boulders such as are not found in the ordinary gravels. But the shapes of the hills are not those of ordinary moraines left by glaciers: they are more of the nature of eskers—that is, low ridges composed of water-worn materials roughly stratified in places.

The Moeraki Downs show no trace of morainic origin. They are ordinary sand and gravel beds, standing out above the plains to a height of about 100 ft. How could they have been deposited? Certainly not under the present conditions. The rivers, when the gravels of the Moeraki Downs were deposited, must have run at a much higher level than they do now.

On the flanks of the Malvern Hills there are also patches of gravels resembling those of the Canterbury Plains. I have examined them in the neighbourhood of Sheffield, and estimated that they go to a height of 50 ft. to 70 ft. above the plains.^* Again, at Fighting Hill, on the south-western side of the Malvern Hills, there is a bed of gravel, 1,700 ft. above the sea, which is not morainic in origin. It resembles a bar across the entrance to a sound or river, and it is impossible to imagine that it could have been built up by the Rakaia.

I have not examined the hills near the gorge of the Ashley on the northern boundary of the plains, and so I cannot say if any high-level patches of gravel are found there; but in the railway-cuttings north of the Waipara gravel-beds are found at a considerable height above the level of the plains.^†

On Banks Peninsula there are no gravel-patches similar to those of the Canterbury Plains—nothing but fragments of local volcanic rocks covered by the silt-deposit which I have already mentioned.

Before trying to explain the origin of these high-level gravels it is necessary to consider two remarkable topographical features in the plains—I mean the lower gorges of the Waimakariri and Waipara Rivers. The first of these I brought to the notice of this Institute in a paper read on the 15th November, 1883; but it is so long ago that perhaps I may be allowed to draw your attention to it once more.

The Waimakariri, after debouching on the plains, is joined by the Kowai, coming from the south-west, and the two united rivers make straight for Gorge Hill and cut it in two, thus forming the lower gorge of the Waimakariri, which is crossed by the railway between Sheffield and Oxford. The hill is composed of sandstones, and rises about 270 ft. above the plains.^‡ Above the

gorge the river-bed is very wide and beautifully terraced, and the same occurs below the gorge; while in the gorge itself the river is confined within a narrow channel.

Now, how can we explain this? Evidently the river must at one time have run at a higher level than the top of the gorge. At that time nearly the whole of Gorge Hill would have been buried by sand and gravel beds, through which the river cut its way down. We can hardly doubt but that the high-level gravel-patches near Springfield are also part of these same gravel-beds, for they go up to about the same height. But how have the intermediate portions been removed? They could not have been removed by the river itself, because to do so the river must have left the gorge and cut a new channel at a lower level round the hill. If it had done this it could not possibly have regained its old level in the gorge, for that would be on a higher level. But the gravel-beds have evidently been removed by water—they could not have been blown away by the wind—and if they were not removed by the river the sea is the only agent left.

Now, take the lower gorge of the Waipara close to where it runs into the sea. This gorge has been cut through a spur from the neighbouring hills, which the river pierces instead of going round, as it certainly would have done if the river-bed had been marked out under the existing conditions. Here also, as in the case of the Waimakariri, we must assume that the river at some former time ran at a higher level than the spur through which it cuts, and that the greater part of the gravel-beds over which it then ran have been subsequently removed by the sea.

If we go outside the Canterbury Plains to look for collateral evidence we find it at Amuri Bluff, where a raised beach has been described by Mr. A. McKay, at a height of 500 ft. above the sea, containing marine shells of still living species. Again, at the mouth of the River Conway gravel terraces go to 300 ft. above the sea, but no shells have been found in them. And lastly, at Motunau marine shells of recent species are found in abundance at 150 ft. above the sea.

I cannot, therefore, agree with Sir Julius von Haast, in his report “On the Formation of the Canterbury Plains” (Christ-church, 1864), that these plains are entirely due to river-action. But I must believe that, although the materials of which they are formed were brought down from the mountains by the rivers and accumulated under river conditions, during a long period of subsidence the contour of the surface has been much modified by the subsequent action of the sea, which swept away a considerable portion of the upper beds when the land was probably 1,000 ft. lower than at present.^*

Two principal objections may be made to this hypothesis: the first is the absence of marine shells on the plains; the second is the absence of high-level gravels on the slopes of Banks-Peninsula. Both, you will notice, are founded on negative evidence, which has less weight in geology than in any other department of science, on account of the imperfection of the record. The first objection I will postpone until later. The absence of marine beaches, or gravel-patches, on Banks Peninsula may be due either to the effects of subsequent atmospheric denudation, which has washed them away; or, which is more likely, the gravel-beds never reached high levels on Banks Peninsula, on account of its distance from the mountains.

