Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 53, 1921
This text is also available in PDF
(2 MB) Opens in new window
– 47 –

Art. V.—The Modification of Spur-ends by Glaciation.

[Read before the Philosophical Institute of Canterbury, 6th October, 1920; received by Editor, 31st December, 1920; issued separately, 27th June, 1921.]

Plates VIIXI.

The subject of the changes which glaciers exert on the form of stream-valleys is such an interesting one that special aspects are worthy of detailed consideration. It has not, however, been fully considered so far as this country is concerned, although Andrews in his classic paper on the glaciation of south-western New Zealand (1905) has drawn attention to certain forms, such as the total truncation of spurs, and the development of sitting-lion and titan-beehive shapes, as well as the formation of a double slope on the valley-sides and especially on the spur-ends. The present author has pointed out certain other features (1907 and 1911), but observations made during the past few years in the alpine region of the South Island of New Zealand have suggested that still other forms exist. The faceting of spur-ends as a general result of the overdeepening of glaciated valleys and the formation of tributary hanging valleys has been dealt with in various places by W M. Davis, G. K. Gilbert, de Martonne, and others; but apart from this, judging by the literature at my disposal, little has been written. Davis has, however, insisted that the detached knobs on the floors of valleys, either separated from or in close proximity to the valley-walls, are remnants of a pre-glacial land-surface which have escaped destruction. He says (1900, p. 274), “On entering the glaciated valley of the Rhue it is found that the regularly descending spurs of the non-glaciated valleys are represented by irregular knobs and mounds, scoured on their up-stream side and plucked on the down-stream side; and that the cliffs formed where the spurs are cut off are sometimes fully as strong as those which stand on the opposite side of the valley. The spurs generally remain in sufficient strength to require the river to follow its pre-glacial serpentine course around them, but they are sometimes so far destroyed as to allow the river to take a shorter course through what was once the neck of a spur.” Again, on page 276 he says, “It is seen that just before the complete obliteration of the spurs some of the remnant knobs may be isolated from the uplands whence these pre-glacial spurs descended. It is out of the question to regard the ruggedness of such knobs as an indication of small change from their pre-glacial form, as has been done by some observers. The ruggedness is really an indication of the manner in which a glacier reduces a larger mass to smaller dimensions by plucking on the down-stream side as well as by scouring on the up-stream side. It is possible that knobs in other glaciated valleys than that of the Rhue may be of this origin; they should then be regarded not as standing almost unchanged and testifying to the incapacity of glacial erosion, but as surviving remnants of much larger masses, standing, like monadnocks above a peneplain, as monuments of the departed greater forms.”

– 48 –

The glaciated knobs of the Central Plateau of France that he notes later on hardly come into this category, but on page 288 he refers to rocky knobs seen in abundance about Ambleside and along the ridge separating Thirlmere from St. John's Vale, in the County of Cumberland, in England. In this paper he everywhere emphasizes the potency of glacier erosion, especially in valleys.

In a subsequent paper (1905, pp. 4–5) he again refers to knobs: “The knobs and ledges may be taken to be so-many unfinished pieces of work, which would have been more completely scoured away had the glacial action lasted longer.” This point he again emphasized in a paper on “American Studies on Glacial Erosion” (1910, p. 423), and refers to it slightly in the discussion on his account of the glacial features of North Wales (1909), and also in an answer to a question on glacial knobs addressed to him by M. Allorge.

This is a summary of Professor Davis's position as far as I can see from the literature at my disposal. It will be noted, however, that nowhere in the papers I have cited has he illustrated his point by showing the various stages by which a spur actually develops into a field of knobs; and this is somewhat surprising, as the method would be one entirely in keeping with the way in which he so frequently presents a physiographic problem.

I have examined other authorities, such as Hobbs and de Martonne, and find that faceting is everywhere recognized, but no other forms are noted. In the report of the Harriman Expedition to Alaska, G. K Gilbert deals exhaustively with the origin of hanging valleys and faceted spurs, but says little or nothing of any other of the various stages of modification However, I have examined the maps and illustrations and can see little evidence of intermediate forms, with the possible exception in the case of Nunatak Glacier (p. 59, and map), where the Nunatak appears to be a detached knob or end of a reduced spur.

