The Geology of Banks Peninsula—A Revision.
[Read before the Canterbury Branch, October 6, 1942; received by the Editor, November 11, 1942; issued separately, June, 1943.]
In 1917 I published an account of the Geology of Banks Peninsula (Trans. N.Z. Inst., vol. 49, pp. 365–392), and during the twenty-five years that have elapsed since it appeared, I have examined the area in some detail and am consequently in a position to correct mistakes and to bring the matter more up to date. In doing so I shall not repeat statements from my original account which I think well founded, but merely modify and amplify them where necessary. I may say at the outset that I have seen no reason to change my opinion in regard to the general origin of the present land-forms of the area, but the account of the geological history needs considerable revision. A consideration of the former will be presented first.
The physiography has attracted considerable attention from visiting geologists and from those resident in New Zealand. Notable among the former are W. M. Davis, who gives somewhat detailed reference to it in “The Coral Reef Problem” (1928); and Howel Williams, who deals with “Calderas and Their Origin” (1941), while among the latter C. A. Cotton may certainly be mentioned, for he discusses the problem in “Some Volcanic Land-Forms in New Zealand” (1941), and refers to examples of landscape features furnished by the area in his latest edition of “Geomorphology” (1942).
The most important physiographic problem presented by the area is that which concerns the formation of the great hollows now occupied by Lyttelton and Akaroa harbours. Both Haast and Hutton regarded them as explosion craters, while the present author has attributed them largely to the action of stream and marine erosion. In this he is followed by the authorities just mentioned, who certainly make various contributions to the discussion while giving general support to the solution suggested.
Davis (op. cit. pp. 191–4) points out that the form of the cliffs on the seaward margin suggests that they were chiefly cut when the land was higher, since they descend into deep water without any definite shore platform, though he admits that they have been cut back slightly at about present sea-level. He also points out that the land must have been higher when the valleys were excavated and that owing to the dominance of strong wave action over short stream erosion in the case of small pelagic islands it is probable that at one stage the mouths of the valleys were of the ‘hanging’ type, that the discordance in grade on the coastline has been hidden by the subsidence of the land, and that this subsidence was relatively rapid, otherwise the cliff-cutting on the coast might have removed the ends of the spurs more completely and shortened the lower reaches
of the valleys. He compares the former condition of the cones to that of Egmont, but this volcano is not so maturely eroded as to allow of the formation of deep indentations in its periphery were the lower slopes depressed below sea-level, so that the comparison is only applicable to a very early stage in the history of Banks Peninsula.
It should be noted that the present small, V-shaped valleys on the more exposed stretches of the coastal fringe end discordantly at the cliff margin, their lower reaches, once extended seaward, having been cut back by marine erosion, and Davis suggests that in the case of the larger valleys, whose lower reaches are now occupied by the sea, similar action took place when the land was at a higher level.
The following additional comments are made on certain aspects of the hypothesis as presented by Davis, and first as regards the cliffs.
Over long stretches of the coast near East Head the cliffs rise sheer for 200 feet and more, and there are occasional short stretches north and south of the entrance to Akaroa Harbour, where they reach nearly 600 feet (Plate 3, fig. 1.). In a few cases their lower slopes are masked with heavy rock-falls, but in most cases they descend almost immediately to deep water. For example, at the entrance to Lyttelton Harbour, Godley Head rises directly some
400 feet from a shore platform from half a chain to a chain in width and covered with blocks fallen from the cliffs above, the under-water slopes descending steeply into water now 7 fathoms deep and formerly much more owing to the reduction in depth arising from the deposit of a thick cover of sediment on the floor of the harbour. There are other cases where the cliffs rise from basal reefs, either submerged or with rocks showing above the surface of the sea and extending discontinuously for a quarter of a mile from the base of the cliffs, as, for example, the south head of Akaroa. This is specially noticeable where the ends of the spurs have been attacked from both flanks by the sea, and some of the reefs may be the remnants of spurs not completely destroyed when the land stood at a higher level. Plate 3, fig. 2 shows platforms on the southern side of the peninsula where the cliffs have been cut back by heavy seas attacking the coast from the south. However, in very many cases on the exposed coastline the cliffs descend sheer into deep water.
