Journal of Comparative and Physiological Psychology 1975, Vol. 89, No. 9, 1070-1076
Scotopic Sensitivity in Coyotes (Cam's latrans) Steven W. Horn and Philip N. Lehner Laboratory of Animal Behavior, Colorado State University The absolute scotopic limen for light intensity was measured for three 10mo-old female coyotes. The methodology was similar to that used by Blough in determining psychophysical thresholds in pigeons. Three coyotes were operantly conditioned to depress one of two foot treadles, left or right, depending on the condition of the stimulus light. Scotopic adaptation curves for each coyote were generated. Nonlinear regression curves were then fitted to the raw data. The mean scotopic thresholds did not differ significantly. However, time to the curves' asymptotes did differ significantly for one of the coyotes. The adaptation curves showed a distinct rod-cone "break," and retinal histology confirmed that the coyote has a duplex retina with a preponderance of rods. In addition, electroretinographic analysis showed the relative contributions of rods and cones at various light intensities and indicated a rod-cone break at approximately 15 min. Scotopic spectral sensitivity curves were also generated. The coyotes' scotopic visual threshold is exceeded by the natural illumination available under many nocturnal conditions. The results are discussed in relation to the ecology of the species.
The purpose of this experiment was to determine scotopic visual sensitivity in the coyote, utilizing the psychophysical technique first described by Blough (1955, 1958, 1961). Blough's procedure was based on a technique developed by von Be'ke'sy (1947) for tracing the human auditory threshold. Blough's procedure was designed to study visual sensitivity in the pigeon (Columba lima). More recently this procedure has been used to study dark adaptation in the American red squirrel (Dippner & Armington, 1971), visual sensitivity in the black-bellied tree duck (Hersloff, Lehner, Bolen, & Rylander, 1974), and scotopic sensitivity in the mallard duck (Wells, Lehner, Bolen, & Rylander, 1975). In the experiments by Blough (1955, 1956, 1958), dark-adaptation curves obThis report is Contribution No. 5 of the Laboratory of Animal Behavior at Colorado State University. The authors are indebted to T. J. Boardman and P. Biondini for their assistance in the data analysis. Mary Fisher assisted in the retinal histology. The authors are also indebted to E. L. Pautler for his cooperation and assistance in the electroretinographic analysis as well as his review of the manuscript. Requests for reprints should be sent to Philip N. Lehner, Laboratory of Animal Behavior, Colorado State University, Port Collins, Colorado 80523.
tained from the pigeon suggested a duplex retina. A distinct rod-cone "break" in the curve appeared after 20 min, followed by a rapid increase in visual sensitivity reaching a maximum at about 50 min. Wells et al. (1975) showed similar results with the mallard and supported their visual adaptation curves with retinal histology. We measured the scotopic visual threshold in the coyote to show the adaptive significance of the visual system under study. To support our inferences we also examined the retinal structure of the coyote and conducted electroretinographic (ERG) analysis during the process of dark adaptation. Both the retinal histology and ERGs are correlated with the adaptation curves in the Discussion section. The coyotes' visual thresholds are discussed in terms of the ecology of the species in relation to the amount of natural illumination present under various nocturnal conditions. METHOD Subjects The experimental subjects were three 10-moold female coyotes obtained from dens in Colorado at approximately 7 days of age. The pups were hand reared in the laboratory and were housed in individual cages. Food deprivation schedules varied throughout the study, and the animals averaged 75%-85% of their free-feeding weights.
1070
SCOTOPIC SENSITIVITY IN COYOTE
1071
fore, the subject could get reinforcement one time in five by depressing the left treadle for a short interval immediately after eating. This ensured that the subjects attended to the stimulus and that the stimulus intensity remained within a narrow range oscillating around the subjects' visual thresholds. The stimulus intensity was regulated by the amount of amperage to the stimulus light. This of course resulted in a change in the lamp's spectral characteristics, which will be considered in the interpretation of the results (see Discussion). A Rustrak analog strip-chart recorder plotted the adaptation curves directly as a function of milliamperes of current to the stimulus lamp. Sample points (milliamperes) taken .every 37 sec from the strip charts were used for the data analysis. A correlation between milliamperes to the stimulus lamp and illuminance (footcandles) was determined by measurements with a photometer (Photovolt Corp.), and a logarithmic relationship was demonstrated. For comparison with other studies, including measurements of illumination from natural situations, the stimulus intensities are reported in both footcandles and milliamperes. In order to confirm the duplex nature of the Procedure coyote retina, histological examinations were perThe procedure used in this experiment was formed. In addition, the domestic dog's (Canis basically the same as the psychophysical tracking famttiaris) retina was also examined for comparaprocedure described by Blough (1958) and Hers- tive purposes. One female coyote, a litter mate of the experiloff et al. (1974), in which the stimulus intensity is under the control of the subject. Each subject mental subjects, was anesthetized while light was preexposed in its home cage under normal adapted under normal room illumination, and an room illumination provided from eight 116.8-cm, electroretinograna (ERG) was recorded during 400-W Cool White General Electric fluorescent the process of dark adaptation. The ERG was obbulbs. The subjects were taken directly from their tained by means of a cotton wick electrode posihome cages and placed in the test chamber, which tioned on the cornea. The electrode was connected was in total darkness. The only source of light