Art. LI.—On the Lower Gorge of the Waimakariri.
[Read before the Philosophical Institute of Canterbury, 15th November, 1883.]
A Triangular portion of the Canterbury Plains projects into the mountains just north of the Malvern Hills. On the north it is bounded by Mount Torlesse, on the south-west by the Malvern Hills, to the east it merges into the main portion of the plains. This triangular portion of the plains slopes rather rapidly to the east, being at an elevation of 1,400 or 1,500 feet at the base of Mount Torlesse, and falling, in a distance of ten or twelve miles, to about 800 feet at the eastern end of the Malvern Hills and at West Oxford. The even surface of the plain is broken by deeply terraced valleys of the rivers that run through it, and by several small hills situated towards the eastern margin. Two of these hills—Racecourse Hill and Little Racecourse Hill—near the south-east corner of the triangle, are entirely composed of rounded shingle like that of the plains, and remind one of the Eskers of Ireland or the Kames of Scotland. I have not closely examined Racecourse Hill, but Little Racecourse Hill contains large angular erratic blocks washed out of some former morainic deposit, and now completely mixed up with river shingle, but there is no boulder clay.
Racecourse Hill is oblong in form and rises about 50 feet above the plain. Little Racecourse Hill is long, narrow, and rather curved, the concavity facing the east end of the Malvern Hills. Its height above the plain is about 30 or 40 feet. Gravel deposits are also found in places round the base of the Malvern Hills, at estimated altitudes of from 50–70 feet above the present level of the plains.* Ice-borne erratics are found in the upper portions of the triangle and extend down as far as the ford just below the junction of the Kowhai and the Waimakariri. These erratics are embedded in river shingle. Dr. von Haast also mentions large morainic accumulations on the western slopes of Abner's Head, a part of the Malvern Hills.† In addition to these shingle hills, three others, formed of rocks of far more ancient date than the gravels, rise through the plains. They are View Hill, Gorge Hills and Burnt Hills.
The Waimakariri River, rising in the heart of the New Zealand Alps, breaks abruptly into the middle of the north side of this triangle through a deep rocky gorge, called the Upper Gorge, at the eastern base of Mount Torlesse, and runs with a southerly course for about two miles, until it is joined by the Kowhai coming in an easterly direction from the apex of the triangle. The Waimakairi then flows in a south-easterly direction for about 4 ½ miles to Gorge Hill, which it cuts into two portions—Gorge Hill on the southwest and Little Gorge Hill on the north-east—forming what is called the Lower Gorge. This is crossed by the bridge for the Oxford and Sheffield railway now in course of construction.
[Footnote] * See Geological Reports for 1873–74, p. 57.
[Footnote] † Geology of Canterbury and Westland, p. 392.
That a river running in a shingle plain should make straight for a rocky hill and, instead of flowing round it, should cut it in two, is a remarkable phenomenon; and it is an account of this phenomenon and an attempt to explain it that forms the subject of the present paper.
First as to the facts which have to be explained. Both Gorge Hill and Little Gorge Hill are formed of thin bedded slates and sandstones dipping 75° to W.S.W. at the bridge, but getting flatter higher up the river. On the right bank is Gorge Hill, 363 feet above the level of the river-bed at the bridge.* From the summit a spur runs towards the river in an E.N.E. direction. Its altitude at the river gorge is about 270 feet. Then, following the same line, comes a depression occupied partly by the river, partly by gravel beds, and then, still continuing the same direction, is Little Gorge Hill, 216 feet high. The depression in the slate rocks between these two hills forms a steep-sided, flat-bottomed col, in which the river has cut a narrow perpendicular gorge about 90 feet in depth; the remainder of the col being covered with beds of river shingle, capped by a deposit of silt which, in colour and in structure, resembles the silt deposit at Lyttelton. No fossils have been detected anywhere in the neighbourhood, except the so-called annelid tubes in the slates.
The plains at the north-west foot of Gorge Hill are 192 feet above the river-bed at the bridge, and they extend from thence to the Malvern Hills without any terraces. The plains at the north-east foot of Little Gorge Hill are also 192 feet above the river-bed, and from here also they stretch over to Oxford without any terraces, except some small ones formed by the River Eyre which have nothing to do with the subject of the present paper. The river-bed, both above and below the gorge, is very wide and deeply terraced; but all the terraces, when they approach the gorge from above, contract suddenly and then, below the gorge, expand again as suddenly as they contracted, thus appearing like an hour-glass, the narrow neck of which is at the gorge. It is evident, therefore, that the river has never left the gorge since it first began to cut into the plains and form terraces. However much it may have swayed from side to side, either above or below, the hard rocks of the gorge have always held it like a vice ever since it began to cut below the present level of the plains. But how is it that the plains are at exactly the same level on both sides of the gorge hills? And how is it that the river has cut through the hill instead of running round it?
[Footnote] * The heights of the two Gorge Hills are taken from the trigonometrical survey of Canterbury as given by Dr. von Haast in his “Geology of Canterbury and Westland,” p. 476; the levels of the terraces are from the survey for the Oxford and Sheffield railway, kindly supplied me by Mr. G. P. Williams. All the levels here given are calculated from the bed of the river at the bridge, taken as the datum. This datum is 808 feet above high-water mark at Sumner. The figures on the plan show the height in feet above the river-bed at the bridge.