In his presentation of the problem, Davis cites the finding of peat beds in the bores of the wells of the Christchurch artesian area as a proof that the land was once higher. The most interesting of these wells is that sunk at the old racecourse at the mouth of Heathcote Valley, and within the line of the periphery of the spurends in that sector of the volcanic cone. The well produced little water and was abandoned. With the exception of one thin layer of gravel, the usual repository of the water supply, the beds encountered consisted of a succession of sands, sandy clays, and clay beds, some containing shells, and no doubt marine or aestuarine in origin. The special point of interest is that solid rock was reached at a depth of 708 feet (Speight, 1911, p. 429).
This seems to indicate that the former Heathcote Valley was eroded to that depth, and that the land once stood over 700 feet higher. This is the deepest bore near the base of the hills, but rock has been encountered at shallower depths in or just off the mouths of valleys radiating from the Lyttelton volcano, for example, in Sumner, Hoonhay, and Gebbies Valleys. Although no specimen of the rock is available for examination, it is reasonably certain that it is basic in character. In a well sunk in the Sumner Estuary off Fisherman's Flat, near Redcliffs, and 20 chains away from the base of the cliffs, peat was struck- at the 366 feet level and solid rock at 416 feet (Speight, op. cit. p. 428 and Plate 12), both occurrences indicating a former land surface now far below sea-level. The nature of the rock surface between the bore and the base of the cliffs cannot be stated definitely, but wells 10 chains from the cliffs reached a depth of 180 feet without meeting rock, so it is probable that the surface between the estuary well and the hidden base of the cliffs is fairly flat, and this would mean that the cliffs here were once nearly 600 feet high. Thus those fringing the peninsula must have been comparable with the coastal cliffs of the Hawaiian Islands, a region with many analogies to Banks Peninsula.
By the kindness of the Christchurch City Engineer I have been furnished with a copy of the log of the well sunk in the South Brighton Domain, about a mile north-east of Fisherman's Flat and
a quarter of a mile from the seashore. This is the well which W. M. Jones (1942, p. 78B) cited as having struck rock at 400 feet, but the log says that a depth of 498 feet was reached without striking any. The record indicates that a total of 163 feet was in gravel, the thickest bed being 73 feet, and the remainder in sand and clay. On comparing this with the record of the Fisherman's Flat well it appears that the amount of gravel is substantially greater, but even this does not approach the proportion encountered in a deep well further north at Kaiapoi (Speight, op. cit., p. 430 and Plate 13), where the record is almost entirely of gravel.
These cases are cited to show the presence of thick beds of gravel close to the present shoreline and presumably extending there-from under the sea-bed as it exists at present. But thick beds of gravel must have been laid down on a land surface or close to its margin, and thus their presence right up to the seashore supports to some extent the former extension of the land to the east.
The extension of the volcanic rock beneath the present alluvial deposits of the Canterbury Plains is also proved by the geophysical survey carried out by W. M. Jones (op. cit. pp. 78B–79B, and Map), which indicates its presence beneath the land surface in the Halswell district and its extension for 4 or 5 miles towards Hornby; no estimate of its depth is given. The occurrence near Halswell is confirmed by the logs of two wells which record the presence of scoria and hard rock about half a mile from the base of the hills and at depths of 182 and 155 feet.
The records of the wells round the fringe of the hills extending from Redcliffs, past Heathcote. Valley to Gebbies Valley, all only a few feet above present sea-level, indicate that a surface of rock or scoria antedating some of the gravels of the plains exists far below tide mark, the maximum recorded to date being 700 feet in the case of the Heathcote well. The presence of a former land surface receives confirmation in some sense from the presence of peat recorded in other wells of the Christchurch area at depths ranging down to 450 feet.
It is unfortunate that the only records of the depth of alluvial deposits within the coastal periphery are those furnished by two wells sunk at Teddington, at the head of Lyttelton Harbour, where a considerable area of flat land lies at sea-level and just above it. The records, such as they are, have been cited by Page and Prideaux (1901, pp. 335–6) on the authority of the well-sinker, who states that “the wells pass through clay, freestone rock, and in one case rubble, but no shingle”; their depths are given as 95 and 177 feet. The reported occurrence of freestone may indicate that the bore penetrated Charteris Bay Sandstone, though no exposures of this rock appear either on the flat or on the hills in close proximity to the sites of the wells. Owing to the uncertainty in this respect the value of the inferences as to the depth of silting within the harbour limits is much reduced. But Mr Orton Bradley, of Charteris Bay, a resident of long standing, and very interested in such matters, tells me that the 95-foot well passed through mud containing shells and resembling the present deposit at the head of the harbour. His
information is, in my opinion, reliable, and it indicates that the solid floor of the hollow lies at a lower level than the present sea-bed at the harbour entrance (depth 7 fathoms), and may be lower than that 15 miles north-east therefrom.