A third objection is the shape of the stones which form the plains, which are not flattened as they should be if they had been exposed to long-continued wave-action on a sea-beach. But this objection would not apply to all the stones of the gravel-beds, only to an upper layer; and it is quite possible that the submergence did not last sufficiently long for the stones to assume the shape characteristic of a sea-beach.

We now come to the silt or loess which lies above the gravels. It is found in places all over the plains, and inland to Gorge Hill or beyond it. It is also found all round Banks Peninsula, where it is quite as well developed on the seaward as on the landward side. Here it occurs chiefly on the tops of the ridges, having been largely removed from the valleys by denudation. It is well seen in Lyttelton Harbour and along the Governor's Bay Road, as well as all along the foot of the hills from the Dyer's Pass Road to Sumner. On the coast between Lyttelton and Akaroa it also forms thick deposits, a section of some of which I gave in my paper in the “Transactions of the New Zealand Institute,” vol. xv., p. 413. Sir Julius von Haast, who examined Banks Peninsula with considerable care, states that it goes up to a height of about 800 ft. above the sea; and, judging from what can be seen when going along the roads to Akaroa, I think that he is right. Near Amberley it is sometimes covered by fine river-gravels brought down by the River Ashley.

At Timaru the same deposit is largely developed, covering the volcanic rocks and extending northwards over the shinglebeds of the plains. It can be traced all the way down to Gamaru, where it is as well developed as at Timaru. But after crossing the Kakanui River we lose it; or, rather, it changes into an ordinary clay, with more iron in it than in the typical deposit. These brown clays extend through Dunedin and South Otago nearly to the mouth of the Mataura, where we again come across a silt exactly like that of Lyttelton, which continues over the Southland plains to Invercargill, and perhaps further—for I am

compelled to limit myself to my personal observations, not being able to find anything definite about this silt - deposit in the reports of the officers of the Geological Survey. I have mentioned the geographical distribution of the silt because it must have had the same mode of origin at Oamaru and Southland as on the Canterbury Plains, so that any theory respecting it must be suitable to all these places.

Its chemical composition is not yet accurately known, as no analysis has been made; but it consists of a mixture of clay and fine quartz sand, with very little iron or lime in it. The sand-grains are not much rounded—not so much as we should expect to find in a wind-blown deposit.

The most remarkable thing about this silt is its capillary structure; the mass of the rock being penetrated in all directions by minute tubules, very irregular in diameter, branching in all directions, and anastomosing with each other. Occasionally, but rarely, these tubules are filled with calcium-carbonate, and in one place, on Banks Peninsula, I found them containing limonite. On the slopes of Banks Peninsula the lower portions of the silt often contain small fragments of volcanic rocks evidently derived from the hills above. These are not found in the upper parts, where the deposit is thick, nor on the plains.

Fossils are very rare. On Banks Peninsula and on the hills of the Oamaru Peninsula moa - bones occur; and inland of Oamaru the skull of an elephant-seal was found many years ago: it is now in the Dunedin Museum. At Raupo Bay, near Little Akaloa, on the property of Mr. J. W. McHale, there are a number of concretions round bones in the silt, about 100 ft. above sea-level. They are numerous, and represent the remains of a very large animal. Most of them cannot now be recognised, but Mr. McHale showed me one which was evidently the first phalanx of the third or fourth finger of a large baleen whale. It was 6 in. long by 2 ½ in. in breadth at its contracted middle. Marine shells occur in abundance at the base of the deposit round the Oamaru Peninsula, and a few are said by Mr. McKay to have been found at Timaru. Also, at the Port Hills, near the Convalescent Home, Miss M. Bridges has found two specimens of a Euthria and one of a Crepidula.

At Lyttelton the silt is distinctly stratified near its base, and occasionally the same is seen near Timaru, but elsewhere its composition is too uniform to show any division into layers.

