1220 J. Opt. Soc. Am./Vol. 71, No. 10/October 1981

Deuteranomalous color matching in the deuteranopic eye

Michael E. Breton* and William B. Cowan

Physics Division, National Research Council of Canada, Montreal Road, Ottawa, Ontario, K1A OR6, Canada

Received January 19, 1981; revised manuscript received June 20, 1981

Two observers were classified as deuteranopes by standard tests including two-degree anomaloscope matches.Color matching similar to the Rayleigh type was then carried out for a 10-degree field size at retinal illuminancesranging from 1 to more than 3000 trolands (td). The results show that at the larger field size and higher levels ofretinal illuminance, a third independent color-mediating mechanism with the sensitivity of the deuteranomalouscone is participating in the color match. The results also confirm participation of a different third mechanism withrod sensitivity at levels below about 100 td. There is a range of transition between the two as the level increasesabove 100 td. Therefore large-field color matching in these deuteranopes is trichromatic at the levels tested, notdichromatic, and a third cone system is found to operate at typical photopic light levels under static viewing condi-tions in a dichromatic eye.

INTRODUCTION

Standard screening techniques for identifying defective colorvision are generally restricted to testing visual fields of twodegrees and smaller (i.e., anomaloscope tests, Farnsworth-Munsell 100-Hue and D-15 tests, pseudoisochromatic plates,etc.). Defects identified by these tests include dichromacyof the protan and deutan types. In particular, the anom-aloscope requires color matching along the Rayleigh line ofthe color diagram, the results of which are considered to bedefinitive in making a diagnosis of protan or deutan dichro-macy.

However, recent work has shown that observers who aresmall-field red-green dichromats in the classic pattern havetrichromatic color-naming and color-matching abilities underlarge-field viewing conditions.'- 5 The work of Smith andPokorny in particular indicates that at low photopic levels,the large-field trichromatic color matching of small-fieldprotanopes and deuteranopes is mediated by a receptor withthe sensitivity of rhodopsin. This result contrasts with otherrecent work by Piantanida and co-workers6' 7 showing evidenceof the activity of a third cone type under transient conditionsin the red-green dichromatic eye.

These facts make uncertain the nature of color discrimi-nation in classically diagnosed red-green dichromats. In thispaper we attempt to clarify some of these points. We presentevidence that the large-field color-matching ability of deu-teranopic small-field observers exhibits trichromacy of a dualnature with respect to the level of illumination. The evidenceindicates that at low photopic levels large-field trichromacyis subserved by a receptor with rod spectral sensitivity, con-sistent with the Smith-Pokorny result, and at higher levelsby a receptor with the spectral sensitivity of a deutroanom-alous cone. The broad range over which reliable trichromaticcolor matching was obtained in conjunction with the unres-trictive viewing conditions indicates further that thesesmall-field dichromats have available to them a trichromaticdiscrimination ability under ambient conditions of large-fieldviewing.

RATIONALE

We examined color matches along the Rayleigh line since theygive definitive information on the number of independentcolor-mediating systems and are well understood for normaltrichromacy as well as for many conditions of abnormal colorperception. To match lights in the long-wavelength portion(540-700 nm) of the spectrum, normal observers require twoprimary lights at the extremes of this spectral range. 8 Suchobservers use a unique mixture of the two primaries to matchany intermediate wavelength, forming in the process theRayleigh equation: if R is the amount of the red primary andG of the green primary, then R + G = C, where C is theamount of the intermediate wavelength.

If the R-to-G ratio used to match a given X is unique, it in-dicates two independent mediating systems in this spectralrange and, given normal color-matching behavior in the shortwavelength region, the observer is trichromatic. We take thisto mean that three cone types, called here the red, green, andblue cone types, produce the match. Unique matching witha shift in the R-to-C ratio relative to the population norm isanomalous trichromacy of either protan (too much red) ordeutan (too much green) type. We take this to indicate a shiftin sensitivity of one of the two long-wavelength cone types:a shift in the red cone for protanomaly and a shift in the greencone for deuteranomaly. 9-"1 The range of acceptable ratioscharacterizes the severity of an anomalous trichromatic defect.Complete broadening of the range so that the endpoints areacceptable matches to each other as well as to each interme-diate A shows monochromacy in this spectral region andoverall dichromacy, given normal blue functioning.12 Weinterpret dichromacy to be the effective loss of one cone type:the red cone for protanopia, the green cone for deuteran-opia. 13,14

