Journal of Comparative and Physiological Psychology 1975, Vol. 88, No. 1, 402-408
COLOR PECKING PREFERENCES IN WHITE LEGHORN CHICKS1 GLORIA J. FISCHER,2 GRANT L. MORRIS,3 AND JOHN P. RUHSAM Washington State University Studies of color pecking preferences in newly hatched chicks (Gallus dom.esticus) have shown unimodal preference in the orange region of the spectrum or bimodal preferences at blue and orange. In the present study, darkhatched White Leghorn chicks were tested in darkness with targets illuminated at 1 of 3 radiant intensity levels. Results showed the least amount of pecking at green (541 nm.) and peak preferences in blue-violet and orangered regions. Findings were similar when other dark-hatched chicks were tested in the light (Experiment 2). Overall, findings suggest unlearned pecking preferences for short and long wavelengths, with minimums at green. Possible evolutionary and photochemical bases for such a bimodal wavelength preference function were discussed. Since bimodality was unaffected by target intensity and background, these variables probably do not account for the unimodal function reported by others.
Findings from color pecking preference (e.g., White Rock and White Leghorn) and studies in chicks (Gallus domesticus) are similar methods of testing pecking preferinconsistent. For example, Hess (1956) re- ences. However, stimulus brightness and ported preferences in the blue and orange prior light experience of the subjects varied. regions of the spectrum, whereas other in- One or both of these factors might influence vestigators found a unimodal preference in color preferences and, thus, may have conthe orange region (Gunther & Jones, 1962; tributed to discrepant findings. The present Oppenheim, Jones, & Gottlieb, 1970). Cur- study attempted to resolve that discrepancy tius (1954) found bimodal preferences when by assessing these possibilities. This was illuminated targets were presented against done in Experiment 1 by (a) eliminating a black background and unimodality when prior visual experience by incubating, hatchthey were presented against a white back- ing, and even testing chicks in darkness ground. However, her results do not account (except for illuminated targets visible durfor the discrepancy in findings (i.e., bimodal ing testing) and by (b) controlling brightness vs. unimodal preferences), since Hess by randomizing stimulus (target) radiance (1956) found bimodality with a gray rather over 3 intensity levels. than black background and the investigators EXPERIMENT 1: COLOR PECKING who found unimodality (Gunther & Jones, PREFERENCES IN DARK-HATCHED 1962; Oppenheim et al., 1970) also used a CHICKS TESTED IN DARKNESS gray background. The chick studies cited used similar breeds 1
Experiment 1 in the present report was a Master's thesis submitted by the second author to Washington State University. The authors wish to thank a number of persons. Linda Ireland and John Nord helped with the pilot work and with data collection, respectively, in Experiment 2. George Leary was very helpful with technical aspects of the research. Ronald Oppenheim provided information and comments that were very helpful in the discussion section of the manuscript. 2 Request for reprints should be sent to Gloria J. Fischer, Department of Psychology, Washington3 State University, Pullman, Washington 99163. Now at the University of Northern Colorado.
Method Subjects. Two or 3 sets of 6 1-day-old White Leghorn chicks (Gallus domesticus) were tested weekly, reaching a total of 138 chicks. The chicks were from weekly hatches of fertile eggs obtained from a pedigreed flock maintained by the Washington State University Department of Animal Sciences. Apparatus. The pecking apparatus (Morris, 1970) consisted of 6 11 X 6 X 12Ji in. sheet-metal chambers, painted a flat gray. A 1-in. circle was cut from the outside wall of each chamber, 5 in. above a hardware cloth grid floor. Behind these 1-in. cutouts, clear, rectangular, Plexiglas keys were suspended pendantly. Masks constructed of
402
403
COLOR PREFERENCES IN CHICKS .997-in. brass shim-stock with a Jff-in. circle cut out of the center were attached at the front of the keys by clips. The front of each mask was sprayed with flat white paint; the back was covered by frosted plastic film taped over the projection spot. Light sources for the 6 test chambers were Bausch and Lomb 100-w. microscope illuminators that contained a single-element lens for light collimation. A ground glass was mounted in the projector between the lamp and the collimating lens. Intensity could be controlled by a calibrated iris. The filters were Photovolt monochromatic interference filters that passed no side bands within the visible spectrum. These filters were fitted in holders and then positioned 1J^ in. behind the target spot and in the projector light beam. Six filters were used, and