Art. XLIII.—The Glaciation of New Zealand.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 42, 1909, Page 334

Art. XLIII.—The Glaciation of New Zealand. By P. Marshall, D.Sc., Professor of Geology, Otago University. [Read before the Otago Institute, 14th September, 1909.] It is needless at this stage of the observation of the surface features of New Zealand to insist on the fact that in the latest geological times all the present glaciers were much extended, and that numerous glaciers existed where there is now no ice. In regard to fundamentals, in recent years few additions have been made to the amount of information that was possessed by Hutton, Hector, and Haast in the early “seventies.” There has, however, been a large amount of detailed information gained in regard to several of the former glaciers in many parts of the country, and in regard to the actual extent of many of those glaciers. All those surface features due to glacial sculpture which have been recorded and described in other glaciated countries have been recognised in New Zealand, especially in the upland and western districts of Otago and Canterbury. Generally it may be stated that in the north-west of the Nelson Province signs of glaciation have not been found at a lower level than 4,000 ft. above sea-level, while in the eastern portion of this province, where the mountains do not often exceed 5,000 ft. in height, there appears to have been no important accumulation of ice whatever. Further south, the glaciers that filled the basins of the present lakes Rotoura and Rotoiti had their terminal faces at an altitude less than 2,000 ft. above sea-level. This relatively small altitude is easily explained by the situation of the lakes, for they lie between mountain-ranges, and in country that is continuously high, which afforded a large gathering-ground for snow and ice. The effects of glaciation are clearly seen in all the large river-valleys of Canterbury. High up on the mountain-sides are seen the terraces with their decided slope outwards from the mountain-axis. In all the valleys roches moutonnées are abundant, and are deeply scored with glacial marks. Smoothed rock-surfaces are to be found on every side. These features have been observed and noted by every geologist who has wandered into any of the great valleys. Marshall and Speight have advocated the idea that the glaciers which filled the valleys are also responsible for the over-deepening which these valleys have evidently undergone. In Canterbury it is recognised that the larger glaciers crept through the present gorges into the plains, and that their terminal faces were outside of the mountain regions. Such elevations as Woolshed Hill and Little Racecourse Hill are considered to be portions of the terminal moraines. Only one locality is yet known where the ice approached the east coast: this is at the Taieri, where the well-known moraine extends for some six miles north-east and at least ten miles south-west of the gorge of the river, and stretches three miles east from the plains. On the west coast the glaciers then, as now, extended to far lower levels than on the east coast. Ice filled all the profound valleys of the western sounds, and in them ice-worn surfaces of rocks, roches moutonnées, glacial cirques, hanging valleys, and all the other features of glacial erosion are to be seen in perfection.

Further north the ice extended some distance westward form the foot of the mountain-range. The terminal moraines now form the coast-line at intervals between Ross and Milford Sound, and are not far inland at Hokitika and Greymouth. These limits of glaciation have been recognised for a long time, and have been embodied in a map published by the Glaciation Research Committee of the Australasian Association for the Advancement of Science. In a recent bulletin of the New Zealand Geological Survey, No. 7, p. 25, the following statements are made: “The features of the land in the Wakatipu region, and from there southwards to the plains of Southland and eastwards to the sea, are everywhere dominated by evidences of ice erosion on a scale of magnificence that is unknown elsewhere in the Southern Hemisphere, and is perhaps without a parallel elsewhere outside of the polar regions. Each valley and depression contained its own glacier, but all were united in one, forming part of the great ice-sheet that covered the land. Its surface formed a vast plateau through which the tops of the highest peaks appeared, looking like lonely islets in a frozen sea. A striking feature of this gigantic glaciation is the remarkable freshness of the ice-grooved surfaces, the beautiful flowing contours being everywhere sharply outlined in their carpets of native grasses.” P. 40: “Evidences of a continuous ice-sheet can be traced along the main axial divide until we reach Boulder Lake, in Collingwood County. The smooth, rounded contours of the ranges between Wakatipu and the sea, the thick sheet of boulder-clay covering the Henley Hills and the area between Saddle Hill and Dunedin, and the silts and glacial till covering the Timaru and Oamaru areas afford convincing proof that the glaciers also reached the sea on the east coast. The ice-sheet appears to have reached its northern limit near Cook Strait.” P. 43: “The Henley deposit is undoubtedly of glacial origin. It is not a terminal moraine, but a true boulder-clay, the formation of which appears to have been partly subglacial and partly subaqueous. A sheet of boulder-clay of great interest covers all the hills and ridges between Dunedin and Saddle Hill. It consists of stiff yellow clay, which often contains boulders of basalt, phonolite, dolerite, and other igneous rocks.” “The question may now be asked, Was the great ice plateau that covered the South Island of New Zealand an extension of the polar ice-sheet ? It is almost certain that this must be answered in the affirmative, for whatever cause or causes originated the glacial period in New Zealand would also operate in the higher latitudes. The exploring expedition of last November proved that the Auckland and Campbell Islands had suffered erosion by an overriding sheet of ice that must have had a polar origin.” The only evidences that are offered in support of these sweeping statements in the publication are contained in the above extracts and in a detailed statement of the thickness of the ice in the Wakatipu basin. This is estimated to have been 7,400 ft. It is advisable here to state the nature of the criteria some or all of which are usually employed in defining the limits of a prehistoric ice-sheet:— (1.) A more or less continuous line of morainic deposits at the extreme limit of the ice-sheet. (2.) The occurrence of till or boulder-clay over almost the whole area that was covered by the ice-sheet.

