Art. XLVIII.—Origin of the Loess Deposit of the Timaru Plateau.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 22, 1889, Page 406

Art. XLVIII.—Origin of the Loess Deposit of the Timaru Plateau. By J. Haudcastle. [Read before the Philosophical Institute of Canterbury, 5th September, 1889.] So far as I am aware, three varying opinions have been published respecting the origin of the clay, silt, or loess which covers the dolerite sheet in the Timaru plateau. The late Dr. von Haast (“Geology of Canterbury and Westland,” p.367) adopted Richthofen's theory, that it is a sub-aerial formation, which has grown up under existing conditions, and is still growing. Professor Hutton raised some strong objections to that explanation, and (“Trans.,” xv., 416) concludes (with special reference to an admittedly similar deposit on the flanks of Banks Peninsula) that “the evidence in favour of the marine origin of this deposit preponderates enormously over the evidence in favour of subaerial origin.” Lastly, Mr. J. Goodall (“Trans.,” xix., 457) declares the Timaru loess to be a volcanic ash. The object of this paper is to state some evidence which proves that the Timaru loess, and presumably also the similar formation on Banks Peninsula, and others elsewhere, is of subaerial origin; that it is a formation of wind-borne dust, entrapped by successive generations of dry-land vegetation; that the whole deposit from base to summit, inch by inch, line byline, film by film, has successively been a dry-land surface; that it accumulated not only slowly, but intermittently, with prolonged periods of pause; and that its growth was dependent upon a set of climatic conditions which no longer prevail in the neighbourhood. For the production of a massive subaerial formation of dust four factors are required—(1) a source of wind-borne dust, (2) winds to transport the dust, (3) vegetation to entrap it, and (4) sufficient time for its accumulation. In respect of this formation, the first of these factors must be indicated, but need not be located. The second will be granted; also the fourth. The third will be admitted if it is proved that the deposit was formed on dry land. I leave the indicating of the source of dust for the present, proceeding first to prove that the loess is a dry-land formation. The Material.—The material of the deposit is remarkably homogeneous in character, and is almost wholly fine enough to be easily taken up and transported by winds. The exceptions are a few elements which could not have been wind-

borne, and these—except, here and there, at low levels, scattered pebbles and boulders of scoriaceous dolerite of local origin—are among the evidences in support of the dust-heap theory. A Buried Water-hole.—Among the numerous excavations made in the loess at Timaru, the most instructive as to the origin of the formation is one made to the southward of the railway-station to form a site for the engine-sheds. The section left by the excavation, say 20ft. deep (all numbers in this paper are rough guesses), shows at the surface the remains of a swampy hollow—one of scores, perhaps hundreds, scattered over the Timaru Downs. Some of these hollows contain pools of water after heavy rains in summer, and continuously through the winter; larger ones are never dry, and contain a growth of peaty vegetation. The hollow cut through by the excavation referred to was of the latter class. It once contained a growth of peat, which had been set on fire in the early days of the settlement, or possibly earlier, by aborigines, and the light ashes now form a layer a few inches thick beneath an ordinary black soil. The clay beneath this old pool is distinctly stained blue by phosphate of iron, through the action of the percolating swamp-water upon the iron contained in the clay. This coloration extends downwards for several feet, and gradually fades out. Rather more than halfway down the face of the cutting the section shows the remains of a similar but smaller water-hole, with a little carbonized vegetable matter lying upon the uppermost of several thin layers of silt which were successively deposited in the hollow, and are marked off by lines of peaty stain; and there is a repetition of the blue phosphatic stain in the clay beneath, the stain extending downwards below the base of the section. In this buried water-hole we have proof positive of the existence of a dry-land surface during its formation and existence. Tracing it outwards, the bed merges insensibly into the general homogeneous mass of the formation, but into a layer of it marked by a character to be described in the next paragraph. Subsoil Iron-ore Granules.—In every railway-cutting and cliff exposed to rainfall there are to be distinctly seen several bands, lft. or 2ft. in depth, curving in the spurs so as to be roughly parallel with the present land-surface. These bands project slightly from the general slope of the cuttings, and are also a little darker in colour. Examination shows that the darker colour of these bands, and their relatively greater power of resisting erosion by rain, are due to the presence of numbers of granules of an oxide of iron—precisely such granules as are to be found in the subsoils of damp lands—numerous and large in the subsoil on the margins of the surface peat-

