It is a well known fact that Seismogram is greatly disturbed by the free oscillation of the penduium, when no damping is introduced to it. In general, the displacement of the index point of a seismometer (ξ) is given by where n0 and n are the numbers of oscillation in 2n seconds, of the earthquake motion and that of the free oscillation of the pendulum respectively, if the displacement of the earthquake motion (x) is given by If the earthquake occurs at t=0 and afterward a simple harmonic motion of ground continues indefinitely, i.e. if its displacement is given by the apparent displacement of seismometer index is given by where The author has computed the numerical values of the maximum displacement of the pendulum for first few oscillatious and corresponding phases. The results are given in table I and II in page 110, and graphically shown in fig. 1 and 2. The followings may be specially noticed: 1. The amplitude of the first motion recorded is greatest for the oscillation with infinitely short period, 2. The amplitudes of the second and other oscillations are greatest for some periods of oseillation shorter than that of the free oscillation of the pendulum, 3. The free oscillation of the pendulum is larger than the forced oscillation, when the period of the earthquake motion is larger than that of the free oscillation. Accordingly, in this case no apparent relation is found between the earthquake motion and the motion recorded.
In this paper, the seismic origin means an underground point where the seismic disturbance sets in earliest. The formula given by Prof. F. Omori showing the relation between the distance of this point from the place of observation and the duration T of the preliminary tremor, is Δ=7.42T, where Δ is measured in km. and T in sec., which is applicable for 0<Δ<1000km. The author applied this formula to find the seismic origin of the Great Kwantô Earthquake of Sept. 1. 1923, referring to the seven seismograms taken at the meteorological observatories at Kôhu, Numadu, Tôkyô, Kumagae, Maebasi, Matumoto and Kôbe, and also to another obtained at the Geophysical Observatory of the University, Sendai. Eight spheres with radii of the calculated Δ's for, and centers at the respective plac_??_s of observation meet together nearly at a point as shown in the figure shown in p. 115, with a depth of 40 km. Thence the author concludes that the seismic origin was situated at a point a little north of Mount Fuzi, and also that the formula of Prof. Omori is exact. (S. F.)
From the durations of the preliminary tremors observed at various stations, the author, in his previous paper, located the seismic origin of the Great Kwanto Earthquake at a point to the north of Mt. Fuzi. In the present paper, he found the following two equations representing the relation between the distance of each station from the above point and the time of arrival of the first shock at that station in which Δ is the distance in kilometer and t the time in second measured from its origin assumed at 11h 58m, Sept. 1st, 1923. The first and second equations are well satisfied by the observed values of t and Δ at stations respectively in western and eastern Japan, the greatest discrepancy between the culculated and observed t being 2.7 seconds, while the total value of t ranges between 39-86 seconds. If the above equations are correct, there must be a time difference of 7 seconds for the commencements of the earthquake shocks at the seismic origin for the two waves, one travelling to the east and the other to the west. The author considers a highly strained state of earth crust, in a finite area near the origin, existing before the great earthquake and supposes that the waves actually started from the both sides of the area, whose diameter is about 50km, so that the apparent difference of time is accounted for.
Since Japanese Islands are stretched to a wide extent from southwest to northeast, from the tropical region to 50°N. of latitude, we can meet with various samples of snow. The author studied the corre_??_ation between the density of snow and the air temperature and humidity observed when snow is falling at Takata in Japan. From the results of 80 observations he found that the density of snow increases with the increase of temperature and with the decrease of humidity. When the air temperature is below the freezing point, the density (ratio of the depth of melted water to the original depth of snow on the ground) is about 0.09-0.10, but when the temperature is above that point, the density gradually increases with it. On one occasion a density amounting to 0.496 was measured. He concludes that the variety of snow may be conveniently classified with regard to the temperature and humidity prevailing at the time of fall.