]\gt ; Theory of Colour Vision.surement of By Sir W. DE W. ABNEY , K.C.B. , F.B.S. ( Received August 17 , 1912 .
) That of the retina by a colour will the hue of the spectrum is well known .
In this paper a method of measuring the amount of fatigue is described , showing that by fatiguing the retina colour-blindness is produced which is the same as that found when one of the three sensations is deficient in sensitiveness .
In the first part of this communication the fatigue is produced in one eye , and change in hue is registered by the unfatigued eye , and from these observations what we may call the factor of the fatigue is calculated , which is the ratio in sensitiyeness of the unfatigued sensation to that of the fatigued .
In the second part of the communication a method is described of measur .
the luminosity of what , for sho1t , we will call the fatigue spectrum , which is , of course , the luminosity of the spectrum as seen by the eye .
In previoos papers the value of the three sensations in luminosity to the normal eye has been given , and the red , green , and blue sensations have been indicated in equations as , and B.S. In this paper the luminosities of the three sensations of the spectrum have to be shown as curves of equal stimulation , and the percentage values of the different colours have been derived from these curves .
The curves of equal stimulation of the spectrum are the rdinary sensation curves , two of which ( the blue and the green ) have to be multiplied by factors to make them of equal area .
Equal ordinates in these curves will then give the amount of white light existing in the different colours .
The sensations of equal stimulus are indicated for the sake of brevity as , and For the purpose of deriving the amount of fatigue existing in the retina after an exposure to some ray of spectrum , the percentage value of the different colours of the spectrum in terms of , and .
will be employed , and the following table is given to show the amounts of each in the different rays .
A column is added to give the proportion of .
to found in the colours tabulated .
Column I gives the spectrum standard scale numbers , II the wave-lengths , III , and the , and , and Column the ratio of .
to unity .
Fig. 1 shows diagrammatically the figures in the table .
Trichromatic Theory of Colour Vision .
shows the arrangement of the viewing stand .
A and are slits in spectrum apparatus , with the collecting lenses and in situ , patches of the same or different colours can be thrown on two white screens and D. ( If a is required on the screen , the is removed , and the whole is collected by , and its can be reduced by sectors or means .
) is a thin screen standon the table , which bears the two screens and , and is so pointed its prolongation would divide the between and equally .
This screen is of such a length that ray from the slit A will not be interrupted by it .
and show the of the eyes , the forehead resting on a pad fixed to for the sake of With this arrangement the right eye sees only , and left eye only C. In the position shown in the figure , the retina of the right eye is that which is to be fatigued .
To bring a fatiguing ray to the eye , a third spectrum , formed by the reflected white beam*of the apparatus coming from is employed , with a lens L The ray from the slit in this spectrum is projected to mirror ( the addition of which was suggested to me by Dr. Watson , who worked me in these observations ) .
The mirror reflects the beam the eye E To fatigue the whole pupil , a lens is inserted in the of the ray , which is there seen as a more or less bright patch of light works on a pivot , and can be rapidly turned against the side when fatiguing gives place to observation of D. When the eyes are in position , the distance apart of and is altered till the two patches to touch when observed , one by one eye and the other by the other .
When it is desired that the same colour shall be in each patch , the scale of A and spectra being known , this is readily effected , and a confirmation the correctness of the match can be obtained by throwing one patch over other , and obtaining two illuminated shadows side by side by the method .
The slits in each spectrum can be adjusted so that appear equally bright .
describing a set of observations made with a fatigued retina it is as a comparison light .
light can be much reduced by substituting a white surface for the , and this has been frequently used .
Sir W. de W. Abney .
well that the following should be borne in mind .
If the retina is white there is no change in its hue ; it merely becomes darker , at the light is well nigh extinguished .
In white light fatigue , it follows the sensations are equally fatigued .
Again , as the curves of equal to every kind of vision , normal and colour-blind ( the only difference the white they make ) , it follows that , supposing a match of some spectrum as seen by a colour-blind eye can be made ( by some means or with another colour by the normal eye , the ratio of the colour-blind stimulus sensations at the place in the spectrum which the erved must be the same as those of the normal eye , in that in the spectrum where the match was made .