(3.) The occurrence of roches moutonnées or rounded hills of solid rock distributed over the area. (4.) The occurrence of grooved or ice-worn surfaces of rock generally throughout the district that has been covered. Such surfaces will naturally be more numerous and conspicuous where any elevations rise in the path of the ice. (5.) Boulders of rocks foreign to the locality will be found in great numbers in positions where they cannot have been deposited by streams that formerly existed or on prehistoric sea-beaches. In the publication to which reference has been made there is no mention of any observations near the coast-line in connection with the subjects of these five headings, with the exception of No. 2. This is very noticeable, for it is explicitly stated that the evidences of glaciation are remarkably fresh even in the mountain regions, where geological processes are most active, and at the level that is described as the upper limit of the ice-sheet, which would therefore be from the first diminution of the glacial covering subject to the healing and concealing processes of nature. It is evident that the lapse of time and the effect of weathering cannot be appealed to as causing the blotting-out or disappearance of other evidence near the coastline. It is, however, advisable in a discussion of this kind to consider what observations have been made by geologists in regard to each one of these criteria. (1.) There is no continuous line of morainic material near the east coast that has yet been described by any geologist. The terminal moraine at Henley is quite isolated. It is opposite the watershed between the Clutha and Taieri basins—that is, opposite to the highest line of country—and it is therefore the most likely place for a glacier to exist if climatic conditions allowed of an accumulation of ice. It is also well known that the material of the moraine dips uniformly to the west at an angle of 15° to 20°. The inclination is constant from Henley to a point three miles to the east. Such an inclination means that the Henley side is now 5,000 ft. lower than when deposited, or that the other side is higher by that amount, or that one is higher and the other lower by 2,500 ft. than in their original position. This shows that earth-movements of some importance have taken place subsequently to the deposition of the moraine, and this may account for its present low-lying position. The materials of the moraine have often been described, and its nature is well known to geologists. Its structure is remarkably similar to the morainic material partly assorted by water through which the Tekapo and Pukaki Rivers flow when they issue from lakes of the same names. The great local thickness—800 ft. at least—and the abrupt termination in all directions show that this mass of fragmental material is not a boulder-clay—a conclusion that would naturally be drawn from an inspection of its material alone. It is only necessary to say, in regard to the latter point, that in many places for great thicknesses, especially on the landward margin of the moraine, there is no clay, but merely an un-assorted mass of large and small angular pebbles. A moraine has been described in the Leith Valley. It is a mound 8 ft. high and a few hundred yards in length. It contains no schist boulders, but some andesite is said to have been found in it. I have seen no specimen of this rock in it; neither do I know of any occurrence of andesite in the Leith Valley basin above this point. There appears to be no

special evidence in favour of the glacial origin of this material. At all events, even if a glacial origin could be proved for the material of this mound, 8 ft. by 5 chains, such an occurrence can hardly be seriously used as an argument in evidence of the existence of an ice-sheet in eastern Otago. (2.) Boulder-clays: The coastal area from Timaru to Kaitangata has been examined by numerous gelogists. Of these, Hector, Hutton, Haast, McKay, and Park have repeatedly failed to record the occurrence of any till or boulder-clay. Hutton regarded all the surface clay of South Canterbury and North Otago as a marine silt. Haast described it as a loess, and his description has been strongly supported by Hardcastle, and adopted by Heim and Speight. Park, in 1904, referred to the Oamaru loess as yellow silts. The loess of Timaru and Oamaru cannot be distinguished from the similar deposits that form a nearly continous covering over the Canterbury Plains, and extend to the height of 800 ft. on Banks Peninsula. My own microscopic examination of this material strongly supports Haast's loess theory. Not only have I seen under the direction of Mr. Hardcastle a few pebbles of small size of well-rolled material, but also well within the mass of the deposit a small amount of bog-iron ore, which was formed in small depressions of the surface on to which the dust was blown from the river-beds. Microscopic examination shows that the loess consists of minute rounded grains, mostly of quartz, and this is very characteristic of wind-blown material. The clays about Dunedin are of a very different character. In all cases the only recognisable mineral grains that they contain are those of the most resistant minerals contained in the underlying rock—a matter that at once suggests that they have been formed in situ. That this explanation is in many cases correct is proved by the occurrence of sections which show the gradual degradation of the rock into clay. One that is particularly instructive is to be seen at the top of Roslyn, where the Rattray Street tramway passes through a cutting at the top of the hill. the boulders in the clay and on its surface are in almost all cases identical in composition with the sound rock below. In those cases in which this identity of nature is not found the boulders belong to another adjacent lava-flow that lies above it; but there are very few instances of these. There are a few places at relatively low levels where there is a layer of well-rounded pebbles and boulders beneath the clay. These mark old marine shore-lines. In previous publications it has been shown that the present shore-line bears unmistakable evidence of elevation, for in many places, especially at Sea View, there are marine terraces now some 200 ft. above sea-level. The clay that covers the boulders in the localities referred to has been washed down the hillside on to them. In a typical instance at Caversham the clay that covers the boulders consists almost entirely of quartz grains and remains of Foraminifera. These are also the main constituents of the underlying rock and of the surrounding rock over a small area. Although this marine sandstone has quite a small outcrop, and is surrounded by volcanic rock, there is very little admixture of other material that could have been derived from volcanic rocks. (3.) Roches moutonnées: No geologist has yet pointed to or described in any scientific publication any roches moutonnées near the coast-line. Until such features have been described and the descriptions discussed, it

