Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 61, 1930
This text is also available in PDF
(3 MB) Opens in new window
– 524 –

Physiographic Features of the Lower Cascade Valley and the Cascade Plateau, South Westland.

[Read before the Otago Institute, 8th July, 1930; Received by Editor, 23rd July, 1930; issued separately, 25th November, 1930].

Plates 81-86.

Contents.
  • Outline of Topography.

  • The Conglomerate Series.

  • Physiographic Features of the Cascade Plateau.

  • Origin of the Cascade Plateau.

  • Glaciation in the Cascade, Martyr and Jackson Valleys.

  • Conclusion.

  • List of Literature.

Outline of Topography.

The area under discussion is a belt of rugged hill and mountain country in southern Westland, lying between the Arawata, Jackson and Cascade Rivers, and bordered on the north by the sea coast. It has twice been visited by the writer during the last two summers, when the following observations were made.

Of the three rivers mentioned above, the Arawata is the largest, swiftest and most formidable. It heads against the Main Divide among the glaciers and ice-fields of the Barrier Range, and the high country in the vicinity of Mt. Aspiring. Its tributary; the Jackson, rises at the north-eastern end of the Olivine Range on the slopes of Mt. Collyer, and, for the greater part of its extent, follows a rectilinear north-easterly course to the Arawata, which it joins about five miles in from the sea.

The Cascade is also a large, swift river, but normally does not carry as much water as the Arawata, since its source lies west of the Main Divide in the vicinity of Red Mountain, where the permanent ice-fields are not so extensive as further east. The river flows through a series of steep-walled gorges with a well marked, general north-easterly trend, which is especially pronounced for a distance of eight or ten miles below McKay Creek, where it continues the north-easterly line of the Jackson Valley. Twelve miles above its mouth the river emerges from the gorge between the Olivine and Hope-Blue River Range and, swinging abruptly through a great bend of 90°, follows a meandering north-westerly course down a wide, alluviated valley to the sea (Pl. 81, Fig. 3; Pl. 82, Fig. 4). Its three tributaries—Martyr, Woodhen and McKay Creeks—occupy deep and precipitous gorges which dissect the western slopes of the Olivine Range (Pl. 81, Fig. 2; Pl. 82, Fig. 5).

– 525 –

This latter is the chief mountain range of the district. It branches in a north-easterly direction from the north end of the Humboldt and the south end of the Barrier Range, and for many miles constitutes the watershed between the Cascade and Arawata Rivers. At its southern end the peaks attain a height of 7,000 ft., but further north-east the range seldom rises much above 5,500 ft. The south-eastern sides and summit of the Olivine Range consist of chlorite-epidote-albite-schist and quartz-muscovite-schist similar to the

Picture icon

Fig. 1. Map of the middle and lower portions of the Cascade and Arawata Valleys.

schists of Central Otago, but its western slopes are carved in the great intrusive mass of peridotite which extends from the Jackson River in the north to Red Mountain and the Red Hill Range in north-west Southland (Turner, 1930).

The Hope-Blue River Range, which lies west of the middle portion of the Cascade Valley and parallel with the Olivine Range, constitutes a belt of densely forested country rising to a little above

– 526 –
Picture icon

Fig. 2.—Block diagram of the Lower Cascade Valley and the Cascade Plateau as seen from the Coast.

– 527 –

4,000 ft., the seaward side of which is drained by the Hope River. The country between the Jackson River and the coast is of a similar nature, and, like the Hope-Blue River Range, consists of gneiss and hornfels invaded by granite-pegmatite (Turner, 1930, p. 174).

Possibly the most striking topographic feature is the Cascade Plateau which borders the Cascade Valley on its northern side for a distance of ten miles in from the sea coast, and extends westward for several miles to Carmichael Creek. This feature will be described and discussed fully in a later section.

The Conglomerate Series.

In a previous paper (Turner, 1930) an account has been given of the metamorphic and ultrabasic rocks of the Lower Cascade Valley. These constitute a basement upon which lies unconformably a great series of bedded clastic rocks—mainly coarse, well-cemented conglomerates—derived from the erosion of the underlying formations. These younger rocks are here grouped together as the Conglomerate Series, the origin of which is closely connected with the early stages in the physiographic development of the Cascade Valley.

