In several places—not less than twelve—the water of the lake penetrates through this barrier and issues in streams of larger or smaller size from its outer slope. The various streams soon unite, and combined with the water from the surface outlet, constitute the Waikaretaheke River, which rushes down the steep slope of the barrier in a series of fierce cascades. During some months of the year the leakages through the barrier more than balance the inflow and no water then flows through the narrow channel of the outlet.

The escape of water through the leakages is of course nearly constant, and the narrow gorge of the outlet does not allow of a great increase of overflow as the level of the lake rises. Consequently in periods of heavy rainfall the inflow is much greater than the outflow, and for this reason the level of the lake is liable to greater variation than that of the majority of lakes.

The indented and irregular outline is also rather peculiar, and the ratio between the area of catchment and lake surface is rather small for a lake that occupies part of a valley-system formed by stream erosion. That this is the nature of the basin of Lake Waikaremoana is at once indicated by its outline, and by its relation to the physiography of the surrounding country. It is clear at a glance that the arms of the lake are merely drowned portions of existing stream valleys.

The rocks that occur in this locality are:—

(1) A sandstone of moderate hardness with its grains partly cemented together by carbonate of lime.

(2) A fine grained bluish-grey claystone or so-called papa. The rocks are arranged in well-defined strata which have a considerable uniformity over a wide area.

These strata strike 40° (magnetic) and have a dip of 18° to the south-east. All the prominent hills on the south side of the lake,—Puketapu 3905 ft., Pukanui, Panikiri and Ngamoko 3644 ft.,—are escarpments in a series of strong sandstone strata which are perhaps 900 feet thick.

The north-west slope of this escarpment in Panikiri has an angle of 78°, though the papa rock below the sandstone escarpment on the same slope has an angle of only 20° on the continuation of the same hillside. The south-east slope of these hills follows the surface of the harder sandstone with great regularity along the dip.

The outlet of the lake is situated at its south-east corner, and the direction of the stream that issues from it closely follows that of the dip of the rock.

The soundings made by Keith Lucas show that the slopes below the water-level of the lake are in general similar to those above it. This suggests that the slopes above and below the present water-level have been formed by the same agency—in other words the contours of the surface-relief acquired their present form before the waters of the lake occupied this area. The floor of the lake has a central relatively flat area about one square mile in extent. There are deep channels extending up each arm of the lake, and the most noticeable of these is in the direction of the outlet. Raekahu, the hill on the

south side of the outlet, is 2421 feet high, or 406 feet above the level of the lake surface. It is formed entirely of broken material which extends outwards on its south-east slope to 800 feet above sea-level or 1600 feet below the summit of the hill. The whole of this slope as well as that from the outlet to the valley of the Waikaretaheke is irregular and hummocky with large angular rocks showing frequently. There are many depressions in this irregular surface which are filled by the waters of small lakes.

The whole physiography shows that the structure of this barrier has been formed by agencies in which the ordinary action of running water took no part. Two agencies at once suggest themselves:—(1) Glacial action and (2) Earth-slides.

(1) Terminal moraines formed by glacial action have in many instances in New Zealand and elsewhere formed barriers across mountain valleys, and behind these morainic structures lakes have been formed when amelioration of the climate has caused the ice-masses to melt. The moraines piled across such valleys, however, constitute firm and impervious dams through which no percolation of water takes place. They have also a gently-sloping outer face, mainly formed of gravels carried by the streams that issued from the terminal face of the glacier. The crest of the barrier of Lake Waikaremoana is far narrower than that of a moraine. Again, a close examination of the rocks of which the Waikare barrier is formed shows that they are all of local origin,—there are no fragments of the rocks of which the north western margin of the catchment area is formed. Glacial action would certainly have carried a large quantity of such material to a terminal moraine. Finally there seems to be a total absence of flat-sided and striated boulders which are characteristic of material that has been deposited by glaciers. It is therefore apparent that this great barrier is not of a morainic nature and opinions in regard to its permanence that might be based upon this idea of its origin must be abandoned.