must be accepted that there is no positive evidence in regard to this. No appeal can be made to the blotting effect of weathering if the statement is true that “a striking feature of this great glaciation is the remarkable freshness of the ice-grooved surfaces.” In the coastal ranges near the Nuggets the development of the many subsequent or strike stream-valleys, eroded to a great depth in hard rocks, is an undoubted proof that those hills have been submitted to uninterrupted stream erosion for a great lapse of time. I have failed to see any effect of glacial erosion in the coastal hills near Dunedin, Timaru, Oamaru, Taieri Mouth, or between the Nuggets and Waikawa, or in the Hokonui Hills. Mr. Hardcastle has cited the north-west face of Mount Horrible, near Timaru, as an example of glacial erosion. The form of this slope and that of the valley of which it forms one of the sides are, however, quite different from those found in glacial areas, and appear to me to be those of a typical valley and escarpment formed by stream erosion. The typical glacial topography is unknown in the coastal regions, and has not been described in two Geological Survey Bulletins dealing with Central Otago.* Since the above was written, Saddle Hill, Mount Watkins, and other points have been called roches moutonnées. The absence of polished and grooved rock-surfaces, and the large number of loose boulders on the summits, appear sufficient to disprove this. These hills are volcanic plugs. (4.) Glacial pavements: There have been no descriptions of ice-striated surfaces of rock or of boulders. Hutton mentions some that had been stated to be ice-worn. He was, however, able to convince contemporary opinion that the grooves were not due to ice, but to unequal weathering of the rock. Actual experiments with artificially polished rock-surfaces of the same kind of rock as that which he mentions have convinced me that these irregular, curved, and branching grooves are produced by chemical reagents which decompose certain of the rock-constituents. (5.) Erratic boulders: No foreign boulders have yet been discovered on the Dunedin hills or elsewhere on the coast. It is absolutely impossible that boulders of schist should be entirely absent if the country had been covered by an ice-sheet that radiated from Wakatipu or from some other point on the main divide, for the whole country is formed of schist to within a very few miles of Dunedin. The importance of this is realised when the results of the late Antarctic expeditions are recalled. In South Victoria Land the coastal country round the bases of the great volcanic cones and well up on their flanks is strewn with granite blocks and other rocks brought apparently from great distances. At the Gaussberg, where the ice completely covers the mainland, so that all the coast and even the mountaintops are invisible, it is still found that the Gaussberg itself is strewn with boulders of much variety brought from the hidden continent. Similar results were obtained in Graham Land by Arctowski. The Swedish expedition in Louis Philippe Land also found numbers of erratic blocks in that ice-bound area. Much emphasis has lately been laid upon the thickness of the ice in the Lake Wakatipu basin,† Bulletin No. 7, N.Z Geological Survey (New Series), p. 36. and upon its effect in removing and transporting material. In general it is well to acknowledge the great extent to which these descriptions have added to our knowledge of the details of glaciation of the district. There are a few matters to which exception may be taken. It is perhaps unfortunate that the Arrow basin should be called a “cirque,” for