their transmittance characteristics, measured with a Beckman Model DU spectrophotometer, are shown in Table 1. Since the orange (598-nm.) filter transmitted the least amount of light (see Table 1), it and an arbitrarily selected chamber were designated the standard for setting relative intensity levels. This was done by using a sensor probe with 22-F characteristics (responsive to 300-800 nm., with an essentially flat response to the radiant energy of wavelengths used in the present study). The probe was placed in front of the standard key, and this standard brightness was used to calibrate a Photovolt photometer (Model 520 M). The photometer was used, in turn, to equate radiance of the targets (i.e., the amount of colored light passed by each filter was equated). Equating individual targets to the photometric standard was accomplished by adjusting the iris on the projector lamp. This was done for all filter-key-light source combinations at the maximum intensity level (bright, B) for orange and for 2 lower levels designated medium (M) and dim (D). The B, measured 9 cm. from the standard key's surface using a Macbeth illuminometer, was 2.58 mL. The M was .275 density units less than B, and D was .75 density units less than B. The M and D for the standard were measured at 1.76 mL. and .68 mL., respectively, using the Macbeth illuminometer. A key peck of 5 gm. operated a microswitch and attached pulse former and counter. The microswitch was mounted behind the lower edge of each key. The first response (peck) on any 1 of the 6 keys activated a timer common to all pulse formers. The timer was set for 15 min. Counters recorded only during this 15-min. period and then automatically locked out of the circuit. Procedure. Incubation and hatching. A flat of 30 fertile eggs, collected on Monday of each week and placed in cold storage, was set in a commercial incubator the next day and/or on Friday p.m. for a total of 10 hatches. When one hatch was being set, eggs from a 2-wk.-old hatch were candled and then transferred in a light-proof container to a dark, forced-air incubator for the last week of incubation and hatching. The hatching
TABLE 1 SPECIFICATIONS OF INTBBFEEBNCB FILTERS Hue
Violet Blue Green Yellow Orange Red
Wavelength Percentage of Half bandwith (in nm.) at peak transmittance (in nm.) transnrittance
423 468 516 582 598 680
25
36 39 12 8 43
10.0 10.5 8.5 7.5
10.5 10.5
incubator was located in a dark room, and an infrared vision device was available for observation of hatching, though most checks were made in the dark, using tactile cues. Observations were begun at midnight of the twentieth day and at 8. a.m., noon and 4 p.m. on the twenty-first day. At these times, chicks that had hatched were isolated individually or in pairs in separate tray compartments for age and identification purposes. Pecking preference tests. Chicks that had, hatched within the same 4-hr, period were grouped into sets of 6 chicks. When the average age of the set was 23J^ hr., the chicks were isolated for J£ hr. in the darkened incubator room, in individual cubicles of a transport carton. At 24 hr., they were transferred in the carton, within a larger lightproof container, to the darkened test room. There the chicks were placed randomly in the 6 test chambers, facing the key, and left for 5 inin. of adaptation. During this time, the 6 interference filters and their respective intensity level (B, M, or D) were assigned to test chambers, according to predetermined random sequences. These sequences included the assignment of intensity level and the order of filter presentation over 6 successive test sessions. Filter order was restricted only by the requirement that each chick receive all 6 filters. Intensity levels were randomized without restriction. After the 5 min. of adaptation to the test chamber in the dark, the targets for all 6 chambers, were illuminated simultaneously. The first peck by any chick activated all counters and a timer. Fifteen minutes later the timer disconnected lights and counters, marking the end of the first test session. The chambers were then covered by a board so that the chicks remained in darkness during a 5-min. intertrial interval (ITI). During the ITI, the number of pecks per chick were recorded from counters, filters were shifted, and intensity levels were changed. Fifteen-minute test sessions and ITIs continued in this manner until the chicks had been exposed to all 6 interference filters.
Results and Discussion Data on one chick were eliminated because the experimenter misaligned the light.
404
G. J. FISCHER, G. L. MORRIS, AND J. P. RUHSAM
35
• 25 in £20
'400423 (violet)
468 (blue)
516 (green)
582598 (yellowXorange)
680 (red)
WAVELENGTH, NANOMETERS
FIGUBE 1. Mean number of pecks by darkhatched chicks, tested in the dark, at each of 6 target colors presented with radiant intensity randomized.
