J. Dairy Sci. 84:807–813 American Dairy Science Association, 2001.
The Perception of Color by Cattle and its Influence on Behavior C.J.C. Phillips* and C. A. Lomas†2 *Department of Clinical Veterinary Medicine University of Cambridge, Cambridge CB3 0ES, UK †School of Agricultural and Forest Sciences University of Wales, Bangor, Gwynedd LL57 2UW, Wales, UK
ABSTRACT
INTRODUCTION
Experiments have suggested that cattle can only discriminate long wavelengths of light (colored red) from short (blue) or medium (green) wavelengths, and not short from medium wavelengths; however, stimuli were inadequately balanced for intensity. In this study, an initial group of calves was trained to discriminate light sources by intensity, and the intensities of short, medium, and long wavelength lights were then varied to determine when the calves perceived them to be isoluminant. A new group of calves was tested for their ability to discriminate between the three isoluminant sources and were able to discriminate between long and short or medium wavelengths (mean correct choice 82 and 89%, respectively) but had limited ability to discriminate between the short and medium wavelengths (three out of seven calves could just discriminate in the first eight tests, but thereafter they all selected at random). The response to three stimuli—novel, fearful, and their handler—was video-recorded in isoluminant short, medium, and long wavelengths and movement was assessed by image analysis. In the fear test (a loud noise behind them), the calves negotiated a barrier and concealed themselves more rapidly in the medium (58 s) than the short wavelength (95 s) light. They performed fewest movements in the medium wavelength light compared with the short and long wavelength lights in the novel stimulus and fear tests. They had stronger movement in the long than the short or medium wavelength light in the novel arena test and in response to the handler, and they took least time to reach the handler in the long wavelength. (Key words: cattle, color, wavelength, perception)
Cattle are believed to possess some color vision, in common with nearly all mammals (Jacobs et al., 1998), and early research (Rochon-Duvigneaud, 1943) suggested a high proportion of cones in the retina (2 to 3:1 rods:cones in the central region of the retina, increasing to 5 to 6:1 at the periphery. This agrees with their predominantly diurnal behavior, with crepuscular peaks for the grazing activity (Phillips, 1993). Peak photopic spectral sensitivities occur at 455 and 554 nm (Jacobs et al., 1998), suggesting dichromatic vision, but color discrimination potential is difficult to predict from knowledge of spectral sensitivity of terrestrial mammals (Jacobs, 1993), partly because of the moderation of wavelength when light is absorbed by the lens and the tapetum (Jacobs et al., 1998). Several discrimination experiments have suggested that cattle only possess the ability to differentiate long from short and medium wavelengths, and that they cannot distinguish short and medium wavelengths (Gere et al., 1981; Gilbert and Arave, 1986; Riol et al., 1989; Soffie et al., 1980). However, none of these tests adequately equated the light sources for intensity (Jacobs, 1993): Gilbert and Arave (1986) and Gere et al. (1981) did not correct for intensity at all, Soffie et al. (1980) varied the reflectance properties of their light stimuli in large steps, and Riol et al. (1989) equated the intensity of their sources spectrophotometrically, but only to the human eye. Gere et al. (1981) attempted to justify their lack of correction by conducting a separate (unpublished) experiment in which cattle were unable to discriminate different shades of gray. Intensity correction is best achieved by conducting an initial discrimination test with a similar group of conspecifics to determine when different light sources are isoluminant (Phillips and Prayitno, 1997). In this study, we sought initially to confirm that cattle only perceive differences between long and shortmedium wavelength light, equating the perceived stimulus brightness, which has never been done in experiments investigating color vision in cattle. The ability of cattle to discriminate light sources on the basis of brightness appears less than that of humans, but this
Received July 31, 2000. Accepted November 29, 2000. Corresponding author: C.J.C. Phillips; e-mail:
[email protected]. 1 Research supported in part by Ministry of Agriculture, Fisheries and Food studentship for C. A. L. 2 Current address: Ottley College, University of East Anglia, Ipswich, Suffolk, UK
807
808
PHILLIPS AND LOMAS
