Full text of DOI:10.1007/s12009-001-0007-z

A B S T R A C T
This study evaluated the color discrimination of adults who were
exposed to long durations of sunlight during their lifetime. The
Farnsworth-Munsell 100-Hue test was used to examine the color
vision of 141 subjects (mean age, 46.5 years), who had no ocular
diseases and whose visual acuities ranged from 20/40 to 20/16.
Results showed higher total error scores than previously published
norms. There was a tendency for blue-yellow axis, especially in men
older than 45 years. Long exposures of ambient light may cause
color vision deficiencies.
O R I G I N A L A R T I C L E
Color Vision Testing in an Area With
High Levels of Ambient Illumination
Abdelmalik Nacer, PhD, & Mosa’ad A.
Al-Abdulmunem, PhD, FAAO
ight exposure in some geographical areas where
the annual average sunshine duration is high can
L
damage the retina and threaten vision. Since the
early 1960s, there has been extensive research in ani-
mals, which shows persistent damage to photorecep-
tor and epithelial cells in the retina.1 Some of the
underlying mechanisms of the damage are believed
to be the result of phototoxicity as well as metabolic
disorders from extended photic overbleach.2,3 Recov-
ery from the damage was found to be incomplete
most of the time, and visual function was therefore
likely to be affected.4
In humans, it has been shown that high amounts of
visible light, especially short-wavelength light (eg, vio-
let and blue) can produce photoreceptoral and pig-
ment epithelial damage in the retina.5 Reports of an
association between sunlight exposure and age-relat-
ed macular degeneration have been conclusive only in
men because of their higher frequency of sunlight
exposure compared with women.6 There have also
been reports of retinal damage produced by long expo-
sures of bright ambient light in desert areas.7
The study of the integrity of the visual processes in
a certain environment requires testing some of the
functions of these processes, for instance, color vision.
One possible site where visual degeneration can lead
to color vision disturbances is the retina. It has been
Reprints:
Abdelmalik Nacer, PhD, Department of Optometry, King Saud University, P.O. Box
25208, Riyadh 11466, Saudi Arabia.
The authors are from the Department of Optometry, King Saud University, Riyadh,
Saudi Arabia.
Acknowledgments
The authors thank Dina Al-Ismail for helping with administration and subject selec-
tion for the Farnsworth-Munsell 100-Hue test and Lotfi Tadj, PhD, and Madjid Amir,
PhD, for statistical assistance.
ANN OPHTHALMOL. 2001;33(2):125–130
125

documented that color vision can be impaired in a
large number of diseases.8 As early as 1912, Kollner9
reported that damage to the retina led to blue-yellow
defects, whereas deeper optic nerve disease caused
red-green defects. Now it is widely accepted that color
vision defects, both blue-yellow and red-green types,
can be found where there is some retinal dysfunction
due to known disease or to retinal degeneration of
unknown etiology.8,10-14
The classification of color vision defects into types
1, 2, and 3 goes back to Verriest10 in 1963 and was later
modified by Pokorny and coworkers.12 In the type 1
defect, red-green impairment comes as a result of
cone degeneration. In type 2, lesions of predominant-
ly postreceptoral origin (optic nerve and beyond) were
found to cause red-green and, to a small degree, blue-
yellow defects. Type 3 color vision defects were found
to be the most common consequence of retinal dis-
ease. Examples of retinal diseases leading to type 3
color defects include toxic retinopathies,15 diabetic
retinopathy,16,17 and optic neuropathies (eg, optic neu-
ritis18 and glaucoma16,19,20).
Type 1 defects are most likely caused by cone dys-
trophy syndromes,21 and type 2 defects probably occur
following optic disc nerve disorders.8,22 However, in
some cases where foveal fixation is maintained (good
visual acuity), it is a predominantly type 3 defect that
is predicted. When foveal fixation is not spared by the
disease, type 2 defect will most likely predominate. It
is therefore not always true that retinal diseases pro-
duce type 3 defects. There are cases where the sepa-
ration into blue-yellow or red-green defect is not
possible.8 The neural pathways that originate from
the blue cones and that convey exclusively color
information to the visual cortex23,24 seem to be less
resistant to disease than do pathways from red and
green cones. Although ultraviolet B (UV-B) radiation
is absorbed nearly fully by the cornea and lens, there
is no doubt that high quantities of short-wavelength
radiation can be more damaging to photoreceptors
than can longer wavelengths.
