in the outer and inner rings, in which are repeated the same devices with slight variations. In my opinion this scroll forms a very essential part of the pattern, and I call special attention to it, inasmuch as it completes the interpretation of the entire design; for I venture to repeat what was stated at the beginning, that every pattern represents ultimately some definite object. I even believe that we have here the germ of the conventional scroll so common as an ornament upon Greek vases, in combination with other well-known devices that still hold their own as stock patterns in this twentieth
century. Of these I give a typical example from a vase in the Berlin Museum (Fig. 16).
Here we have a mere modification of the so-called Acroterion repeated over and over again as in Figs. 14 and 15, together with a modification of the scroll twining itself in all directions. It is easy to understand how the idea of the Acroterion was developed out of the twisting scroll of tentacles, as shown in those figures. We may thus readily perceive how it might become still more conventionalised as a decorative ornamental pattern, such as that I produce from Olympia (Fig. 17). But in every one of these ornaments we cannot fail to note the per-