive their reward soon after making their choice, so their tendency to approach red is immediately strengthened . Those birds that must approach blue have first to overcome their naive preference for the red stimulus, which takes more trials and results in slower learning . Regardless of pre-exposure and positive training colour, all chicks learn that they must
1
leave the start box and run down the alley . It is not possible to say whether they do so because the correction procedure employed tends to reward the making of any choice, or because the stimuli are attractive as imprinting objects . Probably both factors are involved . The age of the chicks (48 h) argues against an interpretation solely in terms of imprinting, because even though largely dark reared these birds are well beyond the most sensitive period for imprinting (Bateson 1966) . The pre-training procedure was designed to help the birds to associate the goal boxes with heat, so one should not be surprised that they run to the goal boxes . The important thing is that light exposure has no effect on the ability of a chick to learn to run down the alley to the goal box for heat . Dark chicks are not incapable of learning to run to a particular place for heat, but they take more trials to learn that the correct place is indicated by a particular colour . The poor performance of the birds running to blue also reveals that dark birds are able to discriminate as well as light birds ; if they were not discriminating they would run to red and blue about equally, whereas in fact they continue to run to red even though rewarded more quickly when they choose blue . It is as though they quickly forget, or never appreciate, that they are rewarded only in the box containing the blue circle . In a similar way dark chicks are not immune to the attractions of an imprinting object, but they take longer to approach and learn more slowly (Bateson & Wainwright 1972, Bateson & Seaburne-May 1973). It may be, as Dimond & Adam (1972) proposed, that the first few minutes of light exposure are necessary to get the visual system working ; once this has happened visual learning is good, but the time taken to achieve full operation makes it appear that learning is defective . Operationally, learning is indeed retarded in the dark birds, but they are fully capable of learning a task that does not involve visual discrimination as is shown by the decrease in latency to leave the start box and time to run down the alley . It should be noted that even the dark/blue chicks show an increase in correct choices in the last block of trials . Dark chicks learn to leave the start box and run down the alley just as effectively as light chicks . They fail to learn that they will get their reward quicker if they go to a particular colour . This is especially true when the positive stimulus is less attractive than the negative stimulus .
CHERFAS : LIGHT EXPOSURE AND DISCRIMINATION LEARNING
In an earlier paper (Cherfas 1977) the activating effects of brief exposure to light were extended from imprinting and visuomotor coordination to one-trial passive avoidance conditioning . The present study widens further the range of activating effects due to light exposure to include extended learning of a simultaneous colour discrimination based on positive reward . Furthermore it confirms the belief that light exposure is enhancing acquisition, not recall, because all birds run for heat but only the lightexposed chicks learn quickly to run to a particular colour. Acknowledgments This work was supported by a Science Research Council Studentship for post-graduate research . I thank Patrick Green for constructing the illuminated stimuli used in the maze . Michael Morgan assisted with the statistics . REFERENCES Bateson, P. P . G. 1966. The characteristics and context of imprinting . Biol. Rev., 41, 177-220 . Bateson, P. P. G . 1976 . Specificity and the origins of behaviour . In : Advances in the Study of Behaviour, 6, 1-20 . New York and London : Academic Press .
