_Cause of microseismic movements._—As to the cause of tromometric movements, we have a field for speculation. Possibly they may be due to slight vibratory motions produced in the soil by the bending and crackling of rocks produced by their rise upon the relief of atmospheric pressure. If this were so we should expect similar movements to be produced at the time of an increase of pressure. Rossi suggests that they may be the result of an increased escape of vapour from the molten materials beneath the crust of the earth, consequent upon a relief of pressure. The similarity of some of the sounds which are heard with the microphone to those produced by boiling water are suggestive of this, and Rossi quotes instances when underground noises like those which we should expect to hear from a boiling fluid have been heard before earthquakes without the aid of microphones. One instance was that of Viduari, a prisoner in Lima, who, two days before the shock of 1824, repeatedly predicted the same in consequence of the noises he heard. A possible cause of disturbances of this order may be small but sudden fluctuations in barometric pressure, which are visible during a storm. During a small typhoon on September 15, 1881, when in the Kurile Islands, I observed that the needle of an aneroid worked back and forth with a period of from one to three seconds. This continued for several hours. With every gust of wind the needle suddenly rose and then immediately fell. At times it trembled. These movements were observed in the open air. The extent of these sudden variations was approximately from ·03 to ·05 inches. Beckoning an increase of barometrical pressure of one hundredth of an inch as equivalent to a load of twenty million pounds on the square mile, during this storm there must have been the equivalent of loads of from 60 to 100 million pounds to the square mile continually placed on and removed from a considerable tract of the earth’s surface. If the period of application of these stresses approximately coincide with the natural vibrational period of the area affected, it would surely seem, especially when we reflect upon the effect of an ordinary carriage, that tremors of considerable magnitude ought to be produced. An inspection of the following few observations taken from my note-book for the same typhoon will suggest that even the large and slower variations are capable of producing tremulous motions. Time Barometer h. m. reading 12 5 P.M. 29·02 12 10 „ 29·05 12 12 „ 29·07 12 13 „ 29·05 12 25 „ 29·10 12 50 „ 29·00 1 10 „ 29·00 1 20 „ 29·07 CHAPTER XX.