The depth of the sea fringing the peninsula as it stands at present is less than this, and to get values comparable with the record of the Heathcote well, distances from the coastline of at least 30 miles have to be reckoned with. For example, 17 miles northeast of the entrance to Lyttelton Harbour the depth is only 18 fathoms; at 30 miles it is 40 fathoms, while the 100-fathom line is reached at a distance of 35 miles (see Admiralty Chart, 2925, corrected to 1938). East of the peninsula the 40 fathom line is reached in about 7 miles, then the sea-bed apparently continues rather flat, and the 50 fathom line is reached after about 25 miles; while after another 7 miles it sinks to 300 fathoms. Off Akaroa the depth is slightly greater, but at a distance of 25 miles it is still under 50 fathoms.
It should be noted that both north and south of the Long Lookout Point, at the north-east corner of the peninsula, the sea is more than usually shallow, the 10 fathom line making-a bulge away from the coast till it reaches a maximum distance of 5 miles.
Although it may be inferred from the soundings given on the chart that the sea bed is fairly flat, irregularities no doubt occur, to be disclosed by a more closely spaced survey. The 10 fathom bank recorded as lying 10 miles south of the entrance to Akaroa Harbour does not exist according to the statements of G. Brasell and other responsible fishermen. At the edge of the continental shelf the sea-bed sinks rapidly to a depth of over 100 fathoms, but (fide G. Brasell) about 100 miles from land it rises again to a depth of only 40 fathoms, and it is possible to trawl in the locality. This shallowing of the depth is important in view of a former land connection with the Chatham Islands, distant about 400 miles from the coast of the peninsula. The comparatively shallow depth of the sea near the coast is in marked contrast to that off the northeast coast of the Hawaiian Islands, but the latter depths are attributable to faulting, and there is no evidence whatsoever of such in the case of Banks Peninsula unless, indeed, it has occurred far out to sea near the edge of the continental shelf.
This shelf has been the dumping ground for a large proportion of the following material:—
(1) Derived from the erosion of the incompetent gravel, sand, and clay beds forming the coast to the south, and into which the sea has made persistent and considerable inroads.
(2) Brought down by rivers rising in the mountain area of Canterbury.
(3) Derived from the surface and coastline of the peninsula itself.
The first of these, as well as that brought down by the rivers discharging on the south side of the peninsula, is transported by strong currents, tides, and waves, and has shallowed the sea-bed and formed sandy beaches at the heads of the bays as at Peraki,
on the south side of the peninsula. The material from the north is swept south by an eddy which runs past Sumner and the Lyttelton Heads, aided at times by northerly winds which frequently coincide with flooded rivers. In my opinion, this accounts for the shallowness of the sea near Long Lookout Point, and the filling of the lower reaches of Okains and Le Bons Bay Valleys. The extent of flat land in these cases is exceptional, and exceeds that at the head of the great harbours. A considerable portion of the filling is of marine origin, the deposition on this stretch of coast being aided no doubt by the variability in strength and direction of the currents and tides alongshore. According to G. Brasell the current from the north reaches three knots at times, while the chart shows one from the south of one and a-half knots.
A considerable amount of the third class of material has contributed to the filling of the lower reaches of the valleys, both above and below sea-level, the excess being moved off into deeper water and distributed in the same manner as the first two classes. All three would tend to mask any irregularity arising from erosion of the land surface before depression set in, such as an extension of the valleys across the shelf or a discordance in grade if Davis's hypothesis of the presence of hanging valleys be correct.
Deposition results in the formation of extensive areas with a sandy or muddy bottom, and therefore suitable for trawling, in a broad belt parallel to the shore, but there are rocky patches either included in it or lying further off-shore, so that the cover of recent sediments is not complete. Those that exist merge into the marine and estuarine beds of the Christchurch artesian area, specially those round the base of the hills, and these pass into gravels of the land surface, so that the shelf deposits may be regarded as an extension of the land surface where gravels are the dominant constituent. The nature of the rock of the patches off-shore is not known, and they may be an extension of the volcanics of the peninsula or may be beds of sedimentary or other origin on which the mass of the peninsula has been built.