5
Bateson, P. P . G . & Seaburne-May, G. 1973 . Effects of prior exposure to light on chicks' behaviour in the imprinting situation. Anim. Behav., 21, 720-725 . Bateson, P . P . G . & Wainwright, A . A . P . 1972 . The effects of prior exposure to light on the imprinting process in chicks. Behaviour, 42, 279-290. Cherfas, J. J . 1977. Prior exposure to light improves avoidance learning in day-old chicks . Anim. Behav., 25,732-735 . Dimond, S. J . & Adam, J. H . 1972. Approach behaviour and embryonic visual experience in chicks : studies of the rate of visual flicker . Anim. Behav., 20,413-420 . Fellows, B . J. 1967 . Chance stimulus sequences for discrimination tasks . Psychol. Bull., 67, 68-92 . Sluckin, W . 1972 . Imprinting and early learning . 2nd edn . London : Methuen . Wekstein, D . R . & Zolman, J. F. 1967 . Homeothermic development of the young chick . Proc. Soc . Exptl. Biol . Med., 128, 294-297 . Winer, B . J . 1971 . Statistical principles in experimental design . 2nd edn. New York : McGraw-Hill . Zolman, J . F . 1968. Discrimination learning in young chicks with heat reinforcement . Psychological Record, 18, 303-309 . Zolman, J. F . & Martin, R . C . 1967. Instrumental aversive conditioning in newly hatched domestic chicks . Psychon . Sci ., 8, 183-184. Zolman, J. F . & Becker, D . 1968. Spatial discrimination learning in young chicks . Psychon. Sci., 10, 361-362 . (Received 20 December 1976 ; revised 23 February 1977 ; MS. number : 1594)
SIMULTANEOUS COLOUR DISCRIMINATION IN CHICKS IS IMPROVED BY BRIEF EXPOSURE TO LIGHT By JEREMY JON CHERFAS Sub-Department of Animal Behaviour, Madingley, Cambridge CB3 8AA
Abstract . Dark-reared chicks were exposed to white light for 1 h at age 24 h. At 48 h these chicks, and others kept in the dark throughout, were given the opportunity to obtain a heat reward by choosing the correct arm in a Y maze . The discriminanda were red and blue illuminated circles, each being the positive stimulus for half the chicks . Light-exposure resulted in faster learning, especially when the positive stimulus was blue . This is taken as further support for the view that brief exposure to light activates the visual system non-specifically, affecting many kinds of visually guided behaviour . Brief exposure to light is known to have an effect on visually mediated behaviour in the chick . For example, imprinting takes place more readily (Bateson & Seaburne-May 1973) and results in a stronger preference (Bateson & Wainwright 1972) if chicks are first exposed to light for 30 min . Avoidance learning, too, is enhanced by prior exposure to light (Cherfas 1977) . These, and other, results have been interpreted as evidence for a non-specific activation of the visual system by prior exposure (Bateson 1976) . In particular, Bateson & Wainwright (1972) suggested that the light-exposed chick is able to learn faster because its visual system is already, as it were, primed to process incoming information. Dimond & Adam (1972) said much the same thing, that light `acts on the learning mechanism itself and brings it into a state of readiness' (p . 419) . Further evidence of the non-specific activating effects of light exposure was provided in an earlier experiment (Cherfas 1977) . Light-exposed and dark-reared chicks were allowed to peck at a coloured target coated with methyl anthranilate . Subsequent retention tests, consisting of successive presentations of the training target colour and a neutral target, revealed that lightexposed chicks increase their pecks to the neutral colour and decrease their pecks to the training colour, but dark-reared chicks do not alter their behaviour as a result of their experience with methyl anthranilate . This last study, and those concerned with aspects of imprinting, suggest that light-exposed birds pay more attention to visual stimuli and thereby learn more about those stimuli . If this is the case one would expect enhanced performance on any task where the chick is required to learn about visual stimuli . The present study had two aims ; to extend the `enhanced learning'
observations to a simultaneous discrimination based on conventional positive reward, and to look at performance in a task involving learning over a number of trials . A positive reward to motivate the behaviour of young chicks is not easy to find . They need neither food nor water, so the performance of a group of chicks working for these rewards is very variable . Working for an imprinting object involves complex perceptual rewards, and it would be difficult to establish a discriminative stimulus independent of the imprinting object . Young chicks, however, are not very good at regulating their body temperature (Wekstein & Zolman 1967) and heat has proved to be a potent reward for young chicks (Zolman & Martin 1967 ; Zolman and Becker 1968 ; Zolman 1968) . Dark-reared chicks were exposed to light, or held in darkness, for 1, h . Twenty-four hours later, when the chicks were 2 days old, they were given 20 trials in which to learn to approach a particular coloured stimulus in a Y maze for a heat reward . Methods Rearing and exposure techniques have been described elsewhere (Cherfas 1977). Briefly, chicks were removed from a dark hatching incubator about 6 h after hatching and marked to enable subsequent recognition . They were transferred to individual compartments in a second light-tight incubator until they were 24 h old . Half the birds were then moved to similar compartments in an incubator with a transparent lid. A 100-W tungsten bulb behind an opal screen 50 cm above the birds was switched on for 1 h, after which the chicks were returned to their compartments in the light-tight incubator . The other half of each batch was moved between compartments in light-tight incubators . There 1
