The Geology of the West Coast from Abut Head to Milford Sound.
Part 2. Glaciation.
[Read before the Wellington Branch, October 9, 1941; received by the Editor, August 1, 1942; published separately, December, 1942.]
The Pleistocene geology is discussed with special regard to the composition and shape of the moraines, which extend from the foot of the Alps to the sea. A marine Pleistocene deposit is described, and a map presented showing the correspondence between the glaciated alpine valleys, when freed from Recent gravels, and the Southland Fiords.
Every geologist who has visited South Westland has been impressed by the height and extent of the moraines. Haast stated (1879, p. 93) that “if anything will give the geologist an insight into the power which glaciers have of destroying gigantic mountains, and of carrying their debris away into lower regions, a journey to that part of the West Coast will easily effect this object.” The moraines were first described in 1868 by Hector, who published an account of the coast and harbours between Milford Sound and Hokitika. He gave a rough account of the rocks, recognized that the ridges and hummocks along the coast between Bruce Bay and Abut Head were morainic accumulations, and concluded that their rounded form was due to their having been submerged since formation.
In 1877 MacFarlane described the valleys of the Jackson and Cascade Rivers in South Westland and remarked on the thick bed of conglomerate that forms the Cascade Plateau; but he was unable to explain its origin.
The greater part of South Westland was described in a report on the Westland district presented by Cox in 1877. The direct relation between the morainic accumulations and the higher part of the Alps was remarked on in this report (p. 85) where he stated that the glacial drifts “occur largely in a district radiating from Mount Cook, and form all the bluffs from the Mahitahi north, having evidently in former times been the lateral moraines of glaciers which have long since retreated.”
Haast's detailed description of the coastal moraines was published in 1879. This masterly account contains a wealth of interesting and valuable detail together with many important generalizations. Further north, in areas actually outside that discussed in this paper, Bell and Fraser (1906) and Morgan (1908) mapped and discussed the northern extension of the South Westland moraines.
The Cascade River district was mapped and described by Turner (1930), who recognized that the peculiar drainage pattern of this plateau was controlled by successive accumulations of lateral moraines. He maintained that these moraines overlie the conglomerate, which he held to be of late Pliocene age.
The glacial topography is discussed in two parts:—
Topography caused by glacial erosion.
Topography formed by glacial deposition.
1. Topography caused by glacial erosion.
In this paper it is not proposed to discuss the glaciation of the Southern Alps proper, but merely that of the lower land west of the Alpine Fault, the position of which has been described in the first part of this paper. The rounded, beehive-shaped hills standing above the lower land are the most conspicuous glaciated features of this area. They extend for many miles north of the area discussed here and are wholly or partly composed of granitic or gneissic rocks. They were probably shaped by a combination of the following processes:—
(a) Normal weathering of massive granitic rocks.
(b) Differential erosion of an area composed of greywackes and argillites intruded by granites.
(c) Glacial erosion.
It is probable that as a result of normal differential erosion the more resistant parts of these areas of granitic, gneissic, and hornfelsic rocks became isolated from one another and more or less rounded before the advance of the ice, and that when the area was invaded by ice they were further rounded by ice action. If this were the case, then the top of the ice must have been more than 2,000 ft. above present sea-level and the ice must have had considerable erosive powers away from the mouths of the alpine valleys, for Mount McLean and Mosquito Hill are well rounded and over 2,000 ft. high. The hills do not generally lie opposite the mouths of the valleys.
Similar granitic hills occur at Lake Manapouri, and similar but smaller hills lie close to the terminal faces of the Franz Josef and Fox Glaciers; these latter have almost certainly been overridden by ice in comparatively recent times.
Many of these hills have been enumerated in the first part of this paper in the discussion of the igneous rocks, and many of them, particularly in the southern part of Westland, must have formed islands when the ice retreated, but have since been connected with the mainland by the formation of the Westland gravel plain.
2. Topography formed by glacial deposition.
The long smooth ridges which slope at a gentle grade towards the sea (see plate 16) form the most striking feature of the West Coast moraines. These sloping ridges usually extend from the eastern margin of the morainic material, usually close to the scarp of the Alpine Fault, right to the sea, where they end, generally, in cliffs. It is clear that they once extended some distance beyond the present coastline, and that they have been cliffed back by wave attack to their present position. The characteristic shape is well shown by the moraines that enclose the Cook River Flats, which can be viewed to advantage from the Waiho Gorge-Weheka Road a few miles north of Weheka. (See plate 16, fig. 1.)
The characteristic slope is also well shown by the moraines that form the following bluffs (heights approximate):—
|Waiho Bluff||256 ft.|
|Omoeroa Bluff||149 ft.|
|Galway Point||465 ft.|
|Gillespies Point||70 ft.|
|Otorokua Point||81 ft.|
|Malcolms Knob||428 ft.|
|Karangarua Bluff||171 ft.|
|Heritaniwha Point||294 ft.|
|Long Reef Point|
Some of these are illustrated on plates 15 and 16.
The main external and internal features of the moraines are:—
A seaward slope of from 2°–5° presenting an even profile when viewed from a distance, but somewhat irregular when examined in detail, and often with flattish parts rising a few hundred feet above the general level.
The moraines form the interfluves of the rivers and some extend from the bedrock at the scarp of the Alpine Fault to the sea, but more terminate 3–4 miles west of the scarp.
Terraces closely parallel to the upper surface extend for considerable distances along the sides of some of the moraines. These are well developed north of Cook River and on the sides of the Cascade Plateau, where they have been described by Turner (1930).
The moraines decrease in height from a maximum of 1500–2000 ft. at their eastern boundary to 100–400 ft. at the sea coast.
At several places near the coast, namely Mount One One, Omoeroa Bluff, and Malcolms Knob, the moraines instead of maintaining their regular grade rise steeply to form promontories. This may be due to cliffing having extended back to an unusually large rise in the morainic surface.
The cross-sections of the moraines as exposed on the coast show that they are in general convex, with no great area of flat top and sides which differ considerably in steepness.
The layers of material that form the moraine are generally sub-parallel to the surface of the moraine.
This last feature is so general and so peculiar as to deserve special mention. It was first observed by Haast. (1879, p. 394 and plate 8), who described this section of the coast as “a true lateral moraine consisting of the usual detrital matter, blocks of all sizes embedded in sand and silt, the whole having a rough anticlinal arrangement.” On the next page, while discussing Puerua Bluff, he states that this bluff “has at its southern extremity also the rough anticlinal arrangement peculiar to lateral moraines.” The anticlinal arrangement is well shown by the narrow, steep-sided moraines, but in all there is a rough parallelism between the top of the moraine and the bands of till, gravel, sand, and silt of which they are composed.
