depth, the resulting hodograph retains its character. Although the evidence that there is an increase in the velocity of propagation of waves as we trace them beneath the surface is by no means so complete as might be desired (see p. 161), Dr. Schmidt compares the advantages which the curvilinear propagation presents over that of the rectilinear transmission employed by Seebach. It will be observed that in fig 18 there is a great concentration of earthquake rays in the epifocal region which would correspond to the destruction which is so noticeable in such districts, while with rectilinear radiation the absence of such concentration is not in accord with the results of experience. Although both hypotheses agree in showing a higher apparent velocity near to the epicentre, in Seebach's hyperbola an identical limit is reached for the apparent horizontal velocity for all earthquakes, while Schmidt's modification of the laws shows that the apparent velocity on the surface cannot be less than that between the focus and the first coseismal, with which it varies. From this it follows that for limited areas the latter method admits the possibility of very high velocities resulting from earthquakes originating at reasonable depths. With rectilinear propagation on the contrary, to obtain such high velocities as have been observed, it is necessary that origins should be situated at enormous depths. Should a disturbance originate near the surface, Schmidt's hodograph consists of two symmetrical concave branches which meet in an angle at the centre, indicating that the velocity increases from the epicentre outwards. After indicating the above and other advantages presented by the hodograph over the hyperbola as representing the velocity with which earth waves are apparently propagated, Dr. Schmidt takes a number of earthquakes for which good time observations have been obtained and plots the resulting curves. These which refer to earthquakes felt over moderate areas show the characteristic inflexion point denoting an increased velocity in the outer portion of the disturbed tract. The following are examples of his results : Middle Germany, March 6, 1872.-Longest wave path 400 km. Here the hodograph is distinct in character with a point of inflexion at about 11 miles from its vertex, having a slope indicating a velocity of 2.5 miles per minute. At a distance of 36.7 miles the velocity is 15 miles per minute. The curve passes much more closely through the points representing time and distance than the hyperbola of Seebach. Possible depth, 5 to 10 miles. Mallet's method, dependent upon a single observation, gives 1.9 to 2.9 miles. Herzogenrath, October 22, 1873.-Longest wave path 150 km. In this case the hodograph is practically concave, throughout its length indicating an origin near the surface. It is indicated over a radius of 17 miles. Possible depth is less than 3 km. By Seebach's method it may be from 0 to 14 km. Swiss Earthquake, January 7, 1887.-Longest wave path, 150 km. The general character of this hodograph is like the last. At the point of inflexion the velocity is 170 m. per second, and at 150 km. it is 1,300 m. The depth of the centrum is from 1 to 6 km. Charleston Earthquake, August 31, 1886.-Longest wave path, 1,500 km. Here the hodograph is nearly a straight line. The depth of the centrum may exceed 120 km. In the case of the first three of these examples, their hodographic character may be due to the fact that observations were made in epifocal areas, within which disturbances radiating from a centrum were recorded ; but in the last example this character has been lost, because most of the times which were noted probably refer to the arrival of a surface disturbance capable of being felt, and which might have been recorded by ordinary seismographs. These latter records are therefore such that we could not expect them to conform with the hypothesis under consideration, and until a number of stations separated by long distances are provided with instruments capable of recording minute tremors which may go through the earth. Until these have been established, it would seem that the confirmation of the attractive theory put forward by Dr. Schmidt must remain in abeyance. 3. A Suggestion that there are three Classes of Movement. The last hypothesis is one that takes into consideration three classes of movements which immediately round an epicentre are hopelessly confused. These are the truly elastic disturbances which from a focus reach the surface of the earth along rectilinear or curvilinear paths, forced displacements, and quasi-elastic waves, causing tumultuous movements in the centre of a megistoseismic area, and long undulatory elastic-gravity waves which are propagated over the surface of the earth. The escape of energy is most pronounced along the paths of least resistance, that is round the seismic vertical to an epifocal area, and then radially over the surface of the globe. The rate of propagation of the surface waves seems to be about 2 or 3 km. per second, and it may be fairly constant. The minute tremors which have been observed at stations 6,000 miles distant from their originating cause, if they travelled through the superficial crust of the earth they did so at a rate of perhaps 