As a rule the upper reaches of the indentations round the coastline have shown little material change owing to deposition by stream or sea action in recent times. There have certainly been cases of filling such as that at the head of Pigeon Bay, this being due to large land slides. It is maintained that Lyttelton Harbour shows no shoaling near the waterway, but some filling has taken place in the upper reaches, and according to the 1926 Annual Report of the Lyttelton Harbour Board, the 4, 5, and 6 fathom contour lines moved seaward during the interval between 1881 and 1897, the first two a distance of nearly a mile, which implies some silting during that time.
One feature of the sea-bed off the Canterbury coast should certainly be mentioned in this connection, viz., the presence of coal measures near the edge of the submarine shelf (Speight, 1908, p. 40). Large masses of brown coal or lignite were reported as having been brought up in the trawl of the Nora Niven at stations both north and south of the peninsula, from depths of between 25 and 40 fathoms, and at a distance from the coastline of some 25 to 30 miles.
Although one occurrence has been reported from 27 miles north-east of Lyttelton Heads, none has been recorded from the east of the peninsula, though it is quite possible, and even probable, that similar beds, or those at an equivalent stratigraphical horizon exist there. If the masses brought up by the trawl have come from outcrops in position exposed on or near the edge of the submarine shelf, the measures from which they were derived probably form the eastern wing of the syncline, the western wing lying towards the base of the foothills of the Southern Alps with the peninsula itself located near its axis. It is possible that the coal-bearing beds may have been brought to their present position by faulting, but the former alternative appears more reasonable, except that faulting on a minor scale might have accompanied the deformation. It is tempting to attribute this deformation to the loading caused by the deposition of the great mass of material by the Canterbury rivers, but it may also be attributed to the subsidence which so often attends the post-eruptive phase of volcanic activity. Whatever the cause, such a sagging of the crust serves to explain several of the most striking features of the area. The present drowned condition of the peninsula follows readily, and there will be no difficulty as regards the relative level of the edge of the submarine shelf and the floor of Heathcote Valley as well as other buried surfaces lying at lesser depth.
Considerable stress has been laid on the special features of the known deposits indicated by the records of the artesian wells, and on the character of the sea-bed off the peninsula, because they give confirmation to changes in land elevation inferred from purely morphological considerations.
It must be remembered that during the Pleistocene Glaciation, owing to eustatic changes in ocean level, the land at times showed greater emergence, and as a consequence it may have reached as far as the edge of the continental shelf. The cutting of existing cliffs and their downward continuation commenced during a period of emergence that perhaps coincided with the height of the latest phase of glaciation. There are no remains of old terraces round the coast to show that the land was relatively lower, but numerous incidents in changing level may be hidden by the sea. It may be noted that Davis considered the present still-stand of the land has been of short duration owing to the narrowness of the shore platforms. If he is correct, then the recent rise of the sea can be attributed to eustantic change following on the decline of the latest phase of glacial intensity, which probably began about 25,000 years ago.
In his very illuminating and complete account of Calderas and their Origin (1941), Howel Williams expresses the opinion that explosion calderas are of very rare occurrence, and that most large hollows associated with volcanic cones are due to collapse arising from (1) the evacuation of material from the magma chamber beneath the cone and the consequent withdrawal of support for the roof during the final phases of eruption, or (2) to the assimilation of overlying material by hot magnas as in the special case of Katmai
(op. cit. pp. 296–300). Williams groups the volcanic hollows of Banks Peninsula with those of Tahiti (pp. 307–9), and attributes them to stream erosion. He notes the difficulty of explaining why in their cases one master stream has dominated the erosion pattern and has been responsible for the barranco or entrance to the hollows, which is so clear-cut in the cases of Lyttelton and Akaroa. He refers to the La Palma caldera, which was supposed by Gagel to be entirely due to stream erosion, but which is believed by Friedlaender and Reck to be due to modification by stream erosion of a hollow and an outlet therefrom primarily caused by collapse. This explanation does not fit the Banks Peninsula calderas as far as the origin of the hollows is concerned, for there is no positive evidence of collapse, but the evidence for the formation of the entrances to the harbours must be seriously considered.
It may be noted that the orientation of the outlet of Lyttelton Harbour is on E.N.E.-W.S.W. lines, which is the aproximate orientation of the dominant fault-lines of the north-eastern portion of the South Island. There is, however, no evidence of faulting on a major scale furnished by exposures near the harbour. The known faulting is on a minor scale and appears on the Gebbies Pass ridge and on the peninsulas stretching down into the head of the harbour, and this is incompetent to account for the entrance.
