COLOUR PERCEPTION IN FISH
(_Rewritten from "Land and Water," November 6, 1897_)
Many interesting problems constantly come before the fisherman, but
certainly one of the most interesting which has recently attracted his
attention is Sir Herbert Maxwell's theory on the power of fish to
discriminate between various colours.
His theory is, that though fish can undoubtedly discriminate between
different shades of light and dark, they cannot distinguish one colour
from another. The only conclusion that can be drawn from this theory is
the conclusion at which Sir Herbert Maxwell has apparently arrived. This
is, that if the same relations of light and shade be maintained in the
artificial which exist in the natural fly, the colour of the imitation
is quite immaterial.
The facts upon which he based this theory were (1) that during the
May-fly season he used several artificial May-flies, some of which were
coloured scarlet, some bright blue, and some coloured to imitate the
natural fly, all of them being similarly graduated with regard to the
shade of their various component parts; (2) that he caught trout with
all these flies, no particular one of them being decidedly more
successful than the others.
This experience of his no doubt would at first strike one as being very
strongly in favour of his theory; but on going deeply into the matter,
its bearing on the fish's powers of vision is not so great as it
appears.
To begin with, we must consider whether, judging from experience in the
past, trout have been known to rise at things on the water which were
not only unlike in colour to any flies on the water, but also unlike
them in shape and gradations of shade. This we know they will sometimes
do. I have on several occasions seen a trout which refused a fairly
accurate imitation of the flies which were on the water rise at and take
below the surface a swan's feather. There are also many other much more
extraordinary but similar cases on record. Thus, the fact that these
trout took an abnormally coloured fly is not a conclusive proof that
they mistook it for the natural fly, particularly as this experiment was
made during the May-fly season, when the trout sometimes appear to be
quite mad, but are at any rate always much less shy than at any other
time of the year.
The experiment, too, was made upon a private water, and I think that
there is great doubt that the same result would have occurred had it
been made upon a well-fished water where the trout were more shy and
better educated.
We must then consider whether, in what we know of the natural history of
fish, there are any facts which point towards the probability of their
being able to discriminate between different colours. Here we find that
there are cases in which in certain species the males are more
brilliantly coloured than the females, either at the spawning season or
always. This is probably a process in evolution which tends to make them
more attractive to the female. We also know that fish sometimes assume a
colour similar to their surroundings. This colour is, no doubt, evolved
for their protection from enemies, and surely a very large proportion of
these enemies are other and larger fish. Many of the larvae of water
insects and other creatures upon which fish feed are also coloured
according to their surroundings, in order to facilitate their
concealment. These facts would naturally lead us to come to a conclusion
opposed to that of Sir Herbert Maxwell, as the probabilities all point
towards the power of fish to discern various colours.
Another very important point is the structure of the fish's eye in
comparison with that of man, who we know has the power of discriminating
between colours. This power is, in the human eye, probably situated in
the layer of rods and cones of the retina. Had the fish's retina not
contained this layer, as is stated by Sir Herbert Maxwell, there would
certainly have been most excellent grounds for supposing that his theory
was true; but this layer _is_ contained in the fish's eye, though it is
not the same as in man. If the fish's eye did not contain it, fish would
have been totally blind.
How far this difference in the retina of the fish bears on its sense of
colour is, at present, a moot point, though I believe researches are
being made in this direction. At present, our knowledge is too limited
with regard to it for any definite statement to be made. The probability
is, that fish have the power of distinguishing colour from colour. A
probability, however, is not a certainty, though one is more inclined
towards it than towards an improbability.
Even should Sir Herbert Maxwell's theory prove true, in spite of
probabilities to the contrary, I do not see that we should have
progressed very much further with regard to facilities in imitating the
natural fly. We know that the relative values of light and shade in
various colours contiguous to each other, is not actually the same as
the impression conveyed to our eyes. We have an example of this always
with us in the photograph, where red and blue, in relation to each
other, certainly do not produce the same effects on the plate as they do
on the eye; and as we have no accurate knowledge as to the effect of
contiguous colours upon a normally monochromatic eye, we could hardly be
certain of producing an accurate monochromatic imitation of a
multi-coloured object, which would deceive that eye.