to a Grass EEG preamplifier by means of a salt was from the 20-mm stimulus patch which was bridge. A needle scalp electrode served as the always preset at a constant intensity of 22 X 10~3 reference.2 The stimulus intensity provided by a ftc. (.2368 Ix) at the beginning of each session. Grass PS Photostimulator was raised in fractions Each depression on the right foot treadle auto- of log units until a 60-MV b-wave was recorded. matically reduced the stimulus intensity by ap- This was the operational definition of the ERG proximately 1 X 10~4 ftc. (.0011 lx)/sec of treadle threshold. This process continued at various indepression time, while left treadle depressions tervals over a period of 40 min, at which time the increased the stimulus intensity by the same process of dark adaptation was essentially comamount. The subjects depressed the right treadle plete. The b-wave amplitude of the ERG was deuntil the stimulus was no longer visible, at which termined in response to chromatic stimulation time they would then depress the left treadle until (equal energy) from 400-650 nm. This action specthe stimulus was again within their visible range. trum was analyzed in terms of the established inAs a result of this, the stimulus intensity oscil- tensity-amplitude relationship to provide a speclated above and below the subjects' visual thresh- tral sensitivity curve, according to the method of old for light intensity during the process of dark Tomita, Kaneko, Murakami, and Pautler (1967). adaptation. Depressions on the right treadle also turned RESULTS the stimulus light off on a modified variable inThe mean adaptation curves for each terval schedule of 22 sec (VI 22); depressions on the left treadle then provided reinforcement on a coyote are plotted in Figure 1, along with modified variable interval 10-sec (VI 10) schedule. the mean curve for all three subjects. The The VI schedule was modified such that it was vertical axis of the graph is presented both not in force unless the subject kept the foot treadle in the depressed condition. After reinforcement in milliamperes, since these were the units on the left treadle, the stimulus light remained used in the statistical tests, and in footoff on a variable ratio of 5 (VR 5) schedule. There- candles, since these units depict amounts of
Apparatus A modified Skinner box was used as the test chamber. The test chamber contained two foot treadles (40.6 cm apart) placed at one end of the chamber and separated by Plexiglas partitions. A 20-mm-cireular stimulus patch was centered above and between the treadles at the coyotes' eye level, and a food tray was centered directly below the stimulus patch. Three-gram Gaines Prime dog food pellets dispensed individually from a Davis Model No. 310 Universal Feeder were used as reinforcers. The treadles were connected through electrical relays to a Harvard cumulative recorder, a Rustrak analog recorder, and counters. The stimulus light source was a No. 1820-28VDC General Electric miniature bulb in a Lehigh Valley Electronics multistimulus rear-screen projector. The stimulus intensity and the various reinforcement contingencies were controlled and monitored from an adjacent room, using standard electromechanical modules and relays. A BRS Foringer audio generator provided masking noise and a 3-sec tone which accompanied the presentation of .reinforcement.
1072
STEVEN W. HORN AND PHILIP N. LEHNER 218 120
65
COYOTE # 2
60
68
COYOTE* 4 COYOTE* 5 •—/VERAGE
55 50
38 Q x 22
45
12
40
7
35
4
30
2
1.2 5
10
15
20
25
30
35
40
45
50
55
60
TIME (minutes) FIGURE 1. Individual and mean scotopic sensitivity curves.
illumination irradiating from the stimulus patch. In order to predict both a base level amperage (i.e., the subject's dark-adapted light-intensity thresholds) and the time at which the base level was reached (the point at which the process of dark adaptation was completed), a single exponential nonlinear regression of the form y = a + f}eri was generated for use as a model of a curve of best fit. The variables were denned as follows: a = the base level amperage; 0 = the difference between the starting amperage and the base level; ? = the slope of the curve; t = time. A one-way analysis of variance was performed on the differences between the adaptation curves for all three coyotes. No reliable differences emerged between the curves' base level intensities, F(2, 6) 1.39, p > .31. However, there was a reliable difference between the curves' slopes, F(2, 6) 6.13, p < .04. This difference was the result of one coyote which consistently reached base level amperage (threshold intensity) faster than the other two subjects (Figure 1). The base levels (thresholds) for all three subjects had a predicted grand mean of 26 mA or 1.4 X 10~3 ftc. (.0151 Ix). Figure 2 depicts the predicted average 'dark-adaptation curves for each subject and for all three coyotes together, which may be
more representative of the species as a whole. The curves are very similar with the exception that Coyote 2 reached base level amperage within 20-30 min, while the other subjects were consistent in reaching threshold in 30-40 min. The coyotes' mean adaptation curve as plotted in Figure 1 does not show a definite break during the process of dark adaptation, but a partial plateau is evident between 20 and 35 min. However, when the coyotes' mean adaptation curve is plotted with stimulus intensity in log units (Figure 3) a definite break was found at 15 min. Blough (1955) interpreted the break as a reflection of the initial adaptation of the cones and subsequent adaptation of the rods. Wells et al. (1975) also found a break for the mallard at about 25 min. When electroretinographic analysis was performed, the break described in the preceding paragraph was clearly evident at approximately 15 min after the start of dark adaptation (Figure 3). Figure 4 presents the scotopic spectral response and sensitivity curves obtained for the coyote. The spectral response curve (Figure 4A) should be distinguished from the spectral sensitivity curve (Figure 4C), which is a plot of the reciprocal of the spectral light intensity at each wavelength just
SCOTOPIC SENSITIVITY IN COYOTE
1073
COYOTE NO. 2
COYOTE NO. 4 COYOTE NO. 5 OVERALL AVERAGE
25.
