Colour-blind

conditions of atmosphere, or if the mind were disturbed by some imminent danger.

In colour-blindness, induced by disease or injury, although the ness induced loss of colour sense is usually confined to a small area of the by disease. retina, yet, as it is the central area, and therefore the part on which the image of small objects naturally falls, the danger of mistaking a colour is as great, and even more so than in congenital colour-blindness; for loss of colour-sense is in this case as already has been stated accompanied by loss of form-sense.

Colour-blind persons should be rejected for certain occupations.

Most suitable colours for signals, and causes which

On the general grounds that have been explained, the Committee are of opinion that it would, under any circumstances, be dangerous to trust the reading of signals to anyone who is totally or even partially colour-blind to the extent indicated above, and this opinion is fortified by practical tests which they have carried out. They consider that such a person under no circumstances should be allowed to take a post for which this defect renders him physically unfit, and with this object in view the tests employed in the examination should be of a nature to at once detect, not only pronounced colour-blindness but defective colour-vision of the above character.

On some railways white lights instead of green have been used as safety signals, but the former are liable to be confounded with other white lights which are not signals, more particularly in the modify their neighbourhood of towns. At sea the evidence shows that the use of a second coloured light in addition to a red is a necessity, and that a white light could not be substituted for it.

selection.

It has been suggested, on theoretical grounds, that all danger of misreading signals would be avoided by using for one a red and for the other a pure blue, as each of these colours is recognized by the red-green blind. Certain difficulties, however, present themselves in practice which preclude the employment of the blue, more especially for night signals. The desiderata for signals are, that they should be as bright as possible, and that their colour should be distinct when viewed at a distance. A red glass transmits about 10 per cent. of the luminosity of the lamp-light behind it; it is also a saturated colour, and appears unaltered in hue from whatever distance it may be viewed. A blue glass, as ordinarily met with, will appear purple, or even whitish, by lamp-light, as it transmits, besides blue, a large proportion of red rays, and, if it be pale, it will also transmit a variable quantity of all the colours of the spectrum : moreover, the luminosity of the light transmitted is, at the best, only some 4 per cent. of the naked light. If two glasses, one of blue-green and another of cobalt blue, be placed together, in front of the light, the red rays will be cut off, and the light will be a fairly pure blue, but the luminosity will be reduced to about 2 per cent. When the effect of foggy weather on the carrying power of different lights is considered (see Note a, page 303), it will be understood how this small luminosity will be again diminished, and that it will become practically nil. In making

The

the selection of signal colours, these facts have to be taken into account. The choice of a red light as a signal light is one in which theory and practice really agree, and it is in the selection of a colour for a second signal that the difficulty arises. only colour for the latter, which the red-green blind would be able with certainty to distinguish from the red, is the pure blue, and this has been shown to be an impracticable choice. This being the case, the second signal should be of the kind most suitable for normal colour-vision without regard to the requirements of those who are colour-defective. Evidently for carrying power it should be as near the brightest part of the spectrum as possible, but far enough away from the red to render the signals easily distinguishable. A yellow or greenish-yellow is inadmissible, as it might be mistaken for a white light under some circumstances, as is also the case with those greens which, when sufficiently light to be effective, allow some red rays to pass. It is for reasons such as these that most railway companies have Colours for adopted as a danger signal a rich ruby-red, and for a safety signals signal (where a white light is not used) a blue-green, which adopted by varies slightly in hue on different lines, as was shown in the railways. table given at page 292.

Navy.

The sealed pattern standards of red and green glasses used in Standard the Royal Navy are the best that have come before the Com- signal glasses mittee, and they suggest their adoption both for railways and used in Royal the mercantile marine. The sealed pattern green inclines to blue and cuts off all red light. The blue-green of the spectrum, when mixed with about 25 per cent. of white light, matches the hue of this glass, and owing to this comparatively small dilution it will also appear as a fairly saturated colour. Its luminosity also approaches that of the standard red light, which is very desirable.

colour

blindness.