The observation of the colour by both eyes independently , one being fatigued , is amode of registering such matches .
When a set of observations is to be made the eyes are placed in position as shown in the figure .
The fatiguing ray is thrown on the mirror and the right eye*is fatigued .
There is a time after which the fatiguing action and the action of ecovery from the fatigue practically balance one another .
When this staoe is reached the mirror is rapidly turned against and the two patches having the same colour inally are viewed together .
The slit of is quickly moved till the colour viewed by the left eye appears to approximate to that seen by the fatigued eye .
Four or five " " fatigues\ldquo ; are given by the mirror , and after each the match is made more closely correct .
When no alteration is found necessary , the observation is at an end and the position of the matching colour is read off on the scale .
If a match goes no further than the full green of the spectrum when a colour in the red or yellow of the spectrum is viewed by the fatigued eye a good match in hue can be made .
When a match , however , falls in the blue-green , or blue , or purple only the general hue of the fatigue colour can be matched , as the sensation colours in these parts of the spectrum are contaminated with the white .
A glance at fig. 1 will at once show how this arises .
When red is the fatiguing colour and is fairly bright and a green ray in the spectrum is observed it is sometimes difficult to form an exact match , as the fatigue colour is more saturated in hue than any colour seen by the unfatigued eye .
This question of the saturation of colour , it is hoped , will be treated of shortly .
For the present I must confine myself Go the method of measuring the fatigue .
A good illustration will be found in the following record of matches by eye of fatigue colours .
When we use the expression " " fatiguing an eye\ldquo ; it is intended always to fatiguing the retina .
The first point to call attention to in the above is that from S.S.N. 56 ( in which there is a small quantity of ) to S.S.N. , where there is no change in hue capable of being measured accurately , the matches are throughout lower in S.S.N. than the fatigue colour .
In Column VI , in Table I , it is seen that at that S.S.N. the ratio of red to green is at its minimum .
From this number to S.S.N. 60 the readings of the matches are always higher .
In both divisions of the spectrum , less red in the fatigue oolour is shown in its match .
This is a direct general confirmation of the truth of the percentage curves of the three sensations , and therefore of the luminosity curves from which they were derived .
Studying the observations in detail so far as is necessary , it will be found that the amount ( or factor ) of fatigue on the retina is a fixed one when the observations are made as described .
One will now be taken in which there is a large proportion of .
to , say S.S.N. .
The match to this colour is at S.S.N. and have for their sensation compositions 7822 ( 49.64 ) 5644 The only alteration made in S.S.N. when the fatigued eye observes it is a diminution of the .
; the .
remains tered .
In the match there is , of course , the normal proportion of .
to .
If , then , we make the .
of the fatigue and the match composition the same , we can directly compare the .
in the colour with its .
when the eye is not In the case in point , it happens that the .
in the match colour is Sir W. de W. Abney .
exactly half of that in the fatigue colour .
Its composition is equally expressed as 2822 .
The fatigue .
is therefore 28/ 78 of the normal in S.S.N. , and the factor is .
Taking S.S.N. , its match at S.S.N. 39 .
The compositions of these two are and or From this we see that the factor is of Tn the same way , taking S.S.N. 's and , having composition 49492 and , we obtain a factol of At we have a match with , with composition of and .
The factor for this , after deducting the white present in the fatigue colour , is At S.S.N. 56 , where the match is at S.S.N. , the compositions are and these we get a factor for .
of If we examine the match to S.S.N. , which is at S.S.N. , we have the following compositions and 164440 .
After the white in the fatigue colour from both , we have .
White . .
White .
and Treating these compositions in the same way .
we get as a factor This increase equires an explanation .
The .
to .
in the match colour is 1 to , and turning to Table I it will be seen that there is no colour which has so low a ratio , hence the eye has to do the best it can in finding a match .