375
25.0
TIME (MINUTES)
FIGURE 2. Predicted average dark adaptation curves.
sufficient for eliciting an equal amplitude response. Conversion from the spectral response curve to the spectral sensitivity curve is necessary for the comparison of data from different sources. The procedure (based on Tomita et al., 1967) was as follows: Upon recording the spectral response at a given wavelength, the intensity-amplitude relation was obtained utilizing white light; for example, a response amplitude equal to the peak value (40 at 500 m/i) in Figure 4A is produced by white light of
— 1.5 log-unit intensity (Figure 4B). Alsor the response amplitude (10) at 625 m/j is reproducible with white light of —3.5 logso i 40-
]30-
S20-
• PSYCHOPHYSICAL (BthovicrilO EPS
1.5
400
450
500
55O
600
WAVELENGTH (nyi)
0.5
15
20
25
TIME Mnuto)
FIGUBE 3. Dark adaptation curves obtained from electroretinagraphic (ERG) and psychophysical procedures.
FIGURE. 4. Data utilized in deriving the spectral sensitivity curves from the spectral response curve: A, spectral response curve; B, the intensityamplitude relation of the unit shown in A to white light; C, spectral sensitivity curve derived from A and B (solid line) along with Dartnall's nomogram (1953) with the same peaking wavelength, 500 HIM (broken line).
1074
STEVEN W. HORN AND PHILIP N. LEHNER
unit intensity. This means that a two logunit (20-fold) increase in light intensity is necessary for the 625-rmt light to produce a response equal to the peak value (40 at 500 m/i) or, conversely, that the sensitivity at 625 nui is one twentieth of that at 500 nut. The sensitivity to other wavelengths was determined likewise and plotted as a solid line in Figure 4C. For comparative purposes the broken line illustrates Dartnall's nomogram (1953) with the same peaking wavelength of 500 m/j. At this time no quantitative data have been collected from the histological examinations made on the coyote retina. The coyote definitely has a duplex retina (both photoreceptor cells present) and appears to have a preponderance of rods. Examinations of sections taken from the domestic dog showed no great differences. DISCUSSION Because of the coyote's nocturnal-crepuscular behavior, one might hypothesize that it would require a duplex retina with rod dominance (Chesness, Note 1). This hypothesis is only weakly supported by the original adaptation curves (Figure 1) obtained during psychophysical testing. The curves in Figure 1 show no pronounced break that would clearly delineate predominantly rod and cone functions in the retina. However, there appears to be a plateau between 20 and 35 mm and a subsequent gradual attainment to a lower base level, which may reflect the duplex retina. When an electroretinogram was made, this apparent rod-cone break was most evident at about 15 min into the process of dark adaptation. This apparent contradiction was corrected when the curves were plotted with footcandles converted to log units (Figure 3). Probably the single most important factor involved in the differences between the curves obtained psychophysically (Figure 1) and the one obtained during electroretinographic analysis has to do with what was actually being measured. During the psychophysical testing procedure, the subjects were required to make a value judgment as to whether the stimulus evoked a visual
sensation. When we made an electroretinogram on an anesthetized animal, we were not asking for the perception of a sensory response but were simply measuring the receptory capabilities of the coyote retina through the change in the corneal potential. Therefore, we were measuring two completely different processes: one dealing with perception giving rise to a sensation and the other with reception regardless of sensation. Perception necessitates reception, but the converse is not true. Attention and motivation varied throughout the training and test sessions. Attention to the stimulus was occasionally tested by manually driving the stimulus intensity well above or below the coyotes' threshold and observing that the coyote's resultant behavior was to return it to threshold. The food deprivation levels and reinforcement schedules were occasionally altered, according to the coyotes' apparent motivation levels. One explanation for the plateau seen in Figure 1 between 20 and 35 min is that the subjects would occasionally depress the left treadle (the treadle from which they received food reinforcement) continuously, regardless of the stimulus intensity, thus driving the stimulus intensity to its maximum. This rapid increase in stimulus intensity may have caused a bleaching of the photosensitive pigment rhodopsin which was previously formed during the process of dark adaptation. With increased time spent at the higher stimulus intensity, a more complete degradation of rhodopsin to vitamin A would occur (Tansley, 1965). When the subject returned to normal responding, the slow thermal reaction of vitamin A combining with the protein opsin to form rhodopsin may have resulted in a gradual increase in sensitivity. Therefore, the subject had to become partially dark adapted once again. This, of course, would have then resulted in a more gradual attainment of a base level rather than a distinct drop (break) from a higher threshold to a lower one. This would account for the plateau seen in Figure 1. However, as mentioned earlier, when the average adaptation curve was plotted with footcandles in log units, this plateau became
SCOTOPIC SENSITIVITY IN COYOTE
a definite break, delineating both rod and cone functions. This curve is comparable with the one obtained during electroretinographic analysis (Figure 3). One plausible explanation, aside from actual individual differences in the coyotes' visual capabilities, for the difference in the slopes of the adaptation curves (Figure 1) between Coyote 2 and the others has to do with preexposure. Coyote 2 was housed in its individual cage directly underneath the other two subjects. Preexposure intensity was provided from overhead lighting located in the ceiling of the animal holding room. Coyotes 4 and 5 were housed directly under the lights at a distance of approximately 2 ft. (.6 m) at 50 ftc. (538 Ix) while Coyote 2 was housed at ground level under the other cages approximately 8 ft. (2.4 m) at 8 ftc. (86 Ix) from the light source. This difference in preexposure intensity may account for Coyote 2's faster attainment of threshold. In addition, another contributing factor may have been Coyote 2's higher response rate and proportionately fewer incorrect responses, as recorded during the training procedure. The alteration of amperage to the stimulus lamp resulted in a change in the spectral characteristics of the light emitted. A decrease in amperage to the lamp resulted in a corresponding decrease in color temperature, which resulted in a shift in radiant power toward longer wavelengths so that more energy was emitted as heat than as light (Graham, 1965). This shift was probably confounded by a Purkinje shift in maximum spectral sensitivity of the eye toward the short wavelength end of the visible spectrum in scotopic vision (Blough, 1957). The scotopic spectral sensitivity curve of the coyote was determined after dark adaptation, and as with humans, the peak of the curve lies at around 500 m/x in the blue-green (Figure 4C). Since the coyote has a duplex retina, its photopic spectral sensitivity curve probably lies to the right of the scotopic curve. However, even with less light energy being emitted at the shorter visible wavelengths, the coyotes wore able to respond to quite low intensities during scotopic adaptation.
