THE EFFECT OF ROD ACTIVITY ON COLOUR MATCHING FUNCTIONS’ ULF STABELLand BJP)RN STAE~ELL Institute of Psychology, University of Osto, Norway (Received 3 lMay 1974;
in
revised
form14 November
1974)
Abstract-When an achromatic scotopic stimulus was superimposed upon a red test light in a dark-adapted eye, the hue of the test light changed toward yellow. Moreover. when the chromaticities of a test beam were measured twice during dark adaptation: (if at the moment when the cone threshold and (2) when the sengitivity of the rods were assumed to have reached their absolute dark-adapted values, colours of orange and green-yellow were found to change toward yellow, while colours of blue-green and violet were found to change toward blue. The results appear to oppose the assumption of the Duplicity Theory of vision that the rod activity contributes a constant achromatic colour quality to the visual sensation, but are consistent with the suggestion that the chromatic response of rods may change as a function of selective chromatic stimulation of cones.
Putkinje ft825) reported that pigment colours appear black or grey at dawn. Aubert (1865) accomplished
Second, spectral sensitivity curves of the scotopic contrast hues-recorded at threshold and suprathreshold levels-show a close correspondence to the scotopic visibility curve. Third, the scotopic contrast hues are independent of the wavelength used in the test. Hence, it has been concluded that the scotopic contrast hues are triggered upon test-stimulation by light signais initiated in the rods. Furthermore, since the scotopic contrast hue has been shown to be approx~tely opponent to the hue of the pre-stimulation, it has been suggested that the scotopic hue basically reflects the ratio of the primary hue-related processes of the pre-stimulation (Stabell and Stabell, 1971b). The colour quality initiated by the rods upon test-stimulation at scotopic intensity levels has thus been assumed to be controlied by chromatic pre-st~ulation of cones (Stabell and Stabell, 1973a). Similarly, when the intensity of the test-stimulation increases above the absolute cone threshold, the cone impulses initiated by the test light have been assumed to change the disposition for chromatic rod activity (Stabell and Stabell, 1973b, 1974). Accordingly, it might be suggested that the- colour quality initiated by the rods at mesopic intensities depends on the wavelength of the test-stimulation.
the first thorough experimental study of this phenomenon. He noted that the pigment colours, after dark adaptation and at low levels of illumination, appear colourless, differing only in brightness. Chodin (1877) made observations both with pigment and spectral colours and confirmed the results of Aubert. Later, the observation of an achromatic intensity interval has been repeatedly demonstrated (Cohn, 1882; Hillebrand, 1890; Tschermak, 1920; Kohlrausch, 1931; Hecht, Haig and Chase, i937; Mandelbaum, 1941; Lie, 1963). The achromatic or so-called photocbromatic interval has generally been explained on the basis of the Duplicity Theory of vision, which assumes that (a) rods alone are activated by light in night vision, and (b) rods mediate only achromatic vision (Schultze, 1866). Hence, the colourless appearance of the spectrum under scotopic conditions is ascribed to the function of the rods. In opposition to the Duplicity Theory, however, night vision appears achromatic only under conditions where the eye is in a chromatically neutral state of adaptation (Stabell and Stabell, 1973a). Subsequent to selective chromatic pre-stimulation, hue may be observed upon test-stimulation at intensities throughout the phot~~o~tic interval. Hence, if rods alone are activated by light in night vision, rods may METHODS mediate chromatic sensations. There appears to be To investigate the effect of the rod intrusion at mesopic well-founded evidence in f&our of the latter assumpintensity levels, an achromatic scotopic siimulus was supertion (cf. &bell and Stabell, 1975). imposed upon a test beam above cone threshold. The First, subsequent to selective chromatic. prechromaticities of the mixture as well as the chromaticities stimulation, scotopic contrast hues may be observed of each of the two components were determined. Between upon t~t-st~ulation at an intensity level of about 3 difFerent sessions the wavelength of the test light was log units below the absolute threshold of dark-adapted varied, while the wavelength of the scotopic stimulus was cones, estimated from the absolute threshold of the held constant at 510 nm. The achromatic scotopic stimulus did not a&t the c~o~ticities of the test beam at dark-adapted fovea and the cone pkteau of the longmedium and short wavelengths, however, pr&umably term dark-adaptation curve. because such wavelengths activate the rod mechanism to
’ The Norwegian Research Council for Science and the Humanities supported this study financially.
