is observed that pendulum A is steady on dull wet days, but on warm days the daily curves are well defined. Farther, although A and E move at the same time, they move in opposite directions, but each usually moves towards the side from which the greatest load is being removed by evaporation.

III. THE TOKIO EARTHQUAKE OF JUNE 20, 1894.

On June 20, at a few minutes past two in the afternoon, Tokio, Yokohama, and the surrounding districts were shaken by an earthquake which was more violent than any which has been recorded since 1855. On June 25 it was reported that in Tokio alone 33,940 buildings had suffered damage, some being entirely ruined, 140 persons had been wounded, and that 26 had been killed. In Yokohama, where many chimneys fell and houses were unroofed, 6 people were killed. When statistics are completed and it is known what has happened in other places these numbers will be increased. Small fissures were formed in the ground at Tokio in 96 places, many walls were shattered, while stone lanterns and tombstones were overthrown, twisted, or deranged.

The following facts are taken or deduced from the records obtained at the Central Observatory and the University Laboratory, both of which are in Tokio, at a distance of about 1 mile from each other. To these are added observations from the Hitotsubashi Observatory, which is situated on soft ground lying between the Central Observatory and the University, at which places the ground is comparatively hard.

At the University for the first 10 seconds the horizontal motion was slight, when it suddenly became severe, reaching 80 millimetres. The severe motion continued for about a minute, during which time there were more than 10 pronounced movements. As the range of motion was outside the limits of seismographs with multiplying indices these were deranged or broken, and complete diagrams were only obtained at the University and Hitotsubashi, where there are seismographs without such indices, the recording surfaces for which are only set in motion at the time of violent disturbances. Until the diagrams have been carefully analysed I am inclined to think that the recorded horizontal motions may represent the angles through which the seismographs have been tilted, rather than the range through which a given point suffered horizontal displacement.

Assuming for the present that these quantities are what they are represented as being, then at the University and at Hitotsubashi the maximum accelerations were respectively 400 and 1,000 millimetres per sec. per sec. In the Nagoya-Gifu earthquake of 1891, when nearly 10,000

lives were lost, the maximum accelerations, calculated on more certain data, varied between 3,000 and 8,000 millimetres per sec. per sec. At the University there is a seismometer, consisting of a number of iron shot, arranged on a ledge round the top of a strong post, beneath which there is a bed of sand. These shot were not projected, but all of them, excepting one on the N.E. side, simply fell. The duration of the disturbance is of course that given by seismographs. Horizontal pendulums may have been tilted backwards and forwards for one or two hours. For some time after the shock it was observed that the Sunida River, which runs through Tokio, rose and fell as if its bed continued to be agitated. The direction of motion, as with nearly all earthquakes, was varied, and the direction given is that which was most pronounced.

The times at which the commencement of the disturbance was recorded at places some distance from Tokio are as follows:

From these times and the distribution of destruction it may be assumed that Tokio was well within the epicentral area.

A remarkable feature distinguishing this earthquake from most others is that during the next three days instead of a long series of after shocks only three disturbances were recorded. The primary shock does not appear to have been accompanied by any sound, while one of the secondary shocks, at 4.25 P.M., on the 20th was preceded by a roaring sound.

At many places telegraph and telephone wires were broken. Underground the pipes of the Yokohama Waterworks were caused to leak, drains were deranged, and there was a falling in of material in a railway tunnel. A curious fact communicated to me by my colleague Professor W. K. Burton is that in his house, where he was barely able to keep his feet while the shaking was going on, several decanters were not upset, but their stoppers were shot out. This is similar to what has occurred on more than one occasion with the lamp glasses at the Kannonsaki lighthouse in Tokio Bay.

IV. MISCELLANEOUS.

In addition to the foregoing work two numbers of the 'Seismological Journal' have been issued and the manuscript of a catalogue of Japanese earthquakes between 1885 and 1892 has been completed. This catalogue gives the date, the time, the area shaken, and the position of the origin for 8,337 shocks. Appended to each shock are a series of numbers, and

a line traced through these, as shown on a key map, is an outline of the land area which was disturbed. The object of this catalogue, which is different from previous publications of the same description, was stated in the last report.

Investigation of the Earthquake and Volcanic Phenomena of Japan.— Fifteenth Report of the Committee, consisting of the Right Hon. Lord KELVIN, Professor W. G. ADAMS, Mr. J. T. BOTTOMLEY, Professor A.H. GREEN, Professor C. G. KNOTT, and Professor JOHN MILNE (Secretary). (Drawn up by the Secretary.)

II. Observations with Horizontal Pendulums

(a) The Instruments, Installation, Character of Movements
(b) Daily Wave Records

(c) Tremors, Microseismic Disturbances, or Earth Pulsations

(d) The Slow Displacement of Pendulums.

(e) Periodic movements of several days' duration, and wandering of 129

the pendulums.

III. Description of a Catalogue of 8,331 Earthquakes recorded in Japan 149 between January 1885 and December 1892.

(d) Results already obtained or shown by the Catalogue and Map of 155

(e) Object of the Catalogues

Centres.

IV. On the Velocities with which Wares and Vibrations are propagated on 158 the Surface of and through Rock and Earth. (Compilation).

Introduction

153

(a) Observations on Artificially Produced Disturbances. Experiments 159 of MALLET, ABBOT, FOUQUÉ and LÉVY, GRAY and MILNE. (b) Observations on Earthquakes. Where the rare paths have been 163 short: (MILNE and OMORI). Where the ware paths have been long: (NEWCOMв and DUTTON, AGAMENNONE, RICCO, CANCANI, VON REBEUR-PASCHWITZ. MILNE).

(e) The Probable Nature and Velocity of Propagation of Earthquake 170 Motion. The suggestions of Dr. C. G. KNOTT, LORD RAYLEIGH, LORD KELVIN.

(d) The Paths Followed by Earthquake Motion. Hypotheses of HOPKINS 173 and SEEBACH, SCHMIDT, and a suggestion by the writer.

(e) Conclusions.

179

V. Miscellaneous Notes relating to Large Earthquakes, &c.
APPENDIX.-On Causes producing Movements which may be Mistaken for 182
Earth Tremors.

I. THE GRAY-MILNE SEISMOGRAPH.

The first of the above seismographs, constructed in 1883, partly at the expense of the British Association, still continues to be used as the standard instrument at the Central Observatory in Tokio.

I am indebted to Mr. K. Kobayashi, the Director of the Observatory for the following table of its records :

1895.

I

Catalogue of Earthquakes recorded at the Central Meteorological Observatory in Tokio between May 1893 and February 1894.

II. OBSERVATIONS WITH HORIZONTAL PENDULUMS.

(a) The Instruments, Installation, Character of Movements. Since 1893 nineteen sets of records have been obtained from horizontal pendulums installed either in Tokio or its vicinity. In the following description the different installations or sets of instruments are indicated by letters of the alphabet. The instrument at A, which was in my house, occupied the same position from the commencement of the observations until February 17, 1895, when it was destroyed by fire. The instruments at other stations were kept in position until they had given continuous records for a period of from one to four months, when they were moved to a new locality. Although these periods may appear short, they seem to have been amply long to determine the general character of the movements to be expected at any given station. Instruments within a mile of