Go to National Library of New Zealand Te Puna Mātauranga o Aotearoa
Volume 50, 1918
This text is also available in PDF
(1 MB) Opens in new window
– 212 –

Art. XXIII.—The Geomorphology of the Coastal District of Southwestern Wellington.

[Read before the Wellington Philosophical Society, 12th December, 1917; received by Editors, 31st December, 1917; issued separately, 10th June, 1918.]

Plates XIV, XV.

Contents.
  • Introduction.

  • The Foundation or Old Land.

  • The Coastal Lowland.

  •    Theoretical Discussion of the Growth of a Coastal Lowland under Conditions of Fluctuating Waste-supply.

  •    An Alternative Explanation.

  •    The Paekakariki Coast.

  •    Another Alternative Explanation.

  • Subdivisions of the Lowland.

  •    The Otaki Series.

  •       Topography of the Otaki Series.

  •       Distribution of the Otaki Series.

  •       Lithology and Structure of the Otaki Series.

  •    The Fans or Gravel Plains.

  •    The Delta of the Manawatu.

  •    The Modern Dunes.

  •    Lakes and Swamps.

Introduction.

The coastal district of south-western Wellington (fig. 1), forming part of the fertile “Manawatu” and largely comprised within the limits of Horowhenua County, presents a considerable variety of physiographic features all of comparatively modern growth* and explicable on an assumption of a geological history of a somewhat unusual kind.

I have been informed by Dr. L. Cockayne and Mr. A. H. Cockayne that the ecology and agriculture of the district are closely related to the physiography, and I am tempted to present this somewhat sketchy description and attempted explanation of the forms in the hope that it may be of use as a basis for further studies.

The Foundation or Old Land.

The skeleton or foundation of the coastal district of south-western Wellington is the upland block of somewhat old rocks sculptured into strong relief that forms the Tararua Mountains and the lower ridges farther south. Lithologically these rocks are greywacke-sandstones with occasional bands of argillite, the latter much sheared by earth-movements. All are closely

[Footnote] * Geologically the whole history of the lowland is comprised within a portion of the Notopleistocene period.

[Footnote] † The early geologists termed these rocks the “Rimutaka series,” a local name that might be now appropriately used for them. Subsequently to the issue of the geological map of 1873 the name “Rimutaka” gave place to “Maitai,” as it was believed that the two formations were identical; but it was used occasionally by McKay in the “seventies” for the whole or part of the rocks of the Rimutaka and neighbouring ranges.

– 213 –

folded, so that the strata are now everywhere nearly vertical, and the strike, though variable from point to point, averages a little to the east of north. Some bands of the rock are thoroughly shattered and allow of the percolation of water along closely spaced joints, so that they yield rather readily to deep weathering and erosion. Other hands have escaped shattering or have had the network of crevices re-sealed by a deposit of secondary mineral matter, so that there are now few joints, and weathering and erosion, go on relatively slowly. These latter relatively resistant bands of rock stand out boldly as ridges, while the bands of shattered rock are now followed by streams, and, where of considerable breadth, are reduced to somewhat subdued forms with moderate relief.

Picture icon

Fig. 1.—Map of the coastal district of south-western Wellington, including the upland ranges eastward to the main divide. Scale, 10 miles to 1 in. The inset map shows the locality.

It is not known with certainty that the alternating bands of resistant and weak rocks, which owe their varying strength, as just noted, to the extent to which they are fissured, correspond exactly to the bedding; but they are at least elongated in the same general north-north-easterly direction, and the shattering in the weakened bands was probably caused by

– 214 –

the ancient folding. (An exceptional belt of shattering close to the city of Wellington, which crosses the strike diagonally, is not here referred to. It appears to be connected with fault movements of a much later date than the folding of the rocks.*)

The texture of the dissection on both the higher ridges of resistant rock and the lower belts of subdued topography is somewhat fine, owing, no doubt, to the low permeability of the mantle of residual clay that results from the weathering of the greywacke.

The development of the present mature topography seems to have been interrupted from time to time by renewed uplift, as in the Wellington peninsula farther to the south-west, but the high valley-floors of the earlier cycles are here maturely dissected, and the land-forms of the various cycles merge almost or quite completely into one another. Some of the later pauses in the uplift are probably recorded by the fragmentary terraces in some of the valleys. These are generally rock terraces, but some are formed of alluvium—e.g., in the Otaki valley a thick mass of fluviatile gravel underlying a portion of a terrace and extending below the present level of the river suggests trenching, due to uplift, and later refilling of the trench, perhaps during a temporary submergence preceding an uplift to which the cutting of the present narrow inner valley of the river is due. The terrace features cannot be ascribed wholly to vertical movements of the land, however, for they must be closely connected with certain to-and-fro movements of the shore-line which will be described below.