a considerable extent. To investigate the effect of rod intrusion at the blue, green and yellow regions of the spectrum, therefore, a
1119
supplementary method was employed. taking advantage of the generally accepted assumption that the cones adapt to darkness at a more rapid rate than the rods (Rushton. 1957). On this basis, the chromaticities of a test beam may be measured twice during dark adaptation: (1) when the sensitivity of the cones; and (2) when the sensitivity of the rods have reached their absolute dark adapted thresholds. A chromaticity shift might then be ascribed to the change in the sensitivity of the rods (Lie. 1963). Apparatus
and general
procedure
The Wright calorimeter and calibration procedures used have been described in detail by Wright (1946). The colorimeter was arranged with three reflecting prisms in the W? spectrum, giving the instrumental primaries 650, 530 and 460nm. The unit matches were made symmetrically at the fovea on wavelengths 494 and 5825 nm. In the following experiments the comparison and test fields were applied at the fovea and 6’ temporally to the fovea of the right eye, respectively, each subtending 2’ x I”. The test and comparison fields were presented in succession using 0.5~set flashes. The authors served alternately as subjects. Experiment
I (scotopic stimulus superimposed)
The successive phases of the operation were as follows: (I) After 30 min dark adaptation, the absolute thresholds to the lights of 510 and 65Onm. applied 6” temporally to the fovea, were determined in succession. The absolute thresholds were taken as the lowest intensity at which the subject was certain of seeing light. The intensity was increased in small steps from about @5 log unit below the expected thresholds. using 0.5.see flashes. (2) IO mill dark adaptation. l.3) Stimulation for 0.5 see with the e\trafo\cal test hcam of 650nm at an intensity 1.3 log units above the absolute threshold as measured in phase 1, representing an intensity level of about 2 photopic td. (4) @5 set dark adaptation. (5) Stimulation for 05 set with the fovea1 comparison field consisting of the instrumental primaries. The subject tried to establish a match by increasing or reducing the intensity of each of the three primaries. (6) Repetitions of phases 2-5 until a reasonably good match existed between the test and comparison fields. Occasionally the extrafoveal stimulus appeared too saturated to be matched by any fovea1 patch within the colour gamut of the RGB system. In such cases the test field could be matched by a monochromatic wavelength observed at the fovea. When the chromaticities of this wavelength had been established the chromaticities of the wavelength observed extrafoveally could be plotted in the fovea1 WDW system. (7) Phases 2-6 were repeated, except that in phase 3 only the scotopic stimulus (the 510~nm light) illuminated the test field. The intensity was 1.5 log units above the absolute threshold as measured in phase 1, representing an intensity level of about 1 log unit below the absolute threshold of dark-adapted cones, estimated, for the given stimulus, from the cone plateau of the long-term darkadaptation curve and from the absolute threshold of the dark-adapted fovea. (8) Phases 2-6 were repeated, except that in phase 3 the scotopic stimulus of 510nm was superimposed upon the extrafoveal test stimulus of 650nm. (9) Phases 1-8 were repeated, except that the .test field was illuminated with 600nm instead of 65Onm.
Red
Fig. I. Mean chromaticities within the fovea1 WDW system of an achromatic scotopic stimulus C,, a photopic stimulus C?, and their mixture C3. The wavelength of the photopic stimulus was 65Onm (0) and 6OQnm (0). The data points in Figs. 1-6 give the mean of four observations of subject US. being the same in all essentials for subject BS.
beam), the photopic stimulus C2 (the GO- or 600-nm test beams), and the mixture Cs. It will be seen that the chromaticities of Cs do not lie on or close to
the straight mixture locus C,C2 predicted by chromatic additivity, but are displaced toward yellow. DISCLSSION
The results clearly demonstrate that an achromatic scotopic stimulus, assumed to be well below the absolute threshold level of dark-adapted cones, may influence the sensation of hue observed at mesopic intensities. Thus, when the light of the 510-nm stimulus was superimposed upon the red test lights, the hue changed toward yellow. If the scotopic stimulus represents pure rod activity, it would be expected from the Principle of Univariance that the change in hue toward yellow would be independent of the wavelength of the scotopic light. In order to test the suggestion that the scotopic stimulus represents rod activity, Experiment 1 was repeated. except that the wavelength of the scotopic stimulus was varied between 420 and 550nm. In support of the suggestion, the chromaticity shift was independent of the wavelength used. Experiment
la,b
(dark adaptation)
Comparing a fovea1 and an extrafoveal field at a moment immediately before the cone-rod break of the long-term dark-adaptation curve, Lie (1963) found that the two fields could lx closely matched both with respect to saturation and brightness. However, after 35 min of dark adaptation the extrafoveal field appeared more bright and less saturated at various intensity levels above the specific threshold. As the intensity of the two fields in-
creased, the brightness and saturation differences gradually decreased until the two fields appeared identical at about 2 log units above the absolute threshold of dark-adapted
RESLZTS
The results of Experiment 1 are given in Fig. 1. The data points represent the chromaticities of the achromatic scotopic stimulus C, (the SlO-nm test
cones. Presupposing that the “potential cone threshold.” i.e. the pure cone threshold, remains at the cone plateau during the second step of the dark-adaptation curve, the change in saturation and brightness observed between 5 and 35 min of dark adaptation was ascribed to the increased sensitivity of the rods.
