colours. Hering says that the black-white substance is most acted upon by the yellow rays of the spectrum. The red-green substance is acted upon by three parts of the spectrum, red at each end, and green in the middle. The blue-yellow substance is acted upon by the whole of the spectrum with the exception of the green. It will be seen that spectral red acts upon the blue-yellow substance as well as upon the red-green substance, and that spectral violet also acts upon both substances. The only spectral colour which is capable of giving rise to a pure sensation of colour, according to this theory, is green.

1. Perception of Shade of Colour-blind Persons.-The arguments which were used with reference to the YoungHelmholtz theory equally apply to this theory, substituting 'chemical substance' for 'set of nerve-fibres.' When we find that there is no alteration in the perception of shade there can be no loss of importance of any light-perceiving substance.

2. Spectral Examination of the Colour-blind.-According to the theory a redgreen-blind should see spectral red and orange as yellow, and spectral green as white, but a redgreen-blind sees red, orange, yellow, and green as colours which in no way differ from yellow. Also, if a redgreen-blind person saw spectral red and orange as yellow, and spectral green as white, he would not confuse red and green. But he does confuse red and green, therefore the theory is fallacious.

According to the theory a blue-yellow blind should see spectral red and orange as red; spectral yellow and blue as white and black; spectral green as green, and spectral violet as red. I have not met with a colour-blind person who has seen the spectrum in the above way; and I am not aware of any recorded case in which spectral blue has been seen as black. From this it is seen that the theory

is inconsistent with the established facts of colour-blind

ness.

3. Complementary Colours.-There is no evidence to show that red and green, and blue and yellow are opposed. White and black may be opposed, and each of the colours may be opposed to black, which is the negative of white and of all colours. The simple colours are not complementary as Hering states, thus the complementary of red is bluish green, not green. The antagonistic colours of Hering do not annihilate each other, but combine into a white manifestly more intense than either of the colours.

It is unnecessary to discuss any further objections to this theory.

It may be well to consider the restrictions which facts. put upon theory in connection with colour. From the fact that many colour-blind persons are able to match colours with regard to shade with perfect accuracy it is evident that there can be no loss of any light-perceiving portion of the retina. It is therefore evident that any theory which assumes this loss must be a fallacy. A retinal theory, to be consistent with this fact, must assume the existence of separate organs in the retina for the perception of colour and light. There are two separate organs in the retinathe rods and the cones; but then the rods are absent from the fovea centralis where the perception of light and colour are both most distinct. Then, as colour is not a physical property of bodies, it is very difficult to comprehend how it could exist apart from light. It is easy enough to understand how shortening of the spectrum might be due to some retinal defect, because, in this case, a certain number of rays are not perceived, and there is light as well as colour loss.

Deductions made from a few cases are useless, because

the essential facts cannot be separated from the individual details; thus I am not aware that any previous observer has separated shortening of the spectrum from colourblindness due to defective perception, and until this is done the facts are very difficult to explain. Then, the colour may be modified by the media of the eye. From this it will be seen that it is necessary to obtain the details of a large number of cases, and note those facts which are common to all.

CHAPTER X.

THE CLASS OF THE SEVEN-UNIT.

I HAVE described the colour-perception of the normalsighted, but there still remains a class who have a colourperception different from the normal-sighted, and yet are not colour-blind. These persons see seven colours in the spectrum, and the division is different from that of the normal-sighted. See Fig. 1 of the frontispiece. The percentage of these persons is very small. There is probably not more than one in every two or three thousand. I have met with three cases. One was a mechanic, who was pointed out to me as colour-blind because he had persisted in calling a colour blue which his fellow-workmen all agreed in calling a violet. I examined him with the spectrum, and he marked it out as in Fig. 1. He persisted in saying that there was a dark blue visible between the blue and the violet, though I could see no definite colour at this point, neither was the spectrum particularly dark. This is, in all probability, how Newton marked out seven colours in the spectrum. He tells us that he and a friend of his, who was noted for having a particularly good eye for colour, obtained similar results. Newton would not have been likely to have marked out a colour which he could not see.

The seven-unit also see the orange band of a larger size than the normal-sighted.

Many become aware that they differ from the normal in this respect. Thus, when I was testing a man who possessed very accurate colour-perception, and whose occupation consisted in matching and arranging colours, I picked out a violet and asked him what colour it was. He said, "I should call it a blue; some people would call it a violet." I then picked up a very dark blue, and said, "Would you put it with this?" "No," he replied, "I should not; it is a distinct colour. People seem to class so many colours under one name.'

All the persons I have found belonging to this class have been very fond of colours. The first individual whom I came across was an amateur artist, who, his employer told me, painted very well.

It is perfectly certain that individuals belonging to this class see more in colour than the normal-sighted. They can match colours with greater ease, and have a better memory for varieties of colour. They can also recognize differences in colour which are not perceptible to the normal-sighted. This we should expect, as their absolute psycho-physical units are smaller. In order to prove this increased power of colour-perception I constructed a test in the following way: I painted a ground of Prussian blue, and then, when the colour was dry, I put an after-tint of another colour over it, keeping a careful record of what I had done. In most instances I put the after-tint only over half of the disc. On looking at these coloured discs I am unable to tell, without referring to my note-book, the colour of the after-tint.

I took these discs to a well-known artist, but he failed to tell the colour of the after-tints. A pupil of his, however, a boy of fourteen, looked at the discs, and told the colour of the after-tint correctly in each case. He apologized for seeing the colours like this (because his