On the other hand, the orientation of Akaroa Harbour is on N.N.W.-S.S.E. lines, a direction corresponding to that which characterises the faults affecting large areas of the mountain region of Canterbury, and lying at right angles to the direction of Lyttelton. There may be some tectonic connection between these directions of orientation and their coincidence with major fault lines, but they may, after all, have no diagnostic significance.
There are also possibilities that the entrances may have been determined initially (1) by a line of minor craters such as stretched south-west from Tarawera during the eruption of 1886, or (2) by collapse following on the rapid outflow of magmas along tectonic lines, the caldera in both cases being modified by erosion with the master stream draining through the sector graben if formed according to the second hypothesis (Williams, op. cit. pp. 309–10, referring to a discussion of the origin of such valleys by Friedlaender). It is unfortunate that the evidence which might substantiate either of these hypotheses lies buried under the floor of the harbour.
In my opinion the entrances have been formed chiefly by consequent streams which have eroded radial trench-like valleys across the outward dipping beds of a volcanic cone, and the sides of the valleys have been modified especially near the entrances by marine erosion. The latter effect is seen at a maximum in the view of the entrance to Akaroa Harbour (Plate 3, fig. 1). As Williams has suggested, there is some difficulty in accounting for the location and initiation of these master streams, and all I can say is that some accident in topography or a variation in the local intensity of rainfall has favoured one stream more than its fellows. Such an accident might be a low place in the crater ring formed either by a moderate summit explosion or some accidental feature arising during the final
stages of the construction of a volcanic cone, as for example, the low place on the northern side of Ngauruhoe, or that on the eastern side of Ruapehu.
All stages of valley development exist on the external slopes of the Banks Peninsula cones from juvenile forms with V-shaped cross-section to mature forms comparable with those of the great harbours, but it is only those at the mature end of the series which need be considered in this connection. As an example of the submature form Price's Valley may be cited. (Plate 4, fig. 1.) This lies on the southern facing of the peninsula and is external to the Akaroa cone. It is a rather widely opened V in cross-section, and is being widened further by steep gullies at right angles to its axis. These are developed to a slight extent near the mouth of the valley, but are larger and more mature towards its head, a feature which may be attributed to heavier rainfall on the higher slopes.
A further stage in development is shown by Pigeon Bay Valley (Plate 4, fig. 2), also belonging to the Akaroa cone. Its head is more enlarged, so much so that Haast, in a map dated 1879, marked it as a definite centre of eruption. Although the entrance is not so bold as those of Akaroa and Lyttelton, yet the form is exactly the same. And yet another stage is reached with Little River Valley, which both Haast and Hutton considered as a major centre of eruptive activity. This valley has an enlarged head towards whose centre radial spurs stretch down from the rim of the Akaroa caldera. The lower reaches of the valley are narrower, floored with alluvial flats, and near the entrance for some six miles occupied by the shallow and brackish Lake Forsyth, the counterpart in some respects of the great harbours. The entrance is flanked by cliffs, once washed by the sea but now cut off by a shingle bank built of material drifted from the south. The floor of the valley gives no evidence in support either of the hypotheses cited earlier to account for the location of a master stream; the formation of the barranco is not due to faulting; and the valley itself has not been initiated from a low place in the crater ring. It has developed by the normal processes of stream erosion on the exterior of a volcanic cone, and it is merely a further stage in development beyond those illustrated by Price's and Pigeon Bay valleys. The formation of the caldera and barranco presents a further evolutionary stage. In this case the external valley has broken into the heart of the mountain, attacked the outward dipping beds from the rear, enlarged a hollow, collected the drainage therefrom, and thus for a time has secured dominance. But in the case of Lyttelton the master stream has had a rival elsewhere, for the combination of Gebbies and McQueens valleys on the outside of the cone have attacked it from the south-west, destroyed the easily-eroded volcanics of that sector, but found the underlying greywackes too strongly resistant. Thus no second barranco has been formed, although the crater-ring has been completely broken down; only one small fragment of the Lyttelton basic rocks, a few chains in area, and lying on the end of the spur between the two valleys, has escaped destruction, while they have been completely removed from the slopes facing the harbour. But at one stage these valleys or their coalesced
heads must have exerted considerable influence in modifying the south-western part of the hollow now forming the caldera. It will be readily recognised that this account of its formation agrees substantially with that advanced by Cotton (1941, pp. 303–4), when he discusses the origin of the master stream of an eroded volcanic cone.
