THE INTERNATIONAL SCIENTIFIC SERIES
VOLUME LV

THE INTERNATIONAL SCIENTIFIC SERIES

EARTHQUAKES
AND
OTHER EARTH MOVEMENTS

BY
JOHN MILNE
PROFESSOR OF MINING AND GEOLOGY IN THE IMPERIAL COLLEGE OF ENGINEERING,
TOKIO, JAPAN

WITH THIRTY-EIGHT FIGURES

NEW YORK
D. APPLETON AND COMPANY
1, 3, and 5 BOND STREET
1886

PREFACE

In the following pages it has been my object to give a systematic account of various Earth Movements.

These comprise Earthquakes, or the sudden violent movements of the ground; Earth Tremors, or minute movements which escape our attention by the smallness of their amplitude; Earth Pulsations, or movements which are overlooked on account of the length of their period; and lastly, Earth Oscillations, or movements of long period and large amplitude which attract so much attention from their geological importance.

It is difficult to separate these Earth Movements from each other, because they are phenomena which only differ in degree, and which are intimately associated in their occurrence and in their origin.

Because Earthquakes are phenomena which have attracted a universal attention since the earliest times, and about them so many observations have been made, they are treated of at considerable length.

As very much of what might be said about the other Earth Movements is common to what is said about Earthquakes, it has been possible to make the description of these phenomena comparatively short.

The scheme which has been adopted will be understood from the following table:—

In some instances the grouping of phenomena according to the above scheme may be found inaccurate, as, for example, in the chapters referring to the effects and causes of Earthquakes.

This arises from the fact that the relationship between Earthquakes and other Earth phenomena are not well understood. Thus the sudden elevation of a coast line and an accompanying earthquake may be related, either as effect and cause, or vice versâ, or they may both be the effect of a third phenomenon.

Much of what is said respecting Earthquake motion will show how little accurate knowledge we have about these disturbances. Had I been writing in England, and, therefore, been in a position to make references to libraries and persons who are authorities on subjects connected with Seismology, the following pages might have been made more complete, and inaccuracies avoided. A large proportion of the material embodied in the following pages is founded on experiments and observations made during an eight years’ residence in Japan, where I have had the opportunity of recording an earthquake every week.

The writer to whom I am chiefly indebted is Mr. Robert Mallet. Not being in a position to refer to original memoirs, I have drawn many illustrations from the works of Professor Karl Fuchs and M. S. di Rossi. These, and other writers to whom reference has been made, are given in an appendix.

For seeing these pages through the press, my thanks are due to Mr. Thomas Gray, who, when residing in Japan, did so much for the advancement of observational Seismology.

For advice and assistance in devising experiments, I tender my thanks to my colleagues, Professor T. Alexander, Mr. T. Fujioka, and to my late colleague. Professor John Perry.

For assistance in the actual observation of Earthquakes, I have to thank my friends in various parts of Japan, especially Mr. J. Bissett and Mr. T. Talbot, of Yokohama. For assistance in obtaining information from Italian sources I have to thank Dr. F. Du Bois, from German sources Professor C. Netto, and from Japanese sources Mr. B. H. Chamberlain. For help in carrying out experiments, I am indebted to the liberality of the British Association, the Geological Society of London, the Meteorological and Telegraph departments of Japan, and to the officers of my own institution, the Imperial College of Engineering.

And, lastly, I offer my sincere thanks to those gentlemen who have taken part in the establishment and working of the Seismological Society of Japan, and to my publishers, whose liberality has enabled me to place the labours of residents in the Far East before the European public.

John Milne.

Tokio, Japan; June 30, 1883

CONTENTS.

EARTHQUAKES.

CHAPTER I.
INTRODUCTION.

Relationship of man to nature—The aspect of a country is dependent on geological phenomena—Earthquakes an important geological phenomenon—Relationship of seismology to the sciences and arts—Earth movements other than earthquakes—Seismological literature—(Writings of Perrey, Mallet, Eastern writings, the Philosophical Transactions of the Royal Society, the ‘Gentleman’s Magazine,’ the Bible, Herodotus, Pliny, Hopkins, Von Hoff, Humboldt, Schmidt, Seebach, Lasaulx, Fuchs, Palmieri, Bertelli, Seismological Society of Japan)—Seismological terminology.

In bygone superstitious times lightning and thunder were regarded as supernatural visitations. But as these phenomena became better understood, and men learned how to avoid their destructive power, the superstition was gradually dispelled. Thus it is with Earthquakes: the more clearly they are understood, the more confident in the universality of law will man become, and the more will his mental condition be advanced.

In his ‘History of Civilisation in England,’ Buckle has laid considerable stress upon the manner in which earthquakes, volcanoes, and other of the more terrible forms in which the workings of nature reveal themselves to us, have exerted an influence upon the imagination and understanding; and just as a sudden fright may affect the nerves of a child for the remainder of its life, we have in the annals of seismology records which indicate that earthquakes have not been without a serious influence upon the mental condition of whole communities.

To a geologist there are perhaps no phenomena in nature more interesting than earthquakes, the study of which is called Seismology. Coming, as shocks often will, from a region of volcanoes, the study of these disturbances may enable us to understand something about the nature and working of a volcano. As an earthquake wave travels along from strata to strata, if we study its reflections and changing velocity in transit, we may often be led to the discovery of certain rocky structures buried deep beneath our view, about which, without the help of such waves, it would be hopeless ever to attain any knowledge.

By studying the propagation of earthquake waves the physicist is enabled to confirm his speculations respecting the transmission of disturbances in elastic media. For the physicist earthquakes are gigantic experiments which tell him the elastic moduli of rocks as they exist in nature, and when properly interpreted may lead him to the proper comprehension of many ill-understood phenomena. It is not impossible that seismological investigation may teach us something about the earth’s magnetism, and the connection between earthquakes and the ‘earth currents’ which appear in our telegraph wires. These and numerous other kindred problems fall within the domain of the physicist.

It is of interest to the meteorologist to know the connections which probably exist between earthquakes and the fluctuations of the barometer, the changes of the thermometer, the quantity of rainfall, and like phenomena to which he devotes his attention.

Next we may turn to the more practical aims of seismology and ask ourselves what are the effects of earthquakes upon buildings, and how, in earthquake-shaken countries, the buildings are to be made to withstand them. Here we are face to face with problems which demand the attention of engineers and builders. To attain what we desire, observation, common sense, and subtle reasoning must be brought to bear upon this subject.

In the investigation of the principle on which earthquake instruments make their records, in the analysis of the results they give, in problems connected with astronomy, with physics, and with construction, seismology offers to the mathematician new fields for investigation.

A study of the effects which earthquakes produce on the lower animals will not fail to interest the student of natural history.

A study like seismology, which leads us to a more complete knowledge of earth-heat and its workings, is to be regarded as one of the corner-stones of geology. The science of seismology invites the co-operation of workers and thinkers in almost every department of natural science.

We have already referred to the influence exerted by earthquakes over the human mind. How to predict earthquakes, and how to escape from their dangers, are problems which, if they can be solved, are of extreme interest to the world at large.

In addition to the sudden and violent movements which we call earthquakes, the seismologist has to investigate the smaller motions which we call earth tremors. From observations which have been made of late years, it would appear that the ground on which we dwell is incessantly in a state of tremulous motion.

A further subject of investigation which is before the seismologist is the experimental verification of the existence of what may be called ‘earth-pulsations.’ These are motions which mathematical physicists affirmed the existence of, but which, in consequence of the slowness of their period, have hitherto escaped observation.

The oscillations, or slow changes in the relative positions of land and sea, might also be included; but this has already been taken up as a separate branch of geology.

These four classes of movements are no doubt interdependent, and seismology in the widest sense might conveniently be employed to include them all. In succeeding chapters we will endeavour to indicate how far the first three of these branches have been prosecuted, and to point out that which remains to be accomplished. It is difficult, however, to form a just estimate of the amount of seismological work which has been done, in consequence of the scattered and uncertain nature of many of the records. Seismology, as a science, originated late, chiefly owing to the facts that centres of civilisation are seldom in the most disturbed regions, and that earthquake-shaken countries are widely separated from each other.

As every portion of the habitable globe appears to have been shaken more or less by earthquakes, and as these phenomena are so terrible in their nature, we can readily understand why seismological literature is extensive. In the annals of almost every country which has a written history, references are made to seismic disturbances.

An idea of the attention which earthquakes have received may be gathered from the fact that Professor Alexis Perrey, of Dijon, who has published some sixty memoirs on this subject, gave, in 1856, a catalogue of 1,837 works devoted to seismology.[1] In 1858 Mr. Robert Mallet published in the Reports of the British Association a list of several hundred works relating to earthquakes. Sixty-five of these works are to be found in the British Museum. So far as literature is concerned, earthquakes have received as much attention in the East as in the West. In China there are many works treating on earthquakes, and the attention which these phenomena received may be judged of from the fact that in a.d. 136 the Government appointed a commission to inquire into the subject. Even the isolated empire of Japan can boast of at least sixty-five works on earthquakes, seven of which are earthquake calendars, and twenty-three earthquake monographs.[2] Besides those treating especially of earthquakes, there are innumerable references to such disturbances in various histories, in the transactions of learned societies, and in periodicals. To attempt to give a complete catalogue of even the books which have been written would be to enter on a work of compilation which would occupy many years, and could never be satisfactorily finished.

In the ‘Philosophical Transactions of the Royal Society,’ which were issued in the eighteenth century, there are about one hundred and eighty separate communications on earthquakes; and in the ‘Gentleman’s Magazine’ for 1755 there are no less than fifty notes and articles on the same subject. The great interest shown in earthquakes about the years 1750–60 in England, was chiefly due to the terrible calamity which overtook Lisbon in 1755, and to the fact that about this time several shocks were experienced in various parts of the British Islands. In 1750, which may be described as the earthquake year of Britain, ‘a shock was felt in Surrey on March 14; on the 18th of the same month the whole of the south-west of England was disturbed. On April 2, Chester was shaken; on June 7, Norwich was disturbed; on August 23, the inhabitants of Lancashire were alarmed; and on September 30 ludicrous and alarming scenes occurred in consequence of a shock having been felt during the hours of Divine service, causing the congregations to hurry into the open air.’[3] As might be expected, these occurrences gave rise to many articles and notes directing attention to the subject of earthquakes.