pools, and very sparsely scattered through the present subsoil elsewhere. Similar granules are to be found scattered thinly throughout the mass of the loess, but they are distinctly numerous in these bands. These granules must have taken some time to form, and other evidence is forthcoming in proof that the layers in which they occur respectively formed a soil and subsoil for very long periods of time. It is to be inferred, too, that the ore-granules indicate the prevalence of a very wet climate during the whole or some portion of each of these periods, as such granules do not appear to be formed except in wet soils. The climates must have been very wet, as the curvature of the granule-bands is such that the land-surface which each represents was well drained. The buried peat-bed already described is easily traced outwards into one of these granule-bands. Humus-stains.—The vestiges of the successive generations of vegetation have been almost but not wholly obliterated. It would seem that the growth of the deposit was so slow as to nearly allow the rootlets of each generation of plants to suck up the last remnant of the decay of previous ones. Nearly, not quite. Whenever a cutting is newly made, or a fresh face is formed on a sea-cliff, there is observable from top to bottom a brownish stain in the clay, which is intensified a little in certain bands. This stain disappears after a short exposure to the air, the surface soon assuming a bright-yellow “clay” colour. The uppermost of these dark bands, which varies in position from 5ft. to 8ft. or so beneath the present soil (and in some places also the second band, some feet lower down), shows in old “backs,” or natural crevices, abundant stains caused by the decay of rootlets, the ramifications of which can be easily traced. This evidence, taken alone, would not be of much value, as recent rootlets—very few in number, however—penetrate these cracks. The obscure brown stains in deeper bands, to which recent rootlets do not extend, are evidently of the same character as the more strongly marked upper ones, and I have no doubt that these darker bands, and the more diffused stain of the spaces between them, are due to a trace of humus remaining in the successive subsoils. The ore-granule bands previously described are the representatives in a weathered face of the bands of darker stain in a fresh face of the clay. Such granules of ore are frequently formed in the present subsoil in the shape of pipes around decaying roots. I have found in the granule-bands specimens which bore on a concave surface plain imprints of vegetable form. At a low level in the deposit there is to be seen here and there a stratum of fine, dense, somewhat plastic clay (as Sir J. von Haast well described it in one of his reports), which, from the positions in which it occurs, can be nothing else than

a fine mud settled from pools of storm-water. This low-lying, exceedingly tough and impervious material has in some places retained stain-impressions of roots. Worm-borings.—Very satisfactory proof that the loess is a dry-land formation is afforded by the fact that worm-borings are to be found in it plentifully, in the dark bands, from top to bottom. Most of them have been filled up by the worms, as are recent borings; but some are partly open, and the upper part of the filling consists of loose distinguishable “asts.” It cannot be asserted that the lower borings were made by worms working down from the present surface. The borings are unmistakably more numerous in and just beneath each brown-stain band than in the spaces just above these bands, and they are in some places to be found beneath, and terminating abruptly at, the mud-beds just described, which have not been bored through at all. There are also to be found in the stratum beneath it worm-holes plugged with this mud, the plugs preserving the characteristic irregularity, the departures from cylindrical form, in the shape of the borings. The “capillary texture” of the loess, mentioned by Sir J. von Haast, is evidently due more to the multitude of wormborings than to the decay of roots, though this, no doubt, had some influence in producing it. Evaporation-veins.—Another evidence of the dry-land origin of the loess, and of long pauses in its growth, is the existence, beneath some of the lower granule-bands, of what, for want of a known name, I must call “evaporation-veins.” Most clay-formations contain what excavators call “backs”—natural vertical cracks—and these are usually lined with a film of greater or less thickness of finer, whiter, and denser material than the clay between them. In many roadside cuttings these vertical veins are seen to streak the clay quite thickly. If the vegetable soil be cleared away these veins are seen to divide the subsoil into irregular figures, rude pentagons being the most common form. In size the figures vary from a few inches up to 2ft. in longest diameter. A small number may be larger. I have never met with an explanation of these veinings, and must attempt one. Whenever a soil cracks through drought the cracks extend into the subsoil. The cracks in the subsoil must be fine, as one never finds on digging into it streaks of dark vegetable soil fallen into them. As the drying of the ground in a drought proceeds, the moisture at the surface of the fissures will evaporate; and that remaining in each prism of clay will constantly endeavour, so to speak, to maintain its capillary level, and will keep up the supply for evaporation at the fissures. In thus moving towards the fissures, I conceive that the water drags with it such fine