may be due to the measurement technique (Phillips and Weiguo, 1991), because their anatomical apparatus is more sensitive than ours (Walls, 1942). Jacobs (1993) reported that the utilization of color vision has received scant attention compared with investigations of its nature, and so experiments were also conducted to investigate whether cattle exhibit differences in behavior in isoluminant short, medium, and long wavelength colors. MATERIALS AND METHODS Experimental Design Initially a group of calves were trained to discriminate between different brightness levels in light sources. This was achieved with white lights, and after this the calves were offered a variety of intensities of short, medium, and long wavelength lights until there was no difference in perceived brightness levels. A new set of calves was then rewarded if they could discriminate between isoluminant short, medium, and long wavelength lights (with peak wavelengths of 415, 525, and 635 nm, respectively), and finally their behavioral response to different stimuli was compared in the three wavelengths of light. All calves were housed and managed in accordance with UK Codes of Recommendations for the Welfare of Cattle (MAFF, 2000) and were protected by the UK Animals (Scientific Procedures) Act, 1986. Experiment 1: The Ability of Calves to Discriminate Among Short, Medium, and Long Wavelengths Training to discriminate light sources on the basis of intensity. Eight female Friesian calves of 8 wk of age commenced the experiment. For this purpose, white lights were used for the training, which followed the procedure of Phillips and Weiguo (1991). Tests commenced after each subject had become adapted to the light intensity in a light-proof experimental room, assumed to occur after 30 min. The eight calves were positioned in front of two chambers containing a bright and dim white light 0.9 m from the ground. The initial intensities were 1.2 × 1018 and 2.4 × 1020 photons, as measured in the six directions of the faces of a cube (Smith, 1989) with a quantum sensor with microvolt integrator. Four calves were trained to select the brighter of the lights, and the other four were trained to detect the dimmer light. Calves in both groups were offered a reward of 50 g of concentrate when they entered the correct chamber and were prevented access to the incorrect chamber for the first 24 tests. After this, the calves were tested 32 times with access to both chambers, in which the light intensities were varied Journal of Dairy Science Vol. 84, No. 4, 2001
(
Figure 1. Transmission spectra for the short ( ), and long ( ) wavelength light sources.
), medium
between 1.2 × 1018 and 2.4 × 1020 photons. During all tests both chambers contained a feed reward in a bucket, but in the incorrect chamber the reward was made inaccessible by placing a mesh over the feed, thus ensuring that calves could not use olfaction as a cue. The correct chamber was randomly alternated between the two chambers to prevent side preferences influencing the results. Successful discrimination was deemed to have occurred when at least a 75% correct choice was achieved over 16 continuous tests. Equating the intensity of colored lights by calf discrimination. Different wavelengths of light were produced by fitting three filters (filter no. 106, primary long wavelength, transmission 9.3%; no. 139, primary medium wavelength, peak wavelength 525 nm, transmission 1.3% and no. 120 deep short wavelength, peak wavelength 415 nm, transmission 15%; Lee Filters, Andover, Hampshire, UK) over a tungsten filament bulb (Figure 1). The intensity of the short, medium, and long wavelength lights were varied in pairs in two 2- × 1-m lightproof chambers. The difference in intensity of the two lights was gradually reduced until the calves were unable to discriminate between them, with one half of the calves selecting the brighter of the two lights and one half the dimmer, as in white light selection. Initially, long wavelength and short wavelength lights were compared, then long wavelength and medium wavelength and finally short wavelength and medium wavelength. The intensity recordings from the quantum sensor were corrected for the proportional transmission of light through the filters and its spectral composition (Figure 1), the spectral output of the 100-W tungsten filament lamp at 2700°K (Figure 2), and the spectral sensitivity of the quantum sensor to determine the quantity of light reaching the calves’ eyes from the
PERCEPTION OF COLOR AND EFFECTS ON BEHAVIOR
(
Figure 2. Energy transmission (%) of the tungsten filament lamp ), and light meter ( ).