Little River Valley has another feature worth mentioning, viz., its head is not placed radially as far as the cone is concerned, but tends to sag away till it becomes almost parallel with the edge of the caldera. This unsymmetrical enlargement of the valley head as compared with the axis of the valley is noted by H. T. Stearns (1942, pp. 1–19) in his account of the surface features of the Haleakala volcano of the Hawaiian Islands. Not only the main valleys leading up to the summit depression on the mountain, but also the subordinate external valleys which do not reach it show this feature typically (op. cit., Plate 1 and Figs. 2 and 3), the latter being entirely analogous to those of Pigeon Bay and Little River Valleys. Now Stearns attributes the summit depression of Haleakala chiefly to the coalescence of the heads of external valleys enlarged in the way suggested. He thus continues his advocacy of stream erosion as the dominant agent responsible for the formation of the amphitheatre-headed valleys on volcanic cones which he first advanced in his account of the Geology of Oahu (Stearns and Vaksvik, 1935, pp. 22–29). However, Hinds (1931, p. 186) appears to be the latest advocate of the thesis that they are due to faulting and marine abrasion.
No reference in this account has been made to two other main valleys on the peninsula, viz., those of Port Levy and Kaituna, since they are only partially radial and lie for the greater part of their course along the line of junction of the rocks extruded from each centre.
One further point must be considered in this connection, viz., that the presence of one entrance to a central hollow marks an intermediate stage in the dissection of a volcanic cone. But for the resistance of the greywacke barrier at Gebbies Pass the entrance to the Lyttelton caldera might have been at the south-west end, especially as the southern slopes of the peninsula receive a heavier rainfall; and it is conceivable that two openings might have existed at the same time. The case of Carnley Harbour in the Auckland Islands is interesting in this connection (Speight, 1909, pp. 707–19). The main entrance to this caldera faces east, but the western wall has been broken down largely as the result of the heavy westerly storms that prevail in those latitudes, and through the gap strong waves and tides sweep furiously. There are thus already two openings to the caldera, but another is threatened at the head of North Arm, where the defence against attacks by the sea from the west is wearing thin. Carnley is thus the case of a volcanic cone in a more advanced stage of dissection than either Akaroa or Lyttelton. Thus there is no reason why in the later stages of erosion of a volcanic cone more than one entrance to the central hollow should not be formed. Akaroa presents a caldera of more typical form than Lyttelton because it is a younger volcano and dissection has not been carried as far.
By the kindness of Dr. Barnett, Director of the Meteorological Office, Wellington, I have been furnished with the record for the last 10 years of the various stations on and near Banks Peninsula. These show a much heavier fall at higher levels, and especially those on that slope of the peninsula facing south and south-east. The highest amount, 78.94 inches, an average for three years, is given by a station, height about 500 feet, on the eastern slope of Carew Peak. The stations in Okuti and Puaha valleys of the Little River basin have 52.23 and 43.75 respectively, while Akaroa and Onawe have 47.67 and 43.59—averages for 10 years. The records show a low intensity on the external coastline, and specially so on the north-east and north; for example, the amount at Magnet Bay, east of the entrance to Lake Forsyth, is only 28.52 inches, a little more than half that given for the two Little River stations, while at Little Akaloa (800 feet) the amount is only 33.70, and those given by places on the north side are lower still.
This higher rainfall at the upper levels of a volcanic cone as compared with those at lower levels must exert a marked effect in opening out the heads of valleys and thus promoting the development of the amphitheatre form, due allowance being always made for the structure of such a cone; while the lower rainfall near the mouths of the valleys is responsible to some extent for the maintenance of the barranco-form of the entrance, whether it is exposed to wave action or not.
In his paper entitled “Some Volcanic Landforms in New Zealand” (1941, pp. 297–306) Cotton substantially supports and extends the thesis for the erosion origin of these calderas. One or two comments may be made in connection therewith. First of all, it must be remembered that Lyttelton is the older volcano of the two, so that landscape forms have reached a more mature stage in its case than in that of Akaroa. Thus planezes (adopting Cotton's 1941, p. 300 term, anglicised from de Martonne's) are more maturely developed in the case of the former and cannot be regarded as striking features of the latter volcanic cone. This may also be partly due to the heavier rainfall on the south and south-east side of Banks Peninsula as compared with that on other parts of the periphery, as a result of which the radial valleys belonging to the Akaroa cone are more closely spaced and more deeply incised, the planezes thus becoming indefinite. But even if they were strongly developed at a former stage, then the more active marine erosion of the southern and eastern coasts has removed the broadly extended distal sectors of the marginal belt, leaving those parts nearer the centre of the cone where the feature is less apparent. When it was an island the western margin must have experienced weaker wave action, and even this ceased when the island was joined on to the mainland, so planezes are more typically preserved on the strip of country extending round the periphery from the Sumner Estuary to the north-eastern shore of Lake Ellesmere.
