J Neurol (1992) 239 : 451-454

Journal of

Neurology
© Springer-Verlag1992

A n e w w a y to use the Ishihara test

Dilogen V. de Alwis and Chee H. Kon
Sussex Eye Hospital, Eastern Road, Brighton BN2 5BF, UK

Received August 1, 1991 / Received in revised form January 9, 1992 / Accepted January 21, 1992

Summary. The Ishihara plates are widely used as a test normalities [5]. None of the various editions of the test
for colour vision. Originally designed for the purpose of include plates that explore the blue-yellow (tritan) axis,
detecting congenital red-green colour blindness, the test thus limiting its value in testing for acquired colour
also has some value in demonstrating acquired colour vision defects. Despite this, the plates are often used for
vision defects. There are, however, several disadvan- this purpose (e.g. in optic neuritis), where the number of
tages in the present arrangement of the plates. A modifi- errors made are counted as a score.
cation of the test, involving the rearrangement of the The plates in the Ishihara test are arranged in a
order of the plates, is presented which, together with a haphazard order with some producing similar informa-
new recording chart, simplifies both the administration tion interleaved with some that produce other forms of
and the interpretation of the test. information (see below), making scoring and interpreta-
tion of results difficult. This modification of the Ishihara
Key words: Ishihara - Colour vision - Daltonism test involves rearrangement of the order of the plates, a
recording chart with improved diagnostic facility and
new instructions for its use.
Introduction

The early attempts by Dalton [1] in 1947 to elucidate Rearrangement of the plates
colour blindness led to the recognition of the term dal-
tonism to mean an inherited red-green colour vision For the purpose of this study, a 1987, 38-plate edition of
defect. The Ishihara plates, probably the best known of the Ishihara test for colour blindness [3] is used. The
the colour vision tests, were first produced in 1917. They plates are rearranged in such a way that those producing
were designed to detect and to classify congenital red- similar information are grouped together.
green defects, and were based upon the pseudo-iso-
chromatic principles devised by Stilling [2]. Coloured Group 1: Discrimination plates
spots are used to form a figure within a background also
composed of coloured spots. The colours of the figure These include plates numbered 10-17 and 30-33 in the
and the background of each plate have been chosen so original test. They contain orange patterns on a greenish
that they differ only in the relative stimulation of either background or vice versa, and were designed to be read
long or medium wave length cones (i.e. they lie along by the normal observer, but misread or not seen at all by
protan or deutan confusion axes). Deliberate variation any colour abnormal observer, i.e. they discriminate be-
of spot size, in both pattern and background, eliminates tween colour-normal and colour-abnormal subjects.
visual clues which depend upon the detection of edges to
the pattern. Similarly, the brightness of the spots is Group 2: Transformation and "hidden figure" plates
varied to overcome problems associated with differences
in luminosity between figure and background. The test Transformation plates. Plates 2-9 and 34-37 again con-
is thus randomly anisoluminant and is free of problems tain orange figures on a greenish background or vice
related to edge artefact [6]. versa, but were designed to reveal one pattern to the
It must be remembered that despite widespread use normal observer, and quite another to the daltonic ob-
in a variety of clinical situations, Ishihara designed the server. Subjects with acquired defects may see all or part
plates only as a test for congenital red-green "colour of a figure as seen by the normal observer, or none at all.
blindness" [3]. The plates are extremely effective in de- (They will not, however, see the figure as seen by the
tecting severe protan and deutan defects, but can fail in daltonic observer.)
attempting to detect or classify minor forms of these ab-
"Hidden figure" plates. These include plates 18-21 and
Correspondence to: D.V. de Alwis, 3 Maytree Walk, Hove BN3 28-29, and contain figures discernible only to the dal-
7JB, UK tonic observer. The normal subject, or one with acquired
452

