Page images
PDF
EPUB

assistant of the Meteorological Bureau, who has worked with me, pointing out doubtful information, translating papers, calculating areas, determining centres, filling in maps, and in carrying out other tedious operations for the last twelve months.

The chief reason for terminating the catalogues at the end of 1892 is because the material subsequent to that date has not yet been reduced to the map form, and to examine all the documents necessary to accomplish this would have occupied at least another year; also it may be added that what has been done is in all probability sufficient to determine whether work of this description is likely to lead to results of sufficient importance to guarantee its continuance.

(b) Explanation of the Catalogues.

In the first catalogue the shocks are placed in chronological order from 1 to 8,331. When disturbances have apparently been simultaneous in two distant localities, they are included under a single number.

In the second and third columns the date and time for each disturbance are given. When the latter is noted to seconds, the record refers to the commencement of motion at an observatory, like that in Tokio, which is provided with automatic chronographs. Until the end of 1887, these records, which are practically correct, refer to Tokio mean time, or 9 hr. 19 min. 1 sec. before Greenwich mean time. Subsequently to this date the times given are those of Long. 135° E., or nine hours before Greenwich mean time. The other time records are only approximately correct, and cannot be used in any investigation relating to the velocity with which earthquake motion is propagated.

The fourth column gives in square ri (1 square ri=5.96 square miles) the land area which was shaken. For small shocks which were only felt at one or two stations the determination of this quantity has largely depended upon the judgment of the observer. The figures given are those obtained from the maps by means of a planimeter and entered in the records of the Meteorological Department. In the second catalogue, based upon a second inspection of the maps, it will be noticed that many material alterations have been made in these quantities. In many instances the land areas of the first catalogue are total areas, but in others they only represent an insignificant portion of a disturbed tract, the centre of which was beneath the bed of the ocean. The limits of the areas given are those places round an origin up to which the movement was perceptible to people or sufficiently strong to have been recorded by ordinary seismographs. With instruments like delicately adjusted horizontal pendulums, there is no doubt that movements might have been detected far beyond these arbitrary limits. For example, shock number 4,145 has assigned to it a land area of 15,750 sq. ri, when we have good reasons for believing that with suitable instruments it might have been noted at any point upon the surface of our globe.

The number in the fifth column approximately indicates, as shown upon the key map, the epicentre of a disturbance, or a number on the coast line nearest to a submarine origin. In the second catalogue, the position of a submarine origin, by means of a distance in tens of miles and the direction in which it is to be measured from a central number, is defined more closely. On the key map the numbers referring to squares,

each 10 miles by 10 miles, commence at the top and run from left to right down to the bottom of the same.

A line drawn on the key map through the numbers in the sixth column gives the boundary of the land surface which was shaken. The area of this should be equal to the quantity in the second column. By completing, when it may be necessary, this outline seawards, a total area is obtained, which is indicated by its major and minor axes in the second catalogue.

In the small map, which is a photographic reproduction of a map the same size as the key map, the small dots indicate the position of all the epicentral numbers, and the large numerals ranging from 1 to 15, districts in which earthquakes are frequent. Districts 6 and 7 are bounded by straight lines because there was not sufficient space in which to place all the dots. For example, in District 7 all the dots indicate earthquakes which originated about the centre of this district. Until October 28, 1891, the disturbances in this district were not more numerous than they are in District 8.

When an earthquake has been felt at the extremities of the Empire, and at the same time not along a great length of coast line, as in Districts 1 and 10, it is often difficult to determine the direction or distance from the coast line of its origin. In these cases the assumption made has been that the shocks just reaching the coast have originated from about the same locality as the larger shocks which have spread some distance along the shore line, these stronger disturbances being severe at places just reached by their feebler successors. The signs and along the coast line indicate that near these places there are evidences of secular elevation or depression. This information was obtained by the help of Professor D. Kikuchi, who kindly assisted in the distribution of a circular to various towns and villages round the coast of the Empire inquiring whether from maps, traditions, or observations there were reasons to believe that changes had taken place in the relative position of the land and water. The large black dots on the map indicate the positions of more or less active volanic cones, in the neighbourhood of which there are huge bosses of volcanic rocks and many ancient craters. The dotted lines show the boundaries of provinces, which are usually the ridges of high mountains dividing one seismic region from another.

