Rock Fans in South-East Wellington

Transactions and Proceedings of the Royal Society of New Zealand, Volume 85, 1957-58, Page 101

Rock Fans in South-East Wellington By J. B. Waterhouse Victoria University College, Wellington [Received by the Editor, January 30, 1956.] Abstract Large rock fans with a veneer of alluvium are described from the eastern flank of the southern Haurangi Mountains in south-east Wellington. These have been carved by small, heavily loaded streams from the mountains in relatively unresistant Cretaceous and Tertiary beds at the foot of the range. Following the formation of the fans the streams became incised, and new rock fans were carved into the older ones. This process has recurred several times, and now up to five rock terraces border the lowest rock fan. On the lowest rock fan, alluvium to a depth of over 40ft is tentatively correlated with the rise of sea level following the Wisconsin (Würm) Glaciation. It is suggested that most of the lateral planation to form the rock fans was effected during the glacial episodes of the Pleistocene, followed by aggradation on the lower slopes of the rock fans in the early part of the warm episodes with the rise in sea level. Later, dissection occurred during the later part of each warm and early part of the next cold period. Tectonic uplift has also been involved. Introduction Rock fans have been described so rarely in New Zealand that it may be well to emphasize that a rock fan is cut—carved out of rock by a stream or river—not built of alluvium. The features closely resemble alluvial fans in shape, but are developed by erosion on bedrock. Cotton (1948, p. 178) describes rock fans as follows: “The process of lateral corrasion or ‘planation’ at a stable gradient … cuts a bedrock floor (only thinly veneered with a flood-plain deposit of alluvium) which … in theory must be slightly convex in transverse profile, as it is part of the surface of a very flat cone with its apex at the source of the river… Small streams of steep gradient have indeed occasionally succeeded in cutting laterally far enough to develop small ‘plains’ that are quite strongly convex. Davis has called these ‘rock fans’.” Johnson (1932, p. 392) seems independently to have recoined the descriptive term “rock fan”. In his discovery of examples in the field, and in descriptions by McGee (1897), he finds confirmation of his earlier deduction that there must be such forms. The Germanised form Felsfächer, or alternatively Felskegel, is used by von Wissmann (1951, pp. 17, 31) in describing examples from China. Location and Distribution of the Rock Fans The Haurangi (or Aorangi) Range is a forest-clad rugged chain of mountains in south-east Wellington. The mountains rise gradually eastwards from the Waira-rapa Plains, to peaks over 3,000ft in height. On the eastern side, the range forms an imposingly high and steep wall that drops abruptly for 2,000ft to a lowland with hills only 300ft to 400ft above sea level. The range consists of earth blocks tilted westward towards the Wairarapa Plains and bounded on the east by high fault scarps or steep monoclinal flexures. The eastern lowland slopes eastwards away from the scarps. Making up the range are indurated, but locally closely jointed, greywackes and argillites, without diagnostic fossils. To the north similarly indurated rock occurs in

the lowland on the downthrown side of the tectonic scarp, but south of Boars Spur the rock east of the scarp is soft and easily eroded. It is here that the rock fans are developed. The easily eroded rock belongs chiefly to the Cretaceous Mokoiwi Formation, which occurs throughout the east coast of the North Island. The formation comprises alternating jointed bands of sandstone and soft mudstone. It is subject to intensive badland erosion, for the beds are highly fractured and offer little resistance to vertical corrasion. At this locality the Mokoiwi Formation is faulted against soft Miocene mudstones, correlated with the Bell Creek beds of Vella (1954) on the western side of the range. The fact that there has been faulting on both sides of the Mokoiwi Formation in this neighbourhood may have further contributed to crushing and fracturing of the beds. There is no scarp, between the Cretaceous and Tertiary beds because the faulting took place some time ago and the Cretaceous beds are weaker than the Tertiary rocks. Near the coast these Tertiary beds have also suffered planation, but there is no visible contrast between the two formations on the surface of the rock fans. The difference in relief between the lowland, with its weak