Art. XXXII.—The Earthquake in the Aimuri.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 21, 1888, Unnumbered Page

Art. XXXII.—The Earthquake in the Aimuri. By Professor Hutton. [Read before the Philosophical Institute of Canterbury, 6th September, 1888.]* A rough draft of this paper was read on 6th September, 1888, but it has been added to and revised up to 20th November, 1888. Plates XV.—XVII. The earthquake that took place on Saturday, the 1st September, 1888, was felt from Invercargill in the south to New Plymouth and Masterton in the north, a distance of about six hundred miles, but was most severe in the neighbourhood of the Hanmer Plains, which are nearer to the northern limit of the disturbed area than to its southern limit by about fifty miles. The shock commenced soon after 4 a.m., with a rumbling noise and slight shakes for a second or two, followed by the main shock, lasting from forty to sixty seconds, or even more in some places. Judging from my own feelings at Christ-church, I should say that the shock was a backward-and-forward oscillation that began gradually and as gradually died away after about forty-five seconds' duration, and that it was not accompanied by any sharp jerks. It was followed within the next quarter of an hour by two much smaller shocks, while other slight ones occurred continually until 5 or 6 a.m., these slighter shocks being only felt in the Amuri, at Boatman's, Reefton, and Westport. All Saturday, Sunday, and Monday the ground at the Hanmer Plains was quivering; with smarter shocks, felt on the west coast and at Christchurch, at about 3·55 a.m. and 4·25 p.m. on Saturday, at 11.15 a.m. on Sunday, and at 8·15 a.m. on Monday. At Westport small earthquakes occurred almost every day for a fortnight; at Reefton every day for nine days; while in the Hanmer Plains they were tolerably frequent up to the end of the month, with heavier shocks, felt on the west coast and at Christchurch, on the 9th and 28th September, and again on the 12th October. A slight shock was felt in the Amuri on the 28th October, and another at about 11 a.m. on the 13th

November. It lasted for thirty or forty seconds, and was an easy, swaying kind of movement. On Thursday, the 30th August (that is, before the main earthquake), a shock occurred at a few minutes past 10 p.m. which was felt from Hokitika and Westport, on the west of the island, to Christchurch and Kaikoura, on the east; but it was not severe at the Hanmer Plains, and the time observations indicate that it originated more to the south—probably in the Upper Hurunui. Another sharp earthquake took place at about a quarter past eight on the morning of the 23rd October; but this was more severely felt at Nelson, and the time-observations show that it certainly originated north of the Hanmer Plains and probably in the neighbourhood of Mount Owen. Still another but slighter shock on the 28th October was, according to Mr. A. McKay's report, felt more severely at Kaikoura than elsewhere. At any rate, none of these shocks originated from the same place as that of the 1st September. At the time of the shock it wanted just five days to the new moon. The weather was fine and cold, a sharp frost being felt at Hanmer Plains. Over the north-west portion of the island, from Nelson to Hokitika and Bealey, the barometer was slowly rising; at Lyttelton it appears to have remained steadily at 30·45 from 9 a.m. on Friday to 9 a.m. on Saturday; while at Kaikoura it fell slightly, from 30·44 at 5 p.m. on Friday to 30·41 at 9 a.m. on Saturday. The self-registering barometer at the Agricultural College, Lincoln, was falling from 30·70 at 1.30 a.m. to 30·45 at 2 p.m. on Saturday, the height at the time of the earthquake being 30·65. The sky is reported as clear and star-lit. The air was nearly calm in the interior, and with a slight easterly breeze on both coasts, changing to north-west at Bealey. The humidity of the atmosphere at 9 a.m. on Saturday is given at 92 at Nelson, 63 at Bealey, and 41 at Tophouse in the Upper Wairau. The shock, therefore, appears to have been quite unconnected with the weather or with the position of the moon. Description of the District. The watershed of the New Zealand Alps is here formed by the Spencer Mountains, from Mount Franklin in the north to the Hurunui Saddle and Arthur's Pass in the south. It lies at a distance of about twenty-five miles north-west of the hot springs on the Hanmer Plains. Between Mount Franklin and the Hurunui Saddle the range is broken by three low passes called Cannibal Gorge, the Amuri Pass, and the Hope Saddle; and it lies almost exactly half-way between the east and west coasts of the island, which is here about 110 miles broad. Both east and west of the watershed the country is mountainous, undulating hills and plains being rare.

The main range of the Spencer Mountains is formed by contorted sandstones and slates, which to the west are followed by a narrow band of micaceous schists. Beyond these granite forms a range called the Victoria Mountains, which is parallel to the Spencer Mountains. East of the Spencer Mountains, much-jointed sandstones and mudstones form the lower ranges surrounding the Hanmer Plains, the only known eruptive rock being a small syenite boss at Hurunui Peak and the Mandamus River. These sedimentary rocks are of carboniferous and triassic age, some, perhaps, being jurassic. Bordering them on the north side of the Hurunui Plains, and stretching north-west towards Kaikoura, tertiary limestones, sandstones, and clays are found, which are of oligocene and miocene age, and among them volcanic rocks, not younger than miocene, occur in three places—(1) Where the River Pahau enters the Hurunui Plain; (2) at Lyndon, about nine or ten miles due east of Hanmer Hot Springs; and (3) up the Mason River, on the Highfield Station, about five miles east of the last. The Hanmer Hot Springs occur on a clay terrace on the north side of the plain, and are at a height of about 1,200ft. above the sea. There are ten springs, two of which are cold, the rest warm; the hottest having a temperature of about 117° F. In cutting a ditch for laying pipes to take away the overflow from the baths, a layer of black peat, some 6in. or 7in. thick, with tough clay on each side, was found. This layer would pass about 10ft. or 11ft. under the bathhouse, and probably some 15ft. below Spring No. 1. One of the springs (No. 9) often brings up small fragments of this peat; so that it probably spreads under the whole. Complete analyses have been made of the water from three of the hot springs by Professor Bickerton. They are all alkaline. The salts in the springs are chiefly sodium-chloride, but in addition there are alkaline sulphates and carbonates in about equal quantities. The ammonia and albuminoid ammonia are no doubt derived from the layer of peaty matter which the waters pass through; and, no doubt also, this organic matter reduces part of the alkaline sulphates to the condition of sulphides, which are decomposed by the action of carbonic acid derived from the peat, and changed into alkaline carbonates with the disengagement of sulphuretted hydrogen: some of the latter may, however, be derived from the albuminoid ammonia. The passage of the water through the peat-bed is too rapid to allow of the whole of the sulphates being changed into carbonates; but in all probability no carbonates and no sulphuretted hydrogen, and certainly no ammonia, exist in these waters below the peat-bed. The heat of the water makes these reactions go on energetically, but the reactions