Seismic literature has not, however, at all times been a measure of seismic activity: thus, in Japan, the earthquake records for the twelfth and sixteenth centuries scarcely mention any shocks. At first sight it might be imagined that this was owing to an absence of earthquakes; but it is sufficiently accounted for by the fact that at that time the country was torn with civil war, and matters more urgent than the recording of natural phenomena engaged the attention of the inhabitants. Professor Rockwood, who has given so much attention to seismic disturbances in America, tells us that during the recent contest between Chili and Peru a similar intermission is observable. We see, therefore, that an absence of records does not necessarily imply an absence of the phenomena to be recorded.

Perhaps the earliest existing records of earthquakes are those which occur in the Bible. The first of these, which we are told occurred in Palestine, was in the reign of Ahab (b.c. 918–897).[4] One of the most terrible earthquakes mentioned in the Bible is that which took place in the days of Uzziah, king of Judah (b.c. 811–759), which shook the ground and rent the Temple. The awful character of this, and the deep impression produced on men’s minds, may be learned from the fact that the time of its occurrence was subsequently used as an epoch from which to reckon dates.

The writings of Herodotus, Pliny, Livy, &c., &c., show the interest which earthquakes attracted in early ages. These writers chiefly devoted themselves to references and descriptions of disastrous shocks, and to theories respecting the cause of earthquakes.

The greater portion of the Japanese notices of earthquakes is simply a series of anecdotes of events which took place at the time of these disasters. We also find references to superstitious beliefs, curious occurrences, and the apparent connection between earthquake disturbances and other natural phenomena. In these respects the literature of the East closely resembles that of the West. The earthquake calendars of the East, however, form a class of books which can hardly be said to find their parallel in Europe;[5] while, on the other hand, the latter possesses types of books and pamphlets which do not appear to have a parallel elsewhere. These are the more or less theological works—‘Moral Reflections on Earthquakes,’ ‘Sermons’ which have been preached on earthquakes, ‘Prayers’ which have been appointed to be read.[6]

Speaking generally, it may be said that the writings of the ancients, and those of the Middle Ages, down to the commencement of the nineteenth century, tended to the propagation of superstition and to theories based on speculations with few and imperfect facts for their foundation.

Among the efforts which have been made in modern times to raise seismology to a higher level, is that of Professor Perrey, of Dijon, who commenced in 1840 a series of extensive catalogues embracing the earthquakes of the world. These catalogues enabled Perrey, and subsequently Mallet in his reports to the British Association, to discuss the periodicity of earthquakes, with reference to the seasons and to other phenomena, in a more general manner than it had been possible for previous workers to accomplish. The facts thus accumulated also enabled Mallet to discuss earthquakes in general, and the various phenomena which they present were sifted and classified for inspection. Another great impetus which observational seismology received was Mr. Mallet’s report upon the Neapolitan earthquake of 1857, in which new methods of seismic investigation were put forth. These have formed the working tools of many subsequent observers, and by them, as well as by his experiments on artificially produced disturbances, Mallet finally drew the study of earthquakes from the realms of speculation by showing that they, like other natural phenomena, were capable of being understood and investigated.

In addition to Perrey and Mallet, the nineteenth century has produced many writers who have taken a considerable share in the advancement of seismology. There are the catalogues of Von Hoff, the observations of Humboldt, the theoretical investigations of Hopkins, the monographs of Schmidt, Seebach, Lasaulx, and others; the books of Fuchs, Credner, Vogt, Volger; the records and observations of Palmieri, Bertelli, Rossi, and other Italian observers. To these, which are only a few out of a long list of names, may be added the publications of the Commission appointed for the observation of earthquakes by the Natural History Society of Switzerland, and the volumes which have been published by the Seismological Society of Japan.

Before concluding this chapter it will be well to define a few of the more ordinary terms which are used in describing earthquake phenomena. It may be observed that the English word earthquake, the German erdbeben, the French tremblement de terre, the Spanish terremoto, the Japanese jishin &c., all mean, when literally translated, earth-shaking, and are popularly understood to mean a sudden and more or less violent disturbance.

Seismology (σειμός an earthquake, λόγος a discourse) in its simplest sense means the study of earthquakes. To be consistent with a Greek basis for seismological terminology, some writers have thrown aside the familiar expression ‘earthquake,’ and substituted the awkward word ‘seism.’

The source from which an earthquake originates is called the ‘origin,’ ‘focal cavity,’ or ‘centrum.’

The point or area on the surface of the ground above the origin is called the ‘epicentrum.’ The line joining the centrum and epicentrum is called the ‘seismic vertical.’

The radial lines along which an earthquake may be propagated from the centrum are called ‘wave paths.’

The angle which a wave path, where it reaches the surface of the earth, makes with that surface is called the ‘angle of emergence’ of the wave. This angle is usually denoted by the letter e.

As the result of a simple explosion at a point in a homogeneous medium, we ought, theoretically, to obtain at points on the surface of the medium equidistant from the epicentrum, equal mechanical effects. These points will lie on circles called ‘isoseismic’ or ‘coseismic’ circles. The area included between two such circles is an ‘isoseismic area.’ In nature, however, isoseismic lines are seldom circles. Elliptical or irregular curves are the common forms.

The isoseismic area in which the greatest disturbance has taken place is called the ‘meizoseismic area.’ Seebach calls the lines enclosing this area ‘pleistoseists.’

These last-mentioned lines are wholly due to Mallet and Seebach.

Many words are used to distinguish different kinds of earthquakes from each other. All of these appear to be very indefinite and to depend upon the observer’s feelings, which, in turn, depend upon his nervous temperament and his situation.

In South America small earthquakes, consisting of a series of rapidly recurring vibratory movements not sufficiently powerful to create damage, are spoken of as trembelores.

The terremotos of South America are earthquakes of a destructive nature, in which distinct shocks are perceptible. It may be observed that shocks which at one place would be described as terremoto would at another and more distant place probably be described as trembelores.

The succussatore are the shocks where there is considerable vertical motion. The terrible shock of Riobamba (February 4, 1797), which is said to have thrown corpses from their graves to a height of 100 feet, was an earthquake of this order.

The vorticosi are shocks which have a twisting or rotatory motion.

Another method of describing earthquakes would be to refer to instrumental records. When the vibrations of the ground have only been along the line joining the observer and the epicentrum, the disturbance might be called ‘euthutropic.’ A disturbance in which the prominent movements are transverse to the above direction might be called ‘diagonic.’ If motions in both of these directions occur in the records, the shock might be said to be ‘diastrophic.’ If there be much vertical movement, the shock might be said to be ‘anaseismic.’ Some disturbances could only be described by using two or three of these terms.

CHAPTER II.
SEISMOMETRY.

Nature of earthquake vibrations—Many instruments called seismometers only seismoscopes—Eastern seismoscopes, columns, projection seismometers—Vessels filled with liquid—Palmieri’s mercury tubes—The ship seismoscope—The cacciatore—Pendulum instruments of Kreil, Wagner, Ewing, and Gray—Bracket seismographs—West’s parallel motion instrument—Gray’s conical pendulums, rolling spheres, and cylinders—Verbeck’s ball and plate seismograph—The principle of Perry and Ayrton—Vertical motion instruments—Record receivers—Time-recording apparatus—The Gray and Milne seismograph.

Before we discuss the nature of earthquake motion, the determination of which has been the aim of modern seismological investigation, the reader will naturally look for an account of the various instruments which have been employed for recording such disturbances. A description of the earthquake machines which have been used even in Japan would form a bulky volume. All that we can do, therefore, is to describe briefly the more prominent features of a few of the more important of these instruments. In order that the relative merits of these may be better understood, we may state generally that modern research has shown a typical earthquake to consist of a series of small tremors succeeded by a shock, or series of shocks, separated by more or less irregular vibrations of the ground. The vibrations are often both irregular in period and in amplitude, and they have a duration of from a few seconds to several minutes. We will illustrate the records of actual earthquakes in a future chapter, but in the meantime the idea that an earthquake consists of a single shock must be dismissed from the imagination.

To construct an instrument which at the time of an earthquake shall move and leave a record of its motion, there is but little difficulty. Contrivances of this order are called seismoscopes. If, however, we wish to know the period, extent, and direction of each of the vibrations which constitutes an earthquake, we have considerable difficulty. Instruments which will in this way measure or write down the earth’s motions are called seismometers or seismographs.

Many of the elaborate instruments supplemented with electro-magnetic and clockwork arrangements are, when we examine them, nothing more than elaborate seismoscopes which have been erroneously termed seismographs.

The only approximations to true seismographs which have yet been invented are without doubt those which during the past few years have been used in Japan. It would be a somewhat arbitrary proceeding, however, to classify the different instruments as seismoscopes, seismometers, and seismographs, as the character of the record given by certain instruments is sometimes only seismoscopic, whilst at other times it is seismometric, depending on the nature of the disturbance. Many instruments, for instance, would record with considerable accuracy a single sudden movement, but would give no reliable information regarding a continued shaking.

Eastern Seismoscopes.—The earliest seismoscope of which we find any historical record is one which owes its origin to a Chinese called Chôko. It was invented in the year a.d. 136. A description is given in the Chinese history called ‘Gokanjo,’ and the translation of this description runs as follows:—

Fig. 1.

