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