do not themselves give rise to sufficient heat to heat the water. The water must be hot before it reaches the peat-bed.* The following are the analyses (referred to on p. 271), which have not before been published:— Grains per Gallon. Spring No. 1. Spring No. 7. Spring No. 8. Specific gravity at 60° F. 100064 100103 Sediment (silica and free sulphur) 3·15 1·4 Ammonia, free 0·156 0·193 0·112 " (albuminoid) 0·44 0·016 0·058 Potash 1·63 1·47 1·06 Soda 34·83 31·61 32·38 Lithium Trace Trace Lime 1·72 0·70 4·11 Magnesia 0·07 1·00 0·17 Iron 2·52 0·25 0·26 Alumina 0·08 0·07 0·03 Carbonic anhydride 5·39 3·69 5·23 Sulphuric " 4·69 10·42 7·00 Nitric " 0·139 0·215 Phosphoric " Trace Trace Trace Chlorine 35·78 33·60 34·29 Sulphuretted hydrogen 3·29 ? 3·43 Bromides and iodides nil nil nil Total 93·885 83·019 89·745 Deduct oxygen equivalent to chlorine 8·06 7·57 7·72 Grains per gallon 85·825 75·449 82·025 I mention these details because it has been supposed that the presence of sulphuretted hydrogen in these springs indicates the presence of volcanic energy below the Hanmer Plains, which has been supposed to be connected with the origin of the earthquake. Other hot springs occur in Cow Creek, a branch of the Edwards River;† Analysis made in the Colonial Laboratory, Wellington, of incrustations from the hot springs at Cow Creek, Edwards River, Amuri District, forwarded by Professor Hutton, November, 1888:— Nos. 1 and 2 are pure alum, which has a sweetish astringent taste, and is entirely soluble in water. No. 3 is a dark pitchy substance, having a disagreeable pungent odour. The following is the result of a partial analysis of it:— Organic matter 16·00 Sulphur (free) 16·25 Iton-oxide 13·00 Sand and clay 36·00 Water 18·75 100·00 in Cannibal Gorge; in the Upper Hope; in

the Upper Hurunui; and in the Otira Gorge; in all cases lying in the old sedimentary sandstones and slates. The waters of these have not been analysed, but Mr. F. Stephen son Smith, who surveyed the district, informs me that the Cow Creek hot springs issue from the solid rock, which is coated over with a red deposit. The temperature at the point of issuing from the rock was 135° F. The temperature of the Hurunui springs is 139° F. The Cow Creek springs are said to smell of sulphuretted hydrogen like those at Hanmer Plain, as also do those in the Otira. Evidently none of these hot springs are connected with the miocene volcanic rocks of Lyndon or the Pahau, but owe their heat, in all probability, to the crushing of rocks under the mountains. It was in this district, known as the Amuri district, that the principal force of the shock was felt. It is thinly inhabited, and is bounded on all sides but the south for a distance of from thirty to ninety miles by country which is almost or quite uninhabited; large parts being mountains, which on the western side are covered with dense forest. Under these circumstances we cannot expect to obtain a full knowledge of the nature of the earthquake or the position of the centrum; but, at the request of the Council of the Institute, I have put together all the reliable information that I have been able to obtain. On the east side of the Alps we have very good reports from newspaper correspondents and especially from Mr. A. McKay, Assistant Geologist, and as soon as I could leave Christchurch I paid a visit myself to the Hanmer Plains. From the west coast I have been most liberally supplied with copies of the newspapers published at Greymouth, Reefton, Westport, and Lyell; and from these as well as from other places I have received information from many people in reply to a circular that I sent out; and Mr. McKerrow, Surveyor-General, has supplied me with some excellent maps of the district. Damage done to Buildings. Amuri District. Glynn Wye.—The manager's house, a wooden building, was pushed several degrees out of the perpendicular to the east, and the chimneys were thrown down. The men's cottage, also wood, was shifted bodily 15in. to the north. The woolshed was also much damaged. Hopefield.—Chimneys thrown down and roof twisted out of shape. Jones, Waiau Station.—Woolshed and house badly damaged. Men's cob hut destroyed.

Woodbank.—Partly brick and partly wood. The brick portion, which was very old, had two walls thrown down. The wooden building was not much damaged, but shifted bodily 2½in. to the east. The chimney fell to the west. A concrete chimney between the two parts of the house was broken off at the roof, and the upper portion was thrown upwards and fell to the east. A cob hut has the south-west corner knocked out; the west end would have fallen, but was held up. Hot Springs.—No damage was done; but the only chimney is low, and of concrete. Jack's Pass Hotel.—No damage done, and very little breakage of glass. St. James'.—Chimney-tops fell on earthquake of the 28th September; but it is thought that they were cracked through by the shock of the 1st September, but did not then fall. Jollie's Pass Hotel.—-No damage done, and very little breakage of glass. St. Helen's.—The three chimneys were thrown down. They fell in different directions. Hams were thrown off hooks, as also was a birdcage. Ferry Hotel, Upper Waiau-ua.—This is an old cob building, but it was not much damaged. Tekoa.—A brick house. The upper portions of the walls fell, it is thought because they were not so well built as the lower portions. Balmoral.—No damage reported. Culverden.—At the station three or four chimney-pots fell. At the township no damage is reported. Montrose.—The tops of two chimneys fell. Leslie Hills.—A stone building. The walls were cracked in various directions, and the five chimneys fell, the north-west portion being the most damaged. The old building with cob walls, in very good condition, stood better than the stone. The men's hut, ten chains from the house, and built of cob, received no damage—even an old chimney, partly cob and partly brick, was not injured. Lyndon.—A chimney was thrown down. Waiau Township.—Several chimneys were thrown down, and a granite monument in the cemetery was overturned. It was a pyramid standing on a granite base. The pyramid only was overthrown. Highfield.—Three chimneys fell. An outside one was completely wrecked; the two inside ones were broken off at the roof. They fell in three different directions. Kaikoura.—One or two chimneys were thrown down, and others were cracked.

To illustrate Paper by F. W. Hutton.