A specially good example is that near the Woolston station on the Christchurch-Lyttelton railway. This planeze has been separated from the crater ring by the approach of the heads of two valleys lying on either side, and it seems certain that the stream running through
the Avoca Valley on the east will ultimately capture the head of the stream on the west, and thus carry the isolation further. It may be noted that the surface of this planeze is modified by subordinate streams in a very early stage of development, and it retains its flattish form, a form determined primarily by a persistent covering flow of the olivine-andesite typical of the area. Further west the Cashmere settlement occupies the lower levels of another planeze, more maturely dissected, with the Bowenvale and Cashmere valleys on either flank converging at their heads, and thus tending to isolate the distal portion. On other parts of the periphery, both to the east and to the west, there are further examples, but those on the southern side of the peninsula have experienced a more mature insequent erosion, and consequently their surface is more completely dissected; this effect may be due to heavier rainfall. A good example of this mature dissection is furnished by the sector between Gebbies and Kaituna valleys.
These planezes have been modified to a greater or less extent by secondary and ternary streams which have eroded short valleys on their fringes, the lower reaches being frequently masked by alluvial deposits of the plains and by material carried down by the streams themselves after rain—they seldom carry permanent water. They also show features characteristic on a major scale of the dominant valleys on the exterior of the volcanic cones, viz., they are frequently double- and treble-headed. Thus Pigeon Bay is double-headed, while the Little River and Gebbies valleys are treble-headed. The discontinuous character of the interstratified hard and soft beds, with a scanty lateral extension as a rule, and dipping outward from a composite volcano, is no doubt chiefly responsible for the development of the primary and secondary valley system. As noted by Cotton (1942, p. 391) the crest of the divide between adjacent valleys is frequently formed by the convex surface of a lava tongue. When lava issues in broad sheets the planeze is more typically developed.
The slope of Mount Herbert as it reaches down to Lyttelton Harbour is an excellent example of this feature. Owing to its being formed of sheets of resistant basalt with wide lateral extent and probably of very late date in the volcanic episode, it has been very slightly scored by surface streams and their channels are in a very juvenile stage of development. They are deeply incised only on the sea margin of the slope. It has been attacked from either flank by the Purau and Charteris Bay streams, whose heads are approaching each other. Its seaward fringe has been partly removed by the weak action of waves inside Lyttelton Harbour, and like the cliffs on the exposed coastline of the peninsula the shore platforms are only slightly developed. So the cliffs were cut back when the land stood at a higher level, and one cannot say what was the depth of water in the harbour at that time, since its bed has been raised substantially by deposition of material shed from the slopes of the hills.
The foregoing presentation has dealt seriatim with the considerations raised by Davis, Williams, and Cotton as they bear on the morphological features of the area, and especially on the mode
Fig. 1. South-west head of Akanioa Harbour, view taken from north-east head looking south. The cliffs are over 500 feet high There is a very narrow shore platform, but outlying rocks [ unclear: ] from deep water show above the surface a quarter of a mile from the base of the cliffs
Fig. 1. Phoes valley looking down the valley from the head [ unclear: ] coming in at right angles show towards the mouth of the valley and are more developed near the head These are to be distinguished from more or less [ unclear: ] valleys which are slightly developed near the outlet of 1 main valley a typical minstance of which is to be seen near Budling Flat at the outlet of Little River Valley
of formation of the great hollows or calderas now occupied by Lyttelton and Akaroa harbours. It will be readily recognised that these authorities are in substantial agreement among themselves and with the present author in regarding the major land-forms, including the calderas, as the result of the erosive action of streams on the interior of the craters and the exterior slopes of the two volcanic cones, this action having been initiated and continued when the land stood at a higher level than at present.
