Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 72, 1942-43
This text is also available in PDF
(4 MB) Opens in new window
– 148 –

A Detail of the Pukaki Moraine.

[Read before the Canterbury Branch, July 2, 1941; received by the Editor, March 9, 1942; issued separately, September, 1942.]




Outlet of the Lake.


Structure of the Ridge near the outlet of the Lake and its Special Features under headings:—


River Gravels.


Silts, etc.




Lake-Beach Deposits.


Summary of Evidence.


Discussion of the Problem.


General Sequence of Events, including references to other parts of Canterbury and to the Patagonian Shingle-Formation.

A. Introductory. (See Sketch Map and Plate 7.)

At the height of the Pleistocene Glaciation of New Zealand, the Tasman Glacier extended from the main divide of the Southern Alps north-east of Mount Cook to about four miles beyond the southern shore of Lake Pukaki, a total distance of about sixty miles. Its end moraine, with adjacent laterals and outwash deposits bounded by moraine, covers an area of over twenty square miles and once formed a barrier extending right across the mouth of the Tasman Valley, but fairly early in its glacial history it was breached near the western end, and the gap so formed acted as the main outflow channel for the glacial and other waters of the Tasman watershed. The water flowing through this gap was in all probability not responsible to any marked degree for the building up of the main outwash plain which forms the floor of the upper part of the Mackenzie intermont. The surface deposits of this are to be credited chiefly to streams which issued from various and fluctuating points of the moraine east of the gap, their former presence being evidenced by numerous abandoned water-courses, some of large size, which originate within or on the outskirts of the moraine, intersect the upper fringe of the plain fronting it, and gradually merge into its general surface.

After a recession of the ice, during which the gap was fully maintained if not actually widened, a re-advancing tongue occupied its floor. Subsequently this was deeply incised by the Pukaki River, and the sections furnished by the latter's steep banks give some insight into the succession of pre-glacial, glacial, and other deposits, characteristic of the area.

East of the river, the southern shores of the existing Lake Pukaki rise gently at first, and then more steeply, for some 450 feet, and the land-surface then falls with progressively lower broken cross-ridges having moderate and irregular southern inclination and passing

– 149 –
Picture icon

Sketch. Map Outlet, Lake Pukaki

into more gentle slopes which merge ultimately into the outwash plain. A considerable amount of destruction has been caused by the former outflowing streams, and remnants of old morainic dumps fringe the upper part of the plain, getting less and less definite in form till they disappear entirely, or are represented by sporadic masses, and these in turn become rare and finally are quite absent when the surface of the plain is maturely developed. However, about four miles from the shore of the lake, a series of disconnected, well-developed dumps composed generally of large blocks rises well above the surface. Its line has in general an east-west orientation and just crosses the river to the western bank near the old ford. I know of no definite evidence of the former presence of ice further downstream from this line of dumps. After their deposition the ice must have retreated, become stabilised on the line of the great moraine for a considerable period and developed the outwash plain between it and the dumps just referred to.

– 150 –

On the western side of the river lies a smaller area covered by the main moraine; it is not so high as that on the east, is more hummocky, and was evidently formed of material dumped on ground removed from the direct action of the main ice-stream. It is really latero-terminal moraine.

The surface of the morainic complex facing the lake is covered in general with angular blocks, from the highest point right down to lake level, and it might be a reasonable inference that the whole deposit was composed of similar material, did not the sections alongside the Pukaki River disclose facts discrediting such a conclusion. It should be noted that the present lake lies behind the great moraine, but is not ponded by it. It may have ponded back an older lake, but there is no definite evidence of its existence.

It will be appropriate at this stage to record the heights above sea-level of various features under consideration as given by the records of the Lands and Survey Department.

Height of summit of moraine east of the river 2029 feet
Height of summit of ridge west of the river 1914 feet
Height of outlying dumps of main moraine east of the river 1655 feet
Height of upper surface of the outwash plain 1589 feet
Height of plain at maximum extension of the ice 1492 feet
Height of surface of the lake 1588 feet

It should be noted with respect to the last that the level of the lake varies considerably with the season, so the figure must be taken as an approximation. Other heights, as well as the thickness of the beds cited in this account, have been determined as accurately as possible by means of an abney level.