‘In the first year of Yōka, a.d. 136, a Chinese called Chôko invented the seismometer shown in the accompanying drawing. This instrument consists of a spherically formed copper vessel, the diameter of which is eight feet. It is covered at its top, and in form resembles a wine-bottle. Its outer part is ornamented by the figures of different kinds of birds and animals, and old peculiar-looking letters. In the inner part of this instrument a column is so suspended that it can move in eight directions. Also, in the inside of the bottle, there is an arrangement by which some record of an earthquake is made according to the movement of the pillar. On the outside of the bottle there are eight dragon heads, each of which holds a ball in its mouth. Underneath these heads there are eight frogs so placed that they appear to watch the dragon’s face, so that they are ready to receive the ball if it should be dropped. All the arrangements which cause the pillar to knock the ball out of the dragon’s mouth are well hidden in the bottle.’

‘When an earthquake occurs, and the bottle is shaken, the dragon instantly drops the ball, and the frog which receives it vibrates vigorously; any one watching this instrument can easily observe earthquakes.’

With this arrangement, although one dragon may drop a ball, it is not necessary for the other seven dragons to drop their balls unless the movement has been in all directions; thus we can easily tell the direction of an earthquake.

‘Once upon a time a dragon dropped its ball without any earthquake being observed, and the people therefore thought the instrument of no use, but after two or three days a notice came saying that an earthquake had taken place at Rōsei. Hearing of this, those who doubted the use of this instrument began to believe in it again. After this ingenious instrument had been invented by Chōko, the Chinese Government wisely appointed a secretary to make observations on earthquakes.’

Not only is this instrument of interest on account of its antiquity, but it is also of interest on account of the close resemblance it bears to many of the instruments of modern times.

Another earthquake instrument also of Eastern origin is the magnetic seismoscope of Japan.

On the night of the destructive earthquake of 1855, which devastated a great portion of Tokio, the owner of a spectacle shop in Asakusa observed that a magnet dropped some old iron nails and keys which had been attached to it. From this occurrence the owner thought that the magnet had, in consequence of its age, lost its powers. About two hours afterwards, however, the great earthquake took place, after which the magnet was observed to have regained its powers. This occurrence led to the construction of the seismoscope, which is illustrated in a book called the ‘Ansei-Kembun-Roku,’ or a description of the earthquake of 1855, and examples of the instrument are still to be seen in Tokio. These instruments consist of a piece of magnetic iron ore, which holds up a piece of iron like a nail. This nail is connected, by means of a string, with a train of clockwork communicating with an alarm. If the nail falls a catch is released and the clockwork set in motion, and warning given by the ringing of a bell. It does not appear that this instrument has ever acted with success.

Columns.—One of the commonest forms of seismoscope, and one which has been very widely used, consists of a round column of wood, metal, or other suitable material, placed, with its axis vertical, on a level plane, and surrounded by some soft material such as loose sand to prevent it rolling should it be overturned. The fall of such a column indicates that a shaking or shock has taken place. Attempts have been made by using a number of columns of different sizes to make these indications seismometric, but they seldom give reliable information either as to intensity or direction of shock. The indications as to intensity are vitiated by the fact that a long-continued gentle shaking may overturn a column which would stand a very considerable sudden shock, while the directions in which a number of columns fall seldom agree owing to the rotational motion imparted to them by the shaking. Besides, the direction of motion of the earthquake seldom remains in the same azimuth throughout the whole disturbance.

An extremely delicate, and at the same time simple form of seismoscope may be made by propping up strips of glass, pins, or other easily overturned bodies against suitably placed supports. In this way bodies may be arranged, which, although they can only fall in one direction, nevertheless fall with far less motion than is necessary to overturn any column which will stand without lateral support.

Projection Seismometers.—Closely related to the seismoscopes and seismometers which depend on the overturning of bodies. Mallet has described two sets of apparatus whose indications depend on the distance to which a body is projected. In one of these, which consisted of two similar parts arranged at right angles, two metal balls rest one on each side of a stop at the lower part of two inclined

like troughs. In this position each of the balls completes an electric circuit. By a shock the balls are projected or rolled up the troughs, and the height to which they rise is recorded by a corresponding interval in the break of the circuits. The vertical component of the motion is measured by the compression of a spring which carries the table on which this arrangement rests. In the second apparatus two balls are successively projected, one by the forward swing, and the other by the backward swing of the shock. Attached to them are loose wires forming terminals of the circuits. They are caught in a bed of wet sand in a metal trough forming the other end of the circuit. The throw of the balls as measured in the sand, and the difference of time between their successive projections as indicated by special contrivances connected with the closing of the circuits, enables the observer to calculate the direction of the wave of shock, its velocity, and other elements connected with the disturbance. It will be observed that the design of this apparatus assumes the earthquake to consist of a distinct isolated shock.

Oldham, at the end of his account of the Cachar earthquake of 1869, recommends the use of an instrument based on similar principles. In his instrument four balls like bullets are placed in notches cut in the corners of the upper end of a square stake driven into the ground.

Vessels filled with liquid.—Another form of simple seismoscope is made by partially filling a vessel with liquid. The height to which the liquid is washed up the side of the vessel is taken as an indication of the intensity of the shock, and the line joining the points on which maximum motion is indicated, is taken as the direction of the shock. If earthquakes all lasted for the same length of time, and consisted of vibrations of the same period, such instruments might be of service. These instruments have, however, been in use from an early date. In 1742 we find that bowls of water were used to measure the earthquakes which in that year alarmed the inhabitants of Leghorn. About the same time the Rev. S. Chandler, writing about the shock at Lisbon, tells us that earthquakes may be measured by means of a spherical bowl about three or four feet in diameter, the inside of which, after being dusted over with Barber’s puff, is filled very gently with water. Mallet, Babbage, and De la Bêche have recommended the same sort of contrivance, but, notwithstanding, it has justly been criticised as ‘ridiculous and utterly impracticable.’[7]

An important portion of Palmieri’s well-known instrument consists of horizontal tubes turned up at the ends and partially filled with mercury. To magnify the motion of the mercury, small floats of iron rest on its surface. These are attached by means of threads to a pulley provided with indices which move in front of a scale of degrees. We thus read off the intensity of an earthquake as so many degrees, which means so many millimetres of washing up and down of mercury in a tube. The direction of movement is determined by the azimuth of the tube which gives the maximum indication, several tubes being placed in different azimuths.

This form of instrument appears to have been suggested by Mallet, who gives an account of the same in 1846. Inasmuch as the rise and fall of the mercury in such tubes depend on its depth and on the period of the earthquake together with its duration, we see that although the results obtained from a given instrument may give us means of making approximate comparisons as to the relative intensity of various earthquakes, it is very far from yielding any absolute measurement.

Another method which has been employed to magnify and register the motions of liquid in a vessel has been to float upon its surface a raft or ship from which a tall mast projected. By a slight motion of the raft, the top of the mast vibrated through a considerable range. This motion of the mast as to direction and extent was then recorded by suitable contrivances attached to the top of the mast.

A very simple form of liquid seismoscope consists of a circular trough of wood with notches cut round its side. This is filled with mercury to the level of the notches. At the time of an earthquake the maximum quantity of mercury runs over the notches in the direction of greatest motion. This instrument, which has long been used in Italy, is known as a Cacciatore, being named after its inventor. It is a prominent feature in the collection of apparatus forming the well-known seismograph of Palmieri.

Pendulum instruments.—Mallet speaks of pendulum seismoscopes and seismographs as ‘the oldest probably of seismometers long set up in Italy and southern Europe.’ In 1841 we find these being used to record the earthquake disturbances at Comrie in Scotland.

These instruments may be divided into two classes: first, those which at the time of the shock are intended to swing, and thus record the direction of movement; and second, those which are supposed to remain at rest and thus provide ‘steady points.’

To obtain an absolutely ‘steady point’ at the time of an earthquake, has been one of the chief aims of all recent seismological investigations.

With a style or pointer projecting down from the steady point to a surface which is being moved backward and forward by the earth, such a surface has written upon it by its own motions a record of the ground to which it is attached. Conversely, a point projecting upwards from the moving earth might be caused to write a record on the body providing the steady point, which in the class of instruments now to be referred to is supposed to be the bob of a pendulum. It is not difficult to get a pendulum which will swing at the time of a moderately strong earthquake, but it is somewhat difficult to obtain one which will not swing at such a time. During the past few years, pendulums varying between forty feet in length and carrying bobs of eighty pounds in weight, and one-eighth of an inch in length, and carrying a gun-shot, have been experimented with under a great variety of circumstances. Sometimes the supports of these pendulums have been as rigid as it is possible to make a structure from brick and mortar, and at other times they have intentionally been made loose and flexible. The indices which wrote the motions of these pendulums have been as various as the pendulums themselves. A small needle sliding vertically through two small holes, and resting its lower end on a surface of smoked glass, has on account of its small amount of friction been perhaps one of the favourite forms of recording pointers.

The free pendulums which have been employed, and which were intended to swing, have been used for two purposes: first, to determine the direction of motion from the direction of swing, and second, to see if an approximation to the period of the earth’s motion could be obtained by discovering the pendulum amongst a series of different lengths which was set in most violent motion, this probably being the one which had its natural period of swing the most nearly approximating to the period of the earthquake oscillations.

Inasmuch as all pendulums when swinging have a tendency to change the plane of their oscillation, and also as we now know that the direction of motion during an earthquake is not always constant, the results usually obtained with these instruments respecting the direction of the earth’s motion have been unsatisfactory. The results which were obtained by series of pendulums of different lengths were, for various reasons, also unsatisfactory.

Of pendulums intended to provide a steady point, from which the relative motion of a point on the earth’s surface could be recorded, there has been a great variety. One of the oldest forms consisted of a pendulum with a style projecting downwards from the bob so as to touch a bed of sand. Sometimes a concave surface was placed beneath the pendulum, on which the record was traced by means of a pencil. Probably the best form was that in which a needle, capable of sliding freely up and down, marked the relative horizontal motion of the earth and the pendulum bob on a smoked glass plate.