Since the Akaroa volcano shows little departure from the simple and regular structure of a single composite cone, the present form of the central hollow may well be used as the type of an erosion caldera in an early mature stage of development, and in the opinion of the author it has been formed in the following manner. One of the streams on the exterior slope of the cone was favoured by some accidental factor, such as (1) a moderate summit explosion which resulted in the break-down of a portion of the crater-ring, or (2) a higher intensity of rainfall over one sector. In consequence the stream was able to break into the crateral hollow from the outside, take to itself the interior drainage, become the dominant or master stream, and enlarge its head to an amphitheatre form by the headward and lateral erosion of tributaries converging towards the centre of the basin. When subsidence of the land took place the sea invaded the hollow and formed the present harbour. Some of the streams on the exterior slope, such as those occupying the Little River and Pigeon Bay valleys, also enlarged their heads so that they simulate a caldera form, not only in their upper reaches, but also in their cliff-guarded entrances; but up to the present they have not succeeded in breaking into the enlarged crater. When the land was lowered the sea occupied the lower reaches of the valleys and formed the fringe of bays round the present coastline. The distal ends of the spurs dividing the bays have been exposed to heavy seas, and have been cut back into bold cliffs, notably when the land stood at a higher level; a subordinate amount of cutting has taken place at about present sea-level.
The erosion pattern of Lyttelton has been carried a stage further as it is the older volcano of the two. The crater-ring has been broken in two places, and the narrowness of the summit ridge at the head of some of the exterior valleys, as for example that of Heathcote Valley, foretells the location of future breaks in the ring. There are other signs as well, such as the presence of planezes, which indicate a more mature stage of erosion than is shown by Akaroa. Although bold cliffs mark the coast near the entrance to Lyttelton Harbour, and also near Sumner and Redcliffs, the spur-ends of the remainder of the periphery are not so distinguished. This absence is no doubt due to the more sheltered position of the western stretch of coastline when the peninsula was an island, and any cliffs then cut are now masked by the subsequent lowering of the land and by the material brought down by the Canterbury rivers and deposited along the western fringe of the eroded cone.
Finally, in making a comparison between these two calderas and that of Carnley in the Auckland Islands, it should be noted that the second break into the craterial hollow of the last is to be credited
almosts entirely to the action of the sea, and it is not analogous to the break into Lyttelton near Gebbies Pass. Rather is it an example of Davis's maxim quoted earlier that in small pelagic islands strong wave action is more important than short stream erosion. The contrast of the land-forms of the sheltered area east of Carnley with the strongly wave-corded area to the west—both showing drowned topography—convincingly supports Davis's contention.
Chief References to Literature.
Cotton, C. A., 1941. Some Volcanic Landforms in New Zealand. Journ. of Geomorphology, vol. 4, No. 4.
——, C. A., 1942. Geomorphology. Third Edition.
Davis, W. M., 1928. The Coral Reef Problem. Amer. Geog. Soc. Spec. Publ., No. 9.
Hinds, N. E. A., 1931. The Relative Age of Hawaiian Landscapes. Calif. Univ. Dept. Geol. Sci. Bull., vol. 20, No. 6.
Jones, W. M., 1942. Magnetic Disturbances on the Canterbury Plains. Dept. Sci. and Indust. Research, vol. 23, No. 3 B, pp. 73B–83B.
Page, S., and Prideaux, E. B. R., 1901. Notes on an Artesian Well System at the Base of the Port Hills. Trans. N.Z. Inst., vol. 33, pp. 335–6.
Speight, R., 1908. Terrace Development in the Valleys of the Canterbury Rivers. Trans. N.Z. Inst., vol. 40, pp. 16–43.
——, 1909. Physiography and Geology of the Auckland, Bounty and Antipodes Islands. Subantarctic Islands of New Zealand, pp. 705–44.
——, R., 1911. A Preliminary Account of the Geological Features of the Christchurch Artesian Area. Trans. N.Z. Inst., vol. 43, pp. 420–36.
——, R., 1917. The Geology of Banks Peninsula. Trans. N.Z. Inst., vol. 49, pp. 365–92.
Stearns, H. T., 1941. Origin of Haleakala Crater, Island of Maui, Hawaii. Bull. Geol. Soc. Am., vol. 52, pp. 1–19.
——, H. T., and Vaksvik, K. N., 1935. Geology and Ground-water Resources of the Island of Oahu. Terr. of Hawaii, Div. of Hydrography, Bull. 1.
Williams, Howel, 1941. Calderas and their Origin. Univ. Calif. Publ., Bull. Dept. Geol. Sci., vol. 25, No. 6, pp. 239–346.