If analyses of this catalogue show that it is of any value, it is clear that several advantageous changes may be made in a system for its continuation. As it stands it is only a tentative effort to provide investigators with a new kind of data, which may lead to investigations not hitherto possible. None of the facts, excepting a few of the time observations, claim any great degree of accuracy. The object of the list drawn up for me by Dr. E. von Rebeur-Paschwitz is explained in the next section.

The long list of corrections, additions, and suggestions at the end of the volume, inasmuch as they have, so far as possible, been inserted in the second catalogue, almost entirely refers to the first catalogue. Although they show that actual errors occur in work of this description, they also show that from given data different persons may arrive at different results.

The unit is 10 geographical miles.

(c) Object of the Catalogues.

The principal object of the catalogues, as we have indicated, is to furnish investigators with a certain quantity of material relating to the occurrence of earthquakes, different from that which has hitherto been at their disposal, on account of the want of which it has been impossible to make many desirable inquiries.

Many catalogues exist, like those of Perrey, Mallet, Kluge, de Ballore, and Fuchs, in which the actual number of records are equal to, or greater than, the number of earthquakes now noted, and which are equally good as foundations for a particular class of investigations. The incompleteness of these catalogues, however, is seen in the fact that they give for the whole world a frequency less than the present list gives for a small portion of it like Japan. If, for example, we take Dr. C. W. C. Fuchs' 'Statistik der Erdbeben,' 1865 to 1885, giving a list of some 8,000 disturbances, out of these Japan is credited with from three to thirty shocks per year, while a truer estimate would have been from 500 to 1,000. Again, it is often difficult to distinguish between shocks which have shaken a few square miles and those which have disturbed an empire. Large shocks and small shocks, primary shocks and after shocks, are with difficulty separable, and no data have been available enabling an investigator to separate disturbances arising from the yielding of strata in one area from those due to fracturing which might take place in a neighbouring region. Even when the lists of a particular observatory have been examined by themselves, inasmuch as its records are those of shocks of local orgin combined with those of shakings which originated at distances of several hundreds of miles, all that we can expect to find is a relationship between earthquake occurrence and influences of a widespread character. Such investigations have been made for the records of observatories, countries, and the world, with the result that a more or less pronounced annual and semi-annual periodicity and traces of what is apparently a lunar influence have been discovered.

No doubt many and very just objections may be made as to the accuracy of much of the material in the present list; but because it enables us to give approximate weights to the different shocks, to distinguish between primary and secondary disturbances, and to divide the country to which it refers into distinct seismic or natural districts, it is to be hoped that it will open the way for investigations along new lines.

Although the catalogues suggest several investigations hitherto impossible, inasmuch as it so often happens that one inquiry becomes the parent of another, it is impossible to indicate all the paths which may be followed. A suggestion given by the list, which shows that shocks originating in Japan have travelled to Europe, is that a ring of twelve or twenty-four stations situated round our globe would in a very short time give us valuable information, not simply about its crust, but possibly also about its interior.

One set of investigations which may possibly lead to interesting results will be those relating to the frequency and periodicity of earthquake shocks which may be considered as having equal values, or receive values relative to the area they have disturbed. Each of these analyses may be made for Japan as a whole, or for special seismic districts; in the former case the object being to determine whether the occurrence of

earthquakes is dependent upon influences which simultaneously affect Japan as a whole, and in the latter case to determine how far their frequency may be related to phenomena of a more local character.

As an example of an influence which affects Japan as a whole, the difference in the summer and winter barometrical gradients crossing the country may be taken, while tidal loads along the coast would be expected to produce effects in different districts at different times.