Canterbury District. Waikari.—Two chimneys were overthrown. Amberley.—No damage reported. Rangiora.—A few bricks were thrown from the top of a chimney. Some crockery was thrown down and broken. Kaiapoi.—The tops of two or three very old chimneys were thrown down. The woollen-factory chimney was uninjured. Christchurch.—The most noticeable damage was to the Cathedral spire, the upper 26ft., with the cross, having been shaken down. An eye-witness says that his attention was called to the spire by the ringing of the bells (which had been set overnight for ringing); he then saw three or four stones shoot out, after which the top part of the spire swayed for a second or two; and then, after the chief violence of the shock was over, the top, with the cross, fell to the north. No other injury was done to the Cathedral, nor to the spire below the 26ft. that fell. The cross, which was of solid iron, was fastened to an iron bolt which passed through 18ft. of solid masonry, and had an iron plate a foot square at the bottom, the whole weighing not less than 60 tons. Below this the spire was hollow, and from the iron plate four iron stays 16ft. long were carried down inside the spire and secured to iron plates fastened in the masonry. The top of the cross was 210ft. above the ground. It is generally thought that the spire would not have been damaged if the top had not been solid, and if it had not been so firmly tied down. In the Normal School the top of one of the chimneys was shaken down, and four others were split. The East Christ-church school had some of the chimney-tops much shaken, and some chimney-heads fell. The Wesleyan Church was much shaken, and some of the stones moved out of their places. Three chimneys in private houses also fell. A few other buildings were slightly damaged. None of the factory-chimneys were damaged, although that of Scott and Co. is about 80ft. high, and has an iron railing weighing 2 tons on the top. Very little, if any, glass or crockery appears to have been broken. None of the specimens in the Museum were hurt. In the East Belt the main sewer was cracked; but it here passes through a quicksand, and great difficulty was experienced in making it. At Avonside Church the stone cross on the northern gable of the transept was thrown down. A few chimneys were also injured at Avonside, Heathcote, and Linwood. Lyttelton.—No damage was done. The water in the harbour was not disturbed. Ashburton.—The roof of the Borough School was split, and the plaster shaken down. No damage is reported from other parts of Canterbury.

Nelson District. Nelson.—No damage done. A few bricks are said to have been dislodged from the top of a chimney, but this has not been confirmed. Maruia Plains.—The Reefton Guardian says that it is reported that the effects of the earthquake were of a very serious nature at Walker's station, in the Maruia, but I have not been able to learn anything more about it. Lyell.—No damage was done. Westport.—No chimneys were thrown down, but two in course of erection were cracked. No damage was done to the lighthouse at Cape Foulwind. There was no sea-wave. Boatman's.—No damage done. Reefton.—No chimneys suffered, but some glass and crockery was thrown from shelves and broken. The shock was severely felt by miners in the Progress Mine: the timbers creaked, and strange noises issued from the ground. Greymouth.—Several old and badly-built chimneys fell. They seemed, from the indications on the mortar, to have been screwed off from the north-east towards the south, in some instances showing signs of quite a third of the arc between the two points mentioned. The greater portion of the bricks fell on the southern sides of the chimneys. The tall brick chimney of the engine-house of the hydraulic cranes was uninjured, as also was the Catholic church. A good deal of glass and crockery was thrown down and broken. In the teacher's house at Dunganville the school-register was thrown a considerable distance a little south of west. The shock is stated to have been more severely felt in the inland districts. There was no sea-wave. Hokitika.—Chimneys were thrown down in the Odd Fellows' Hall and Police Camp, and several others were cracked. There was no sea-wave. The conclusions that may be drawn from this statement of facts are—(1) that, except in the neighbourhood of the centre of impulse, the only damage was to buildings put up with bad mortar, or faulty in construction; (2) that cob stands the shock better than brick or stone with bad mortar; and (3) that houses on alluvial gravels, &c., suffered more, pari passu, than those on solid rock. Other points connected with this part of the subject will be considered further on. Fissures, Landslips, etc. These were formed chiefly in the valleys of the Hope and Waiau-ua Rivers. On the Hanmer River there were a few small cracks near the edge of the terraces, and a few others on an island in the bed of the Percival: all these appear to

have been filled up again before November. Small landslips occurred in the cutting leading to the bridge over the Waiauua, and two larger ones at the approach to the ferry. Up the Waiau-ua no fissures are reported until opposite the Grantham River, where there are some cracks 4in. or 6in. wide. From here, along the south side of the river, they get more and more abundant to Hopefield and Glynn Wye, but were seldom more than a foot in breadth, the larger ones being generally near the river. On the flat of Shingle Creek there were several fissures 4in. to 6in. wide. A large fissure was reported at the back of the house at Hopefield, and two circular holes about 4ft. in diameter and several feet deep are said to have been formed near Glynn Wye. Near this place fissures were very numerous in the terraces, some being more than a foot wide. Up the Hope they were still larger, some being more than 2ft. broad and several feet deep. Wire fences on the terraces were moved in places from 5ft. to 8½ft. horizontally. All these fissures were in alluvial deposits, and were more or less parallel to the valley of the Hope and Waiau-ua Rivers. Above the junction of the Boyle with the Hope the fissures get smaller and less numerous and more confined to the edge of the terraces, but there are numerous landslips on the sides of the mountains. Beyond Kiwi Creek no fissures have been noticed in the valley of the Hope, but some continue up the alluvium of Kiwi Creek. None are reported in the valley of the Boyle, and none in the Waiau-ua above Hopefield. As a glance at the map will show, all these fissures are confined to the alluvial deposits; none have been detected in solid rock. At Tekoa Station, on the Mandamus River, numerous and large blocks of rock fell from the cliffs, making a great noise. In the Bealey several landslips occurred, and in the Otira Gorge part of the road slipped down. At the accommodation-house at the entrance to the Otira Gorge the shock was felt very severely. Stones and rocks rolled down the mountainside in great numbers, striking each other and leaving long trains of fire behind them—a phenomenon which has been observed before in landslips. A large fissure was formed in Kelly's Creek, but I have not been able to obtain any particulars about it. A miner from the Totara Flat District, between Greymouth and Reefton, reports that a number of trees on both sides of his claim were thrown down; and this was probably due to slips. A shepherd who was in Jones's hut, in the Upper Hope, also reported that dead branches were shaken from the trees, and it appears that many dead trees were also broken off about 10ft. from the ground, some at least a foot in diameter. In some places near here green trees 25ft. to 30ft. in height have been torn up by the roots; and this was probably due to slips.