Rod
activity
effect
on
colour
1121
matching
Experiment 2 may be regarded as an extension of the work of Lie (1963). The chromaticities of a test beam were measured twice during dark adaptation: (I) at the moment when the potential cone threshotd; and (2) when the sensitivity of the rods were assumed to have reached their absolute dark-adapted values. Experiment
21 (determination
of potenrial
cone threshold)
The cone threshold without rod activity involved may be measured during dark adaptation within the rod-free fovea, but also in the periphery provided relatively long wavelengths are used in the test. The successive phases
of the operation were as follows. (1) tight o~prar~on. Subsequent to 10min of preliminary dark adaptation, the light-adapting system was brought into use. The size of the circular adapting field was eight degrees, centered 6’ temporally to the fovea of the right eye. The subject was light adapted for 3 min to a constant brightness of about 30,000 photopic td. The adapting white light was then removed. (2) Threshold measurements. With the subject in complete darkness. absolute thresholds were measured every minute. The size of the calorimeter field was 2” x I”, applied 6” temporally to the fovea. using OSsec flashes. Between different runs the wavelength of the calorimeter field was varied between 650 and 550nm. (3) Phases 1 and 2 were repeated with the differences that (a) the light adaptation and test field were centered at the fovea, and (b) the wavelength of the test field was varied between 650 and 460 nm.
RESULTS
The results of Experiment 2a are shown in Figs. 2 and 3. The dark-adaptation curves indicate that the final values of the potential cone thresholds and the threshold of the rods, under the given conditions of ex~rimen~tion, are reached respectively, after about 6 and 30 min of dark adaptation. Consequently, in Experiment 2b the chromaticities of the test beam were measured between 6 and 7min and between 30 and 31 min of dark adaptation. 4.0 li;
3.6
2.4 fa 2 0 E
2.0
H
I.2
I.6
2
4
6 6 10 Time in darkness.
12 min
14
I6
Fig. 3. Dark-adaptation curves of the wavelengths 6.50, 600, 575. 550. S30. 505, 485 and 46Onm as determined at the fovea. The size and duration of the test-stimulation were as in Fig. 2. The curves were made to coincide at 5 min of dark adaptation.
Experiment
2b (determination
of the chromaticity
shift)
The successive phases of the operation were as follows.
(I) Light adaptation.
This was a repetition of phase
I of Experiment ?a. (2) 6 min dark adaptation. (3) Test-stimulation for 05 set with the extrafoveal test beam of 6OOnm. The intensity was 1 photopic td, representing an intensity of about 1 log unit above the cone plateau as measured in Experiment 2a. (4) @5 set dark adaptation. (5) Stimulation for 05 set with the fovea1 comparison field, consisting of the instrumental primaries. The subject tried to establish a match by increasing or reducing the intensity of each of the three primaries. Only small adjustments were required since pre-experimentation had established an approximate match. (6) Repetitions of phases 3-5 until the subject had established the match. (7) 23 min dark adaptation. (8) Repetition of phases 3-6. (9) Phases l-8 were repeated, except that the intensity of the extrafoveal test beam was 10 and 100 photopic td. (10) Phases l-9 were repeated, except that the wavelength of the extrafoveal test field was: 460. 485. 505. 530, 550 and 575 nm. Occasionally, the extrafovea stimulus appeared too saturated to be matched by any fovea1 patch within the colour gamut of the RGB system. In such cases a reasonably good match could be obtained between the test stimulus and a monochromatic wavelength observed at the fovea. The chromaticities of the wavelength observed at the fovea could then be established. RESLZTS
The results of Experiment 2b are presented in Figs.
2
6
IO lime
14
in darkness
18
,
22 min
26
30
Fig. 2. Dark-adaptation curves of the wavelengths 654 600,575 and 550 nm recorded at 6” temporally to the fovea of the right eye. The duration of the test flash was O+ set and the size 2” x 1”. The dark-adaptation curves were arbitrarily made to coincide at 5min of dark adaptation.
4-6. In Fig. 4 the mean chromaticities of monochromatic wavelengths at an intensity of 1 photopic td observed at about 6 min (filled symbols) and 30 min (plain symbols) of dark adaptation, are given together with the mean chromaticities of the achromatic scotopie stimulus C,. In agreement with the finding of Lie (1963), the results show that the saturation of the different colours was reduced between 6 and 30min of dark adaptation. Furthermore, colours of 6001~~ (orange) and 55Onm (g~e~~llow) changed
I’LF STABELL and
BJCIR~ Sr.kalrLL
600
GO&_
e
Red
Red
Fig. 4. Chromaticity shifts of monochromatic wavelengths between 6 (tilled symbols) and 30min {plain symbols) of dark adaptation. The intensity of the wavelengths was held constant at about 1 photopic td.
markedly toward yellow, while 460nm (violet) and 485 nm (blue-green) changed toward blue. As the intensity of the test field increased, the change in saturation decreased (Figs. 5-6). At the highest intensity level, no differences in hue or saturation were observed for 600 and 575 nm.