particles of earth as can be moved between the coarser particles of the mass, and these fine particles are of course deposited on the sides of the fissures when the water evaporates. The fissures, once formed, remain planes of weakness, the ground cracks in the same lines year after year, and in course of time a considerable thickness of fine material may be thus separated from the mass and collected upon the sides of each clay prism. (Is the whiteness of the veins due, wholly or in part, to a bleaching-power in the water which filters into them from the soil when a rain occurs?) In the present subsoil at Timaru these white veins range up to 1½in. in width, between prisms of clay not exceeding 2ft. in longest diameter, and they extend downwards as much as 8ft. or 10ft. in some places. In the lower portion of the loess there are two or more series of these veins, the highest terminating above in a granule-band about 10ft. above the rock. There is another terminated by the second well-marked band from the surface. There may be other series: the cuttings are too much hidden by rain-wash to allow one to see. If there were any possibility of mistaking the veins in vertical sections, there can be no mistaking the characteristic pentagonal forms seen in horizontal sections. The existence of these veins supports the testimony of the granule-bands as to the occurrence of long pauses in the deposition of the loess. But, while the granule-bands seem to indicate a wet climate, the evaporation-veins indicate a dry one, and to accept the contradiction and explain it we must suppose the pause was sufficiently prolonged to permit of a complete change of climate, either from wet to dry, or from dry to wet. “Bird-stones.”—It would be reasonable to suppose that the surface peat-pool cut into by the engine-shed excavation was formerly more or less frequented by water-fowl, and that they would leave lasting vestiges of their visits or sojourn there in the shape of ejected gizzard-stones. As a matter of fact, they have done so. Many small stones, well worn—unquestionable “bird-stones”—can be picked out of the clay immediately beneath the ashes of the burned peat-bed. The same expectation might be formed regarding the buried peat-pool; and it would be fully justified, for these vestiges of the water-fowl of the period are immeasurably more numerous there than similar ones in the surface-hollow. The majority of the stones are small, though larger than those in the upper bed, but among them are some the size of common marbles. Large and small, there must be a bushel or two of gizzard-stones buried with this old water-hole, the section of which is only about 20 yards long, and the stratum representing the pool only 12in. or 15in. thick. The same proofs of the existence of bird-life are to be found throughout the deposit, from the clay