The mean number of pecks per filter color for the remaining 137 subjects and the standard error of the means are shown in Figure 1. A filter color by subjects repeated-measures design revealed that the differences among the means in Figure 1 were reliable (F = 2.60, df = 5/680, p = .05), These findings affirm unlearned, bimodal preferences for red and blue-violet regions of the spectrum in dark-hatched chicks tested in darkness. The finding of bimodal color pecking preferences agrees with Hess (1956). In fact, the overall shape of the relation between pecking and wavelength resembles closely the func-
tion reported by Hailman (1964) for gull chicks. All prior studies of chick color pecking preference, including, of course, those reporting unimodal preference in the orange region, tested chicks in the light. As indicated previously, the level of illumination in these studies varied. In an effort, then, to reproduce the unimodal color preference found in some of these studies (e.g., Gunther & Jones, 1962; Oppenheim et al., 1970), Experiment 2 tested dark-hatched chicks in the light at 3 levels of target radiance (B, M, and D), varied systematically. Since Curtius (1954) reported unimodality with a white but not with a black background, the targets in Experiment 2 were presented against white keys. EXPERIMENT 2: COLOR PREFERENCES IN DARK-HATCHED CHICKS TESTED IN THE LIGHT WITH DIFFERING TARGET INTENSITY Method Subjects. A total of 150 1-day-old chicks (2-3 sets of 6 chicks per hatch) were run from 9 successive hatches. A random one third (50) were assigned to each of the B, M, and D target intensities. Apparatus. The apparatus was the same as in Experiment 1 except for changes relating to testing in the light. For example, a 15-min. period of light adaptation was given immediately prior to the first test session. To prevent the subjects from seeing or pecking at the key during this time, metal doors, painted a flat gray, were suspended over the key wall of each cubicle. The B, M, and D levels were recalibrated as in Experiment 1, with new bulbs in all projectors. Dim, central fluorescent lighting 1.38 m. above the wire floor of the pecking apparatus provided 3.23 mL. at the pecking surface of the key in the standard chamber. When the standard filter, 598 nm., was projected at a bright intensity level onto this key, illumination at the key's surface measured 4.21 mL. Procedure. Within each hatch an equal number of cubicles were assigned randomly to 1 of the 3 radiance levels designated B, M, or D. Chicks were placed randomly into the cubicles (with metal doors covering the keys) for 15 min. of adaptation to the cubicles and to light. The metal doors were then removed, and the projector lights and automatic counters turned on. As in Experiment 1, these counters activated only after the first peck by any 1 of the 6 chicks. Six successive 15-min. test sessions over the 6 filters proceeded
COLOR PREFERENCES IN CHICKS as in Experiment 1 except that each chick pecked at all 6 colors at the same (B, M, or D) intensity.
Results On one test session the counter recording from the violet filter was cleared accidentally, and the number of pecks by that chick on that filter was treated as a missing observation. Data for 20 of the 150 chicks (12.6%) were excluded because they failed to peck at any 1 of the 6 filters. Of these 20 nonresponders, half (10) had been exposed to dim target intensity, 6 had medium, and 4 had bright. For the remaining 130 chicks that pecked at least once, the number of pecks per filter, per intensity level were subjected to a 3 X 6 (Intensity X Filter Color) repeated-measures design analysis of variance. The magnitude of pecking declined absolutely with decreasing target intensity, but the effects of intensity and its interaction with filter color were statistically insignificant (Fa < 1.00). The effects of filter color approached significance (F = 1.96, df = 5/634, p < .08), and the mean rates of pecking per filter wavelength and their standard error are shown in Figure 2. Individual comparisons among the means in Figure 2 showed that the only single color preferred over green was orange (F — 4.97, df = 1/634, p < .05), but violet and yellow, taken together, also were significantly preferred to green, and thus, to red (F = 4.10, df = 1/634, p < .05). These results affirm unlearned, bimodal pecking preferences for violet and yellow-orange in dark-hatched chicks tested in the light. DISCUSSION Present findings show bimodal color pecking preferences in dark-hatched domestic chicks, both when they are tested in darkness (with illuminated targets) and when they are tested in the light. Why the discrepancy between these and Hess's (1956) findings and those showing only a unimodal pecking preference at orange (e.g., Gunther & Jones, 1962; Oppenheim et al., 1970)? Curtius (1954) found unimodality when chicks were tested in the light with targets presented against a white background. In Experiment 2 of the present
405
35-
400423
468
516
582598
680
WAVELENGTH, NANOMETERS
FIGURE. 2. Mean number of pecks by darkhatched chicks, tested light adapted and in the light, at 6 target colors, all presented against white and at a random 1 of 3 intensities.