Since there is this absence of statements concerning intermediate forms, I have attempted to supply some evidence as to their occurrence which I have come across during years of intimate acquaintance with the alpine region of the South Island of New Zealand. Incidentally this will be found to support Davis's contention that fields of knobs in the floor of a glaciated valley represent the remnants of spur-ends.

The main effect of ice-action on valley-spurs is due to abrasion, although no doubt plucking is very important at times, and especially in its more mature stages, when the spur-ends have become faceted At this stage, too, the excavating-power of a glacier has a dominating influence on the resulting landscape-form. But the depth of the ice, its velocity, and the time to which the surface has been subject to its action all exert important influences; and, further, the direction in which the tributary valley meets the main valley also controls to some extent in its initial stages the result of ice-action on the spur-ends.

As its dominating influences are those of thickness, velocity, and time, the modification of valleys, and therefore of the spurs running into them, will be different in different parts of the valley, being more pronounced in the upper portions, owing to the fact that these agencies are there at their maximum. Those parts of the valley where the ice is thickest, its velocity greatest, are just those parts which have experienced its action for longest time, and therefore modifications will be carried further than in the lower reaches. It will follow also that the character of the pre-glacial topography will be most easily arrived at by a study at the fringe of the glaciated district,

Picture icon

Fig. 1.—Lake Manapouri, looking east, showing notched spurs with islands formed by ice moving across the spur from west to east.
Fig. 2.—Lake Manapouri, looking east, showing islands with characteristic profile, remnants of dissected spurs.

Picture icon

Fig. 1.—Lake Manapouri, looking west. Island in foreground with profile similar to those in Plate VII, figs. 1 and 2, but more rounded. A still more rounded form in the background farther west, it having been more exposed to erosion.
Fig. 2.—Semi-detached knob, Thompson Sound.

– 49 –

where the action has not been intense, owing to the thinness of the ice and the shortness of the period during which the area has been covered. Also, there will be a progression of phenomena, varying in intensity on moving from the outskirts of the glaciated area; and phenomena characterizing the areas where glaciation has been intense, inexplicable in themselves, may be elucidated from the intervening regions where glaciation has been intermediate in its intensity.

The region of the South Island whence most of the instances to be mentioned later are drawn had reached a submature stage in the cycle of erosion before the incidence of the glaciation. Valleys had been cut in an elevated area, and a well-developed stream-system had been established with long spurs trailing down into the main valleys; but the district was one of alpine character, with peaks approaching in elevation, if not exceeding, the present European Alps.

A most interesting case illustrating the nature of the slight modification to which spurs may be subjected on the outskirts of a glaciated area is furnished by Lake Manapouri. The chief complex of spurs entering the basin occupied by the lake reaches down from the north, the spurs running in a north-and-south direction, whereas the direction of the chief ice-stream was from the west, and in its passage eastward it cut across the long trailing ridges of the pre-glacial land-surface. Erosion was most marked in the western reaches of the lake, where the ice was thickest and had acted for a longer time, so that a great trough or hollow was formed, with precipitous sides carried far down below the present surface of the lake (depth 1,458 ft.). Not all of this is to be credited to excavation by glacier-action, but some portion to the damming-back of the water by the morainic bar of the combined Te Anau and Manapouri glaciers. While the ice has profoundly modified the western portion of the lake and removed the spurs of the pre-glacial valley-system, the change in the eastern spurs has been slight, merely cutting them into a series of notches placed one below the other down the backbone of the ridge, all with the same characteristic profile, and continued down to lake-level, where exactly the same landscape-form is reproduced in the islands that dot the lake. (See Plate VII, figs. 1 and 2.) These notches form a kind of stairway with the treads inclined backward so that the level of the tread is lower at the foot of the riser than on the edge of the tread (cf. glacial stairway in a valley). The spur has thus been little modified, so that its original form can be restored. The slight modification suggests that the ice, though deep, as is evidenced by the height up the spur to which the series of notches reaches, can have exerted its action for a comparatively short time or it would have produced a profounder impression. Although signs of ice-action are found some twenty miles to the east of these spurs, the period of advance must have been quite short, or the spurs would have been more profoundly modified. Traces of this peculiar landscape-form are to be found on all the spurs to the eastern end of Manapouri where they were likely to be exposed to the full force of the ice-flood, so that it can hardly have been an accidental feature. Still farther eastward the spurs are unmodified. In Plate VIII, fig. 1, which is a view of the lake looking west, there are also signs of the same form, with a more developed knob in the background.