this term is almost restricted to the amphitheatre at the head of a glacial valley. The Arrow basin is at the junction of at least two glacial valleys of considerable length. The lateral thrust of the Greenstone Glacier is a difficult matter to understand.* Bulletin No. 7, N.Z. Geological Survey (New Series), p. 31. The fact that the Greenstone River flows for five miles of its course through a profound narrow gorge before it joins the Caples appears to be a sufficient proof that the ice of the Greenstone Glacier never passed into the Wakatipu basin. Such a conclusion is strongly supported when it is recognised that the upper Greenstone Valley has a direct continuation to the Mararoa over a pass formed of stratified gravels only a few hundred feet high. Large moraines exist at the lower end of the Mavora Lake. The ice from the Caples undoubtedly joined the Lake Wakatipu Glacier and deposited morainic matter at Rere Lake, but it did not join it at right angles, and cannot have exerted the thrust which is stated to have produced such important effects on the opposite slopes of the Richardson Mountains. Such a thrust is hard to understand, for it is known that the ultimate strength of ice is so small that it would be reduced to the physical state of a fluid long before the pressure was sufficient to bend or break the least-resistant schist. The described effect of the Von Glacier is equally difficult to follow, for there is a low pass from this valley to the Oreti, and it is certain that the Wakatipu ice, if it reached the thickness, would have flowed over into the Oreti, for it is stated that it “strode over” Mount Nicholas, which is 4,827 ft. above sea-level. There was certainly no opposing mass of ice of great magnitude to prevent an overflow into the Oreti Valley. At the present time it is rash to make ass rtions as to the source of the boulders found in moraines, for nothing is at present known of the structure of the country at the head of the Dart or even of the Shotover. In the case of the former, at least, it is quite possible that gabbros and other plutonics may exist, for they are known on the western side quite close to the watershed. It is still more likely that greywacke rocks may yet be found in many localities within the Dart watershed. The actual thickness of the ice is stated to have been 7,490 ft. near the Kingston end of Lake Wakatipu. Judging from analogy with the surface of the Greenland ice-sheet, the slope of the upper surface of the Wakatipu Glacier, even if it is regarded as an ice-sheet, cannot have been less than 90 ft. per mile to the west of this point. Mount Crichton is about twentyfive miles away from it in the direction of the flow of the ice, and the surface of the ice there would be 2,250 ft. higher. In other words, the ice-surface would be 9,950 ft. above sea-level. As Mount Crichton is only 6,185 ft. high, the ice would have been 3,765 ft. above its summit. To any one who has ascended the hills around Lake Wakatipu it is evident that ice has never flowed over their tops. One of the first effects of the movement of an ice covering is to remove all loose blocks and boulders; but when it melts it will leave behind a few foreign boulders which it has carried with it from elsewhere. The photograph on page 80,† Loc. cit, p. 80. of Mount Aurum, which is 7,315 ft. high, shows that it has not been covered by a moving sheet of ice. Every tourist knows of the litter of loose blocks on the summit of Ben Lomond, which is 5,967 ft. high. I have found the same thing on the top of the Livingstone Mountains, and on that of Mount Dick, 6,021 ft., which rises above the point where Lake Wakatipu

begins to shallow. The only summit mentioned in the report,* Bulletin No. 7, N.Z. Geological Survey (New Series) p. 31. apart from the Hector Mountains, as showing ice erosion is Mount Nicholas, 4,827 ft. high. I have personally seen no effect of ice erosion due to a moving ice-stream at a greater height than 4,000 ft. on Mount Dick or 5,000 ft. on Ben Lomond. There is here evidently a difference of opinion as to facts; yet the aspect of the Tooth Peaks, near the Greenstone; of the Bayonet Peaks, 5,213 ft., at Halfway Bay; of the Cecil Peak, 6,477 ft.; of the Walter Peak, 5,946 ft., is in every case sufficient to prove to a geologist that none of these peaks have been covered by ice. If we proceed still further into the hinterland there is everywhere the distinct difference between the glaciated topography of the lower heights and that of the unglaciated topography above, as is most clearly seen in the Routeburn, the head of which is situated on the main watershed, where the ice must have had its greatest thickness. The statements that have been made show that a maximum thickness of 5,240 ft. of ice is all that can be allowed for the great Wakatipu Glacier. Even this thickness will probably appear surprising at first, and may even seem to justify the statement that an ice-sheet overspread the lowlands. It is therefore as well to compare this with the probable thickness of the ice of our present existing glaciers. In the first place it must be recognised that when an ice-sheet descends to a low level or has a great thickness there is no reason to conclude that all the country, or even adjacent valleys, are similarly affected. The Tasman Glacier has its terminal face at present 2,358 ft. above sea-level, and the terminal face of the Franz Josef Glacier is 697 ft. above sea-level. Yet there are in each case neighbouring valleys either without ice or with only small glaciers at their heads. On the western side the Douglas River rises to over 4,500 ft. before it disappears beneath the ice of the Marchant Glacier; on the east side the Murchison rises to 3,308 ft. and the Cass to 5,209 ft. before the glacier is reached; while further south, on both sides of the main range, there are scores of valleys rising with gently sloping floors to 3,500 ft., and yet they are innocent of glaciers. The mere mention of these facts shows that it is obvious that the height of the valley-floor above sea-level does not settle the question as to whether the valley will be occupied by a glacier. It is also obvious that the existence of long glaciers in some valleys does imply that other neighbouring valleys will be filled with ice. It is an undoubted fact that the size of the névé, the slope of the valley-floor, the amount of snowfall, are all characters that have great importance in relation to the size of glaciers. It is seldom realised that the ice in our present glaciers is of great thickness. Actual estimates based on measurements in New Zealand are not at present available. It is, however, probable that we shall not err gravely if we adapt measurements that have been made in Switzerland to the slightly different dimensions of our own glaciers. Of all the Swiss glaciers the Aar appears to be most similar to the Tasman; but its length is only ten miles, while that of the Tasman is 17.5 miles. The thickness of the ice in the Aar Glacier is 1,312 ft. (400 metres: De Lapparent). If the thickness of the Tasman has the same ratio to its length as that of the Aar, it must be 2,296 ft.