rock, and the Haurangi Range is not due to erosion, but to the upheaval of the range in the late Pliocene and Pleistocene Kaikoura Orogeny, when the major landforms of New Zealand were blocked out. The scarps on the eastern face of the range are no longer fresh and linear, however, for they have been dissected by large graded streams that have worked headwards for more than a mile into the range. The valleys of these streams widen as they cross the belt of soft rock, but the volume of water is too great for the valley floors to show any conspicuous convexity in the transverse profile. The facets of the mountain scarp between the large streams have retreated a short distance from the fault, but are still high, steep, and roughly aligned. On each of the segments a few ravines arise, most of them with small streams that flow constantly except in an unusually dry summer. The courses of the small streams are in gullies broken by small waterfalls and rapids, and the walls of the gullies are very steep, exposing bare rock in places, as well as thin soil with scattered trees and scrub. These small streams, because they are heavily loaded with detritus from the range, have been compelled to adopt steeply graded courses across the soft rock at the foot of the range, and have planed the soft belt, each forming a rock fan. In the past, the rock fans have coalesced to form an apron of such fans interrupted only by the valleys with larger streams. Effective lateral planation* “Planation” is not used in its literal sense, but in G. K. Gilbert's sense of cutting an “even surface” which is not strictly plane, and which in this case is distinctly convex (Gilbert, 1877, p. 127). has been almost wholly confined to the weak belt, and there is thus scarcely any back-wearing of the mountain scarp, such as has been described by Johnson and other authors. The original shape of the coalescing fans is no longer preserved, for the streams which carved the rock fans have later been compelled to degrade and to incise new valleys through the first-formed rock fans. These valleys have been widened by lateral corrasion and other processes into new fan-like surfaces, leaving remnants of the previously carved rock fans as steeply sloping terraces. Entrenchment and lateral planation have recurred a number of times, so that the rock fans are now conspicuously terraced, “storied” (Cotton, 1948, p. 256), or “telescoped” (Blissenbach, 1954), a possibility envisaged by Johnson (1931). Features incised in this manner in earlier valley forms are described by French authors as emboité, anglicised to “inset” by Barbour (1933). In an example described by von Wissmann (1951) from central China, two incised rock fans make terraces bordering the lowest rock fan. In plan, the inset surfaces flare out away from the mountain scarp, or are “trumpet-shaped”, and so are compared to the Trompetentälchen of Troll (1926), eroded in glacial outwash on valley floors of the German Alps. More recently, Troll (1954, p. 291, Abt. I, p. 290) refers to Trompetentäler in a map of the floor of the Inn Valley.

Fig. 1.—View of the lowest rock fan. and tenace. east of Mt. Baton from a tenance at the coast, looking northwards. The example of diffluence seen in the foreground of Fig. 2 is at X marked on the photograph. (Photo by G. R. Stevens.) Fig. 2.—View of the high rock fan noith of Boars Spur, showing the hilrocky surface, the allavial veneer, and the marginal streams on each side of the rock fan. Note the small area of drainage for the stream that caived this rock fan. Fig. 1—Ti section of the rock fan tenace nea the sea (shownm Fig. 3) A vencei of alluvum with channelled base hes over Mokoiwi beds. reverse fault and soft Tetian beds. high-coloured in the photograph Fig. 2.—The angulan allwum of the high rock fan at Boars Spur

The map, Fig. 1, shows the belt of unconsolidated Mokoiwi rock with which the apron of rock fans almost, but not quite, coincides. The fans overlap the belt to the west on the fault-brecciated older greywacke, particularly north of Boars Spur, where the two streams that arise near the crest of the hill enclose a large rock fan. Near the coast, the Tertiary mudstones are also planed. Not all of the fan-cutting streams are shown on the map, because many are ravines too small to be mapped, particularly those east of Boars Spur. Fig. 1.—Map showing locality with rock fans (topographic maps N168, N169). The small numbers with arrows correspond approximately with the position of the observer in the following figures: 1, Fig. 2; 2, Fig. 3; 3, Plate 4, fig. 1; 4, Plate 4, fig. 2.