colour vision defect, sees only a random mixture of spots Instructions for use with modified Ishihara test
with no meaningful pattern. These plates therefore can- 1. The test should be conducted under a bright uniform light source
not be used in scoring for acquired defects. corresponding to illuminant C (e.g. MacBeth easel lamp). In the ab-
sence of such a source, the test may be conducted under bright dif-
fuse daylight or a daylight-balanced artificial light source.
Group 3. Classification plates 2. The plates should be held at a distance of about 75cm from the
observer, except in the case of presbyopic subjects, who should hold
This group, consisting of plates 22-25 and 26-27, at- them at a distance suitable for their reading correction. Each eye
should be tested separately. Until the examiner is familiar with the
tempts to differentiate between the protan and deutan test, the "instructions to the subject" (see below) should be read out
forms of daltonism, and between the trichromatic and aloud.
dichromatic types of deficit. Each plate has two figures. 3. With illiterate subjects, only those plates represented within brack-
ets in italics on the scoring chart (the "illiterate series") should be
It is intended that the protan dichromat would see one used. With literate subjects, either the "literate" series only or all of
figure only, and the deutan dichromat the other. The the plates may be used.
plates detect deutan defects better than protan [4]. 4. When testing specifically for the presence of inherited red-green
disease, test only with plates in groups 2 and 3, and for the presence
Anomalous trichromats may see both figures, but one of of acquired defects, only with plates in groups 1 and 2. If the type of
them more clearly than the other, thus identifying pro- defect is unknown, all of the plates should be used for the test.
tanomalous and deuteranomalous observers. Amongst 5. Each reply given by the subject is recorded by circling the cor-
subjects with acquired defects, only those with severe responding response on the chart. However, in groups 1 and 2, the
response is underlined in the "abnormal" row in group 1 and the
loss would find it difficult to identify these patterns. "acquired" row in group 2 when:
Rearrangement of the plates into the three groups, a) The pattern is partially seen or seen to be broken
as described above, greatly simplifies analysis of the b) The response is correct, but considered to be hesitant
(in group 3, in a similar situation, the response is marked with a circle
results. in the "normal" row). In marking responses to "hidden figure" plates
In the original test, plates of identical design (e.g. in group 2 (plates 20, 21, 19, 18, 29, 28), if the subject does not see a
plates 10 to 14 with orange figures on a greenish back- pattern, the response is only marked on the "normal" row (no re-
sponse is offered on the "acquired" row).
ground) are grouped together. This might lead to anxiety 6. If any of the responses fall into the "daltonic" row in group 2, the
and frustration, as a subject experiencing difficulty with defect is most likely to be an inherited one and responses to group 3
one plate might continue to do so with subsequent plates plates must also be tested. The category of the inherited defect may
become apparent in group 3.
of the same design. It is also possible that a subject might 7. If none of the responses to group 2 plates fall into the "daltonic"
develop a recognition strategy helping identification of a row, the subject either has normal colour vision, or an acquired defect
series of plates of similar design if they are presented and the test is concluded on completion of group 2 (i.e. group 3
plates are not used).
consecutively. This might interfere with the validity of 8. In counting the error score for acquired defects, assign 1 point to
counting an error score in testing subjects with acquired each circled response and 1/2 point to each underlined one. For this
colour vision defects. Therefore, in this modification, purpose, count only the responses on the "abnormal" row in group 1
within each group, the plates are further rearranged in a and the "acquired" row in group 2.
random manner, so that plates of identical design do not Instructions to the subject
lie alongside each other. (Instructions for the illiterate series appear in italics and those for the
Plates 30-33 (discrimination), 34-37 (transforma- combined test within brackets)
tion), 26-27 ("hidden figure") and 28-29 (classification) I am going to show you several plates such as this one (show plate 1).
are grouped separately at the end of each series. They Can you see a number here? When I show you the rest of the plates,
were originally designed for use with illiterate or in- I want you to tell me if you see any numbers. You may sometimes see
only part of the number, or find the number only with some difficulty.
numerate subjects who were asked to trace the pattern In such a case tell me that you find it difficult. There are some plates
(a line) on each plate. However, as about 8% of male that do not have any numbers. If you do not see any number, say so.
clinicians suffer from some degree of red-green defi- Do not worry if you do not see a number in several of the plates. You
are not expected to see numbers in all of the plates.
ciency, it may sometimes be difficult to administer the
test. In such a case, it is possible to ask the subject to I am going to show you several plates such as this one (show plate
38). Can you see that the line connects the two small x signs on the
name the colour (or colours) of the line (or lines) and edge of the design? When I show you the rest of the plates, I want
this variation has been incorporated into the recording you to tefl me if you see any lines. If you see a line, tefl me what colour
scheme. While the two sets of plates (i.e. the "literate" the line is. You may sometimes see only part of the line, or find the
line only with some difficulty. In such a case tell me that you find it dif-
and the "illiterate") may be used independently of each ficult. There are some plates that do not have any lines. If you do not
other, it is possible to use both sets together in subjects see any line, say so. Do not worry if you do not see a line in several of
with acquired disease in order to increase the number of the plates. Your are not expected to see lines in all of the plates.
questions asked, and hence improve the validity of the (I am going to show you several plates such as this one (show plate
error score. 1). Can you see a number here? Some of the plates may have a line
as in this one (show plate 38). Can you see that the line connects the
Plates 1 and 38, designed for instruction and demon- two small x signs on the edge of the design? When I show you the
stration, are placed at the beginning of the test. Failure rest of the plates, I want you to tell me if you see any pattern. If you
to read these indicates a severe acquired defect, hysteria see a number, tell me what it is. If you see a line, tell me what colour
the line is. You may sometimes see only part of the line or the
or malingering. number, or find the pattern only with some difficulty. In such a case
tell me that you find it difficult. There are some plates that do not have
any pattern. If you do not see any pattern, say so. Do not worry if you
The recording chart do not see a pattern in several of the plates. Your are not expected to
see patterns in all of the plates.)
The new chart simplifies the recording of the responses
and interpretation of the results. As the subject reads Fig. 1. Instructions
 453

