AN ECONOMIC INVESTIGATION OF THE MONTANE TUSSOCK – GRASSLAND OF NEW ZEALAND.

New Zealand Journal of Agriculture, Volume XXV, Issue 1, 20 July 1922, Page 1

AN ECONOMIC INVESTIGATION OF THE MONTANE TUSSOCK - GRASSLAND OF NEW ZEALAND.

XII. THE REGRASSING EXPERIMENTS IN CENTRAL OTAGO.

Dr. L. COCKAYNE,

F.R.S., F.N.Z.Inst.

PART II.—OBJECTS, PRINCIPLES, AND METHODS. THE preliminary examination of the pastures of Central Otago, made by me in the late summer and early autumn of 1919, showed, that, apart from such ground as could be irrigated, no really sound knowledge regarding their improvement could be gained except by experimental methods. It is true that, in the valleys, enclosures which had been fenced from stock and rabbits for some years exhibited, according, to the length of time during which they had been enclosed and without water other than rain, every stage of regeneration from the original almost bare, ground to a close covering of vegetation important matter already dealt with in this series of. articles.* But this regeneration occurred only in the valleys or low down on the bases of the hills, so no information was available regarding. the upper and greater part of the. depleted areathat is, the all-important winter country of the sheep-runs, where the conditions for regeneration, as shown in Part I of this article,' are different from those of the valleys, and are themselves far

from uniform. Nor are precise particulars available concerning the actual plant contents of these areas in the valleys at the time of their enclosure, and their subsequent histories. For instance, a most important and fundamental point is whether the new vegetation has arisen, in the first instance, from seed, or.from plants which.were already in the ground when it was enclosed, though hardly visible through their close cropping by rabbits. In other words, would spelling, as so many believe, alone suffice to bring about regeneration ? In the summary to the article just cited the opinion is expressed that “ Most likely a good deal of regeneration arises from seeds being blown on to the enclosed area, though in many cases it arises from plants already on the ground ” (foe. cit., p. 94). But even if spelling would suffice it did not seem to me, in view of the capabilities of the rich mica-schist soil, that a mere return to the original pasture of low feeding-value ought to be the destiny of the depleted lands.. Rather, if regrassing were feasible, would it not be possible to clothe the now barren slopes with plants of high rather than of low pasture-value ? Even if none of the pasture-plants at present in New Zealand were suitable, there would be the pastures of the whole world available for material. These facts, and others detailed under the next head, clearly show that experiment was most desk able. This view I laid before the Director-General of Agriculture in a report. After due consideration he arranged that experiments on the lines therein indicated should be carried out. These have been in progress since the end of January, 1920, so far as spelling is concerned, but with regard to sowing, &c., only since the middle of May, 1920— that is, for about two years and two months. Here, and in Part III of this article, the results of the experiments up to 16th May,. 1922, are dealt with .. . . About twelve years ago the Department of Agriculture commenced a series of experiments on the Mackenzie Plain and near Clyde, but of a kind fundamentally different from those described below. A fair amount of surface-sowing on depleted ground exposed to stock and rabbits was also carried out. This sowing showed that usually no grass was established ; nevertheless in specially favoured spots, as I have noted, more or less grass, evidently due to this sowing,, is still to be seen. The Clyde experiments, which alone concern this article, have clearly demonstrated that many grasses, clovers, and other plants can be successfully established, and will increase, if sown in a well-prepared seed-bed on ploughed or cultivated ground and kept free from stock and rabbits.* Of course, this was already known from ordinary farm practice in the district; but the important part of the work near Clyde was the establishing of many species possessing pasture-value which had not hitherto been grown in Central Otago, as also was the successful growing of lucerne and cocksfoot for ten years or more under conditions apparently of extreme aridity. The area has since proved of great value in grazing experiments and subsequent spelling,! and in demon-