The northernmost rock fans and remnants lie north of Boars Spur. The largest of these is more than 200ft above the present stream bed, and is more than 350 yards in length. (Fig. 3; Plate 4, Fig. 2.) There are remnants of rock fans at several levels between Boars Spur and Whawanui Stream, but all of them are small and insignificant. They have been cut by a number of small streams flowing from the eastern flanks of Boars Spur. The valley of the Whawanui Stream is broad, with a low gradient. There are no conspicuous rock fans immediately south of Whawanui Stream, but halfway between the stream and the coast are great “plane” surfaces. These were not closely examined, as they are covered in dense scrub. Farther south is a large inset series of rock fans east of Mt. Barton. The series is more than three-quarters of a mile wide, and extends half a mile from the mountain scarp. Terraces occur at at least five different levels (Fig. 2; Plate 4, fig. 1). Not one but several small subequal gullies lie at the head of the lowest rock fan of this series. These may have carved the rock fan, but it seems more likely that the stream east of Mt. Barton, and marked as “X” on the map, was the chief agent. This stream no longer flows over the rock fan, but behind the series, against the mountain scarp; it has apparently been diverted by a subsequent stream working headwards from the Waiarakeke Stream to the south. Adjoining the seaward side of the three lowest terraces at corresponding heights are narrow remnants of rock-fan surfaces. At each level a gentle depression is formed, where the two convex surfaces meet. The apices of this southern series when projected lie above the present shore. Evidently, these are remnants of a series of inset rock fans that have been largely destroyed by marine cliffing. The stream which carved the rock fans is no longer visible: it may have been Waiarakeke or some other stream. An illustration by King (1930, PI. 78, fig. 2) shows some of the surfaces of these two series of rock fans in the middle ground. Waiarakeke Stream is a steep, rapidly flowing stream, heavily laden, but of considerable volume; it has carved a number of low-level “rock benches” (Tator, 1953) up to 100ft above the stream. Description Surface and Profile Generally the convexity of the cross-valley profile of the rock fan is pronounced. In longitudinal profile the surface is steeply concave near the scarp at its rear, with an inclination sometimes as much as 25°, passing into a progressively more gentle slope away from the scarp until the profile is little steeper than that of streams in the vicinity. There is one striking exception, that of the high rock fan north of Boars Spur, which is convex in longitudinal profile at the outer edges. The original concave profile may have been altered by landslides or solifluxion at the outer edges above the incised valley at the lower margin of the rock fan. The surfaces of even the high remnants remain fairly smooth, except for scattered zones of sometimes single sometimes interlacing shallow depressions less than 12 inches deep. These resemble stream channels on recent flood plains, and probably represent channels of the streams that formerly flowed over the surfaces. In view of the rarity of deep channels it seems likely that the streams flowed in shallow, rapidly changing, braided courses while carving the rock fans. Except beside present streams, traces of undercutting at the valley sides are obscured by soil creep. The high rock fan north of Boars Spur has hillocks and depressions with a relief of about four feet (Plate 4, fig. 1). The rock in which this surface is carved being more resistant than the Mokoiwi beds, some of the relief could be due to cores of hard rock left standing as knolls above the general level of the rock fan. On the other hand, the hillocks could be mounds of alluvium left between interlacing stream channels, or the result of modification by solifluxion.

Fig. 2.—View of the inset rock fans east of Mt. Barton, looking west. In the foreground, as arrows suggest, the lowest surface extends into two valleys, diffluence having occurred. Fig. 3.—View of the extensive rock fan north of Boars Spur. The Alluvial Veneer. The rock underlying the surface of the rock fan is normally covered with a thin veneer of alluvium, derived chiefly from the Haurangi Range. The alluvium commonly forms low indefinite escarpments on the valley sides, acting as a cap rock over the underlying Mokoiwi beds. Though extensive sections are rare, a good but weathered section of the high rock fan north of Boars Spur shows unconsolidated angular and muddy debris in a widespread discontinuous sheet up to 20ft thick (Plate 5, fig. 2). The rubble is angular, for it lies close to its source, and includes blocks up to 20 inches in diameter. The base of the detritus is irregularly undulating, due to the presence of stream channels.