Rocks of this type, sometimes containing boulders as much as four feet in diameter, are exposed abundantly along the old track, between the bridge across the gorge of Martyr Creek and the ford where it crosses the same stream, about four miles below. The constituent boulders are imperfectly rounded and are set in a hard cemented matrix of fine blue clay (Pl. 81, Fig. 1). The majority consist of fairly fresh peridotite derived from the rocks of the Olivine Range, though masses of gneiss, schist and hornfels also are present. In this locality the gneissic basement is not far beneath, and outcrops from under the mantle of drift in the gorge of Martyr Creek and at intervals along the track. McFarlane, who, in 1877, first explored the Cascade Valley, took advantage of the creek being unusually low to follow down the Martyr Gorge. He notes (1877, p. 30) that “the river having cut clear through to a depth of 150 feet in places, a fine section of the formation is presented, which consists of a heavy conglomerate showing very complete stratification, having a very slight dip to the north-west.”

Cemented conglomerate was also observed along the southern slopes of Red Spur, where it falls away steeply into the valley of Martyr Creek (Pl. 81, Fig. 2). It is exposed, almost from the level of the creek to a height of over 1200 ft. above sea-level, in the bed and walls of a narrow gorge leading down from Red Spur. Although here it directly overlies peridotite, the component boulders of the conglomerate are mainly schist—probably derived from the tract of schist country in which the Martyr rises. This fact and the highly indurated state of the rock preclude any possibility of its being simply a hillside talus.

The rocks of the Conglomerate Series attain their most extensive development on the north-eastern side of the lower part of the Cascade Valley, where they underlie the whole of the Cascade Plateau, an area of about 20 sq. miles. The surface of the Plateau is itself

– 528 –

largely covered with a mantle of glacial moraines, the origin of which will be outlined later in this paper. A fine section is exposed in the gorge of Teer Creek, which is the only large stream to cross the plateau. It rises in the ranges to the south-east and cuts north-west across the plateau through a profound gorge about 1,000 ft. in depth. This was crossed at two points, respectively three miles and one mile above the mouth.

At the first of these, conglomerate outcrops in perpendicular bluffs at intervals between the stream bed (possibly 200 ft. above sea-level) and a height of about 1,200 ft. or more. The boulders are large and consist mainly of gneiss, hornfels and schist, to the almost complete exclusion of peridotite.

At the second point, one mile above the mouth of the creek, hard blue mudstone is exposed in the stream bed and is overlain upstream by conglomerate, consisting of subangular but partially rounded boulders set in the usual matrix of hard blue clay. Here again peridotite is rare. There is a regular dip upstream (i.e. south-east) at about 5°. On the west side of the creek, at a considerable height above the bed, steep bluffs are cut in massive mudstones about 100 ft. thick, through which run occasional thin bands containing subangular boulders. These pass up into conglomerate, which appears to continue to a height of about 1,000 ft. or 1,100 ft. above sea-level. Immediately west of this, in the bed of the tributary marked A (Text-figure 3), similar conglomerate is again exposed not more than 200 ft. below the general plateau level, i.e. between 900 ft. and 1,000 ft. above sea-level.

The above observations indicate beyond doubt that the unconsolidated morainic material which covers most of the plateau is actually a relatively thin cover, beneath which lie at least 1,000 ft. of rocks belonging to the Conglomerate Series. Observation of conglomerates in the gorge of Laschelles Creek, and reports regarding exposures in the sea cliffs along the northern margin of the plateau lend support to this conclusion.

Cox (1877, pp. 94, 95) has described strongly folded Tertiary sedimentary rocks, lithologically very different from the Conglomerate Series, from the vicinity of the old settlement at Jackson's Bay, about five miles east of the edge of the Cascade Plateau. It may safely be assumed that the contorted state of these strata is due to the great earth movements of the Pliocene which Cotton has termed the Kaikoura deformation. It follows, then, that the almost undisturbed strata of the Conglomerate Series were most probably laid down subsequently to this movement, and their age cannot in that case be earlier than late Pliocene.