The margin of the upland block is a mature coast-line rising in places as high cliffs, but fronting the sea now only at the south-western end, beyond the limits of the district here dealt with, and bordered elsewhere by a coastal lowland the width of which increases north-eastward. With respect to this lowland of later growth the upland block may be spoken of as the “old land.” The ancient coast-line of the old land appears to have originated as a fault coast, for its almost straight north-east and south-west trend crosses the grain of the country obliquely. Whatever the initial form may have been, however, a mature coast of simple outline developed by marine erosion now forms a nearly straight boundary-line between the upland block or old land and the strongly contrasted coastal lowland.

The Coastal Lowland.

Different parts of the coastal lowland are of different ages, and their present topographic forms have been developed in different ways, and the materials of which they are composed, though in some cases originally identical, are in different stages of consolidation and decay, so that they yield very different soils. The materials also came originally from two distinct sources.

The divisions of the lowland are best introduced by a discussion of the conditions under which it seems to have come into existence. All the features of the lowland may have been produced by an alternation of retrogradation (or retreat of the shore-line under wave-attack) with progradation (or advance of the shore-line due to accumulation of the waste of the land). Such an alternation is necessarily connected with a fluctuation in the supply

[Footnote] * G. A. Cotton, Supplementary Notes on Wellington Physiography, Trans. N.Z Inst., vol. 46, pp. 294–98, 1914.

[Footnote] † C. A. Cotton, Notes on Wellington Physiography, Trans. N.Z. Inst., vol. 44, pp. 245–65, 1912.

– 215 –

of waste, leading to a fluctuation in the ratio of wave-energy to load, for, when the supply of waste is small, waves attack a coast vigorously, cut it back, and draw much of the waste produced in this process back into the deeper water off shore, where it comes to rest; whereas when there is a large supply of gravel or sand, either brought in by local rivers or transported along-shore by the activity of waves and currents from a more distant source, the energy of the waves is used up in maintaining a graded off-shore profile of the bottom as the abundant waste accumulates at all depths, and some of the material is thrown up on the beach, so that the shore-line advances seawards, leaving a prograded strip of new land.

In the case in question it is probably the supply of sand, which comes from rivers farther to the north-east, that has fluctuated, rather than that of gravel brought down by local streams. The cause of the fluctuation is not apparent. Changes of level of small amount would have an effect, no doubt, by disturbing the graded profile of the neighbouring sand-covered sea-bottom, and would perhaps produce alternate overloading and underloading of the waves at the shore-line. The fluctuation in the supply of sand is too great, however, to be attributed to that cause alone. I have not recognized the concomitant effects of such small movements on the coastal topography, nor succeeded in distinguishing them from the effects of advance and retreat of the shore-line, and so they are necessarily ignored in this account.*

Theoretical Discussion of the Growth of a Coastal Lowland under Conditions of Fluctuating Waste-supply.

Where a coast, perhaps originally a fault coast—though this is not essential—has been cut back to the mature stage, like the ancient coast of the old land of south-western Wellington (fig. 2, A), and a change to progradation takes place, a strand-plain, generally dune-covered, is developed (fig. 2, B). The material is mainly “imported” sand, if it is assumed that the excess of material has been brought from another section of the coast, but there will be mixed with it some gravel of local origin near the mouths of streams. As the streams grow in length seaward, however, with the growth of the prograded strip, they are constrained to aggrade so as to build up channels with sufficient slope to maintain their flow, and a proportion, perhaps the whole, of their gravel is thus used up, and accumulates as fans along the base of the cliffs of the old land (fig. 2, B), Clearly, if these streams are somewhat closely spaced and bring down much gravel their fans will become confluent, forming a piedmont alluvial plain; but, on the other hand, if their supply of gravel is smaller in proportion to the amount of sand being thrown up along the shore-line they will remain separate, and on the spaces between them dune sand may accumulate to a considerable thickness.

There is thus developed a coastal lowland of general seaward slope, with a somewhat irregular surface, and possibly with a width of many miles.