B. Outlet of the Lake. (Plates 7, 8, 9, and 11.)

The river discharges from the present lake through a somewhat open valley cut in gravels and overlying glacial deposits (Plate 7). As the exposures of the beds involved are more clear and complete on the western bank, these will be described first. The surface rises gently from lake level for some seventy feet to an old beach, which is a very persistent feature of the lake shores. There is a narrow bench some ten feet higher, and this is succeeded after another sixty to seventy feet by a shelf formed of irregular mounds, presumably morainic, and then finally by a well-defined, narrow ridge, looking like a terminal moraine (Plates 9 and 11), and ending in a cliff at the river end. It has a fairly even summit, which at Trig. B reaches a height of 226 feet above the lake. Near the cliff-face the ridge bears W.N.W., in its middle section W. 15° N., further west the direction reverts to W.N.W., and this continues as far as the road crossing, beyond which the ridge appears as a detached section with N.W. orientation; and then, after an interval, it is resumed with northerly direction, to terminate about half a mile north-west of the outlet of the lake. At this end lies the deserted bed of an old outflow from the lateral glacier margin (Plate 11), which follows concentrically round the outside of the ridge, channels the outwash plain at first, and finally merges with it.

On the eastern bank of the river, the lake-beach is developed close to the outlet, but modified and reduced in height near the downstream end (Plate 8), while behind it, that is, to the east, stretches

Picture icon

View looking east across the southern end of Lake Pukaki. The middle distance is occupied by the great morainic complex (A). The extension of the ridge with silts to the east of the river is marked by (R), the whole length being shown; the doubtful extension of deposits of the same glacial phase to the north by (R?). Behind the former and to the right of the latter is the site of the old lateral stream bed, while on the extreme right of the picture the boulder-clay under the moraine shows as a white scar at the top of the river terrace. The seventy foot lake-beach in the foreground is marked (C) and lies on both sides of the road; it shows on the east side of the river near the line of the top of the trees.

Picture icon

View looking downstream from just below the hotel The terrace (C) on the left is part of the seventy-foot lake-beach lowered in level to about sixty feet by river erosion, while the true level of seventy feet appears on the west of the river (C). The cliff ending the ridge (R) shows in the middle distance, and the outwash plain (O), composed of gravels, is in the fan distance.

Picture icon

Looking upstream towards the river-cliff. The outwash plain (O) with occasional boulders is on the left; the tussock-covered slopes (B) are probably of gravels, the upper levels of which can be faintly seen a little higher than the track. The silt beds marked (R) lie above them. The older morainic complex west of the liver can be seen (D) in the distance on the right, in front of which lies the extension of the ridge (R') to the north. The flat terrace, mentioned in Section C, 4, shows on the extreme right of the picture between the letters D and R'.

Picture icon

Face of the cliff, view looking south. The junction of the gravels with the overlying silt is marked in places by a broken black line. A mass of included gravel with interstratified silt is marked (A) and masses of angular stones are marked (B); included boulders show in the front of the picture.

Picture icon

View looking east from near the road-cutting in boulder-clay and showing the western and central portions of the ridge (R), the surface of the outwash plain lies on the right (O), which has been incised by the stream whose deserted channel (S) shows in the foreground.

– 151 –

a detached ridge (Plate 7) with N.–S. orientation, composed apparently of morainic blocks, but evidently a remnant of the ridge west of the river. North of this remnant lies an area of morainic hummocks perhaps dating from the same glacial phase. Behind the ridge fragment stretches an old, deserted stream channel, the counterpart of that west of the river, which joins the present river with marked discordance, and close to the junction a fragment of outwash plain persists at a level accordant with that of the plain west of the river. East of the deserted channel, the ground rises fairly steeply to the summit of the great moraine which in this locality reaches a height of 400 feet above the lake.

At the actual outlet of the lake, the cross-section of the channel of the Pukaki River is open (Plate 7), but this rapidly becomes more confined, and finally the banks are steep and almost precipitous. This is the case where the river cuts through the ridge just mentioned (Plate 8); where it enters the outwash plain, a quarter of a mile from the lake, the bed is deeply incised, 110 feet on the west and 180 feet on the east bank, into the surface of the plain.