A knowledge of the origin of this feature is essential to an understanding of the conditions of deposition of the moraines. The deposition of sub-parallel layers of moraine upon an irregular surface by successive ice advances is regarded as a reasonable explanation.
The following alternative explanations were considered and rejected.
Folding of previously level layers by regional compression. Folding upon such a small scale and in such an irregular manner seems improbable, without considerable associated faulting; and no large scale faulting was observed.
Slumping due to melting of sub-adjacent ice. The effects of the melting of sub-adjacent ice were observed both on the moraines of the present-day glaciers and at some horizons of the old moraines. In both cases they were associated with numerous, small, closely spaced faults.
Dumping of morainic material from icebergs upon an irregular surface. No marine organisms or obviously marine gravels were seen in the main body of the moraine, and at Omoeroa Bluff, a peaty layer of definitely terrestrial origin extends a few feet above present sea level.
The highest part of the moraines gives a minimum height for the ice sheet and their gradual slope suggests that the base of the ice sheet sloped seaward at a low angle. The present alluvial-filled river valleys represent areas of thicker ice where little or no moraine was deposited, and are not considered to be formed by post-glacial erosion of the moraines. There is no good evidence for the height of the top of the ice sheet; but Omoeroa Hill (2,237 ft.) and Mount McLean (2,400 ft.), both glaciated, suggest a minimum height of at least 3,000 ft. above present sea-level for the top of the ice near the Alpine Fault. Moraine does not appear to have been deposited under the thicker part of the ice sheet immediately west of the Alpine Fault, but was confined to the thinner, seaward part of the sheet. Lakes and alluvial-filled depressions now mark the area that was once occupied by thick ice.
The method of morainic accumulation under the thinner ice appears more closely comparable with continental glaciation than alpine glaciation, in that the supply of material was smaller and the rate of deposition less.
Haast (1879, p. 394) maintained these moraines to be “true laterals,” an explanation that may be correct for the comparatively narrow moraines on the Cascade Plateau that flank the Cascade Valley (Turner 1930). Such an explanation is unsatisfactory for the enormous body of moraine to the north of the Cook River. The bedding of these moraines shows that they have been formed by successive addition of roughly horizontal layers, not by successive accumulation of lateral moraines.
Similar Morainic Topography in New Zealand.
Topography that appears to be similar to that characterizing the moraines of South Westland has been described from Preservation Inlet district by McKay (1896), Benson (1933), and Benson, Bartrum, and King (1934). Overlying morainic material has been described by McKay (1896, p. 37), who states that the glacial drifts “on Coal Island form fully three-quarters of the area of surface rock, and are also largely developed at Gulches Head Peninsula, and to a lesser extent between Te Whara Beach, the Neck, and Southport. On the long, gentle slope from Treble Mountain they are in part developed to heights corresponding to those at which they are found
on the eastern side of the inlet, while inland of Landing Bay, near Cape Providence, their presence has been reported, and it is possible that they attain to somewhat similar heights to the north of Chalky Inlet, to those reached between Cromarty and Wilson River.” The localities mentioned above by McKay and described as having a varying thickness of morainic covering are figured by Benson, Bartrum, and King (1934, plate 3).
The long, gentle, coastward slope mentioned by McKay (1896, p. 37) and clearly illustrated by Benson, Bartrum, and King's sketches (1934, plate 3) is similar to that of the South Westland moraines. Benson considered this coastward sloping topography as evidence for the presence of a coastal plateau and evidently thought that the moraines had been deposited without much modification of the pre-existing topography. The writers, however, maintain that the strong resemblance between the Preservation Inlet and Westland areas both in topography and glacial deposits is suggestive of a similar, late geological history.
Further north, the Cascade Plateau, which is similar to the more northerly moraines, has been held by Benson, Bartrum, and King (1934, p. 65) and by Benson (1935, p. 399) to be the remnant of an incompletely removed delta that once filled the pre-“Coastal Plateau” valleys, and across which the “Coastal Plateau” was cut. Benson (1935, p. 399) states that this “explanation involves the probability that remnants of incompletely removed deltas might occur high above the mouths of their parent streams. An excellent example of this is afforded by the Cascade Plateau, occurring 20 miles northeast of the limits of Fiordland proper, where a large, gently sloping remnant of a former delta stands beside the mouth of the Cascade River at a height of 1,000 feet or more, quite comparable with that of the coastal plateau.”
Turner (1930), largely because of the induration of the material, concluded that the Cascade Plateau was formed of a veneer of moraine overlying a thick series of late-Pliocene conglomerates. The writers do not agree with the age or origin assigned to these conglomerates by Turner or Benson, for they are very similar to the undoubted moraine exposed on the coast both north and south of the Cascade Plateau and they rest upon smooth and rounded gneisses and hornfels which are well exposed close to the track between the Martyr Bridge and the Cascade Hut. Furthermore, at several places along this track where the overlying material has only recently been removed, well marked, sub-parallel grooves were observed. These were parallel to the adjoining part of the Cascade River, which flows at 10° east of north. It seems almost certain that they are glacier striae cut by the eastern side of the Cascade Glacier which later deposited both the “Conglomerate Series,” and the overlying moraine, and that the Conglomerate Series and the overlying moraine are not separate beds, but that both form part of a thick glacial deposit of about the same age as those described to the north and to the south. The large proportion of serpentine in the Cascade moraine is held to cause the induration of this deposit, for similar induration has been observed with recent serpentine conglomerates at D'Urville Island and North-West Nelson.
In the Martin Bay district, Healy (1938) described a coastal plateau which he correlated with that described by Benson in the Preservation Inlet area to the south and with the Cascade Plateau to the north. Healy (1938, p. 84) maintained that the large morainic boulders distributed over the surface of this coastal plateau are part of a widespread but thin deposit. In this district is Awarua Point, which, when viewed from the south across Big Bay, appears similar in profile to the moraines to the north and has a seaward slope of about 2°. It is convex in cross-section and shows signs of terracing on the south side. Moraine may reach a thickness of 200 ft. on the top of this point, but is represented at sea level only by erratics, which are strewn over the surface of a slightly uplifted bench that the sea has cut in the projecting end of this point. This platform has been cut in the Tertiary rocks upon which the moraine rests further inland. A similar sloping surface has been illustrated by Healy (1938, fig. 4) extending back from Long Reef Point, between Big Bay and Martin Bay, to Sara Hills. Morainic boulders similar to those at Awarua Point veneer a coastal bench for about a mile along the southern side of this point. The surface of both Awarua and Long Reef points is regarded as a part of the coastal plateau by Healy.