Based on these 2 hypotheses, the measurement of
color discrimination in a highly illuminated environ-
ment can be an important indicator in determining
whether there is or there is not damage to the sensory
cells of the retina (ie, photoreceptors). Color blindness
was obtained in the rhesus monkey25 with moderately
strong spectral light stimulation (8 × 10-4 W/steradian 3
hours a day for 21 days), and no recovery was seen 5
months after exposures ended. This cumulative loss
of retinal sensitivity, confined especially to blue
region of the spectrum was later found to be the result
of cone damage.26 Results of color vision testing in
humans are probably not enough to prove that there is
retinal damage but will undoubtedly elucidate the
structural changes that can occur in the eyes of a pop-
ulation exposed to high levels of visible radiation.
In this study, color discrimination was measured in
a large group of healthy adults with normal vision who
lived in the Arabian Peninsula all their life.
Subjects & Methods
Subjects. We randomly selected 168 adults from
patients who visited the refraction unit of the oph-
thalmology service in Riyadh (Saudi Arabia) Central
Hospital between September 1998 and January 1999.
Only 141 subjects fit our criteria of good vision and
ocular health after we performed further ophthalmo-
logic assessment and took a general medical history.
The remaining 27 subjects were excluded because
they showed pathologic signs, which were not previ-
ously reported in their files, although their visual acu-
ity was acceptable. These conditions consisted of the
following, as reported in ascending order of incidence:
uveitis, optic disc edema, early cataract formation,
mild glaucoma, macular degeneration, and recent-
onset hypertension or diabetes. The other criteria for
inclusion in the study were Saudi nationality, age
above 34 years, and visual acuity of at least 20/40.
The subjects were exclusively from a population of
Saudi nationals who lived all their life in the Arabian
Peninsula, mainly around the Riyadh area, which is
known for its arid and dry environment. Their age
ranged from 35 to 71 years, with a mean of 46.46 years
(SD, 7.29 years). Of the 141 subjects, 81 were male
(57.5%) and 60 were female (42.5%). The best-corrected
visual acuity was at least 20/40, or 6/12 (mean SD,
0.896 0.166 or 20/22.3). Appropriate near add was
used, and testing was monocular (in the eye with the
better visual acuity).
Subjects were given a questionnaire that asked
some questions about social and occupational activi-
ties. Forty-nine (61.25%) of the men said they spend a
yearly average of 6 daylight hours outdoors in occupa-
tional and recreational activities. The remainder
(38.75%) reported that they spend a daily average of
only 3.5 hours outdoors annually. Of the women, 33%
reported spending an average of 4 hours outdoors and
67% spend just under 2 hours on average. The regular
use of sunglasses among subjects 50 years and older
was 29% for the men and 22.4% for the women. (The
majority of good quality sunglasses are UV-blocking.)
In the younger group (under 50 years of age), it was
47.6% and 34.1%, respectively.
Methods. The Farnsworth-Munsell 100-Hue (FM 100-
Hue) test (Munsell Color, Macbeth Corp, Baltimore, Md)
was used to test all 141 subjects. The same procedure
was followed as given in the test manual.27 An average
of 2 minutes was allowed for each box, and the test was
repeated twice with a break of 5 minutes between tests.
Two Macbeth Easel fluorescent lamps (Macbeth
Corp) were used, giving 85 foot-candles at 6700°K of
illumination, and a shield around the subject’s head
was used to stop most of the extraneous light coming
from around the room. The testing distance was 40 to
50 cm with appropriate near addition. An adaptation
time of 1 minute preceded each testing session.
The test procedures were clearly explained to the
subjects, and a verbal consent was obtained before the
beginning of the examination. The testing involved no
discomfort and took no more than 15 to 20 minutes.