None of the hot springs were permanently altered, although those on Hanmer Plains were much agitated, became muddy, and emitted more gas, but with no extra flow of water. By the 5th September they were merely discoloured, and they gradually got quite clear again. Sulphuretted hydrogen escaped from the ground in many places near the hot springs; and it was reported that it escaped from other places on the plain, but there is no evidence of this. A small opening was made close to the swimming-bath, which spouted out mud and gas, with very little water, for the three days that the ground was in constant movement. All these phenomena appear to me to be secondary effects of the earthquake—that is, they were not the cause of the shock, but were produced by the reaction of the earth-wave in its propagation through the earth. Fissures which are more or less parallel to some superficial feature of the surface must almost certainly be themselves superficial; and fissures which are confined to alluvial deposits must almost certainly have originated in those deposits. They can, I think, all be explained by the principles laid down by Oldham and Mallet in their paper on the earthquake in Cachar of 1869.* “Quarterly Journal of the Geological Society of London,” vol. xxviii., p. 255. Probable Position of Epicentrum. There are three different kinds of evidence which will help us to find the probable position and shape of the epicentrum:— (1.) The intensity of the shock in different places. (2.) The direction of the shock in different places. (3.) The time the shock was felt at different places. The first kind of evidence will give us true results so far as it goes, and when an earthquake has originated in a well-populated and civilised country this method can be relied on; but when an earthquake originates under the sea or in a thinly-inhabited district it cannot lead to very accurate results. The second and third kinds of evidence are liable to many sources of error; but if all erroneous observations could be eliminated, the remainder would give a much closer approximation to the truth than can, in the cases supposed, be got from the first kind of evidence. In our case the earthquake originated in a district not only very thinly populated, but one very difficult to examine—so much so that the only accurate observations that have been made are along one line—from west to east; all the country to the north, west, and south of the place of origin being as yet unexamined. This being so, it is obvious that

observations on the intensity of the shock cannot do more than give us a rough approximation to the position of the epicentrum; nevertheless, this approximation, although rough, will be undoubtedly correct so far as it goes, and will thus enable us to discard evidently erroneous observations of the second and third class of evidence. I shall therefore begin with the evidence of the first class. Intensity of the Shock at Different Places.—The intensity of shock can be roughly estimated by the damage done to buildings, or to glass and crockery on shelves; but great anomalies occur locally (which will be considered later on), and it is only by taking a comprehensive view that we can arrive at any results. There is no difficulty in concluding that where wooden houses have been wrenched out of shape the shock has been more intense than where chimneys only have suffered. But there are great differences in chimneys—in proportions, in supports, in construction, and in materials—and we cannot make any close comparison between them. Bottles and crockery on shelves are, however, under more similar conditions, and afford a better comparison than chimneys in estimating the relative intensity of the shock. Fissures and landslips also afford good evidence when the conditions are tolerably equal. From the record of facts already given it will be seen that Glynn Wye, on the River Hope, appears to have sustained the greatest shock. It is the only place where wooden houses have been wrenched out of shape; and here the fissures and landslips are greater than elsewhere. Glass and crockery were thrown off shelves at Waikari, Rangiora, Reefton, Westport, Greymouth, Marsden, Notown, Kumara, and Hokitika, all being within a radius of seventy miles from Glynn Wye. Chimneys were thrown down or damaged at Kaikoura and Christchurch, within a radius of eighty miles of Glynn Wye; and slight damages are reported from Ashburton and Nelson, each about a hundred miles from Glynn Wye. At further distances no damage was done to buildings. The greatest damage, however, does not take place at the epicentrum, where the shock is vertical, but where the direction of the wave makes an angle of between 55° and 45° with the horizon; consequently the position of the epicentrum would probably be somewhere to the west of the meridian of Glynn Wye; and Mr. O. Thompson, the manager, says that the shock passed to the eastward, down the valley, with a hoarse crashing sound which gradually died away in the distance, while things were quiet at the place where he stood. At Reefton no chimneys were thrown, so that the shock there must have been less than at the Hanmer Plains. This may have

been due in part to the intervening ranges of mountains, but it prevents us locating the epicentrum very far west of Glynn Wye. If we assume that the meizoseimic band extended three miles on either side of Glynn Wye, and that the angle of emergence was 55° on the western edge of this belt and 45° on the eastern edge, it would indicate that the epicentrum was about seventeen miles from Glynn Wye, and about twenty miles below the surface. Mr. A. McKay, in his report, says that he is of opinion that the shock “commenced at some point to the west of Glynn Wye, perhaps further west than the junction of the Kiwi with the Hope, and that it travelled eastward with increasing force to Glynn Wye and Hopefield, beyond which places, by what appears at the surface, its destructive character began to be less.” The junction of the Hope and the Kiwi is fourteen miles west of Glynn Wye. Direction of the Shock at Different Places.—Reports under this head vary extremely, even from the same place, and in the absence of seismographs no accurate results can be expected. It is known from observation that the normal wave is followed by a transverse wave, and that afterwards the ground oscillates irregularly; so that, even if the direction be estimated right, it would be impossible to distinguish the normal from the transverse wave. Even accurate observations may often give a wrong direction. For example: The movement of cream in a pan at Rangiora gave S.W. and N.E. as the direction. At Ohoka the same kind of seismometer registered the shock as E.S.E. and W.N.W. At Ashburton a lamp was seen to swing east and west. In Christchurch water was thrown out of buckets in different directions in the same building, although in the majority of cases it was to the N.W. In fact things in general seem to have been thrown away from a wall without much reference to the shock. In the Canterbury Museum some unsupported table-legs in the Indian case fell to the east; but I found that the shelf on which they stood had a slight slope in this direction. All these and many others must be rejected as pointing far out of the direction of the normal wave; and, indeed, but little weight can be attached to this kind of evidence at all: but, as it is quite independent of all other evidence, it may be worth while to find out what results it leads to. At Wellington the seismograph is reported in the newspapers as registering the shock N.E. and S.W.; at Christchurch the cathedral-spire is octagonal, and the cross fell over to the side facing N.W. This no doubt shows roughly the true direction of the shock, but it might have come from any point between W.N.W. and N.N.W. I will take it at N.W.

