ON SEISMOLOGICAL INVESTIGATION. 73 



cannot, however, be so regarded. Between these extremely large rein- 

 forcements it must not be overlooked that there are others of less magni- 

 tude separated by intervals of from two to four minutes. 



All that we can conclude from an inspection of the above table is 

 that after all sensible motion of a large earthquake has ceased horizontal 

 pendulums, whether they are situated near to its origin or at a great 

 distance from the same, indicate that the earth waves at intervals of from 

 two to six minutes show marked increments in amplitude. The earth- 

 quake does not die out gradually, but by surgings. In its latter stages, 

 for intervals of one or two minutes, the ground may be entirely at rest, 

 after which movement recommences. This alternation of rest and move- 

 ment may be repeated many times. 



If it can be admitted that large earthquakes result from the collapse 

 of ill-supported portions of the earth's crust upon a more or less plastic 

 layer beneath, it may be imagined that rest is attained by a series of more 

 or less regular surgings, which are propagated to distant places to disturb 

 horizontal pendulums in the way observed, 



11. The Nature of Large Waves. 



To explain the existence of the large waves of earthquakes we are 

 at present left to choose between two hypotheses. One is that the 

 large waves of earthquakes are disturbances travelling partly under the 

 influence of gravity over the surface of our earth, and the latter that they 

 represent the outcrop of distortional waves passing through its mass. 



Near to the origin of a large earthquake earth waves are visible ; some 

 distance away their existence has been inferred from the wave-like motion 

 seen on the tops of forests, at a distance of 300 miles, and even at very 

 much greater distances the feeling occasioned by the moving ground is 

 similar to that which is felt upon a raft moved by an ocean swell. 

 Bracket seismographs, hanging pictures and lamps, water in vessels, ponds, 

 and even in lakes, do not move with their natural periods, but are clearly 

 influenced by a forced tilting. Finally, even as far as the antipodes of an 

 origin, the character of motion assumed by horizontal and other pendulums 

 shows that this is due to slow but repeated changes in the inclination of 

 their supporting foundations. 



If we except the movements observed within the epifocal area, all the 

 other movements are as explicable by the assumption of the outcrop of 

 mass waves as they are by the assumption of surface radiation. 



The explanation that these waves have an increased velocity in their 

 quadrantal region (assuming such to be the case) may perhaps rest on the 

 fact that we are not dealing with radiation in uniformly widening rings, 

 as would be the case over a plane surface. The condition in this region 

 is such that energy is transferred from ring to ring, the diameters of which 

 are but little different from each other. Radiation from a pole to its 

 antipodes over a spherical surface may be likened to that of a wave which 

 runs along a channel, which expands for half its length and then contracts. 



The phenomena which give the greatest support to the idea of surface 

 radiation are, first, the existence of earthquake recurrences or waves which 

 have travelled from an origin to a distant station in opposite directions 

 round the world, the one arriving last having its amplitude reduced to 

 expected dimensions ; and second, the observations which show that waves 

 trftvelliog over a continental surface £|.re pot so rapidly reduced in magni- 