It will be seen that similar, seaward-sloping land-forms constitute a conspicuous feature of the topography of the West Coast from Ross southwards to Bruce Bay, from the Cascade Plateau to Milford Sound and in Fiordland south from Resolution Island. They appear to be absent between Bruce Bay and the Cascade Plateau and between Milford Sound and Resolution Island. Their absence between Milford Sound and Resolution Island is probably due to the Alpine Fault, which has caused deep water to lie off the high land to the east. Their absence between Bruce Bay and the Cascade Plateau is more difficult to explain, but is no doubt connected with the almost complete absence of moraines in that district.
Source and Significance of the Dissimilar Morainic Materials.
The proportion of metamorphic rocks in the moraines increases southwards as has been noted by most of the early workers in this area. Some have ventured an explanation. Hacket (1868, p. 9) stated that in the moraines of the Okarito district “micaceous rocks are however very scarce although they are abundantly found in situ a few miles inland at the heads of the rivers. They are of more common occurrence, however, in the moraines near Cooks River, where the mountains approach nearer to the coast.” Cox (1877, p. 85) commented on the composition of the moraines and stated that “they are composed chiefly of Maitai and auriferous slates and sandstones mixed, however, with the metamorphic rocks which are more plentiful the further we go south, and some blocks of which could not weigh less than 10,000 tons.” The same writer (idem, p. 86) also noted the presence of volcanic rocks, which he regarded as having been derived from volcanics at Paringa and which he stated “points to the fact that during the period of greatest extension of the glaciers the great centre of elevation was from a point south of Mount Cook.” Haast (1878, p. 393–4) commented upon the absence of true meta-
morphic rocks in the moraines at Bold Head and Wanganui Bluff, and pointed out that the presence of such rocks in the alluvial deposits that separate the moraine into two distinct parts are evidence of “the retreat of the glacier to such a high position, that the lower slopes of the Alps were exposed to glacial and fluviatile action.” The same writer observed (p. 393) that “towards the south the metamorphic rocks become gradually more numerous.”
An outline of the basement geology is necessary to understand the origin of the moraines. Somewhat similar greywackes and argillites lie both to the east and west of the belt of metamorphic rocks that extends along the western side of the Alps as far south as Jackson River. The western greywackes and argillites, classed by the writers with the Greenland series, have been intruded by granites, and extend as a coastal strip, flanked, however, by Tertiary rocks as far south as Milford Sound. The eastern or Alpine Greywackes extend east of the metamorphic rocks and form part of the west side of the Alps only as far south as the headwaters of the Haast River, from which point south to Jackson River the entire western slopes of the Alps are composed of metamorphic rocks. Consequently the headwaters of rivers as far south as the Haast lie within the Alpine Greywackes, although to decreasing amount as we go south. The snowfields of the present glaciers also lie within the Alpine Greywackes and although the glaciers pass through the metamorphic belt, the greater part of their moraines is composed of Alpine Greywacke as is the gravel in the rivers which drain them.
Between the Waiho and Cook Rivers the moraines can often be divided into two parts, an upper, consisting of predominantly metamorphic rocks, and a lower, of predominantly greywacke. Further south where the western side of the Alps is mostly schist, such a distinction cannot be drawn.
It is probable that the glaciers have always moved away from the mountains and almost directly towards the sea, and that the materials in the moraines have been derived from the mountains to the east. If this is the case, then the change in the composition of the moraines must be a reflection of the composition of the area where glacial erosion was greatest at that time, and the change from Alpine Greywacke to schist suggests that glacial erosion was first concentrated at the eastern greywacke areas, then moved west to the metamorphic belt, and finally as shown by the present glaciers moved back to the greywacke belt. It is not clear how this change took place, but it is difficult to understand unless there has been some elevation of the schist belt.
This schist belt now forms the seaward-facing scarp of the Alpine Fault, and has been elevated several thousand feet relative to the land to the west. A small part of this elevation would serve to account for the changes mentioned above, and evidence for recent movement along this fault-line has been given in the first part of this paper (1942, p. 292). If the schist belt has been elevated during the Pleistocene, then the following sequence of events is not improbable. The glaciers, when they advanced in the pre-Pleistocene valleys, eroded most at their heads well back in the greywacke belt, there being less erosion in the flatter, lower parts of the valleys, the moraine
deposited by these glaciers was then largely composed of greywacke. Movement then took place along the Alpine Fault; and, over the fault-scarp thus formed, the glaciers plunged as ice falls till erosion along the schist belt re-established the grade of the surface of the glaciers. At this time the glaciers carried mostly schistose rocks to the sea. A possible further small movement along this fault may have taken place also during the retreat of the ice; for, although the beds of the present glaciers at the fault-line are close to or below sea level, the valleys do not seem wide enough to accommodate glaciers that could either extend to the sea or deposit the extensive moraines previously described, and it seems likely that the older, larger glaciers must have had wider valleys. These valleys were probably also cut down to sea level at the fault-line, but have since been elevated, and may now be represented by the upper broader parts of the present valleys. This last elevation does not seem to have extended far south of the highest part of the Alps, for the glaciated Karangarua Valley and most of the other valleys to the south are much wider than those of the present glaciers, and show no signs of recent elevation.
The glacial deposits described cover the larger parts of two areas, the more northerly and larger extending north from Blue River beyond the area described to Ross, and the other as a veneer over the coastal parts of western Otago and as a deposit comparable in thickness with the northern area from Sandrock Bluff to the Cascade Plateau. The varved silts, which are often associated with the other morainic material, are described separately in the following accounts of morainic sections.
Hackett (1869) was the first to describe the cliffs between Waitaki and Cook rivers in any detail. He stated (id., p. 10), “in some places the whole mass of the morainic matter has a stratified appearance, which is pretty distinct when viewed from a distance. In other places are stratified beds, of limited extent, but very decided.” Haast later (1879) gave a fairly detailed description of the morainic sections exposed along the coast from Bold Head south to Heritaniwha Point. In his description of the Bold Head section (1879, p. 393), he noted the “existence of an alluvial deposit 30 to 40 feet thick having a considerable slope to the south and separating the morainic beds into two distinct proportions.” The same writer (p. 393) stated that the material making up the moraines at Bold Head is “derived from the Mount Torlesse and Waiho Formations, typical metamorphic or igneous rocks being of rare occurrence.” Morgan (1908) mapped the northern part of the morainic coast, but did not add to Haast's detailed descriptions.
Omoeroa Bluff to Galway Beach.
Omoeroa Bluff forms the seaward extremity of a small moraine lying between the Waiho and Omoeroa Rivers. At the north side of the bluff the following section is exposed at the coast:—
10 ft. of fairly hard till containing rounded to sub-angular cobbles of greywacke with little schist or granite.
25 ft. of sandy varved silts containing rounded and subangular boulders of greywacke with occasional lenses of well rounded gravels.