(2) The geological structure of the district is precisely that which most effectively promotes earth-slides of large dimensions. This structure is best exemplified in the hills Puketapu, Panikiri-Ngamoko. In each of these there is a hard cap of sandstone at least 700 feet thick resting on impervious relatively-soft papa which readily develops a greasy or unctuous surface. The dip of the whole series is 18° to the south-east, and is markedly uniform.

The prominent and dominating escarpment from Puketapu to Ngamoko is broken by the depression of the outlet at Onepoto. The strata of sandstone were of course at one time continuous between the outcrops on these hills, and probably the escarpment then extended with little interruption across this distance, though it would have been breached by a gorge through which the drainage of the Waikare basin passed. At the point of the breach through the line of the present escarpment the gorge of the river was certainly 1000 feet below the present lake-level. It has to be assumed that the “pre-lake” Waikare River in one of its meanders, or that one of its tributaries, had cut deeply into the sandstone at the low level of that valley some two miles below the present outlet. This erosion must

have been deep enough even to expose the papa below the sandstone (Fig. 1).

The heavy masses of thick sandstone would then have no support at the lower end of the dip-slope, and under a special association of extreme conditions would slip down the slope. If these conditions were so extreme as to enable the heavy rock-mass to acquire some velocity, its movements would be so great that it would smash and pile itself up even in such a mass as that of Raekahu—which though a prominent point is only 406 feet above the level of the lake and 1484 feet below the summit of Puketapu—the level from which it is conceivable that the material that travelled furthest was derived. This slipping rock would consist mainly of sandstone with sufficient papa to fill up the crevices between the rocks and make the whole mass watertight (Fig. 2).

It is probable that this slide filled the gorge of the pre-lake Waikare river to a partial extent only. Its effect would be the formation of a small lake, and it would also force the current of the outflowing river further east, where it would undermine the escarpment then lying north-east of the present outlet. Deposition in the basin of the small lake thus formed might be the origin of the flat central floor of the lake.

The erosive effect forced thus to the north-east by the slipped matter in time rendered the escarpment in that direction insecure, and a second slide took place, for the kind of rock and geological structure are the same here as on the south-west side of the outlet (Fig. 3).

This second earth-slide from the eastern side was, however, of less amount, and the movement was not so great. The rock-masses on this side have been far less smashed and piled up. It still retains something of its original form and structure. This second slide-movement consisted almost solely of the sandstone strata, and started from that portion of the escarpment north-east of Rosie Bay, which actually owes its formation to this movement. The measure of the horizontal component of this movement is to be found in the distance between the summits of the two ridges on either side of Rosie Bay—perhaps rather more that 400 yards. The hill to the south-west of this bay has an escarpment of rock-masses fractured and tumbled in a most remarkable manner, but still all the masses clearly form parts of a former continuous escarpment. The absence of slipped papa, at any rate in the upper part of this slide, renders it less obstructive to the passage of water, which therefore penetrates to a large extent between the fragments of the rock-slide, and issues as leakages at various points over the outer slope of the barrier to a level of 300 feet below the lake surface.

If this explanation is correct it follows that the subterranean channels through which the water passes are of a highly indefinite and irregular nature, and that it is impossible from any examination of the surface to form any reliable idea of their intricacies or the details of their directions. The nature of the ground also shows that the width of these crevices depends on the relative position of various angular rocks which may be more or less unstable, though if it be

correct that in this rock-slide the movement was relatively small and the rock travelled “en masse” and without much crushing, it is probable that the blocks are well keyed-in and fairly stable. That this rock has travelled from a higher level is shown by its nature. Near the outlet the broken rock contains seams which have fossil polyzoa which appear to me to be restricted to the higher strata of the sandstone.

The actual outlet passes through the second slip very close to its south-west margin where it abuts against the material of the first slip.

It is probable that if the excess of rainfall over the evaporation amounts to 60 inches annually over the catchment area, that about ten years would elapse before the basin would be filled. Within that time some consolidation of the slipped matter might have taken place, and some stability of the barrier would have been attained. It is noticeable that the overflowing water at the outlet escapes by one of the fissures that has been formed in the slipped rock of the second period; for in my opinion the overflowing water has not had any appreciable erosive effect on the rock which forms the lip of the lake at the outlet.