[Footnote] * Adkin reports that a trial well-boring on the lowland passed through a swampy layer (i.e., a land surface) far below present sea-level (Trans. N.Z. Inst., vol. 43, p. 498, 1911). I have not been able to obtain official records of any of the bores that have been put down; but if this interpretation is correct it indicates that the lowland, during an early period of outward growth, advanced, or was at least maintained, in spite of considerable subsidence (as in the case of the Canterbury Plain). To make this possible the supply of waste at that stage must have been very abundant.

– 216 –

If now the landward dunes have become fixed by vegetation and somewhat consolidated, they will in time have a normal drainage system developed on them, and their irregularity of surface will be reduced by the filling of the hollows and the wearing-down of the initial hills, so that their surface will become a peneplain (fig. 2, C).

Picture icon

Fig. 2.—Diagram of successive stages in the growth of a composite coastal lowland during two periods of progradation separated by one of retrogradation.

If now after a period of slow progradation, perhaps followed by a period of stationary shore-line, retrogradation begins again, the seaward-sloping coastal lowland, as its toe is cut back to a line of cliffs of growing height, will be dissected owing to the shortening of its streams. Since the dunesand areas lie between the larger streams from the old land, they will be traversed only by systems of small streams arising on them or heading on the cliffs behind. The energy of such streams is slight, but they work on very weak material and so erode quite rapidly (fig. 2, D). The larger streams from the old land will now trench the fans they formerly built, and they may destroy the fansurface altogether or reduce it to fragmentary terraces, developing streameroded plains at lower levels, which may in places be widened so as to plane down parts of the dune-sand areas.

The lowland will now resemble in a general way a dissected coastal plain of subaqueous origin, but there will be differences in detail, for the initial form in this case is a subaerially graded surface, and when the streams entrenched below it have again become mature the slopes of their valley-plains will approximate to that of the interfluves, and also such of the interfluvial areas as are underlain by dune sandstone will not be quite flat, but will retain the small relief of a peneplain.

There may be pauses in the retrogradation of the coast, and even reversals from time to time, and thus several series of terraces or valley-in-valley forms may be developed (fig. 2, E).

It may be postulated that after the coastal lowland has been cut back by

– 217 –

the sea to perhaps half, or less than half, its former width progradation again sets in, and a new strand-plain, dune-covered like its predecessor, grows seaward from the recently cut cliffs. This will give rise to various modifications in the form of the dissected lowland. It will cause vigorous aggradation in all the gravel-bearing streams from the old land, and some of these may completely fill the valleys they have previously excavated in the lowland. Thus the fans are reconstructed (fig. 2, F). Intercalated short periods of retrogradation or other causes may lead to channelling of the fans from time to time, followed by renewed aggradation, and similar results may be produced by fluctuation of level. Thus there may be a considerable amount of complexity in the structure of fans.

Along the border of the inter-fan areas the even seaward slope of the lowland is not so readily restored. Some aggradation will take place, however, along the lines of the small dissecting streams, but irregularly, the greatest effects being seen where the channels, perhaps kept open for a time across the newer strand-plain, becomes blocked by sand-dunes, forming lakes partly within and partly beyond the border of the older dissected lowland (fig. 2, F). These, when filled, become swamps with arms extending up the floors of the tributary valleys, which are in process of aggradation with fine silt. Similar swampy flats will occur as normal features also among the dunes on the newer strand-plain.

Meanwhile the cliffs along the toe of the older lowland will be reduced to gentler slopes by subaerial erosion, which goes on quickly in this soft material, and reduction of the surface will still continue. Parts of it are by now maturely dissected, and there may be close interfingering between the spurs of the older and the dunes of the newer lowland.

In a coastal lowland developed as outlined above four quite distinct physiographic types of surface can be recognized: (1) The older dunesandstone areas of the dissected older lowland, with peneplained tops, mature topography towards the margins, and more or less dissected terraces in the valleys; (2) the gravel fans, which may still be confined between low banks of the dune sandstone or may overlap its peneplained surface; (3) the newer sand-dunes, which still exhibit the forms due to accumulation; and (4) swampy flats, which have accumulated in lakes due to ponding among the newer dunes or between these and the toe of the older lowland, or are the result of aggradation in the silt-bearing small streams trenching the older lowland. In the coastal lowland of south-western Wellington all four of these physiographic elements are important.

An Alternative Explanation.