It is, however, possible to make an independent estimate of the thickness of the ice. It is well known that the ice of a glacier decreases somewhat rapidly in thickness towards its lower end. This is a natural result of the fact that at the low altitudes the relatively high temperature melts it. The rate of melting in Switzerland is between 19.7 ft. and 26.2 ft. per annum. It is probable that the rate is somewhat higher in New Zealand, for the latitude is lower, and in consequence the rays of the sun are stronger and the amount of summer rainfall is higher—and it is known that the rainfall is perhaps the most important agent in melting the ice. The rate of ablation, or of lowering of the ice-surface, mainly due to melting, is in Switzerland 19 ft. to 25 ft. per annum (De Lapparent). Here it is certain that the rate of ablation will not be overestimated at 20 ft. per annum. The rate of flow of the ice of the Tasman Glacier is 12 in. per annum at the Malte Brun spur—say, opposite the De la Beche bivouac. This was measured in summer; and in glaciology the summer rate is supposed to be twice that of the winter rate. The average throughout the year at this point is, then, 9 in. per day. The Mueller Glacier 40 chains from its terminal face moves at the rate of 2.9 in. per day on an average throughout the year. We shall not underestimate the rate of flow between the De la Beche bivouac and the terminal face of the Tasman Glacier at 6 in. per day on an average throughout the year. In other words, the ice travels over its bed in this part of its course a distance of 182.5 ft. per year. If the ablation is supposed to commence to be in excess of accumulation at the De la Beche bivouac, and to amount to 20 ft. per annum at the terminal face, the average rate of ablation between these points will be 10 ft. per annum. If the onward movement is 182.5 ft. per annum, and the terminal face is eleven miles distant, the ice at the De la Beche bivouac will take 5280/1825 × 11, or 311.3 years, to reach the terminal face. During this time its thickness will decrease by 311.3 × 10 ft., or 3,113 ft. In other words, this method shows us that the ice of the Tasman Glacier is 3,113 ft. thick at the De la Beche bivouac. Since the surface of the ice is at this point 4,782 ft. above sea-level, its lower surface must on this estimate be 1,669 ft. above sea-level, or 689 ft. lower than the valley-floor at the terminal face. Or, expressed differently, the Tasman ice must lie in a trough-shaped basin similar to that of Lake Wakatipu. Whichever of these estimates of the ice-thickness is taken—2,296 ft. or 3,113 ft.—it is, at any rate, evident that the thickness of the ice in the Tasman Valley is very considerable; but it is well known that it has little or no effect on the climate of the surrounding country, not to mention New Zealand as a whole. Viewed in the light of these facts of the low level and thickness of ice in a valley of the present day, the statements as to the low level and thickness of the ice in the Wakatipu Valley during the glacial extension becomes less impressive. Whatever may be the origin of the Wakatipu basin, it is, at any rate, evident that its form has not been changed by any faulting since the ice that once filled it disappeared. In the absence of definite information we are justified in assuming that the terraces round it are actually horizontal, and this would almost certainly not be the case if faulting had occurred. Soundings show that the floor of the lake is so destitute of irregularities that no fault scarp can cross it. The reversed slope of its lower end must, then, have existed when the basin was filled with ice.

There is also no question that the main stream of ice flowed towards Athol. In order that there should have been any flow in that direction, the upper surface of the ice must have sloped towards Athol. The Tasman Glacier has a slope of 229 ft. per mile in its flattest part. The Greenland ice-sheet has a surface slope of 220 ft. per mile for the first 3,281 ft. of rise on its eastern side, and 93 ft. per mile for the next 3,281 ft., according to Nansen. If we take the slope of the Wakatipu ice at 100 ft. per mile above Kingston, and 300 ft. per mile from there onwards, we are not likely to err much. We will suppose the thickness to be 7,490 ft. at a distance of five miles from Kingston, where the upward slope of the lake-floor begins. At Kingston, five miles away, the thickness of the ice would be 1,265 ft. less owing to the rise of the lake-floor, and an additional 500 ft. less owing to the downward slope of the surface of the ice. That is, at Kingston the thickness will be reduced by 1,765 ft., and will therefore be 5,725 ft. This reduction in thickness must be due to melting, and it amounts to 388 ft. per mile. There is no reason to suppose that the rate of melting would decrease below Kingston, and, if it did not increase, the terminal face of the ice would be 5,725 × 388, or fifteen miles away—that is, at Athol. Here the terminal moraine is actually situated. This estimate is based on the supposition that the reversed slope had then its present form. It is quite possible that it owes its steepness to the deposition of morainic matter which took place when the ice began slowly to retreat. If so, the reversed slope must be supposed to extend from the floor of the lake opposite Mount Dick to the Dome Pass, where the rock-surface is 1,300 ft. above sea-level. The end of the slope at Athol will then be 1,565 ft. above its commencement twenty miles away. If an average slope of surface of 150 ft. per mile be adopted, 3,000 ft. of thickness is accounted for within this distance. The reduction in thickness is therefore 1,565 ft.+3,000 ft. at this point, which leaves 2,925 ft.; and with a slope of 150 ft. per mile the terminal face would be near Lumsden. At this place there is not a terminal moraine, nor is any known between it and the sea-coast. With my estimate of 4,000 ft. of ice, the terminal face would, under similar conditions of slope, be at Athol. The result of this inquiry is, at any rate, of such a nature as to prove emphatically that, with such a great thickness of ice in a valley glacier of enormous dimensions, all comparison with an ice-sheet is fallacious. The statement that the imaginary ice-sheet of New Zealand was connected with another ice-sheet that simultaneously extended northwards from South Victoria Land is not at present supported by any evidence. At first sight it appears to be reasonable to assume that a glacial advance in New Zealand must have been due to some cause that would have affected lands in the south in a similar manner. This, however, is a conclusion that we are not justified in adopting. If a comparison is made with North America, we find that at the time that ice advanced to a latitude of 37° 30′ in the Mississippi Valley, where now a mild, temperate climate is experienced, Alaska was largely free from ice except on or about the mountains, though now it experiences a climate of great severity. To take another instance, Europe during the period of maximum ice-advance was in the west covered by an ice-sheet as far south as the Thames Valley in England, and to the parallel of 50° north latitude in Poland. At the same time, in northern Siberia, at Werchojansk, where the mean annual temperature is now 1.4° Fahr., and the January temperature is—54.4° Fahr., the surface of the land was not covered with ice.

