this seemed sufficient to put the eastern alternative out of court; and the puzzle now was to reconcile this fact3000 miles from Chicago-with the known distance of the earthquake from England so as to identify the locality. Now, the pieces of the puzzle would not fit just as they stood they assigned a position from which there would almost certainly have been direct news. An earthquake could not occur in San Francisco, for instance, without our hearing of it in the obvious way. It is true that telegraph lines and cables are liable to be broken on such occasions; but, nevertheless, the news comes-as it came promptly about Tokyo and Yokohama a few months ago. To account for the absence of any news of this kind, it was necessary to interpret the American and to strain the English information, so as to bring the intersection into some spot remote from civilisation. Thus 3000 miles might mean 3000 kilometres; for earthquake distances are often specified in kilometres, and some one might have confused the two. A question could, of course, have been asked and answered by cable; but the earthquake was so large that it seemed incredible that it should remain unidentified even for the time required to put the question and get an answer. It was natural to accept the American news as claiming the disaster for North America, in which there are very few corners so remote from the enterprising reporter as to hide an earthquake.

The mystery actually lasted several days, and only by news from China was it discovered that the American news was completely erroneous, and that the earthquake had really taken place in Kansu in Western China. It had been truly appalling. Its very magnitude had deceived the American observers into misreading their records; it seemed impossible that so great an effect should have come from so far away. Moreover, the shock had not taken place under the sea as we had ventured to hope during the time of doubt, but under a thickly populated district, where the poor folk either lived in caves in the hills which fell upon them and buried them alive, or were accustomed to sleep on platforms of clay under which a fire could be kept burning all night. These platforms were so completely wrecked that those who escaped falling into the fire below and

being burnt were left to perish of cold. The loss of life was variously estimated, but must have been between 100,000 and a million.

As another example of the difficulty arising when the stations are not well distributed, let us take the information about the Japanese earthquakes of Sept. 1 and 2, as received to date. We have records from about twenty-five European stations, which in themselves suffice to give a good time determination and an accurate distance from Europe. But beyond this few records have been received. There are two from La Paz and Rio de Janeiro, which are of great value as enabling us to say that the focus was certainly not deep, and probably rather shallow; but very little as yet from the United States or Canada or Australia. Stations nearer the scene of the disaster from which we have records are as yet few (Zi-ka-wei, Taihoku, Simla, Batavia, Honolulu, and Colombo), and they are all to the west, except Honolulu. Hence, we cannot as yet fix a good intersection of the European mean circle. For the purposes of this article I have made the best computation at present possible, which assigns the positions lat. 35°.2 N., long. 139°-7 E. to the great Sept. 1 shock (just south of Yokohama), and lat. 36°3 N., long. 139°.2 E. to that of Sept. 2, the position of Tokyo being 35°7 N., 139°8 E. If these figures are correct, the epicentre for the second shock was well north of Tokyo, and further inland, which might possibly explain the fortunate failure to wreck the work of

rescue.

In such cases of large earthquakes, the time determinations of the main shock generally agree within a few seconds, and their mean or average may be accepted as accurate within narrow limits. We can thus infer when the effects will reach the other side of the eartha region not represented in the table given above, for reasons which need not trouble us here. The usual time to the precise antipodes is 20m. 17s., and allowance can readily be made for stations near, but not at, the exact anti-point. But it is noticed that for some earthquakes-not many-the effect arrives nearly a minute earlier; and for others-even less numerous-it arrives distinctly late on schedule time. It seems a reasonable inference that in the former case the distance travelled