Fig. 6. Same as Fig. 4. except that the intensity of the test field was 100 photopic td. The chromaticities of the wavelengths 460 and 485 nm could not be determined due to the low energy output provided by the apparatus at the short wavelengths.
chromatic stimulation of cones, and appears to oppose the assumption of the Duplicity Theory of vision that the rod response is of constant colour quality and contributes an achromatic component to the visual sensation.
DISCUSSlOX
On the assumption that the potential cone threshold, under the present conditions of experimentation, remained unchanged after about 6 min of dark a&p~tion, the change in hue and saturation observed in Experiment 2b might be ascribed to change in rod activity. In conformity with this suggestion, the colours were found to remain invariant between 6 and 30min of dark adaptation when, as a control, the light adaptation and test-stimulation were centered at the right eye fovea, while the comparison field was centered at the dark-adapted feft eye fovea. The evidence presented in Experiments 1 and 2b is consistent with the suggestion that the chromatic response of rods may change as a function of selective
REFERENCES Aubert H. (1865) Physiologic der Netzhaut. Morgenstern Breslau. Chodin A. (1877) fiber die Abhan~gkeit der Farbenempfindungen von der Lichtsdrke. In Sammlung Pfr~siolog&her A.bhandlungen (Edited by Preyer W.). pp. 375444. D&t, Jena. Cohn H. (1882) Uber Far~nemp~ndungen bei schwicher kilnstlicher Beleuchtung. Arch. Augenheiik. 11, 283-302. Hecht S., Haig C. and Chase A. M. (1937) The influence of light adaptation on subsequent dark adaptation of the eye. J. gen. Phy~~~~.20, 831-850. Wiliebrand F. (1890) Uber die specifisohe Heiligkeit der Farben. Beitrlge zur Psychologie der Gesichtsempfindungen. Sber. Akad. Wiss. Mien Ma&.-natttrwiss. KI. Abt. 98 f3f, 70-120.
Kohlrausch
A. (1931) Adaptation, Tagessehen und Dlmmerunassehen. In Hand&h der Norma/en und ~~r~ologi~~~~ Pb~s~o~offie [Edited by Bethe A., Berg mann G.. Embden G. and Elfinger A.), pp. 1499-159-t. Springer, Berlin. Lie I. (1963) Dark adaptation and the photochromatic interval ~oc~rnen~a ophth. 17, 41 l-510. Mandelbaum J. (1941) Dark adaptation. Archs Ophrhal. X.X 26, 203-239.
600
Red
Fig. 5. Same as Fig. 4, except that the intensity of the test field was 10 photopic td.
Purkinie J. (1825)Beobachrungen und Versuche :ur Physiologi; der Sinne. Neue Beitriige :ur Kenntniss des Sehens in Subiecticer Hinsichr. Bd. 2. Reimer, Berlin. Rushton< W. A. H. (1957) Physical measurement of cone pigment in the living human eye. .Vattrr~, Lo& 179. 571-573. Shultze M. (1866) Zur Anatomie und Physiologie der Retina. Arch. mikrosk. Anal. 2, 176-256. StabeIl B. and Stabell U. (1971a) Chromatic rod vision-l: Wavelength of test-stimulation varied. Stand. J. Psychoi. 12. 175-178. Stabell B. and Stabell U. (1973a) Chromatic rod vision-IX: A theoretical survey. Vision Rrs. X3,449-455. Stabeli 3. and Stabetl U. (1974) Chromatic rod-cone interaction. tision Rex 14, 1389-1392.
Rod activity effect on colour matching Stabeli U. and Stabell B. (1971b) Chromatic rod vision-II: Wavelength of pre-stimulation varied. SC&. J. Psychol. 12, 282-288. Stabell U. and StabelI B. (1973b) Chromatic rod activity at mesopic intensities. Msion Res. 13, 2255-2260. Stab4 U. and Stabell B. (1975) Scotopic contrast hues triggered by rod activity. Vision Res. 15 1115-l 118.
