We think we have made it perfectly clear that it is only in exceptional circumstances that the fish can see its prey against the brightness of the sky. The normal case is only where the background is the bed of the river, the reflection of that bed or a blackness, depending more or less on the mass of water between the fish and the object. The object itself is invariably lighted by rays from the surface, which rays are reflected to the eye of the fish. The amount of light reaching the fish depends on the depth of the object under the surface and on its distance from the fish as well as on the clearness of the water. The fish has, it is true, an eye suitable for a weak light, but if the fish be colour-blind, and the object be of the same tone and relative lighting as the background, how can the fish perceive the object? A pike will rush twenty feet at a fly in a piece of water only three feet deep; he cannot have seen the fly against the sky. That could happen only if the pike were at the bottom and the object not more than five feet from him. Whereas, in the case supposed the fish is near the surface and the fly is seen by reflected light. The form is probably not seen distinctly but the “colour” is certainly attractive. It is conceivable that various states of light, and various states of water favour a particular colour and make it more alluring—at one time black, at another blue, or yellow, or red, or even white. We do not assert that a fish can see colour as we see it, but we hold that a fish can hardly be colour-blind when we consider the conditions under which it lives and moves. For the salmon to see its prey against the brightness of the sky is, as we have shown, exceptional, so that the argument which insists on the uselessness of colour to a creature which sees only dark silhouettes falls to the ground. We conclude that the salmon has a colour perception; whether the scale is like that of human beings we cannot say, but it is enough that the fish recognises one colour as different from another. Further than this we cannot go as yet, but to that conclusion we think we are certainly entitled.
J. Allan Stewart, M.A.
In the case of any of the lower animals an enquiry regarding its power of appreciating different colours is usually conducted by making experiments calculated to show whether or not the animal in question behaves as if it had this power. Whether it actually sees colours as we do, or is even conscious of seeing them at all, is, necessarily, beside the question.
So far as this method of investigation carries us in the case of the salmon, there would seem, from what is mentioned elsewhere, to be trustworthy evidence that it is influenced by differences in the colour of the artificial fly. But there remains a considerable difference of opinion, I understand, as to the value of these practical observations.
That the different colours and light intensities of an object, situated within a definite area of the water surface, are theoretically visible by the fish’s eye, has been clearly shown by Mr. J. Allan Stewart, who has thus answered those who argued that any body on the surface of the water would only be seen as a dark object against the bright sky-ground.
The fact of an animal choosing one part of the spectrum in preference to another can only be accepted as evidence that it recognises some difference. Now, it is only from the facts known regarding human vision that we can draw comparative conclusions. Judging from these facts we find that certain parts of the spectrum affect us differently from other parts, not only by virtue of their colour, but also of their brightness. We find, indeed, that we have two distinct visual sensations dependent on light, colour and brightness, both influenced by differences in wave-rapidities or wave-lengths. Thus, if we gradually reduce the intensity of the spectrum, the colours all finally cease to be recognisable. In this colourless spectrum the brightest part is slightly to the violet side of the line E (corresponding to green), so that it is quite possible that this part of the spectrum would appear distinct from the rest, even in the case of an animal destitute of all colour perception. (In proof of this it may be mentioned that in a case of total congenital colour-blindness in man, this part of the spectrum is found to be recognised as different from the rest by its brightness alone.)
From arguments which I have adduced elsewhere, there is good evidence in favour of the view that light perception (brightness) is the more primitive of these two sensations, and that it is dependent upon changes induced by light in the retinal pigment epithelium. Also, it is practically proved, so far as the human retina is concerned, that this light impression is communicated to the rod-cells of our retinal neuro-epithelium and so transmitted by conducting channels in retina, nerve and ganglion, to the sight centre in the brain. In the case, therefore, of an animal with a low form of retina, in which the above-named cells are alone represented, any differences in its behaviour on exposure to different parts of the spectrum may be presumed to be explained by the amount of brightness or light effect thereby induced.
But, on the other hand, we now believe that the finer gradations of differences in stimulation due to variations in wave-length which we recognise as colours are dependent on the presence of another kind of retinal element called cone-cells. It is sufficient here merely to say that these cone-cells seem to derive the initial stimulus through some effect produced on the pigment epithelium, and that they also transmit the impulse by conducting paths to the brain in a similar fashion to that above-mentioned.
Accordingly, should the retina of an animal contain, not only pigment and rod-cells, but also cone-cells, we must admit that, so far as we know, it is anatomically and physiologically capable of being influenced by colour in a similar way to the human retina. Now, the retina of the salmon does undoubtedly possess all these anatomical elements—pigment epithelium, fine rod-cells, well-marked cones.