126
ANN OPHTHALMOL. 2001;33(2)

The Riyadh area is a region where the average
number of clear sunny days per year far exceeds that
of any European or North American city. Table 1
shows a comparison between ambient light levels in
Riyadh and the global average earth levels. The ambi-
ent light levels that prevail in summer (June through
September) on a clear day can approach, if not exceed,
maximum eye sensitivity levels.
Results
The results of this study were automatically comput-
ed using a software program (manuscript in prepara-
tion) developed entirely in our department (see case
example in Fig 1). The FM 100-Hue data with total
error scores for the 141 subjects are illustrated in a
histogram (Fig 2). More than half the sample had
scores higher than 102 errors (mean SD, 113.3 2.12)
and at least 25% of the sample had scores greater than
164 errors.
The data, illustrated in this histogram, show a nor-
mal distribution, with 73 subjects (51.8%) scoring more
than 100 errors. The remaining 68 subjects (48.2%)
may be considered to have medium to high color dis-
crimination. The majority of subjects, therefore,
seemed to have difficulties in making good color judg-
ments in conditions where color discrimination was
truly important (at least experimentally). Statistically,
the probability that anyone will score more than 100
errors is quite significant (P = .008, all ages combined).
When presented as a function of age, the same data
showed an increase of total scores with age (Fig 3). The
Pearson product-moment correlation coefficient r is
equal to 0.36, indicating a nonnegligible dependence
relationship with age (Student t test: P<.01). Subjects
in the lower age range showed relatively lower scores
than did the older age group (mean, 93 68 for sub-
jects 45 years and less; mean, 139 67 for subjects
above 45 years). To be statistically significant (we take
P = .05 for the z test), the inferred mean for the first
(younger) group was around 81 errors, whereas for the
older group it was around 125. The dependence of
these data on sex also was analyzed, and the data
Fig 1.—Results of Farnsworth-Munsell (FM) 100-Hue test in a 60-year-old man
(patient 8), showing error distribution mostly along a blue-yellow axis. Calculation of
both total and partial (blue-yellow or red-green) scores is illustrated. VA indicates
visual acuity.
showed a statistically significant relationship (P =
.043) in favor of the male population (mean total error
score, 123 for men vs 104 errors for women).
To differentiate between subjects who made errors
along the red-green axis or the blue-yellow axis, total
T A B L E 1
Ambient Luminance Levels Measured in Riyadh, Saudi Arabia, Area, 1997
Area and
Month
Earth on Cloudy Day Sky on Cloudy Day Earth on Clear Day Sky on Clear Day
(ft-c)
(ft-c)
(ft-c)
(ft-c)
Riyadh in December
Riyadh in March
Riyadh in July
700
1100
1200
2300
1900
NA
80
3700
NA
3200
4400
600
6400
11500
1100
Earth average
200
Data for Riyadh area are the average of 4 readings obtained at separate times of the day (10 AM, noon, 2 PM,
and 4 PM), using a photometer (General Electric, Cleveland, Ohio). NA indicates not available; ft-c, foot-candles.
ANN OPHTHALMOL. 2001;33(2)
127

Fig 2.—Frequency histogram showing the distribution of total error scores. The num-
ber of subjects scoring between 40 and 150 scores (midpoint at approximately 100
scores) is highest in this normal distribution.
Fig 4.—Scatter graph for age-corrected axis of the color defect. The graph is
obtained by plotting corrected square roots of {(B-Y) - (R-G)} as a function of age,
where B-Y indicates blue-yellow and R-G is red-green. The correction formula28 is:
1
1
⁄
2
⁄
2
(B-Y ) - (R-G) +Age × (–0.061)+2.035.