On the other hand, the well cemented nature of the conglomerates and mudstones, their regular stratification, and the slight inclination of the strata (north-west in the gorge of Martyr Creek and south-east in Teer Creek) all point to a pre-Pleistocene age. Morgan (1928) and Marwick (1928) both regard the Pleistocene of New Zealand as a period of extensive glaciation comparable and contemporaneous with the European Pleistocene, and uphold the principle that, in the rocks of the South Island, the Pleistocene-Pliocene boundary should be

Picture icon

Fig. 1.—Conglomerate, ¼ ml. west of Martyr Bridge, Cascade Valley. [G. J. Williams photo
Fig. 2.—Red Spur and the valley of Martyr Creek, seen from the crest of Martyr Spur. The light coloured area, bare of vegetation, is peridotite, while the forested area to the right is underlain by schist. [J. S. Thomson photo.

Picture icon

Fig. 3.—The Cascade Valley and the Olivine Range, looking south from the Cascade Plateau (1900 ft.). On the right, in the middle distance, is the mouth of the Cascade Gorge, and on the left, the valley of Martyr Creek. [J. A. Bartrum photo.

Picture icon

Fig. 4.—The lower portion of the Cascade Valley, seen from the end of Martyr Spur (500 ft), looking Seawards The Cascade Plateau forms the sky-line on the right and in the middle of the photograph, while on the left are seen the lower spurs of the Hope-Blue River Range. [G. J. Williams photo.

Picture icon

Fig 5—Mt. Richards and the gorge of Woodhen Creek, seen from the peridotite-schist junction on the crest of Martyr Spui (4,000 ft) Note the bare peridotite in contrast with the bush-covered schist on the left The snow-covered peaks beyond Mt Richards are the summit of the Olivine Range [G. J. Williams photo

Picture icon

Fig. 6—The surface of the Cascade Plateau looking south to Twin Blocks Trig Nearly all the boulders are peridotite [J. A. Bartrum photo.
Fig. 7.—Glacier-borne blocks of schist on the surface of the Cascade Plateau, near the source of Duncan Creek [G. J. Williams photo.

Picture icon

Fig. 8—Eastern boundary of the Cascade Plateau near the head of Teer Creek, looking northward On the right is the spm leading up to Colin Hill, while the gorge of Teer Creek cuts-across the middle of the photograph from right to left [J A. Bartrum photo.

Picture icon

Fig 9—The eastern boundary of the Cascade Plateau, with Colin Hill rising beyond to form the skyline. At this point the upper portion of Teer Creek separates the bare moraines of the Plateau. from the bush-covered hornfels country, of which Colin Hill is the highest point [J A Bartrum photo.

Picture icon

Fig. 10.—Regular marginal moraine, forming a sharp ridge 1 ½ mis. long, ¾ ml. north-west of Twin Blocks Trig., summit of Cascade Plateau. Note the meandering course of the stream. Photograph taken looking north-west [J. A. Bartrum photo.

Picture icon

Fig 11—Regular morames on the north-eastern side of the valley shown in Fig 10, summit of Cascade Plateau The almost horizontal line AB is the erest of the ridge marked × in Text-Fig. 3 Beyond it lies another inter-morainic valley, the drainage from which reaches the main valley through an incipient break in the ridge AB, just to the left of A. Note the difference between the height of AB and that of the ridge CD beyond. [J. A. Bartrum photo

Picture icon

Fig 12—Surface of moraine at Twin Blocks Trig Station, summit of Cascade Plateau. The two large blocks are hornfels, the smaller boulders in the foreground being peridotite [J. A. Bartrum photo

Picture icon

Fig 13.—Junction of streams B and C (Text-Fig 3), summit of Cascade Plateau, looking north-east The line × Y Z is the erest of a regular moraine, beyond which lies the deep gorge of Teer Creek The streams in the foreground reach the latter through the gap Y The even surface of the Plateau east of the gorge constitutes the sky-line [J. A. Bartrum photo.

Picture icon

Fig 14—Southern end of the Cascade Plateau, as seen looking north from the Cascade Hut, Cascade Valley On the right is the mouth of Laschelles Creek [J. A. Bartrum photo.