The direct evidence before the Committee is not sufficient to Accidents enable them to say that accidents, either by land or by water, through have conclusively been traced to defective colour-vision, yet this by no means disproves the high probability that accidents have really occurred from such defects.* There can be no doubt that every colour-blind person employed afloat, or upon railways, in certain capacities, must of necessity be a source of danger to the public. As is known, colour-blindness is hereditary to a large extent, and we have it in evidence before us that in the training vessels in which the orphan children of sailors are educated there are about 4 per cent. of colour-blind boys. may therefore take it, apart from all other evidence, that a considerable number of the fathers of these orphans who were employed as sailors must have suffered from the same defect; and we have it in direct evidence that a considerable number of colour-blind people, officers and seamen, are actually at sea at

We

* In Dr. Joy Jeffries' book on "Colour-blindness; its Dangers and its Detection," the case of the loss of the "Isaac Bell" is fairly conclusively traced to colour-blindness. Other cases are mentioned in Mr. Bickerton's evidence.

VOL. LI.

X

Board of

Trade tests for colourvision.

Naming colours, a

defective test.

the present time. Allowing for those whose colour-vision has been found defective by the inadequate tests used, and who may not be afloat, it is certain that out of the 120,000 seamen who are employed, there must be a large number who are colourdefective, and consequently a source of danger to life. The statistics of the examinations of eyesight on railways, so far as they have come before the Committee, are eminently unsatisfactory. Although candidates for employment are occasionally rejected for defective colour-vision, yet the percentages of the rejections on different railways differ widely from each other, and from the average percentage of colour-blindness of the male population. The evidence taken on this subject points to these differences being due to the variation in efficiency of the tests employed, and the Committee have been forced to the conclusion that some men, whose vision is defective for colour and for form, are in all likelihood employed in positions where normal vision is essential for public safety.

The evidence, moreover, points to the fact that steps have not hitherto been taken (at least, as a rule) in judicial inquiries relating to the causes of accidents, to ascertain whether they were due to defective vision. The Committee are strongly of opinion that in cases of collision or accident, where the evidence is conflicting as to the recognition of a coloured light, witnesses should be examined both for colour- and form-sense.

The Committee have had before them evidence regarding the colour-vision testing of the marine service as laid down by the Board of Trade.

Tests may be divided into two classes: one dependent upon the correct naming of a colour, and the other on its correct appreciation. The first class are intended to combine with the detection of colour-blindness that of colour-ignorance, or the defective knowledge of the names of colours. The last class are intended to detect colour-blindness alone, colour-ignorance being independently tested. The tests which the Board of Trade have officially adopted, are described in Appendix II. The examination consists in requiring the examinee to name correctly the colours of cards by day-light, and of coloured glasses by lamp-light. The correct naming of the colours is alone insisted upon.

The Committee consider that the tests themselves and the method of applying them are necessarily open to very grave objection. The Board of Trade test cards and coloured glasses can be procured from dealers, and the Committee have no hesitation in saying that the colours may be correctly named in the testing room by colour-blind persons after a certain amount of instruction, which would consist in teaching them to distinguish the different cards or test glasses by their different luminosities. The glasses are red, pink, three kinds of green, yellow. neutral, standard blue, and pale blue, all of which are viewed by artificial light, usually that of an oil lamp. In trials made

before the Committee, several people, whom Holmgren's test had proved to be colour-blind, passed this lantern test, a fact sufficient to show that it is unsafe to trust to it. But besides this uncertainty as to the rejection of the colour-blind, it appears to the Committee that an injustice may also be done to the candidates by its use. They believe that a perfectly normal-eyed person, who has been educated to observe colours, would not be able to speak positively as to the precise names of the colours of some of these glasses when illuminated by lamp-light. Less educated candidates would be much more liable to make mistakes in these puzzling tints (which the Committee consider have neither use nor significance), and, from sheer confusion, to misname those colours which are the only real tests, and thus to fail to pass the examination. The only safeguard to a candidate thus rejected lies in the fact that he can be re-examined, and that more than once. Cases have been brought before the Committee's notice where a candidate who has failed at first has passed in a subsequent examination. If the test for colour-blindness used by the Board of Trade were fair to the candidate, and perfectly efficient, such a re-examination would be unnecessary, and passing upon reexamination would be impossible.