1075
The ecological significance of the coyotes' scotopic visual threshold can be interpreted in at least two contexts. First, the visual threshold can be compared with the illumination present under various nocturnal conditions. Second, it can be compared with the thresholds of other crepuscular and nocturnal species occupying similar or different ecological niches. The amount of natural illumination present at night under various moon phases and cloud conditions was measured and calculated by Dice (1945). His calculation indicates to us that the coyote has the capability of detecting natural illumination (stellar and lunar) available under many nocturnal conditions. However, this should not be mistaken as evidence of the coyotes' ability to actually function under these low levels of illumination. "Whether these low levels of celestial illumination are sufficient to provide for adequate visual acuity and depth perception for various activities such as prey capture is not known. Osterholm (1964) demonstrated that the fox (Vulpes vulpes) relies in varying degrees on its different sensory receptors, depending on the amount of illumination available. It is likely that the coyote uses a similar strategy. Chesness (Note 1) reported that coyotes in natural habitats are primarily crepuscular and nocturnal during spring, summer, and fall, with sporadic activity during the daylight hours. In the winter most activity occurs during the night. This species may be presently adapting to a more diurnal or nocturnal niche. However, regardless of the coyotes' evolutionary path, it has either retained or obtained the ability to perceive low levels of illumination under many nocturnal conditions. This probably has definite survival value for a species that has been under such heavy pressure from man, a diurnal species. REFERENCE NOTE 1. Chesness, R. A. Home range and territorially of coyotes in North-Central Minnesota. Paper presented at the 34th Midwest Fish and Wildlife Conference, Des Moines, Iowa, 1972. REFERENCES Blough, D. S. Method for tracing dark adaptation in the pigeon. Science, 1955,121,703-704.
1076
STEVEN W. HORN AND PHILIP N. LEHNER
Blough, D. S. Dark adaptation in the pigeon. Journal of Comparative and Physiological Psychology, 1656, 49, 425-430. Blough, D. S. Spectral sensitivity in the pigeon. Journal of the Optical Society of America, 1957, 47, 827-833. Blough, D. S. A method for obtaining psychophysical thresholds from the pigeon. Journal of the Experimental Analysis of Behavior, 1958, 1, 31-43. Blough, D. S. Experiments in animal psychophysics. Scientific American, 1961, 206(1), 3-10. Dartnall, H. J. A. The interpretation of spectral sensitivity curves. British Medical Bulletin, 1953, 9, 1, 24-30. Dice, L. R. Minimum intensities of illumination under which owls can find dead prey by sight. American Naturalist, 1945, 79, 385-416. Dippner, R., & Armington, J. A behavioral measure of dark adaptation in the American red squirrel. Psychonomic Science, 1971, 24, 43-45. Graham, C. H. (Ed.). Vision and visual perception. New York: Wiley, 1965.