study, however, targets were presented against a white background at different levels of radiant energy, and bimodal preferences were found. Thus, the unimodal preference found by others cannot be attributed to any simple function of background and/or target intensity. Perhaps, investigators who did not find bimodality (e.g., Gunther & Jones, 1962; Oppenheim et al., 1970) might have, had they included violet as a test filter. Also, if target intensity is not important, then filter bandwidth is a possibility, as suggested by Hailman (1969). Specifically, narrow band-pass Wrattan filters presented at equal quantum intensity produced bimodal pecking preferences in gull chicks, whereas wide band-pass Wrattan filters and uncontrolled intensity produced unimodality in the red region. The interference filters used in the present study also passed narrow bands (see Table 1), but so did those used by Oppenheim et al. (1970; Oppenheim, personal communication, July 1973). This would seem to rule out the salience of filter bandwidth. Still another possibility is ambient light intensity, which varied in the studies cited and is known to affect the rate at which chickens peck for food. Why might domestic chicks show unlearned color pecking preferences for short.
406
G. J. FISCHER, G. L. MORRIS, AND J. P. RUHSAM
and long wavelengths with minimum at green? A negative pecking preference for green might have adaptative significance since green is a salient habitat color but is the background color for feeding. Thus, so-called preferences at violet and yelloworange (e.g., when tested in the light) might be preferences instead for objects that contrast maximally with green. Hailman (1968) suggested a similar explanation of bimodal color pecking preferences in gull chicks. Using a photochemical approach to account for gull chick color preferences Hailman (1964) predicted the V-like function relating pecking amplitude and wavelength from a ratio of transmission spectra for red and orange oil droplets. The form of Hailman's (1964), and of the present, function agrees reasonably well, too, with what might be expected from the reverse of the sensitivity curve for the photopigment, iodopsin. lodopsin from the chicken eye has maximum sensitivity, X max , at 562 nm. and closely resembles the photopic sensitivity curve for the chicken (Wald, Brown, & Smith, 1955). In the present study, pecking tended to be lowest for niters nearest to X max for iodopsin. This was especially so when testing was in darkness (see Figure 1), where iodopsin should be more sensitive. The Xmax at 562 nm. implies maximal sensitivity at green. As indicated, this would not be unreasonable, since green is a salient habitat color. As a background color for feeding, however, green might be approached readily but pecked at least. A problem for the behavioral and photochemical approaches discussed is their in-
ability to account for an unexpected and perplexing finding of the present study. This is an apparent shift in pecking preference away from red when colors are presented in the light (cf. Figures 1 and 2). Though minimum pecking always is at green, preferences peak at blue-violet and orange-red when testing is in darkness. When testing is in the light, preferences peak at violet and yellow-orange. REFERENCES Curtius, A. Uber angeborene Verhaltensweissen bei Vogeln inbesondere bei Huhnerkucken. Zeitschrift fur Tierpsychologie, 1954, 11, 94-109. Gunther, W. C., & Jones, R. K. Effect of nonoptimally high incubation temperature on frequency of pecking and on color preferences in the chick. Proceedings of the Indiana Academy of Sciences, 1962, 72, 290-299. Hailman, J. P. Coding of the colour preference of the gull chick. Nature, 1964, 204, 710. Hailman, J. P. Spectral reflectance of gull's bill: Physiological and evolutionary implications for animal communication. Science, 1968, 162, 139140. Hailman, J. P. Spectral pecking preference in gull chicks: Possible resolution of a species difference. Journal of Comparative and Physiological Psychology, 1969, 67, 465-467. Hess, E. H. Natural preferences of chicks and ducklings for objects of different .colors. Psychological Reports, 1956, 2, 477-483. Morris, G. L. Color preference in dark-hatched White Leghorn chicks. Unpublished Master's thesis, Washington State University, 1970. Oppenheim, R. W., Jones, J. R., & Gottlieb, G. Embryonic motility and posthatching perception in birds after prenatal gamma irradiation. Journal of Comparative and Physiological Psychology, 1970, 71, 6-21. Wald, G., Brown, P., & Smith, P. lodospin. Journal of General Physiology, 1955, 38, 623. (Received November 12, 1973)
COLOR PECKING PREFERENCES IN WHITE LEGHORN CHICKS1 GLORIA J. FISCHER,2 GRANT L. MORRIS,3 AND JOHN P. RUHSAM Washington State University Studies of color pecking preferences in newly hatched chicks (Gallus dom.esticus) have shown unimodal preference in the orange region of the spectrum or bimodal preferences at blue and orange. In the present study, darkhatched White Leghorn chicks were tested in darkness with targets illuminated at 1 of 3 radiant intensity levels. Results showed the least amount of pecking at green (541 nm.) and peak preferences in blue-violet and orangered regions. Findings were similar when other dark-hatched chicks were tested in the light (Experiment 2). Overall, findings suggest unlearned pecking preferences for short and long wavelengths, with minimums at green. Possible evolutionary and photochemical bases for such a bimodal wavelength preference function were discussed. Since bimodality was unaffected by target intensity and background, these variables probably do not account for the unimodal function reported by others.