At the south end of the bluff the exposed section is thicker and shows:
20 ft. of sand and gravel layers interbedded in a hard till. The gravel ranges from 3–6 inches and is practically all greywacke.
6 inches of ancient peaty soil containing lumps of woody material, some of which is changed to lignite and some almost unaltered.
The upper 100 ft. of the bluff lies above the sections described and is composed of coarse angular boulders of greywacke set in a sandy till and partly hidden by vegetation. Numerous, large, sub-angular boulders of greywacke lie on the beach below the bluff. They are the unconsumed remnants of the previous, seaward extension of the moraine. Similar boulders, many of huge size, form small islands off the ends of most of the bluffs, being particularly well developed opposite the highest morainic cliffs, those at Galway Point.
To the south of Omoeroa Bluff moraines are next exposed on the coast at Neds Creek, 7 miles south of Waiho River, and from this point to Otorokua Point, a mile north of Cook River, the moraines are unbroken by large rivers and generally form cliffs along the coast. Between Omoeroa Bluff and Neds Creek the moraines do not reach the sea, but are separated from it by a triangular area of gravel through which the Omoeroa and Waikukupa rivers flow in their lower reaches. The morainic cliffs are not confined to the present coast; for, both north and south of Neds Creek, cliffs extend back behind beach gravels and show that marine cliffing has been followed by marine deposition and retreat of the sea. This has taken place without obvious change of sea level and is due to the sea having rapidly modified the irregularities left by the retreat of the ice. The sea being supplied by the rivers with vast quantity of gravel, soon formed the present mature coast. The high, continuous, morainic cliffs between Neds Creek and Otorokua Point represent deposits of the glaciers which advanced furthest from what is now and was probably then, the highest part of the Alps. Moraines along this part of the coast consist of fine, bedded silts, sand, and gravel together with cobbles and boulders of greywacke and schist. The proportion of schistose rocks is greater than at Omoeroa Bluff and represents over 40 per cent. of the moraine, occasional granite cobbles being less weathered than at Omoeroa Bluff.
The small streams that drain the swampy surface of the moraine discharge over the cliffs as waterfalls with little erosive power, for these falls are covered with vegetation and no fresh rock is exposed. The larger streams have, however, caused some modification of the face of cliffs; but have not, as might be expected, graded courses, but fall vertically from the top of the cliffs down the side of hollow vertical cylinders eroded out of the moraines and enter the sea through comparatively small gaps in the fronts of these cylinders. These features are probably due to the streams eroding an unconsolidated stratum at the base of the cliff and undermining the more compact, cemented material on the surface and front of the cliffs.
Morainic cliffs extend the entire length of the coast from Neds Creek to Gillespies Point, jutting out into the sea at Galway Point.
At Galway Point the following section exposed in the cliff face was noted:—
40 ft. (+) Well-consolidated till, containing cobbles and small boulders of greywacke and schistose rocks. No large boulders. Schistose rocks and greywacke rocks in equal proportions. The cliff above extended another 200 feet and appeared to be of similar material.
20 ft. Grey-white till, containing angular boulders up to four feet of non-schistose greywackes and schistose rocks.
20 ft. Horizontal sandy micaceous layers with a six-inch interbedded layer of well-rounded gravel. Here and there an odd boulder of schistose rocks.
20 ft. Till containing fairly well-rounded cobbles of greywacke, but a marked absence of any large angular boulders. The general appearance of this moraine is much more weathered than that overlying and in particular the occasional granitic cobbles are fairly well decomposed.
Along Galway Beach, which lies between Galway Point and Gillespies Point, the between tide beach at the base of the cliffs is covered with huge, sub-angular, schistose boulders up to 20 feet in diameter. Greywacke boulders are rare along this part of the beach.
Gillespies Point is the seaward projection of the moraines that extend from the Cook Saddle, about 12 miles inland, in a westerly direction toward the coast. (Fig. 2, plate 16.)
The moraines represented at the point are composed of a wellconsolidated till containing rocks largely derived from the greywacke of the Greenland Series, schistose rocks being rarely represented.
At Gillespies Point, the moraine forms a long, curved embankment, having a triangular cross-section which is a striking feature of the topography.
At Otorokua Point, two moraines composed of dissimilar material can be seen in contact. The line of junction between the two moraines dips to the south at 30°, being parallel to the rude bedding of the upper moraine, and truncating the bedding of the lower. Hence it may be regarded as a morainic unconformity.
The Point rises about 80 feet above sea-level and is a south-west extremity of the moraine immediately north of the Cook-Fox River Flats, of which Gillespies Point is a part.
The lower or older moraine at Otorokura Point is made up chiefly of greywacke pebbles and cobbles together with a few boulders of granitic material and a rare volcanic cobble. The boulders range up to 8 feet in diameter and are embedded in well-consolidated till, composed of the fine silts and sub-angular gravels. Where the silts occur, the bedding observed dipped steeply (45°) to the north. The whole moraine is deeply weathered and decomposed, the granitic rocks in particular showing the effects of weathering.
At the base of the overlying or younger moraine is a layer of fine silt with small gravel lenses all dipping parallel to the contact, that is, to the south, at 65°. These silts truncate the bedding of the older moraine at right angles. The silt-band varies in thickness from three inches to three feet, and at the base of the bluff the dip of the silt flattens. Nearly 95 per cent. of the upper moraine is composed
of schistose rocks, greywackes being extremely rare. This moraine is much less weathered and decomposed than the underlying one, all the schist cobbles and boulders being much more angular and less rounded than the boulders of the lower moraine. Where greywacke predominates in the lower moraine, the upper is almost entirely composed of schistose rocks.
Wave attack along the base of the cliff has removed the till and cobbles, leaving the between-tide beach covered with huge greywacke and schistose boulders, the largest observed being about 20 feet by 15 feet. A boulder of Tertiary sandstone was also noted.
The surface slope on the south side of the moraine at Otorokura Point is parallel to the bedding of the upper moraine and to the silts at the line of contact between the two moraines.
On the north side, the surface is parallel to the bedding in the lower moraine. The junction between the two moraines extends from the top of the cliff to the base.
At several points along the Gillespies Beach-Weheka Road the moraines were observed in road cuttings. Nearer the beach the moraines seen were of the older type, being composed largely of greywacke cobbles and boulders; but eastwards and further inland, later moraines of schistose rocks were noted.
The Karangarua Bluff, rising 171 feet above sea-level, immediately to the north of the mouth of the Karangarua River, is the coastal end of a moraine that extends inland in a south-east direction for about six miles and rises in height to 1,424 feet.
The form and composition of the moraine at this point is similar to the others described above, the rocks being largely greywackes with a few igneous cobbles with a rare schist cobble. Haast (1879, p. 395) noted that the Weheka-Karangarua moraine contained some metamorphic and igneous rocks at its northern end. This probably refers to the moraine at Malcolms Knob.