which fills the larger crevices in the dolerite to the top, very thinly scattered through the mass generally, but in astonishing numbers in the granule-bands. In each of at least three of these bands superposed at one spot bird-stones exist in such quantities that one may well speak of them as “so many bushels to the acre.” Certainly no farmer could afford to sow grain so thickly. The buried peat-pool runs out into the uppermost and most prolific of these. The majority of stones found in the buried pool are of quartz or of other mineral whose whiteness deceived the birds; those found in the granule-bands, while containing a not-inconsiderable proportion of white stones, are of all colours, but for the most part of hard materials. Water-fowl usually possess considerable powers of flight, and from frequenting watercourses they have a better chance of obtaining quartz or other white or whitish pebbles. This would account for the larger proportion of such pebbles in the buried pool. Land-birds, whether of powerful flight or not, have not the same opportunities for selection; hence the mixed character of the pebbles found in the granule-bands—the old land-surfaces. The whole of the pebbles can, I believe, be matched as to mineral character from the drift-deposit overlain by the dolerite. Bird-stones are to be found in the present soil and subsoil, but they are so rare that the search for them is disheartening work. This refers to the dry-land surface. In the old peat-pool at the engine-shed very small stones are somewhat numerous, and beneath the mud in the bed of the Waimataitai Lagoon, now being cut back by the sea, white stones, of such size that they must have been used by some of the moa tribe, are not uncommon. In two places—one near the base of the deposit, the other in the lowest granule-band—I obtained a few large, mostly well-rounded pebbles of brown sandstone, such as could only have been used by gigantic birds.* This section will suggest a solution of the “very puzzling geological problem” stated by Mr. J. C. Crawford in “Trans.,” xvii., p. 341. Moa-bones.—Moa-bones have been obtained from the formation. Some were dug out in making the excavation for the passenger-station, but I cannot say from what position. In the somewhat low sea-cliff of clay at Dashing Rocks moa-bones are occasionally weathered out, about one-third of the way down the face. These have no relation whatever to the remains of a moa-hunters' encampment near that spot, these remains being wholly contained in black soil. Exceptional Bedding.—The loess generally is quite devoid of stratification in the ordinary sense, but there are small portions here and there which show a perfect bedding. Sir

J. von Haast mentions one of these in a report on the water-supply of Timaru. A much larger patch was cut into in “stripping” off the clay at the North Mole Quarry, on the bank of the Wai-iti Creek. This patch rises upwards from the creek as though laid down upon a sloping bank; the bank of clay it overlies is well veined and worm-bored. During the formation of a mass such as this from wind-borne dust, it could not but happen that now and then storm-waters would scour away from one place, and deposit in stratified form in another, some of the surface-soil. These rearrangements of material would naturally be made most frequently near the large watercourses, and it is in such positions that the only two cases I am aware of are found. The one mentioned by Sir J. von Haast is in the cliff bounding the Waimataitai Valley, on the south side. It consists of a small patch, as seen in section in the sea-cliff, 10ft. or 12ft. wide, 4ft. or 5ft. deep, and situated about midway up the cliff. The other case, as already stated, is beside the Wai-iti Creek. Granting that this bedded layer was laid down by a flood in the creek, it must appear that the stream had not then cut through the 50ft. of dolerite rock as it has done, and through some 150ft. of the underlying drift-formation besides. The Waimataitai Creek has similarly, but in less degree, cut down its channel since the patch of stratified loess was laid down.* Since this paper was written it has occurred to me that the Waiiti Creek patch was more probably a wind-drift. Unfortunately this idea did not occur to me soon enough, for on going to the quarry to re-examine it I found it had been nearly all carted away. In dealing with the vestiges of vegetation mention was made of a fine, somewhat plastic clay which occurs here and there, and regarding which it was suggested that it must be a mud deposited by pools of storm-water. This view is suggested by the fact that the best specimens of it are found in hollows in the rock-surface. In the North Mole Quarry a depression in the rock-surface was evidently almost filled by it, the deposit exceeding 2ft. in depth in the centre, and thinning out towards the margin of the hollow. In many places the mud-bed is formed, not at the base, but at some higher and apparently not constant level, yet never far from the base. A hint is supplied by the case of Wai-iti Creek, and the stratified silt at Waimataitai (which rests upon a similar mud), that the cutting-down of the channels after a time prevented the overflowing of storm-waters, and the production of more mud-beds and bedded rearrangements of the silt. Summary of Evidence.—I think it will be conceded that the evidence herein adduced is sufficient to prove the growth of the Timaru loess as a strictly dry-land—an æolian—deposit. There are good evidences of several old land-surfaces in