It generally happens that at the time of a moderately severe earthquake the whole of these forms of apparatus are set in motion, due partly to the motion of the point of support of the pendulum, and partly to the friction of the writing point on the plate.

Among these pendulums may be mentioned those of Cavallieri, Faura, Palmieri, Rossi, and numerous others. It is possible that the originators of some of these pendulums may have intended that they should record by swinging. If this is so, then so far as the determination of the actual nature of earthquake motion is concerned, they belong to a lower grade of apparatus than that in which they are here included.

A great improvement in pendulum apparatus is due to Mr. Thomas Gray of Glasgow, who suggested applying so much frictional resistance to the free swing of a pendulum that for small displacements it became ‘dead beat.’ By carrying out this suggestion, pendulum instruments were raised to the position of seismographs. The manner of applying the friction will be understood from the following description of a pendulum instrument which is also provided with an index which gives a magnification of the motion of the earth.

b b b b is a box 113 cm. high and 30 cm. by 18 cm. square. Inside this box a lead ring r, 17 cm. in diameter and 3 cm. thick, is suspended as a pendulum from the screw s. This screw passes through a small brass plate p p, which can be moved horizontally over a hole in the top of the box. These motions in the point of suspension allow the pendulum to be adjusted.

Fig. 2.

Projecting over the top of the pendulum there is a wooden arm w carrying two sliding pointers h h, resting on a glass plate placed on the top of the pendulum. These pointers are for the purpose of giving the frictional resistance before referred to. If this friction plate is smoked, the friction pointers will write upon it records of large earthquakes independently of the records given by the proper index, which only gives satisfactory records in the case of shocks of ordinary intensity. Crossing the inside of the pendulum r there is a brass bar perforated with a small conical hole at m. A stiff wire passes through m and forms the upper portion of the index i, the lower portion of which is a thin piece of bamboo. Fixed upon the wire there is a small brass ball which rests on the upper side of a second brass plate also perforated with a conical hole, which plate is fixed on the bar o o crossing the box.

If at the time of an earthquake the upper part of the index i remains steady at m, then by the motion at o, the lower end of the index which carries a sliding needle at g, will magnify the motion of the earth in the ratios m o : o g. In this instrument o g is about 17 cm.

The needle g works upon a piece of smoked glass. In order to bring the glass into contact with the needle without disturbance, the glass is carried on a strip of wood k, hinged at the back of the box, and propped up in front by a loose block of wood y. When y is removed the glass drops down with k out of contact with the needle. The box is carried on bars of wood c c, which are fixed to the ground by the stakes a a.

The great advantage of a pendulum seismograph working on a stationary plate is, that the record shows at once whether the direction of motion has been constant, or whether it has been variable. The maximum extent of motion in various directions is also easily obtained.

The disadvantage of the instrument is, that at the time of a large earthquake, owing perhaps to a slight swing in the pendulum, the records may be unduly magnified.

On such occasions, however, fairly good records may be obtained from the friction pointers, provided that the plates on which they work have been previously smoked. It might perhaps be well to use two of these instruments, one having a comparatively high frictional resistance, and hence ‘dead beat’ for large displacements.

Many attempts have been made to use a pendulum seismograph in conjunction with a record-receiving surface, which at the time of the earthquake should be kept in motion by clockwork. In this way it was hoped to separate the various vibrations of the earthquake, and thus avoid the greater or less confusion which occurs when the index of the pendulum writes its backward and forward motion on a stationary plate. Hitherto all attempts in this direction, in which a single multiplying index was used, have been unsuccessful because of the moving plate dragging the index in the direction of its motion for a short distance, and then allowing it to fall back towards its normal position.

In connection with this subject we may mention the pendulum seismographs of Kreil, Wagener, Ewing, and Gray.

In the bob of Kreil’s pendulum there was clockwork, which caused a disc on the axis of the pendulum to continuously rotate. On this continually revolving surface a style fixed to the earth traced an unbroken circle. At the time of an earthquake, by the motion of the style, the circle was to be broken and lines drawn. The number and length of these lines were to indicate the length and intensity of the disturbance.

Gray’s pendulum consisted of a flat heavy disc carrying on its upper surface a smoked glass plate. This, which formed the bob of the pendulum, was supported by a pianoforte steel wire. When set ready to receive an earthquake, the wire was twisted and the bob held by a catch so arranged that at the time of the earthquake the catch was released, and the bob of the pendulum allowed to turn slowly by the untwisting of the supporting wire. Resting on the surface of this rotating disc were two multiplying indices arranged to write the earth’s motions as two components.

In the instruments of Wagener and Ewing, the clockwork and moving surface do not form part of the pendulum, but rest independently on a support rigidly attached to the earth. In Wagener’s instrument one index only is used, while in Ewing’s two are used for writing the record of the motion.

A difficulty which is apparent in all pendulum machines is that when the bob of such a pendulum is deflected it tends to fall back to its normal position. To make a pendulum perfect it therefore requires some compensating arrangement, so that the pendulum, for small displacements, shall be in neutral equilibrium, and the errors due to swinging shall be avoided.

Several methods have been suggested for making the bob of an ordinary pendulum astatic for small displacements. One method proposed by Gray consists in fixing in the bob of a pendulum a circular trough of liquid, the curvature of this trough having a proper form. Another method which was suggested, was to attach a vertical spiral spring to a point in the axis of the pendulum a little below the point of suspension, and to a fixed point above it, so that when the pendulum is deflected it would introduce a couple.

Professor Ewing has suggested an arrangement so that the bob of the pendulum shall be partly suspended by a stretched spiral spring, and at the same time shall be partly held up from below by a vertically placed strut, the weight carried by the strut being to the weight carried by the spring in the ratio of their respective lengths. As to how these arrangements will act when carried into practice yet remains to be seen.

Another important class of instruments are inverted pendulums. These are vertical springs made of metal or wood loaded at their upper end with a heavy mass of metal. An arrangement of this sort, provided at its upper end with a pencil to write on a concave surface, was employed in 1841 to register the earthquakes at Comrie in Scotland. In Japan they were largely employed in series, each member of a series having a different period of vibration. The object of these arrangements was to determine which of the pendulums, with a given earthquake, recorded the greatest motion, it being assumed that the one which was thrown into the most violent oscillation would be the one most nearly approximating with the period of the earthquake. The result of these experiments showed that it was usually those with a slow period of vibration which were the most disturbed.

Bracket Seismographs.—A group of instruments of recent origin which have done good work, are the bracket seismographs. These instruments appear to have been independently invented by several investigators: the germ from which they originated probably being the well-known horizontal pendulum of Professor Zöllner. In Japan they were first employed by Professor W. S. Chaplin. Subsequently they were used by Professor Ewing and Mr. Gray. They consist essentially of a heavy weight supported at the extremity of a horizontal bracket which is free to turn on a vertical axis at its other end. When the frame carrying this axis is moved in any direction excepting parallel to the length of the gate-like bracket, the weight causes the bracket to turn round a line known as the instantaneous axis of the bracket corresponding to this motion of the fixed axis. Any point in this line may therefore be taken as a steady point for motions at right angles to the length of the supporting bracket. Two of these instruments placed at right angles to each other have to be employed in conjunction, and the motion of the ground is written down as two rectangular components. In Professor Ewing’s form of the instrument, light prolongations of the brackets form indices which give magnified representations of the motion, and the weights are pivoted round a vertical axis through their centre.

In the accompanying sketch b is a heavy weight pivoted at the end of a small bracket c a k, which bracket is free to turn on a knife-edge, k, above, and a pivot a, below, in the stand s. At the time of an earthquake b remains steady, and the index p, forming a continuation of the bracket, magnifies the motion of the stand, in the ratio of a c : c n.

Fig. 3.

In an instrument called a double-bracket seismograph, invented by Mr. Gray, we have two brackets hinged to each other, and one of them to a fixed frame. The planes of the two brackets are placed at right angles, so as to give to a heavy mass supported at the end of the outer bracket two degrees of horizontal freedom.

In all bracket machines, especially those which carry a pivoted weight, it is doubtful whether the weight provides a truly steady point relatively to the plate on which the record is written for motion parallel to the direction of the arm.

Fig. 4.

Parallel motion Instrument.—A machine which writes its record as two components, and which promises great stability, is one suggested by Professor C. D. West. Like the bracket machines it consists of two similar parts placed at right angles to each other, and is as follows: A bar of iron a is suspended from both sides on pivots at c c by a system of light arms hinging with each other at the black dots, between the upper and lower parts of the rigid frame b c. The arms are of such a length that for small displacements parallel to the length of the bar, c c practically move in a straight line, and the bar is in neutral equilibrium. A light prolongation of the bar d works the upper end of the light index e, passing as a universal joint through the rigid support f. A second index e′ from the bar at right angles also passes through f. The multiplying ends of these indices are coupled together to write a resultant motion on a smoked glass plate s.

Conical Pendulums.—Another group of instruments which have also yielded valuable records are the conical pendulum seismographs. The idea of using the bob of a conical pendulum to give a steady point in an earthquake machine was first suggested and carried into practice by Mr. Gray. The seismograph as employed consists of a pair of conical pendulums hung in planes at right angles to each other. The bob of each of these pendulums is fixed a short distance from the end of a light lever, which forms the writing index, the short end resting as a strut against the side of a post fixed in the earth. The weight is carried by a thin wire or thread, the upper end of which is attached to a point vertically above the fixed end of the lever.

Rolling Spheres and Cylinders.—After the conical pendulum seismographs, which claim several important advantages over the bracket machines, we come to a group of instruments known as rolling sphere seismographs. Here, again, we have a class of instruments for the various forms of which we are indebted to the ingenuity of Mr. Gray.