The other places from which I have records pointing more or less in the true direction are—Greymouth, E. and W.; Notown, S.E. to N.W.; Westport, S.E. to N.W.; Reefton, S.E. to N.W.; Boatman's, first shock E. and W., second S.E. to N.W.; Lyell, S. to N., or S.E. to N.W.; Nelson, S.W. to N.E.; Blenheim, S.W. to N.E.; Kaikoura, N.N.W. to S.S.E.; Waikari, N. and S.; Leeston, N. and S.; and Kirwee, N. and S.; or fourteen stations in all. I have also ten other stations, in which the directions given are too wide of the mark to be of any use. They are—Rangiora, S.W. to N.E.; Ohoka, E.S.E. to W.N.W.; Ashburton, E. and W.; Lauriston, S.E. to N.W.; Glentunnel, E. and W.; Timaru, between W. and N.; Queenstown, N.W. to S.E.; Dunedin, E. to W.; Invercargill, W. to E.; Manaia, S. to N. (nearly correct). If we project the fourteen fairly accurate directions on a map, and then describe the smallest circle possible which will touch or cut all the lines, it comes out that the circle has a radius of about thirty miles, and its centre is situated at the Amuri Pass, at the head of the Doubtful and Ahaura Rivers, about seventeen and a half miles W.N.W. of Glynn Wye. This approximation is nearer the truth than could have been expected. Time of the Shock at Different Places.—Time-observations are subject to error from the clock not showing correct time, from incorrect readings, and from observations being taken at different periods of the shock. The first source of error is got over by comparing the clock with telegraph-time as soon as possible after the shock. When a clock is stopped by the earthquake the second source of error is eliminated; but the first and third remain. If, however, the time of the shock is correctly given to the nearest minute, and the stations, are sufficiently distant from each other, fairly accurate determinations may be made from them; and experience has shown that in a civilised country, with telegraphs and railways, these time-observations are of great value. The following are the times reported:— H. M. New Plymouth 4 15 Manaia 4 10 Wanganui 4 10 Feilding 4 10 Masterton 4 15 Wellington 4 15 Nelson 4 12 Havelock 4 15 Blenheim 4 15 Kaikoura 4 12 Hanmer Plains 4 12? Waikari 4 13 Rangiora 4 12 Christchurch 4 12 Lyttelton 4 13 Akaroa 4 10 Selwyn 4 15 Lauriston 4 13 Ashburton 4 13.5 Leeston 4 20 Kirwee 4 23 Bealey 4 10 Fairlie Creek 4 12 Timaru 4 11 Lyell 4 11 Westport 4 10 Boatman's 4 8 Reefton 4 10 Notown 4 11 Greymouth 4 10 Hokitika 4 12 Queenstown 4 10? Dunedin 4 15 Invercargill 4 15 A cursory inspection of this list will show that many of

the localities must be rejected as too inaccurate. After careful consideration, and from information obtained, I judge the following eight places to give the most trustworthy records for the purpose of discovering the epicentrum—viz., Christchurch, Ashburton, Lauriston, Kaikoura, Bealey, Boatman's, Westport, and Greymouth. I first tried to draw coseimal lines; but found the data far too meagre for any useful purpose. I next attempted to find the position of the epicentrum by the method of straight lines,* For this and the following methods, see Milne's “Earthquakes,” International Scientific Series, p. 200. using three pairs—viz., Westport and Greymouth, Greymouth and Bealey, Christchurch and Kaikoura. By this means, on the Government map of twenty-five miles to the inch, I found the position of the epicentrum to be in the Upper Grey, six or seven miles west of Lake Christabel, or five to six miles north-west of the Amuri Pass. I then tried the method of circles, with the following results, all being taken on the twenty-five-miles-to-the-inch map. Greymouth, Ashburton, and Kaikoura gave it in the Upper Grey, near Lake Christabel; Bealey, Boatman's, and Christchurch gave it one mile north of Lake Christabel; Bealey, Christchurch, and Kaikoura gave it between Lake Christabel and the Amuri Pass; Greymouth, Boatman's, and Kaikoura, half-way between Lake Christabel and the Amuri Pass. All these localities lie within a circle the radius of which is five miles, and the centre about four miles N.E. of the Amuri Pass. I subsequently obtained a more recent and accurate map, on a scale of eight miles to the inch, and on trying on this the stations Greymouth, Boatman's, and Kaikoura, I found that it gave the position four miles more to the S.E. than the twenty-five-miles-to-the-inch map: thus putting the position of the epicentrum between three and four miles E. of the Amuri Pass, or fourteen miles W.N.W. of Glynn Wye. Taking all these different methods into consideration, I conclude that the epicentrum was not of an elongated form, but more or less circular, with a radius of perhaps five miles, and the centre a little east of Amuri Pass and about sixteen miles W.N.W. from Glynn Wye. Depth of the Centrum. When we consider that the earthquake-wave spread for a distance of three hundred miles from the epicentrum, it becomes evident that the centrum was deeply seated, and this conviction is strengthened by the fact that the wave passed below high mountain-ranges, to Greymouth and Westport on the one hand and to Kaikoura on the other, without any apparent effect

Distribution of Earthquake Shock. September 1st 1888.

being produced by them. It was also felt strongly in the mines at Reefton. In order to obtain some idea of its depth I tried Professor Milne's method of co-ordinates, but without success. As each of the simultaneous equations is more or less inaccurate, different results are obtained by combining them in different ways, and I found, after many attempts, that these results for velocity of transit and depth of centrum were so discordant that I lost all confidence in them. The method is a very good one when the data are sufficiently accurate; but it is not adapted for obtaining the most probable result from a mass of incorrect data. We have seen that the effects of the shock were most severe in the neighbourhood of Glynn Wye, so that probably the angle of emergence was here between 45° and 55°, and this, with the epicentrum situated sixteen miles away, would give the depth of the centrum between sixteen and twenty-three miles. When we have discussed the velocity of propagation we shall be able to make another approximation to the position of the centre of the centrum. Velocity of Propagation. Westport, Boatman's, and the epicentrum are nearly in a straight line; and, if d is the distance of the centrum from Westport, and d1 its distance from Boatman's, t and t1, being the time taken by the wave to pass to each place respectively, then we have, on the assumption that the velocity is the same in both cases— d2 = v2t2; d12 = v2t12; and t − t1 = 2 minutes. ∴ d2/t2 = d12/t12. (1.) d1t − dt1 = 0. (2.) t − t1 = 2. Eliminating t1 from these two equations, we have— t = 2d/d − d1. Also v = d/t. The values of d and d1 depend upon the depth at which we place the centrum, and consequently t and v depend upon it also. The following results are obtained for different values of z, which is the depth of the centrum below the surface; the distance of Westport from the epicentrum being sixty-two miles, and Boatman's thirty-six miles:— If z = 15, then t = 5.15 and v = 12.39, or 1,090ft. per second. If z = 20, then t = 5.46 and v = 11.92, or 1,049ft. per second. If z = 25, then t = 5.81 and v = 10.74, or 945ft. per second.