The question at once arises of the stability and security of the upper portion of this naturally-formed dam. It must be recognised that in a geological sense this barrier is only a temporary feature of the physiography. In time the outflow will corrode its channel in the loose debris, and ultimately will completely drain the lake.

In forming an opinion as to the stability of the barrier the following matters must be considered.

1. The structure of the dam. 2. The age of the dam. 3. Its stability in the past. 4. The leakages. 5. Erosion by the cascades. 6. Heavy rainfall. 7. Chances of further landslides.

1. The structure of the dam.—As already stated that portion of the dam which was formed by the first landslide (i.e., on the south side) appears to be almost impervious, and quite secure. On the other hand the portion that lies to the north-east of the outlet, which was due to the second landslide, is formed of far larger and more solid blocks of rocks, but between these there are numerous crevices. Along these water flows and issues as leakages on the outer side of the barrier in at least twelve places. This implies a loose setting of the rocks and a condition of relative instability, the absolute amount of which cannot be stated with any certainty.

2. The age of the dam.—Information in regard to the age may be obtained by observations made on the following:—

(a) The distribution of the pumice covering of the ground; (b) The size and extent of the wave platforms on the lake shores; (c) The nature of the lake beaches; (d) The amount of erosion by the stream which runs over the barrier.

(a) It is noticeable that the distribution of pumice is general, and as uniform as could be expected over the whole of the slipped material. This fact indicates clearly enough that the earth-slides occurred before the great fall of pumice in this district. This cer-

tainly took place before the occupation of the land by the Polynesian people. It is probable that the fall took place several thousand years ago. This, from the human standpoint, gives a great antiquity to the barrier.

(b) The wave-platforms on the lake-shores are of small dimensions. They occur only in those localities where soft papa rock constitutes the lake-shore. These ridges of soft rock are worn back into truncated spurs. In such places a wave-platform not more than 20 yards wide in any observed cases has been formed. Where sandstone forms the lake margin no platform whatever has been formed, nor does the sandstone appear to be waterworn. These observations indicate that from the geological standpoint the water of the lake has been at the present level for an extremely brief period.

(c) It is said that on the beach-shores sand is to be found at One-poto only. The sand that lies on the beach at this place consists of angular quartz grains and crystals of hypersthene; substances that are clearly derived from pumice. The presence of this sand does not, therefore, suggest that wave-action has taken place for any but a very brief period. The absence of sand elsewhere goes far in support of this conclusion. Generally speaking there are no beaches on the sandstone frontages of the lake, and it is also noticeable that the small rock-fragments have not been appreciably worn by wave-action. In the lee side of the head-lands formed of papa small beaches have been built up. There are also flats formed of detrital matter in those places where small streams enter the lake. All of these that were seen are however quite small when the softness and rapid yield of the papa to erosion are considered.

(d) Erosion of the outlet: It is a well-known fact that lake-water erodes a channel through rock very slowly. There is an absence of suspended matter which is the most effective tool in the ordinary course of erosions by running water. As previously stated the actual outlet of the lake passes through a rift which has not been formed by erosion.

3. The stability of the dam.—It is of course recognised that such a structure as this dam, narrow and composed of loose incoherent material, cannot in the nature of things have a long life from the geological standpoint. The present question however is whether reliance can be placed on the dam to contain the water of the lake permanently from the human standpoint.

It must first be considered whether the dam has suffered any partial collapse since its formation.

Evidence has to be sought on the lake shores. The existence of terraces round the lake above the present water-level would at once indicate that the dam had partly failed at some time in the past. Search, however, failed to trace any terraces. At the boat-landing at the Accommodation House there is an alluvial flat at a level that is seldom reached by the lake-water. It is, however, quite possible that this could be formed at the present time, for as has been previously stated the gorge-like form of the narrow outlet causes the level of the water to vary more than is usually the case in lakes.