C. Structure of the Ridge. (Plates 8, 9, and 10.)

This ridge resembles a typical moraine in external appearance, and it may be so in general, for large angular blocks lie scattered almost everywhere on its surface, but there are no clear-cut sections except that furnished by the cliff facing the river, and this shows marked departures from the ordinary structure of a moraine; and it is possible that, if other sections could be obtained, additional discrepancies might be observed. As the section of the cliff is most important, it will be described in detail. The sequence exposed is as follows:—

1. River Gravels. (Plates 9 and 10.)

These gravels extend from river-level to a height of 110 feet. Their lower slopes are masked with debris, but there is no suggestion that beds other than gravel occur at this horizon. The upper slopes are steep and form the lower members of the actual cliff-face. The clear-cut section here displayed shows that the deposit is composed of subangular and rounded stones in a finer matrix, and with no sign of included morainic material. The uppermost layer, however, contains numerous water-worn blocks, some up to three feet in diameter, and the deposit may be fluvio-glacial in origin. On the cliff-face the upper surface of the gravel appears to be slightly irregular (Plates 9, 10), and near the downstream end it is a few feet higher than the river terrace in close proximity. The material of both shows a close resemblance in texture and composition, and the continuity of the two deposits does not appear to be broken. It is, therefore, reasonable to conclude that, as the gravels under the cliff must have been in position before the overlying beds were laid down, so the gravels of the outwash plain were likewise in position, and any modification they have experienced subsequently is only superficial.

The gravels cannot be traced upstream beyond the face of the cliff, and the beds exposed at a corresponding level are whitish glacial silts very feebly stratified. These appear at various levels down

– 152 –

to that of the river, and they are almost certainly a veneer masking other deposits in this locality. In a trial well sunk near the hotel on the lake-beach, similar silts were passed through, interstratified with rounded stones and angular debris. The gravels do not show for certain on the eastern side of the river near the lake, and higher than the lake-beach in that locality rounded gravels are exposed in a short road-cutting. The exposure is somewhat obscure and they may belong to the boulder-clay to be mentioned later.

2. Silts, etc. (Plates 8, 9, and 10.)

These beds form the top portion of the cliff which cuts right across the eastern end of the ridge (Plates 8 and 9). The length of the exposure is about ten chains, the height of the top of the cliff above the river 175 feet, and the maximum thickness of the beds exposed is sixty feet. They consist of stratified silts, somewhat sandy in texture, containing numerous isolated angular boulders, and also lenses and masses of gravel and angular stones usually in a matrix of silt and showing rude stratification. The lines of stratification pass through the gravel with directions corresponding to those of the adjoining silt. At the southern end of the section, the beds have a general synclinal arrangement with minor irregularities; but from about the middle of the exposure they dip upstream and show such distortions as might have been produced by strong local eddies or varying currents operating during deposition. At the extreme northern end of the section, is a small, well-defined syncline in the silts. The inclusions sometimes consist of masses of angular and sub-angular, occasionally scratched, stones, in a matrix of silt, and they thus closely resemble a boulder-clay. They occur at all levels and occasionally rest directly on the surface of the underlying gravels. The finer material has not the tenaciousness of a true boulder-clay, so they were not formed under a glacier in the manner characteristic of such deposits. It is here suggested that they, as well as the silts, were laid down in a lake or large pond lying alongside an ice-face, from which the silts were derived directly as the ice melted, and from which the boulders and other coarser material fell into still, but occasionally strongly agitated, water. The silts occurring at lower levels upstream have apparently no stratigraphical relation to those exposed in the cliff-face. Further west on the ridge exposures are absent; but, judging from the spoil thrown out of the numerous rabbit burrows on the slopes of the ridge, the silty material occurs some distance away from the cliff, though the continuity between the occurrences cannot be established. At the western end of the ridge boulder-clay is exposed in a road-cutting, and it also occurs further west under the cover of the old morainic complex. Both occurrences contain rounded and angular stones, sometimes scratched, and that lying under the western end of the ridge appears to correspond, in stratigraphical position, with the silts exposed at the eastern end, though they may antedate them entirely and belong to the boulder-clay definitely underlying the older moraine further west. It should be noted that similar deposits are exposed on the face of the high terrace on the east side of the river below the gorge, and here they lie on gravel and are capped by older moraine.