strating • the establishment under arid conditions of plants from. seed buried by the trampling of sheep (see Fig. i). In 1912 Mr. D. Petrie, M.A., F.N.Z.Inst., in a report to the Department of Agriculture,* had intimated that the regrassing problem demanded experiment. He suggested that an area of land be fenced in above the race feeding the Cromwell dam.f Such a station,” he writes, “seeded from the plants growing on the edge of the race, would soon show whether sun-scorching and wind-desiccation or the eating-out of the seedling grasses is the dominant factor in preventing the native grasses from reclaiming the desert lands on which they formerly flourished. I know of no locality where this important question can be tested more conveniently or in a more conclusive way.” Farther on, the same writer expresses his conviction that spelling alone will do no good; nevertheless he considered that “The question is of sufficient importance to justify the enclosure of a typical and fairly large patch of the desert hillside, to ascertain what spontaneous improvement a few years’ rest could effect.” He considered the Earnscleugh area of 20 acres as altogether too small, and not varied enough in its stations for plants, to give reliable results. It will' be seen from what follows that Petrie’s suggested experimental methods and mine are essentially different, except in the matter of natural sowing from growing plants ; but even in that regard Petrie relies on seed coming from plants outside the enclosure,. while I choose areas which already contain living plants suitable for producing and scattering seed. OBJECTS OF THE EXPERIMENTS. The general object of the regrassing experiments is either to find out methods which may be used for regrassing the depleted area, or to gain evidence that regrassing at anything like a reasonable cost is impossible. In order to procure such information precise knowledge on certain matters must be ascertained. The following appear the most important : — 1. To find out the effect of spelling alone, and the length of time which must elapse before results of moment come about. Here comes in the spelling of a large area with various soil conditions and aspects, and the spelling of uniform areas of different kinds. It is also important to find the effect of keeping out sheep alone and rabbits alone—a matter which Mr. Middleton considers of fundamental importance. 2. To ascertain the results brought about by sowing the seeds of various kinds of pasture-plants, apparently suitable for regrassing operations, both in the presence and absence of stock and rabbits. Here misleading information may easily arise owing to the number of stock or rabbits being a highly variable quantity.

3. To study the effects of different soil or climatic conditions with regard to the germination - of seeds, the establishment of plants, and their subsequent rate of growth. ■

4. To learn at what season seeds of the various pasture-plants can be sown with the best results. Here one is confronted with a climate varying more or less year by., year. ... 5. To ascertain the results of different' methods of sowing the seeds. 6. To learn* definitely whether certain species desirable for introduction will or will not grow. , 7. To gain information concerning the rate of growth and the amount of fodder produced by the 1 various pasture-plants when growing under arid conditions.* • . 8. To discover what intensity of grazing the various species will tolerate when fully established. . . . 9. To ascertain the results of seeding from plants .in situ. 10. To find out if trees can be established either by planting or sowing. ■ a ' . ■ THE PRINCIPLES UPON WHICH THE EXPERIMENTS ARE BASED. • The principle which lies at the root of the experiments is that of applying the exact knowledge gained from a close study of the history, flora, and vegetation of the depleted areas (the results of which have been briefly stated in Part I of this article) —that is, the methods of nature are to be followed as far as is feasible. f For instance, it has been seen that nature has sown seeds broadcast upon the bare inhospitable ground; and from such plants as appeared the rigid selective power of the environment has allowed only those to survive which are so constructed as to be able to withstand not only excessive drought, but also grazing of the hardest character. It has also been seen that some of the plants thus established are more or less palatable, and that some of the most palatable of the primitive grassland species still linger here and there, though so small and closely eaten as to generally escape notice. As the vegetation established by nature varies according to aspect, altitude, and the quality, physical condition, water content, and depth of the soil, such vegetation of any area is a reliable index of its capabilities for regrassing ; and the dominant species of this vegetation may be looked upon as indicators of what will or will not grow. For instance, the occurrence of scabweed in profusion is a certain indicator of the almost maximum adverse conditions for regrassing (see Fig. 2)' —the absolute maximum being denoted by either completely bare, much wind-eroded grourid, or land on which the scabweed is being rapidly eradicated. On the other hand, florin (Agrostis alba) speaks equally plainly of a soil constantly moist. In the depleted country, as seen in Part I, nature has no large number of species tolerant Of arid conditions on which she can draw ; also, she is hampered by the extreme slowness with which many such species extend their range. Experiment, on the contrary, can proceed far more rapidly and has a considerable amount of material available.