(1) bands of soil and subsoil ore-granules, (2) humus-stains; (3) worm-borings, (4) bird-stones—all in the same bands; (5) a buried water-hole in connection with one of them, (6) series of “evaporation-veins” in connection with more than one, (7) and, lastly, the exceptional occurrence of current-bedding, in silt and mud-beds near watercourses. All these evidences are mutually corroborated, and point to but one conclusion—the conclusion above stated. Marine, lacustrine, or fluviatile agencies are entirely out of the question. The Timaru plateau stands above all the surrounding country, and on the summit of Mount Horrible the loess rises to over 1,000ft. above the sea. Neither lake nor river could have lain or flowed over this region. Nor can submergence beneath the sea be invoked in explanation. The bands containing the several characters above described are all flat or curved with definite relation to the present drainage-lines. This relation would be impossible were the loess a marine deposit. Not a single shell or other mark of the sea has presented itself to me in the original deposit. I am informed that Mr. McKay found a marine shell in the loess. I am compelled to conclude that it must have been obtained from a slope-deposit or other rearrangement of the original formation; in such cases they are not uncommon. The Timaru region has not been beneath the sea for long ages. The dolerite buries an older land-surface, with distinguishable soil, impressions of plants, worm-borings, and moa-bones; this soil being formed upon the surface of a considerable thickness—some hundreds of feet—of river-gravels, sands, and clays. Having found the loess to be a dry-land deposit, we must find the still-missing factor in the dust-heap theory. The Source of the Dust.—There was only one source possible in these latitudes for such a quantity of dust; and a mere hint as to its nature will suffice. If we consider the loess to belong to the great Ice Age there is no difficulty. The dust was “rock-meal,” produced by the great ice mill, and spread out by rivers of sludge for the winds to dry, and pick up, and bear away, losing more or less of their load whenever they passed over a vegetated region. The material itself to-day bears testimony that such was its origin. Under the action of running water it tends to separate into darker- and lighter-coloured layers, the darker being the more oxidized particles. Sir J. von Haast remarked the resemblance of the stratified silt he saw to glacier-silt, and quite recent rearrangements by rains of the material of the cliff present precisely the same appearance. No other agent than ice could have produced so great a quantity of such fine material. Other Loess Deposits.—The small dolerite sheet of the

Geraldine Downs—which in all probability was emitted simultaneously with the Timaru dolerite—is covered by a bed of clay several feet in depth. I have not examined this with instructed eye, but, as it could have been deposited in no other way, its origin may safely be declared to have been similar to that of the Timaru loess. It is a stiffer, “colder” clay than this of Timaru, and this fact may be accounted for by difference in the character of the dust supplied to the winds which formed it. I cannot doubt that examination of the loess on Banks Peninsula (described by Professor Hutton as aqueous) would result in like proofs of its subaerial origin being obtained. Other loess-fields in Canterbury would doubtless furnish similar evidence. The descriptions I have read of the great loess-formations of the Rhine and of the “terrace” formations of North America, the origin of which, I understand, still puzzles European and American geologists, suggest to me in every detail that the dust-heap theory which explains the Timaru loess will also perfectly explain them. In a small volume on the “Geology and Physical Geography of Brazil,” by C. F. Hartt (Trübner, 1870) (which appears to be one of a series published under the title “Scientific Results of a Journey in Brazil; by Louis Agassiz and his Travelling Companions”), the author describes a “sheet” of arenaceous clay, very uniform in composition, “absolutely structureless,” “totally devoid of stratification,” usually quite free from pebbles or boulders. This sheet covers the coast provinces of Brazil from the level of recent estuarine formations—even extending beneath these—to the tops of the highest hills Mr. Hartt ascended. From his numerous topographical observations and his summary description of this superficial formation, which varies in thickness from a few feet to 100ft., I gather that it is a loess, and that it is probably of similar origin to the Timaru loess. The author and Agassiz could find no mode of deposition competent to account for all its features save a general glaciation of the country. It was a boulder-clay without boulders. The agency of dust-bearing winds, however, was not one of those taken into consideration by them. A superficial deposit in Lower La Plata, described by Darwin (“Voyage of the ‘Beagle’”) as estuarine, “with concretions and bones,” may also well be a wind-drifted dust formation.