Makawhio Point, formerly known as Jacobs Bluff, forms a coastal projection 250 feet high immediately north of the Makawhio River. The moraine extends from the coast inland to the line of the Alpine Fault, a distance of about six miles, and is separated from the Karangarua moraine by low swampy coastal flats. The till forming this moraine shows a large increase in schistose rocks. In fact, according to Haast (1878, p. 396) “the metamorphic rocks form at Makawhio Bluff the greatest portion of the morainic accumulations.” The same writer (idem, p. 396) states that at Makawhio Bluff an interesting feature is “the occurrence of an ancient river bed about 20 feet thick, deposited against a lateral moraine, covered by a younger morainic accumulation.”
On the south side of Bruce Bay, a moraine projects seaward to form Heretaniwha Point, 294 feet above sea-level. (Fig. 1, plate 15.) This moraine presents the typical semi-circular cross-section and is made up largely of schist cobbles and boulders, greywacke rocks being rather rare. The base of the bluff is fringed with large schist and greywacke boulders, derived from the moraine.
From Heretaniwha Point to Cascade Point, a distance of about 60 miles, moraines are entirely absent, with the exception of Jackson Bay, where the between-tide beach is covered with large morainic boulders, which may have been derived from the erosion of the moraines at Cascade Point and transported round the point into Jackson Bay.
Coastal moraines form a line of cliffs from Seal Rocks (seven miles east of Cascade Point) to Sandrock Bluff with the exception of Barn Bay, where the Cascade alluvial plains reach the coast. The cliffs from Seal Rocks to the mouth of Cascade River represent the coastal end of the moraine that forms the Cascade Plateau. This plateau extends inland for a distance of about 12 miles to the line of the Alpine Fault.
Mount Iota, between Cascade River and Barn Bay, is a morainic hill extending seaward to form Halfway Bluff.
Another morainic mass forms the coastline from the Hope River south to Bluff Creek, immediately south of Sandrock Bluff. This moraine extends inland for about two miles to Steep Head and forms a series of high cliffs.
McFarlane, reporting on the Jackson River-Cascade River district, described the moraine forming the Cascade Plateau as “a heavy conglomerate showing very complete stratification, having a slight dip to the north-west” (McFarlane, 1877, p. 30). Haast (1879) did not record the presence of any moraines there, but mapped the rocks of the Cascade Plateau as Tertiary.
Charles E. Douglas, in an unpublished and undated map of Westland, shows the Cascade Plateau as being formed of morainic drift, resting directly upon the Maitai Slates. This map also shows a similar plateau to the east of Mount Malcolm, between the Jerry and Gorge rivers, which is mapped as having a cover of morainic material.
Turner (1930) was the first to record the presence of moraine in this area, but he maintained that “the morainic material which covers most of the plateau is actually a relatively thin cover, beneath which lie at least 1,000 feet of rocks belonging to the Conglomerate Series.”
For reasons stated in a previous section of this paper, the authors maintain that the morainic accumulations on the Cascade Plateau include what Turner (1930) regards as his Pliocene Conglomerate Series.
As the coastal section at Cascade Point and Halfway Bluff was not examined, the writers are unable to give any descriptions. Further south, however, they examined the coastal section from Barn Bay to Sandrock Bluff.
McFarlane (1877, p. 30) stated “at Teer's Creek, on the coastline, at high-water mark, the underlying rock is exposed consisting of a blue clay, passing into rock,” while Haast's map (1879) shows the Tertiary rocks forming a belt along the coastline from Jackson Bay south to Big Bay. Turner (1933) mapped the basement rocks in this area in the Oligoclase zone of the Maniototo metamorphic series. McFarlane (idem, p. 28) shows in his cross-sections along the
Cascade Plateau the conglomerates at Cascade Point resting directly upon clay passing into slate. This suggests that the Greenland greywackes may underlie the moraines at this point and that the Middle-Tertiary rocks mapped by the writers (see plate 48, fig. 1, p. 302, 1942) are confined to a belt about two miles wide and were eroded away on the west prior to the deposition of the moraines. The writers regard the presence of Greenland rocks below the moraines at Cascade Point as highly probable.
The moraine of Cascade Point Plateau area is composed largely of cobbles and boulders of ultra-basic rocks, but boulders of schist were occasionally noted. The cobbles are fairly well rounded and the whole is extremely well cemented with a light blue silt. It is so strongly cemented that, at several places on the Martyr River-Cascade track, outcrops of the morainic material are actually overhanging. Turner (1930, p. 527) gave a further description of these deposits and commented on the degree of cementation and the composition of the constituent cobbles. The same writer (idem, p. 528) noted that in the section exposed in Teer Creek the beds dip to the south-east at 5°.
Sandrock Bluff rises to a height of about 120 feet, and is the most southern point in Westland where thick morainic accumulations reach the sea.
The section exposed shows the following features:—
80 ft. Varved sandy silts, containing lenses of coarse gravel, with occasional schist or ultra-basic angular cobblcs.
25 ft. Varved sandy silts, showing contortion and minor faulting. Some lenses of fine gravels and occasional cobbles in the varves.
15 ft. Well-cemented till, containing subangular cobbles of schist and ultra-basic rocks set in a fine-grained grey silt.
The varved silts dip to the south at about 10° and many show well-preserved ripple marks when split along the bedding planes. In some places they exhibit peculiar contortions of the varves in a series of small folds. Each fold was faulted. The beach along the cliff base is covered with huge angular masses of consolidated moraine, and even varved silts are sufficiently cemented to withstand pounding by the waves for a considerable time without disintegrating. There are only a few large boulders of schist or ultra-basic rock. From Barn Bay to Sandrock Bluff, the beach is covered with rounded boulders that have evidently been derived from moraines at the back of these beaches. The majority of these boulders are ultra-basic boulders.
The beach extending south from Madagascar Beach to Yates Point is strewn with boulders up to 8 feet in diameter of igneous and schistose rocks. These must be derived from the covering of moraine that forms a plateau known as the Tableland between the Wolf and John o' Groats rivers. Although this area has not been examined by the writers, it appears to be similar in form and origin to the other morainic-covered, plateau-like areas to the north, such as the Gorge River and Cascade plateaus. The long, gentle, seaward slope of Yates Point, shown by Healy (1938, p. 84, fig. 6) strikingly resembles the long coastward slopes of the moraines to the north.
Varved silts were observed at (1) Pug Creek, (2) on the Paringa Haast track, (3) near Moeraki River, and (4) at Sandrock Bluff. The Pug Creek deposit is the most extensive and, being fossiliferous, is of considerable interest.
Pug Creek Varved Silts.