12 km. per second, while if they were created on the passage, their velocity, which is increased, becomes more abnormal. Assuming that they came as condensational waves through the earth, then their velocity is reduced to 8 or 10 km. per second, a quantity which, as suggested by Dr. E. von Rebeur-Paschwitz, may possibly throw new light upon the nature of materials constituting the interior of our earth. At present the facts bearing upon this latter question are both few and imperfect. To confirm or dispel the important conclusions indicated by the few facts at our disposal, it would seem desirable that investigations should be extended in such a manner that the results obtained by different observers would be comparable. With a set of stations situated round the globe at intervals of 15° or 30° apart, provided with instruments similar in character, similarly installed and similarly worked, which are capable of recording not simply small changes in level but also minute vibrations, we might easily extend our present knowledge, not simply respecting the propagation of surface undulations, but possibly of motion transmitted through the rigid globe. An indication of the latter phenomenon would be an enormous increase in the apparent velocity of a disturbance as it approached the antipodes of its origin, while the concentration of energy in such a region would suggest internal refraction. Other phenomena which might be recorded would be the diurnal and longer period wanderings of the instruments, local earthquakes and earth tremors. The latter, although important in themselves, because they so 1895. N often eclipse the effects of earthquakes, should, by proper instalment of the instruments, be as far as possible minimised. (e) Conclusions. If we except the curious results respecting the velocity of propagation of motion which we might expect to find, and which apparently exists in an epifocal area, the phenomenon of greatest interest, the study of which may lead us to important conclusions respecting the physical constitution of the interior of our globe, are the so-called preliminary tremors of earthquakes which are often continued as superimposed serrations on the quasielastic motions. In Japan these have been recorded and studied for the last 15 years, but it has only been within the last year or two that their appearance has been recognised in Europe. All that I know about these latter records is what I learn by letters from Dr. E. von Rebeur-Paschwitz, and what I have seen in the publications of Dr. Agamennone and other Italian observers, and the conclusion is that these tremors are the reappearance of a phenomenon which has for so many years puzzled seismologists in Japan. If this is so, and if they really possess the abnormally high velocities attributed to them, seismologists may be on the verge of probing our earth to depths greater than it was thought probable that the study of earthquakes could possibly lead. Although something farther may yet be learned by studying the elastic gravitational surface disturbances, we know that whether they are recorded on paths of about 600 miles in length in Japan, or on paths of 6,000 miles in length between Japan and Europe, they travel at a rate of about 3 km. per second. All that is now required is to increase the accuracy of the observations by adopting such methods of noting the arrival of these disturbances that the records at each station refer to the same phase of motion. Before the short list below was completed, I unfortunately lost my library and everything else by fire. It is therefore possible that some of my quotations may be incomplete and perhaps inaccurate. Such writings as have been referred to, so far as I am able to give them, are as follows:- References. Dr. A. Schmidt (Stutt- Wellenbewegung und Erdbeben. Jahreshefte des Vereins gart). P. Tacchini Dott. A. Cancani für vaterl. Naturkunde in Württ.,' 1888. Untersuchungen über zwei neuere Erdbeben, das Schweizerische vom 7. Januar 1889 und das Nord-Amerikanische vom 31. August 1886. Ibid., 1890. Terremoto calabro-messinese del 16 Novembre 1894. Sulla registrazione a Roma del terremoto calabro-messinese del 16 Novembre 1894. Ibid., p. 365. Sulla velocità di propagazione del terremoto di Constantinopoli del 10 Luglio 1894. Ibid., p. 409. Sugli strumenti più adatti allo studio delle grandi ondulazione provenienti da centri sismici lontani. Ibid., p. 551. Sulle ondulazioni provenienti da centri sismici lontani. 'Annali dell Officio Centrale di Meteorologia e Geodinamica,' vol. 15, part 1, 1893. Dott. G. Agamennone. Velocità di propagazione delle principali scosse di terre moto di Zante nel recente periodo sismico del 1893. 'Reale Accademia dei Lincei,' vol. 2, p. 392. Dott. G. Agamennone Alcune considerazioni sulla velocità di propagazione delle principali scosse di terremoto di Zante nel 1893. Ibid., vol. 3, p. 383. Robert Mallet William Hopkins Prof. Simon Newcomb and Capt. Dutton, C. E. Alcune considerazioni sui differenti metodi fino ad oggi adoperati nel calcolare la velocità di propagazione del terremoto andaluso del 25 Dicembre 1884. Ibid., p. 303. Velocità superficiale di propagazioni delle onde sismiche, Report on the Facts of Earthquake Phenomena. 