Findings from color pecking preference (e.g., White Rock and White Leghorn) and studies in chicks (Gallus domesticus) are similar methods of testing pecking preferinconsistent. For example, Hess (1956) re- ences. However, stimulus brightness and ported preferences in the blue and orange prior light experience of the subjects varied. regions of the spectrum, whereas other in- One or both of these factors might influence vestigators found a unimodal preference in color preferences and, thus, may have conthe orange region (Gunther & Jones, 1962; tributed to discrepant findings. The present Oppenheim, Jones, & Gottlieb, 1970). Cur- study attempted to resolve that discrepancy tius (1954) found bimodal preferences when by assessing these possibilities. This was illuminated targets were presented against done in Experiment 1 by (a) eliminating a black background and unimodality when prior visual experience by incubating, hatchthey were presented against a white back- ing, and even testing chicks in darkness ground. However, her results do not account (except for illuminated targets visible durfor the discrepancy in findings (i.e., bimodal ing testing) and by (b) controlling brightness vs. unimodal preferences), since Hess by randomizing stimulus (target) radiance (1956) found bimodality with a gray rather over 3 intensity levels. than black background and the investigators EXPERIMENT 1: COLOR PECKING who found unimodality (Gunther & Jones, PREFERENCES IN DARK-HATCHED 1962; Oppenheim et al., 1970) also used a CHICKS TESTED IN DARKNESS gray background. The chick studies cited used similar breeds 1
Experiment 1 in the present report was a Master's thesis submitted by the second author to Washington State University. The authors wish to thank a number of persons. Linda Ireland and John Nord helped with the pilot work and with data collection, respectively, in Experiment 2. George Leary was very helpful with technical aspects of the research. Ronald Oppenheim provided information and comments that were very helpful in the discussion section of the manuscript. 2 Request for reprints should be sent to Gloria J. Fischer, Department of Psychology, Washington3 State University, Pullman, Washington 99163. Now at the University of Northern Colorado.
Method Subjects. Two or 3 sets of 6 1-day-old White Leghorn chicks (Gallus domesticus) were tested weekly, reaching a total of 138 chicks. The chicks were from weekly hatches of fertile eggs obtained from a pedigreed flock maintained by the Washington State University Department of Animal Sciences. Apparatus. The pecking apparatus (Morris, 1970) consisted of 6 11 X 6 X 12Ji in. sheet-metal chambers, painted a flat gray. A 1-in. circle was cut from the outside wall of each chamber, 5 in. above a hardware cloth grid floor. Behind these 1-in. cutouts, clear, rectangular, Plexiglas keys were suspended pendantly. Masks constructed of
402
403
COLOR PREFERENCES IN CHICKS .997-in. brass shim-stock with a Jff-in. circle cut out of the center were attached at the front of the keys by clips. The front of each mask was sprayed with flat white paint; the back was covered by frosted plastic film taped over the projection spot. Light sources for the 6 test chambers were Bausch and Lomb 100-w. microscope illuminators that contained a single-element lens for light collimation. A ground glass was mounted in the projector between the lamp and the collimating lens. Intensity could be controlled by a calibrated iris. The filters were Photovolt monochromatic interference filters that passed no side bands within the visible spectrum. These filters were fitted in holders and then positioned 1J^ in. behind the target spot and in the projector light beam. Six filters were used, and their transmittance characteristics, measured with a Beckman Model DU spectrophotometer, are shown in Table 1. Since the orange (598-nm.) filter transmitted the least amount of light (see Table 1), it and an arbitrarily selected chamber were designated the standard for setting relative intensity levels. This was done by using a sensor probe with 22-F characteristics (responsive to 300-800 nm., with an essentially flat response to the radiant energy of wavelengths used in the present study). The probe was placed in front of the standard key, and this standard