A second section is exposed by badland erosion that is working into the highest of the three terraces truncated by marine erosion. The alluvium, approximately a quarter of a mile from the mountain scarp, includes well rounded boulders of grey-wacke up to 15 inches in diameter. The matrix, which is not compacted, consists of sand and mud, with plant fragments. The alluvium is about 15ft thick, and tapers in thickness laterally, thinning out against the valley sides. At the base are well defined channels, between 2ft and 4ft deep, and 2ft to 8ft wide, cut in the bedrock by the stream that shaped the rock fan. The third section lies along the banks of the incised valley in the gentle outer slopes of the lowest rock fan east of Mt. Barton, half to three-quarters of a mile from the mountain scarp. The alluvium comprises well sorted chips and partly rounded pebbles, approximately half an inch in diameter, of greywacke from the range set in an uncompacted matrix of mud and sand, together with leaves and stems of plants. An obscure fine banding is present. The valley sides, more than 40ft high, expose alluvium throughout their height, and the full thickness of the alluvium is unknown, for the base is not exposed. A thickness of alluvium that exceeds 40ft, as in this instance, is too great to be explained by deposition of debris by the fan-cutting stream during the normal course of planation (cf. Tator, 1953, p. 121). Evidently after the planation of the rock fan the stream was compelled to aggrade because of a rise in base level, the most likely cause being the rise of sea level after the last Pleistocene glaciation. No section of the upper part of this rock fan is known, but in view of the steepness of the profile it is likely that the alluvium tapers towards the apex of the rock fan to a depth of perhaps 2 or 3 feet. If the development of the rock fans has been influenced by climatic change, and by repeated oscillations of base level, as seems likely, it is probable that the lowest rock fan east of Mt. Barton is a type of most or all of the rock fans in the vicinity. Certainly the depth of 15ft to 20ft of alluvium for the two other extensive sections is excessive for ordinary stream-bed deposits laid down in the normal course of lateral planation. As these sections appear to be some distance from the former lowest part of the rock fans, the foot of these fans also may have been covered with a considerable depth of alluvium, possibly deposited in interglacials. Diffluence and Preservation. Half a mile from the apex the lowest inset rock fans east of Mt. Barton enter two valleys, on either side of a hill. A spillover must have developed where a low divide between the fan-cutting stream and a neighbouring valley was either planed away or buried by alluvium. In other examples of diffluence the fan-cutting stream has isolated a mesa-like island of an older rock fan by flowing first to one and then the other side of the island. Von Wissmann (1951) refers to such examples of diffluence as Felskegel-Bifurkation. Another possible origin for such mesa-like islands is by the action of two or three streams flowing over the one rock fan. Commonly, there is a natural depression between the mountain scarp or valley side and the rock fan, because the surface of the rock fan is strongly convex in transverse profile. In the past streamlets and rills from the valley side could have gathered in such depressions, just as described in the Uinta Mountains by Bradley (1936, pp. 193–4; 1940). Such marginal streams are named Randgerinne by von Wissmann (1951, p. 31). The marginal stream, as well as the main stream that carved the rock fan, would have responded to lowering of base level and both streams could have later carved inset surfaces below the older rock fan. The presence of marginal streams has undoubtedly helped to preserve the large high rock fan north of Boars Spur. The good preservation of this surface is in contrast to the relatively small remnants at similar heights that have survived to the south. Whereas these southern rock fans have become entrenched by a single stream flowing erratically over the surface, the high rock fan near Boars Spur is

bordered by two small streams which have become deeply incised, leaving the remainder of the surface intact and also protecting it from soil creep, land slides, rills, and (in the past), possibly solifluxion, from the hillsides above. The Origin of the Rock Fans Before discussing their origin it is necessary to know whether the rock fans are relict, inherited from the past, out of harmony with their present environment, and undergoing destruction, or whether they are still developing. The answer is that, in general, the rock fans are relict; they are now being dissected, chiefly by concentrated vertical corrasion during heavy downpours of rain, in gulches and badlands that have developed in the soft Mokoiwi beds. The badlands are developing as amphitheatre-like embayments, situated at the heads of small, frequently dry gullies, and also small trenches along the length of some larger streams. These badlands are encroaching rapidly on higher rock fans, not only the lowest ones. Birot (see Discussion, Joly, 1950, p. 125) has remarked how commonly badland erosion is developing in the same rock that has previously been subject to lateral planation. The causes of badland erosion in this neighbourhood are manifold No doubt degradation by the main streams has been a contributing factor, but not all of the streams are rejuvenated as yet; the nick may still lie downstream Some responsibility lies with the deforestation of the land by Europeans—“man has tipped the scales” “in favour of the existing climatic processes” (Peltier, 1950). The existing climatic processes favour vertical corrasion. But during glaciations these rocks appear to have been impermeable owing to frozen ground and to have been prone to mass movement that has left broadly convex hills wherever this formation lies at the surface. With the onset of warmer conditions, when the rock has regained its natural permeability, vertical corrasion has commenced, perhaps independently of entrenchment by larger streams. The recent entrenchment of the streams to a depth of 40ft at the coast is due to tectonic uplift, occurring five times, which will be described below. In the stream of the inset series east of Mt. Barton, two or three of the nick points are still more than a quarter of a mile from the apex of the rock fan. Within this quarter of a mile lateral planation has occurred very recently. The stream has meandered for about 50 yards over the rock fan to the valley side, shaving off several feet from the surface and leaving very fresh scars of anastomosing channels and low banks (Plate 4, fig. 1). It is thus evident that neither degradation nor the postglacial amelioration in climate has immediately halted lateral corrasion. But where the streams are deeply incised little or no lateral planation has occurred for some time: vertical corrasion is predominant. From this it appears that the general cessation of lateral planation has been due in part to a lowering of base level and not to change in climate alone. The Influence of Climatic Changes. Nevertheless it is likely that climate has played a part in the shaping of the rock fans. In recent years the part that changes of climate have had upon the processes of pedimentation has been emphasized. In many cases, pediments are relict landforms now undergoing dissection by vertical corrasion. In semi-arid North Morocco, for example, Wiche (1955) finds that lateral planation in the Pleistocene was restricted to the glaciations, when the production of detritus by frost action was favoured, and the annual precipitation was increased, allowing the development of close-set ravines that worked back into the wall of rock at the rear of the pediments, and of rills and sheetfloods that scoured the rock plains. In south-east Wellington the climate is at present temperate and humid, with a well-distributed rainfall exceeding 35 inches a year, and with few frosts. Falls of snow are very rare on even the highest rock fans, but the peaks of the Haurangi Range in most winters receive a few inches and the snow remains for a few days.

The climate was undoubtedly more severe during the Pleistocene glaciations, and, as in south-west Wellington (Cotton and Te Punga, 1955, A; 1955 B), solifluxion deposits occur over the Haurangi Range, particularly in the saddles and on the western slopes, and over the hills of the lowland also, though the deposits are thin and restricted in distribution. Irrespective of an increased or decreased precipitation the cold times should, theoretically, have favoured lateral planation. The amount of detritus must have increased following the thinning of vegetation and also the development of freeze-and-thaw. Hence aggradation must have been favoured, with the possible development of braided shifting courses by the stream. It is also likely that melting of ice and snow occurred in the spring, suddenly increasing the volume of the streams and thereby imparting to them much power available for lateral corrasion, and also giving rise to solifluxion, mudflows, and landslides, which helped to lower the valley sides and further increased the loads carried by the streams. Special Factors in the Formation of the Rock Fans. Recognizable rock fans are rare except in regions that have or have had a dry climate or a climate with a dry season. The reason is mainly because, where water flows abundantly and continuously, stream gradients are so gentle that such forms are inconspicuous. Nevertheless, a theory of pedimentation by the middle courses of rivers debouching from uplands has been advocated by Johnson (1931), regardless of climate. Cotton (1917: 1939) applied this concept to the Maniototo Plains, and similar planation has been reported along the eastern base of the south end of the Seaward Kaikoura Range, the north-west side of the middle Clarence valley, and the Takapau Plain at the foot of the Ruahine Range, in Hawke's Bay (Professor C. A. Cotton, pers. comm.). Otherwise, lateral planation by streams has rarely been invoked to explain other extensive plains in New Zealand, aggradation being more often assumed. (Late Tertiary pedimentation under a somewhat dry climate has been widespread