out the plates the corresponding response in circled or the plates are held at a distance suitable for the subject's
underlined (see below) on the chart. When the test is near refraction.
completed, the pattern of deficit in a daltonic subject With illiterate subjects, only those plates represented
becomes obvious. In the case of acquired loss, an error within brackets in the recording chart (the "illiterate"
score of the abnormal responses is totalled, producing a series) should be used. With literate subjects, only the
numerical record of the subject's performance. "literate series" need be used, especially when testing
for congenital disease. In the case of acquired disease,
however, by using the entire series the number of ques-
New instruction chart tions asked may be increased, thus increasing the valid-
ity of counting an error score.
The new instruction chart (Fig. 1) contains two sections; The plates in group 1 only differentiate between nor-
the first instructs the examiner on conducting the test mal and abnormal colour vision and therefore may be
and the second contains instructions for performing the omitted when testing specifically for inherited red-green
test that may be directly read out to the subject. defects. The plates in group 3 are interpreted correctly
As with all non-illuminated colour vision tests, the by most subjects with acquired disease and, therefore,
test should be conducted under standardised lighting these plates may be excluded when testing specifically
corresponding to illuminant C (e.g. MacBeth easel lamp). for acquired colour vision defects. However, should some
In the absence of such, bright diffuse daylight or a day- of the subject's responses appear in the daltonic column
light-balanced artificial light source should be used. of group 2, it must then be assumed that the subject has
In the original version, it is recommended that the congenital red-green abnormality (in addition to any ac-
test is performed at a distance of 75 cm [3]. This, how- quired defect) and the test should include group 3 plates.
ever, can be unsuitable for a presbyopic subject, being In subjects with acquired disease, an error score is to-
incompatible with both near and distance refractive cor- talled from incorrect responses to plates in groups 1 and 2.
rections. It is therefore suggested that, in such a case, Both normal subjects and those with acquired disease do

ISHmARA RECORDING CHART ISHIHARA RECORDING CHART

Name : ~ Date : S ~ 9 ~ Name : s.w. Date : ~ 9 _ ~

ID No. : .... Eye : R i g h t @ IDNo. : Eye @ Left

Sex : @ F DUB : ~ 3 9 Indication: ? Colour B l i n d Sex : M @ DUB : J ~ _ ~ _ Indication: ? Qptic Neuritis

GROUP 1 DISCRIMINATION PLATES GROUP 1 DISCRIMINATION PLATES

Pla[¢no 1 38 1O 11 17 16 12 1,5 14 13 (31 30 32 33 ) Plateno i 38 10 11 17 16 12 15 14 13 (31 30 32 33 )

Normal @ @ 2 6 73 16 97 7 5 45 (G G O O)