Picture icon

Fig 15—Looking north-west into the valley of the Cascade River, from the margin of the Cascade Plateau, west of Twin Blocks Trig Station The floor of the valley is hidden in mist, but commencing from the point A, the edge of the lateral moraine which constitutes the teriace at 1,350 it above sea-level, is clearly visible, merging gradually into the surface of the plateau in the distance. [J. A Bartrum photo

– 529 –

drawn at the base of the earliest deposits of glacial origin. There is no reason to suppose that the Conglomerate Series of the present area represents outwash from Early Pleistocene glaciers. On the contrary, as will be shown later, two series of moraines are definitely known to post-date the deposition and uplift of these rocks.

The Conglomerate Series is therefore regarded as being of late Pliocene age.

From the foregoing description it will be seen that the rocks of this series are limited to areas bordering the Cascade and Martyr Valley, where they occur between sea-level and a height of about 1,200 ft., and are entirely absent from the higher slopes of the adjacent ranges. It follows, therefore, either that the conglomerates originally had a wider distribution, and owe their present limited extent to preservation from erosion in down-faulted areas, or else that they were deposited in a late Pliocene valley, the site of which is still occupied by the Lower Cascade and Martyr Rivers. The field evidence strongly supports the latter view. The writer, therefore, suggests that, towards the close of the Pliocene, a wide triangular depression, probably the result of erosion along lines determined by faulting in the Kaikoura Orogeny, extended across the area which to-day is occupied by the Cascade Plateau and the lower portion of the Cascade Valley. This depression narrowed inland, and continued southward some distance beyond the present great bend in the Cascade River, along the line of what is now the valley of Martyr Creek. As a result of long-continued slow sinking of the land, alluvial gravels, and to a less extent finer sediments, together with pluvial and talus debris, accumulated continuously at the foot of the ranges.

In this way a thickness of over 1,000 ft. of strata was built up. When eventually this phase of slow subsidence and accompanying deposition of gravels came to a close, the land surface must have consisted of dissected mountain-ranges bordering a broad infilled depression, the surface of which sloped gently seaward to sea-level at the coast. The present elevation of the remnants of this ancient surface indicates that at this time the land stood considerably lower (possibly 1,000 ft.) than to-day.

Physiographical Features of the Cascade Plateau.

Between the mouth of the Laschelles Creek and the sea coast the north-eastern wall of the Cascade Valley rises steeply from only a few feet above sea-level to the summit of the Cascade Plateau (Pl. 86, Fig. 14). This is an extensive triangular tableland, about 20 sq. miles in area, which stretches between the Cascade Valley and the eastern side of the Carmichael Creek. At its highest point—Twin Blocks Trig. Station, just above the gorge of Laschelles Creek—it reaches an elevation of 1,900 ft. above sea-level, and thence slopes gently northward to a height of between 800 ft. and 1,100 ft., where it terminates in lofty and precipitous cliffs along the sea coast. This seaward slope is very noticeable as seen from the floor of the Cascade Valley or the lower spurs of the Olivine Range, whence the plateau presents a remarkably even profile in striking contrast with the

– 530 –

ragged sky-line of the adjacent mountain ranges (Pl. 82, Fig. 4). On its eastern border the regular surface of the plateau terminates abruptly against the steep, heavily-bushed slopes of the spurs leading up to Colin Hill and Mt. Alpha (Pl. 83, Fig. 8).

As shown in the previous section, the Cascade Plateau is everywhere underlain by a great thickness of late Pliocene conglomerates. The surface, however, is covered—probably to a depth of over 300 ft. in some places—with unconsolidated moraine. It is littered with large boulders (Pl. 82, Fig. 6) which average about 3 ft. to 4 ft. in diameter, but which frequently are much larger (Pl. 85, Fig. 12) and may even attain 20 ft. or more in average dimension. The majority consist of fresh dunite, wehrlite and harzburgite brought down from the Olivine Range (Turner, 1930, p. 190), but hornfels and schist are also represented (Pl. 82, Fig. 7). As on the peridotite belt itself, the surface is bare of vegetation, except for tussock grasses, rushes, and patches of low scrub, but in the gorge of Teer Creek, where the underlying conglomerates are exposed, the steep slopes are heavily bushed.