– 153 –

3. Moraine.

This forms a discontinuous capping to the ridge. Blocks lie scattered on the slopes, probably having rolled from a higher level, and they occur on the surface of the outwash plain to the south, becoming scarcer and scarcer as the distance from the ridge increases. They end about half a mile away, but are resumed about three miles downstream, near the old crossing of the river, in a line with the series of dumps mentioned previously. Both these occurrences belong to the earlier complex and do not indicate a former wider extension of the ridge; there is no satisfactory evidence that this ever occurred.

4. Lake-Beach Deposits. (Plates 7, 9, and 11.)

Although these deposits have apparently no relation to the problem suggested by the ridge just described, they do deserve some further notice. The chief beach lies at a height of about seventy feet above present lake-level, and another of less importance ten feet higher still. The first has a variable width, sometimes as much as three or four chains, and this suggests a somewhat long period when the level of the lake experienced little change. The higher shelf must be regarded as an ephemeral feature. The tops of both are masked in general by typical beach gravel, on which morainic blocks lie or through which they protrude. Silts appear to be interstratified with the gravels and must have been deposited in a manner similar to that which obtains at present at the southern ends of lakes like Tekapo. The slopes leading down from the main beach occasionally show subordinate shelves, but in some cases they are absent, and this suggests that the lowering of the barrier behind which the lake was ponded was fairly continuous and rapid. It must be noted that the surface of the outwash plain below the ridge lies at a higher level than the main beach, and even than the narrow one ten feet higher, so the gravel of the plain could have functioned as an efficient barrier when the lake stood at its highest level without calling in the aid of moraine as a ponding agent.

A peculiar terrace also occurs on the west side of the lake about half a mile from the outlet, and lying at a height of approximately 165 feet above it (Plate 9). The surface is almost flat and is floored with morainic blocks, though pebbles of typical beach form occur very occasionally, supporting the suggestion that it is a lake terrace. However, it is not clear whether or not it belongs to a higher level of Lake Pukaki, or to a marginal lake ponded between the edge of the ice and the older morainic complex which bounds it on the wese. This shelf heads the stream-bed which leads south, outside the ridge, on to the main outwash plain referred to earlier (Plate 11). There are several similar shelves at a fairly accordant level around the lake; but they generally lie in indentations of what would be the former shore-line, had a lake existed at that level; and there is no definite sign of the presence of shelves on the points of a shoreline, where they should have been cut rather than in the sheltered indentations. Therefore I conclude that they are accidental features, or may have been formed in marginal lakes as I have described.

– 154 –

D. Summary of Evidence.

These seem to be the main geological facts on which a discussion of the precise mode of origin of the ridge must be based. It is certainly not a typical terminal moraine taking it as an entity, though sections of it may be classed as such. This applies to the western end and may also apply to that section east of the river. In the case of the former, near the road crossing, morainic material rests directly on boulder-clay, which is what one might expect if it were a typical end moraine, but the cliff-face and, to judge from the rabbit burrows, the central portion also, cannot be explained in this way. The problem therefore is not simple. Although some destruction must have taken place when the river breached the ridge and cut down through the gravels beneath it to present water-level, the form of the ridge does not indicate that it is part of an extensive deposit which has survived erosion, but that its present form closely resembles that which it had originally.

E. Discussion of the Problem.

The special difficulty presented by the case is that of furnishing a satisfactory explanation of the conditions under which the stratified silts, presumably lacustrine, could have been laid down in the apparent absence of any barrier, at present or formerly existing, behind which the water of the lake or pond could have been empounded. The only barriers that the geological circumstances of the locality permit of consideration are (1) gravel, (2) moraine, (3) boulder-clay, and (4) ice, and to be an effective ponding agent, the barrier must have reached a height of nearly 200 feet above the present surface of the lake. However, there is no evidence whatsoever for the presence of any of the first three at the necessary height or in the required position. There remains the fourth suggestion, viz., that the barrier was of ice, and temporary in character. In this case deposition of the silts could have taken place either in a broad and partially water-filled crevasse near the ice-front, or in a lake lying between the residual of a former ice front and the main body of the glacier. Such a case has been described by Matthes (1940, p. 399) as occurring in connection with the Conness Glacier of the Sierra Nevada. Any silts derived from such a glacier and deposited in the lake would have the features described in Section C, pt. 2, of this article. This suggestion is advanced with all due diffidence and with a full realisation of its uncertain character, but some explanation has to be furnished. It may be commented in this connection that if an adequate barrier, not being of ice, had really existed, why were not the incompetent silt beds removed while it was being destroyed?

