VII
COMPOSITION
There are so many millions of good compositions in the world that it seems strange any one should ever waste time on a bad one. The good ones lie about us at every turn of the road. All that is necessary is the eye to see them. There are no fixed and immutable laws of composition at least, none that cannot frequently be broken to advantage by a man of genius. All of the old conventional rules are explanatory rather than constructive. They may prevent an utterly bad arrangement, but they can hardly enable us to create a masterpiece; for the all-essential note of personality would be absent. In my own opinion, about all of the rules of composition which are of any practical value to a painter, are negative rather than positive, and can best be expressed in a series of "don'ts."
The first and by far the most important of these is, "don't try to say two things on one canvas." Any motive that is worth painting must have a central point of interest. Concentrate on that and sacrifice everything else to it. If there chance to be another attractive feature in the same subject, ruthlessly suppress it, in order that the one thing which you have to say may be said strongly. It often happens in nature that there are two points of nearly equal interest in the same scene. In this case divide the motive into two separate pictures, or else paint some other motive. If you try to paint both on the same canvas you will fall between two stools; 
H. W. Ranger—"Landscape"
By permission of E. S. Bennett
for the human mind is capable of receiving but one impression at a time. An instance of this double motive which recurs constantly in nature is the scene where some handsome landscape is reflected in a pool or stream, the reflection being often more beautiful than the scene which it reflects. It would be fatal to attempt to reproduce both in one picture. The eye of the spectator would not know upon which of the two pictures to rest and neither would make its full impression.
An excellent example of the correct way to treat this motive is to be found in the river views of the Norwegian painter, Fritz Thaulow, who never gives more of the landscape itself than a suggestion at the top of the picture, thus concentrating the attention on the beautiful swirling expanse of water below. The water itself tells all that is needful of the thing it reflects, and the attention is not distracted in the effort to see two things at once.
I have seen many a poor picture in which two very excellent pictures had been painted upon the same canvas, either of which would have been beautiful by itself. If you wish your message to carry, don't confuse your audience with irrelevancies. Make your single statement clear and forceful and convincing—and let it stand by itself. Don't try to give too much for the money. This is even a worse mistake in art than it is in business.
Secondly. "Don't divide your picture into spaces of equal size and proportion." For some psychological reason of which we have not the explanation, the human mind abhors an equal division of space in a picture. Therefore don't put either your horizon line or your principal object of interest in the exact centre of the canvas. How far above or how far below, the centre the horizon should be placed, will of course depend upon the character of the motive and its various units. Unless there is some very convincing reason for the high horizon, however, all experience points to the lower division as best. A vast sky always lends nobility to a picture; while the suppression or nearly total elimination of the sky tends to convert the canvas into a sort of transcendent still-life. This is the case with the water pictures of Thaulow. They are the very apotheosis of still-life, it is true, but they are held within the still-life class by the fact that they are a representation of near-by objects, that they make no appeal to the infinite—translate no mood or effect.
The low horizon line is peculiarly essential when the principal motive of the picture is found in the sky itself—some vast composition of rolling clouds, some gorgeous sunburst radiating its luminous streamers athwart the canvas, some castle in the air towering up and up to the zenith. In this case, a mere line of land is often sufficient—enough to give the dark and solid value that lends light and air to the upper reaches of the sky.
"Don't have anything in the picture which does not explain itself." Because a thing happens to exist in nature is no reason why it should be allowed a place in your picture—which is a work of art. Treat nature with respect and affection, but don't let her rule you. And, moreover, don't paint any motive that is so unusual and outré that it will not explain itself without a pamphlet attached to the frame. I once asked Mr. Lhermitte, the veteran French master, what he proposed to call an important picture which he had just then completed for the Salon. "I don't know," he replied. "A picture which needs a title should never have been painted. What would you call it yourself?" We had best not poach upon the preserves of the story-teller, because he can always beat us at his own game. No beauty was added to a certain picture of the Cornish coast which I once saw in the Royal Academy, by the fact that it was entitled "Where the Phœnicians came for tin."
"Don't repeat the main line of your picture with another important line parallel to it." If you have a mountain form swinging up to the left, have your clouds swing up to the right; or tend in that direction. If you are painting in a flat country like Holland, and your horizon line is forcedly horizontal, make this straight line beautiful by adjusting the cloud forms to it in agreeable contrast. The sky is in this respect a wonderful resource to the painter, for its lines may sweep in any one of an hundred different directions; and they can thus always be made to balance or accentuate or modify the lines of the solid earth, which cannot change.
Above all, "don't let the dominant line of your picture end aimlessly in mid-air." With the sky to help, there is no excuse for this. It should be picked up and carried on in a sinuous, living line, like the sweep of a winding brook or the curve of a mountain road. The psychological effect of this living line in a picture is one of the most potent, though one of the most mysterious, things in art.
As I have already said, however, there is not one of these rules, nor one of the old conventional tenets, that cannot occasionally be disregarded to advantage. No! in this I am mistaken. There is one rule at least which must never be broken—the rule which says "thou shalt not paint two pictures upon one canvas"; for the house which is divided against itself inevitably falls to the ground.
But I have seen an excellent picture in which the horizon line bisected the canvas exactly in the centre—the necessary balance being achieved by other means. I have also seen pictures in which the repetition of the dominant line added a strange beauty to the canvas.
"Don't crowd your composition." Let your tree or your mountain have breathing space. Keep them away from the edge of the frame. They will gain in dignity and apparent bigness by diminishing rather than increasing their proportions.
"Don't put in a single unnecessary feature." Everything which does not contribute to the grace, or the beauty, or the force, or the sentiment of your picture detracts from it.
But unquestionably the best rule of all is to keep the eyes always wide open and observant of the things about you, for the most beautiful compositions in the world are always the daring and unexpected arrangements of nature. It behooves us to see them.