different filtered lights. In addition, knowing the perceived intensity of each light source and the luminescence, the proportion of the spectral output of each light to which the cattle were sensitive was calculated, and the approximate maximum perceived wavelength of light was determined. Testing the ability of calves to discriminate between the three wavelengths of light. In the second part of this experiment, we attempted to train a new group of 11 Friesian calves to discriminate between the three colors of light, which were made isoluminant according to the results of the previous test (1.4, 0.68, and 2.2 × 1020 photons, for the short, medium, and long wavelength lights, respectively, when measured at calf eye level). The calves, five steers and six heifers, were 8 wk old at the start of the experiment, and the lights were again compared in pairs. They were allocated at random to be trained as short wavelength selectors
Figure 3. Chamber for testing the response of calves to a novel stimulus.
809
(three calves), medium wavelength selectors (four calves), and long wavelength selectors (four calves) and used in comparisons of first, long and medium wavelength lights, then long and short wavelength lights, and finally short and medium wavelength lights. Initially, each test commenced with 24 training runs, conducted as for white light intensity discrimination training, then each calf was tested eight times/day for its ability to discriminate. When all calves had achieved a correct success proportion of 75% or more for each of 2 consecutive days, the test was terminated. One long wavelength-selecting calf had to be removed from the experiment after 3 d as it was distressed by the testing procedure. Statistical analysis. The difference in the mean correct choice proportion for each day from that expected by chance was examined for significance by the chi-square test. The difference between the two groups of calves in each test was examined by an analysis of variance of the mean correct choice for each calf, using the Minitab statistical package (Ryan et al., 1985). Experiment 2: The Behavior of Calves in Short-, Medium-, and Long-Wavelength Light Nine female Friesian calves of approximate age 8 wk at the start of the experiment were housed in two 3.5 × 3.5 m lightproof chambers for three 16-d periods. The lighting provision was either a sequence of medium, long, and then short wavelength light (five calves) or short, medium, and then long wavelength light (four calves). The light sources were the same as those used in experiment 1, which were set to provide the same intensities determined by calves in the previous experiment as isoluminant (1.4, 0.68, and 2.2 × 1020 photons, for the short, medium, and long wavelength lights, respectively). The light sources, which remained on for 24 h/d, were suspended from the roof of the chamber at a height of approximately 2 m. During the last 2 d of each period, the calves were individually taken to a 2.5- × 3.5-m lightproof test chamber with the same lighting provision as the calves housing chambers to investigate their responses to three different types of stimuli—novel, fearful, and their handler. The response to each stimulus was measured over a 4-min period, and the series of three tests were run in triplicate over two 8-h sessions in the 2 d. Calves were selected at random for each replicate, which lasted approximately 20 min/calf, and were returned to their housing afterwards. The novel stimulus was the first test for each calf in each replicate and was provided by allowing the calf to enter the chamber (Figure 3), which they had not experienced before. A rectangle 0.7 m from the chamber Journal of Dairy Science Vol. 84, No. 4, 2001
810
PHILLIPS AND LOMAS
boundaries was painted on the floor, and the time that the calf had its front feet in the outer section was recorded as an estimate of the time spent by the calf in the vicinity of the boundary wall of the chamber. It is currently unclear whether extended time spent near the walls indicates anxiety (McGrath et al., 1999), increased locomotory behavior (Harley and Martin, 1999), or exploration. The fear stimulus was created by dropping a 0.5- × 1-m wooden board to the floor behind the calf at the entrance to the chamber (Figure 4). The response measured was the time taken for the calves to conceal themselves from the stimulus by moving to the diagonally opposite corner of the chamber, negotiating two 1- × 0.8-m barriers en route. In the test of response to the calves’ handler, the time that the calves took to negotiate two internal barriers and reach their handler at the opposite side of the chamber was recorded (Figure 5). The handler was familiar to the calves, having previously provided the positive stimulus of food and water and being responsible for the well being of the calves. The barriers were constructed of steel bars so that the calf could see the person from the entrance to the chamber.