1123
Tschermak A. (1929) Licht- und Farbensinn. In Handbuch der Normulen und Pathologixhen Physiologic (edited by Bethe A., Bergmann G., Embden G. and Ellinger A.), Bd. 12 (1X pp. 679-713. Springer, Berlin. Wright W. D. (1946) Researches on Normal and Defictiw Colour &ion. Henry Kimpton, London.
in
revised
form14 November
1974)
Abstract-When an achromatic scotopic stimulus was superimposed upon a red test light in a dark-adapted eye, the hue of the test light changed toward yellow. Moreover. when the chromaticities of a test beam were measured twice during dark adaptation: (if at the moment when the cone threshold and (2) when the sengitivity of the rods were assumed to have reached their absolute dark-adapted values, colours of orange and green-yellow were found to change toward yellow, while colours of blue-green and violet were found to change toward blue. The results appear to oppose the assumption of the Duplicity Theory of vision that the rod activity contributes a constant achromatic colour quality to the visual sensation, but are consistent with the suggestion that the chromatic response of rods may change as a function of selective chromatic stimulation of cones.
Putkinje ft825) reported that pigment colours appear black or grey at dawn. Aubert (1865) accomplished
Second, spectral sensitivity curves of the scotopic contrast hues-recorded at threshold and suprathreshold levels-show a close correspondence to the scotopic visibility curve. Third, the scotopic contrast hues are independent of the wavelength used in the test. Hence, it has been concluded that the scotopic contrast hues are triggered upon test-stimulation by light signais initiated in the rods. Furthermore, since the scotopic contrast hue has been shown to be approx~tely opponent to the hue of the pre-stimulation, it has been suggested that the scotopic hue basically reflects the ratio of the primary hue-related processes of the pre-stimulation (Stabell and Stabell, 1971b). The colour quality initiated by the rods upon test-stimulation at scotopic intensity levels has thus been assumed to be controlied by chromatic pre-st~ulation of cones (Stabell and Stabell, 1973a). Similarly, when the intensity of the test-stimulation increases above the absolute cone threshold, the cone impulses initiated by the test light have been assumed to change the disposition for chromatic rod activity (Stabell and Stabell, 1973b, 1974). Accordingly, it might be suggested that the- colour quality initiated by the rods at mesopic intensities depends on the wavelength of the test-stimulation.
the first thorough experimental study of this phenomenon. He noted that the pigment colours, after dark adaptation and at low levels of illumination, appear colourless, differing only in brightness. Chodin (1877) made observations both with pigment and spectral colours and confirmed the results of Aubert. Later, the observation of an achromatic intensity interval has been repeatedly demonstrated (Cohn, 1882; Hillebrand, 1890; Tschermak, 1920; Kohlrausch, 1931; Hecht, Haig and Chase, i937; Mandelbaum, 1941; Lie, 1963). The achromatic or so-called photocbromatic interval has generally been explained on the basis of the Duplicity Theory of vision, which assumes that (a) rods alone are activated by light in night vision, and (b) rods mediate only achromatic vision (Schultze, 1866). Hence, the colourless appearance of the spectrum under scotopic conditions is ascribed to the function of the rods. In opposition to the Duplicity Theory, however, night vision appears achromatic only under conditions where the eye is in a chromatically neutral state of adaptation (Stabell and Stabell, 1973a). Subsequent to selective chromatic pre-stimulation, hue may be observed upon test-stimulation at intensities throughout the phot~~o~tic interval. Hence, if rods alone are activated by light in night vision, rods may METHODS mediate chromatic sensations. There appears to be To investigate the effect of the rod intrusion at mesopic well-founded evidence in f&our of the latter assumpintensity levels, an achromatic scotopic siimulus was supertion (cf. &bell and Stabell, 1975). imposed upon a test beam above cone threshold. The First, subsequent to selective chromatic. prechromaticities of the mixture as well as the chromaticities stimulation, scotopic contrast hues may be observed of each of the two components were determined. Between upon t~t-st~ulation at an intensity level of about 3 difFerent sessions the wavelength of the test light was log units below the absolute threshold of dark-adapted varied, while the wavelength of the scotopic stimulus was cones, estimated from the absolute threshold of the held constant at 510 nm. The achromatic scotopic stimulus did not a&t the c~o~ticities of the test beam at dark-adapted fovea and the cone pkteau of the longmedium and short wavelengths, however, pr&umably term dark-adaptation curve. because such wavelengths activate the rod mechanism to
’ The Norwegian Research Council for Science and the Humanities supported this study financially.