The time the shock took place at the centrum will be between 4h. 4·85m. and 4h. 4·19m., or, say, 4h. 4m. 30s. But at stations like Ashburton and Christchurch, which are at a considerable distance from the epicentrum, the depth of the centrum will affect the distance very little, and therefore the velocity of propagation calculated from these places will be almost independent of z. Assuming that the time of shock at the centrum was 10h. 4·5m., and that the depth of the centrum was 20 miles, the distance of Christchurch from the centrum will be about 79 miles, and that of Ashburton about 104 miles. The time of shock at Christ-church was 4h. 12m., and at Ashburton 4h. 13·5m.; consequently the velocity of propagation to Christchurch was 10·5 miles per minute, and to Ashburton 11·5 miles per minute, the mean being 11 miles per minute, or 968ft. per second. This indicates the depth of the centrum at 24 miles, and probably about 20 miles is as near an approximation as the nature of the data at our disposal will admit of. The size of the centrum we have no means of estimating. From this it follows that the wave arrived at the epicentrum at about 4h. 6m., and that the average velocity of propagation along the surface was, from the epicentrum to Boatman's, 1,584ft. per second, and from Boatman's to Westport 1,232ft. per second. In attempting to locate the epicentrum from time-observations it is assumed that the rate of propagation was the same in different directions; and the result of that attempt agrees so closely with the result arrived at by the methods of greatest intensity and of direction of shock that we may conclude that, to the places used for this purpose, the rate of propagation was approximately the same, and that it was about 12·3 miles per minute, or 1,082ft. per second, along the surface. If, however, we take the distant stations, we find a much faster rate: to Timaru, 28·4 miles per minute; to Dunedin, 27·4; to Invercargill, 36·1; and to New Plymouth, 29·7 miles per minute. As it is impossible to suppose that the earthquake travelled faster at a distance than it did near its origin, it looks at first as if there must be errors in the time. But if we assume that it travelled at the rate of 12 miles per minute all round, it should have arrived at Timaru at 4h. 18m.; at Dunedin at 4h. 26m.; at Invercargill at 4h. 33m.; and at New Plymouth at 4h. 28m. This supposes errors in time of from seven to eighteen minutes, which could not have been the case. The only possible explanation that occurs to me is that the shock was transmitted with great velocity along the mountains in a south-west direction to Queenstown, and that Invercargill and Dunedin received the shock from there. This would agree with the direction of the shock given at Dunedin,

but it would not agree with the time given at Bealey. The postmaster at Queenstown informs me that the shock occurred there at about 4h. 10m., but he cannot guarantee the accuracy of the clock observed: the direction he gives is N.W. to S.E. The same explanation, however, will not apply to New Plymouth, although, on the other hand, we cannot believe that there is an error here of thirteen minutes in the time. I give this problem up. Judging from the slow rate of propagation, this earthquake ought to be considered as a small one, notwithstanding the great area over which it was felt; but until we have seismographs to register the amplitude of the wave it will not be possible to compare our earthquakes with those of other countries. Sounds Heard. In several places the rumbling sound which often precedes an earthquake was heard—caused, no doubt, by the fracture of rocks, and transmitted as a sound-wave through the earth, the noise which accompanies or follows an earthquake being a sound-wave through the air. In the Otira loud rumbling noises like thunder were heard before the shock; and at Jackson's accommodation-house, on the Teremakau, there was a long-continued roll, as of artillery, during the greater part of the night. This latter, however, like the noise heard at Tekoa station, was probably produced by falling rocks. In the Amuri District noises like the falling of avalanches or the firing of cannon were very frequent and loud on the Saturday and Sunday following the earthquake of the 1st September. By the end of the week they had become faint, and at distant, irregular intervals only; but they are heard occasionally up to the present time. There can, I think, be no doubt but that these sounds were heard occasionally for many months before the earthquake; but before that date they were never followed by a shock, and consequently must have been small. Mr. Stewart, who has charge of the baths at Hanmer, told me that on the 19th August he heard a number of sharp booms at regular intervals, none of which was accompanied by a shock. He heard no more until after the main shock on the 1st September. Even on that day the noises were not very loud. They were loud enough to be heard in a coach when travelling, but it is a great exaggeration to say that they were so loud that people could not hear each other speak. Earthquakes often occurred without any sounds; but immediately after the severe shock of 11.30 p.m., 28th September, the booming sounds became again very frequent, more than twenty-five loud booms being counted within an hour after the first shock, and these booms continued for two days. Again,

the shock of 2.30 a.m., 12th October, was followed by incessant booming like the fire of artillery in the distance, but some of the explosions seemed quite close at hand. A visitor thus describes them: “On Friday, the 12th October, at 2.25 a.m., I was awakened out of sleep by a most violent shock, or, rather, double shock, as there was a break of some two or three seconds in it. As soon as this had subsided these underground explosions began, and followed each other at intervals of, say, five seconds for some minutes, when they diminished in quantity but increased in strength, until every explosion made the house (a galvanized-iron one) quiver and rattle. This continued until 8 o'clock a.m., during which time we had seventeen shocks, five of which may be termed sharp.” I myself heard two booms on the 3rd November and three on the 9th November. Those of the 3rd November were at about twenty seconds' interval, and each lasted about five seconds. The sound was like that of a distant avalanche; they were not loud, and were not followed by any shock. Those on the 9th November were of quite a different character: they were short and sharp, like the explosion of a cannon at a distance. They were not followed by a shock, but I fancied that there was a slight shake simultaneous with the sound and quite as short. This, of course, would be an air-reverberation, and not an earth-wave. One of these booms was at 10.30 a.m., another at 3.14 p.m., and the third at 3.20 p.m. There was also another of a similar nature at about 2 a.m. the next morning. Mr. McKay mentions having heard noises of two different characters—one on the 7th October, which resembled the rumble of a distant avalanche, and was accompanied by a slight shock; the other, which was later on the same day, resembled a strong blasting-shot in a mine, and was not succeeded by a shock. Inquiries made on the ground lead me to think that these two kinds of sounds have been heard all through, and that each kind, when loud, was followed by a slight shock at about two seconds' interval. On the 13th November there was a loud boom at 2.10 a.m. and another at 10.5 a.m. Both these were followed, at between one and two seconds' interval, by a sharp short shake, like the blow of a hammer, quite distinct in character from the earthquakes unaccompanied by a boom, one of which took place at 11 a.m. on the same day, and has already been mentioned as a swaying movement lasting for thirty seconds. These sounds have been heard in the valleys of the Hope and the Edwards, and doubtfully at Cannibal Gorge, as well as on Hanmer Plains and the hills immediately surrounding them; but they were not heard at Culverden or Waikari on the south, nor at Tarndale on the north, nor at Reefton or Boatman's or Lyell on the west. At the Hanmer Plains it is