The same is the case with the next two numbers .
In my paper on the " " Change in Hue of Spectrum Colours by the Addition of White Light , will be seen that from S.S.N. 36 to about 56 the blue which is added by the white light is ignored by the eye and the hue of a colour is judged by .
and .
only .
After S.S.N. 34 is passed towards the blue ) this no longer holds good .
It is the same also in the matching of the fatigue colours .
For instance , let us consider S.S.N. 32 , which matches itself .
composition is .
The average factor for the .
is 'Roy .
Soc. Proc 1909 , , vol. 83 .
Trichromatic Theory of Colour Vision .
multiply the 12 .
by this number we have .
in the fatigue There is therefore no change in it except a greater degree of in these two equations into sensation and white , we get The mixture of white and the green and blue sensations is paler in the first than in the econd equation , but the general hue would be the same .
This is the in all degrees of fatigue tried ; the match at S.S.N. is invariably that S.S.N. itself , the only difference being paleness , the differing but little in each equation .
One more match must be considered , that at S.S.N. , which , when it was a fatigue colour was matched at S.S.N. The composition of is , the in excess of the Taking as the factor we get ( nearly ) .
It is all used up in making the white S.S.N. has a composition of .
White .
or the .
being greater than the Reducing the last composition to make the .
the same as that of the fatigue colour we get The ratio of white to .
is not very different in the two , and the eye ignores the excess of .
in the first .
( It must be recollected that the luminosity of the .
is 190 times smaller than that shown as ) In this case the fatigue alters ths ratio of .
to .
to such a degree that the match lies on the other side of the second point of intersection of the red and green sensation curves , viz. , at S.S. When the factor of is much smaller , the fatigue colour has to be closer to S.S.N. before the match lies on the other side .
) I give a record of observations made with the fatiguing red reduced to of that used in the other record , and factor of red fatig ue in the same way as above is shown .
Sir W. de W. Abney .
Alongside of this record is one when fatigue was induced by a red lithium light eflected from a card which had a luminosity for the of two candles at 1 foot distance from the screen .
The factors derived this record are also given .
I have only taken , as an example , red fatigue .
I might have given examples reen and violet fatigues , but I believe it is more satisfactory to deal with a single kind of fatigue that other fatigues can be dealt with in exactly the same manner and the same kind of results will be obtained .
place in the spectrum for fatiguing the retina which is of reat theoretical interest is that of S.S.N. , where the .
and .
curves cut .
The only effect this fatigue hss is to heighten the blue curve , the proportion of red and green sensations remains the same .
In fatiguing the retina by any colour the general sensitiveness ( regardless of any particular sensation or sensations with which the fatigue is induced ) appears to be lowered .
Feeble luminosities are extinguished by the condition of the retina .
It is somewhat curious to note the advent of a colour which is quite ) arent and measurable by an unfatigued retina , but which has vanished when one which is fatigued looks for it .
As the fatigue lessens the colours gradually come into view .
It has been stated to myself that the retina has some sort sympathetic action with the fatigued one .
Personally , I can find no trace of such synlpathy .
It is satisfactory to know that after observations have been Trichromatic Theory of Colour Vision .
to match a fatigued colour , the addition of the fatiguing colour will give hue of the original colour to the unfatigued eye .
I now turn to another class of observation in which only the fatigued eye be employed .
If the fatigue can be reduced to a factor , as has been in the foregoing pages , it follows that the luminosity curve of the observed with a fatigued retina should be the same as that of a colour-blind person .
This is a direct method of getting the factor of fatigue Dr. Watson suggested .
His suggestion I have carried out , modifying the arrangement shown in fig. 2 to a certain extent .
Most of the description of fig. 2 may be read as applicable to this , Only one screen , , is employed , and the fatiguing colour is obtained from the second spectrum , B. On leaving the lens a mirror is inserted in the path of the rays issuing from the slit .
This is again reflected by another mirror on to the mirror on the screen .