Here-again field observations come in, and the greater the knowledge of plants in their natural habitats possessed by the experimenter, the more easy -it is for him to avoid that loss of time brought about by using empirical methods, for many of those most likely to thrive are known already and can at once be made use of. But empirical, methods, quite different though they , are from those under discussion, are not to be despised, since plants apparently most unlikely to succeed according to nature's teaching may be of considerable value. Thus it will be seen in Part III that tall fescue (Festuca elatior) — a swamp promises to be specially suitable for arid areas. Fit seems, quite sound to assume, with regard to the experimental plots, , that what will take place on a small area—say, a quarter of an acre in extent—in the matter of seed-germination, growth of plants, and spelling, will happen on a .much larger area so long as the conditions for vegetation are exactly the same on both areas. If this be so there is no need for large experimental plots. In point of fact, exact similarities of environment in two or more such areas do not exist in nature, but the actual differences may be so slight over much of ' each area that it is justifiable to consider the conditions identical. Small plots have'also-the great advantage that those changes in the vegetation which take place from time to time, due to the conditions imposed by the experiment, can be far more easily and accurately estimated than would be the case in the larger ones. As there are various types of habitat on the depleted slopes of the mountains, there can be no uniformity in the behaviour of the plants introduced thereon. Any experiments, then, designed to furnish a true idea either of spelling a large area extending from the base of the mountain to the grass-line, or of the possibility of growing plants over.;the whole area, must be sufficient in number to allow at least one plot; for each type of habitat. But between the typical habitats there exist . many others of an .intermediate or transitional character. Some of these are certain to be represented in almost any area of a quarter of an acre, while plots can also be selected on account of their lack of uniformity, and thus various lessons could be learned on the one plot (see. Fig. 3). ? ; . - : J ' ■ ..' . ’ ' : Spelling, that highly important object of the experiments, demands a : . careful survey of the vegetation of each plot in the first instance so as to learn exactly what each contains, and periodical surveys to note what is taking place.. Then, if regeneration is in progress, its rate of progression- is the point of special moment, for it may easily be too slow to be' a paying proposition. With spelling may go, hand-in-hand,. seed-sowing, so that each plot can serve at any rate a dual purpose. . . ■ • .

.4 No method of treating the experiment plots can be applied which is likely, when used on a large scale, to cost more than will be profitable. Thus ploughing’ even if possible, is but of the question. On the other hand, there is much country where, if it is proved that sufficiently good pasture can be established, by that method, a light harrowing might well be a payable proposition. ■ • - ■ . As it has been shown clearly by the Eamscleugh area and by many observations that various palatable grasses, clovers, and other herbs grow well on badly depleted land, even ; where there is the minimum rainfall, an important principle is the using of species of higher pasture-

value than those of the original grassland. It should be ' only after failure to establish these better plants that a restoration of the primitive plant-covering be aimed at. The latter object has, up to the present, been , the ideal of those who have , so frequently suggested spelling as the panacea. In the arid depleted areas of the United States, too, the restoration of the original grassland appears to be the chief aim.* .. . THE METHODS OF THE EXPERIMENTS. ' In the report submitted to the Director-General it was suggested that twelve rabbit-proof enclosures, each about a quarter of an acre in area, be erected in some part of the worst-depleted area, these enclosures . to represent the typical habitats of such country from an altitude of nearly 1,000 ft. to the border-line between the depleted and'the grassed areas at about 3,000 ft. altitude. According to the principle already mentioned, such small areas if well. chosen should give an accurate account of what would eventually happen were the entire area on which they were situated to be spelled. Also, they could be used for experiments in sowing and for various other purposes. . , Mr. J. L. Bruce, of the Department of Agriculture, who had had many years’ experience of Central Otago, was deputed to accompany me, and together we were to choose the sites for the plots and arrange for the fencing to be carried out. After going into the matter, in . its various bearings we decided to explain our views and seek advice from certain of the local runholders, whose practical knowledge would be of great advantage in the work. Accordingly, we consulted Messrs. R. K. Smith, of'Tarras, and D. Middleton, of Northburn, who willingly consented to act with us as a committee. Finally, it was decided to carry out.. the experiments on Mr. Middleton’s Northbum Run, since it was easy of access and typical of much of the depleted area. Mr. .Middleton also kindly consented to supervise the fencing, and to arrange that the plots should be visited at comparatively short intervals in order to keep them free from rabbits. The following day Mr. Middleton, Mr. Bruce, and myself visited Northbum Run and selected the sites for nearly all the plots. These varied from the most stony, exposed, and highly depleted ground (see Fig. 2) to deep soil on dark faces where Californian. thistle (Cnicus arvensis). % is frequently abundant (see Fig. 4). In one case two slopes of different aspect, and facing each other, were selected, while another of a saddle shape gave two distinct aspects and classes of soil (see Fig. 3). In some were rocks among which grew certain indigenous grasses,' especially tall blue-tussock (Poet intermedia) and' blue-grass (Agropyron scabrum), whose seeds at times would reach the bare ground in their vicinity. The final plot, situated at nearly 3,000 ft. altitude,, was later on selected by Mr. Middleton himself in order to show the effect of spelling country near the upper limit, of depletion, where there is more or less nearly dead tussock. ’ : The fencing of the plots was no easy matter, as some are in positions difficult of access for fencing-material, and. frequently the ground was extremely hard and rocky. By the end of January, 1920, the fencing was completed; and in February, with my assistant, Mr. W. D. Reid,

