122 EEPOBT— 1892. 



of a town, special attention ought to be given to the construction of shop 

 and other frontages, and the streets be made wide. 



Another indication of wave movement is the statement of people who 

 say that after they had been thrown upon the ground the movements of 

 the earth rolled them from side to side. A station-master, who tumbled 

 on the line as the station-house fell close behind him, showed the writer 

 the manner in which he seized one of the rails whilst lying on the 

 ground, the rail passing between his legs. While in this position he 

 was tumbled from side to side, first striking the ground with one 

 shouldei', and then with the other. 



Reasons for believing that in Tokio the ground was thrown into long 

 undulations have already been given. First, there was the evidence of 

 our sensations ; secondly, the observation of the manner in which water 

 moved in ponds ; and, thirdly, the observations on the movements of 

 bracket seismographs, which were tipped from side to side. The most 

 certain evidence about the tilting is, however, that which is furnished by 

 the diagrams of many seismographs, which, rather than showing a series 

 of irregular waves with superimposed irregularities, in almost all cases 

 show a series of clean-cut curves. In one instrument which was tested 

 the periodicity of these curves did not agree with the period of the 

 instrument, from which we may conclude that they had not been formed 

 by swinging. Further, the periods of a consecutive series of waves are 

 not constant. For example, one set of east and west tiltings followed 

 each other, with periods measured in seconds of 3'4, 2'0, 2'7, 17, 4-1, 3'1, 

 3'1, 2'7. On another instrument another set of waves, taken at random, 

 followed each other at intervals of 1'9, 2*5, 1'3, and 2'6 seconds. These 

 observations also preclude the idea that the records were obtained by 

 swinging. The most interesting observation, however, is that a pair of 

 conical pendulums, the bobs of which were supposed to be steady points, 

 and which had no pointers for multiplication, gave diagrams about twice 

 as large as similar, but smaller, conical pendulums which carried pointers 

 to multiply any motion relative to their bobs six times. 



The actual records are as follows : — 



I 



On the assumption that the bobs of these machines had acted as 

 steady points, we should come to the conclusion that the range of north 

 and south motion had been 8 inches, as given by one instrument, whilst 

 it was only "66 inch as given by another, both instruments being in the 

 same building. It is clear that these two instruments had not behaved 

 as modern seismographs are supposed to behave at the time of an earth- 

 quake, but because the displacements indicated are practically proportional 

 to the length of boom, or the length of boom and pointers, that the 

 instruments had been tilted, and the extent of the displacement measures 

 maximum slopes of earth-waves. To interpret these measurements, it is 

 necessary to place a level on the stand of the seismograph, and determine 

 by experiment the angular values of tilting corresponding to measured 