Alternatives can no doubt be put forward in opposition to this suggestion, and they include:—

  • (1) The silts may represent a portion of the surface drift laid down by streams from the ice-front, but their stratification and thickness are against such an explanation.

  • (2) The ridge might be considered as a normal deposit laid down under the frontal portion of the glacier. This may be correct for the western end, where a veneer of moraine rests on boulder-clay, and perhaps for the portion east of the river, but such an explanation does not fit the case of the deposits exposed in the river-cliff.

– 155 –
  • (3) The silts could be regarded as a moat deposit, laid down near the ice front, but, as pointed out earlier, there is no indication of a barrier behind which the moat could lie, if ice be ruled out as a possible barrier.

  • (4) The winding disposition of the ridge and the form of its cross-section suggest that it might be an esker, but the structure as exposed in the cliff, and the orientation of the ridge across the valley and not roughly longitudinal with the direction of ice-flow are against such a hypothesis. In neither particular does it resemble the typical eskers of Sweden and Ireland.

So, to sum up, I consider that a portion of the ridge may be explained according to hypothesis (2), but that the position and character of the silts require an explanation similar to that advanced in the first part of this section. The extensions of the ridge northwards on both banks of the river are no doubt latero-terminal moraine perhaps resting in places on boulder-clay.

F. General Sequence of Events.

As the area provides evidence of at least two glacier advances, it may not be out of place to give in conclusion a brief summary of the events which took place in the neighbourhood, during and just after, the Pleistocene glaciation. It seems reasonably certain that gravels were in position near Pukaki before the great glacier extension, for they occur under the boulder-clay and moraine on the banks of the river just below the gorge. This agrees with the conclusion arrived at in the case of the analogous Tekapo moraine (Speight, 1940, pp. 183–4), a conclusion supported by Mr. T. G. Beck, Public Works Engineer, in a personal communication. He first of all states that the tunnel at Tekapo, mentioned in my account of the locality as having been driven for a short distance near its outlet in gravel underlying moraine, was continued in gravel under moraine throughout its entire length, and, secondly, he states that a shaft sunk to a depth of 80 feet in order to cut this tunnel near the margin of the lake passed through a thin veneer of moraine, and then through ordinary river gravel, some of it hardly consolidated. These observations give strong support to the hypothesis that gravels, post-dating those correlated with the Kurow Series of Marwick (1935, pp. 329–39) were in position on the floor of the Mackenzie intermont before the deposition of the great morainic complex, and that this complex with its boulder-clay, though reasonably thick, perhaps some hundreds of feet in places, is really only a veneer resting on pre-existent undeformed gravels.

The relations just mentioned are thus analogous to those of similar deposits laid down on the western margin of the Canterbury Plains near the localities where the Rangitata, Rakaia, and Waimakariri Rivers issue from the mountain area east of the main range of the Southern Alps. In these cases moraine forms a veneer over pre-existent gravels, and is not interstratified with them. Further, in only one case that I have observed in connection with the deposits near the mouths of the valleys of the Canterbury rivers, is there evidence that glacial beds were laid down before the deposition of the gravels of which the plains are composed. This occurs at the

– 156 –

lower end of the Rakaia Gorge where, on the south bank, the basal beds are pre-Senonian rhyolite covered unconformably by boulder-clay, fluvio-glacial material, and varved silts in that order. On the north bank of the river, above the rhyolite, there is an obscurity and then 130 feet of silts covered by 200 feet of gravels. Gravels also cover the silts on the south bank, and downstream from the exposure, thick gravel beds occur at a higher altitudinal level than the silts, and may be presumed to be more recent. This evidence seems to be conclusive; but there are difficulties in the way of accepting it at its face value, especially as regards the existence and composition of the barrier behind which the silts were deposited. The only possible barrier appears to have been a mass of pre-existing gravels, and this implies that a considerable deposit of this material must have antedated the glacier advance responsible for the formation of the boulder-clay and silts. On the southern bank, too, the section connecting the silts with the gravels downstream is not clear. In the other rivers no similar glacial deposits can be observed to rest on Late Palaeozoic or Mesozoic greywackes where they might reasonably be expected.