2
ANIMAL BEHAVIOUR, 26,
are thus two groups of birds, both of which received the same non-visual stimulation, but only one of which received the hour of exposure to light ; the two groups will be called the dark and light groups . Chicks were assigned to the groups at random . Training began about 24 h after the end of priming. The apparatus was a discrimination maze divided into a start box, a choice area, and two goal boxes . The maze was constructed of hardboard with a wire-mesh floor . Except in the goal boxes there was a false floor of hardboard beneath the wire mesh . The start box was 15 . 2 x 15 . 2 cm and opened onto a choice area 30 . 5 x 58 . 4 cm . The goal boxes were each 15 . 2 x 17 . 8 cm, and the whole apparatus was 38 cm deep . The internal walls and floors were painted matt black . On the rear wall of each goal box was a circular hole, 10 . 2 cm in diameter, with its bottom edge 3 . 8 cm above the wire mesh floor . Behind the hole was an opal diffusing screen which could be illuminated from behind by a 21-W tungsten bulb . Coloured acetate filters could be inserted between the diffusing screen and the hole, so that the rear wall of each box contained an illuminated red or blue circle . The colour of the stimulus in the goal box was easily changed by changing the filters . The room in which the maze was situated was in complete darkness, and the only light in the maze was provided by the stimuli. Ambient temperature in the room was about 15 C . Heat was provided by a fan heater mounted 30 cm below the goal box floor . For each correct choice the heater was switched on for 15 s . This raised the ambient temperature in the goal box from 15 to 25 C in about 2 s . Temperature returned to baseline within 15 s of the end of heat presentation. (All measurements were made with a Grant continuous recording thermometer connected to a probe 2 cm above the goal box floor.) Training involved pre-training and acquisition of a simultaneous colour discrimination . Pretraining consisted of four trials in which the chick was shaped to approach the goal box to receive heat . The subject was removed from the dark incubator and placed in a box cooled to 10 C . After 1 min it was removed and placed in one of the goal boxes . The light, without filter, was switched on and the chick was given 15 s of heat. After an inter-trial interval of 30 s, spent in the dark in the cooled box, the subject was placed in the other goal box and given another heat reward. On the remaining two trials the
1
chick was placed directly in front of the goal box and rewarded when it entered the box . Acquisition trials began 30 s after the end of the last pre-training trial . Subjects were required to approach and enter the goal box containing a particular coloured stimulus. For half the subjects the positive stimulus was the red circle and the negative the blue (referred to as red chicks) . For the other half (referred to as blue chicks) the significance of the colours was reversed . The position of the stimuli was varied for each acquisition trial according to a predetermined schedule (Fellows 1967) . On each training trial the chick was placed in the start box facing the choice area . The stimuli were illuminated and the latency before the chick left the start box was recorded. Time to run down the alley and initial choice were also noted . If the initial choice was correct the chick immediately received 15 s of heat reward. A correction method was used . If the first choice was incorrect the chick was allowed to retrace or, after 60 s, was guided into the other goal box where it received a reward . After each reward had been delivered the stimuli were switched off and the chick was moved into the cooled box where it remained in the dark for the intertrial interval of 30 s . All chicks were given 20 acquisition trials with a 15-s heat reward on every trial. Light and dark chicks were trained alternately with red or blue stimulus positive, according to a predetermined schedule . A total of 32 birds from four batches were trained, eight in each group (light/red, light/blue, dark/red, dark/blue) . Results Data were analysed using an analysis of variance, with repeated measures where appropriate . The r Llght
Dark Bed BIue
1-5
6-10
11-15
16201rials
Mean time to leave Start-box
1-5
6-10
11-15
16dOtrials
Mean time to run down Alley
Fig . 1 . Mean values for (left) time to leave start box and (right) time to run down alley . Scores on blocks of five consecutive trials have been averaged to give four values for each subject . For both measures only the main effect of blocks of trials is significant .