Pug Creek is a small stream that flows in from the north-west to join Omoeroa River about five miles from the sea. Pug Creek has in very recent time changed its course, probably as the result of capture by a small creek eroding headward in the soft silts. The unusually good exposure of these soft beds is largely due to this geological accident. The silts dip to south-west at 10°.
The following is a generalized section of the beds exposed:—
30 ft. Coarse, angular, morainic material composed of schist and grey-wacke.
30 ft. Varved, sandy silts.
2 ft. Peaty layer, with lignified wood-fragments.
40 ft. Sandy silts passing down into varved, sandy silts.
6 in. Peaty layer.
5 ft. Cross-bedded, marine sands.
60 ft. Fine, varved silts, showing about ten varves per inch.
30 ft. Extremely fine, varved silt, with irregularly distributed fragments of schist up to 6 inches, fossiliferous.
10 ft. Coarse, angular, morainic material, consisting of greywacke boulders up to 2 ft., with smaller, rounded cobbles.
Basement. Greywacke of the Greenland Series.
The fossils are confined to bands in the lower silts, but are not well preserved, having apparently been broken during the compaction of these even-grained beds. The writers are indebted to Mr. C. A. Fleming, of the Geological Survey, for the following determinations and notes.
Three species of marine molluscs are represented, all recent, benthic forms. Their known depth-ranges suggest deposition in water of a minimum depth of about 10 fathoms and a maximum depth of probably no more than 100 fathoms, though the maximum is less definite than the minimum.
1. Chlamys radiatus (Hutton).
The form of this species present is comparable with specimens from deep water off Stewart Island rather than from northern localities, but in the absence of full data from recent bottom-faunas off Westland, this should not be taken to imply any post-glacial southward migration.
Depth-range: 13 to 25 fathoms.
Recorded from Preservation Inlet.
2. Nemocardium pulchellum (Gray).
Depth-range: 10 to 120 fathoms.
Recorded from Wet Jacket Arm (Preservation Inlet), Milford Sound in 100–120 fathoms.
3. Eximiothracia cf. vitrea (Hutton).
Depth range: 2 to 15 fathoms, but probably to a greater depth also—evidence is scanty.
All the species require highly saline conditions, none being able to tolerate any degree of freshening; for example, they do not occur in landlocked sea arms that are receiving large quantities of river
waters. Such forms are not present, for instance, in Auckland Harbour, but bottom conditions in fiords seem to be suitable.
A sample of the lower Pug Creek varved silts was examined by Dr. H. J. Finlay, who prepared the following notes on the foraminiferal fauna.
The sample of somewhat greasy siltstone from Pug Creek washed down with some difficulty owing to the very large amount of fine mica. The residue consisted principally of this, which made flotation of the foraminifera impracticable, and a micro-fauna was obtained only after prolonged tapping on a sloping bench. The best that could be obtained was still a poor fauna, of only nine species, as follows:—
Bolivina probably n.sp. (small, slender, with regular, 45°, sloping chambers projecting in blunt spires, somewhat like the Cretaceous decurrens, Ehrenb.).
*Bulimina aculeata d'Orb.
*Cassidulina cf. carinata Cush.
*Cassidulinoides orientalis (Cush.)
*Nonion aff. labradoricum (Dawson)
*Elphidium advenum Cush.
*Notorotalia n.sp. aff. zelandica Fin.
Anomalina cf. parvumbilia Fin.
*Globigerina sp. (very small, 5 globular chambers)
This is too small and peculiar a fauna to allow much deduction as to the age and affinities. The species marked with an asterisk all persist to Recent times and all the species were very rare with the exception of the Nonion (abundant), Notorotalia (9), and Elphidium (7). The only age remark that can be made with certainty is that the fauna is not older than Pliocene, almost certainly not older than Waitotaran. Its occurrence between glacial moraines would suggest late Pliocene or Pleistocene age, but there are some discrepant features in the actual fauna. It does not, for instance, compare at all well with a fauna from George Sound, a recent fiord condition fauna; this is quite rich, also has a spiked Bolivina (but of the beyrichi type), and has a distinctive Loxostomum kanerianum (Brady) abundant, a notable feature of all our Recent faunas, when contrasted with the Pliocene. The Castlecliff, late Pliocene, has also a rich fauna, but the spiked Bolivina present is as in most later Pliocene faunas, difformis (Will.). A striking difference, certainly due to facies, is the abundance of Miliolids in Castlecliffian and Nukumaruan faunas. They are not actually common in the Waitotaran.
Although the Bolivina is the most striking species in the small fauna, one hesitates to use it too much because of the known likelihood of distinct species of this genus appearing when the facies differ considerably. Yet it cannot be argued that the Pug Creek fauna is different solely because of the icy water conditions; cold water faunas are usually rich, especially in arenaceous species, and the present assemblage does not in the least resemble the communities recorded by Heron-Allen and Earland from Antarctic areas. But it is not essentially different from the 30–40 fathom faunas at present met with, even in the Hauraki Gulf, if one considers that some inhibiting factor prevented the appearance of the Miliolidae, Textulariidae,
Lagenidae, etc., and left only the hardier, fairly shallow-water forms. Similar, very poor faunas with overwhelming abundance of the same Nonion have been seen in the Urenuian and Opoitian, where the sediments were also very micaceous and sometimes carbonaceous, indicating poor conditions for rhizopod life, so that the question of facies must loom large over the whole matter.
A probably more important species than the Bolivina is Anomalina parvumbilia. This has a wide range of habitat, and is usually a common species from the Hutchinsonian onwards, but has not been observed above the Petane. The Castlecliff and the shallow water Kai-Iwi faunas lack it in all cases seen, rare species are present in the upper and lower Nukumaruan, and it is not so common till the Waitotaran. An extraordinarily rich and varied fauna (300 species) from Martinborough contains practically all the Petane species from a variety of habitats, including some 35 Bolivinas, but the Pug Creek Bolivina is not there.
Until a larger fauna is obtained, or a match is found elsewhere for the present fauna, it would be hazardous to carry the discussion any further.
Condition of Deposition.
The fossiliferous beds were deposited after an advance of the ice had deposited a predominantly greywacke moraine upon the greywacke basement. This material was either deposited below sea-level, or the land was lowered relative to the sea before the deposition of the 170 feet of fine bedded marine silts. This fine bedded material may represent the infilling of the sheltered landward end of a fiord that was supplied by the finer products of glacial erosion, the upper peaty layers being formed close to sea-level when the infilling was almost completed.
The angular fragments of various types of schist unassociated with other coarse material which are scattered through the fossiliferous beds, are strongly suggestive of materials dropped during the melting of floating ice, and indicate that during this period the ice margin was not far from the sea.
