I therefore commend to the Reader the now somewhat unfashionable hypothesis of Semper and his school, according to which the first preference for beauty of shape must be sought for in those arts like stone and metal work, pottery and weaving, which give opportunities for repetition, reduplication, hence rythm and symmetry, and whose material and technique produce what are called geometric patterns, meaning such as exist in two dimensions and do not imitate the shapes of real objects. This theory has been discredited by the discovery that very primitive and savage mankind possessed a kind of art of totally different nature, and which analogy with that of children suggests as earlier than that of pattern: the art which the ingenious hypothesis of Mr Henry Balfour derives from recognition of accidental resemblances between the shapes and stains of wood or stone and such creatures and objects as happen to be uppermost in the mind of the observer, who cuts or paints whatever may be needed to complete the likeness and enable others to perceive the suggestion. Whether or not this was its origin, there seems to have existed in earliest times such an art of a strictly representative kind, serving (like the spontaneous art of children) to evoke the idea of whatever was interesting to the craftsman and his clients, and doubtless practically to have some desirable magic effect upon the realities of things. But (to return to the hypothesis of the aesthetic primacy of geometric and non-representative art) it is certain that although such early representations occasionally attain marvellous life-likeness and anatomical correctness, yet they do not at first show any corresponding care for symmetrical and rythmical arrangement. The bisons and wild boars, for instance, of the Altamira cave frescoes, do indeed display vigour and beauty in the lines constituting them, proving that successful dealing with shape, even if appealing only to practical interest, inevitably calls forth the empathic imagination of the more gifted artists; but these marvellously drawn figures are all huddled together or scattered as out of a rag-bag; and, what is still more significant, they lack that insistence on the feet which not only suggests ground beneath them but, in so doing, furnishes a horizontal by which to start, measure and take the bearings of all other lines. These astonishing palaeolithic artists (and indeed the very earliest Egyptian and Greek ones) seem to have thought only of the living models and their present and future movements, and to have cared as little for lines and angles as the modern children whose drawings have been instructively compared with theirs by Levinstein and others. I therefore venture to suggest that such aesthetically essential attention to direction and composition must have been applied to representative art when its realistic figures were gradually incorporated into the patterns of the weaver and the potter. Such