Our test investigates the Rayleigh equation for small (2degrees) and large (10 degrees) field size. We varied theoverall field illuminance from low levels (1 td) to very highlevels (10,000 td). Our deuteranopic observers were able tomatch the R and G primaries to each other under the small-

0030-3941/81/101220-04$00.50 © 1981 Optical Society of America

M. E. Breton and W. B. Cowan

Vol. 71, No. 10/October 1981/J. Opt. Soc. Am. 1221

field conditions as required for dichromacy. Our test condi-tions were designed to see if this dichromacy persists forlarge-field sizes and to indicate in quantitative terms how anyadded independent mediating system was affecting the matchequation.

METHODS

We used the NRC trichromator in the classical color-matchingparadigm for a 2- and a 10-degree field.'5 The trichromatoris a multichannel color mixer employing Maxwellian viewing,prism monochromators, and xenon-arc sources. It is capableof presenting stimuli of high spectral purity at high levels ofretinal illuminance through the entire range of visible wave-lengths.

The arrangement used in this experiment presents a hori-zontailly bipartite 2- or 10-degree field (Fig. 1) to one eye of theobserver. Red (M: 645 nm; half-band width, 8 nm) and green(G: 526 nm; half-band width 6 nm) spectral stimuli arecombined in the top field half to be matched to the testwavelength (X: 555, 570, 589, and 600 nm; half-band width,7-10 nm) added to a small amount of spectral blue (B: 444.4nm; half-band width, 12 nm). The visual field surround andobserver compartments are dark. Field size is adjusted byplacement of an aperture in the collimated portion of thebeam. Reduction in field size to a 2-degree field by placementof the appropriate aperture permits accurate comparison ofexperimental with classical small-field results. Luminanceis adjusted by neutral-density wedges and fixed filters. Ra-diance measurements are made directly, using a calibratedradiometer, by diverting the entire stimulus beam to the ra-diometer with an inserted first surface mirror betweentrials.

Most of the experimental data were gathered on one ob-server (MEB), and the trends were then confirmed on a secondobserver (JM). Both observers are classical deuteranopesaccording to the visual screening techniques, including theNagel anomaloscope (Type II), AO-HRR pseudoisochromaticplates, Farnsworth-Munsell D-15 and 100-Hue tests, andneutral point matches of a spectral stimulus at 504 nm to amixture of spectral stimuli of 444 and 645 nm. In addition,the Rayleigh match behavior obtained with the Nagel ano-maloscope was confirmed on the NRC trichromator set upappropriately using a 2-degree split-viewing field. This rig-orous screening procedure was used to make certain that thesesubjects behaved as classically acceptable dichromats.' 6

RESULTS AND DISCUSSION

Figures 2-5 show the results for observer MEB (those of ob-server JM are less complete but show the same trends),plotting on the x axis the red-green ratio needed to matchyellow (Y) at the brightness shown on the y axis. We havechosen to define log(R/G) as the red-green ratio for the fol-lowing reasons:

1. It is invariant under a change of units that affects R andG in the same way (e.g., trolands to kilotrolands, lumens totrolands).

2. It is invariant up to an additive constant under a changeof units that affects R and G in different ways (e.g., scotopicto photopic troland, watts to trolands). This means thatcurves produced by plotting raw data in instruments units

CIE 1931 CHROMATICITY DIAGRAM

Fig. 1.in thenm).

Bipartite stimulus field arrangement shown with the locationCIE diagram of the primaries and one test wavelength (589

3.51

.10

0

2.51

1. .

0 5

i- -re . M

X =555 nm

.0--\ 0-

-I\

-0.- 0.6 -0.4 -0.2 0 0.2D A NORMAL

LOG R/G

0.4 0.5 0.5ROD

EEB

1.0

Fig. 2. Plot of log (RIG) as a function of log retinal illuminance inphotopic trolands for individual matches to a 555-nm stimulus (dots)and the average of all matches at one intensity (circles). The arrowson the x axis indicate the log (RIG) for a deuteranomalous observer(DA), a normal observer (NORMAL), and an observer using a rod asthe third color receptor (ROD). Solid line is to guide the eye.