American geologists at the present time believe that the glacial advance took place at different times in different parts of the glacial region. It was earliest in the west, spread eastwards, and is now at its maximum in Greenland. Another wholly different line of reasoning will naturally make us hesitate to adopt the idea that a reduction of temperature in New Zealand, associated with a similar reduction in South Victoria Land, would cause an ice-sheet from the former to extend northwards to the latter. Captain Scott, as a result of the “Discovery” expedition, recognised that the Great Barrier during the greater glaciation in the south extended over the whole of the Ross Sea and this is the greatest ice-extension in the south that has yet been suggested by any of those who have visited these regions. The Ross Sea, it must be remembered, is a relatively shallow area of water, for only two soundings have been made that indicate a greater depth than 300 fathoms. Captain Scott states that at the time of this ice-extension the mean temperature must have been higher than to-day. This is accepted as a “natural conclusion” by Philippi, and is also adopted by David. It therefore appears that a decrease of mean temperature in New Zealand was not necessarily associated with a lower temperature in South Victoria Land; and, even if there was this association, a glacial advance would probably not result. There is at present no known instance of a continental ice-sheet extending far over deep oceanic water. That of Greenland stops short of the coast everywhere except at the heads of some of the fiords. In South Victoria Land the Great Barrier terminates in water that is almost everywhere less than 300 fathoms. The “Belgica” expedition, in longitude 90°, as it approached the edge of the ice got frequent soundings less than 300 fathoms. Bruce, as he approached Coats Land, got two soundings less than 200 fathoms. The German expedition in the neighbourhood of Kaiser Wilhelm II Land found similarly shallow water: in two cases the depth was less than 200 fathoms. Thus, even in this greatest known of ice-sheets the northern face of the ice in no place where soundings have been made extends into deep water—in fact, it nowhere extends beyond the shallowest fringe. The importance of this is realised when it is stated that, so far as known, 1,400 miles of water more than 1,500 fathoms deep separates the Ross Sea from New Zealand. It is also a remarkable fact that the ice seldom extends from island to mainland except where the water is quite shallow. There are many instances of this in the south. The Palmer Archipelago, Coulman Island, the Balleny Islands, are all examples. In the north, Spitzbergen is separate from Greenland, while its islands and those of Franz Josef Land are separate from one another. Still more remarkable, Grinnell Land is not united to northern Greenland, though the strait between them is extremely narrow and shallow. In Europe, the Scandinavian ice-sheet extended to Britain during the Pleistocene glaciation. As the water of the North Sea is almost everywhere less than 100 fathoms deep, this cannot be quoted as an instance of the extension of an ice-sheet across deep water, especially as geologists agree that at that time the whole region was more elevated. A consideration of these facts shows us that remarkably strong proof is necessary before we are justified in countenancing the idea that New Zealand and South Victoria Land were in the past united by an ice-sheet. No geological evidence of this connection has yet been put forward, unless we accept the statements of the thickness of ice at Wakatipu as an