Though the plateau appears from below to be regular in the extreme, in reality this is by no means the case; for its surface, upon actual examination, is seen to have a complex and strikingly unusual drainage system, the main details of which are sketched in the accompanying map (Text-Fig. 3). In the north-western corner the major features only are indicated, since the time available was not sufficient to map the whole plateau in detail.

The most important drainge channel is Teer Creek, which rises in the bush-covered hills to the south-east and cuts north by west across the plateau, through the deep gorge described in an earlier section. On either side, the gorge is flanked by a well-defined terrace about 100 ft. below the general level of the plateau, and this appears to represent a bench, where the cemented rocks of the Conglomerate Series outcrop from under the morainic cover.

All the streams west of Teer Creek, including the tributaries of Teer Creek itself, have several striking peculiarities in common. In the first place their valleys and the intervening ridges show perfect parallel disposition along a distance of from five to ten miles. This trend of the topography, though roughly north-west, in reality varies round the are of a circle, between north, at the southern vertex of the plateau, to almost due west at its north-western corner in the vicinity of Cascade Point. Again, the floors of the valleys are without exception only about 100 ft. or 200 ft. below the general level of the plateau, along the greater part of their extent. The cross-profile is typically V-shaped, though sometimes, as in the case of Creeks B and C, the valley floor may be nearly flat for a width of a quarter of a mile or more (Pl. 85, Fig. 13).

The tributary streamlets are in every case parallel to the major drainage channels; and junction between the two is effected by the tributary turning sharply at right angles to its former course, and cutting through an incision in the intervening ridge. In such cases, it will be noted that the trend of the tributary is usually continued beyond such an outlet, by another stream flowing in the opposite direction (e.g. stream B continuing the trend of stream C).

– 531 –
Picture icon

Fig. 3.—Map showing the main drainage features of the Cascade Plateau. (The course of the stream, which is shown terminated by the arrow-head, was not traced below that point).

– 532 –

Origin of the Cascade Plateau.

It might at first sight be suggested that the regular drainage pattern of the surface of the Cascade Plateau is the result of the establishment of consequent streams, flowing down the slope of an uplifted, slightly tilted plain. There are, however, a number of striking and persistent peculiarities, which cannot be explained on the assumption that we are dealing with normal valleys of erosion. These may be outlined briefly as follows:—

(1) The regular sweep of the drainage trend, from north almost to due west, is not in accordance with the due northerly slope of the plateau surface. (The latter has been verified accurately from the surveyed heights of a number of trig. stations at various points on the plateau).

(2) Though the drainage channels have the typically V-shaped cross-profile of the juvenile valley of erosion, the streams themselves are sluggish, and show on a small scale a perfect system of meanders across a narrow belt only four or five yards in width (Pl. 84, Fig. 10). In some cases even there are swamps and small lakelets at the confluence of two or more such streams.

(3) Though the morainic material comprising the intervening ridges consists of boulders of all sizes, yet there is no concentration of the larger masses towards the valley floors, such as must certainly have taken place if steep-walled valleys had been cut in unconsolidated moraine.

(4) The valleys and ridges are unusually closely spaced.

(5) In some cases the summit of a ridge, though regular and obviously not reduced by erosion, may lie from 20 to 50 ft. below the summits of the adjacent ridges (Pl. 84, Fig. 11).

(6) In almost every case the ridges are covered with boulders of peridotite, amongst which masses of schist and gneiss occur to the extent of only about 5 per cent. Nevertheless a single minor ridge (X, Text-Fig. 3), about 400 yds. in length, was found to consist almost entirely of blocks of foliated schist, similar to the distant schists that crown the Olivine Range. In this case, the streams on either side of the ridge × form sharp lines of separation from the adjacent peridotite-covered ridges on either side.

The only hypothesis which will readily explain the above facts is that the plateau is an elevated plain, the surface of which is covered with an immense series of regular, parallel marginal moraines, the spaces between which now act as drainage channels. These moraines have for the most part been but little eroded, since, with the exception of the deeply entrenched Teer Creek, the intervening streams do not rise beyond the confines of the plateau, and hence have little erosive power. Whereas the general uniformity of the plateau surface is due to the fact that the morines have accumulated upon an ancient uplifted plain, the present curious drainage pattern is due to the disposition of the overlying moraines themselves. The writer, therefore, offers the above suggestion as the solution to the problem.