Conditions of deposition are abruptly changed in the upper 30 feet of the section, which is composed of schistose morainic material similar to that at Otorokua Point. This suggests that during or after the deposition of the fine beds glacial erosion had shifted from grey-wackes, probably those east of the schist belt, to the schist. This change was possibly caused by elevation along the Alpine Fault, exposing the schist belt to rapid glacial erosion.
Varved Silts at Moeraki River.
Varved silts associated with morainic till extend along the Paringa-Haast track immediately north of Moeraki River. They are similar to those at Pug Creek, but are not fossiliferous. They dip to the south-west at 15°. This dip is likely to have been caused by movement along the Alpine Fault, for it is too regular to be a depositional feature.
Varved Silts at Sandrock Bluff.
These beds have been described with the rest of the section at Sandrock Bluff. They are more indurated than any further north, a feature that may be caused by the presence of serpentine rock flour.
Fig. 1—Heretaniwha Point, view from north side of Bruce Bay, a seaward sloping moraine on the south side of the Mahitahi River.
Fig. 2—Glaciated granitic knobs near Wataroa, South Westland.
Fig. 3—Longridge Point, view north from the mouth of the Hacket River, a seaward sloping moraine extending from Mount Malcolm.
Fig. 4—Awarua Point, view from south side of Big Bay, another seaward sloping morainic covered point.
Fig. 1—View west from the Cook Saddle on the Waiho-Weheka Road, showing the Cook River flats flanked by seaward sloping moraines.
Fig. 2—Gillespies Point, view from Sandfiv Beach showing the long seaward sloping moraine.
Fig. 3—View north along coast line from Bruce Bay, showing a series of seaward sloping moraines and morainic cliffs, Malcolms Knob at extreme left.
Fig. 4—Malcolms Knob, cliffed seaward end of a moraine that flanks the southern side of the Cook River Flats.
The Relation between the West Coast Valleys and the Fiords of Southland.
The Southland fiords, Lake McKerrow, and the main valleys of the South Westland rivers are glaciated valleys that differ in original depth and in the amount of post-glacial alluviation. They may be considered as representing stages in post-glacial evolution. The way in which this change is taking place is well shown by Lake McKerrow, which has reached a stage intermediate between the fiords and the South Westland valleys.
The hard gneisses and hornfels on the east side of the lake bear ample evidence of the glacial erosion to which they have been subjected, ice scratched and rounded surfaces being common close to the edge of the lake. The sides plunge steeply below the waters of the lake without change of slope and deltas formed by side streams are small.
Alluviation is taking place at both ends of this lake, at the upper end by delta growth and at the lower end by widening of a bar that has formed close to the seaward end of the old fiord. This bar now has a width of about two miles and is largely composed on its inner side of outwash fans from the sides of the valley. The lake is now so far above the sea that tidal effect is much reduced, but the water is still brackish and the marine fauna not entirely replaced by fresh.
Mr. C. A. Fleming, of the Geological Survey, has kindly prepared the following notes on mollusca collected from the delta at the upper end of Lake McKerrow by Mr. J. Healy in 1937 and on one species found on the beaches at the lower end of the lake by the writers. It is likely that all these species lived in the comparatively recent past when there was connection with the sea.
A. From silts at head of Lake McKerrow, about 10 feet above the present lake level:—
1. Chione (Austrovenus) stutchburyi (Gray) (collected by J. Healy).
The specimens are moderately large and have fairly thick shells, suggesting a fairly high salinity, for in brackish, acid water conditions near the limit of tolerance for the species the shells are far thinner and stunted.
Depth range: Half-tide to 2 fathoms, tolerant of low salinity.
2. Mactra cf. rudis (Hutton).
Found in similar conditions to the above and also with a wide salinity tolerance.
B. Near the mouth of Lake McKerrow, at mouth of Hokuri Creek, found on shingle beaches, evidently derived (collected by the authors):—
Chlamys sp. probably zelandiae (Gray).
Found in marine waters, but never in estuarine conditions of very decreased salinity; that is, less tolerant than the above.
Depth range: Low tide to 25 fathoms.
Te Whanga Lagoon, Chatham Island, receives little increment of fresh water streams and is under tidal influence; here both species of A, above, just manage to exist. With the greater inflow of water from the Hollyford River into Lake McKerrow, similarly influenced by tides, it is unlikely that these species could exist at the present time; certainly the Chlamys could not do so.
If the chief difference between the Southland fiords and the glaciated South Westland valleys is in the amount of post-glacial debris that has accumulated in them, then they have to be considered together when discussing their method of formation. Advocates of the glacial origin of fiords are, in Gregory's opinion (1913, p. 361–2), faced with the difficulty of explaining the absence of fiords on the West Coast in the Mount Cook area where the maximum development of the glaciers took place. Gregory (1913, p. 361) states that “the west coast near the chief glacial centre is straight and unbroken instead of being fiord indented.” This straight and now almost mature coast is largely due to post-glacial alluviation having filled the indentations; and if the recent material were removed the coast would be indented as much as that of Fiordland. An attempt has been made to show the form of this coast shortly after the retreat of the ice by a text map, in which the sea is shown as extending into those areas where recent alluvium now extends below sea-level. As the floors of the main valleys do not usually rise far above sea-level and as the sides are only less steep than those of the fiords, little alluvium has accumulated above sea-level and the floors of the valleys almost represent the position previously occupied by the sea.
Independent evidence for fiord conditions during glaciation is provided by the fossiliferous, varved silts at Pug Creek, which is almost opposite the highest part of the Alps. These fine-bedded silts could not have been deposited in other than sheltered water and the
fossils they contain show that the water was, like that of the fiords, of normal salinity. It is almost certain that the lower part of Omoeroa Valley was then an arm of the sea.
It is likely that both the fiords and the glaciated valleys of South Westland were freed from ice at about the same time; consequently, the difference in amount of post-glacial alluviation is not due to the difference in time which has elapsed since the ice retreated from these areas but must depend upon other factors. The chief factors to be considered are:—
(a) The volume of the valleys below sea-level when freed from ice;
(b) The rate at which material was transported to these valleys.
It is not possible directly to compare the volumes of the sea-filled fiords with that of the alluvium-filled Westland valleys; but although the depth may be somewhat less, the width of the valleys is about the same, so the volumes will be of the same order. There are, however, much greater differences in the quantities of material now being carried down by the rivers than there is in the volumes of the valleys in the two areas. For this there are two reasons; the first is the difference in the rocks being eroded and the second is the difference in relief. In western Southland, granite and other material, almost as resistant to erosion, form the mass of the rocks while in South Westland much less resistant schist forms the headward part of the valleys. The difference in relief is also considerable, for in South Westland the main rivers drain areas which rise to 7,000 or 8,000 feet, while in western Southland the maximum height is little over 6,000 feet. This height difference becomes more significant when we consider the area above timber line where erosion is much more rapid, this area must be ten times as large in South Westland as in western Southland.