Now, as to its origin: It is, I think, evident that the material of the silt-deposit was brought down by the great rivers draining old glacier districts. It resembles closely the glacial mud which fills up some of the older lakes, as in the Rakaia Valley. Commencing a little north of the Waimakariri and ending a little

south of the Waitaki, and commencing again near the mouth of the Mataura and continuing, probably, to the mouth of the Waiau, it appears to be connected with the rivers which come down from the glaciers. We may perhaps ask why it is not found also at the mouth of the Clutha. In answer I would say that it has never been looked for in that district. And if it is really absent we should probably find an explanation in the alteration of the courses of some of the rivers in the interior of Otago.

But, granting that the material came down the rivers, how did it become diffused and spread out over the surface of the land near the embouchures of the rivers? Evidently it is not accumulating now, for it is covered up near Amberley by the gravels of the Kowhai; and it is everywhere being washed away by the rain. Also it is clear that it is not a river deposit, for it is too evenly spread, and a river could not possibly have taken it up the hills of Banks Peninsula. Hence we are driven to suppose that it must be either a marine or a wind-borne deposit. It must be one or the other—it cannot be both.

Sir Julius von Haast was the first advocate of the wind hypothesis.^* After reading Richthoven's theory of the origin of the loess in China he came to the conclusion that our silt-deposit was also a loess, and had a similar origin to that of China; and in this opinion he was supported by Mr. J. Hardcastle.^† Richthoven's theory is this: He supposes that in the arid interior of a large continent the decomposition of the rocks would produce a great quantity of dust, which would be blown away by the wind. On the outskirts of this arid district a more or less rainy district would be found where grass would grow. The grass would catch the dust, hold it together and grow over it, so that the dust would accumulate and form the deposit called loess. The capillary structure he thought to be due to the roots of the grass which had decayed away.

In applying this theory to a small island like New Zealand we meet with considerable difficulties. Where are we to find the desert area capable of producing so large an amount of dust? If we suppose that the land formerly extended much farther on each side of the mountains so as to make an arid area, how can we account for the absence of the silt from the interior of Otago and the neighbourhood of Dunedin? But the silt may have been brought down by the rivers and deposited in their beds, and may have been subsequently blown away as dust-storms over the plains. Here we have a possible source of dust in the old glacier muds, which did not require a desert to form them in. Let us

therefore see whether the æolian origin will account for the deposit, and what is the evidence in its favour. First comes the capillary structure. But even a cursory examination shows that it in no way resembles the fibrous roots of grasses. The tubules do not radiate downwards from centres where the grass-plants should have grown; neither do they taper downwards like roots, but change rapidly and irregularly in diameter. And the anastomosing is very different from anything seen in roots. No doubt some of these tubules are now occupied by roots, which have grown into them; but they are comparatively rare, and evidently do not belong to the time of the formation of the tubules. Then we have the occurrence of moa-bones, and, as stated by Sir Julius von Haast, of land-shells. But the remains of both birds and land-shells are found in marine deposits, having been carried down by the rivers into the sea; and the most important evidence is the general absence of marine fossils, even at the base of the deposit. But there is also the difficulty of accounting for the entire absence of decaying vegetation and carbonaceous compounds, which ought to exist if the wind-borne theory is true. If grasses were constantly getting covered over with dust the deposit must at one time have contained nearly as much vegetable matter as sand. What has become of it all, and where are the marks of the former plants? Why should the deposit be full of the marks of roots and yet not have retained any marks of leaves, stalks, or fruit, which form the largest part of the plant? The question has only to be stated to show the inadequacy of the wind hypothesis.

On the other hand, the absence of marine fossils may have been due to the rapidity with which the deposit accumulated and the short time the land was submerged, so that shells and slow-moving animals could not people the constantly changing shore-line.

The evidences in favour of the marine origin of the silt are—first, the actual occurrence of marine fossils at Oamaru, Banks Peninsula, and perhaps at Timaru; secondly, its stratification at Lyttelton and Timaru; thirdly, its position on the crests of the hills; fourthly, the occurrence in it of pebbles too large for the wind to blow away; fifthly, its equal development on both the lee and weather sides of Banks Peninsula; and sixthly, the angular shape of the grains. None of these facts can be explained on the theory of the æolian origin of the silt, and its advocates have failed to bring forward any crucial evidence in its favour.

According to Professor Boehm of Freiberg, who visited New Zealand in 1899, this pseudo-loess, as he calls it, is certainly of marine origin at Oamaru. “Whether it is so everywhere in New Zealand must be decided by the examination of abundant