An alternative explanation which would account for the existing forms in the broader part of the lowland fairly well would make the “older lowland” of the foregoing a coastal plain of subaqueous sands, which was subjected to subaerial erosion with the shore-line stationary for a period long enough to allow of peneplanation, and afterwards cliffed at the margin and dissected. The remaining events would be the same as those outlined in the previous explanation. An objection to this explanation is found in the nature of the material of the older dissected lowland, which will be referred to on a later page; and an argument in favour of the explanation previously given is found in the clearly decipherable history of the lowland at its extreme south-western end (at Paekakariki), where there is evidence of alternating progradation and retrogradation on a scale sufficiently large to warrant the assumptions made.

– 218 –

The Paekakariki Coast.

South-westward from the tapering end of the lowland at Paekakariki the cliffed margin of the old land is now being cut back, and there is no evidence to show whether this part of the coast has ever been prograded. The dune-covered strand-plain is now extending in that direction along the base of a line of high, fresh cliffs cut diagonally across one of the resistant ridges of the old rocks (Plate XIV).

A mile or two north-eastward, however, there is evidence of the former presence of a strand-plain and of its removal prior to the growth of the present lowland. This evidence is found in the presence of cliffed remnants of fans of locally-derived gravel, one of which is of such dimensions that the fan when complete must have extended seaward about a mule beyond the margin of the old land.

These fans were built forward in what may be termed the first (strictly the nth) progradational phase (fig. 3, B), following the first (strictly nth) retrogradational phase, in which the cliffs of the ancient coast-line of the old land were cut (fig. 3, A). These cliffs are now subdued and rounded, and pass by smooth concave curves at the base into the fans and talus slopes.

Picture icon

Fig. 3.—Diagram of successive stages of a coast alternately retrograded and prograded.

The fans are irregularly truncated by a younger line of cliffs developed in a second [(n+1)th] retrogradational phase, and these are cut back far enough in places to intersect the line of older cliffs (fig. 3, C). The cliffs previously referred to as lying behind the extreme end of the lowland are continuous with these of the second [(n+1)th] retrogradational phase.

In front of this newer line of cliffs lies the modern lowland or dune-covered strand-plain, a belt of dunes enclosing between themselves and the cliffs a narrow strip of marshy plain (fig. 3, D). Fig. 4 is a panoramic sketch of this portion of the coast, illustrating the features described. Plate XV, fig. 1, is a photographic view showing the truncated fans.

Picture icon

Plate XIV.
General view of the narrow southern end of the coastal lowland, looking northward from the cliffs south of Paekakariki.

Picture icon

Fig. 1 —Fan with cliffed seaward margin, north of Paekakariki
Fig. 2.—Dissected bench of Otaki sandstone near Shannon

– 219 –
Picture icon

Fig. 4.—The Paekakariki coast, viewed from the modern fixed dunes. Angle of view, south-east to south-west.

Another Alternative Explanation.

An account of the geological history of the coastal lowland which diverges considerably from that assumed in the explanation of the physiography here adopted was given by Adkin,* whose account deals in particular with that part of the lowland adjoining the Ohau River.

Adkin's classification of various stages in his concept of the history of the lowland as Early Pleistocene, Middle Pleistocene, Later Pleistocene (First, Second, and Third stages), and Recent must be discarded, as he had no means of correlation with deposits of Pleistocene age elsewhere; and for the present purpose the stages “Early Pleistocene” to “Recent” may be renamed stages 1 to 6.

According to his interpretation, at stage 1 the Ohau River built a great fan over hypothetical undissected uplifted Pliocene formations with a plane surface, the latter being vaguely “inferred from the configuration and character of the superimposed fluviatile deposit.” Large parts of the surface of the fan formed at this time are regarded as surviving to the present day, though buried by a marine deposit and re-exposed by erosion in the intermediate historical stages.

Stage 2 was a period of complete submergence of the lowland beneath the sea. This was followed by a period of still-stand, succeeded by uplift continuing to the present day. During the submergence at stage 2 a “raised-beach formation” was deposited, consisting of beach sands spread over the whole area of the lowland partly during the advance of the sea and partly during its retreat. This “raised-beach formation” comprises the partially consolidated sands of the older lowland. Adkin states that its present level surface is not the original one, as it has been lowered by erosion. It is not clear, however, to what base-level he ascribes the planation, or at what stage of the history it occurred. Dissection of the surface by small streams is mentioned in the description of an illustration.