a considerable extent. To investigate the effect of rod intrusion at the blue, green and yellow regions of the spectrum, therefore, a
1119
supplementary method was employed. taking advantage of the generally accepted assumption that the cones adapt to darkness at a more rapid rate than the rods (Rushton. 1957). On this basis, the chromaticities of a test beam may be measured twice during dark adaptation: (1) when the sensitivity of the cones; and (2) when the sensitivity of the rods have reached their absolute dark adapted thresholds. A chromaticity shift might then be ascribed to the change in the sensitivity of the rods (Lie. 1963). Apparatus
and general
procedure
The Wright calorimeter and calibration procedures used have been described in detail by Wright (1946). The colorimeter was arranged with three reflecting prisms in the W? spectrum, giving the instrumental primaries 650, 530 and 460nm. The unit matches were made symmetrically at the fovea on wavelengths 494 and 5825 nm. In the following experiments the comparison and test fields were applied at the fovea and 6’ temporally to the fovea of the right eye, respectively, each subtending 2’ x I”. The test and comparison fields were presented in succession using 0.5~set flashes. The authors served alternately as subjects. Experiment
I (scotopic stimulus superimposed)
The successive phases of the operation were as follows: (I) After 30 min dark adaptation, the absolute thresholds to the lights of 510 and 65Onm. applied 6” temporally to the fovea, were determined in succession. The absolute thresholds were taken as the lowest intensity at which the subject was certain of seeing light. The intensity was increased in small steps from about @5 log unit below the expected thresholds. using 0.5.see flashes. (2) IO mill dark adaptation. l.3) Stimulation for 0.5 see with the e\trafo\cal test hcam of 650nm at an intensity 1.3 log units above the absolute threshold as measured in phase 1, representing an intensity level of about 2 photopic td. (4) @5 set dark adaptation. (5) Stimulation for 05 set with the fovea1 comparison field consisting of the instrumental primaries. The subject tried to establish a match by increasing or reducing the intensity of each of the three primaries. (6) Repetitions of phases 2-5 until a reasonably good match existed between the test and comparison fields. Occasionally the extrafoveal stimulus appeared too saturated to be matched by any fovea1 patch within the colour gamut of the RGB system. In such cases the test field could be matched by a monochromatic wavelength observed at the fovea. When the chromaticities of this wavelength had been established the chromaticities of the wavelength observed extrafoveally could be plotted in the fovea1 WDW system. (7) Phases 2-6 were repeated, except that in phase 3 only the scotopic stimulus (the 510~nm light) illuminated the test field. The intensity was 1.5 log units above the absolute threshold as measured in phase 1, representing an intensity level of about 1 log unit below the absolute threshold of dark-adapted cones, estimated, for the given stimulus, from the cone plateau of the long-term darkadaptation curve and from the absolute threshold of the dark-adapted fovea. (8) Phases 2-6 were repeated, except that in phase 3 the scotopic stimulus of 510nm was superimposed upon the extrafoveal test stimulus of 650nm. (9) Phases 1-8 were repeated, except that the .test field was illuminated with 600nm instead of 65Onm.
Red
Fig. I. Mean chromaticities within the fovea1 WDW system of an achromatic scotopic stimulus C,, a photopic stimulus C?, and their mixture C3. The wavelength of the photopic stimulus was 65Onm (0) and 6OQnm (0). The data points in Figs. 1-6 give the mean of four observations of subject US. being the same in all essentials for subject BS.
beam), the photopic stimulus C2 (the GO- or 600-nm test beams), and the mixture Cs. It will be seen that the chromaticities of Cs do not lie on or close to
the straight mixture locus C,C2 predicted by chromatic additivity, but are displaced toward yellow. DISCLSSION
The results clearly demonstrate that an achromatic scotopic stimulus, assumed to be well below the absolute threshold level of dark-adapted cones, may influence the sensation of hue observed at mesopic intensities. Thus, when the light of the 510-nm stimulus was superimposed upon the red test lights, the hue changed toward yellow. If the scotopic stimulus represents pure rod activity, it would be expected from the Principle of Univariance that the change in hue toward yellow would be independent of the wavelength of the scotopic light. In order to test the suggestion that the scotopic stimulus represents rod activity, Experiment 1 was repeated. except that the wavelength of the scotopic stimulus was varied between 420 and 550nm. In support of the suggestion, the chromaticity shift was independent of the wavelength used. Experiment
la,b
(dark adaptation)
Comparing a fovea1 and an extrafoveal field at a moment immediately before the cone-rod break of the long-term dark-adaptation curve, Lie (1963) found that the two fields could lx closely matched both with respect to saturation and brightness. However, after 35 min of dark adaptation the extrafoveal field appeared more bright and less saturated at various intensity levels above the specific threshold. As the intensity of the two fields in-
creased, the brightness and saturation differences gradually decreased until the two fields appeared identical at about 2 log units above the absolute threshold of dark-adapted
RESLZTS
The results of Experiment 1 are given in Fig. 1. The data points represent the chromaticities of the achromatic scotopic stimulus C, (the SlO-nm test
cones. Presupposing that the “potential cone threshold.” i.e. the pure cone threshold, remains at the cone plateau during the second step of the dark-adaptation curve, the change in saturation and brightness observed between 5 and 35 min of dark adaptation was ascribed to the increased sensitivity of the rods.