The form of the modified spurs suggests another point. Judging from the shape of the islands which lie off the ends of the spurs, it is clear that before the ice-advance the spurs continued down below the present level of the lake. It therefore negatives the idea that the hollow in which the lake

– 50 –

now lies has been entirely due to glacial erosion. In my opinion, the hollow is primarily tectonic, but the surface so formed has been modified by stream-action, succeeded by glaciation, and that now a new cycle of stream erosion has commenced.

The form of the notches cut in these spurs is also characteristic of ice erosion, since glaciers always appear to exert their maximum erosive effect at the base of the valley-sides or shelves along which they move. Thus the notches have the backward slope which results from this mode of action. When this becomes more pronounced and ice-action has been more prolonged, the outstanding portions of the ridge tend to become rounded eminences. If a stairway was attacked further the notches would become a string of knobs, gradually getting higher as the spur is followed upwards. This stage of development is seen in the Waimakariri Valley to the west of the Cass River, where Mount Horrible and Mount Misery owe their rounded form to the great Waimakariri Glacier crossing a spur which runs parallel with the present Cass River and enters the main valley nearly at right angles. (Plate IX.)

The formation of a well-developed series of notches generally occurs where the spur has great length; but if it is shorter in the pre-glacial stage only one or two notches may be cut, and the resulting form becomes a semidetached knob or titan beehive noted by Andrews in the Sounds region. (Plate VIII, fig. 2.) This form is typically developed in the Upper Rakaia Valley at Mein's Knob and Jim's Knob, the latter being formed by the Ramsay Glacier passing over the terminal spur of the Butler Range. (Plate X, fig. 1.) Numerous illustrations in all stages of development can be seen in the Upper Waimakariri Valley, especially where the action of the main glacier has not been interfered with by the weight of the ice issuing from a tributary comparable in size to the main stream. When the tributary becomes large the modification of the spur is attributable chiefly to its action, and not to the erosion of the main stream.

From the slight difference in the form of the notches in the higher part of the series as compared with those at floor-level it is evident that all the notches were cut during one period of ice-advance. Had there been more than one ice-flood, reaching various levels, there would have been some difference in the form of the higher members of the series from that of the lower. As the lower members would have experienced more than one ice-flood, their stage of erosion would have been more mature. Also, if the ice had not reached so high in the later floods as in the earlier the exposed notches or knobs at higher levels would show more the effects of subaerial erosion, by rain, frost, &c. For example, if the first flood were the highest, then while the lower levels were being subsequently glaciated the higher and exposed levels would have been differentially modified by subaerial erosion, and glacial erosion of the lower slopes would have been carried to a more mature stage. If the last ice-flood had been the highest, the modification of the higher levels would have been different in that the glacial surface would have been juvenile, while the lower would have been mature. If an intermediate flood had been the highest, a differential modification partaking of both characters would have occurred, depending on the relative importance of the two phases. But the only difference—and that is a very slight one—is that which might have been expected in the lower parts of a glacier, where, under the influence of greater weight of ice, abrasion, plucking, sapping, and other glacial agencies are more intense.

The knobs of the Cass Range show very markedly the modifying effect of frost erosion, as their plant covering is of the scantiest—in marked contrast

Picture icon

Range west of Cass River, Waimakariri Valley, showing partially dismembered spur with rounded knobs.