Since writing the above, I have received from Dr. C. Caldenius, of Stockholm, Sweden, an account of the Patagonian Shingle Formation (1940) in which he notes (p. 161) that Nordenskjold compared this deposit to the Canterbury Plains. Although the former author was unable to show the relationship of this Patagonian deposit to the Quaternary glacial beds of that region (p. 177), in his summary he assigns the spreading out of the shingle-formation to a climatic deterioration which took place immediately anterior to the great Quaternary glaciations. This conclusion as to date agrees with that determined for the gravels of the Mackenzie intermont lying under the moraine, and for those of the Canterbury Plains, provided that in the latter case one can explain the deposition of the varved silts in the Rakaia Gorge as post-dating that of the gravels as a whole or of a considerable thickness of their lower levels. This may present some difficulty, though perhaps not an insuperable one.

After this digression I return to the consideration of the conditions at Pukaki.

As the ice advanced over the gravel substratum near Pukaki, the streams issuing from various sections of the front modified the surface of the outwash plain, not only by adding fluvial material, but also by carrying forward blocks dropped from the ice-front. The possibility that these may have been carried considerable distances renders it difficult to state how far forward the glacier really came, but it certainly reached the series of dumps lying about four miles from the lake shore. During retreat, the outwash plain above the dumps gradually developed, the process continuing while the ice-front was stabilised on or near the line of the great moraine. This plain then reached a mature stage and coalesced in accordant level with that below the dumps, partially overwhelming the latter in the gravel flood. Some, no doubt, were completely covered, a few to be brought to light when the Pukaki River incised its bed into the plain at a later date. During this period the surface of the plain took its present form, and its maturity of development indicates that considerable time elapsed while the ice was retreating from the line

– 157 –

of dumps to that of the main moraine, and while the ice-front was stabilised in that position. The enormous mass of material constituting this feature also implies a relative stability of front or but slight variation therefrom for a long period.

This stability was followed by retreat when the gap was opened near the south-western corner of the lake, through which the concentrated drainage of the Tasman-Pukaki valley and its tributaries discharged. It is possible, too, that a lake was then in existence during this episode, and the conditions may have resembled those now existing. Additional gravels would be laid down on the outwash plain, building up its surface further and extending upstream through the jaws of the gap. The ice advanced again and was in some way responsible for the formation of the ridge under special consideration. Whether this advance was a mere modification of the ice-front, or a second glacial episode of major importance, cannot be stated; but the width of the gap suggests that a considerable interval of time elapsed between the formation of the complex and the ridge. If the circumstance is of major importance it adds another to the critical events of the Pleistocene Glaciation. It has been maintained (Speight, op. cit., p. 180) that the cross-section of the Tasman Valley indicates that there were two periods of glacier advance, the great moraine at Pukaki belonging to the latter. So, if another is added, there may have been three distinct phases of the Pleistocene Glaciation of the region, and if the line of dumps previously referred to is of major importance, there may be yet another.

After the formation of the ridge at Pukaki, a retreat of the glacier occurred once more, and the hollow behind the ridge and the older complex was occupied by a lake whose surface, for a comparatively short period, lay eighty feet above the present lake level. It then fell some ten feet, when a broad persistent beach was formed, and then it fell gradually to what it stands at now. The surface of this lake does not appear ever to have reached the level of the upper-most gravels under the ridge, or the surface of the plain just in front of it, so the gravels functioned as the barrier behind which the lake was ponded, moraine playing a relatively unimportant part in its formation. The cutting down of the gravel barrier and the formation of the gorge of the river were no doubt responsible for the fall in level of the lake. Downstream the plain is deeply terraced, but the terraces gradually decrease in height till they disappear as such and merge into the general surface of the plain, which at this stage is being aggraded, not only by the Pukaki, but also by the Tekapo and Ohau Rivers, before they unite to form the Waitaki River and to cut the deep, rock-bound gorge through the mountains which rim the intermont on its southern margin.


Caldenius, C., 1940. The Tehuelche or Patagonian Shingle-Formation, Geografiska Annaler, H. 3–4.

Marwick, J., 1935. The Geol. of the Wharekuri Basin, Waitaki Valley, N.Z. Journ. Sci. and Tech., vol. 16, pp. 321–338.

Matthes, François, 1940. Report of the Committee on Glaciers, Amer. Geophys. Union Trans. 1939–40.

Speight, R., 1940. The Gravels of the Mackenzie Intermont, Trans. Roy. Soc. N.Z., vol. 70, pp. 175–187.