The general arrangement and principle of one of these instruments will be readily understood from the accompanying figure. s is a segment of a large sphere with a centre near c. Slightly below this centre a heavy weight b, which may be a lead ring, is pivoted. At the time of an earthquake c is steady, and the earth’s motions are magnified by the pointer c a n in the proportion of c a : a n. The working of this pointer or index is similar to that of the pointer in the pendulum.

Fig. 5.

Closely connected with the rolling sphere seismographs, are Gray’s rolling cylinder seismographs.

These are two cylinders resting on a surface plate with their axes at right angles to each other. Near to the highest point in each of these cylinders, this point remaining nearly steady when the surface plate is moved backwards and forwards, there is attached the end of a light index. These indices are again pivoted a short distance from their ends on axes connected with the surface plate. In order that the two indices may be brought parallel, one is cranked at the second pivot.

Ball and Plate Seismograph.—Another form of seismograph, which is closely related to the two forms of apparatus just described, is Verbeck’s ball and plate seismograph. This consists of a surface plate resting on three hard spheres, which in turn rests upon a second surface plate. When the lower plate is moved, the upper one tends to remain at rest, and thus may be used as a steady mass to move an index.

The Principle of Perry and Ayrton.—An instrument which is of interest from the scientific principle it involves is a seismograph suggested by Professors Perry and Ayrton, who propose to support a heavy ball on three springs, which shall be sufficiently stiff to have an exceedingly quick period of vibration. By means of pencils attached to the ball by levers, the motions of the ball are to be recorded on a moving band of paper. The result would be a record compounded of the small vibrations of the springs superimposed on the larger, slower, wave-like motions of the earthquake, and, knowing the former of these, the latter might be separated by analysis. Although our present knowledge of earthquake motion indicates that the analysis of such a record would often present us with insuperable difficulties, this instrument is worthy of notice on account of the novelty of the principle it involves, which, the authors truly remark, has in seismometry been a ‘neglected’ one.

Instruments to record Vertical Motion.—The instruments which have been devised to record vertical motion are almost as numerous as those which have been devised to record horizontal motion. The earliest form of instrument employed for this purpose was a spiral spring stretched by weight, which, on account of its inertia, was supposed at the time of a shock to remain steady. No satisfactory results have ever been obtained from such instruments, chiefly on account of the inconvenience in making a spring sufficiently long to allow of enough elongation to give a long period of vibration. Similar remarks may be applied to the horizontally placed elastic rods, one end of which is fixed to a wall, whilst the opposite end is loaded with a weight. Such contrivances, furnished with pencil on the weight to write a record upon a vertical surface, were used in 1842 at Comrie, and we see the same principle applied in a portion of Palmieri’s apparatus. Contrivances like these neither give us the true amplitude of the vertical motion, insomuch as they are readily set in a state of oscillation; nor do they indicate the duration of a disturbance, for, being once set in motion, they continue that motion in virtue of their inertia long after the actual earthquake has ceased. They can only be regarded as seismoscopes.

Fig. 6.

The most satisfactory instrument which has yet been devised for recording vertical motion is Gray’s horizontal lever spring seismograph.

This instrument will be better understood from the accompanying sketch. A vertical spring s is fixed at its upper end by means of a nut n, which rests on the top of the frame f, and serves to raise or lower the spring through a short distance as a last adjustment for the position of the cross-arm a. The arm a rests at one end on two sharp points, p, one resting in a conical hole and the other in a v-slot; it is supported at b by the spring s, and is weighted at c with a lead ring r. Over a pin at the point c a stirrup of thread is placed which supports a small trough, t. The trough t is pivoted at a, has attached to it the index i (which is hinged by means of a strip of tough paper at h, and rests through a fine pin on the glass plate g), and is partly filled with mercury.

Another method of obtaining a steady point for vertical motion is that of Dr. Wagener, who employs a buoy partly immersed in a vessel of water. This was considerably improved upon by Mr. Gray, who suggested the use of a buoy, which, with the exception of a long thin style, was completely sunk.

Among the other forms of apparatus used to record vertical motion may be mentioned vessels provided with india-rubber or other flexible bottoms, and partially filled with water or some other liquid. As the vessel is moved up and down, the bottom tends to remain behind and provides a more or less steady point. Pivoted to this is a light index, which is again pivoted to a rigid frame in connection with the earth. Instruments of this description have yielded good records.

Record Receivers.—A large number of earthquake machines having been referred to, it now remains to consider the apparatus on which they write their motions. The earlier forms of seismographs, as has already been indicated, recorded their movements in a bed of sand; others wrote their records by means of pencils on sheets of paper. Where we have seismographs which magnify the motion of the earth, it will be observed that methods like the above would involve great frictional resistances, tending to cause motion in the assumed steady points of the seismographs. One of the most perfect instruments would be obtained by registering photographically the motions of the recording index by the reflection of a ray of light. Such an instrument would, however, be difficult to construct and difficult to manipulate. One of the best practical forms of registering apparatus is one in which the record is written on a surface of smoked glass. This can afterwards be covered with a coat of photographer’s varnish, and subsequently photographed by the ‘blue process’ so well known to engineers.

To obtain a record of all the vibrations of an earthquake it is necessary that the surface on which the seismograph writes should at the time of an earthquake be in motion. Of record-receiving machines there are three types. First, there are those which move continuously. The common form of these is a circular glass plate like an old form of chronograph, driven continuously by clockwork. On this the pointers of the seismograph rest and trace over and over again the same circles. At the time of an earthquake they move back and forth across the circles, which are theoretically fine lines, and leave a record of the earthquake. Instead of a circular plate, a drum covered with smoked paper may be used, which, after the earthquake, possesses the advantage, after unrolling, of presenting the record in a straight line, instead of a record written round the periphery of a circle, as is the case with the circular glass plates. Such records are easily preserved, but they are more difficult to photograph.

The second form of apparatus is one which is set in motion at the time of a shock. This may be a contrivance like one of those just described, or a straight smoked glass plate on a carriage. By means of an electrical or a mechanical contrivance called a ‘starter,’ of which many forms have been contrived, the earthquake is caused to release a detent and thus set in motion the mechanism which moves the record receiver.

The great advantage of continuously-moving machines is that the beginning and end of the shock can usually be got with certainty, while all the uncertainty as to the action of the ‘starter’ is avoided. Self-starting machines have, of course, the advantage of simplicity and cheapness, while there is no danger of the record getting obliterated by the subsequent motion of the plate under the index.

Time-recording Apparatus.—Of equal importance with the instruments which record the motion of the ground, are those instruments which record the time at which such motion took place. The great value of time records, when determining the origin from which an earthquake originates, will be shown farther on. The most important result which is required in connection with time observations, is to determine the interval of time taken by a disturbance in travelling from one point to another. On account of the great velocity with which these disturbances sometimes travel, it is necessary that these observations should be made with considerable accuracy. The old methods of adapting an apparatus to a clock which, when shaken, shall cause the clock to stop, are of little value unless the stations at which the observations are made are at considerable distances apart. This will be appreciated when we remember that the disturbance may possibly travel at the rate of a mile per second, that its duration at any station may often extend over a minute, and that one set of apparatus at one station may stop, perhaps, at the commencement of the disturbance, and the other near the end. A satisfactory time-taking apparatus will therefore require, not only the means for stopping a clock, but also a contrivance which, at the same instant that the clock is stopped, shall make a mark on a record which is being drawn by a seismograph. In this way we find out at which portion of the shock the time was taken.

Fig. 7.

Palmieri stops a clock in his seismograph by closing an electric circuit. Mallet proposes to stop a clock by the falling of a column which is attached by a string to the pendulum of the clock. So long as the column is standing the string is loose and the pendulum is free to move; but when the column falls, the string is tightened and the pendulum is arrested. The difficulty which arises is to obtain a column that will fall with a slight disturbance. The best form of contrivance for causing a column to fall, and one which may also be used in drawing out a catch to relieve the machinery of a record receiver, is shown in the accompanying sketch.

s is the segment of a sphere about 4·5 cm. radius, with a centre slightly above c. l is a disc of lead about 7 cm. in diameter resting upon the segment. Above this there is a light pointer, p, about 30 cm. long. On the top of the pointer a small cylinder of iron, w, is balanced, and connected by a string with the catch to be relieved. When the table on which w p s rests is shaken, rotation takes place near to c, the motion of the base s is magnified at the upper end of the pointer, and the weight overturned. This catch may be used to relieve a toothed bar axled at one end, and held up above a pin projecting from the face of the pendulum bob. When this falls it catches the projecting pin and holds the pendulum.

Another way of relieving the toothed bar is to hold up the opposite end to that at which it is axled by resting it on the extremity of a horizontal wire fixed to the bob of a conical pendulum—for example, one of the indices of a conical pendulum seismograph. The whole of this apparatus, which may be constructed at the cost of a few pence, can be made small enough to go inside an ordinary clock case.

The difficulty which arises with all these clock-stopping arrangements is that it is difficult for observers situated at distant stations to re-start their clocks so that their difference in time shall be accurately known. Even if each observer is provided with a well-regulated chronometer, with which he can make comparisons, the rating of these instruments is for all ordinary persons an extremely troublesome operation.

In order to avoid this difficulty the author has of late years used a method of obtaining the time without stopping the clock. To do this a clock with a central seconds hand is taken, and the hour and minute hands are prolonged and bent out slightly at their extremities at right angles to the face, the hour hand being slightly the longest. Each hand is then tipped with a piece of soft material like cork, which is smeared with a glycerine ink. A light flat ring, with divisions in it corresponding to those on the face of the clock, is so arranged that at the time of a shock it can be quickly advanced to touch the inked pads on the hands of the clock and then withdrawn. This is accomplished by suitable machinery, which is relieved either by an electro-magnet or some other contrivance which will withdraw a catch. In this way an impression in the form of three dots is received on the disc, and the time known without either stopping or sensibly retarding the clock.