[Footnote] * G. L. Adkin, The Post-tertiary Geological History of the Ohau River and of the Adjacent Coastal Plain, Horowhenua County, North-Island, Trans. N.Z. Inst., vol. 43, pp. 496–520, 1911.

– 220 –

The remaining stages are marked by changes in the courses of the Ohau and other streams over the Ohau fan, for the explanation of which Ferrel's law is invoked, though it is much more probable that such changes as have taken place in stream-courses have resulted from spilling over as a normal accompaniment of aggradation. At the same time the shore-line advanced steadily seaward. It is not clear whether this is regarded as entirely the result of an inferred movement of uplift, but this is probably what the author had in mind, as he speaks of the whole lowland as a coastal plain. The new land formed thus was progressively covered with sand-dunes, which impounded lakes.

The account is somewhat difficult to follow, but the foregoing is a fair summary. Adkin regards the gravel plains, such as the Ohau fan, as the oldest, instead of placing them among the youngest elements of the lowland physiography, as is done in the explanation now offered; and his conclusion that any portions of the existing fan-surfaces or gravel plains were in existence prior to the deposition of the sands of the older lowland is an extremely doubtful one, whatever the correct explanation of the mode of accumulation of those sands may be.

It must be added that Adkin's work is obviously based on a large amount of careful field-work; and his mapping in the Ohau River district (Horowhenua) is extremely useful.

Subdivisions of the Lowland.

The Otaki Series.

The oldest physiographic element in the lowland is, then, the dissected peneplain of soft sandstone corresponding to the “older dune-sand areas” of the theoretical discussion previously given, and comprising the “raised-beach formation” of Adkin. To the lithological formation of sandstone thus indicated the name Otaki series may be applied as a local formation name, as it is well developed just north of the Otaki River and town.

Topography on the Otaki Series.—The topography is that described in the theoretical section as developed on the “older dune-sand areas,” and shown diagrammatically in fig. 2, F (see also Plate XV, fig. 2). I agree with Adkin* that the gently undulating tops of the broad benches of this formation are parts of a surface of erosion and not of deposition. In some parts of the district a considerable area of the surface of the Otaki series consists of broad terrace-remnants of valley-floors at intermediate levels developed generally by very small streams which now make their way along the flat, swampy floors of inner valleys. A striking characteristic of the small dissecting streams arising within this formation is the steepness of their valley-sides, which remains practically constant as the width of the floor increases, and the same slope continues around the valley-heads. To their very heads the valleys are box-shaped rather than V-shaped. The slopes separating the broad terraces at intermediate levels were evidently once exactly similar to those of the inner trenches, though they have become somewhat dissected and broken down since the streams were revived and undercutting of these slopes ceased. Where the terraces have been developed by gravel-bearing streams heading in the old land they are gravel-covered.

Distribution of the Otaki Series.—Between the southern end of the coastal lowland and the Manawatu River the Otaki series, with its characteristic

[Footnote] * Loc. cit., p. 509.

– 221 –

topography, covers an area of perhaps sixty square miles, of which about half consists of nearly flat summits. One considerable area occurs between the Otaki and Ohau Rivers; an “island” of it, surrounded by gravel plains evidently parts of the Ohau fan spread by distributaries of that river, occurs at Weraroa, where the Central Development Farm of the Department of Agriculture is situated partly on this formation and partly on the fan; and there is then a nearly continuous bench, broken only by some gravel-covered valley-floors, extending north-eastward for twenty miles.

Lithology and Structure of the Otaki Series.—The prevailing material in the Otaki series is grey sand similar in mineral composition to that of the present beach and the associated dunes. In addition to quartz, the sand contains a considerable proportion of feldspathic, ferromagnesian, and iron-oxide grains. The mineral grains, including those of quartz, though not completely rounded, have their angles smoothed off, and they thus contrast very strongly with the sharp, angular grains of the present beach. This suggests aeolian accumulation. All the samples examined are somewhat weathered, however, and the rounding of grains may be ascribed in part to weathering.

The more or less coherent sandstone formed of this material weathers at the surface to a residual sandy clay, usually containing scattered spheroidal masses of the sandstone. The permeability seems not to be great, for the water-table is generally close to the surface. Much water seeps out along the bases of even low scarps, and necessitates draining.