brightness was used to calibrate a Photovolt photometer (Model 520 M). The photometer was used, in turn, to equate radiance of the targets (i.e., the amount of colored light passed by each filter was equated). Equating individual targets to the photometric standard was accomplished by adjusting the iris on the projector lamp. This was done for all filter-key-light source combinations at the maximum intensity level (bright, B) for orange and for 2 lower levels designated medium (M) and dim (D). The B, measured 9 cm. from the standard key's surface using a Macbeth illuminometer, was 2.58 mL. The M was .275 density units less than B, and D was .75 density units less than B. The M and D for the standard were measured at 1.76 mL. and .68 mL., respectively, using the Macbeth illuminometer. A key peck of 5 gm. operated a microswitch and attached pulse former and counter. The microswitch was mounted behind the lower edge of each key. The first response (peck) on any 1 of the 6 keys activated a timer common to all pulse formers. The timer was set for 15 min. Counters recorded only during this 15-min. period and then automatically locked out of the circuit. Procedure. Incubation and hatching. A flat of 30 fertile eggs, collected on Monday of each week and placed in cold storage, was set in a commercial incubator the next day and/or on Friday p.m. for a total of 10 hatches. When one hatch was being set, eggs from a 2-wk.-old hatch were candled and then transferred in a light-proof container to a dark, forced-air incubator for the last week of incubation and hatching. The hatching
TABLE 1 SPECIFICATIONS OF INTBBFEEBNCB FILTERS Hue
Violet Blue Green Yellow Orange Red
Wavelength Percentage of Half bandwith (in nm.) at peak transmittance (in nm.) transnrittance
423 468 516 582 598 680
25
36 39 12 8 43
10.0 10.5 8.5 7.5
10.5 10.5
incubator was located in a dark room, and an infrared vision device was available for observation of hatching, though most checks were made in the dark, using tactile cues. Observations were begun at midnight of the twentieth day and at 8. a.m., noon and 4 p.m. on the twenty-first day. At these times, chicks that had hatched were isolated individually or in pairs in separate tray compartments for age and identification purposes. Pecking preference tests. Chicks that had, hatched within the same 4-hr, period were grouped into sets of 6 chicks. When the average age of the set was 23J^ hr., the chicks were isolated for J£ hr. in the darkened incubator room, in individual cubicles of a transport carton. At 24 hr., they were transferred in the carton, within a larger lightproof container, to the darkened test room. There the chicks were placed randomly in the 6 test chambers, facing the key, and left for 5 inin. of adaptation. During this time, the 6 interference filters and their respective intensity level (B, M, or D) were assigned to test chambers, according to predetermined random sequences. These sequences included the assignment of intensity level and the order of filter presentation over 6 successive test sessions. Filter order was restricted only by the requirement that each chick receive all 6 filters. Intensity levels were randomized without restriction. After the 5 min. of adaptation to the test chamber in the dark, the targets for all 6 chambers, were illuminated simultaneously. The first peck by any chick activated all counters and a timer. Fifteen minutes later the timer disconnected lights and counters, marking the end of the first test session. The chambers were then covered by a board so that the chicks remained in darkness during a 5-min. intertrial interval (ITI). During the ITI, the number of pecks per chick were recorded from counters, filters were shifted, and intensity levels were changed. Fifteen-minute test sessions and ITIs continued in this manner until the chicks had been exposed to all 6 interference filters.
Results and Discussion Data on one chick were eliminated because the experimenter misaligned the light.
404
G. J. FISCHER, G. L. MORRIS, AND J. P. RUHSAM
35
• 25 in £20
'400423 (violet)
468 (blue)
516 (green)
582598 (yellowXorange)
680 (red)
WAVELENGTH, NANOMETERS
FIGUBE 1. Mean number of pecks by darkhatched chicks, tested in the dark, at each of 6 target colors presented with radiant intensity randomized.