For ordinary observers, if a time-taker is not used in conjunction with a record receiver, as good results as those obtained by ordinary clock-stopping apparatus are obtainable by glancing at an ordinary watch. Subsequently the watch by which the observation was made should be compared with some good time-keeper, and the local time at which the shock took place is then approximately known.

From what has now been said it will be seen that for a complete seismograph we require three distinct sets of apparatus—an apparatus to record horizontal motion, an apparatus to record vertical motion, and an apparatus to record time. The horizontal and vertical motions must be written on the same receiver, and if possible side by side, whilst the instant at which the time record is made a mark must be made on the edge of the diagram which is being drawn by the seismograph. Such a seismograph has been constructed and is now erected in Japan. It is illustrated in the accompanying diagram.

The Gray and Milne Seismograph.—In this apparatus two mutually rectangular components of the horizontal motion of the earth are recorded on a sheet of smoked paper wound round a drum, d, kept continuously in motion by clockwork, w, by means of two conical pendulum seismographs, c. The vertical motion is recorded on the same sheet of paper by means of a compensated-spring seismograph, s l m b.

The time of occurrence of an earthquake is determined by causing the circuit of two electro-magnets to be closed by the shaking. One of these magnets relieves a mechanism, forming part of a time-keeper, which causes the dial of the timepiece to come suddenly forwards on the hands and then move back to its original position. The hands are provided with ink-pads, which mark their positions on the dial, thus indicating the hour, minute, and second when the circuit was closed. The second electro-magnet causes a pointer to make a mark on the paper receiving the record of the motion. This mark indicates the part of the earthquake at which the circuit was closed.

Fig. 8.

The duration of the earthquake is estimated from the length of the record on the smoked paper and the rate of motion of the drum. The nature and period of the different movements are obtained from the curves drawn on the paper.

Mr. Gray has since greatly modified this apparatus, notably by the introduction of a band of paper sufficiently long to take a record for twenty-four hours without repetition. The record is written in ink by means of fine siphons. In this way the instrument, which is extremely sensitive to change of level, can be made to show not only earthquakes, but the pulsations of long period which have recently occupied so much attention.

CHAPTER III.
EARTHQUAKE MOTION DISCUSSED THEORETICALLY.

Ideas of the ancients (the views of Travagini, Hooke, Woodward, Stukeley, Mitchell, Young, Mallet)—Nature of elastic waves and vibrations—Possible causes of disturbance in the Earth’s crust—The time of vibration of an earth particle—Velocity and acceleration of a particle—Propagation of a disturbance as determined by experiments upon the elastic moduli of rocks—The intensity of an earthquake—Area of greatest overturning moment—Earthquake waves—Reflexion, refraction, and interference of waves—Radiation of a disturbance.

Ideas of Early Writers.—One of the first accounts of the varieties of motion which may be experienced at the time of an earthquake is to be found in the classification of earthquakes given by Aristotle.[8] It is as follows:—

1. Epiclintæ, or earthquakes which move the ground obliquely.

2. Brastæ, with an upward vertical motion like boiling water.

3. Chasmatiæ, which cause the ground to sink and form hollows.

4. Rhectæ, which raise the ground and make fissures.

5. Ostæ, which overthrow with one thrust.

6. Palmatiæ, which shake from side to side with a sort of tremor.

From the sixth group in this classification we see that this early writer did not regard earthquakes as necessarily isolated events, but that some of them consisted of a succession of backward and forward vibratory motions. He also distinguishes between the total duration of an earthquake and the length of, and intervals between, a series of shocks. Aristotle had, in fact, some idea of what modern writers upon ordinary earthquakes would term ‘modality.’

The earliest writer who had the idea that an earthquake was a pulse-like motion propagated through solid ground appears to have been Francisci Travagini, who, in 1679, wrote upon an earthquake which in 1667 had overthrown Ragusa. The method in which the pulses were propagated he illustrated by experiments.

Hooke, who, in 1690, delivered discourses on earthquakes before the Royal Society, divides these phenomena according to the geological effects they have produced; thus there is a genus producing elevations, a genus producing sinkings, a genus producing conversions and transportations, and a genus which produces what, in modern language, we should term metamorphic action.

Woodward, in his ‘Natural History,’ written in 1695, speaks of earthquakes as being agitations and concussions produced by water in the interior of the earth coming in contact with internal fires.

Stukeley observed that an earthquake was ‘a tremor of the earth,’ to be explained as a vibration in a solid. The Rev. John Mitchell, writing in 1760, says that the motion of the earth in earthquakes is partly tremulous and partly propagated by waves.

From these few examples, to which might be added many more, it will be seen that an earthquake disturbance has usually been regarded as a concussion, vibration, trembling, or undulatory movement. Further, it can be seen in narratives of earthquakes that it had been often observed that these tremblings and shakings continued over a certain period of time. Although it had been noticed that large areas were almost simultaneously affected by these disturbances, no definite idea appears to have existed as to how earthquake motion was propagated. Usually it was assumed that the disturbance spread through subterranean channels.

The first true conception of earthquake motion and the manner of its propagation is due to Dr. Thomas Young, who suggested that earthquake motion was vibratory, and it might be ‘propagated through the earth nearly in the same manner as a noise is conveyed through the air.’ The same idea was moulded into a more definite form by Gay Lussac.

The first accurate definition of an earthquake is due to Mr. Robert Mallet, who, after collecting and examining many facts connected with earthquake phenomena, and reasoning on these, with the help of known laws connected with the production and propagation of waves of various descriptions, formulated his views as follows:—

An earthquake is ‘the transit of a wave or waves of elastic compression in any direction from vertically upwards to horizontally, in any azimuth, through the crust and surface of the earth, from any centre of impulse or from more than one, and which may be attended with sound and tidal waves, dependent upon the impulse and upon circumstances of position as to sea and land.’

In brief, so far as motion in the earth is concerned. Mallet defined an earthquake as being a motion due to the transit of waves of elastic compression. In many cases it is possible that this is strictly true, but in succeeding pages it will be shown that earthquake motion may also be due to the transit of waves of elastic distortion.

To obtain a true idea of earthquake motion is a matter of cardinal importance, as it forms the key-stone of many investigations.

If we know the nature of the motion produced by an earthquake, we are aided in tracking it to its origin, and in reasoning as to how it was produced. If our knowledge of the nature of the motion of an earthquake is incorrect, it will be impossible for us intelligently to construct buildings to withstand the effects of these disturbances. We have thus to consider, in this portion of seismology, a point of great scientific importance, and shall deal with it at some length.

Nature of Elastic Waves and Vibrations.—When it is stated that an earthquake consists of elastic waves of compression and distortion, the student of physics has a clear idea of what is meant and a knowledge of the mechanical laws which govern such disturbances. The ordinary reader, however, and the majority of the inhabitants of earthquake countries, who of all people have the greatest interest in this matter, may not have so clear a conception, and it will, therefore, not be out of place to give some general explanation on this point.

The ordinary idea of a wave is that it is a disturbance similar to that which we often see in water. Waves like these must not, however, be confounded with elastic waves. A disturbance produced in water, say, for instance, by dropping a stone into a pond, is propagated outwards by the action of gravity. First, a ridge of water is raised up by the stone passing beneath the surface. As this ridge falls towards its normal position in virtue of its weight, it raises a second ridge. This second ridge raises a third ridge, and so on. The water moves vertically up and down, whilst the wave itself is propagated horizontally.

To understand what is meant by elastic waves, it is first necessary to understand what is meant by the term elastic. In popular language the term elastic is confined to substances like india-rubber, and but seldom to rock-like materials, through which earthquake waves are propagated. India-rubber is called elastic because after we remove a compressive force it has a tendency to spring back to its original shape. The elastic force of the india-rubber is in this case the force which causes it to resist a change of form. Now, a piece of rock may, up to a certain point, like the india-rubber, be compressed, and when the compressing force is removed it will also tend to resume its original form. However, as the rock offers more resistance to the compressing force than the india-rubber offers, we say that it is the more elastic. It may be here observed that a substance like granite offers great resistance, not only to compression or a change of volume, but also to a change of form or shape; whereas a substance like air, which is also elastic, only offers resistance to compression, but not to a change of shape.

With these ideas before us we will now proceed to consider how, after a body has been suddenly compressed or distorted, this disturbance is propagated through the mass. For the elastic body let us take a long spiral spring hung from the ceiling of a room and kept slightly stretched by a weight. If we give this weight an upward tap from below, say with a hammer, we shall observe a pulse-like wave which runs up the spring until it reaches the ceiling of the room. Here it will, so to speak, rebound, like a billiard ball from the end of a table, and run towards the weight from which it started. Whilst this is going on we may also observe that the weight is moving up and down.

Here, then, we have two distinct things to observe—one being the transmission of motion up to the ceiling, which we may liken to the transmission of an earthquake wave between two distant localities on the earth’s surface, and the other being the up and down motion of our weight, which we may compare to the backward and forward swinging which we experience at the time of an earthquake.

These two motions—namely, the pulse-like wave produced by the transmission of motion, and the backward and forward oscillation of the weight or of any point on the spring—must be carefully distinguished from each other.

First, we will consider the backward and forward motion of the weight. The distance through which the weight moves depends upon the force of the blow. The number of up and down oscillations it makes, say in a second, depends upon the stiffness of the spring. The weight, supposing it to be always the same, will move more quickly at the end of a stiff spring than at the end of a flaccid one; that is to say, its velocity is quicker. As in any given spring the number of up and down oscillations are always the same in a given interval of time, if these oscillations are of great extent, the weight must move more quickly with large than with small oscillations.

At the time of an earthquake the manner in which we are moved backward and forward is very similar to the manner in which the weight is moved. If we stand on a hard rock-like granite, we are to a great extent placed as if we were attached to a stiff quickly-vibrating spring. If, however, we are on a soft rock, it is more like being on a loose flaccid spring.

All that has thus far been considered has been a backward and forward kind of motion, where there is a rectilinear compression and extension amongst the particles on which we stand.