Hersloff, L., Lehner, P. N., Bolen, E. G., & Rylander, M. K. Visual sensitivity in the blackbellied tree duck (Dendrocygna autwnalis), a crepuscular species. Journal of Comparative and Physiological Psychology, 1974, 86, 3, 486-492. Osterholm, H. The significance of distance receptors in the feeding behavior of the fox Vulpes vulpes. Acta Zoologica Fennica 1964, 106, 1-31. Tansley, K. Vision in vertebrates. London: Chapman & Hall, 1965. Tomita, T., Kaneko, A., Murakami, M., & Pautler, E. L. Spectral response curves of single cones in the carp. Vision Research, 1967, 7, 519-531. von Be'ke'sy, G. V. A new audiometer. Acta OtoLaryngologica, 1947, 35, 411-422. Wells, M. C., Lehner, P. N., Bolen, E. G., & Rylander, M. K. Comparison of scotopic sensitivity in diurnal (Anas platyrhynchos) and crepuscular (Dendrocygna autumnalis) ducks. Journal of Comparative and Physiological Psychology, 1975, 88, 940-944. (Received December 31, 1974)
Scotopic Sensitivity in Coyotes (Cam's latrans) Steven W. Horn and Philip N. Lehner Laboratory of Animal Behavior, Colorado State University The absolute scotopic limen for light intensity was measured for three 10mo-old female coyotes. The methodology was similar to that used by Blough in determining psychophysical thresholds in pigeons. Three coyotes were operantly conditioned to depress one of two foot treadles, left or right, depending on the condition of the stimulus light. Scotopic adaptation curves for each coyote were generated. Nonlinear regression curves were then fitted to the raw data. The mean scotopic thresholds did not differ significantly. However, time to the curves' asymptotes did differ significantly for one of the coyotes. The adaptation curves showed a distinct rod-cone "break," and retinal histology confirmed that the coyote has a duplex retina with a preponderance of rods. In addition, electroretinographic analysis showed the relative contributions of rods and cones at various light intensities and indicated a rod-cone break at approximately 15 min. Scotopic spectral sensitivity curves were also generated. The coyotes' scotopic visual threshold is exceeded by the natural illumination available under many nocturnal conditions. The results are discussed in relation to the ecology of the species.
The purpose of this experiment was to determine scotopic visual sensitivity in the coyote, utilizing the psychophysical technique first described by Blough (1955, 1958, 1961). Blough's procedure was based on a technique developed by von Be'ke'sy (1947) for tracing the human auditory threshold. Blough's procedure was designed to study visual sensitivity in the pigeon (Columba lima). More recently this procedure has been used to study dark adaptation in the American red squirrel (Dippner & Armington, 1971), visual sensitivity in the black-bellied tree duck (Hersloff, Lehner, Bolen, & Rylander, 1974), and scotopic sensitivity in the mallard duck (Wells, Lehner, Bolen, & Rylander, 1975). In the experiments by Blough (1955, 1956, 1958), dark-adaptation curves obThis report is Contribution No. 5 of the Laboratory of Animal Behavior at Colorado State University. The authors are indebted to T. J. Boardman and P. Biondini for their assistance in the data analysis. Mary Fisher assisted in the retinal histology. The authors are also indebted to E. L. Pautler for his cooperation and assistance in the electroretinographic analysis as well as his review of the manuscript. Requests for reprints should be sent to Philip N. Lehner, Laboratory of Animal Behavior, Colorado State University, Port Collins, Colorado 80523.
tained from the pigeon suggested a duplex retina. A distinct rod-cone "break" in the curve appeared after 20 min, followed by a rapid increase in visual sensitivity reaching a maximum at about 50 min. Wells et al. (1975) showed similar results with the mallard and supported their visual adaptation curves with retinal histology. We measured the scotopic visual threshold in the coyote to show the adaptive significance of the visual system under study. To support our inferences we also examined the retinal structure of the coyote and conducted electroretinographic (ERG) analysis during the process of dark adaptation. Both the retinal histology and ERGs are correlated with the adaptation curves in the Discussion section. The coyotes' visual thresholds are discussed in terms of the ecology of the species in relation to the amount of natural illumination present under various nocturnal conditions. METHOD Subjects The experimental subjects were three 10-moold female coyotes obtained from dens in Colorado at approximately 7 days of age. The pups were hand reared in the laboratory and were housed in individual cages. Food deprivation schedules varied throughout the study, and the animals averaged 75%-85% of their free-feeding weights.
1070
SCOTOPIC SENSITIVITY IN COYOTE
1071
fore, the subject could get reinforcement one time in five by depressing the left treadle for a short interval immediately after eating. This ensured that the subjects attended to the stimulus and that the stimulus intensity remained within a narrow range oscillating around the subjects' visual thresholds. The stimulus intensity was regulated by the amount of amperage to the stimulus light. This of course resulted in a change in the lamp's spectral characteristics, which will be considered in the interpretation of the results (see Discussion). A Rustrak analog strip-chart recorder plotted the adaptation curves directly as a function of milliamperes of current to the stimulus lamp. Sample points (milliamperes) taken .every 37 sec from the strip charts were used for the data analysis. A correlation between milliamperes to the stimulus lamp and illuminance (footcandles) was determined by measurements with a photometer (Photovolt Corp.), and a logarithmic relationship was demonstrated. For comparison with other studies, including measurements of illumination from natural situations, the stimulus intensities are reported in both footcandles and milliamperes. In order to confirm the duplex nature of the Procedure coyote retina, histological examinations were perThe procedure used in this experiment was formed. In addition, the domestic dog's (Canis basically the same as the psychophysical tracking famttiaris) retina was also examined for comparaprocedure described by Blough (1958) and Hers- tive purposes. One female coyote, a litter mate of the experiloff et al. (1974), in which the stimulus intensity is under the control of the subject. Each subject mental subjects, was anesthetized while light was preexposed in its home cage under normal adapted under normal room illumination, and an