The evidence given by representatives of various railway Railway comcompanies shows that very few have any adequate system of panies' tests. testing. Nearly all the methods employed are defective, and even where the wool-test is applied it usually breaks down from a choice of improper colours, both for standards and comparisons. In some instances, a person, whom the Committee know to have very defective colour-vision, has been passed in their presence by railway examiners as possessing normal eyesight, and the impression made on the Committee is that many have probably been passed into the service who should most certainly have been excluded.

The Committee have had the opportunity of examining the Tests in the different tests carried out by the Royal Navy, and are glad to Royal Navy. find that they are most efficient, and of such a nature that it may be presumed that no one can pass them who is sufficiently defective in colour-vision to be any source of danger. The long periods over which the examination lasts, however, precludes the adoption in their entirety of these tests used for railways or the mercantile marine. The sealed pattern glasses for signals are excellent, and, as already stated, the Committee would suggest their adoption as the universal signal colours.

The Committee are of opinion that the tests for colour- Tests for blindness should be of such a character that they will readily colourdetermine whether a man is or is not colour-blind, but that, except blindness. for scientific purposes, it is not necessary that they should indicate what kind of mistakes he is likely to make. The fact that a person is found to be colour-blind by an efficient test, properly applied, is amply sufficient to show that his employment in certain occupations is a danger to the public. We lay some

Tests recom

mended by the Committee.

Holmgren's

test.

stress on this point, as, if it were required from the examiner that he should specify what would be the nature of a mistake that an examinee would be likely to make, it would open the door to controversy, and thus defeat the ends for which an examination is instituted. What should be required of the examiner is merely a statement that the candidate has either passed or failed in the examination. In cases of failure, where the candidate is under the impression that a mistake has been made, an appeal to some properly appointed expert should be allowed, and his decision should be final.

The Committee have carefully considered the question as to what tests should be recommended for general adoption on railways and for the marine service.

They are of opinion that tests which involve the naming of colours should be avoided in deciding the question of colourblindness. Failure to satisfy these tests may be due to colourignorance, and lead to the rejection of persons who are not really colour-blind. A candidate who fails should be informed to what cause his failure is due, whether to colour-blindness or to colourignorance, with a view to subsequent re-examination in the latter

case.

On the other hand, if the objects which the examinee is required to name are few in number and accessible to the public, since the chances are that no two of them are exactly alike even to a colour-blind person, he might be instructed as to the names which he is expected to give them, and thereby persons who are really and seriously colour-blind might be passed by the examiner as being free from any defect. Besides trustworthiness, the tests should be adapted for the examination of large bodies of men, and, provided efficiency be not sacrificed, they should be of an inexpensive nature. After practical trials, and also from theoretical considerations, the Committee are of opinion that the simplest efficient test is the wool-test of Holmgren, applied either in the form which Holmgren himself recommends, or in that of Jeaffreson, which is based on precisely the same principles. A full description of Holmgren's test, and of the proper methods of applying it, extracted from Holmgren's work on the subject, is given in Appendix III, page 375.

It is most important that the standard test-colours should be of a proper character both as to hue and also as to dilution with white, the efficiency of the test depending almost entirely on a proper selection. The Committee recommend that sealed patterns of all three test-colours should be kept by some central authority -such as the Board of Trade; and that every set of test-wools should be officially passed as fulfilling the necessary conditions as to these standard colours, and also as to the sufficiency and variety of confusion colours.

The standard test-colours which have been approved by Professor Holmgren have been referred to the spectrum. The first standard is a light green colour, which can be matched with a green in the spectrum (X 5660), when 40 per cent. of white is added