evidence. Even if we accept the statements previously quoted in regard to this, it is known that there are no surface-features that show that this ice ever extended to the coast. The stated union of this supposed mass of ice with an imaginary extension of the South Victoria ice is nothing more than a guess. If evidence is wanted in its favour, it must be sought mainly on the islands between New Zealand and South Victoria Land. If the ice-sheet extended northwards as suggested, it must have passed over these islands, and on their surface should be found the usual effects of a moving sheet of ice. It is also reasonable to expect to find similar effects of a southern ice-sheet on the low-lying parts of the southern and eastern coastline of New Zealand. In 1907 an expedition from New Zealand visited these islands with the object of studying all their natural features. The geologists attached to the expedition paid special attention to the evidence that the surface features of the islands offered in favour of or against the idea of a period of glaciation. The decision was quite unanimous in regard to this matter. All were agreed that there was every indication that glaciers had flowed out radially from the higher lands, and had eroded the deep valleys that now form secure harbours. Agreement was quite as unanimous that the islands showed no indications that they had been subject to the erosion of a sheet of ice of polar origin: there are no erratics, or perched blocks, or rocks of foreign origin; there are no ice-worn surfaces on the hill-tops; there was no development of that peculiar topography characteristic of the work of an ice-sheet. The whole aspect of these islands was considered to negative the idea that a polar ice-sheet had passed over them. There is absolutely no evidence that an ice-sheet from the south reached New Zealand: there are no rocks from that land described; there are no moraines on the lowlying coastal lands; there are no ice-worn surfaces near the coast; there are no characteristic features of glacial erosion near the southern coast. There is yet another line of reasoning, widely different from any of the preceding. It depends upon the well-known peculiarities of the fauna and flora of the outlying islands in the south. The amount of material available in connection with this matter is very great, and in this paper it is impossible to do more than state the general conclusions. Many species of birds belonging to genera well known in New Zealand and not characteristic of ts colder districts have developed peculiar modifications that indicate that the different island groups have been isolated for a long period, and during this isolation have not been subjected to a specially cold climate. Exactly similar results have been arrived at from a study of the plants; but in this case there are genera that are unknown outside the islands. Geologically interpreted, these facts mean that any elevation that has affected these islands since at least the Pliocene period must have been very slight, for an elevation of less than 1,000 ft. would connect the Auckland Islands with New Zealand, and the striking peculiarities of its animals and plants would, if such an elevation had taken place, have entirely disappeared. Glaciation of the North Island. In a paper read at the last meeting of this Institute a statement was made about glaciation in the Ruapehu region. The evidence offered in proof of this glaciation consisted in the occurrence of supposed boulder-clay between Utiku and Waiouru, and the so-called glacial topography on the hills near Taihape.

In regard to the topography, the statement was made that the form of the surface features has only lately been revealed, owing to the effect of the advancement of settlement, and the consequent destruction of forest which previously concealed it. This is certainly not true of the country between Turangarere and Ruapehu, a distance of fifteen miles, which has always been open grass land, and must have been crossed by any supposed glacier which extended from Ruapehu to Utiku. Yet this country has been visited by numerous geologists, all of whom have failed to record any indication of glacial topography. The last of these, Mr. Speight, in an article on the geology of the Tongariro National Park, makes no mention of any glacial features on Ruapehu or the surrounding country outside the area of the present glaciers. The country between Ruapehu and Waiouru consists near Waiouru of river-gravels covered with a thin deposit of volcanic matter, and nearer to the mountain lava-flows take its place. Southwards from Waiouru, hills for the most part flat-topped, and all formed of Cainozoic rocks of late Miocene or Pliocene age, extend to the limits of the supposed glaciation. Near Waiouru the flat-topped hills have a stratum of hard shell-limestone forming the summit. This dips south-eastward 10°, and on its north-westward face there is a steep scarp such as is normally formed in such country by stream-action. I have carefully examined the surface of two of these scarps, and I can assert that they bear no signs whatever that they have been submitted to glacial action. The topography of the land near Taihape does not in any way differ from that of country of a similar age and structure in the neighbourhood of river-valleys in many other districts. In many places in the Hautapu Valley and other river-valleys hillocks of many different sizes and altitudes are covered with a deposit that is at times well-stratified gravels with boulders of many sizes, and in other places finer and more clay-like deposits, still with some boulders in them. These are the “boulder-clays” of the paper referred to. Road-cuttings outside of the Hautapu Valley and other river-valleys do not show these gravels or clays. The boulders in the deposits of the Hautapu Valley consist, except for a few fragments of Cainozoic rock, entirely of andesitic blocks, and doubtless come from Ruapehu. My observations showed that the blocks are not markedly angular, and no striated boulders have yet been found in the deposits. The occurrence of these gravels and boulders is easily accounted for The whole of this part of the land has been elevated between 2,000 ft. and 4,000 ft. within the latest Cainozoic times. During the movement of elevation the rivers have been filing their way downwards. When the upward movement temporarily ceased or was relatively slight, the rivers widened their beds, and deposited gravels or finer matter over their flood-plains. Volcanic action being them in progress, eruptions provided fine and coarse matter in a most irregular mixture, as at Martinique and St. Vincent. When the upward movement again became more rapid they filed their beds down; and during another period of little movement the old flood-plain was dissected until in some places no trace remains of it except a few isolated hillocks. This perfectly normal action is in all countries known to account for the occurrence of gravels and other river-deposits at many different levels in the same valley. It has been stated that the boulders referred to are larger at Utiku than they are nearer to Ruapehu. My observations certainly do not agree with this. I saw larger boulders at Turangarere than elsewhere; and accurate comparative measurements are certainly necessary to establish the reverse statement.

