calculate, not only the number of shocks which will be experienced before the district settles to its normal state of seismic activity, but also the interval of time that will be involved in such an operation. For the earthquakes considered by Mr. Omori it may be concluded that the earth's crust had been so far fractured that there was an approximate similarity in the heterogeneity of the disjointed material, which therefore, as it settled, gave rise to after shocks following a somewhat similar law. Another observation was that the larger of the after shocks travelled to greater distances than their smaller companions, and in consequence there was a marked difference in frequency at places situated at different distances from the primitive origin. If there is any law in this decrease in frequency with distance, then the frequency of what are evidently after shocks observed upon a coast line, as in Districts 1 and 10, might enable an observer to make a rough estimate of the distance of an inaccessible submarine origin. That satisfactory results would be obtained from such an investigation is, however, doubtful.

Distribution of Earthquakes.-An inspection of the map of earthquake origins or centres shows that the central portions of Japan, which are the mountainous districts where active volcanoes are numerous, are singularly free from earthquakes. The greater number of disturbances originate along the eastern coast of the Empire, and many of these have a submarine origin. That very few earthquakes are shown on the coast line between Districts 1 and 2 is in a great measure due to the fact that in this region there are but few observing stations, the island of Yezo in which these districts are situated being sparsely populated. A line drawn from N.N.W. to S.S.E., or from numbers 7 to 557, is the chief anticlinal axis of the northern island, and from the southerly prolongation of this beneath the ocean, earthquakes from time to time originate, which shake, not only the eastern coast of Yezo, but also many of the districts on the main island. Although districts like 11, 9, 8, and then through 7, suddenly northwards up to 13 or 14, lie along the strike line of the southern portion of the Empire, a greater number of earthquakes seem to originate from the face of the steep monoclinal slope which Japan presents towards the Pacific Ocean.

Lines, 120 geographical miles in length, running in an easterly or south-easterly direction from the highlands of Japan into the Pacific Ocean, like similar lines drawn from the Andes westwards into the same ocean, have a slope of 1 in 20 to 1 in 30, and in both of these districts earthquakes are frequent. On the contrary, along the face of flexures which are comparatively gentle, being less than half these amounts, which may be seen along the borders of most of the continents and islands of the world, earthquakes are comparatively rare. The inference from this is that, where there is the greatest bending, it is there that sudden yielding is the most frequent. In the case of many of the Japanese earthquakes, this takes place along the face of a monoclinal feature of the world's surface, and the intimate relationship between monoclines and faults is known to all geologists, the former being, in the words of Sir Archibald Geikie, an incipient stage of the latter.

Earthquakes and Secular Movements.-Another feature indicated by the map or known to the writer from personal observation is that earthquakes are frequent in those districts where there are evidences of secular elevation or depression, that is to say, in those districts where movement of the earth's crust is yet slowly taking place.

In Districts 1, 2, 5, 6, and 7 the writer knows from repeated observation that there are evidences of very recent elevation, and certainly in these districts earthquakes are extremely frequent. The signs + and in the neighbourhood of Districts 8, 9, 11, 12, and 13, and along the Inland Sea, lying to the north of 8 and 9, but to the south of 12, also show a like relationship.

The only exceptions to the general rule appear to be the westerly portion of the district between 12 and 13, where there are evidences of secular movement, and earthquakes are of rare occurrence, and 1 in 5 cases where these conditions are reversed. The district No. 14 presents a series of earthquakes originating along the line of a valley between high mountains running from N.N.E. to S.S.W. Another good example of earthquake fracturing following a line of weakness down a valley between high mountains until it reached the plain was the disturbance of October 28, 1891, which, as has been explained, resulted in the abnormal conditions shown in District 7.

In Japan, therefore, earthquakes have been frequent along the steep monoclinal face of the country, in the synclinal trough of deep valleys, possibly along the continuation of the Yezo anticlinal, and in districts where secular movement is in progress. In Italy earthquakes originate along the anticlinal of the Apennines, and from what we know of the geological history of the country, which had its greatest growth in Tertiary times, and from the bradyseismic movements on the coast, it is not unlikely that the shakings it experiences announce the fact that secular yielding is yet in progress. The earthquakes of Switzerland and those which shake the Himalayas, and the younger mountains of the world, may also be taken as due to orogenic causes which seem to be so actively in operation in Japan.

Earthquake Sounds.-A map which has been prepared, but which has not been reproduced with the catalogue, shows the distribution of earthquakes accompanied by sound phenomena. To indicate that a sound was heard, a dot is used, for a sound with a shock the sign+, for a sound before a shock the sign, while for a sound after a shock the sign | . After a volcanic explosion it might be expected that a sound wave propagated through the atmosphere would succeed a trembling of the ground.

As the latter sign occurs but seldom, although there are one or two cases of its occurrence in Districts 6, 7, 12, and 14, generally near active or old volcanoes, and about two cases in District 8, it may be assumed that earthquake sounds, rather than representing atmospheric waves radiating from an epifocal area, represent elastic vibrations transmitted through the ground, and therefore arrive at a given station in advance of any quasi-elastic surface undulation. Inasmuch as earthquake sounds only travel a few miles from their origin, the intervals between them and an earth movement, which can be felt are very small. The result of this is that it often appears that the two phenomena are simultaneous, and therefore on the map we find nearly as many signs indicating 'sound with shock' as those which indicate sound before shock.' Sounds are often heard which cause people to run from their houses, expecting a shock which does not come. The dots on the map represent sounds which have been to ordinary observers simultaneous with an actual shaking of the ground. Taking the districts in order, we find the sound phenomena distributed as follows:

1. Sounds fairly frequent on the coast at the most easterly and most southerly portions of the district. Inland and on the northern coast they are rare. This may indicate that the majority of earthquake origins lie to the S.E. and are submarine.