The scatter plot in Figure 5 shows the square root of
blue-yellow partial scores minus the square root of
red-green partial scores as a function of age and cor-
rected for age-related tritanopia (Corrected Axis =
Relative Axis – 0.061 × Age+2.035).14,28 Positive values
indicated the blue-yellow axis, whereas negative val-
ues indicated the red-green axis. The tendency was for
the blue-yellow axis, especially at older ages
(mean SD, +0.61 1.41, r = 0.28, P = .003 all ages
combined). In the younger age group (less than or
equal to 45 years, the probability that the axis was
toward blue-yellow was not statistically significant. In
the older group, however, the probability that the axis
of the defect was toward blue-yellow was significant
(P = .044) when the overall data of all sexes were ana-
lyzed together, but not as significant when female data
were analyzed separately (P = .097).
Fig 3.—Square roots of total error scores plotted against age. Straight line drawn
through the data points is the least squares best fit (slope = 0.32).
score data were split into 2 separate columns. A col-
umn consisting of errors along the blue-yellow oppo-
site quadrants and included partial scores from color
cap numbers from 1 to 12, 54 to 74, and from 76 to 84.
Another column consisting of errors along the red-
green opposite quadrants included color cap numbers
ranging from 13 to 33 and from 55 to 75.28
In Figure 4 the square roots of blue-yellow scores
are plotted against red-green scores to show the over-
all axis tendency of the data. As blue-yellow scores
increased, red-green scores increased. In 83% of the
cases, however, the trend was toward the blue-yellow
axis, considering data left or above the straight line
(slope 1) as having a blue-yellow axis and those right
or under the line as having a red-green axis. With a
high Pearson product-moment correlation coefficient
of 0.79, the partial scores of both red-green and blue-
yellow seemed to correlate in a clear and evident man-
ner. This means that subjects, who made large
numbers of color misjudgments, made them in both
directions but with more in the blue-yellow than in the
red-green axis.
Discussion
The FM 100-Hue test is not a simple screening color
vision test. It has been credited for its ability to detect
small amounts of congenital and acquired color vision
defects.10,11 In our study, total scores of the FM 100-Hue
test for 141 normal subjects showed higher means
than those previously published.28,29 Our data also
showed an overall trend for higher error scores with
age, which is a logical developmental process as in
many other biological processes. There is no doubt
that the aging factors underlying retinal sensitivity
will account for these higher scores, a proportion of
which will come from early media changes, such as
lens yellowing.30 However, the amounts of the defects
reported in these data suggest more than just normal
physiologic changes of age. The younger group (under
45 years of age) shows significant scores (mean, 97)
that would not be appropriate to justify by the effect of
the aging process alone. Although great care was taken
in the selection of our sample, which was limited to
healthy adults free of ocular disease, our main concern
128
ANN OPHTHALMOL. 2001;33(2)

adaptation might show increased thresholds. These
features remain to be tested. Unlike some diseases in
which the color vision defect depends on the type of
retinal or postreceptoral site of the disease, the defects
that we have shown in this study of healthy subjects
bear the Verriest type 3 characteristics (or Marré type
1 acquired color defect).34 This supposes that midpe-
ripheral aspect of the retina may be more affected than
the central retina,14,35 which explains the high visual
acuity levels and yet poor color discrimination in our
subjects. It is not surprising, therefore, to see that, in
an environment like the desert of the Arabian Penin-
sula, older individuals (especially men) who surely
must have been exposed to the effects of high ambient
illumination, will most probably show more color
vision deficiencies than will residents of less arid parts
of the world. Whether the environment affects color
discrimination is a plausible scenario, but we do not
rule out with absolute certainty other influencing fac-
tors, such as undetected ocular or systemic diseases.
Fig 5.—Scatter graph for blue-yellow (B-Y) partial error scores vs red-green (R-G)
partial scores. The line in the graph is a slope 1 line, which separates between scores
of blue-yellow axis tendency (points above the line) and red-green axis tendency
(points below the line).
was to limit the number of factors that could play a
role in the deterioration of color vision performance.
When the data are looked at more closely, it is found
that a predominantly blue-yellow axis is singled out in
more than 50% of the cases. However, there are fewer
cases that show an exclusive distribution of errors
along the blue-yellow or the tritanopic axis. In fact,
and from a broad angle, the higher the errors along
the blue-yellow axis, the higher the errors along the
red-green axis. This leads to the presumption that
color perception in its widest sense has been affected
without clear or specific indication as to its type.