If this solution is accepted, it is easy to picture the material composing the ridge × in Text-Fig. 3 (Pl. 84, Fig. 11), as a small

– 533 –

moraine brought down to the main ice-sheet by some tributary glacier, heading back into the schists of the Olivine Range. The hummocky topography developed round the point Y (Text-Fig. 3), not far from the source of Dougal Creek, also lends support to the idea that the original moraine topography has not been materially altered by post-Glacial erosion.

On its south-western flank the steep slope of the plateau, where it falls away rapidly into the Cascade Valley 1,800 ft. below, is broken by rather narrow, well-defined terraces at heights of 1,350 ft., 750 ft., 450 ft., and 400 ft. above sea-level. Characteristically each terrace is bounded along its inner margin by a streamlet running parallel to the terrace edge about 30 ft. below the terrace level. The uppermost, as seen from the summit of the plateau, is almost horizontal, and bends regularly north-west along the are of a quarter-circle, until ultimately it intersects the sloping surface of the plateau, and blends into the regular parallel sweep of the surface moraines (Pl. 86, Fig. 15). South of the plateau it is continued as a gently sloping terrace on the flank of the high hills between Laschelles Creek and the Jackson-Martyr Saddle. The above facts, taken in conjunction with the fact that the terraces seem to be made up largely of peridotite boulders, indicate that the terraces are probably of glacial origin, and are to be regarded as the lateral moraines of the ancient Cascade Glacier.

The physiographic evolution of the Cascade Plateau may be summed up briefly as follows:—

At the close of the Pliocene, after the in-filling of the Cascade Depression, elevation of the land, possibly by 800 ft. to 1,000 ft., accompanied by very slight warping, took place. The ancestral Cascade River cut down rapidly through the uplifted rocks of the Conglomerate Series, and became incised in a deep valley, approximately along its present course, flanked on the north-east by an uplifted plain—the Cascade Plateau.

Closely following this came the Pleistocene glaciation. A large glacier flowed down the Cascade Valley, the lower and wider portion of which must have become filled with slow-moving ice. Eventually, as the glacier increased in volume, the ice, which further upstream was hemmed in by lofty mountain walls, appears to have spilled over the valley rim upon the surface of the plateau, across which it pushed out an extensive ice-lobe, which stretched to its eastern and northern boundaries, and so possibly reached the sea coast. During the slow retreat of the ice-front in the later part of the Pleistocene, an immense series of parallel marginal moraines accumulated, marking intermittent periods when the ice neither advanced nor retreated. Further shrinkage of the ice resulted in its withdrawal from the surface of the plateau, though a very large glacier still occupied the valley below. It was at this stage that the lateral moraines, which now form terraces along the north-eastern wall of the Lower Cascade Valley, were deposited by the slowly sinking glacier.

Teer Creek was probably a well-established stream by the time the ice first spread across the plateau, but its gorge has doubtless been deepened considerably since the covering ice sheet withdrew. Otherwise, the only modification of the plateau in Recent times has

– 534 –

been the establishment of the present system of drainage, in which streams now occupy the less elevated areas between adjacent morainic ridges. The latter have been pierced in a number of places to give a more connected stream system—a feature which is well shown in the tributaries on the western side of Teer Creek (Pl. 85, Fig. 13). An incipient outlet of this type, observed at the north-west end of ridge × (Text-Fig. 3), is shown in Pl. 84, Fig. 11.

Glaciation in the Cascade, Martyr and Jackson Valleys.

Corroborative evidence of glaciation is also to be found in the valley of the Cascade River itself. Cirque remnants, often enclosing small tarns, are to be seen at a number of points along the flank of Martyr Spur, high above the gorge of the Cascade, which itself has the U-shaped profile characteristic of glaciated valleys. Again at the north-eastern end of the Hope-Blue River Range, the spurs on either side of the Colin Creek exhibit the steep faces and triangular shape characteristic of ice-shorn spurs. Finally at the base of the steep hill face on the western (inner) side of the great bend in the Cascade River, there is a large patch of bush-covered hummocky moraine with numerous undrained lakelets and ponds interspersed between the hummocks. The accumulation is not more than 100 ft. above sea-level, and was deposited as a terminal moraine just below the entrance of the Cascade Gorge just before the final retreat of the much-diminished glacier as it withdrew from the lower part of the valley. This moraine is therefore considerably younger than those already described from the Cascade Plateau.