Picture icon

Fig. 1.—Jim's Knob, Upper Rakaia Valley. A semi-detached knob, showing channel where ice has attacked and cut down the termination of a spur.
Fig. 2.—Ice-cut bench on lower side of Bealey River at its junction with the Waimakariri River.

Picture icon

Fig. 1.—Jumped-up Downs, Upper Rangitata Valley, showing spur reduced to a field of knobs. View taken looking down-stream.
Fig. 2.—Jumped-up Downs, Upper Rangitata Valley, showing spur reduced to a field of knobs. View taken looking up-stream.

– 51 –

to the forest-covered slopes near Manapouri; but it is noteworthy that erosion has reached a similar stage in each individual of the series of knobs, suggesting that they were all formed by the same ice-flood, as is the case of those near Manapouri.

The tact that the series of notches in these spurs has been cut all at the same time suggests that the shelves existing in valleys of the European Alps may, in some cases, have been cut during one period of ice-advance. These are referred to by de Martonne in his Géographie Physique (p. 641). After describing the shoulders which are so characteristic of these valleys, and the location of villages on them, he says, “Les replats multiples indiquent que l'érosion des vallées alpines est le résultat d'une série de phases d'érosion glaciaire et d'érosion fluviale alternantes, produisant un enfoncement progressif du thalweg et un encaissement de plus en plus grand de la vallée, malgré les efforts faits par le glacier pour reculer le pied du versant par sapement à chaque période glaciaire !” Although it is dangerous to express an opinion without having seen the locality, it seems possible that these flats and shoulders may—in some cases, at all events—have been formed at one glacial effort, like those at Manapouri.

An important factor which affects the resulting form of the spur-remnant is the angle at which the pre-glacial valley of the tributary meets that of the primary. It will be most convenient to take the simple case when they meet at right angles or nearly so. Good illustrations of this case are furnished by the Bealey and Hawdon Valleys at their junction with that of the Waimakariri. The two tributaries come in from the north, whereas the main stream runs from west to east. The tributary valleys are subequal in size, and the size of the glaciers issuing from them at the height of the glaciation, judging from the present cross-section of the valleys, would be about one-fourth of that of the main stream. As a result of the greater weight of the ice in the main valley, the tributaries were crowded over the shoulder of the spur on the down-stream side of the tributary, with the result that they have both a flattish shelf about 100 ft. above the present floor of the valley and about 200 yards in length, formed by the cutting-down of the end of the spur, so that it terminated in a kind of platform analogous to the wide shore-platforms sometimes seen off a point on a coast-line composed of moderately soft rocks. (See Plate X, fig. 2.) The two spur-ends are so similar in position, shape, and extent that they might easily be mistaken, and photographs taken from the opposite bank of the Waimakariri are almost interchangeable. The similarity in form is no doubt to be attributed to similarity in the conditions under which the spur-ends were reduced by the glaciers as erosive agents.

If erosion proceeds further the shelf is cut down near its proximal end, and the beehive form again results, but it is then flatter than that resulting from the passage of the main stream at right angles over a trailing spur.

If the tributary meets the principal valley at an angle greater than a right angle, as in the case of Harrison Arm and Milford Sound, or the Sinbad Valley with Milford Sound, then the form becomes accentuated. The formation, not of a shelf, but of the couchant-lion shape, takes place, but ultimately this must develop into the beehive form. This form is, of course, subject to profound abrasion, and is liable to be reduced by attack from both sides and also on top, so that it ultimately becomes a mere roche moutonnée, standing in the floor of a glacial trough, and apparently without genetic connection with the valley-sides. In most cases, however, such isolated rocks were once connected directly with the valley-sides, the

– 52 –

connecting ridges having been completely removed by glacial abrasion. All the different stages in the formation of such isolated rocks from spur-ends can be seen in the valleys of the Southern Alps.