in Southland, according to Mr. B. L. Wood (pers. comm., see also Wood, 1956).) For some coastal benches in New Zealand a theory of pedimentation by lateral stream planation was suggested by Johnson (1932; 1944), but these are usually explained as marine terraces. In the vicinity of the Hauraki Range several special factors, involving the relief and lithology, have permitted the development and preservation of planed surfaces. These factors are: the direction of drainage, the heavy load and steep gradient of the streams, and the softness of the rock. The lowland slopes away from the mountains, allowing the streams to cross the soft belt of rock at right angles to the length of the belt and the scarp. It is more usual in New Zealand for rivers to flow parallel to tectonic lineaments; and had the slope of the lowland been towards the range it would have encouraged a consequent fault-angle river to develop in the soft belt, which would encroach upon the mountain scarp (Cotton, 1951) and allow little development of rock fans. To maintain a steep gradient, streams must be heavily laden with coarse debris. The small streams from the Haurangi scarp are steeply graded, because the volume of water is small, and there is an abundant supply of detritus from the range. The rock on which the surfaces are carved is not resistant to stream cutting, and it has readily permitted the streams to respond to shifts of base level and changes of climate, so that episodes of lateral planation and entrenchment have been rapidly accomplished in short intervals. Eustatic Changes of Sea Level, and Tectonic Uplift. In examples (as cited in the literature) of lateral planation that has operated most effectively during cold or pluvial episodes of the Pleistocene the pediments thus developed are far from the coast, so that there is apparently no need to allow for the complication that lowering of sea level during the cold episode must have favoured degradation rather than lateral planation. In the area here discussed the rock fans

are too close to the shore to have escaped the effects of the eustatic rise and fall of sea level during the Pleistocene. The thick alluvium on the lower slopes of the lowest rock fan east of Mt. Barton was probably deposited during the last rise of sea level, after the Wisconsin glaciation. It follows that the rock surface under the alluvium must have been carved when the sea level was low, probably during that glaciation. In addition, there has been some tectonic uplift, in contrast, apparently, with many areas in the world where pedimentation has developed under conditions of tectonic calm. The coast is bordered by a raised marine platform, with five raised beaches, that extends around the south-east and south coast of Wellington Province (King, 1930). At Pukiamuri Stream, near the Haurangi Range, stream deposits the surface of which is graded to the lowest raised beach contain fossil bones of sheep, which were introduced into the country not more than 120 years ago. This beach can possibly be correlated with the lowest of the five raised beaches near Wellington, raised during an earthquake in 1855 (Cotton, 1954). Series of high, tilted, raised marine platforms with recent, Pleistocene, or Pliocene coverhead on the coasts east and west of the Haurangi Range show that tectonic upheaval has continued throughout the Pleistocene (King, 1930; Kite, 1952; Waterhouse, 1954). Tectonic upheaval has therefore been at least partly responsible for the entrenchment of the streams that carved the rock fans, and for the exposure of the alluvium in the lowest rock fan east of Mt. Barton. The effective shifts of base level have been the compromise between eustatic changes of sea level and long continued tectonic upheaval acting at short intervals. The streams that carved the rock fans must have been affected by the changes of the strand line. During the rise of sea level after the Wisconsin glaciation the sea evidently rose more rapidly than the coast was uplifted, and so caused the fan-cutting streams to aggrade. When aggradation commenced, some solifluxion and frost work may still have been supplying debris to the streams, but even in the mildness of the present climate there is an abundant supply of detritus from the range. Remains of leaves and stems of plants in the alluvium show that some woody vegetation was growing in the vicinity, but whether on the range, rock, fan, or valley side is unknown. In the meantime lateral planation was probably continuing on the higher parts of the rock fans, and the alluvium probably never encroached on to their upper reaches, thinning to feather edges because of the steepness of the profile. Aggradation ceased, and vertical incision of the streams began, when tectonic uplift prevailed over the rise of sea level. This occurred either because activity recommenced after a long stillstand, or because the rate of uplift was more constant when sea level no longer continued to rise, or to rise as rapidly. The slight fall in ocean level since the Thermal Maximum has contributed to the fall of the strand line. Clearly, tectonic uplift has not prevented lateral corrasion, but has rather increased