The phenomena that we are certain of in connection with earth vibrations, whether these vibrations are produced artificially by explosions of dynamite in bore-holes, or whether they are produced naturally by earthquakes, are, first, that a disturbance as it dies out at a given point often shows in the diagrams obtained by seismographs a decrease in period; and, secondly, a similar decrease in the period of the disturbance takes place as the disturbance spreads. As examples of these actions I will quote the following. The diagram of the disturbance of March 1, 1882, taken at Yokohama, shows that the vibrations at the commencement of the disturbance had a period of about three per second, near the middle of the disturbance the period is about 1·1, whilst near the end the period has decreased to ·46. That is to say, the backward and forward motion of the ground at the commencement of the earthquake was six times as great as it was near the end, when to make one complete oscillation it took between two and three seconds. Probably the period became still less, but was not recorded owing to the insensibility of the instruments to such slow motions.[153] We have not yet the means of comparing together diagrams of two or more earthquakes, one having been taken near to the origin, and the other at a distance. The only comparisons which I have been enabled to make have been those of diagrams taken of the same earthquake, one in Tokio and the other in Yokohama. As this base is only sixteen miles, and the earthquake may have originated at a distance of several hundreds of miles, comparisons like these can be of but little value. Other diagrams illustrating the same point are those obtained at three stations in a straight line, but at different distances from the origin of a disturbance produced by exploding a charge of dynamite in a bore-hole. A simple inspection of these diagrams shows that at the near station the disturbance consisted of backward and forward motions, which, as compared with the same disturbance as recorded at a more distant station, were very rapid. Further, by examining the diagram of the motions, say, at the near station, it is clearly evident that the period of the backward and forward motion rapidly decreased as the motion died out. These illustrations are given as examples out of a large series of other records, all showing like results. An observation which confirms the records obtained from seismographs respecting the increase in period of an earthquake as it dies out I have had opportunities of twice making with my levels. After all perceptible motion of the ground subsequent upon a moderately severe shock had died away, I have distinctly seen the bubble in one of these levels slowly pulsating with an irregular period of from one to five seconds.