2 and 3. Sounds are rare. Many of the origins of these shocks are submarine. The coast between 2 and 3 is composed of soft materials.

4 and most easterly part of 5. Here the coast is rocky, built up of Palæozoic strata. Sounds are fairly frequent. In the southern part of 5, where there is much soft Tertiary material, sounds are rare.

6. Sounds are frequent in the northern part of the district, which is mountainous, while in the plain of Musashi, constituting the southern part, they are rarely heard.

7. Amongst the Paleozoic hills of the district, and extending down into the plain, sound phenomena accompany about 30 per cent. of the disturbances.

8 and 9. Although the districts are mountainous, sounds are rarely heard. Possibly the shocks originate beneath the ocean.

10, 11, and 12. Sounds are fairly frequent.

13. Here, which is another mountainous region, sound phenomena

are common.

14. Sound is occasionally heard.

15. Along a sandy coast bordering a plain, sound phenomena seem never to be heard

Generally sound is heard in rocky mountainous districts, while on the alluvial plains it is but very rarely observed.

Earthquakes which have been propagated to Europe. The object in appending to the catalogue a list of earthquakes which was kindly drawn up for me by Dr. E. von Rebeur-Paschwitz is to show that some of the Japanese disturbances have travelled as far as Europe, where for minutes or hours, although they were unfelt by persons, they caused movements in delicately adjusted horizontal pendulums. A similar series of unfelt disturbances originating in distant countries or beneath the oceans have been recorded in Japan.

IV. ON THE VELOCITIES WITH WHICH WAVES AND VIBRATIONS ARE PROPAGATED ON THE SURFACE OF AND THROUGH ROCK AND EARTH. (A COMPILATION.)

Introduction.

Because the observations which have been made upon the rate at which waves and vibrations are transmitted through rock and earth are so varied and often apparently contradictory, it has been thought advisable to select from the vast amount of material which is at our command a series of illustrations from experiments upon artificially produced disturbances, and from the records of actual earthquakes in which personal and instrumental errors have been small.

Amongst the real or apparent difficulties are the following

1. Along the same path, earth waves, originating from a powerful

impulse, travel at a higher rate than those resulting from an effort of lower intensity. 2. Near to an origin, the velocity of propagation is greater than it is between points at a distance. 3. After a disturbance has decreased in its speed of transmission it may be accelerated, and this acceleration cannot be with certainty attributed to its having entered a more elastic medium. 4. As an earthquake radiates, it is preceded by a series of minute tremors, the velocity of propagation of which is certainly very much higher than that of the main disturbance.

(a) Artificially produced disturbances. 1. Experiments of Mr. R. MALLET.

The

In the experiments of the late Robert Mallet conducted at Killiney Bay, Dalkey, and Holyhead (British Association Report,' 1861) the initial impulse was caused by the explosion of charges of gunpowder. electrical contact which caused the explosion released a chronograph which was stopped by an observer directly he saw, by means of a microscope magnifying 11:39 times, an agitation caused by the resulting waves in a dish of mercury. After corrections for the intervals of time thus noted, in round numbers the results obtained were as follows:

The charges of powder employed varied between 25 lb. and 12,000 lb., and with but one exception it was clearly shown that the velocity of wave propagation increased with the force of the initial impulse. For example, at Holyhead the relationship between the quantity of explosive and the resulting velocities was as follows:

2. Experiments by General II. L. ABBOT.

In 1885 when Flood Rock was destroyed by the explosion of 240,397 lb. of rack-a-rock and 48,537 lb. of dynamite, the most distant observing station was 182-68 miles off. The instant of the explosion was noted at all the points of observation by means of electrical connections and chronographs, while the arrival of the first tremors and their duration was recorded by observers who watched the disturbance of an image reflected from the surface of mercury.

The Hallet's Point observations, where the initial impulse was due to the explosion of 50,000 lb. of dynamite, and others made in connection with subaqueous explosions at the school of submarine mining at Willet's Point, were conducted in a somewhat similar manner. In the following table, which has been drawn up from the scattered writings of General Abbot, the velocities have been reduced to uniform units:

From the above data it is clear, as Abbot shows, that the rate at which a shock is transmitted increases with the intensity of the initial explosion; that when a high magnifying power has been used, tremors in advance of those revealed by a low power have been noticed, with the result that the apparent velocity in the former case is greater than in the latter; and that the velocity of propagation has been higher through rock than through soft material like drift.

A query put forward by General Abbot is whether still higher velocities would have been recorded had telescopes with a greater magnifying power been used. The answer is apparently in the affirmative, and therefore if we wish to compare the observations amongst themselves, not only must we choose those in which the initial impulse has been the same, but where the observers have employed similar instruments. Comparing observations 10 and 12, but not overlooking the fact that No. 10 was largely transmitted through water, and again 16 and 17, it might be concluded that as a wave advances its velocity is diminished; but from the first five observations it would seem that there is at the commencement an increase in the initial velocity until it reaches a maximum, after which there is a diminution. This increase in the rate of transmission at the outset of a wave from its origin is again seen in experiments 9 and 11. The difference in the velocities recorded for experiments 18 and 19 may be due to the fact that in the case of the shallow torpedo much of the initial energy was