was less, in the latter more, than usual; in other words, that the shock occurred in one case deeper down in the earth than usual, and hence nearer the opposite side; in the second case not so deep, and hence farther from the opposite side. This throws valuable light on the problem of focal depth. We have seen that Mallet put the depth of the Neapolitan earthquake at about 8 miles, and though estimates made since his time have been usually larger than this, they have not exceeded 80 miles. If, however, the above argument is sound the depth of one earthquake may exceed that of another by some hundreds of miles, and though as yet we cannot definitely specify the actual depth in any one case, the evidence suggests that 200 miles for the average earthquake is not far from the truth, some being much deeper (say 400 miles) and others nearer the surface. Further, it seems probable that the destructive earthquakes belong to this latter class-near the surface-and that a really deep earthquake, though it may produce a huge effect on distant seismograph records, may cause little damage at the surface. Such was presumably the case with the Formosa earthquakes of Sept. 1 and Sept. 14, 1922, which caused comparatively little damage, and yet shook the seismographs in England considerably. On the other hand, the big China earthquake of Dec. 16, 1920, and the recent Japanese earthquakes seem to belong to the shallow class, though the evidence is not yet fully collected. According to the slight evidence available, the Tokyo shock of Sept. 2 was distinctly deeper than that of Sept. 1, which is a possible alternative explanation of its less destructive character.

These examples may serve to illustrate what the new instrumental methods can teach us in individual cases. But they have also a much wider application. For the first time we are obtaining statistics of a systematic character as to the times and places where earthquakes occur. All those collected up to the end of the 19th century are affected by accidental circumstances to an extent which would scarcely be credited by those unacquainted with the subject. As an example, consider the mere numbers of large earthquakes which have been put on record. Milne made a Catalogue of Destructive Earthquakes from which the following figures have been

taken for the numbers of large earthquakes in successive periods of sixty years:

We should be prepared for a moderate increase in these numbers, owing to the greater vigilance of observers and increased facilities for communication; but that we should have in 1860 treble the number for 1800, and sixteen times the number for 1440, is so excessive a change as to deprive the numbers of any real value. We must begin counting again on a more systematic plan. Only one or two series of the old records seem deserving of attention; and in particular those collected by the placid Chinese in an apparently uniform way for nearly 2000 years.

The statistics of locality have naturally suffered in the same way as those of number. The accident of the occurrence of an earthquake near a town or village has given it a spurious importance from the scientific point of view, so that a comparatively small shock would find its way into the catalogue as 'destructive' when a much larger shock would pass unnoticed. All this can be set right by patient work, but the new seismology is in its infancy, and many of those who were nursing its tender years have been taken from us-Milne, Galitzin, G. W. Walker, C. G. Knott are serious losses. But the first gathering of seismologists in international conference since the war, at Rome in May 1922, gave promise of renewed earnestness and fresh energy; and we are to meet again in October 1924 at Madrid. We shall hope to have the general good wishes for our work and for the future of the new science.

H. H. TURNER.

Art. 8.-JUTLAND-THE REAL STORY.

1. Naval Operations. Vol. III. (History of the Great War.) By Sir Julian Corbett. Text and Maps. Longmans, 1923.

2. The Fighting at Jutland. Edited by H. W. Fawcett and G. W. W. Hooper. Macmillan, n.d.

So much ignorant criticism and misrepresentation have been displayed in connexion with the story of the Battle of Jutland, that the complete and illuminating account of it published in the new volume of the Official History of the War,' compiled under the directions of the Committee of Imperial Defence, is doubly welcome and valuable. The last chapters of the account, we are told, were completed by the late Sir Julian Corbett only a few hours before his death. It was fitting that this celebrated naval historian should close his life's work with an epic of the greatest naval action of our times. The narrative is one which worthily maintains his high reputation for marshalling facts, for the logic of his deductions and the fairness of his opinions.

Sir Julian Corbett had at his disposal all the essential official documents, the logs of the ships which took part in the battle, copies of the signals and messages sent and received, as well as the despatches and reports of the commanders. Moreover, a mass of material information from the German side was available to him. In addition, he was assisted by a staff of expert Naval officers, both in the technical aspects of the account and in the preparation of the detailed and admirably clear plans which accompany it. Altogether we may take it that here we have presented, for the first time, all the essential features of the great drama of Jutland.

In the light of the whole true story now unfolded before the world, we see the folly of criticism without knowledge of facts or the reasons which inspired events. We see, too, the danger of attempting to write history prematurely. One-sided accounts are shown up with all their faults and personal bias. Ill-balanced records of the battle, like that to be found in the latest supplement to the Encyclopædia Britannica,' are seen with all their inaccuracies and misinterpretations. It is not the pur