We might, however, imagine our rock, which for the moment we will consider to be a square column, to be twisted, and thus have its shape altered. When the twisting force is taken off it seems evident that the column would endeavour to untwist itself or regain its original form. Now the force which a body offers against a change of volume may be very different from that which it offers against a change of form.

In disturbances which take place in the rocky crust of our earth, it would seem possible that we may have vibrations set up which are either compressions and extensions or twistings and distortions. These may take place separately or simultaneously, or we may have resultant motions due to their combination.

The following are examples of possible causes which might give rise to these different orders of disturbance:—

1. Imagine a large area stretched by elevation until it reaches the limit of its elasticity and cracks. After cracking, in consequence of its elasticity, it will fly back over the whole area like a broken spring, and each point in the area will oscillate round its new position of equilibrium. In this case there will be no waves of distortion excepting near the end of the crack, where waves are transmitted in a direction parallel to the fissure.

2. The ground is broken and slips either up, down, or sideways, as we see to have taken place in the production of faults. Here we get distortion in the direction of the movement, and waves are produced by the elastic force of the rock, causing it to spring back from its distorted form. In a case like this the production of a fissure running north and south might give rise to north and south vibrations, which would be propagated end on towards the north and south, but broadside on towards the east and west. With disturbances of this kind, on account of the want of homogeneousness in the materials in which they are produced, we should expect to find waves of compression and extension.

3. A truly spherical cavity is suddenly formed by the explosion of steam in the midst of an elastic medium. In this case all the waves will be those of compression, each particle moving backward and forward along a radius.

Should the cavity, instead of being truly spherical, be irregular, it is evident that, in addition to the normal vibration of compression, transverse waves of distortion will be more or less pronounced, depending upon the nature of the cavity.

The combination of these two sets of vibrations may cause a point in the earth to move in a circle, an ellipse, the form of a figure eight, and in other curves similar to these, which are produced by apparatus designed to show the combination of harmonic motion. From these examples it will be seen that we have therefore to consider two kinds of vibrations—one produced by compression or the alteration of volume, and the other produced by an alteration in shape.

Now the resistance which a body offers, either to a change in its volume or in its shape, is called its elasticity, and the law which governs the backward and forward motion of a particle under the influence of this elasticity may be expressed as follows:

If t be the time of vibration, or the time taken by a particle to make one complete backward and forward swing, d the density of the material of which this particle forms a part, and e the proper modulus of elasticity of the material, then,

t = 2π √^d/_e

From this formula, t = 2π √^d/_e, we see that the time of vibration of the earth during an earthquake, or the rate at which we are shaken backwards and forwards, varies directly as the square root of the density of the material on which we stand, and inversely as the square root of a number proportional to its elasticity.

Velocity and Acceleration of an Earth Particle.—Another important point, which the practical seismologist has often brought to his notice, is the question of the velocity with which an earth particle moves. According to the formula, t = 2π √^d/_e, we should expect that a particle would make each semi-vibration in an equal time, and from a knowledge of the density and elastic moduli of a body this time might be calculated. Although the time of a semi-oscillation may be constant, we must bear in mind that, like the bob of a pendulum during each of its swings, the particle starts from rest, increases in velocity until it reaches the middle portion of its half swing, from which it gradually decreases in speed until it reaches zero, when it again commences a similar motion in the opposite direction.

These pendulum-like vibrations are sometimes spoken of as simple harmonic motions. If we know the distance through which an earthquake moves in making a single swing, and the time taken in making this swing, on the assumption that the motion is simple harmonic we can easily calculate the maximum velocity with which the particle moves.

Thus, if an earth particle takes one second to complete a semi-oscillation, half of which, or the amplitude of the motion, equals a, the maximum velocity equals π × a.

Again, assuming the earth vibrations to be simple harmonic, the maximum acceleration or rate of change in velocity will come about at the ends of each semi-oscillation; and if v be the maximum velocity of the particle, and a the amplitude or half semi-oscillation, then the maximum acceleration equals ^v2/_a.

Later on it will be shown, as the result of experiment, that certain of the more important earth oscillations in an earthquake are not simple harmonic motion. Nevertheless the above remarks will be of assistance in showing how the velocity and other elements connected with the motion of an earth particle, which are required by the practical seismologist, may be calculated, irrespective of assumptions as to the nature of the motion.

Propagation of a Disturbance.—We may next consider the manner in which a disturbance, in which there are both vibrations of compression and of distortion, is propagated. The first or normal set of vibrations are propagated in a manner similar to that in which sound vibrations are propagated. From a centre of disturbance these movements approach an observer at a distant station, so to speak, end on. The other vibrations have a direction of motion similar to that which we believe to exist in a ray of light. These would approach the observer broadside on.

If the disturbance passed through a formation like a series of perfectly laminated slates, each of these two sets of vibration might be subdivided, and we should then obtain what Mallet has termed ordinary and extraordinary normal and transverse vibrations.

In consequence of the difference in the elastic forces on which the propagation of these two kinds of vibration depends, the normal vibrations are transmitted faster than the transversal ones—that is to say, if an earthquake originated from a blow, the first thing that would be felt at a point distant from the origin of the shock would be a backward and forward motion in the direction to and from the origin, and then, a short interval afterwards, a motion transversal, or at right angles to this, would be experienced.

From the mathematical theory of vibratory motions it is possible to calculate the velocity with which a disturbance is propagated. As the result of these investigations it has been shown that normal vibrations travel more quickly than transverse vibrations.

Deductions from experiments on small specimens are, however, invalidated by the fact that the specimens used for experiments are, of course, nearly homogeneous, whilst the earthquake passes through a mass which is heterogeneous and more or less fissured. Mallet, by experiments ‘on the compressibility of solid cubes of these rocks, obtained the mean modulus of elasticity,’ with the result that ‘nearly seven-eighths of the full velocity of wave-transit due to the material, if solid and continuous, is lost by reason of the heterogeneity and discontinuity of the rocky masses as they are found piled together in nature.’ The full velocities of wave-transit, as calculated by Mallet from a theorem given by Poisson, were—

For slate and quartz transverse to lamination, 9,691 feet per second.
„ „ in line of lamination, 5,415 „ „

This more rapid transmission in a direction transverse to the lamination, Mr. Mallet observes, may be more than counterbalanced by the discontinuity of the mass transverse to the same direction.

The Intensity of an Earthquake.—The intensity of an earthquake is best estimated by the intensity of the forces which are brought to bear on bodies placed on the earth’s surface. These forces are evidently proportional to the rate of change of velocity in the body, and, as the destructive effect will be proportional to the maximum forces, we may consistently indicate the intensity of an earthquake by giving the maximum acceleration to which bodies were subject during the disturbance. On the assumption that the motion of a point on the earth’s surface is simple harmonic, the maximum acceleration is directly as the maximum velocity and inversely as the amplitude of motion, or as ^v2/_a where v indicates velocity and a amplitude.

The next question of importance is to determine the manner in which earthquake energy becomes dissipated—that is, to compare together the intensity of an earthquake as recorded at two or more points at different distances from the origin. First let us imagine the origin of our earthquake to be surrounded by concentric shells, each of which is the breadth of the vibration of a particle. Going outwards from the centre, each successive shell will contain a greater number of particles, this number increasing directly as the square of the distance from the origin. Let the blow have its origin at the centre, and give a vibratory movement to the particles in one of the shells near the centre.

This shell may be supposed to possess a certain amount of energy, which will be measured by its mass and the square of the velocity of its particles. In transferring this energy to the neighbouring shell which surrounds it, because it has to set in motion a greater number of particles than it contains itself, the energy in any one particle of the second layer will be less than the energy in any one particle in the first layer; the total energy in the second shell, however, will be equal to the total energy in the first shell. Neglecting the energy lost during the transfer, if the energy in a particle of the first shell at any particular phase of the motion be k₁, and the energy in a particle of the second shell k₂, these quantities are to each other inversely as the masses of the shells—that is, inversely as the squares of the mean radii of the shells.

In symbols, ^k₂/_k1 = ^r₁2/_r2² (1)

Assuming that energy is dissipated,

^k₂/_k1 > ^r₁2/_r2² = f ^r₁2/_r2² (2)

where f < 1 is the rate of dissipation of energy which is assumed to be constant.

Area of greatest Overturning Moment.—Although the rate of dissipation of the impulsive effects of an earthquake may follow a law like that just enumerated, it must be remembered that if the depth of the origin is comparable with the radius of the area which is shaken, the maximum impulsive effect as exhibited by the actual destruction on the surface may not be immediately above the origin where buildings have simply been lifted vertically up and down, but at some distance from this point, where the impulsive effort has been more oblique.

At the epicentrum we have the maximum of the true intensity as measured by the acceleration of a particle, or the height to which a body might be projected, but it will be at some distance from this where we shall have the maximum intensity as exhibited by an overturning effort.

This will be rendered clear by the following diagram.

In the accompanying diagram let o be the origin of a shock, and o c the seismic vertical equal to r. Let the direct or normal shock emerge at c, c₁, c₂, and at the angles θ₁, θ₂, &c.

Assuming that the displacement of an earth particle at c equals c b, and at c₁ equals c₁ b₁, and at c₂ equals c₂ b₂, &c., and let these displacements c b, c₁ b₁, c₂ b₂, &c., for the sake of argument, vary inversely as r, r₁, r₂, &c.

Fig. 9.

The question is to determine where the horizontal component c a of these normal motions is a maximum.

First observe that the triangle o c c is similar to a, b, c.

Also r = ^h/_{sin θ}, and therefore the normal component c₁ b₁ at c₁ is equal to c ^{sin θ}/_h.

Also c₁ a₁ = c₁ b₁, cos θ.

∴ c₁ a = c ^{sin θ cos θ}/_h = ^c/_h ∙ ^{sin 2θ}/₂,

and sin 2θ is greatest when 2θ = 90° or θ = 45°.