The case of a colour-blind human being is certainly not a normal case,
so the shade value of the various colours to this eye could hardly be
taken as a safe standard.
Even if we assumed that all these difficulties had been surmounted, and
that the exact relative shade values to this monochromatic eye of every
colour were estimated, I think that there can be no doubt that it would
be easier to imitate the colours, with the various shades in these
colours, than to calculate out the relative shade values of the
different colours, in one particular colour, and that the result of the
former and easier, would be much more likely to be accurate than the
latter and more difficult attempt.
Besides this, possibly, as the eyes of some families of fish are more
highly developed than those of others, the relative shade values of
colour might be different to the different families, so that if we
eliminate colours from our lures, we must have different shading for
different fish.
Having considered all these things carefully, I have come to the
conclusion that it will be much safer and easier to keep on using
colours in our imitations, even if we do present these imitations to a
monochromatic eye.
* * * * *
Since writing the above article, I have been able to collect some
further information with regard to the probable power of the trout's eye
to discriminate between colours.
These researches, though I have not yet had time to carry them as far as
I had hoped, have led me to believe more firmly than ever that I am
right in recommending the use of colours in our imitation flies. I have
prepared some sections of the retina of the trout, and examined them
carefully in comparison with the retinae of several other fish. A short
account of what is known at present of colour-vision is, I think,
advisable to make my meaning clear to those of my readers who may not be
sufficiently well versed in this particular subject.
The sensation of an individual colour is produced by rays of light of a
particular wave-length falling upon the retina. A sensation of "white"
is produced by rays containing all the wave-lengths which are able to
affect it. When, on looking at an object, we find that neither a colour
nor white sensation is produced, this sensation is called "black."
The white sensation may be mixed with the sensation of any colour of the
spectrum, as also may the sensation of black, and when these two are
mixed they produce a sensation of "grey." Some colours of the spectrum
are probably produced by a mixture of various wave-lengths of different
primary colours, and many colours in nature do not exist in the
spectrum.
The word "tone" expresses variations of wave-lengths within a named
colour, and "brightness" is used to indicate the intensity of the
sensation produced upon the retina.
The enormous difficulty of working out into a monochrome the
shade-values of a collection of colours, with several tones and shades
of brightness in each of the variously coloured parts of the object we
wish to imitate, can be imagined on considering these facts only; but
there are more facts which lead me to believe that to do this is not
only difficult, but impossible.
Two theories have been propounded to explain the sensation of colour
produced upon the retina.
The Young-Helmholtz theory teaches that there are three primary
sensations--red, green, and violet. Other colours are a mixture of these
sensations; white is produced when all three sensations are excited
together, and black is an absence of sensation.
Hering's theory is that there are six primary sensations arranged in
three pairs--white and black, red and green, and yellow and blue. He
assumes the existence of three visual substances which undergo metabolic
changes when subjected to the action of light. These are the red-green,
the yellow-blue, and the white-black substances. The white-black
substance is influenced by all the rays of the spectrum, while the
red-green and yellow-blue substances are differently influenced by rays
of different wave-lengths. When all the rays together fall upon the
retina, no metabolism takes place in the red-green and yellow-blue
substances, but only the white-black substance is affected. Thus the
white-black substance is the most active.
Any discussion as to the relative value of these theories would in this
work be out of place and unnecessary.
The ordinary form of colour-blindness in human beings is the inability
to discriminate between red and green. This shows that the visual power
of these people is dichromatic and not trichromatic, as their power is
limited to two colours, or pairs of colours, and does not extend to
three.
The individuals who belong to this class of the colour-blind may be
divided into two sub-classes--those who are red-blind and those who are
green-blind.