Rod
activity
effect
on
colour
1121
matching
Experiment 2 may be regarded as an extension of the work of Lie (1963). The chromaticities of a test beam were measured twice during dark adaptation: (I) at the moment when the potential cone threshotd; and (2) when the sensitivity of the rods were assumed to have reached their absolute dark-adapted values. Experiment
21 (determination
of potenrial
cone threshold)
The cone threshold without rod activity involved may be measured during dark adaptation within the rod-free fovea, but also in the periphery provided relatively long wavelengths are used in the test. The successive phases
of the operation were as follows. (1) tight o~prar~on. Subsequent to 10min of preliminary dark adaptation, the light-adapting system was brought into use. The size of the circular adapting field was eight degrees, centered 6’ temporally to the fovea of the right eye. The subject was light adapted for 3 min to a constant brightness of about 30,000 photopic td. The adapting white light was then removed. (2) Threshold measurements. With the subject in complete darkness. absolute thresholds were measured every minute. The size of the calorimeter field was 2” x I”, applied 6” temporally to the fovea. using OSsec flashes. Between different runs the wavelength of the calorimeter field was varied between 650 and 550nm. (3) Phases 1 and 2 were repeated with the differences that (a) the light adaptation and test field were centered at the fovea, and (b) the wavelength of the test field was varied between 650 and 460 nm.
RESULTS
The results of Experiment 2a are shown in Figs. 2 and 3. The dark-adaptation curves indicate that the final values of the potential cone thresholds and the threshold of the rods, under the given conditions of ex~rimen~tion, are reached respectively, after about 6 and 30 min of dark adaptation. Consequently, in Experiment 2b the chromaticities of the test beam were measured between 6 and 7min and between 30 and 31 min of dark adaptation. 4.0 li;
3.6
2.4 fa 2 0 E
2.0
H
I.2
I.6
2
4
6 6 10 Time in darkness.
12 min
14
I6
Fig. 3. Dark-adaptation curves of the wavelengths 6.50, 600, 575. 550. S30. 505, 485 and 46Onm as determined at the fovea. The size and duration of the test-stimulation were as in Fig. 2. The curves were made to coincide at 5 min of dark adaptation.
Experiment
2b (determination
of the chromaticity
shift)
The successive phases of the operation were as follows.
(I) Light adaptation.
This was a repetition of phase
I of Experiment ?a. (2) 6 min dark adaptation. (3) Test-stimulation for 05 set with the extrafoveal test beam of 6OOnm. The intensity was 1 photopic td, representing an intensity of about 1 log unit above the cone plateau as measured in Experiment 2a. (4) @5 set dark adaptation. (5) Stimulation for 05 set with the fovea1 comparison field, consisting of the instrumental primaries. The subject tried to establish a match by increasing or reducing the intensity of each of the three primaries. Only small adjustments were required since pre-experimentation had established an approximate match. (6) Repetitions of phases 3-5 until the subject had established the match. (7) 23 min dark adaptation. (8) Repetition of phases 3-6. (9) Phases l-8 were repeated, except that the intensity of the extrafoveal test beam was 10 and 100 photopic td. (10) Phases l-9 were repeated, except that the wavelength of the extrafoveal test field was: 460. 485. 505. 530, 550 and 575 nm. Occasionally, the extrafovea stimulus appeared too saturated to be matched by any fovea1 patch within the colour gamut of the RGB system. In such cases a reasonably good match could be obtained between the test stimulus and a monochromatic wavelength observed at the fovea. The chromaticities of the wavelength observed at the fovea could then be established. RESLZTS
The results of Experiment 2b are presented in Figs.
2
6
IO lime
14
in darkness
18
,
22 min
26
30
Fig. 2. Dark-adaptation curves of the wavelengths 654 600,575 and 550 nm recorded at 6” temporally to the fovea of the right eye. The duration of the test flash was O+ set and the size 2” x 1”. The dark-adaptation curves were arbitrarily made to coincide at 5min of dark adaptation.
4-6. In Fig. 4 the mean chromaticities of monochromatic wavelengths at an intensity of 1 photopic td observed at about 6 min (filled symbols) and 30 min (plain symbols) of dark adaptation, are given together with the mean chromaticities of the achromatic scotopie stimulus C,. In agreement with the finding of Lie (1963), the results show that the saturation of the different colours was reduced between 6 and 30min of dark adaptation. Furthermore, colours of 6001~~ (orange) and 55Onm (g~e~~llow) changed
I’LF STABELL and
BJCIR~ Sr.kalrLL
600
GO&_
e
Red
Red
Fig. 4. Chromaticity shifts of monochromatic wavelengths between 6 (tilled symbols) and 30min {plain symbols) of dark adaptation. The intensity of the wavelengths was held constant at about 1 photopic td.
markedly toward yellow, while 460nm (violet) and 485 nm (blue-green) changed toward blue. As the intensity of the test field increased, the change in saturation decreased (Figs. 5-6). At the highest intensity level, no differences in hue or saturation were observed for 600 and 575 nm.