It is possible that at some former period a glacier may have overflowed these hills although there is no evidence of it, for the steepest slopes face the mountains; but if such has been the case it could not in any way account for the facts we wish to explain. A glacier certainly did not deposit rounded stones in large flat plains, and as these stones have evidently been brought down by the river, it follows that any glacier action that might at one time have existed here must have been entirely obliterated by the river gravels long before the phenomena we are now discussing came into existence. We may therefore dismiss all notions of a glacier as a cause. All the gravels in the neighbourhood have certainly, as I have said, been brought down by the river. But is it possible that the river could have deposited them at exactly the same level on both sides of the hills? I think not. But supposing, for the present, that this is possible, can river action alone account for the rest of the facts?
The steep sides of the col in the Gorge Hills and underneath the gravels, as shown by the railway cutting on the left bank of the river, prove that the col itself was cut by the river. This col must have been subsequently filled up with shingle and silt to the level of the highest terrace inside the col, that is to 148 feet above the present river-bed, or 58 feet above the bottom of the col. After that the river must have once more cut down through the shingle and rock and excavated its present gorge. It is evident that the river could not have commenced cutting the col until shingle had been piled up as high as the rocky spur out of which the col has been cut; for until that was done, it could not overflow the spur, but must have run round it. Now not only is there no trace of a river-bed on either side of the hills, but the shingle is not sufficiently high on either side to cover the spur; and as the river could not have piled up shingle over the spur without also piling it up to at least the same height on either side of the hill, it follows that this shingle, which must formerly have existed on both sides of the hill, has been since removed.
There is evidence in the two racecourse hills, and in the gravel beds round the eastern base of the Malvern Hills, previously mentioned, that the shingle did, at one time, reach a level some 50 or 70 feet higher than at present, and consequently it must at that time have completely covered Little Gorge Hill and a large part of the spur; but not Gorge Hill itself, which would still have projected more than 100 feet above the plain. The river may, therefore, at that time have run over the spur at the eastern base of Gorge Hill, and gradually excavated the col. But it seems to me impossible that the river after cutting the col, could have left it and swept off all the shingle to exactly the same level on each side of the hills, and then have returned to the col to cut the gorge. My reasons for thinking so are,
firstly, because there are no traces of terraces between the Gorge Hills and the margins of the plains; secondly, because having once cut the col the river would be held there by the hard rocks on either side exactly as it is held now; and, thirdly, because in order to remove the shingle-beds above the present level of the plains, the river must have first filled up the col with gravel to at least this height before it left the col, and there is no evidence of its having done so; the highest terrace inside the col being only 148 feet above the river-bed. Consequently the river can never have left the col since it first excavated it, and it is therefore impossible that the river could have removed the missing gravel beds on each side.
If now we take the only other possible view; that is, we suppose—although there is no evidence in favour of it—that the col was formed by some other agency than the river, which merely filled it and the surrounding plains to their present level; and then, happening to run through the col when it commenced cutting down—which is very improbable—it formed the present gorge. If we take this view then we shall not be able to explain, first, how the river could deposit flat beds of shingle, some ten miles wide, between Sheffield and Oxford; nor, second, can we explain the origin of Racecourse, and Little Racecourse Hills, and the gravel beds round the eastern base of the Malvern Hills; which are shown by the form and arrangement of the materials composing them not to be portions of old moraines, although Little Racecourse Hill contains re-arranged morainic matter. Neither of these hypotheses will explain the facts. If we suppose that the col was cut by the river, large masses of gravel must have been removed without any apparent cause. If we assume that the river did not cut the col, then we cannot account for the gravel hills rising above the plains. Evidently river action alone cannot explain the whole of the phenomena.
The following is my explanation:—The shingle brought down by the rivers not only formed the present plains, but accumulated to a thickness of some 50 or 70 feet higher than at present. At this time Gorge Hill projected above the shingle, but Little Gorge Hill and a large portion of the spur between the two were buried. The river, swaying about in the plains between Oxford and the Gorge Hill, was arrested in its southerly swing by the latter, and then, running along the north-easterly base of the hill over the spur, it cut down through the gravel beds into the rocks and formed the col. A period of general subsidence of the land followed, during which the river partly filled up the col again with shingle, but at length, the depression still going on, the sea reached, and passed, this portion of the plains, sweeping away the upper 50 or 70 feet and reducing all to a common level, except the Racecourse Hills. The col, remaining as the estuary of the
river, never filled up to the height of the plains at their reduced level, and, on the land being re-elevated, it remained as the river channel and was gradually cut down into the gorge.
There are several quite independent reasons for thinking that the sea has flowed over the Canterbury Plains; but these I need not repeat here as they will be found in the Transactions of the New Zealand Institute, vol v., p. 387, and in the Geological Reports for 1873–74, pp. 56–58.
On the other hand two objections may be fairly raised to my hypothesis. First, that there is no trace of flattened sea-shingle; and second, that there are no marine fossils. With regard to the first objection we must remember that the form of beach shingle depends a good deal on the structure of the rocks from which it has been derived, and that if the rocks have not a schistose or slaty structure the shingle must remain for a considerable time within reach of wave action before it can acquire a flattened form. This objection may therefore be got over by supposing that subsidence and re-elevation were continuous and tolerably rapid. The second objection rests only on negative evidence as no fossils of any kind have been found. It is of no value here where the beds, except quite the upper portions, are allowed to be of fluviatile origin, and are composed of shingle through which water easily percolates; for under these circumstances all calcareous remains are soon dissolved. I think, therefore, that the objections to the hypothesis here put forward are of far less weight than the objections to any other hyphothesis; but, whatever value may be attached to the argument, it is certain that no theory of the formation of the Canterbury Plains can be satisfactory if it fails to account for the lower gorge of the Waimakariri.