Picture icon

Generalised longitudinal section down the valley of Martyr Creek. The shaded portion on the left represents the riegel cut through by the Martyr Gorge, while that on the right shows the incision by which the stream descends the glacial step, which is developed 3 ½ mls. upstream from the bridge.

During the period of maximum glaciation a large cirque was formed at the head of Martyr Creek, from which a small tributary glacier descended through a hanging valley to the main Cascade Glacier, which it joined near the point where the bridge now spans the Martyr Stream. The lip of this valley, which to-day hangs about 500 ft. to 600 ft. above the floor of the Cascade Valley, is defined by a typical riegel, developed in the gneiss which here outcrops from beneath the less resistant rocks of the Conglomerate Series. At the present time the Martyr cuts through this riegel by means of a very narrow vertical-sided canyon—the Martyr Gorge—which varies between 50 ft. and 150 ft. in depth, and extends for a distance of two miles below the bridge (Text-Figs. 1 and 4). Immediately

– 535 –

above the bridge, the valley of the Martyr opens out into a comparatively wide basin, the floor of which is some 50 ft. below the riegel which shuts it in. This basin has been cut partly in the somewhat easily eroded rocks of the Conglomerate Series.

Further upstream, some 3 ½ miles south of the bridge, a well developed glacial step is shown in the valley floor, just above the south-eastern boundary of the peridotite belt. A generalised longitudinal section indicating the above features is given in Text-Figure 4.

The U-shaped cross-profile of the original hanging valley of Martyr Creek, surmounting the notch of the outlet gorge, may still be plainly traced, especially when viewed from some such distant elevated point as the summit of the Cascade Plateau (Pl. 81, Fig. 3).

The long straight valley of the Jackson River appears certainly to owe its present form to glacial erosion, though the rectilinear north-easterly course of the original pre-Glacial valley was doubtless, as explained previously (Turner, 1930), determined by the major fault zone which follows along this line, across the Martyr-Jackson Divide, and up the line of the Cascade Gorge. During the period of glaciation an offshoot from the Martyr Glacier continued along this line of structural weakness, across the present divide between the Jackson and the Martyr Valleys, and down the Jackson Valley to the Arawata. This influx of ice accounts for the unusually wide floor of the present Jackson Valley. The point where the ancient glacier cut across the divide is to-day occupied by a wide low saddle only 500 ft. above sea-level, through which the track from the Jackson crosses over into the valley of Martyr Creek. On each side of the saddle almost vertical walls rise to a height of 2,000 ft. or more.

Conclusion.

In conclusion the writer wishes to extend his sincere thanks to his companions, Professor J. A. Bartrum, and Messrs. W. E. La Roche, G. J. Williams, J. S. Thompson and G. Simpson.

The expenses incurred on the first expedition (January-February, 1929) were largely met by a Government Research Grant obtained through the New Zealand Institute, while a grant was received from the Otago University to defray part of the cost of the 1930 trip. I wish to express my appreciation of the financial help thus afforded by these two bodies.

List of Literature.

Cox, S. H., 1877. Report on Coal Measures at Jackson's Bay, Rept. Geol. Expl., 1874-1876, pp. 94-95.

Macfarlane, D., 1877. Notes on the Geology of the Jackson and Cascade Valleys, Repts. Geol. Expl., 1876, pp. 27-30.

Marwick, J., 1928. Pliocene-Pleistocene Boundary in New Zealand, Proc. 3rd Pan-Pacific Sci. Cong., 1926, vol. 2, pp. 1767-1775.

Morgan, P. G., 1928. The Definition, Classification and Nomenclature of the Quaternary Periods, ibid., p. 1776.

Turner, F. J., 1930. The Metamorphic and Ultrabasic Rocks of the Lower Cascade Valley, South Westland, Trans. N.Z. Inst., vol. 61. pp. 170-201.