The occurrence of the largest boulders at Utiku proved too much for the glacial theory, for the material was described as subglacial, and hence would be subject to wear which would be greater the further it was moved. It may be objected that the source of the Hautapu is not at present to be found in Ruapehu. While this is true, it is a fact that the source of one branch, the Waiouru, is in the volcanic gravels derived from the mountain. At any rate, this objection is more apparent than real. It must be obvious that when the elevation of this area began, the streams from Ruapehu would flow radially outwards as they now do from Mount Egmont. As elevation and erosion continued, interference of stream-courses Would result. One effect of this is to be found in the present course of the Wangaehu, which now takes nearly all the western drainage of Ruapehu. The air gap at Waiouru, through the Cainozoic country, may be cited in proof of this, as well as the similar air gap at the head of the Mangaio and of the Turakina. The mountain-ranges of the North Island have been studied by a number of geologists from time to time, but none of them have recorded any instances of the erosive effects of glaciation. The present writer is familiar with the features of the following mountain-ranges—Rimutaka, Tararua, Ruahine, Kaimanawa—as well as the volcanoes Ruapehu, Ngauruhoe, Tongariro, and Egmont. As he has made many expeditions into the glaciated districts of Otago, Westland, and Canterbury, he is acquainted with the glacial features of mountain-ranges, and can assert, as has been done by others, that the North Island mountains have not been glaciated, or have not supported glaciers of any size at any time in the later geological periods. Neither in the paper to which reference has been made nor in any other publication has there been any mention of glacial valleys, ice-worn surfaces, erratic blocks, or moraines. In the absence of these there is ample justification for geologists to refuse to accept the statement of the past existence of a huge glacier extending over forty-five miles of country to explain the formation of a thin deposit of clay with boulders in it, and of gravels, the occurrence of which can be readily explained by a simple appeal to ordinary geological action. Summary and Conclusions. A. That an ice-sheet did not reach the east coast is proved by— (1.) The absence of moraines near the coast, except at the landward end of the Taieri Gorge, where there have been important earth-movements. (2.) The absence of till and boulder-clay. The substances that have been claimed as being of this material are— (a.) Loess in Canterbury and North Otago; (b.) Residual clays from volcanic rocks at Dunedin; (c.) Moraine and valley-trains at the Taieri; (d.) Old sea-beaches at Caversham and at Green Island. (3.) Absence of roches mountonnées near the coast-line. (4.) Absence of striated glacial pavements near the coast-line. (5.) Absence of erratics and striated boulders. (6.) Absence of a glacial topography in all coastal districts. B. The ice of the Wakatipu Glacier:— (1.) The occurrence of thick ice in one valley does not justify the conclusion that the ice extended over the surrounding country.

(2.) The thickness of the ice in the ancient Wakatipu Glacier appears less impressive when it is realised that the ice in the Tasman Glacier is at the present day 2,000 ft. thick, and that its floor is only 1,600 ft. above sea-level. (3.) If the ice was 7,940 ft. thick at the Hector Mountains, and if it had the same surface-slope as the Greenland ice-sheet, it must have extended far over the tops of all the western mountains. These, however, show no signs of having been covered by moving ice. (4.) The reversed slope of the Wakatipu basin must have existed during the glaciation. This implies a rapid thinning, which must have been due to surface-melting. Adopting the surface-slope of the eastern side of the Greenland ice-sheet, the terminal face cannot have been east of Lumsden. (5.) The author's estimate of the level of the surface of the ice is 4,000 ft. at Mount Dick. The terminal face would then be at Athol. (6.) No ice erosion has been described in the Hokonui Hills, nor any moraines or erratics in the Waimea Valley. C. The suggested extension of the Antarctic ice-sheet:— (1.) At the present time the ice extends nowhere into deep water. (2.) There was no ice-sheet in Siberia when there was an ice-sheet in central Europe. (3.) There was no ice-sheet in Alaska when the Mississippi Valley was glaciated. (4.) The glaciated portion of North America was not covered by ice simultaneously in all its parts. These facts show the worthlessness of general statements in regard to glaciation of southern lands. (5.) There is an absence of all effects of the erosion of an ice-sheet at the Auckland and Campbell Islands. (6.) No erratic blocks from South Victoria Land have yet been found in New Zealand. D. The suggested glaciation of the North Island is disproved by the following:— (1.) The mapping of the river-courses shows no indication of the existence of a glacial topography. (2.) No striated rock-surfaces. (3.) No glacial moraines. (4.) No erratic blocks with striations. (5.) No roches moutonnées. (6.) The boulder-clays referred to are deposits laid down by rivers that flowed from active volcanoes during periods of activity. (7.) The deposits at different heights are due to rivers; for, as the land was gradually raised, the base-level of erosion would lie deeper and deeper in the rocks, and portions of the different base-levels would be left at different heights. (8.) No valleys with the characteristic forms of those due to glacial erosion have yet been described on Ruapehu. (9.) No glacial valleys have been described in other parts of the North Island.

In general the writer sees no reason to depart from the opinion long held by all New Zealand geologists that the amount of Pleistocene glaciation of New Zealand did not attain to the magnitude of an ice-sheet. On the western side of the South Island the ancient glaciers reached the coast-line in many places. On the eastern side they threaded far through the mountain-valley towards the coast. This result happens to be in harmony with the descriptions that have been written in regard to the glaciation of Tasmania and Australia.