the vertical intervals between the treads of successive surfaces of planation and has combined with falling sea level to encourage degradation. Some of the interval between terraces is perhaps to be ascribed to the fall of sea level throughout the Pleistocene suspected by Baulig (1950) and by Zeuner (1945). Synthesis One important effect of the eustatic oscillations of sea level has been the aggradation over the planed rock surface of the lower courses of the fan-cutting streams. The rock surfaces under the alluvium were cut during the time the seas were withdrawn, in glaciations. The alluviation coincides approximately with the onset of warmer conditions, when the ocean surface began to rise. When sea level rose more slowly or remained constant the influence of tectonic uplift began to make itself apparent, the base level sank, and degradation comenced. Vertical corrasion could have then continued for a long time, and indeed may have been accelerated at the

onset of the ensuing cold time, when sea level began to fall again. It is not necessary to assume that lateral corrasion ceased entirely when entrenchment began, but it gradually lost importance: each stream merely pared back the valley sides, and shaved down the bedrock of its restricted valley floor. The effect of the gradual onset of the cold climate was, first, to reduce the cover of vegetation, which would expose more rock, increase the run-off, and encourage the formation of shallow ravines and landslides. Freeze-and-thaw, probably much more active in the range than the lowland, would increase the supply of detritus, and this contribution, together with weathered rock, might be transported to the valleys by solifluxion. In spring thaws the influx of melt water into the streams would be sudden and violent, and saturation of the valley sides must have led, in such weak rock, to numerous mudflows and landslides. Such an increase in the supply of detritus, and the sudden influxes of muddy water, must have encouraged the streams to assume rapidly shifting courses, and to corrade laterally. Especially under a thinned cover of vegetation the sides of the incised streams would retreat, as a result of the mudflows, ravines, and undercutting by the streams. The ravines on the valley sides might be sufficiently numerous to effect backwearing, but if so, the pediment which would be concave upwards, must have been further planed by the streams, for no such concave pediments are visible now. Shaving of the valley floor by a braided, rapidly flowing stream, often changing course, and not too much closed in by valley sides, was probably very effective in planing the rock. It is necessary to assume that the effects of the cold climate outweighed those of the lowness of sea level and tectonic uplift; but with amelioration of climate the sea level rose. For a time this may have further assisted the streams to plane laterally, tectonic uplift balancing the rise in sea level, but eventually the streams were forced to aggrade. When the rise in sea level became slow or ceased, uplift caused dissection, and this, with some lateral corrasion, continued into the following cold period. In this fashion the storied rock fans have been formed in this particular landscape, with soft bedrock and overloaded streams, controls being tectonic uplift, eustatic rise and fall of sea level, and climatic change. It appears likely that the tread of each terrace of rock fan dates from the early part of an interglacial. If this is so the usually recognized interglacials are represented, together with additional surfaces that may be related to the earlier little-known interglacials Zeuner (1945) refers to or simply to tectonic uplift. On the other hand, a less likely possibility is that the surfaces represent shorter phases of warmth and cold within certain of the glacials or interglacials. At present, there is little information available on the absolute or the relative chronology of these rock fans, but, seeing that plant remains suitable for radiocarbon dating are available at several places, they offer a promising field for research on Pleistocene changes of sea level and the rate of tectonic uplift on the coast of south-east Wellington. Occurrences Elsewhere in New Zealand It is likely that there are rock fans elsewhere in New Zealand, although some may remain long unrecognised for want of a section to show that the alluvium on the surface is only a veneer. Observers who study rock fans in arid or semi-arid climates are more fortunate in this respect, in that the rock fans not only have less vegetation on them, but are often without the alluvial mantle Examples of rock fans can be seen in the previously glaciated areas of the Southern Alps, where streams flow from the steep valley walls across lateral moraines—e.g., down-valley from the shrunken Tasman Glacier. There, as in the examples described, steepness of gradients controlled by abundance of load makes such features conspicuous. Much more gently sloping rock fans have been carved by large streams flowing from the western side of the Haurangi Range in soft upper Tertiary deposits at the foot and on the lower slopes of the range.