In

a:

0zI

0

20

0

D A NORMAL RODLOG R/G

Fig. 3. Same as for Fig. 2 except the match is to a 570-nm stim-ulus.

M. E. Breton and W. B. Cowan

1222 J. Opt. Soc. Am./Vol. 71, No. 10/October 1981

35

-0 4 -0 2 0 02 04 06 0.8 1.0DA NORMAL

LOG RIG

4. Same as for Fig. 2 except the match is

1 2 1 4 1.6ROD

to a 589-nm stim-

MEB

0 0.2 0.4 0.6 0.8 1.0 1.2DA NORMAL

LOG RiG

1.4 1.6 1.8 2.0ROD

Same as for Fig. 2 except the match is to a 600-nm stim-

1:.

I':

I 1'.,

""""-- ......................... , I

"' I

',1I

1I

C I I I I I I T 1 I IRECEPTOR I

LOG RIG

Fig. 6. Qualitative plot of log (RIG) versus log retinal illuminance,where the third receptor is the same at all retinal illuminances (solidline), one receptor at low retinal illuminances switching abruptly toanother at high retinal illuminances (long-dashed line), one receptorat low retinal illuminances switching gradually to another at highretinal illuminances (short-dashed line).

MEB

X I 589 nm

aa

-0, . , Y

. I T 12 14 16A

2.5-Inz0z

0

0to

1.5

05

(amperes from a photomultiplier, slit openings on an anomalo-scope, etc.) are identical with those in photometric units, ex-cept for an additive constant.

3. It treats R and G in a symmetric fashion.4. It is simply related to R/(R + G), the ratio defined by

Smith and Pokorny5; i.e., RIG = [(R + GIR) -11-1.

In this work we chose log RIG in preference to RIR + G mainlybecause of property 2, which the latter ratio does not pos-sess.

In Figs. 2-5 each point represents a match, and the line isto guide the eye. Arrows on the lower axis indicate the red-green ratios for trichromatic observers with normal blue andred cones and a third receptor that is a rod, a normal greencone, or a deuteranomalous cone. Action spectra for thenormal cones are taken from Vos and Walvaven14; for the rodwe took V', from Wyszecki and Stiles17 ; for the anomalouscone we took pigment absorption spectra from Pokorny et al. 9and corrected for intraocular absorption.1 7

To see the significance of these curves, note that a dichro-matic observer will find a match at any red-green ratio. Hismatches do not lie on a line but are scattered throughout thediagram. Note further that a trichromatic observer who usesthe same three receptors at all levels will always have the samered-green ratio. His matches lie on a vertical line at thered-green ratio of his receptor. A more complicated tri-chromatic observer who has two choices for the third receptor,of which he uses one at low light levels and the other at highlevels, will have matches lying on two vertical lines, with abreak between them at the light level at which he switchesreceptors. It is unlikely that there will be an abrupt switchfrom one to the other; instead a region of compromise willoccur where the match switches gradually from one receptorto the other. These patterns are shown in Fig. 6. ComparingFigs. 2-5 with Fig. 6 we see that observer MEB (and JM) fitsthe last pattern, with his two receptors being a rod at low levelsand a deuteranomalous cone at high levels.

Finally, we wish to point out the possibility of deviationsfrom this pattern at the highest levels (>3000 td). MEBfound it hard to make these matches, believing that no settingof the trichromatic controls produced a satisfactory match.This difficulty increased with level so that no data could betaken at higher levels. This effect obviously warrants furtherinvestigation.

In summary, we have confirmed the result of Smith andPokorny 5 that classical deuteranopes are trichromatic inlarge-field color matching. We have shown further that thetrichromacy seems to be mediated by two additional receptors,a rod and a deuteranomalous cone. A close investigation oflarge-field dichromatic color matching using a four primaryarrangement like that of Trezona and Clarke' 8' 1 9 is nowneeded to isolate the contributions of all four receptors.