Worthy of special mention are the detached hills which lie in the angle between the Poulter and Esk Rivers near their junction with the Waimakariri. They are the remnants of the spur which once came down between the two former rivers, and whose end was dismembered by the large glaciers which issued from the Poulter Valley and Boundary Creek Valley, crossing it near its termination.

Spurs are eroded on the up-stream side in a somewhat different way. There is no overriding except in the case of the main stream entering a distributory valley, as in the case of the Rakaia branching off into the Lake Heron Valley, or when a glacier crosses the mouth of a tributary valley which is bare of ice. When, however, both are full of ice the end of the spur is modified by an action which is analogous to the whirlpool that forms when two rivers join, as a result of which the end of the spur is ground back below the surface of the glacier, so that it presents a steep face at the angle between the streams.

When the tributary meets the main valley at an angle less than a right angle the spur-ends are cut back, though with less overriding of the end than when the angle is greater. Narrow shelves, somewhat resembling terraces, are the common resultant form. Excellent illustrations of these can be seen at the junction of the Macaulay River with the Godley, and in the angle between the Potts and the Rangitata.

When valleys are subparallel, then there can be little or no truncation of the dividing ridges, but these are dismembered and cut into lengths as the result of lateral corrasion, chiefly by means of small tributary glaciers of the corrie type whose heads ultimately meet and lower the divide. Thus we get the elongated rocky hills which are so frequent in our ice-enlarged intermontane basins, which if submerged would produce elongated islands in parallel or linear arrangement, such as those which add to the scenie beauty of the West Coast Sounds, notably Dusky and Doubtful Sounds.

In the figures given by Davis illustrating partially destroyed spurs, fields of knobs appear to be a common feature. I have noticed occurrences similar to these in places where spurs have been partially destroyed—e.g., in the valley of the Harper River to the north-east of Lake Coleridge; but the most characteristic occurrence is in the valley of the Rangitata at the place called by the somewhat striking name of the “Jumped-up Downs.” (Plate XI, figs. 1 and 2.) This is evidently the residual of a destroyed spur, and its irregular appearance is well described by the name given by the early settlers. Right out in the floor of the Rangitata Valley is an isolated rocky mound in a line with the hummocky area; this is evidently the remnant of a spur which reached a considerable distance into the wide basin now occupied by the river.

The surface of these hummocks is characteristically worn into smaller roches moutonnées, often well striated, forming rounded oval masses with dimple-like hollows in between. When the general surface is flat, as is frequently the case when shelves are formed from the terminations of spurs, shallow rock-bound pools are formed containing the characteristic bog-vegetation of these regions, which passes into peaty masses. Excellent examples of these can be seen on the platforms at the junction of the Bealey River with the Waimakariri, and on the reduced spur-ends farther up-stream opposite the mouth of the Crow River.

– 53 –


Andrews, C. E., 1905. Some Interesting Facts concerning the Glaciation of Southwestern New Zealand, Rep. Austr. Assoc. Adv. Sci., vol. 10, pp. 189–205.

Davis, W. M., 1900. Glacial Erosion in France, Switzerland, and Norway, Proc. Bost. Soc. Nat. Hist., vol. 29, No. 14, pp. 273–322.

—– 1905. Glaciation of the Sawatch Range, Colorado, Bull. Mus. Comp. Zool., vol. 49, Geol. Ser., vol. 8, No. 1, pp. 1–11.

—– 1909. Glacial Erosion in North Wales, Quart. Jour. Geol. Soc., vol. 65, pp. 281–350.

—– 1912. American Studies on Glacial Erosion, Compte Rendu du XIme Congres Géologique International, vol. 11, pp. 419–27.

de Martonne, E., 1913. Traité de Géographie Physique.

Gilbert, G. K., 1904. Alaska, vol. 111, Glaciers and Glaciation.

Speight, R., 1908. Notes on some of the New Zealand Glaciers in the District of Canterbury, Rep. Austr. Assoc. Adv. Sci., vol. 11, pp. 285–87.

—– 1911. The Mount Arrowsmith District, Part I, Physiography, Trans. N.Z. Inst., vol. 43, pp. 317–42.