The white beam ( which in the last set of observations was utilised to form a third spectrum ) now proceeds direct , and by placing a rod in the paths of rays coming from A and of the white beam , a patch of the colour is placed alongside a patch of white as in the ordinary measures of luminosity .
The eye is placed at and the retina can be fatigued as before , when the mirror ( or card ) is in the position shown .
When the fatigue induced is steady is turned back and the colour and the white are made of apparently the same luminosity , by means of an annulus*which I have described in other being placed in the white beam .
Before commencing fatiguing the eye , the luminosity of the spectrum may be taken in the usual way , or if more convenient the luminosity of the coloured ray on the screen may be measured by the unfatigued eye .
When the whole spectrum at close intervals has been measured by the fatigued eye and also by the unfatigued , the annulus readings are converted into luminosities , and from this two luminosity curves can be plotted and compared with each other .
The example I give will again be one dependent on red fatigue , the fatiguing colour being at the place of the red lithium line .
The hues of the fatigued colour and white are noted .
This is reaUy a screen , graduated from light to dark , through which the white has to ' Phil. Trans 1900 .
From these observations two curves were drawn .
The curve from the unfatigued eye is identical with that obtained previously when the source of light for forming the spectrum was the arc light , with the positive pole in a horizontal position , the current being 100 volts and 22 amperes .
In fig. 4 this curve with that of the adopted curve for the " " fatigue\ldquo ; luminosity is shown together with the readings in the preceding tabla ( Table ) .
We can from these curves ascertain the factor of red sensation fatigue as was done in my communication*to the Royal Society .
The first method of using the two curves is employed .
*Blind and the Trichromatic Theory of Colour Vision .
Part mplete Red or Green Blindness ' Prpc .
Roy .
Soc , vol. 84 , 1910 .
Sir W. de W. Abney .
When a colour-blind has to give a luminosity curve of the spectrum , is always a doubt as to amount of pigmentation of the yellow spot which possesses , and it is not sa fe to assume that any factor derived from an below 42 will agree with one obtained from higher S.S.N. 's .
In the case of fatigued eye by which the normal luminosity curve has been measured difficulty doss not arise , since the pigmentation is the same .
Commencing with two S.S.N. far apart , we may find a factor should hold good throughout .
Taking S.S. and 38 for the bwo points required by Method I in paper referred to , we get two equations , where is the factor which the curve has to be multiplied by to reduce it to the proper scale for the colour-blind ctlrve , and is the factor by which the sensation ( in this case the red sensation ) has to be multiplied .
The equation for S.S.N. is as follows:\mdash ; 21 .
The numbers of the left-hand member of the equation are the luminosities of the normal and fatigue blind shown on the curve and tables .
The right- hand member is the value of the red sensation at S.S.N. In a similar way the equations below are arrived at .
The two equations are For S.S.N. 58 S.S.N. 38 Solving the equation , we get By the method employed , is the deficiency of the sensation .
The factor of the sensation ( which we will call x ) is Taking S.S.N. and 36 , which are again far apart , we get the following equations:\mdash ; S.S.N. 56 . .
S.S.N. 36 This gives , and S.S.N. and 46 give and S.S.N. and 40 give , and of Colour Vision .
S.S.N. and 44 give and As a final pair of equations , we will take S.S.N. 's 52 and 42 , The mean of the sensation factors ( x ) is .
and the curve must be multiplied by to compare it with the normal curve .
The luminosity curve of the spectrum , as measured by an eye fatigued with pure red sensation , is the same as that of a red-blind eye , which has a factor 0 .
Other curves with larger or less factors of fatigue could be giyen , all of which would stand the same test as that just given .
The same applies to the retina fatigued with green and other colours .
But it may be stated that the luminosity curves obtained with eyes fatigued by any colour as far as the blue-green give indications that they can be sorted into two classes , red- and green-blind , as the sensation curves indicate , except in the case of the greenish-yellow ray at , where the curve retains the characteristics of the normal , modified slightly in the extreme red ( where only the red constituent is active ) , and in the blue end of the spectrum .

