green and the green-violet junctions are situated at points occupying positions nearer the violet end of the spectrum than the corresponding junctions of those with a spectrum of normal length. The red-green junction, therefore, in a case of this kind, will be situated in the yellow-green instead of in the yellow as it is in three-unit cases with an unshortened spectrum.

On referring to the details of the cases which I have given, it will be found that these theoretical predictions are verified in every particular.

CHAPTER V.

THE FUNCTION OF THE RETINA IN THE PERCEPTION OF COLOUR.

The Structure of the Retina.-The retina may be conveniently divided into ten layers.

1. The Pigmentary Layer. Many writers describe this layer as part of the choroid, but the study of its development shows that it forms part of the retina. It consists of a single layer of hexagonal nucleated cells. These cells are pigmented, and have processes which extend into the layers formed by the rods and cones.

2. The Rods and Cones.

Each of these bodies consists

The rods are The cones are

of an outer and an inner segment. Their long axis is perpendicular to the surface of the retina. cylindrical and nearly of uniform size. flask-shaped, the smaller end of the flask pointing to the external surface of the retina. At the yellow spot there are no rods.

3. The External Limiting Membrane. This is formed by the expansion of the external extremities of the fibres of Müller. These fibres of Müller form the connective-tissue framework which supports the various layers of the retina.

4. The Outer Granular Layer. This layer is made up of nucleated oval cells which are connected with the rods and cones on the one hand and the cells of the inner granular layer on the other. The cone granules-that is,

those cells which are in connection with the cones- are situated nearer the external limiting membrane than the rod granules.

5. The Outer Molecular Layer. This consists of molecules and fine fibrillæ.

6. The Inner Granular Layer. This layer chiefly consists of nerve-cells which are bipolar and nucleated. The outer processes anastomose with fibres from the rod and cone granules. The other process is lost in the inner molecular layer.

7. The Inner Molecular Layer. This consists of molecules and fine fibrillæ, being similar in character to the outer molecular layer.

8. The Layer of Ganglion Cells. These are multipolar cells. They are connected with the axis cylinders of the optic nerve-fibres, and externally with the inner molecular layer.

9. The Nerve-fibre Layer. This consists of the expansion of the fibres of the optic nerve.

10. The Internal Limiting Membrane. This separates the nerve fibres from the vitreous humour. It is intimately connected with the fibres of Müller.

The Physiology of the Retina.-The rods and cones of the retina are practically the terminations of the optic nerve, and directly sensitive to the influence of light. It is not known what are the relative functions of the rods

and cones. It is stated that the rods cannot be necessary for colour vision, because they are absent at the yellow spot, the region in which form and colour vision are most distinct.

The ordinarily accepted theory with regard to the conversion of light undulations into a sensation, is that light passes through the retina, and is reflected back from the choroid. The undulations on their return through

the retina influence the rods and cones. The effect of this influence is a molecular movement which is transmitted to the brain, and becomes evident as a sensation of light.

It is difficult to understand how it is that the light waves do not influence the rods and cones in their first passage through the retina.

In the outer limbs of the rods a purple substance is found which is sensitive to light. This purple is sensitive to monochromatic as well as to white light. It is bleached most rapidly by the greenish yellow rays, those to the blue side of these coming next, the least active being the red. This visual purple is found exclusively in the rods. Under the action of light the visual purple first becomes yellow and then colourless.

The chief arguments which have been used against the view that the visual purple is an essential element in vision are its conspicuous absence from the cones, and frogs whose retinas have been bleached by exposure to light appear to be as sensitive to colour as other frogs.

I am inclined to think that this visual purple is one of the essential elements in vision, and that the process might take place in the following manner. That light acting upon the visual purple causes its liberation from the rods. This, becoming diffused at the back of the retina, forms an actual photograph of the objects included in the visual field. The function of the cones might be that of conveying this impression to the brain. The absence of rods from the yellow spot might be explained by the assumption that the visual purple which is liberated from the rods surrounding the yellow spot is diffused into this spot in sufficient quantity for distinct vision, and that if there were rods in the yellow spot the quantity liberated would be too great, and would interfere with distinct vision.

This view is supported by the following experiments,

which I instituted on purpose. If we look at a gas flame through a pin hole in a sheet of paper, the flame appears to have lost the greater portion of its luminosity. If we remove the paper with the pin-hole further from the eye, and so that light from adjacent objects can enter the retina, then the gas flame again resumes its original luminosity. For the purpose of experimenting in this way, I obtained a piece of black velvet, and, having cut a square hole in the centre, I pasted a piece of cardboard on the velvet over the hole. The internal surface of this cardboard I blackened with ink. I then pricked a pinhole in the card from within outwards. When this hole is close to the eye, we can see not only the gas flame, but the whole of the gas-globe. Now, the eye is constituted so that any given portion of the retina can only receive light from one object in the visual field. Therefore, whether we are looking at the gas flame directly, or through a pinhole, the portion of retina upon which the image of the gas flame falls receives nearly the same amount of light. When we are looking directly at the gas flame, the image of this flame falls directly on the yellow spot. Now let other objects be viewed through this pinhole, and it will be found that unless an object be brightly illuminated, it will not be visible at all. On looking at the fire the yellow flames can be seen of greatly diminished luminosity, but the red glow of the fire is not visible at all. We could understand how a partial diminution in luminosity might be produced by cutting off some of the rays from the retina, namely those which under ordinary circumstances would fall on the peripheral portions of the cornea, but I do not think that this could account for the very great diminution in luminosity which occurs, and the fact that dimly illuminated surfaces and the red glow of the fire are not visible at all.

E