'British Report of the Experiments made at Holyhead, &c. Ibid., Report on the Geological Theories of Elevation and Earthquakes. Ibid., 1847. The Speed of Propagation of the Charleston Earthquake. Am. Journ. of Science,' vol. 35, January, 1888. Also other publications lost by fire. F. Fouqué et Michel Lévy Expériences sur la vitesse de propagation des secousses dans les sols divers. L'Académie des Sciences de l'Institut de France.' Tome 30. Earthquakes and Earthquake Sounds as illustrating the Dr. E. von Rebeur-Pasch- Many papers lost by fire. witz. V. MISCELLANEOUS NOTES RELATING TO LARGE EARTHQUAKES, &c. 1. Large Earthquakes.-During the past twelve months Japan has been visited by several destructive earthquakes, seismographic records of which are given in the list of shocks recorded by the Gray-Milne seismograph in Tokio. The velocities with which certain of these disturbances were propagated in Japan will be found in the fourth section of this report. At least three of the shocks were recorded in Europe, and to determine the velocities with which they were propagated diagrams and notes relating to observations made in Japan have been forwarded to Dr. E. von Rebeur-Paschwitz and Dr. P. Tacchini. The most notable disturbances were as follows. June 20, 1894. Reference was made to this shock in the report for 1894. In Tokio it was felt very severely, destruction being equally great amongst foreign-built brick buildings on the high ground, and amongst similar structures in the foreign concession on the low ground. An excellent diagram, obtained from a seismograph without multiplication, has been published in the 'Journal of the College of Science' (Imperial University of Japan). The after shocks were extremely few in number. October 22, 1894, 5.20 P.M. This shock, which created great destruction within an area not more than thirty miles in diameter round Shonai on the N.W. coast of the main island, does not appear to have been recorded by the Gray-Milne seismograph in Tokio. The destruction it occasioned in the vicinity of its origin was enormous. More than 300 people lost their lives, while in some respects the country was more fissured and broken up than it was around Gifu in 1891, at which time nearly the whole of Japan was sensibly shaken. Sand hills or dunes, which have a breadth of 3,000 to 4,000 feet at their base, were fissured and sunk along their crests for a breadth of between 200 and 300 feet, and these openings extended for several miles. Fissures of great length were formed in the plains, whilst water and sand were shot upwards, and ring-like craterlets produced. One of the most curious phenomena was the filling up of wells with sand, and the shooting upwards for a height of several feet of their wooden linings. The after shocks were few in number. January 18, 1895. At 10.48 P.M. on this date Tokio was again severely disturbed, and from the feelings of the inhabitants it was difficult to say whether the movement was more or less severe than that of June. The fact that many buildings which escaped the latter shock were on this occasion more or less shattered suggests the idea that the distribution of movement throughout the city was somewhat different. The origin of the disturbance was apparently from 60 to 100 miles to the north or north-east of Tokio, and from this centre the preliminary tremors recorded by a seismograph outraced the main shock by 6 or 8 seconds. Had the writing pointer of the seismograph recorded its movements photographically, it is likely that this interval would have heen increased. It will be of great importance to determine the interval between the preliminary tremors and the elastic surface gravitational waves for this shock as recorded in Europe. Owing to the number of destructive earthquakes which occurred prior to the three here mentioned, so many observations have been accumulated that up to the present no time has been available for their analysis. The observations and notes collected by the writer relating to disturbances which have taken place during the last few years, which in themselves would have formed a voluminous report, were unfortunately destroyed by a fire referred to in the next paragraph. 2. The Destruction of Books and Pamphlets relating to Seismology.-It is with regret that I have to announce that on February 17 my house and observatory were entirely destroyed by fire. The losses consisted of collections of books, instruments, and other things accumulated during the last twenty years, the stock of the Transactions of the Seismological Society, which at the time were packed ready for shipment to Europe, and about 1,500 books and pamphlets relating to Earthquake and Volcanic Phenomena. All that I saved was the clockwork of a new seismograph and a bundle of photograms. The analysis of the latter forms the chief portion of this report. 3. Alterations in the Construction of Chimneys.-One effect of the recent earthquakes in Tokio has been to cause householders to rebuild the upper part of their chimneys with thin iron plate, while factory chimneys from 50 to 100 feet in height have for a length of 20 or 30 feet at about two-thirds up from their base been strengthened with a series of strong iron bands connected vertically by iron straps, it being observed that it |