SYMMETRY AND BALANCE

187

be answered as in Fig. 33b. Crane says: "One cannot draw a line or define a form without demanding an answer that is, a corresponding, re-echoing line or mass." It is interesting to notice the analogy between this con

ception and the principle of musical design which we met as "antecedent and consequent."

Balance of Interest. In many pictures which show but little formal symmetry the arrangement of elements is such that there is a virtual symmetry. Puffer has called this a "substitutional symmetry," and for the sake of illustrating the principle she has offered the following classification of the items of "weight" in a picture. She says we may have (1) mass, (2) depth or vista, (3) direction of line, or of motion, or of attention (e.g. the direction in which a person in the picture is glancing), and (4) interest. In good pictures one will probably find an equation in which two of these items are set over against the other two, unless it happens that one item is extraordinarily strong, and in this case it will be balanced by the other three. In a portrait, for instance, if the mass of the person's form is on one side of the canvas together with some interesting object like a flower or an animal, one would expect to find on the opposite side some vista or depth, or direction of line. In "substitutional"

symmetry one finds that a small, interesting object may balance a larger one of lesser interest.

Although symmetry of form and symmetry of interest are agreeable features of pictorial composition, they are neither of them absolutely essential. A strikingly unbalanced picture would probably always be uncomfortable and displeasing, but in some compositions the thought of, or the feeling for, symmetry seems not to come up at all. In Watts's portrait of Ellen Terry, for example, nearly everything of interest, as well as most of the mass and direction of attention, are far out on the left side, and in some Japanese prints it would be far-fetched, if not impossible, to point out any bilateral symmetry. In these cases there is no felt lack of balance, the conception of balance is merely irrelevant.

Balance on the Vertical Axis. Another phase of the problem of balance is the distribution of masses and spaces between the upper and lower parts of a composition. An arrangement may be symmetrical in its right and left halves, but wholly unsymmetrical as between upper and lower halves. In general, to prevent top-heaviness and give, as it were, enough ballast to a composition, there should be more below the center than above it. Pierce's experiments show that the principle of stability is of even more moment than that of left and right balance. An inverted pyramid would be an unpleasant and precarious-looking structure. The visible sign of a sure equilibrium is breadth of base, and most massive things are built to slope by more or less obvious degrees toward their tops. It is not true, though, that all beautiful and well-poised forms are

VERTICAL BALANCE

189

larger at the bottom; very good effects are sometimes secured by putting the mass of the thing represented near the upper limit of a picture. A mass of graceful flowers may fill the upper part, with only their slender stalks below; a drift of clouds or a flight of birds may be shown high up in a picture, with only a few faint landscape lines below, the nearest possible approach to empty space. Why do not such pictures look as topheavy and unstable as the inverted pyramid? The reason is that they represent things which are not dead, inanimate weights, but are delicate and light. Placing the flowers or clouds or birds above the center of the picture, with the empty space below, is just what suits their character, and brings out their lightness and buoyancy. These two facts, then, are part of the same. truth: to gain stability, large masses must lie below the center, and this is appropriate when the masses are supposed to be heavy; to gain freedom and buoyancy, masses may lie above the center, and this is appropriate when the masses represent something light. It would be interesting to have experimental evidence on this question of how far the ideational element may modify our feeling of "weight" in a picture.

Central versus Axial Balance. In Fig. 34 there is a balance on the horizontal axis between 1 and 4 on the left as against 2 and 3 on the right. On the vertical axis 1 and 2 balance with 4 and 3. In distinction from these two types of balance on an axis, there is also a relationship of balance between 1 and 3 and between 2 and 4. Ross designates this as balance around a center, or balance of double inversion. It is evident that if I

were revolved around the vertical axis it would coincide with 2, and if then revolved around the horizontal it would coincide with 3. An illustration of this kind of balance is apparent in Hokusai's wave (Fig. 35).

1

2

4

3

FIG. 34.

This wave picture illustrates about all the points that we have made on visual form. The sides of the inclosing rectangle are as 3: 2 (the proportion which Fechner places next to the golden section in

beauty). The two most prominent lines, as well as some of the subordinate ones, are serpentine in form. There

X

FIG. 35.

Is a substitutional symmetry in the balance between the mass of the wave on the left and the effect of aerial depth which (in the colored print) is brought out on

OPTICAL ILLUSIONS

191

the right. There is also a shock of opposed forces visible in the water sliding down from the right, but tossed back from the left. The buoyancy of the wave is increased by the fact that the crest is so near the upper edge of the picture. There is fine repetition in the serpentine lines, and, finally, there is the relationship of double inverted balance between the two main lines x and y.

Effect of Optical Illusions. It is important to the artist to recognize the character of certain optical illusions in order that he may know either how to make use of them in his work, or how to compensate for them. Among the most common and patent of such illusions are the following: (1) There is a (1) There is a tendency to overestimate vertical as compared with horizontal distance. Hence a figure which really is square looks a trifle taller than it looks broad, and in order to get a figure to look square the designer must make it a little broader than the geometrical square. Experiment shows that the apparent squares are esthetically more pleasing than the figures which measure square but do not look it. (2) There is also a tendency to overestimate size in the upper part of the field of vision. The letter S looks as if its upper and lower parts were about equal in size; but when it is inverted we can see that we must ordinarily overestimate the part that belongs above. (3) This exaggeration of the upper part of a figure is sometimes reversed by other illusions. Thus in Fig. 36a 1 looks slightly larger than 2, though they are the same size, but in the second arrangement, Fig. 36b, I seems rather smaller than 2, though again they are the same. (4) The