Fig. 6. Same as Fig. 4. except that the intensity of the test field was 100 photopic td. The chromaticities of the wavelengths 460 and 485 nm could not be determined due to the low energy output provided by the apparatus at the short wavelengths.
chromatic stimulation of cones, and appears to oppose the assumption of the Duplicity Theory of vision that the rod response is of constant colour quality and contributes an achromatic component to the visual sensation.
DISCUSSlOX
On the assumption that the potential cone threshold, under the present conditions of experimentation, remained unchanged after about 6 min of dark a&p~tion, the change in hue and saturation observed in Experiment 2b might be ascribed to change in rod activity. In conformity with this suggestion, the colours were found to remain invariant between 6 and 30min of dark adaptation when, as a control, the light adaptation and test-stimulation were centered at the right eye fovea, while the comparison field was centered at the dark-adapted feft eye fovea. The evidence presented in Experiments 1 and 2b is consistent with the suggestion that the chromatic response of rods may change as a function of selective
REFERENCES Aubert H. (1865) Physiologic der Netzhaut. Morgenstern Breslau. Chodin A. (1877) fiber die Abhan~gkeit der Farbenempfindungen von der Lichtsdrke. In Sammlung Pfr~siolog&her A.bhandlungen (Edited by Preyer W.). pp. 375444. D&t, Jena. Cohn H. (1882) Uber Far~nemp~ndungen bei schwicher kilnstlicher Beleuchtung. Arch. Augenheiik. 11, 283-302. Hecht S., Haig C. and Chase A. M. (1937) The influence of light adaptation on subsequent dark adaptation of the eye. J. gen. Phy~~~~.20, 831-850. Wiliebrand F. (1890) Uber die specifisohe Heiligkeit der Farben. Beitrlge zur Psychologie der Gesichtsempfindungen. Sber. Akad. Wiss. Mien Ma&.-natttrwiss. KI. Abt. 98 f3f, 70-120.
Kohlrausch
A. (1931) Adaptation, Tagessehen und Dlmmerunassehen. In Hand&h der Norma/en und ~~r~ologi~~~~ Pb~s~o~offie [Edited by Bethe A., Berg mann G.. Embden G. and Elfinger A.), pp. 1499-159-t. Springer, Berlin. Lie I. (1963) Dark adaptation and the photochromatic interval ~oc~rnen~a ophth. 17, 41 l-510. Mandelbaum J. (1941) Dark adaptation. Archs Ophrhal. X.X 26, 203-239.
600
Red
Fig. 5. Same as Fig. 4, except that the intensity of the test field was 10 photopic td.
Purkinie J. (1825)Beobachrungen und Versuche :ur Physiologi; der Sinne. Neue Beitriige :ur Kenntniss des Sehens in Subiecticer Hinsichr. Bd. 2. Reimer, Berlin. Rushton< W. A. H. (1957) Physical measurement of cone pigment in the living human eye. .Vattrr~, Lo& 179. 571-573. Shultze M. (1866) Zur Anatomie und Physiologie der Retina. Arch. mikrosk. Anal. 2, 176-256. StabeIl B. and Stabell U. (1971a) Chromatic rod vision-l: Wavelength of test-stimulation varied. Stand. J. Psychoi. 12. 175-178. Stabell B. and Stabell U. (1973a) Chromatic rod vision-IX: A theoretical survey. Vision Rrs. X3,449-455. Stabeli 3. and Stabetl U. (1974) Chromatic rod-cone interaction. tision Rex 14, 1389-1392.
Rod activity effect on colour matching Stabeli U. and Stabell B. (1971b) Chromatic rod vision-II: Wavelength of pre-stimulation varied. SC&. J. Psychol. 12, 282-288. Stabell U. and StabelI B. (1973b) Chromatic rod activity at mesopic intensities. Msion Res. 13, 2255-2260. Stab4 U. and Stabell B. (1975) Scotopic contrast hues triggered by rod activity. Vision Res. 15 1115-l 118.
1123
Tschermak A. (1929) Licht- und Farbensinn. In Handbuch der Normulen und Pathologixhen Physiologic (edited by Bethe A., Bergmann G., Embden G. and Ellinger A.), Bd. 12 (1X pp. 679-713. Springer, Berlin. Wright W. D. (1946) Researches on Normal and Defictiw Colour &ion. Henry Kimpton, London.