In the few sections where bedding has been noted the beds are inclined at about 35°. It is quite clearly cross-bedding on a large scale, and again suggests subaerial accumulation. Cross-bedding and also ripple-mark are noted by Adkin.* A horizontal pseudo-stratification, due, apparently, to deposition of iron, is also generally present. Though less prominent, it seems to resemble that noted by Berkey in the aeolian San Juan formation of Porto Rico.

The non-discovery of fossils, though a negative character, points also to the possibility of subaerial accumulation of the sand of the Otaki series.

Clay lenses and bleached soil-beds, which are interbedded with the sandstone, accumulated, no doubt, in lakes and swampy areas impounded among dunes. The clay-bed between upper and lower sands noted by Adkin,§ which he ascribes to marine deposition at a period of maximum depression, is perhaps one of these.

In one section near Shannon layers of small pebbles occur interbedded with the sand. Here is probably the course of one of the smaller streams from the old land, or perhaps the margin of one of the larger fans. If gravel fans and dune sands accumulated side by side, as is assumed in the theoretical section, there must be a considerable amount of intermixture of material along the transition lines where gravel passes laterally into sand. Along these lines, indeed, a complex interfingering of gravel and sand beds may be expected.

The Fans or Gravel Plains.

Throughout the length of the lowland there are numerous gravel fans, both great and small. The largest are those of the Otaki and Ohau Rivers,

[Footnote] * Loc. cit., p. 507.

[Footnote] † C. P. Berkey, Geological Reconnoissance of Porto Rico, Ann. N.Y. Acad. Sci., vol. 26, pp. 1–70 (see p. 50), 1915.

[Footnote] ‡ G. L. Adkin, loc. cit., pp. 497, 507.

[Footnote] § Loc. cit., p. 507.

– 222 –

which have a combined area of about forty square miles. Some parts of their surfaces are thickly covered with small boulders or coarse gravel; others have a gravelly soil; while others, again, have a superficial layer of silt overlying gravel. Most of the fans are trenched and terraced to a small extent. The surfaces of the fans and of terraces cut in them are very similar to one another, as are also gravel-covered terraces within the border of the old land. All these may be classed as gravel plains. The gravel-bearing streams are at present aggrading as though to refill the trenches in the fans. The actual stream-beds are, therefore, areas of bare gravel over which the streams flow in changing, braided channels. As previously mentioned, the upper surface of the Otaki formation passes in some places into that of a fan without any abrupt break of slope.

The Delta of the Manawatu.

More or less analogous with the fans of the southern part of the lowland is the delta of the Manawatu River; but this is one, perhaps the chief, of the sand-supplying rivers. Its delta is composed mainly of fine material, and its gradient is very gentle as compared with that of the gravel fans. The Manawatu delta forms a plain of wide extent lying at present almost entirely on the north side of the river, and continued up-stream by a wide flood-plain, below which the river is now slightly entrenched, and above which there are broad terraces on the northern side. The seaward margin of the delta is covered with dunes, some belts of which extend inland many miles. The Manawatu River at present bends to the south-west after emerging from its gorge across the old land, and at a not very distant date it swung still farther to the south. The toe of the bench formed by the Otaki series is here cut back to a line of cliffs by the action of the river, and at the base of these a considerable area of ill-drained flood-plain, now abandoned by the river owing to its slight entrenchment, forms the great Makurerua Swamp (see fig. 1). The whole of the delta plain was formerly swampy, but a great part has been artificially drained.

The Modern Dunes.

The modern dunes are built of grey sand similar to that forming the sandstone of the Otaki series. All except a narrow belt close to the sea are fixed by vegetation, but beneath the superficial layer of humus the sand is still quite loose. The belt of dunes has a width of from three to six miles, and their average height is 170 ft. Adkin notes that their general arrangement is in ridges at right angles to the coast-line.* The shore-line of the dune-covered foreland advances as a broad cusp towards Kapiti Island (a high island of old rocks some four miles from the mainland). This is evidently an early stage of island-tying.

Lakes and Swamps.

Several lakes and many small ponds and swamps formed by the silting-up of ponds lie between the modern dunes and the margin of the other physiographic elements of the lowland, and there are many swampy areas among the modern dunes. The valley-floors in the Otaki formation are practically all swampy, as a result either of normal aggradation with fine silt or of ponding by sand-dunes followed by accumulation of silt. The largest swamp in the district—the Makurerua Swamp—has been referred to above

[Footnote] * G. L. Adkin, loc. cit., pp. 514–15.