generally agreed that they came from various directions between west and north; and they appear to me to proceed from an elongated area some thirty miles in length, between the hot springs in the Hope and the hot springs in Cow Creek, or perhaps from the neighbourhood of these two localities only. As a rule the hot springs at Hanmer showed no sympathetic action with the noises, the only exception being a sharp boom, like a cannon-shot, at about 11.15 a.m. on the 14th September, accompanied by a shock which appeared to be nearly vertical. On this occasion a small quantity of mud and water was thrown from one of the smaller springs only. It is difficult to offer a satisfactory explanation of these noises: they have been heard with other earthquakes, but never explained. In our case it is evident that the main earthquake, and all those of the same character that followed it, were quite independent of the cause of the booms. This is shown by the fact that many shocks were not accompanied by any noises, although they were heavier than those following the booms, and also by the heavier shocks of the 1st September, 28th September, and 12th October being followed, not preceded, by noises for several days. On the other hand, as the booms were heard before the main earthquake, their origin must be independent of it; but, as they were far more frequent and much louder after the shocks, it is evident that to a large extent they were secondary effects of the earthquake. Mr. Mallet suggests that the noises heard after the Cachar earthquakes of 1869 were due to grinding or crushing of rocks; but this explanation will hardly do for our case, because many of the earthquakes were not accompanied by noises, and the booms do not come from the direction of the epicentrum of the earthquakes. The sounds appear to me to be much more like explosions of steam than crushing of rock; and this seems to be the only other explanation. There is no direct evidence to show that they are connected in any way with the hot springs, but their geographical distribution strongly suggests it. Hot underground water undoubtedly exists in the district in which the sounds have been heard, and at a comparatively small depth this water may be above the boiling-point, but kept fluid by pressure. If this pressure were removed, part or the whole might flash into steam and produce an explosion which would cause a boom. An earthquake might first compress this water, and then, on the backward swing of the wave, the pressure would be relieved and explosions take place; or part of the heated water might be expelled by the shock, which would reduce the pressure on the rest. It seems useless to offer such speculations as these, and I should not have done so if it had not been suggested that these explosions

were caused by “the gradual upheaval of a molten dyke through the upper strata of rocks”—an hypothesis which rests on no evidence whatever. It appears, therefore, that there have been in the Hanmer Plains two kinds of earthquakes, due to different causes and originating in different places. Luminous Appearances in the Sky. At Reefton, in the early morning and in the evening of the 1st September, a “luminous appearance” is reported to have been seen in the eastern sky in the direction of Christchurch, and it was again highly visible on the evening of the 8th September. In Dunedin, on the evening of the 1st September, an extraordinary glow was observed in the western sky, noticeable until after midnight, and it travelled southwards. I mention these things, but I do not think that they were in any way connected with the earthquake. Unequal Effects of the Earthquake. It is well known that the effects produced by an earthquake are often apparently capricious. Sir C. Lyell says that in the Calabrian earthquake of 1783 “in some streets of Monteleone every house was thrown down but one, in others all but two; and the buildings which were spared were often scarcely in the least injured.” And many other examples could be given. Much of this may be due to the different materials of which houses are built, to their different plans of construction, or to their different foundations; still, when due allowance has been made for all these things, a balance often remains over which can only be explained on the supposition that the shock was actually more severe in some places than in others, irrespective of their distance from the place of origin. More than forty years ago Mr. Robert Mallet proposed a theory to account for these apparent eccentricities. He said, “Where a wave of elastic compression, such as our earth-wave, passes through a body varying in specific elasticity in several parts of its course, or passes from one body to another of different elasticity, at each such change of medium the wave changes its velocity and in part changes its course, a portion being reflected and a portion refracted, analogous to a wave of light in passing through media of variable density or of different refractive indices.”* “Dynamics of Earthquakes,” Pro. Royal Irish Academy, 1846, p. 26. This explanation has been universally received as correct; but it can only be applied to particular cases when the local details of geological structure are well known; and before attempting

to apply it to our earthquake it will be useful to explain the theory rather more fully. Rock, of all kinds, is a more highly elastic material than alluvial gravel or sand, and when an earth-wave passes from rock into alluvium it will, unless it be perpendicular to the plane of junction of the two formations, be partly reflected downward and partly refracted towards the perpendicular to the plane of junction (Pl. XVII., fig. 1). If, however, the direction of the wave was very oblique to the plane of junction, the whole wave might be reflected down into the earth, and no shock would be felt on the alluvium (Pl. XVII., fig. 2). On the other hand, when the wave passes from alluvium into rock the refracted portion will be bent away from the perpendicular to the plane of junction, and the reflected portion will have its angle of emergence increased (Pl. XVII., fig. 3); but if the angle is small between the direction of the wave and the plane of junction, then total reflection of the wave in an upward direction will take place (Pl. XVII., fig. 4). This upward reflection might be in the same azimuth as the direction of the earth-wave, but more commonly the wave will be diverted to the right or left according to the inclination of the plane of junction. It is only the cases of total reflection that need be considered here. The slopes of old valleys covered up with alluvium vary very much; but, as the earth-wave is always more or less emergent, the angle formed by the wave with the plane of junction on entering alluvium will generally be greater than the same angle when the wave is leaving alluvium: consequently, total reflection will be rare where the wave enters an alluvial plain, but will be common where the wave leaves it. A glance at Plate XVII. will explain this. It follows, therefore, that along those margins of alluvial plains where the rocky slopes face the origin of the earthquake the shock may be doubled or trebled in force; while along those margins where the rocky slopes are turned away from the origin the shock will either be normal or will be diminished in intensity. This does not apply to a narrow valley, for in that case the whole contents of the valley would be forced to vibrate as one system with its rocky walls, and there would be neither refraction nor reflection. I will now try to apply these principles. Jack's Pass Hotel (31 miles from the epicentrum), Jollie's Pass Hotel (34 miles), and Culverden Station (34 miles) are in narrow valleys, and would receive the normal shock only. Balmoral (30 miles) and Montrose (32 miles) are on alluvial plains near the margin where the wave entered alluvium from rock, and consequently the shock in these places was probably normal also.