Acknowledgments The field observations embodied in this paper were made in the course of a study of the geology in part of south-east Wellington, while the writer was a student of Victoria University College. The writer is grateful to Dr. M. T. Te Punga for his advice and interest during field work, to Professor C. A. Cotton for his advice in preparing the manuscript, and to Dr. C. A. Fleming for his assistance in checking the manuscript. Fig. 1 has been drafted by Mr. C. T. Webb, and Mr. G. R. Stevens has kindly allowed a photograph belonging to him to be incorporated (Plate 4. fig. 2). References Barbour, G. B., 1933. Pleistocene History of the Huangho. Bull. Geol. Soc. Am., 44: 1148–1160. Baulig, H., 1950. Essais de Géomorphologie. Paris: Les Belles Lettres. Blissenbach, E., 1954. Geology of Alluvial Fans in Semi-arid Regions. Bull. Geol. Soc. Am., 65: 175–189. Bradley, W. H., 1936. Geomorphology of the North Flank of the Uinta Mountains. U.S. Geol. Surv., Prof. Paper. 185 (1). —— 1940. Pediments and Pedestals in Miniature. Journ. Geomorph., 3: 244–255. Cotton, C. A., 1917. Block Mountains in New Zealand. Am. Journ. Sci., 44: 249–293. —— 1939. Lateral Planation in New Zealand. N.Z. Journ. Sci. Tech., B. 4: 227–232. —— 1948. Landscape as Developed by the Processes of Normal Erosion. 2nd Edit. Whitcombe and Tombs, Ltd. —— 1951. Tectonic Scarps and Fault Valleys. Bull. Geol. Soc. Am., 61: 717–768. —— 1954. Submergence in the Lower Wairau Valley. N.Z. Journ. Sci. Tech., B. 35: 364–369. Cotton, C. A. and Te Punga, M. T., 1955. A. Solifluxion and Periglacially Modified Land-forms at Wellington, New Zealand. Trans. Roy. Soc. N.Z., 82: 1001–1031. —— 1955. B. Fossil Gullies in the Wellington Landscape. N.Z. Geogr., 11: 72–75. Davis, W. M., 1901. An Excursion into Bosnia, Hercegovina and Dalmatia. Bull. Geog. Philadelphia, 3: 21–50. (See p. 37.) Gilbert, G. K., 1877. Geology of the Henry Mountains. Washington: Government Printing Office. Joly, F., 1950. Pediments et glacis d'erosion dans le Sud-Est du Maroc. C. R. Congr. Int Geog., Lisbon, 2: 110–125. King, L. C., 1930. Raised Beaches and the Features of the South-east Coast of the North Island of New Zealand. Trans. N. Z. Inst., 61: 498–523. Kite., R, 1952. Univ. of N.Z., unpublished M.Sc. thesis, lodged in Victoria University College Library, Wellington. Johnson, D. W., 1931. Planes of Lateral Corrasion. Science, 73: 174–177. —— 1932. Rock Fans of Arid Regions. Am. Journ. Sci., 23: 389–416. —— 1944. Problems of Terrace Correlation. Bull. Geol. Soc. Am., 55: 793–818. McGee, W. J., 1897. Sheet Flood Erosion. Bull. Geol. Soc. Am., 8: 87–112. Peltier, L. C., 1950. The Geographic Cycle in Periglacial Regions as it is related to Climatic Geomorphology. Ann. Assoc. Amer. Geogr., 40: 216–236. Tator, B. A., 1953. The Climatic Factor and Pedimentation. C. R. Cong. Int. Geol., Alger. 7. 121–130. Troll, C., 1926. Die jungglazialen Schotterfluren im Umkreis der deutschen Alpen. Forsch. z. Dt. Ld—u. Volksk., 24: 173, and fig. 2, p. 170. —— 1954. Ueber Alter und Bildung von Talmäandern. Erdkunde, 8: 286–302. Vella, P., 1954. Tertiary Mollusca from South-east Wairarapa. Trans. Roy. Soc. N.Z., 81: 539–555. Waterhouse, J. B., 1954. Univ. of N.Z. Unpublished M.Sc. thesis, lodged in Victoria University College Library, Wellington. Wiche., K, 1955. Fussflachen im Hohen Atlas. Sitz. Ost. Akad. Wiss., Abt. I. 164: 389–417. Wissmann, H. von, 1951. Ueber seitliche Erosion. Bonn: F. Dümmler. Wood, B. L., 1956. The Geology of the Gore Subdivision. N.Z. Geol. Surv., Bull. 53. Zeuner, F. E., 1945. The Pleistocene Period: Its Chronology and Faunal Successions. London: Ray Society Series, No. 130. J. B. Waterhouse, Sedgwick Museum, Cambridge, England.