That is to say, the horizontal component reaches a maximum where the angle of emergence equals 45°.

This question has been discussed on the assumption that the amplitude of an earth particle varies inversely as its distance from the origin of the shock. Should we, however, assume that this amplitude varies inversely as the square of the distance from the origin, we are led to the result that the area of greatest disturbance is nearer to the point where the angle of emergence is 55° 44′ 9″. Both of these methods are referred to by Mallet, but the first is considered as probably the more correct.

Earthquake Waves.—Hitherto we have chiefly considered earthquake vibrations; now we will say a few words about earthquake waves. If we strike a long iron rod at one end, we can imagine that, as in the long spring, a pulse-like motion is transmitted. If the rod be struck quickly, the pulses will rapidly succeed each other, and if struck slowly the pulses will be at longer intervals. Each individual pulse, however, will travel along the rod at the same rate, and hence the distance between any two will remain constant; but that distance will depend on the interval between the blows producing these pulses being equal to the distance travelled by one pulse before the next blow is struck.

From this we see that an irregular disturbance will produce an irregular succession of motions; some will be like long undulations in a wide deep ocean, whilst others will be like ripples in a shallow bay. Again, consider the bar to be struck one blow only, and then left to itself. The bar will propagate a series of pulses along its length, due to the out and in vibration of its end. These will succeed each other at regular intervals, and will be mixed up with the pulses we have previously considered.

From this we see that in an earthquake, if it be produced by one blow, the motion will be isochronous in its character; but if it be due to a succession of blows at regular intervals, the motion will be the resultant of a series of isochronous motions, and will be periodical. If the impulses are irregular, you have a motion which is the resultant of a number of isochronous motions due to each impulse, but these compounded together in a different manner at each instant during the earthquake, and giving as a result a motion which is in no sense isochronous. This approaches more nearly to the actual motions we feel as earthquakes.

If we can imagine the ground shaken by an earthquake, made of a transparent material which transmitted less light when compressed, and we could look down upon a long extent of this at the time of an earthquake, we should see a series of dark bands indicating strips of country which were compressed. The distances between these bands might be irregular. Keeping our attention on one particular band, this would be seen to travel forward in a direction from the source. If we kept our eye on one particular point, it would appear to open and shut, becoming light and dark alternately.

As to the existence of these elastic waves in actual earthquakes we have no direct experimental evidence. The only kind of wave with which we are familiar is a true surface undulation, which, although having the appearance of a water-wave, may nevertheless represent a district of compression.

CHAPTER IV.
EARTHQUAKE MOTION AS DEDUCED FROM EXPERIMENT.

Experiments with falling weights—Experiments with explosives—Results obtained from experiments—Relative motion of two adjacent points—The effect of hills and excavations upon the propagation of vibrations—The intensity of artificial disturbances—Velocity with which earth vibrations are propagated—Experiments of Mallet—Experiments of Abbot—Experiments in Japan—Mallet’s results—Abbot’s results—Results obtained in Japan.

Experiments with Falling Weights.—A series of experiments, as the nature of the disturbance produced in the surface of the earth when a heavy weight is allowed to fall on it, was begun in November 1880 by Mr. T. Gray and the author. These experiments were carried out at the Akabane Engineering Works in Tokio. The weight used was a ball of iron weighing about a ton, which in the different experiments was allowed to fall from heights varying between ten and thirty-five feet. The position of the place where the ball was allowed to fall was such that in one direction the vibrations were transmitted up the side of a steep hill, in another direction across a pond with perpendicular sides, and in another direction across a level plain the material of which consisted for the most part of hardened mud extending to a very considerable depth. The vibrations produced by the fall of the ball were transmitted through this hard mud with considerable intensity to a distance of between 300 and 400 feet.

The object of the experiment was to find the nature of the vibrations produced in the crust of the earth by such a blow, the velocity of transmission through this comparatively soft material, the effect of hills and excavations in cutting off such disturbances, and the law according to which the amplitude of the vibrations diminishes with the distance from the source.

A considerable variety of apparatus was used during these experiments, but the most reliable results were obtained from the records of a rolling sphere seismograph, which wrote the vibrations on a stationary plate, and from the records of two bracket seismographs, similar to Professor Ewing’s horizontal lever seismographs, which gave a record of the vibrations as two rectangular components on a moving plate of smoked glass.

Fig. 10.

The general result as to the nature of the disturbance was that two distinct sets of vibrations were set up by the blow. In one set the direction of motion was along a line joining the point of observation with the point from which the disturbance emanated; in the other set the direction of motion was at right angles to that line. The nature of the resultant motion will be gathered from fig. 10, which is taken from the records drawn by the rolling sphere seismograph at a distance of 50 feet, 100 feet, and 200 feet respectively from the point where the ball struck the ground. The direct or normal vibrations reached the instrument first, and were followed at an interval depending on the distance of the instrument from the origin by the transverse vibrations. From the records of these two sets of vibrations as separated by the bracket seismographs, combined with the known rate of motion of the glass plate, the velocity of transmission was found to be, for normal vibrations 446–438 feet per second, and for transverse vibrations 357–353 feet per second.

The effect of the hill in cutting off the disturbance seemed to be slight, but the direction of the vibrations which ascended the side was mostly transverse. The pond, on the other hand, seemed completely to cut off the disturbance, which, however, gradually crept round the side, so that only a comparatively small triangular area was in shadow.

The amplitude of the vibrations diminished directly as the distance increased for some distance from the origin, but at greater distance the rate of diminution seemed to be slower. The transverse vibration seemed to die out less quickly than the normal vibrations.[9]

These experiments were afterwards very considerably extended by the author. In these later experiments charges of from one to two pounds of dynamite were placed in bore-holes of various depths and exploded by means of electricity. The results obtained confirmed the conclusions already arrived at from the former experiments. The experiments on velocity, however, seemed to indicate that the higher the initial impulse the greater was the velocity. The velocity of propagation of the transverse vibrations seemed to approach more and more to that of the directed vibrations as the distance from the origin of disturbance increased. Fig. 11 shows the nature of the record obtained from the explosion of two pounds of dynamite at the bottom of a bore-hole eight feet deep. These records show the interval of time which elapsed between the arrival of the normal and the transverse vibrations at points distant 100, 250, and 400 feet from the bore-hole. In the case of the 100-feet station it will be observed that the motion towards the origin is greater than that from the origin. It is also to be noticed that the period of vibration becomes greater as the distance from the origin increases.

Fig. 11.—Records obtained at three stations of the motion of the ground produced by the explosion of 2 lbs. of dynamite.

The Intensity of Artificial Disturbances.—The data which we have at our disposal for determining the intensity of an earth particle which has been caused to vibrate by the explosion of a charge of dynamite are a series of records similar to that given on [p. 60]. These disturbances are practically surface movements, and may be compared with the movements of an earthquake which spreads over an area the radius of which is great as compared with its depth.

To find the mean acceleration of an earth particle, which quantity has been taken to represent intensity, during any simple backward or forward motion of the earth, it will be first necessary to determine the amplitude of this motion and its maximum velocity, the mean acceleration being equal to ^v2/_2a.

Fig. 12.

The second and third movements in a shock invariably exhibited the greatest intensity, and to a distance of 400 feet from the origin, where about three pounds of dynamite had been exploded in a bore-hole about six feet deep, these intensities decreased directly as the distance from the origin. The less intense movements also decreased directly as the distance from the origin to a certain point, but after that they decreased more slowly. A mean result of the more prominent vibrations in four sets of experiments is shown in the curve, fig. 12, where the horizontal measurements represent distance from the origin in feet, and the vertical measurements mean acceleration in thousands of millimetres per second.

This curve approximates to an equi-angular hyperbola. The area between the curve and its asymptotes is proportional to the whole energy of the shock. The area of the diagram is proportional to the energy given up to the ground by the explosion of three pounds of dynamite. If we call the unit shock the effect produced by the explosion of one pound of dynamite, the above artificial earthquake had an intensity equal to three.

The only other investigations which have been made in this interesting branch of observational seismology are those by Mr. Robert Mallet,[10] and those by General Henry L. Abbot.[11]

Mallet’s Results.—The velocity with which earth vibrations were transmitted as deduced by Mr. Mallet were as follows:—

A striking result which was obtained by Mallet in his experiments at Holyhead was that the transit velocity increases with an increase in the intensity of the initial shock. Thus with a charge of 12,000 pounds of powder the transit rate was 1,373 feet per second, whilst with 2,100 pounds the transit rate was 1,099 feet per second. In these experiments tremors were observed as preceding and following the main shock.

Abbot’s Results.—The important results obtained by General Abbot are contained in the following table:—

A seismometer of type A means that the telescope used in observing the tremor produced on the surface of a vessel of mercury by the passage of the shock had a magnification of 6, whilst a telescope of the type B had a magnification of 12.

The mean velocity given by six observations with type A is 4,225 feet per second, while that given by the same number with type B is 7,475 feet per second.

Fig. 13.

If we assume that the first tremor observed in the mercury is to determine the true rate of transmission, General Abbot tells us that we must reject all observations made with type A, inasmuch as they do not reveal the velocity of the leading tremor. However, he also tells us that a still higher power above 12 might have detected still earlier tremors.

When gunpowder was the explosive, the observers noted that the disturbance observed in the mercury took a much longer time to reach a maximum than it did when dynamite was employed.

It was also observed that explosions fired beneath deep water gave a higher velocity than similar explosions which took place beneath shallow water. In the latter case much of the energy was probably expended in throwing a jet of water into the air.

Another point which was observed appears to have been that the rate varied with the initial shock. Thus:—

Also it is probable that the rate of a wave diminished with its advance. For,

General Abbot’s general conclusions are:—

1. A high magnifying power of telescope is essential in seismometric observations.

2. The more violent the initial shock the higher is the velocity of transmission.