The mean number of pecks per filter color for the remaining 137 subjects and the standard error of the means are shown in Figure 1. A filter color by subjects repeated-measures design revealed that the differences among the means in Figure 1 were reliable (F = 2.60, df = 5/680, p = .05), These findings affirm unlearned, bimodal preferences for red and blue-violet regions of the spectrum in dark-hatched chicks tested in darkness. The finding of bimodal color pecking preferences agrees with Hess (1956). In fact, the overall shape of the relation between pecking and wavelength resembles closely the func-
tion reported by Hailman (1964) for gull chicks. All prior studies of chick color pecking preference, including, of course, those reporting unimodal preference in the orange region, tested chicks in the light. As indicated previously, the level of illumination in these studies varied. In an effort, then, to reproduce the unimodal color preference found in some of these studies (e.g., Gunther & Jones, 1962; Oppenheim et al., 1970), Experiment 2 tested dark-hatched chicks in the light at 3 levels of target radiance (B, M, and D), varied systematically. Since Curtius (1954) reported unimodality with a white but not with a black background, the targets in Experiment 2 were presented against white keys. EXPERIMENT 2: COLOR PREFERENCES IN DARK-HATCHED CHICKS TESTED IN THE LIGHT WITH DIFFERING TARGET INTENSITY Method Subjects. A total of 150 1-day-old chicks (2-3 sets of 6 chicks per hatch) were run from 9 successive hatches. A random one third (50) were assigned to each of the B, M, and D target intensities. Apparatus. The apparatus was the same as in Experiment 1 except for changes relating to testing in the light. For example, a 15-min. period of light adaptation was given immediately prior to the first test session. To prevent the subjects from seeing or pecking at the key during this time, metal doors, painted a flat gray, were suspended over the key wall of each cubicle. The B, M, and D levels were recalibrated as in Experiment 1, with new bulbs in all projectors. Dim, central fluorescent lighting 1.38 m. above the wire floor of the pecking apparatus provided 3.23 mL. at the pecking surface of the key in the standard chamber. When the standard filter, 598 nm., was projected at a bright intensity level onto this key, illumination at the key's surface measured 4.21 mL. Procedure. Within each hatch an equal number of cubicles were assigned randomly to 1 of the 3 radiance levels designated B, M, or D. Chicks were placed randomly into the cubicles (with metal doors covering the keys) for 15 min. of adaptation to the cubicles and to light. The metal doors were then removed, and the projector lights and automatic counters turned on. As in Experiment 1, these counters activated only after the first peck by any 1 of the 6 chicks. Six successive 15-min. test sessions over the 6 filters proceeded
COLOR PREFERENCES IN CHICKS as in Experiment 1 except that each chick pecked at all 6 colors at the same (B, M, or D) intensity.
Results On one test session the counter recording from the violet filter was cleared accidentally, and the number of pecks by that chick on that filter was treated as a missing observation. Data for 20 of the 150 chicks (12.6%) were excluded because they failed to peck at any 1 of the 6 filters. Of these 20 nonresponders, half (10) had been exposed to dim target intensity, 6 had medium, and 4 had bright. For the remaining 130 chicks that pecked at least once, the number of pecks per filter, per intensity level were subjected to a 3 X 6 (Intensity X Filter Color) repeated-measures design analysis of variance. The magnitude of pecking declined absolutely with decreasing target intensity, but the effects of intensity and its interaction with filter color were statistically insignificant (Fa < 1.00). The effects of filter color approached significance (F = 1.96, df = 5/634, p < .08), and the mean rates of pecking per filter wavelength and their standard error are shown in Figure 2. Individual comparisons among the means in Figure 2 showed that the only single color preferred over green was orange (F — 4.97, df = 1/634, p < .05), but violet and yellow, taken together, also were significantly preferred to green, and thus, to red (F = 4.10, df = 1/634, p < .05). These results affirm unlearned, bimodal pecking preferences for violet and yellow-orange in dark-hatched chicks tested in the light. DISCUSSION Present findings show bimodal color pecking preferences in dark-hatched domestic chicks, both when they are tested in darkness (with illuminated targets) and when they are tested in the light. Why the discrepancy between these and Hess's (1956) findings and those showing only a unimodal pecking preference at orange (e.g., Gunther & Jones, 1962; Oppenheim et al., 1970)? Curtius (1954) found unimodality when chicks were tested in the light with targets presented against a white background. In Experiment 2 of the present
405
35-
400423
468
516
582598
680
WAVELENGTH, NANOMETERS
FIGURE. 2. Mean number of pecks by darkhatched chicks, tested light adapted and in the light, at 6 target colors, all presented against white and at a random 1 of 3 intensities.