room illumination provided from eight 116.8-cm, electroretinograna (ERG) was recorded during 400-W Cool White General Electric fluorescent the process of dark adaptation. The ERG was obbulbs. The subjects were taken directly from their tained by means of a cotton wick electrode posihome cages and placed in the test chamber, which tioned on the cornea. The electrode was connected was in total darkness. The only source of light to a Grass EEG preamplifier by means of a salt was from the 20-mm stimulus patch which was bridge. A needle scalp electrode served as the always preset at a constant intensity of 22 X 10~3 reference.2 The stimulus intensity provided by a ftc. (.2368 Ix) at the beginning of each session. Grass PS Photostimulator was raised in fractions Each depression on the right foot treadle auto- of log units until a 60-MV b-wave was recorded. matically reduced the stimulus intensity by ap- This was the operational definition of the ERG proximately 1 X 10~4 ftc. (.0011 lx)/sec of treadle threshold. This process continued at various indepression time, while left treadle depressions tervals over a period of 40 min, at which time the increased the stimulus intensity by the same process of dark adaptation was essentially comamount. The subjects depressed the right treadle plete. The b-wave amplitude of the ERG was deuntil the stimulus was no longer visible, at which termined in response to chromatic stimulation time they would then depress the left treadle until (equal energy) from 400-650 nm. This action specthe stimulus was again within their visible range. trum was analyzed in terms of the established inAs a result of this, the stimulus intensity oscil- tensity-amplitude relationship to provide a speclated above and below the subjects' visual thresh- tral sensitivity curve, according to the method of old for light intensity during the process of dark Tomita, Kaneko, Murakami, and Pautler (1967). adaptation. Depressions on the right treadle also turned RESULTS the stimulus light off on a modified variable inThe mean adaptation curves for each terval schedule of 22 sec (VI 22); depressions on the left treadle then provided reinforcement on a coyote are plotted in Figure 1, along with modified variable interval 10-sec (VI 10) schedule. the mean curve for all three subjects. The The VI schedule was modified such that it was vertical axis of the graph is presented both not in force unless the subject kept the foot treadle in the depressed condition. After reinforcement in milliamperes, since these were the units on the left treadle, the stimulus light remained used in the statistical tests, and in footoff on a variable ratio of 5 (VR 5) schedule. There- candles, since these units depict amounts of
Apparatus A modified Skinner box was used as the test chamber. The test chamber contained two foot treadles (40.6 cm apart) placed at one end of the chamber and separated by Plexiglas partitions. A 20-mm-cireular stimulus patch was centered above and between the treadles at the coyotes' eye level, and a food tray was centered directly below the stimulus patch. Three-gram Gaines Prime dog food pellets dispensed individually from a Davis Model No. 310 Universal Feeder were used as reinforcers. The treadles were connected through electrical relays to a Harvard cumulative recorder, a Rustrak analog recorder, and counters. The stimulus light source was a No. 1820-28VDC General Electric miniature bulb in a Lehigh Valley Electronics multistimulus rear-screen projector. The stimulus intensity and the various reinforcement contingencies were controlled and monitored from an adjacent room, using standard electromechanical modules and relays. A BRS Foringer audio generator provided masking noise and a 3-sec tone which accompanied the presentation of .reinforcement.
1072
STEVEN W. HORN AND PHILIP N. LEHNER 218 120
65
COYOTE # 2
60
68
COYOTE* 4 COYOTE* 5 •—/VERAGE
55 50
38 Q x 22
45
12
40
7
35
4
30
2
1.2 5
10
15
20
25
30
35
40
45
50
55
60
TIME (minutes) FIGURE 1. Individual and mean scotopic sensitivity curves.
illumination irradiating from the stimulus patch. In order to predict both a base level amperage (i.e., the subject's dark-adapted light-intensity thresholds) and the time at which the base level was reached (the point at which the process of dark adaptation was completed), a single exponential nonlinear regression of the form y = a + f}eri was generated for use as a model of a curve of best fit. The variables were denned as follows: a = the base level amperage; 0 = the difference between the starting amperage and the base level; ? = the slope of the curve; t = time. A one-way analysis of variance was performed on the differences between the adaptation curves for all three coyotes. No reliable differences emerged between the curves' base level intensities, F(2, 6) 1.39, p > .31. However, there was a reliable difference between the curves' slopes, F(2, 6) 6.13, p < .04. This difference was the result of one coyote which consistently reached base level amperage (threshold intensity) faster than the other two subjects (Figure 1). The base levels (thresholds) for all three subjects had a predicted grand mean of 26 mA or 1.4 X 10~3 ftc. (.0151 Ix). Figure 2 depicts the predicted average 'dark-adaptation curves for each subject and for all three coyotes together, which may be
more representative of the species as a whole. The curves are very similar with the exception that Coyote 2 reached base level amperage within 20-30 min, while the other subjects were consistent in reaching threshold in 30-40 min. The coyotes' mean adaptation curve as plotted in Figure 1 does not show a definite break during the process of dark adaptation, but a partial plateau is evident between 20 and 35 min. However, when the coyotes' mean adaptation curve is plotted with stimulus intensity in log units (Figure 3) a definite break was found at 15 min. Blough (1955) interpreted the break as a reflection of the initial adaptation of the cones and subsequent adaptation of the rods. Wells et al. (1975) also found a break for the mallard at about 25 min. When electroretinographic analysis was performed, the break described in the preceding paragraph was clearly evident at approximately 15 min after the start of dark adaptation (Figure 3). Figure 4 presents the scotopic spectral response and sensitivity curves obtained for the coyote. The spectral response curve (Figure 4A) should be distinguished from the spectral sensitivity curve (Figure 4C), which is a plot of the reciprocal of the spectral light intensity at each wavelength just
SCOTOPIC SENSITIVITY IN COYOTE
1073
COYOTE NO. 2
COYOTE NO. 4 COYOTE NO. 5 OVERALL AVERAGE
25.
