It will be seen, on looking at Fig. 3 (Diagram No. 222), that the vase might have been thus tilted upon its own base at the joint , to the extent of nearly 22° 30′, before it would have lost the power of recovering its position, and that when tilted through part of this angle, the effect of the throw of the second shock delivered to it through the base, would tend to increase the rotation commenced, as well as to diminish the extent of horizontal range, or velocity impressed.
The general relation of all these objects as to azimuth, is shown in Fig. 223.
I found here also, as at Polla, an example of that singular circumstance, the keystone or block, of a cut stone semicircular arch over a doorway, which had worked up in place of coming down (by the movements of the earthquake) between the remaining arch stones. This case, of which the (original) Sketch No. 224 is a true representation, (though very imperfectly given by the woodcut,) is obviously due to the rocking to and fro of the whole wall in the plane of the arch, the motion being several times repeated, and hence the alternate partial freeing and gripping of the keystone, at and , and , between the rocking voussoirs, which, moving on the lowest point of the jambs, or at the springing level, are thus at each oscillation, relatively higher and lower than each other, at and at . The