ACKNOWLEDGMENTS

We thank James McLaren for serving as an observer in thisexperiment and Graham Fielder for his invaluable technicalassistance throughout the study.

We also thank Gunter Wyszecki and Alan Robertson forsupport and encouragement in this work.

Fig.ulus

5LU,o 2.5z

Jo0

X

O 1.5

0.5 .T

Fig. 5.ulus.

8,az"I

RECEPTOR 2

. . . . . .

.. . ., 1 , ,

M. E. Breton and W. B. Cowan

3.5 -

Xz600nm

- 1.. . ..

M. E. Breton and W. B. Cowan

* Present address, Department of Visual Physiology, WillsEye Hospital, 9th and Walnut Streets, Philadelphia, Penn-sylvania 19107.

REFERENCES

1. A. L. Nagy and R. M. Boynton, "Large field color naming of di-chromats with rods bleached," J. Opt. Soc. Am. 69, 1259-1265(1979).

2. M. E. Breton and W. C. Cowan, "Large field trichromacy indeuteranopes," in Digest of Topical Meeting on Recent Advancesin Vision (Optical Society of America, Washington, D.C. 1980),paper FB13.

3. A. L. Nagy, "Large field substitution Rayleigh matches of red-green color deficients," in Digest of Topical Meeting on RecentAdvances in Vision (Optical Society of America, Washington,D.C., 1980), paper FB12.

4. A. L. Nagy, "Large field substitution Rayleigh matches of di-chromats," J. Opt. Soc. Am. 70, 778-784 (1980).

5. V. C. Smith and J. Pokorny, "Large field trichromacy in prota-nopes and deuteranopes," J. Opt. Soc. Am. 67, 213-220 (1977).

6. T. P. Piantanida, F. Frome, and D. H. Kelly, "Spectral sensitivityof the forbidden cones in dichromats," Invest. Ophthalmol. 17(Suppl.), 197-198 (1978).

7. F. Frome, T. P. Piantanida, and D. H. Kelly, "Flicker thresholdsreveal forbidden cones in dichromats," Invest. Ophtbalmol. 17(Suppl.), 197 (1978).

Vol. 71, No. 10/October 1981/J. Opit Soc. Am. 1223

8. Lord Rayleigh (J. W. Strutt), "Experiments on colour," Nature(London) 25, 64-66 (1881).

9. J. Pokorny, V. D. Smith, and I. Katz, "Derivation of the photo-pigment absorption spectra in anomalous trichromats," J. Opt.Soc. Am. 63, 232-237 (1973).

10. J. Pokorny and V. C. Smith, "Evaluation of single-pigment shiftmodel of anomalous trichromacy," J. Opt. Soc. Am. 67,1196-1209(1977).

11. M. Alpern and J. Moeller, "The red and green cone visual pig-ments of deuteranomalous trichromacy," J. Physiol. 266, 647-675(1977).

12. J. Pokorny, V. C. Smith, G. Verriest, and A. J. L. G. Pinkers,Congenital and Acquired Color Vision Defects (Grune andStratton, New York, 1979).

13. V. C. Smith and J. Pokorny, "Spectral sensitivity of color blinkobservers and the cone photopigments," Vision Res. 12, 2059-2071 (1972).

14. J. J. Vos and P. L. Walraven, "On the derivation of the fovealreceptor primaries," Vision Res. 11, 799-818 (1971).

15. G. Wyszecki, "Color matching at moderate to high levels of retinalilluminance: a pilot study," vision Res. 18,341-346 (1978).

16. M. Alpern and T. Wake, "Cone pigment in human deutan colourvision defects," J. Physiol. 266, 595-612 (1977).

17. G. Wyszecki and W. S. Stiles, Color Science (Wiley, New York,1967).

18. P. W. Trezona, "The tetrachromatic colour match as a colori-metric technique," Vision Res. 13, 9-25 (1973).

19. F. J. J. Clark and P. W. Trezona, "Towards general systems ofcolourimetry and photometry based on the tetrachromatic colourmatch," C. I. E. Compte Rendu, London, 205-217 (1976).