Figure 5. Chamber for testing the response of calves to the handler (H) (direct route of calf 䊉 䉴).
In each test, the movement of the calf was recorded by an overhead camera (National Panasonic WV-1450/ B, Mitsushita, Uxbridge, Middlesex) connected to a videodigitizing modem (Sallinen and Hatunen, 1993) installed in a Matmos 486 computer. The modem recorded each occasion (termed a ‘movement’) during which 15% of the integer valued picture elements (pixels) changed brightness, reflecting major calf movements in the pen. For each movement, the proportion of pixels changing intensity, which was termed the movement strength, was recorded. Statistical analysis. The total movement data were transformed to normality by taking the square root of the data in both the fear and response to handler tests; the data for time to conceal themselves in the fear test were similarly transformed. The significance of treatment differences was examined by analysis of variance, with wavelength, calf, and replication as fixed effects in the model using the statistical package Minitab (Ryan et al., 1985). RESULTS Experiment 1. The Ability of Calves to Discriminate Among Short, Medium, and Long Wavelengths
Figure 4. Chamber for testing the response of calves to the fearful stimulus of a board dropped behind the calf (direct route of calf = 䊉 䉴). Journal of Dairy Science Vol. 84, No. 4, 2001
All calves successfully learned to discriminate between the bright and dim white lights within 32 tests. The short-, medium-, and long-wavelength lights were found to be isoluminant at a ratio of 1.4, 0.68, and 2.2 × 1020 photons, respectively, as measured by the quantum sensor. After making corrections for the filters, the spectral output of the tungsten filament lamp, the spectral sensitivity of the quantum sensor, and the ratio of isoluminancy of the short, medium, and long wavelength spectra tested were 1.9 × 1020, 1.0 × 1020, and 1.9 × 1020 photons, respectively. The calves’ sensitivity to the middle wavelength light was, therefore,
PERCEPTION OF COLOR AND EFFECTS ON BEHAVIOR
almost twice that of the short or long wavelength lights. The upper limit of spectral sensitivity was calculated to be approximately 620 nm. There were no significant (P < 0.10) differences in the mean correct choice proportions of the two groups of calves over the entire test in any of the three wavelength comparisons. After training, both groups of calves were able to discriminate between long and medium wavelengths within 4 d (Figure 6a). Discrimination between long and short wavelengths was demonstrated by both groups even faster in 2 d (Figure 6b). The medium selectors were just able to distinguish medium from short wavelength light on d 1, but the short wavelength selectors were unsuccessful (Figure 6c). After the first day neither short nor the medium selec-
811
tors were able to select the correct stimulus. The mean correct choice proportions in the final 2 d were 89, 82, and 52% for the long:medium, the long:short, and the short:medium wavelength tests, respectively. Experiment 2. The Behavior of Calves in Short-, Medium-, and Long-Wavelength Light In the novel pen, the calves performed the greatest number of movements and the movements were strongest in the long wavelength light, with the least activity in the medium and an intermediate level of activity in the short wavelength light (Table 1). There was no effect of wavelength on the amount of time spent investigating the chamber boundaries. In the fear test, the calves took less time to negotiate the barrier and conceal themselves in the medium than the short wavelength light. They had fewest movements in this wavelength and the least total movement. In the test of approach to their handler, there was no difference in the total movement between colors, nor in the number of movements made by each calf, but the calves reached the handler faster and performed stronger movements in the long wavelength light than the other two wavelengths. DISCUSSION
Figure 6. Correct choice proportions (%) of calves trained to select short (䊐 䊐), medium (䊉 䊉) and long (䊏 䊏) wavelengths of light, in comparison of a) medium and long wavelength lights, b) short and long wavelength lights and c) short- and medium-wavelength lights, over 4- to 5-d periods. * = P < 0.05, ** = P < 0.01, *** = P < 0.001.