CHERFAS : LIGHT EXPOSURE AND DISCRIMINATION LEARNING
factors included were exposure (light or dark), positive colour (red or blue) and training trials (scores were averaged over four blocks of five successive trials) . Transformations (square root or arcsin) were performed prior to the analyses but the figures all show untransformed data points . Figure 1 shows the means, across blocks of five trials, of time to leave the start box and time to run down the alley for the four groups . Scores were subjected to a square root transform and analysed with a 2 x 2 x 4 repeated measures analysis of variance (Winer 1971, p . 559) . All chicks left the start box more quickly as training progressed (F = 9 .908 ; df = 3, 84 ; P<0.001) and also ran from the start box to the goal box faster (F = 11 .574; df = 3, 84 ; P<0 . 001) . There were no other significant effects . StudentNewman-Keuls tests were used to locate the effect due to trials . For both measures the score for the first block of trials was different from the other three blocks, which did not differ among themselves (Leave start box ; block 1 versus block 2 ; q = 3 . 744, P = 2, P<0 . 05, block 2 versus block 4 ; q = 0 . 99, NS . Run down alley ; block 1 versus block 2 ; q = 3. 707, P = 2, ,,o _ ./. Mean percent Corrcct Choices
P<0.05, block 2 versus block 4 ; q = 1 .68, P = 3, NS .) Figure 2 shows the means across blocks of five trials of the per cent correct choices . Only correct initial choices were included, and the per cent correct in each block of trials was subjected to the arcsin transform before analysis . There is a clear main effect of positive colour (F = 58 .638, df = 1, 28 ; P<0.001) with chicks approaching red making more correct choices . There is also a significant effect of blocks of trials (F = 13 .401 ; df = 3, 84 ; P<0 . 001) with all groups making more correct choices as training progresses. As before, most of this difference is between the first and second blocks of trials (block 2 versus block 1 ; q = 3 . 043, P = 2, P<0.05, block 4 versus block 2; q = 3 . 06, P = 3, NS) . The interaction of exposure and trials is also significant (F = 4 . 535 ; df = 3, 84 ; P<0 . 02) with dark birds making fewer correct choices than light birds in the course of training ; dark birds learn more slowly than light birds . One other measure of learning (trials to criterion) was examined. The criterion was set at five correct choices in succession, and a bird that did not reach criterion was given a score of 20 . A two-factor analysis of variance revealed that both of the main effects, and the interaction between them, are significant . These results are most easily interpreted using Fig . 3, which shows 20 -trials Blue
15
10
5
0
0
I Light
Red
I Dark
Mean Trials to Criterion Fig. 2 . Mean correct choices as a percentage of trials in five-trial blocks . The effects of trials and positive colour are both significant, as is the interaction between preexposure and blocks of trials.
Fig . 3 . Mean trials to criterion (five successive correct choices) for each of the four groups . Both of the main effects and their interaction are significant .
4
ANIMAL BEHAVIOUR, 26,
the mean trials to criterion for the four groups . Birds being trained to approach red required fewer trials than birds approaching blue (F = 19 . 996 ; df = 1, 28 ; P<0.001). In addition birds kept in the dark also took more trials than those exposed to light (F = 7-198 ; df = 1, 28 ; P<0 . 05) but this is due largely to the strong interaction between exposure and positive colour. When the positive colour was blue, darkreared birds required considerably more trials to reach criterion than when the positive colour was red (F= 8 . 751 ; df = 1, 28 ; P<0 . 02) . In summary, there is no difference between groups in latency to leave the start box, nor in time to run down the alley . As training progresses these measures both show a decline for all groups, especially from the first five trials to the second five trials . Birds approaching red make fewer errors than birds approaching blue, and there is an interaction between prior exposure to light and correct choices in each block of trials ; dark-reared birds show a slower increase in correct choices than light birds . These findings are echoed by the measure of trials to criterion. Red-trained birds and light-exposed birds take fewer trials to reach criterion and there is an interaction between positive colour and exposure such that dark, blue-trained birds perform very badly. Discussion The results obtained fall into three categories ; the better performance of birds being trained to approach red, the lack of differences between groups in measures independent of discrimination, and the faster learning of the light-exposed birds . These will be discussed in turn . It is well known that, even in the absence of any contingent reward, chicks are differentially attracted to different colours, and most studies indicate a greater preference for red than for blue (Sluckin 1972) . The question of why birds approach the stimuli at all is discussed further below ; at the moment it is sufficient to note that one would expect naive birds to approach the red stimulus preferentially . In fact all birds do tend to choose red in the initial trials ; those that are rewarded for approaching red receive their reward soon after making their choice, so their tendency to approach red is immediately strengthened . Those birds that must approach blue have first to overcome their naive preference for the red stimulus, which takes more trials and results in slower learning . Regardless of pre-exposure and positive training colour, all chicks learn that they must