Those who are red-blind do not see the red end of the spectrum, and the
blue-green appears grey, though they have distinct colour vision of the
parts of the spectrum on either side of the blue-green. In matching red
with a green, they put a bright red with a dark green.
On the other hand, those who are green-blind see the red end of the
spectrum, while the green appears to them as grey. In matching a red
with a green they put a dark red with a bright green.
No absolutely undoubted cases of blue-yellow blindness have been
recorded, and only one of absolute colour-blindness; but one case is not
sufficient to go upon.
According to the Young-Helmholtz theory, a case in which only shades of
black and white were visible would be impossible, as it would not be
shades of black and white which would be seen, but shades of either red,
green or blue. According to Hering's theory, of course, absolute
colour-blindness would be possible.
In the normal human eye, only the central parts of the retina are
sensitive to colour, the peripheral parts are practically colour-blind.
Anaemia of the retina, which may be produced by pressure on the eye-ball,
will render the retina, first colour-blind and then insensitive to
light. To me it appears that colours in relation to each other assume a
grey tone, and the sensation of black and white disappears last.
The great difference which I have been able to observe between the human
retina and the retina of the trout is, that while the human retina
contains a layer of rods and cones, the retina of the trout only
contains cones, or if it does contain rods, contains very few, as I have
not found any as yet. There exists also at the back of the retina of the
trout a "tapetum," which extends over almost the whole of its posterior
surface. This does not exist in the human eye, but is found in the eyes
of some of the vertebrates. It consists of a layer of "guanin" crystals,
and, presenting as it does a metallic appearance, and having great power
of reflecting light, probably plays an important part in the visual
power of the trout, particularly, I should think, in a dim light.
The fact that the rods are absent from the trout's retina does not bear
the important significance that one would imagine on first realising it.
The fovea centralis of the human retina is the seat of most acute
vision, and in the fovea centralis there are no rods. The cones in the
retina of the trout are very closely arranged, so that they are
practically in contact with each other, and their outer limbs are rather
longer and finer than in the case of man. This layer of cones extends to
the periphery of the retina, and the cones are just as closely arranged
as far as they extend. These facts should lead us to believe that the
vision of the trout is probably extremely acute, in fact, as we find in
the retina of the trout, no material difference from the _fovea
centralis_ of the human retina, we have no reason to suppose that the
visual powers of the _whole_ of the retina of the trout, should differ
in any way from the visual powers possessed by the _fovea centralis_,
the seat of most acute vision both as to colour and light in the human
retina. The retinae of other fishes which I have examined (none of them
were _Salmonidae_) contained only cones; but these cones were some
distance from each other.
The layer of pigment epithelium which is present in the human eye, is
present also in that of the trout. It occupies the same position between
the layer of rods and cones, or cones only, and the choroid. As in the
human eye, it adheres sometimes to the choroid and sometimes to the
retina, when the retina is removed, though perhaps it most often adheres
to the retina.
My space is too limited to enter into any of the theories as to the
possibility of the pigment cells playing a part in colour vision. It is
quite sufficient to state that they undoubtedly do play some part in our
sense of sight, and that they are contained in the eye of the trout.
The retina of a colour-blind person does not show any organic difference
from the normal eye, so we cannot say to what cause colour-blindness is
due; but so far as our knowledge goes, there is no reason to suppose
that the trout is normally colour-blind.
As Michael Foster so ably put it, "No man can tell what are the
sensations of his fellow-man," still less I think can man say what are
the sensations of a trout. All we can do with regard to this question of
colour vision, is to find out all the facts we can relating to it, and
working on comparisons, arrive, not at conclusions, but at
probabilities.
The only thing of which I am sure is that we shall find it safe and
comparatively easy to imitate flies in colours, but to make a
monochromatic imitation of one, which would accurately represent it to a
normally monochromatic eye (about which we know nothing), in a medium of
which we know very little, is practically impossible.