The ability of cattle to differentiate between the long wavelength and both the short and medium wavelengths of light confirms previous observations, even though they had been conducted with inadequate experimental designs (Jacobs, 1993). In our study, it is theoretically possible that the failure of the calves to cooperate could explain the lack of successful discrimination in the third test, short:medium wavelength discrimination. However, there are two reasons why this was not the case. First, the short:medium wavelength discrimination test was the second test for both the short and medium wavelength selectors, with their having previously performed the long:medium wavelength discrimination and long:short wavelength discrimination, respectively. Earlier, the long wavelength selectors had succeeded in discriminating first between long and medium wavelengths and then long and short wavelengths, with no evidence of a reduction in discrimination ability in their second test as a result of boredom. On the contrary, they seemed to improve in the second test. Second, previous research has indicated that visual discrimination in calves improves over a period of about 40 tests, with no evidence of fatigue or boredom becoming established (Phillips and Weiguo, 1991). The evidence of some limited ability to discriminate short from medium wavelengths is novel but not entirely surprising. Color discrimination relies on comparJournal of Dairy Science Vol. 84, No. 4, 2001
812
PHILLIPS AND LOMAS Table 1. Calf responses to novel, fearful, and handler stimuli in short-, medium- and long-wavelength light in experiment 2. Wavelength Novel stimulus Proportion of time investigating chamber boundaries (%) Movement number/min Movement strength (% pixel change) Total movement (√ pixel change/min) Fearful stimulus Proportion of time taken to hide (√ %) Movement number/min Movement strength (% pixel change) Total movement (√ pixel change/min) Handler stimulus Proportion of time to reach person (√ %) Movement number/min Movement strength (% pixel change) Total movement (√ pixel change/min)
Short
Medium
Long
SED1
87 1.9 20.2 3.0
81 1.2 18.5 2.3
87 2.5 21.3 3.5
3.2 0.27*** 0.47*** 0.22***
6.3 1.7 20.6 2.8
4.9 1.2 20.5 2.3
5.5 1.6 21.7 2.8
0.50*** 0.18* 0.67 0.16**
6.4 1.9 19.4 2.8
5.3 1.3 19.3 2.5
3.7 1.5 21.3 2.7
0.65*** 0.27 0.59*** 0.23
1
Standard error of the difference between two means. *P < 0.05, **P < 0.01, ***P < 0.001.
ison of the spectral information from the medium-long wavelength cones (which comprise approximately 93% of the cones, Jacobs, 1993) and the short wavelength cones. Our comparison of short and medium wavelengths would give limited opportunity for such contrast, and the rapid decline in the calves’ ability to make a successful choice suggests that they resorted to random selection as a result of the difficulty of the task. The initial success demonstrated by 40% of the calves suggests that they were all able to make some discrimination or that cattle vary genetically in their ability to make this contrast, in the same way that considerable genetic variation exists in humans in their ability to distinguish green/red coloration. The prediction that spectral sensitivity of the calves terminated at about 620 nm agrees with electroretinogram measurements by Jacobs et al. (1998), who demonstrated that sensitivity of medium-long wavelength cones in cattle declines rapidly above 600 nm. The reduction in activity of calves in the medium wavelength light in the fear test may have been due to greater stereoscopic acuity in the medium wavelength light enabling them to reach the opposite side of the chamber more quickly. All things being equal, there is an increase in human visual acuity in medium wavelength compared with short wavelength stimuli (Koffka and Harrower, 1932) due to better color rendering properties (Yizhong, 1984). However, it has been observed that children are more likely to prefer the color green in fearful circumstances, perhaps because they associate the color of foliage with security (Versobin and Zhidkin, 1980). The reduced activity in medium wavelength light may, therefore, have been due to reduced fear. The stronger movement in the long wavelength light in the novel stimulus and handler tests is similar to Journal of Dairy Science Vol. 84, No. 4, 2001
