The disturbances which have shaken London appear to have been chiefly east and west, or along the valley of the Thames. In South America the line of disturbance is along the western sides of the Andes. Another line is along the northern coast of the continent through Andalusia and Caraccas towards the Antilles and Trinidad. The shocks of the Pyrenees are chiefly felt along the southern side of these mountains. In the middle and on the northern side they are but seldom felt. This propagation in lines or zones may in certain cases be apparent rather than real. Thus the north and south ranges of mountains in Japan are mountains almost simultaneously shaken along their eastern flanks, giving the impression that an earthquake had originated simultaneously from a fissure parallel to this line, or else, starting at one end, had run down their lengths. Time observations have, however, shown that such disturbances had their origin at some distance in the ocean, and, travelling inwards, had reached all points on the flanks of these mountains almost simultaneously. The same explanation will probably hold for the so-called linear disturbances of western South America. All earthquake disturbances have probably a tendency to radiate from their source, and are only prevented from doing so by meeting with heavy mountainous districts, which by their mass and structure absorb the energy communicated to them. Much energy is also lost by emergence on the open flanks of a range of mountains. Rather than say that high mountains often bound the extension of an earthquake, or that earthquakes appear to run along the flanks of such mountains, we might say that earthquakes have boundaries parallel to the strike of the rocks in a given district, that such a direction is the one in which the propagation is the easier. Rossi is of opinion that volcanic fractures play an important part in governing the distribution of seismic disturbances. When a volcano is formed, a series of starlike fractures are formed round the central crater. Secondary craters may indicate the line of these fissures. The mountains about Rome are regarded as typical of this radial structure. The more distant the secondary craters are from the centre of the system, the smaller will they be, and also the younger. If two fissures intersect we get a larger crater at the junction. Earthquakes are propagated along the direction of these fissures, whilst the rising and falling of these lips throw off transverse waves. Rossi adduces observations which appear to meet with explanation on such suppositions.