study, however, targets were presented against a white background at different levels of radiant energy, and bimodal preferences were found. Thus, the unimodal preference found by others cannot be attributed to any simple function of background and/or target intensity. Perhaps, investigators who did not find bimodality (e.g., Gunther & Jones, 1962; Oppenheim et al., 1970) might have, had they included violet as a test filter. Also, if target intensity is not important, then filter bandwidth is a possibility, as suggested by Hailman (1969). Specifically, narrow band-pass Wrattan filters presented at equal quantum intensity produced bimodal pecking preferences in gull chicks, whereas wide band-pass Wrattan filters and uncontrolled intensity produced unimodality in the red region. The interference filters used in the present study also passed narrow bands (see Table 1), but so did those used by Oppenheim et al. (1970; Oppenheim, personal communication, July 1973). This would seem to rule out the salience of filter bandwidth. Still another possibility is ambient light intensity, which varied in the studies cited and is known to affect the rate at which chickens peck for food. Why might domestic chicks show unlearned color pecking preferences for short.
406
G. J. FISCHER, G. L. MORRIS, AND J. P. RUHSAM
and long wavelengths with minimum at green? A negative pecking preference for green might have adaptative significance since green is a salient habitat color but is the background color for feeding. Thus, so-called preferences at violet and yelloworange (e.g., when tested in the light) might be preferences instead for objects that contrast maximally with green. Hailman (1968) suggested a similar explanation of bimodal color pecking preferences in gull chicks. Using a photochemical approach to account for gull chick color preferences Hailman (1964) predicted the V-like function relating pecking amplitude and wavelength from a ratio of transmission spectra for red and orange oil droplets. The form of Hailman's (1964), and of the present, function agrees reasonably well, too, with what might be expected from the reverse of the sensitivity curve for the photopigment, iodopsin. lodopsin from the chicken eye has maximum sensitivity, X max , at 562 nm. and closely resembles the photopic sensitivity curve for the chicken (Wald, Brown, & Smith, 1955). In the present study, pecking tended to be lowest for niters nearest to X max for iodopsin. This was especially so when testing was in darkness (see Figure 1), where iodopsin should be more sensitive. The Xmax at 562 nm. implies maximal sensitivity at green. As indicated, this would not be unreasonable, since green is a salient habitat color. As a background color for feeding, however, green might be approached readily but pecked at least. A problem for the behavioral and photochemical approaches discussed is their in-
ability to account for an unexpected and perplexing finding of the present study. This is an apparent shift in pecking preference away from red when colors are presented in the light (cf. Figures 1 and 2). Though minimum pecking always is at green, preferences peak at blue-violet and orange-red when testing is in darkness. When testing is in the light, preferences peak at violet and yellow-orange. REFERENCES Curtius, A. Uber angeborene Verhaltensweissen bei Vogeln inbesondere bei Huhnerkucken. Zeitschrift fur Tierpsychologie, 1954, 11, 94-109. Gunther, W. C., & Jones, R. K. Effect of nonoptimally high incubation temperature on frequency of pecking and on color preferences in the chick. Proceedings of the Indiana Academy of Sciences, 1962, 72, 290-299. Hailman, J. P. Coding of the colour preference of the gull chick. Nature, 1964, 204, 710. Hailman, J. P. Spectral reflectance of gull's bill: Physiological and evolutionary implications for animal communication. Science, 1968, 162, 139140. Hailman, J. P. Spectral pecking preference in gull chicks: Possible resolution of a species difference. Journal of Comparative and Physiological Psychology, 1969, 67, 465-467. Hess, E. H. Natural preferences of chicks and ducklings for objects of different .colors. Psychological Reports, 1956, 2, 477-483. Morris, G. L. Color preference in dark-hatched White Leghorn chicks. Unpublished Master's thesis, Washington State University, 1970. Oppenheim, R. W., Jones, J. R., & Gottlieb, G. Embryonic motility and posthatching perception in birds after prenatal gamma irradiation. Journal of Comparative and Physiological Psychology, 1970, 71, 6-21. Wald, G., Brown, P., & Smith, P. lodospin. Journal of General Physiology, 1955, 38, 623. (Received November 12, 1973)