The argument that we tried to develop in this
study—that the environment (especially ambient sun-
light) affects color vision—is based on 4 hypotheses.
First, the long-term effect of excessive light on pho-
toreceptors and retinal pigment epithelium, demon-
strated extensively in animals, using various
parameters in the control of the type of light, exposure
duration, animal species, and so forth, is now widely
accepted.1,5 Second, there is an overwhelming number
of studies in the literature linking color vision defi-
ciencies to ocular disease.24 When a known ocular dis-
ease is present (especially in the retina and
postreceptoral pathways), a color vision defect is near-
ly always found. However, one cannot be all that cer-
tain to justify an ocular disease in the presence of a
color vision defect. Third, most cases, apart from a few
(Fig 5), do not show selective axis of the color vision
defect. This presumes that there is a small number, if
any at all, of defects with a congenital origin. Fourth,
the chromatic, color-opponent mechanism, largely
responsible for fine color discrimination, is more vul-
nerable to diseases and retinal changes than the
achromatic luminance system,31,32 especially the blue-
yellow system (ie, short-wavelength sensitive cones).33
The gross vision of a normal person might appear
normal, but some features of fine vision such as color
discrimination, contrast sensitivity, stereopsis, or dark
References
1. Lanum J. The damaging effects of light on the retina: empirical
findings, theoretical and practical findings. Surv Ophthalmol.
1978;22:221–249.
2. Noel WK, Walker VS, Kang BS, Berman S. Retinal damage by
light in rats. Invest Ophthalmol Vis Sci. 1966;5:450–473.
3. Ham WT, Muller HA, Sliney DH. Retinal sensitivity to damage
from short-wavelength light. Nature. 1976;260:153–155.
4. Tso MOM, Wallow IHL, Powell JO, Zimmerman LE. Recovery of
the rod and cone cells after photic injury. Trans Am Acad Oph-
thalmol Otolaryngol. 1972;76:1247–1262.
5. Taylor HR, West S, Munoz B, Rosenthal FS, Bressler SB, Bressler
NM. The long-term effects of visible light on the eye. Arch Oph-
thalmol. 1992;110:99–104.
6. Cruickshanks KJ, Klein R, Klein BEK. Sunlight and age-related
macular degeneration: the Beaver Dam Eye Study. Arch Ophthal-
mol. 1993;111:514–518.
7. Young RW. Solar radiation and age-related macular degeneration.
Surv Ophthalmol. 1988;32:252–269.
8. Hart WM Jr. Acquired dyschromatopsias. Surv Ophthalmol. 1987;
32:10–31.
9. Kollner H. Die Störungen des Farbensinnes: ihre klinische Bedeu-
tung und ihre Diagnose. Berlin: Karger; 1912.
10. Verriest G. Further studies on acquired color vision discrimina-
tion. J Opt Soc Am. 1963;53:185–195.
11. Krill AE, Fishman GA. Acquired color vision defects. Trans Am
Acad Ophthalmol Otolaryngol. 1971;75:1095–1111.
12. Pokorny J, Smith VC, Verriest G, Pinkers AJLG. Congenital and
Acquired Color Vision Defects. New York, NY: Grune & Stratton;
1979.
13. Mullen KT, Plant GT. Colour and luminance vision in human optic
neuritis. Brain. 1986;109:1–13.
14. Silverman SE, Hart WM Jr, Gordon MO, Kilo C. The dyschro-
matopsia of optic neuritis is determined in part by the foveal/per-
ofoveal distribution of field damage. Invest Ophthalmol. Vis Sci.
1990;31:1895–1902.
15. Lagerlof O. Drug-induced colour vision deficiencies. In: Verriest
G, ed. Colour Vision Deficiencies. Bristol, England: Adam Higler
Ltd; 1980:317–319.
16. Adams AJ. Chromatic and luminosity processing in retinal dis-
ease. Am J Optom Physiol Optics. 1982;59:954–960.