Ferry Hotel, Waiau-ua (29 miles distant from epicentrum), is built on rock which is cut off from the earthquake-origin by an alluvial valley; it would therefore, in all probability, receive less than the normal shock due to its distance, because some of the waves may have been totally reflected upwards before reaching it. This hotel is an old cob structure, and manifestly it has not undergone such a severe shaking as Woodbank or St. Helen's were subjected to. Leslie Hills (33½ miles distant from epicentrum) undoubtedly received a more severe shock than did Montrose or Balmoral, and it is situated on the margin of an alluvial plain, where the wave passed onwards into rock, and consequently in a position where we might expect an increase in the violence of the shock from total reflection upwards. The same explanation applies to Highfield (46 miles distant), where several chimneys were thrown down; for it stands on an alluvial terrace, with hills behind which face westerly. St. Helen's certainly received a more severe shock than its distance from the epicentrum (32 miles) would warrant, although it stands nearly in the middle of the eastern half of the Hanmer Plains. But the evidence shows that the wave emerged here at a high angle. Hams and bacon were thrown off hooks; a birdcage was also thrown off a hook, and ice was thrown up out of a pool. Evidently the angle of emergence was greater than usual, and I should account for this, as well as for the increased intensity of the shock, by the supposition that the spur between the Hanmer and the Percival Rivers runs down under the alluvial plain below St. Helen's and acted as an earthquake-reflector upwards. It has been supposed that the ground on which. St. Helen's is built is swampy, and that that would account for the damage done to the house; but it would not account for the increase in the angle of emergence, and, after seeing the locality, I feel inclined to reject the swamp theory altogether. At Woodbank (28 miles distant) the shock was undoubtedly more severe than its mere distance from the epicentrum would explain. I do not take into consideration the brick portion of the building, which was old and put up with bad mortar, but the wooden part of the house, which was shifted bodily 2½in. Here, also, cob huts, not worse built than the Ferry Hotel, were rendered quite uninhabitable, while the Ferry Hotel, only one mile further from the epicentrum than Woodbank, was scarcely injured. At Woodbank, also, a cement chimney-top was thrown up and then fell over on to the roof of the wooden part of the house, which indicates not only a very strong shock, but also a high angle of emergence. This is confirmed by Mr. Atkinson, who says that when standing outside his house immediately after the first shock

Reflection & Refraction of Earth-waves.

he felt a series of strong upward shakes in the ground. I think therefore that at Woodbank the shock must have been locally increased by reflection from below; but it is not easy to say why this should have taken place. The house stands at the northern margin of the alluvial plain, but a spur of the hills comes down between the house and the origin of the earthquake so far that the end of the spur bears south-west from the house. This spur ought to have prevented total reflection upwards from taking place. The only suppositions that occur to me are that either the shocks here came from more to the south—that is, from the valley of the Waiau-ua—or else that an underground spur exists from the hills east of Woodbank, which would act as a reflector of the earth-wave. Neither of these suppositions appears to me to be probable. At the Hanmer Hot Springs the intensity of the shock was probably that due to the distance (31½ miles) from the epicentrum, neither augmented nor diminished. Cause of the Earthquake. All ordinary earthquakes are due to one of two causes—they are the result either of subterranean explosions of steam or of the crushing or fracturing of rocks. There may occasionally be an earthquake due to some other and exceptional cause, but these must be rare. The first cause—explosion of steam—is due to water coming into contact with heated rocks and, as the water must originally have been surface-water—i.e., rain-water—the heated rocks must be near the surface of the earth, and the centrum of the earthquake must also be more or less superficial. These earthquakes are usually found associated with volcanoes or in hot-spring districts, and always occur previous to and during the progress of an eruption; although they also often take place without an eruption. The earthquakes felt in the hot-spring district from Lake Taupo to Rotorua are no doubt of this origin. The second cause—fracture of rocks—is due to the gradual increase of strain on the rocks until at last they give way with a more or less sudden snap and jar. These strains may be due to lateral pressure in the earth's crust, which crumples it up and forms mountain-chains; or they may be due to the gradual relief from pressure owing to the denudation of the surface, which is more rapid in mountain-ranges than elsewhere; or they may be due to the gradual increase of pressure brought about by the deposition of mud or sand on the sea-bottom, a cause which acts most rapidly near the mouths of great rivers. Earthquakes due to this cause may occur in almost any part of the world, but they are most numerous in mountain-ranges and near the mouths of large rivers. Many

of them are small and local, but others are far more violent than earthquakes due to explosions of steam, and, as the centrum is often deeply seated, they are often felt over a very wide area. The earthquake of Wellington in 1855 was one of this kind, as also are, no doubt, most or all of those in the South Island. The Wellington earthquake, however, belongs to a very rare class, in which the centrum extends to the surface, and surface-rocks are moved. In a large majority of cases no movement takes place in the surface-rocks* By term “surface-rocks” I do not include alluvium, &c. except that due to the earth-wave generated below by the fractures. Small earthquakes may not be accompanied by actual fracture of rocks; and when there is no fracture no noise will be heard, although the shock may be felt for a considerable distance: for the waves of sound in the earth are produced by the fractures. In the earthquake we are now considering the shake was of an unusual character, inasmuch as it was long and even, without any violent jerks; but, as it was accompanied by a sound-wave, fractures of some kind must have taken place. These fractures could not have been due to an explosion or to a very sudden break or split: they appear to me to have been due to a slow splitting or crushing of rocks. At first sight the evidence seems to favour the idea of a slow splitting having taken place along an east-and-west line in the valleys of the Waiau-ua and Hanmer, for it was in this direction that most of the damage was done. But this idea is much weakened when we remember that this is the only line which is even fairly well inhabited, and is the only line along which an alluvial valley approaches the epicentrum; and when we examine the evidence attentively I think we must give up the idea altogether. It is certainly in favour of it that a better explanation could then be given of the destruction caused at Woodbank; but this is the only favourable fact that I can find, for no fault or fissure has been proved from other evidence to exist in the valleys of the Waiau-ua and Hanmer, and no fracture or movement of solid rock has been found anywhere in the neighbourhood. On the other hand this line has a distinct meizoseimal band, and if it were a line of fissure reaching to the surface the shock would have commenced at this band and gone both ways, which is distinctly contradicted by both an eye-witness and by time-observations. Again, the Ferry Hotel stands on the very edge of this supposed fissure, and ought to have suffered more than Leslie Hills or Highfield. And, again, no explanation is in this way given of the strength of the shock at Tekoa and in the Otira,

which is from forty to fifty miles south-west of the position of the epicentrum. Certainly, Professor De Rossi has stated that in the neighbourhood of Rome the rock-fissures form axes of propagation of the earth-wave, the movement being at first parallel with and then at right angles to the axis of the fissure; but the even outward spread of our earthquake shows no such connection, and we do not yet know the geology of the district sufficiently well to say where the fissures are. In our case time-observations point to the conclusion that the locus of the centrum was small and without any well-marked seismic radiant like those so often found in other earthquakes, and I should therefore conclude that our earthquake was not due to splitting, or movement along a fissure, but to the crushing of a compact mass of rock.