1
leave the start box and run down the alley . It is not possible to say whether they do so because the correction procedure employed tends to reward the making of any choice, or because the stimuli are attractive as imprinting objects . Probably both factors are involved . The age of the chicks (48 h) argues against an interpretation solely in terms of imprinting, because even though largely dark reared these birds are well beyond the most sensitive period for imprinting (Bateson 1966) . The pre-training procedure was designed to help the birds to associate the goal boxes with heat, so one should not be surprised that they run to the goal boxes . The important thing is that light exposure has no effect on the ability of a chick to learn to run down the alley to the goal box for heat . Dark chicks are not incapable of learning to run to a particular place for heat, but they take more trials to learn that the correct place is indicated by a particular colour . The poor performance of the birds running to blue also reveals that dark birds are able to discriminate as well as light birds ; if they were not discriminating they would run to red and blue about equally, whereas in fact they continue to run to red even though rewarded more quickly when they choose blue . It is as though they quickly forget, or never appreciate, that they are rewarded only in the box containing the blue circle . In a similar way dark chicks are not immune to the attractions of an imprinting object, but they take longer to approach and learn more slowly (Bateson & Wainwright 1972, Bateson & Seaburne-May 1973). It may be, as Dimond & Adam (1972) proposed, that the first few minutes of light exposure are necessary to get the visual system working ; once this has happened visual learning is good, but the time taken to achieve full operation makes it appear that learning is defective . Operationally, learning is indeed retarded in the dark birds, but they are fully capable of learning a task that does not involve visual discrimination as is shown by the decrease in latency to leave the start box and time to run down the alley . It should be noted that even the dark/blue chicks show an increase in correct choices in the last block of trials . Dark chicks learn to leave the start box and run down the alley just as effectively as light chicks . They fail to learn that they will get their reward quicker if they go to a particular colour . This is especially true when the positive stimulus is less attractive than the negative stimulus .
CHERFAS : LIGHT EXPOSURE AND DISCRIMINATION LEARNING
In an earlier paper (Cherfas 1977) the activating effects of brief exposure to light were extended from imprinting and visuomotor coordination to one-trial passive avoidance conditioning . The present study widens further the range of activating effects due to light exposure to include extended learning of a simultaneous colour discrimination based on positive reward . Furthermore it confirms the belief that light exposure is enhancing acquisition, not recall, because all birds run for heat but only the lightexposed chicks learn quickly to run to a particular colour. Acknowledgments This work was supported by a Science Research Council Studentship for post-graduate research . I thank Patrick Green for constructing the illuminated stimuli used in the maze . Michael Morgan assisted with the statistics . REFERENCES Bateson, P. P . G. 1966. The characteristics and context of imprinting . Biol. Rev., 41, 177-220 . Bateson, P. P. G . 1976 . Specificity and the origins of behaviour . In : Advances in the Study of Behaviour, 6, 1-20 . New York and London : Academic Press .
5
Bateson, P. P . G . & Seaburne-May, G. 1973 . Effects of prior exposure to light on chicks' behaviour in the imprinting situation. Anim. Behav., 21, 720-725 . Bateson, P . P . G . & Wainwright, A . A . P . 1972 . The effects of prior exposure to light on the imprinting process in chicks. Behaviour, 42, 279-290. Cherfas, J. J . 1977. Prior exposure to light improves avoidance learning in day-old chicks . Anim. Behav., 25,732-735 . Dimond, S. J . & Adam, J. H . 1972. Approach behaviour and embryonic visual experience in chicks : studies of the rate of visual flicker . Anim. Behav., 20,413-420 . Fellows, B . J. 1967 . Chance stimulus sequences for discrimination tasks . Psychol. Bull., 67, 68-92 . Sluckin, W . 1972 . Imprinting and early learning . 2nd edn . London : Methuen . Wekstein, D . R . & Zolman, J. F. 1967 . Homeothermic development of the young chick . Proc. Soc . Exptl. Biol . Med., 128, 294-297 . Winer, B . J . 1971 . Statistical principles in experimental design . 2nd edn. New York : McGraw-Hill . Zolman, J . F . 1968. Discrimination learning in young chicks with heat reinforcement . Psychological Record, 18, 303-309 . Zolman, J. F . & Martin, R . C . 1967. Instrumental aversive conditioning in newly hatched domestic chicks . Psychon . Sci ., 8, 183-184. Zolman, J. F . & Becker, D . 1968. Spatial discrimination learning in young chicks . Psychon. Sci., 10, 361-362 . (Received 20 December 1976 ; revised 23 February 1977 ; MS. number : 1594)