Early Pleistocene Deposits.
Early Pleistocene beach deposits are confined to narrow, elevated beaches close to the coast. The most southern of these deposits extends from Sardine Terrace, just north of the Haast River, towards Arnott Point. Another area of these deposits is almost continuous from Abbey Rocks to the mouth of Paringa River. They are again represented at a greater elevation at Sandfly Creek and Cement Hill, both near Omoeroa River. The following relations indicate that the deposits are older than the greatest advance of the ice.
(a) The deposit at Sandfly Creek is overlain by morainic material.
(b) Elevated coastal benches are confined to non-glaciated areas (the problem of the “Coastal Plateau” in southern Fiordland is discussed earlier) and none were observed on any of the moraines, although this material is nearly as resistant to erosion as the Tertiary rocks in which the Pliocene benches are cut.
Both the Sardine Terrace deposit and that at Cement Hill have been worked for the gold contained in the lenses of magnetite interbedded with the sand and gravel, consequently, these deposits can now be easily examined. In both cases the material of the deposits is identical with present-day beach deposits, with the exception that some of the magnetite has been oxidised and cemented the surround-
ing sand. This process has so well cemented the sand grains that much of the gold cannot be freed until the black sand is roasted.
The Cement Hill deposit is situated about 400 feet above sea-level and about a mile east of the junction of Gibbs Creek and Omoeroa River. The surface of the deposit is level; and the concentrations of magnetite extend across this flat parallel to the present coastline, being bounded to the north-east by the valley of Gibbs Creek and to the south-west by the valley of Omoeroa River. The seaward-facing front of the hill slopes steeply into Gibbs Creek and on the opposite side the morainic material rises steeply from the edge of the flat. In line with this deposit, on the south side of the Omoeroa River is similar material, well exposed at the head of Sandfly Creek. There, marine sands are interbedded with silts and the whole is overlain by a considerable thickness of moraine. In spite of the fact that these deposits appear to be older than the associated morainic material, they are relatively close to the present shore line, and do not show the extreme weathering so characteristic of the Moutere Gravels in the type locality. Consequently, they may be younger than the Pliocene-Pleistocene orogeny postulated in the first part of this paper (Wellman and Willett, p. 305, 1942) as occurring after the deposition of the Moutere Gravels, and may have been formed immediately prior to or during the main advance of the ice.
From the Mahitahi River north to Ross, sand and gravel beaches alternate with morainic bluffs. The morainic bluffs are connected by even, arcuate beaches. The coastal irregularities left on the retreat of the Pleistocene ice have been reduced partly by erosion of the seaward-sloping moraines by wave attack and partly by the infilling of intervening areas lying below sea-level on the retreat of the ice. The infilling of these areas has not been continuous and regular; and in the past it probably alternated with periods of erosion, as it does at the present day. Haast (1879) held that such a process was responsible for the infilling of the areas lying between the morainic ridges. The initial stage appears to have been the formation of a sand-bar well back from the present coast and separated from the moraine at the back by a depression that is at present occupied by a lagoon.
A series of sub-parallel crescent-shaped ridges composed of sand and gravel extend out from these lagoons to the present beach, and the deposits of magnetite and gold probably concentrated during periods of coastal advance and later buried under these ridges have been extensively worked for gold, first by hand and, later, at Gillespies and Okarito, by dredging. The process of heavy-mineral concentration is still continuing and a considerable advance of the sea is often accompanied by the deposition of several inches of black sand with a gold content ranging from nothing to several ounces perton. The gold is probably derived partly from the moraines and partly from the material brought down by the rapid rivers. Concentration is probably more intense on the coastal beaches than in the rivers, for although payable gold has been worked on the coast as far south as Haast River, little payable gold has been found in the rivers south of Wataroa River.
Bell, J. M., and Fraser, C., 1906. The Geology of the Hokitika Sheet, North Westland. New Zealand Geological Survey Bull., no. 1.
Benson, W. N., 1933. The Geology of the Region about Preservation and Chalky Inlets, South-West Fiordland, N.Z. Trans. N.Z. Inst., vol. 63, pp. 393–432.
—– Bartrum, J. A., and King, L., 1934. The Geology of the Region about Preservation and Chalky Inlets, South-West Fiordland, N.Z. Trans. N.Z. Inst., vol. 64, pp. 51–85.
—– 1935. Notes on the Geographical Features of South-Western New Zealand. Geographical Journal, vol. 86, pp. 393–401.
Cox, S. H., 1877 Report on Westland District. Rept. Geol. Explorations, during 1874–76, no. 9, pp. 63–93.
Gregory, J. W., 1913. The Nature and Origin of Fiords. John Murray, London.
Haast, J. von, 1879. Geology of the Provinces of Canterbury and Westland, New Zealand. The Times, Christchurch.
Hacket, J., 1869. Geology of the Okarita District. Rept. Geol. Explorations, no. 5, pp. 8–15.
Healy, J., 1938. The Geology of the Coastal Strip from Big Bay to Professor Creek, North-West Otago. Journal Sci. and Tech., vol. 20, pp. 80B–94B.
Hector, J., 1868. Notes on the Coast and Harbours from Milford Sound to Hokitika. Rept. Geol. Explorations, no. 4, pp. 42–48.
MacFarlane, D., 1877. Notes on the Geology of the Jackson and Cascade Valleys. Rept. Geol. Explorations, no. 10, pp. 27–30.
McKay, A., 1896. The Wilson River and Preservation Inlet, Fiord County, Otago. Parliamentary Paper, Mines Statement, 11C, pp. 31–45.
Morgan, P. G., 1908. Geology of the Mikonui Subdivision. New Zealand Geological Survey Bull., no. 6.
Speight, R., 1940. The Gravels of the Mackenzie Intermont. Trans. Roy. Soc. N.Z., vol. 70, pp. 175–186.
Turner, F. J., 1930. The Physiographic Features of the Lower Cascade Valley and the Cascade Plateau, South Westland. Trans. N.Z. Inst., vol. 61, pp. 524–535.
Wellman, H. W., and Willett, R. W., 1942. The Geology of the West Coast from Abut Head to Milford Sound, Part 1. Trans. Roy. Soc. N.Z., vol. 71, pp. 282–306.
Dr. C. A. Cotton has drawn the authors' attention to an omission, in Part 1 of this paper, of reference to R. Speight (1910) in which the alpine drainage is considered to have developed upon the surface of an arched alpine peneplain.
Speight, R., Cockayne, L., and Laing, R. M., 1910. The Mount Arrowsmith District; a Study in Physiography and Plant Ecology. Trans. N.Z. Inst., vol. 43, pp. 315–378.