results in other species. In tests that used identical luminaires and filters to those employed in these experiments, chickens were more active in long than medium or short wavelength light, but they were not more aggressive (Prayitno et al., 1997). This could have been due to long wavelength light penetrating the skull and directly stimulating the pineal gland, a phenomenon observed in birds (Hartwig and van Veen, 1979). However, there is evidence of increased activity in long wavelength light in other species—mice show greater defensive activity in long wavelength than white light (Kemble and Goblirsch, 1997) and in humans, long wavelength light has been found to be arousing (Hamid and Newport, 1989) and evocative of active feelings (Levy, 1984) and anger and (or) tension (Gardano, 1986). There may, therefore, be an arousal response to long wavelengths that is common to higher animals, which relates to the survival value of a vigorous response to the color of blood. The discovery that cattle approached a person faster and with stronger movements in long-wavelength light may explain why toreros challenge bulls with a red cape in the final stages of bull fights. In humans, short-wavelength light evokes the opposite feelings to long-wavelength light—relaxation (Gardano 1986), sadness, and fatigue (Levy, 1984), suggesting a low level of arousal (Walters et al., 1982). Thus ‘warm’ colors evoke active feelings, and ‘cool’ colors are experienced as sedate (Levy, 1984). In terms of work performance, long wavelength coloration is distracting to humans (Kwallek et al., 1988). These differences are more pronounced in females than males (Knez, 1995; Peretti, 1974).
PERCEPTION OF COLOR AND EFFECTS ON BEHAVIOR
CONCLUSIONS Cattle were able to distinguish long from short or medium wavelength light, or in human color terms, red from blue or green lights. However, compared with humans, they showed very limited ability to differentiate medium from short wavelength light (green from blue light). This is consistent with dichromatic vision in cattle, whereas most humans are trichromats. The maximum wavelength of light that could be perceived was approximately 620 nm, which is comparable to the rapid decline in sensitivity of long wavelength receptors in humans above 600 nm (Schnapf et al., 1987). Cattle were able to negotiate a barrier and conceal themselves faster in medium (green) than short (blue) wavelength light in response to a fearful stimulus. They approached their handler faster and were more active in long wavelength (red) than medium wavelength (green) or short wavelength (blue) light. These changes in behavior in different wavelengths of light, although only recorded over a short period, have the potential to be used to improve the environment for cattle in specific situations. For example, the use of green lighting in abattoirs, where cattle experience fear, could expedite movement, and the use of blue lighting in the parlor might reduce the cows’ activity and level of arousal, making them easier to milk. Further research is required to test these hypotheses in situ. REFERENCES Gardano, A. C. 1986. Cultural influence on emotional response to color: A research study comparing Hispanics and non-Hispanics. Am. J. Art Ther. 24:119–124. Gere, T., S. Kiss, and Z. Szildgyi. 1981. Study of Colour Perception in Cattle. Proc. Res. Center Anim. Husb. Prod., Godollo, Hungary. Gilbert, B. J., and C. W. Arave. 1986. Ability of cattle to distinguish among different wavelengths of light. J. Dairy Sci. 69:825–832. Hamid, P. N., and A.G. Newport. 1989. Effect of color on physical strength and mood in children. Percept. Motor Skills 69:179–185. Harley, C. W., and G. M. Martin. 1999. Open field motor patterns and object marking, but not object sniffing, are altered by ibotenate lesions of the hippocampus. Neurobiol. Learn. Mem. 72:202–214. Hartwig, H. G., and T. van Veen. 1979. Spectral characteristics of visible radiation penetrating into the brain and stimulating extraretinal photoreceptors. J. Comp. Physiol. 130:277–282. Jacobs, G. H. 1993. The distribution and nature of color vision among the mammals. Biological Reviews 68:413–471.