Not only is it open for us to determine effects due to external influences, but these, so far as possible, must be distinguished from effects resulting from internal conditions. The great frequency in District 7 was entirely due to the shocks succeeding a terrible disturbance which took place on October 28, 1891; and if these after shocks, which at first occurred at the rate of 1,700 per month, and which apparently result from the settlement of disjointed strata, are included in any general list, it is clear that they might accentuate or destroy any law respecting a long period frequency. What is true for District 7 is also true for District 11. By themselves they yield information about the rate at which an enormous quantity of broken-up strata settles to a state of equilibrium, and because the district around the epicentrum is for some time after the primary disturbance in an extreme state of seismic sensibility, it is quite possible that there may be fluctuations in the rate at which quiescence is approached, due to external influences. Other problems which suggest themselves are the possible relationships between the seismic activity of the various districts, the times taken for different areas under the influence of secular movement to attain varying degrees of seismic sensibility, and the connection between earthquake occurrence and the geotectonic character of the country. If the object of an analysis is to discover a relationship between earthquake frequency and exogenous phenomena which recur at long intervals, it would seem advisable to omit long lists of after shocks, and only to take into consideration disturbances which occur in districts where seismic activity is in a normal state. On the contrary, should we seek a relationship between the occurrence of earthquakes and phenomena which recur at intervals of not more than a few days, as, for example, barometrical fluctuations or the rising and falling of the tide, this precaution is hardly necessary.

[merged small][merged small][merged small][merged small][merged small][merged small][subsumed][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][subsumed][merged small][subsumed][merged small][merged small][subsumed][merged small][subsumed][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small]

For the latter investigation, the most desirable lists to use would be those referring to shocks originating beneath the ocean or along the seaboard, and as an assistance to this I give the preceding table, showing the times of high water at full and new moon on the coasts for the fifteen seismic districts shown in the small map.

Nothing has been said about the possible relationship between earthquakes and volcanic eruptions, first, because we have no reason to believe that, with the exception of a few feeble shocks which may precede or accompany an eruption, there is any marked direct connection between these two phenomena, and secondly, because the present catalogue does not extend over a sufficiently long period of years to lend itself to such an investigation. Although one or two new investigations have been here suggested, the principal work will be a repetition of old analyses, taking advantage of the fact that we are now able to deal with natural districts, to give earthquakes, where required, relative weights, and to distinguish between after shocks, the occurrence of which is but little influenced by epigenic actions of long periodicity, and those of a district where seismic strain is in a normal condition.

As to whether seismology will be advanced by carrying out these and other inquiries which may present themselves is a question which cannot yet be answered. It may be or it may not be, but the catalogue, which could not have been compiled without the generous assistance of the Royal Society of London and the kindness of the director and officers of the Imperial Meteorological Department of Japan, by allowing access to their unequalled store of valuable facts, will, it is hoped, settle the question as to whether it is desirable to continue in its present form the largest and probably the most perfect seismic survey which has hitherto been attempted.

I am glad to say that some of the features presented by the catalogues are now being analysed by Dr. C. G. Knott, of Edinburgh.

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

After Shocks.-About the time that the catalogue was commenced, Mr. F. Omori undertook an examination of the shocks succeeding the great earthquake of October 28, 1891, which are now indicated upon the map in District No. 7. This he did, following up the investigation by an analysis of the disturbances since 1889 in District 11, a series which recently occurred in District 10, and another series belonging to a region lying between 8 and 9, which, although now quiescent, about forty years ago was unusually active. As an outline of Mr. Omori's investigations is published in the Seismological Journal,' vol. iii. p. 71, and in greater detail in the 'Journal of the College of Science, vol. vii. Part II., it would be out of place to give any detailed reference to them here. Briefly, it seems that when a large disturbance is followed by a long series of after shocks the number of these is roughly proportional to the area first shaken, or what may provisionally be called the intensity of the initial impulse. The character of the curves which represent the frequency of the after shocks in relation to time is remarkably similar, and having determined by observation the form of the earlier portions of a frequency curve, it seems possible to roughly

« PreviousContinue »