375
25.0
TIME (MINUTES)
FIGURE 2. Predicted average dark adaptation curves.
sufficient for eliciting an equal amplitude response. Conversion from the spectral response curve to the spectral sensitivity curve is necessary for the comparison of data from different sources. The procedure (based on Tomita et al., 1967) was as follows: Upon recording the spectral response at a given wavelength, the intensity-amplitude relation was obtained utilizing white light; for example, a response amplitude equal to the peak value (40 at 500 m/i) in Figure 4A is produced by white light of
— 1.5 log-unit intensity (Figure 4B). Alsor the response amplitude (10) at 625 m/j is reproducible with white light of —3.5 logso i 40-
]30-
S20-
• PSYCHOPHYSICAL (BthovicrilO EPS
1.5
400
450
500
55O
600
WAVELENGTH (nyi)
0.5
15
20
25
TIME Mnuto)
FIGUBE 3. Dark adaptation curves obtained from electroretinagraphic (ERG) and psychophysical procedures.
FIGURE. 4. Data utilized in deriving the spectral sensitivity curves from the spectral response curve: A, spectral response curve; B, the intensityamplitude relation of the unit shown in A to white light; C, spectral sensitivity curve derived from A and B (solid line) along with Dartnall's nomogram (1953) with the same peaking wavelength, 500 HIM (broken line).
1074
STEVEN W. HORN AND PHILIP N. LEHNER
unit intensity. This means that a two logunit (20-fold) increase in light intensity is necessary for the 625-rmt light to produce a response equal to the peak value (40 at 500 m/i) or, conversely, that the sensitivity at 625 nui is one twentieth of that at 500 nut. The sensitivity to other wavelengths was determined likewise and plotted as a solid line in Figure 4C. For comparative purposes the broken line illustrates Dartnall's nomogram (1953) with the same peaking wavelength of 500 m/j. At this time no quantitative data have been collected from the histological examinations made on the coyote retina. The coyote definitely has a duplex retina (both photoreceptor cells present) and appears to have a preponderance of rods. Examinations of sections taken from the domestic dog showed no great differences. DISCUSSION Because of the coyote's nocturnal-crepuscular behavior, one might hypothesize that it would require a duplex retina with rod dominance (Chesness, Note 1). This hypothesis is only weakly supported by the original adaptation curves (Figure 1) obtained during psychophysical testing. The curves in Figure 1 show no pronounced break that would clearly delineate predominantly rod and cone functions in the retina. However, there appears to be a plateau between 20 and 35 mm and a subsequent gradual attainment to a lower base level, which may reflect the duplex retina. When an electroretinogram was made, this apparent rod-cone break was most evident at about 15 min into the process of dark adaptation. This apparent contradiction was corrected when the curves were plotted with footcandles converted to log units (Figure 3). Probably the single most important factor involved in the differences between the curves obtained psychophysically (Figure 1) and the one obtained during electroretinographic analysis has to do with what was actually being measured. During the psychophysical testing procedure, the subjects were required to make a value judgment as to whether the stimulus evoked a visual
sensation. When we made an electroretinogram on an anesthetized animal, we were not asking for the perception of a sensory response but were simply measuring the receptory capabilities of the coyote retina through the change in the corneal potential. Therefore, we were measuring two completely different processes: one dealing with perception giving rise to a sensation and the other with reception regardless of sensation. Perception necessitates reception, but the converse is not true. Attention and motivation varied throughout the training and test sessions. Attention to the stimulus was occasionally tested by manually driving the stimulus intensity well above or below the coyotes' threshold and observing that the coyote's resultant behavior was to return it to threshold. The food deprivation levels and reinforcement schedules were occasionally altered, according to the coyotes' apparent motivation levels. One explanation for the plateau seen in Figure 1 between 20 and 35 min is that the subjects would occasionally depress the left treadle (the treadle from which they received food reinforcement) continuously, regardless of the stimulus intensity, thus driving the stimulus intensity to its maximum. This rapid increase in stimulus intensity may have caused a bleaching of the photosensitive pigment rhodopsin which was previously formed during the process of dark adaptation. With increased time spent at the higher stimulus intensity, a more complete degradation of rhodopsin to vitamin A would occur (Tansley, 1965). When the subject returned to normal responding, the slow thermal reaction of vitamin A combining with the protein opsin to form rhodopsin may have resulted in a gradual increase in sensitivity. Therefore, the subject had to become partially dark adapted once again. This, of course, would have then resulted in a more gradual attainment of a base level rather than a distinct drop (break) from a higher threshold to a lower one. This would account for the plateau seen in Figure 1. However, as mentioned earlier, when the average adaptation curve was plotted with footcandles in log units, this plateau became
SCOTOPIC SENSITIVITY IN COYOTE
a definite break, delineating both rod and cone functions. This curve is comparable with the one obtained during electroretinographic analysis (Figure 3). One plausible explanation, aside from actual individual differences in the coyotes' visual capabilities, for the difference in the slopes of the adaptation curves (Figure 1) between Coyote 2 and the others has to do with preexposure. Coyote 2 was housed in its individual cage directly underneath the other two subjects. Preexposure intensity was provided from overhead lighting located in the ceiling of the animal holding room. Coyotes 4 and 5 were housed directly under the lights at a distance of approximately 2 ft. (.6 m) at 50 ftc. (538 Ix) while Coyote 2 was housed at ground level under the other cages approximately 8 ft. (2.4 m) at 8 ftc. (86 Ix) from the light source. This difference in preexposure intensity may account for Coyote 2's faster attainment of threshold. In addition, another contributing factor may have been Coyote 2's higher response rate and proportionately fewer incorrect responses, as recorded during the training procedure. The alteration of amperage to the stimulus lamp resulted in a change in the spectral characteristics of the light emitted. A decrease in amperage to the lamp resulted in a corresponding decrease in color temperature, which resulted in a shift in radiant power toward longer wavelengths so that more energy was emitted as heat than as light (Graham, 1965). This shift was probably confounded by a Purkinje shift in maximum spectral sensitivity of the eye toward the short wavelength end of the visible spectrum in scotopic vision (Blough, 1957). The scotopic spectral sensitivity curve of the coyote was determined after dark adaptation, and as with humans, the peak of the curve lies at around 500 m/x in the blue-green (Figure 4C). Since the coyote has a duplex retina, its photopic spectral sensitivity curve probably lies to the right of the scotopic curve. However, even with less light energy being emitted at the shorter visible wavelengths, the coyotes wore able to respond to quite low intensities during scotopic adaptation.