813
Jacobs, G. H., J. F. Deegan, and J. Neitz. 1998. Photopigment basis for dichromatic color vision in cows, goats and sheep. Vis. Neurosci. 15:581–584. Kemble, E. D., and M. J. Goblirsch. 1997. Effect of illumination condition on risk assessment behaviors of mice. Psych. Rec. 47:167–174. Knez, I. 1995. Effects of indoor lighting on mood and cognition. J. Exp. Psychol. 15:39–51. Koffka, K., and M. R. Harrower. 1932. Contributions to Gestalt psychology XXII Color and organization, Part II. Psych. Forsch. 15:194–274. Kwallek, N., C. M. Lewis, and A. S. Robbins. 1988. Effects of interior color on workers’ mood and productivity. Percept. Mot. Skills 66:123–128. Levy, B. I. 1984. Research into the psychological meaning of color. Am. J. Art Ther. 23:58–62. McGrath, M. J., K. M. Campbell, M. B. Veldman, and F. H. Burton. 1999. Anxiety in a transgenic mouse model of cortical-limbic neuro-potentiated compulsive behavior. Behav. Pharmacol. 10:435–443. Ministry of Agriculture, Fisheries and Food (MAFF) 2000. Codes of Recommendations for the Welfare of Cattle. http://www.maff. gov.uk/animalh/welfare. Accessed Oct. 4, 2000. Peretti, P. O. 1974. Effects of sex and experimental background on color-mood association. Pak. J. Psych. 7:33–40. Phillips, C.J.C. 1993. Cattle Behavior. 232 pp. Farming Press, Ipswich, United Kingdom. Phillips, C.J.C., and D. S. Prayitno. 1997. Equating the perceived brightness of short wavelength and long wavelength lights to hens. Br. Poultry Sci. 38:136–141. Phillips, C.J.C., and L. Weiguo. 1991. Brightness discrimination by cattle relative to that of humans. Appl. Anim. Behav. Sci. 31:25–33. Prayitno D. S., C.J.C. Phillips, and D. K. Stokes. 1997. The effects of color and intensity of light on behavior and leg disorders in broiler chickens. Poult. Sci. 76:1674–1681. Riol J. A., J. M. Sanchez, V. G. Eguren, and V. R. Gaudioso. 1989. Color perception in fighting cattle. Appl. Anim. Behav. Sci. 23:199–206. Rochon-Duvigneaud, A. 1943. Les Yeux et la Vision des Vertebrates. Masson, Paris, France. Ryan, B. F., B. L. Joiner, and T. A. Ryan 1985 Minitab Handbook 2nd ed. Duxbury Press, Boston, MA Sallinen, S., and E. Hatunen. 1993. Kettu System User’s Manual, Software Version 2.2. Oulu, Finland. Schnapf, J. L., T. W. Kraft, and D. A. Baylor. 1987. Spectral sensitivity of human cone photoreceptors. Nature 325:439–441. Smith, A. 1989. Measurement of light intensity. Pages 27–29 in Photoperiodic Manipulation of Cattle Production. C.J.C. Phillips and J. M. Forbes ed. Dairy Research Unit, Tech. Publ. No. 4, University of Wales, Bangor, UK. Soffie M., G. Thines, and U. Falter. 1980. Color discrimination in heifers. Mammalia 44:97–121. Versobin, V. N., and V. N. Zhidkin. 1980. A study of the color preference in preschool children who are experiencing positive and negative emotion. Voprosy Psikhologii 3:121–124. Walls, G. L. 1942. The Vertebrate Eye and Its Adaptive Radiation. Cranbook Inst. Sci., Bloomfields Hills, MI. Walters J., M. J. Apter, and S. Sveback. 1982. Color preference, arousal, and the theory of psychological reversal. Motiv. Emotion 6:193–215. Yizhong, Z. 1984. The effect of color-rendering properties of light sources on visual acuity. Acta Psychol. Sin. 16:193–203.
Journal of Dairy Science Vol. 84, No. 4, 2001