17. Greenstein VC, Shapiro A, Hood DC, Zaidi Q. Chromatic and
luminance sensitivity in diabetes and glaucoma. J Opt Soc Am.
1993;10:1785–1791.
18. Menage MJ, Papakostopoulos D, Dean Hart JC, Papakostopoulos
S, Gogolitsyn Yu. The Farnsworth-Munsell 100-Hue test in the
first episode of demyelinating optic neuritis. Br J Ophthalmol.
1992;77:68–74.
19. Hart WM, Silverman SE, Trick GL, Nesher R, Gordon MO. Glau-
comatous visual field damage: luminance and color-contrast sen-
ANN OPHTHALMOL. 2001;33(2)
129

sitivities. Invest Ophthalmol Vis Sci. 1990;31:359–367.
20. Verriest G, Sucs F. Colour vision defects and colour perimetry in
ocular hypertension and glaucoma. In: Kulikowski JJ, Dickinson
CM, Kurray IJ, eds. Seeing Contour and Colour. Oxford, England:
Pergamon Press; 1987:399–402.
21. François J, De Rouck A, Verriest G. Progressive generalized cone
dysfunction. Ophthalmologica. 1974;169:255–284.
22. Pokorny J, Smith VC. Eye disease and color defects. Vision Res.
1986;26:1573–1584.
29. Verriest G, van Laethem J, Uvijls A. A new assessment of the nor-
mal ranges of the Farnsworth-Munsell 100-Hue test scores. Am J
Ophthalmol. 1982;93:635–642.
30. Lakowski R. Age and colour vision. Adv Sci. 1958;15:231–236.
31. King-Smith PE, Vingrys AJ, Benes SC, Havener WH. Differences
between perimetric thresholds for white and equiluminous red,
blue and yellow in a nerve fiber bundle defect. In: Drum B, Verri-
est G, eds. Color Vision Deficiencies. Vol 9. Dordrecht, The Nether-
lands: Klewer Academic Publishers; 1989:301–308.
23. Gouras P, Evers Hans U. Neural systems detecting spectral con-
trast independently of effective energy contrast: where and how
does colour begin? In: Kulikowski JJ, Dickinson CM, Kurray IJ,
eds. Seeing Contour and Colour. Oxford, England: Pergamon
Press; 1987:41–58.
24. Mahon LE, Vingrys AJ. Normal saturation processing provides a
model for understanding the effects of disease on color percep-
tion. Vision Res. 1996;36:2995–3002.
25. Harwerth RS, Sperling HG. Prolonged color blindness induced by
spectral lights in Rhesus monkeys. Science. 1971;174:520–523.
26. Sperling HG. Spectral sensitivity, intense spectral light studies
and the color receptor mosaic of primates. Vision Res.
1986;26:1557–1564.
27. Farnsworth D. The Farnsworth-Munsell 100-Hue test for the
examination of color discrimination. Manual, revised. Baltimore,
Md: Munsell Color; 1957:2–17.
32. Drum B, Armaly MF, Huppert WE. Chromatic and achromatic sen-
sitivity in glaucoma. In: Drum B, Verriest G, eds. Color Vision
Deficiencies. Vol 9. Dordrecht, The Netherlands: Klewer Academ-
ic Publishers; 1989:262–272.
33. De Monasterio FM, McCrane EP, Newlander JK, Schein SJ. Den-
sity of blue-sensitive cones along the horizontal meridian of
macaque retina. Invest Ophthalmol Vis Sci. 1985;26:289–295.
34. Marré M. The investigation of acquired colour vision deficiencies.
In: Marré M, ed. Colour 73. London, England: Adam Hilger;
1973:99–135.
35. Marré M, Marré E. Different types of acquired colour vision defi-
ciencies on the base of CMV patterns in dependence upon the fix-
ation mode of the diseased eye. Mod Probl Ophthalmol. 1979;19:
248–252.
28. Smith VC, Pokorny J, Pass AS. Color-axis determination on the
Farnsworth-Munsell 100-Hue test. Am J Ophthalmol. 1985;100:
248–256.
130
ANN OPHTHALMOL. 2001;33(2)