1075
The ecological significance of the coyotes' scotopic visual threshold can be interpreted in at least two contexts. First, the visual threshold can be compared with the illumination present under various nocturnal conditions. Second, it can be compared with the thresholds of other crepuscular and nocturnal species occupying similar or different ecological niches. The amount of natural illumination present at night under various moon phases and cloud conditions was measured and calculated by Dice (1945). His calculation indicates to us that the coyote has the capability of detecting natural illumination (stellar and lunar) available under many nocturnal conditions. However, this should not be mistaken as evidence of the coyotes' ability to actually function under these low levels of illumination. "Whether these low levels of celestial illumination are sufficient to provide for adequate visual acuity and depth perception for various activities such as prey capture is not known. Osterholm (1964) demonstrated that the fox (Vulpes vulpes) relies in varying degrees on its different sensory receptors, depending on the amount of illumination available. It is likely that the coyote uses a similar strategy. Chesness (Note 1) reported that coyotes in natural habitats are primarily crepuscular and nocturnal during spring, summer, and fall, with sporadic activity during the daylight hours. In the winter most activity occurs during the night. This species may be presently adapting to a more diurnal or nocturnal niche. However, regardless of the coyotes' evolutionary path, it has either retained or obtained the ability to perceive low levels of illumination under many nocturnal conditions. This probably has definite survival value for a species that has been under such heavy pressure from man, a diurnal species. REFERENCE NOTE 1. Chesness, R. A. Home range and territorially of coyotes in North-Central Minnesota. Paper presented at the 34th Midwest Fish and Wildlife Conference, Des Moines, Iowa, 1972. REFERENCES Blough, D. S. Method for tracing dark adaptation in the pigeon. Science, 1955,121,703-704.
1076
STEVEN W. HORN AND PHILIP N. LEHNER
Blough, D. S. Dark adaptation in the pigeon. Journal of Comparative and Physiological Psychology, 1656, 49, 425-430. Blough, D. S. Spectral sensitivity in the pigeon. Journal of the Optical Society of America, 1957, 47, 827-833. Blough, D. S. A method for obtaining psychophysical thresholds from the pigeon. Journal of the Experimental Analysis of Behavior, 1958, 1, 31-43. Blough, D. S. Experiments in animal psychophysics. Scientific American, 1961, 206(1), 3-10. Dartnall, H. J. A. The interpretation of spectral sensitivity curves. British Medical Bulletin, 1953, 9, 1, 24-30. Dice, L. R. Minimum intensities of illumination under which owls can find dead prey by sight. American Naturalist, 1945, 79, 385-416. Dippner, R., & Armington, J. A behavioral measure of dark adaptation in the American red squirrel. Psychonomic Science, 1971, 24, 43-45. Graham, C. H. (Ed.). Vision and visual perception. New York: Wiley, 1965.
Hersloff, L., Lehner, P. N., Bolen, E. G., & Rylander, M. K. Visual sensitivity in the blackbellied tree duck (Dendrocygna autwnalis), a crepuscular species. Journal of Comparative and Physiological Psychology, 1974, 86, 3, 486-492. Osterholm, H. The significance of distance receptors in the feeding behavior of the fox Vulpes vulpes. Acta Zoologica Fennica 1964, 106, 1-31. Tansley, K. Vision in vertebrates. London: Chapman & Hall, 1965. Tomita, T., Kaneko, A., Murakami, M., & Pautler, E. L. Spectral response curves of single cones in the carp. Vision Research, 1967, 7, 519-531. von Be'ke'sy, G. V. A new audiometer. Acta OtoLaryngologica, 1947, 35, 411-422. Wells, M. C., Lehner, P. N., Bolen, E. G., & Rylander, M. K. Comparison of scotopic sensitivity in diurnal (Anas platyrhynchos) and crepuscular (Dendrocygna autumnalis) ducks. Journal of Comparative and Physiological Psychology, 1975, 88, 940-944. (Received December 31, 1974)