Abnormal * * * * * * * * (* * * ~') Abnormal * @ --* * @ ~ C) * (* * *

GROUP 2 TRANSFORMATION & "HIDDEN FIGURE" PLATES GROUP 2 TRANSFORMATION & "HIDDEN FIGURE" PLATES

Plate no 8 20 3 2 21 9 19 5 4 6 18 7 (35 28 36 37 34 29 ) Plate no 8203221 91954 (,187 (,t5 28 36 37 34 29 )

Normal ~) * 6 8 * 74 * ( ~ ) 2 9 @@@ ( B G * PO PO BG * ) Normal @@@8 @ 74@57@ 5 @ 3 ( B G *@PO PO BG * )

Daltonic 17 @ @ 3 @ @ 35@ 2 5 5 (PB Y P " PB PB Y ) Daltonie 17 45 5 3 73 21 2 35 70 2 5 5 ( PB Y PB PB PB Y )

Acquired • ,@ . . . . . ( . . . . ) Acquired • * *__ * * • ~_ * (*._. • • , )

GROUP 3 CLASSIFICATION PLATES GROUP3 CLASSIFICATION PLATES

Plate no 24 23 22 25 (2627) Plate no 24 23 22 25 (26 27 )

Anomalous @ 42 26 96 ( R P RP ) Anomalous 35 42 26 96 ( R P RP )

Protanopic 5 2 6 6 (P P ) Protanopic 5 2 6 6 (P P )

Deut . . . . . pic 3 (~)~)@ (R R ) Deuteranopic 3 4 2 9 (R R )

• = misreador not seen * = misreador not seen

O = orange line, G = green line, R = red llne, p = purpleline O = orange line, G = green line, R = red line, P = purple line
BG = blue & green line, PB = purple & blue line BG = blue & green line, PB = ptmple& blue line
PO = purple & orange line, RP = red & ptmpleline PO = purple & orange line, RP = led & purpleline

DIAGNOSIS DIAGNOSIS

C O N G E N I T A L ~ '¢ or A C Q U I R E D Error score: CONGENITAL or ~ACQUIRED Error score:

Anomalous "literate" /16 Anomalous "literate" /16
Protan defect 7 "illiterate" /8 Protan defect "illiterate"
i

Deutan defect ~ , ¢ both /24 Deutan defect both

<22)
Fig. 2. Ishihara recording chart (daltonic observer) Fig. 3. Ishihara recording chart (subject with optic neuritis)
454

not perceive a pattern in the "hidden figure" plates in play format readily lends itself to rearrangement. Rear-
group 2. These plates are therefore excluded f r o m the rangement of the plates, as described above, combined
error score count. W h e n testing, each reply given by the with a new recording chart and scoring strategy, en-
subject is recorded by circling the appropriate response hances the usefulness of the test.
on the chart. H o w e v e r , in practice, it is seen that in ad-
dition to correct and incorrect responses, answers might
sometimes be hesitant or partially correct. By under- References
lining these responses and assigning half a point for
these, the usefulness of the error score system in record- 1. Abney W de (1895) Colour vision. (The Tyndall Lectures)
ing acquired disease is increased. Sampson Low, Marston, London, pp 58-61
2. Duke-Elder S (ed) (1962) System of ophthalmology, vol VII.
A set of instructions that can be read out to the sub-
Kimpton, London
ject has been formulated. While it is of particular benefit 3. Ishihara S (1987) Ishihara's test for colour blindness; 38 plates
to the examiner who is unfamiliar with the test, it is sug- edn. Kanehara, Tokyo
gested that the same format is adhered to in all cases so 4. Labowski R (1966) A critical evaluation of colonr vision tests.
as to standardise the administration of the test. Separate Br J Physiol Opt 23 : 186-209
instructions for the "literate" and "illiterate" series, as 5. Pickford RW (1950) A study of the Ishihara Test for colour
blindness. Br J Psychol 40 : 71-80
well as for conducting both series together, are included.
6. Reffin JP, Astell S, Mollon JD (1991) Trials of a computer-con-
Figure 2 shows a completed recording chart for a trolled colour vision test that preserves the advantages of
daltonic observer, and Fig. 3 for a subject with optic pseudoisochromatic plates. In: Drum B (ed) Colour vision de-
neuritis. ficiencies, vol 10. Kluwer, Dordrecht, pp 69-76

Conclusions

The original order of the Ishihara plates m a k e s the re-

sults of the test difficult to interpret. Their present dis-