I spent several days in dividing the plots into sections and in taking detailed notes of the plant contents of each, or, if the plot was almost bare, with little but the scabweed mats, the notes were more general. In order to divide . the plots into their sections certain of the standards were labelled. In every case the north-west corner post was marked A with lead-pencil on a patch of wet white paint, and then, at regular intervals, standards were marked B, C, D, &c. The parallel fence was similarly labelled, and so, too, were the two other parallel fences, but in their case the lettering was a, b, c, &c. Then pegs were put into the ground to mark the lines AA, BB, &c., and the cross-lines aa, bb, &c., so that each plot was eventually divided into sections named Aa, Ab, Ac, and so on. Notes of the labelling of each plot were taken, so that if the letters on the standards became illegible it would be easy to find the exact limits of the sections. Notes were taken of the plants growing in each section. Later, in the spring October, 1920 —when many of the plants were easier to recognize, the results of the first survey were checked and additional plants recorded, so that there is an accurate enough account available of the original plants of each plot. Their increase from seed and rate of growth can be readily estimated, and no fresh species can come in from outside without being detected. To each plot a number was given, commencing with the lowest, to be known as plot 1, and ending with the uppermost, plot 12. The altitudes have been noted by aneroid on various occasions, so are correct enough for all purposes, for..a few feet either way make no difference.

In accordance with the principles already explained, the treatment so far extended to the plots has been as follows : Each was divided into three parts, two of which were for autumn and spring sowing respectively, and the third for the planting of certain grasses, clovers, and other herbs, in order to see, firstly, if such would grow if planted with care, and, secondly, to establish plants in order to find out if the seeds from such would germinate under the conditions supplied by the environment, the seedlings being readily observed on the bare giound. Evidently, if . there were no such germination, it would be useless to use those particular species for regrassing, since this must depend upon the establishment of a few plants in the first instance, which would populate the ground by means of seed, rather than upon the sowing of the very large quantity of seed required to make a close covering. Further, the planted species should seed much earlier than the same species raised from seed, and so time be saved.

The main sowings and planting were carried out solely by Mr. Reid, since not only had he the necessary experience in such work, but because the employment of one person could alone secure uniformity in the different sowings. In order to save time and to economize space the seeds of the different species used were mixed together. Such a mixture was not in the least designed to test its. value as a mixture for regrassing purposes, but only to ascertain the power and rate of establishment of its various constituents. The composition of the mixtures will be best dealt with in Part III of this article, when discussing the results of the experiments. .

As only the cheapest methods were allowable two alone were followed —namely, broadcast and raked. The latter was designed to

prove if harrowing 'far more efficacious* method—could be used, since there are many thousands of acres where harrowing with one stroke, or probably with two strokes, should not be - too expensive. In the raked method the seed was first scattered broadcast, and then the ground was lightly raked once with an ordinary garden rake. In many places the. surface was , so hard that the rake made virtually no impression. The raked method was fully approved by Messrs. Smith and Middiet - . . ■

The planting was carried out with plants specially prepared for the purpose by Mr. J. Beverley, at the Weraroa State Farm, and with trees.supplied from the nursery of the State Forest Service at Tapanui. A list of these will be given in Part III of this article. As the object of growing these grasses and trees was to get them-well established as rapidly as possible for the purposes already indicated, and not to show how such should be cheaply planted for regrassing or forestry purposes, at the time of planting water was put into the holes where they were to be placed. Bringing this water on to the ground was no easy task, for not only had it usually to be brought from a considerable distance, but also from a deep gully far below. Nor was it always easy to make holes in the hard stony ground. Horses for carrying the seed, plants, water, and various implements were generously supplied by Mr. Middleton; indeed, throughout this experimental work his willing assistance has been invaluable, and I take this opportunity of most sincerely thanking him. ■ THE SOWINGS. The first sowing took place in the middle of May, 1920, at which time the planting of pasture-plants and trees was also carried out. The second sowing was made at the beginning of the September following, and the third early in March, 1921. Thus there have, been, two autumn (early and late) and one spring (early) sowing. The March, 1921, sowing was made on a portion' of the ground of the May J 1920, sowing where the germination had been the poorest. Mr. Reid has supplied an excellent series of maps drawn to scale, showing the exact position of each sowing and of the plants of the planting. . ' In the March sowing I sowed carefully in nearly all the plots, in short rows, each species by itself, covering the seeds with soil, in order to ascertain its germinating-power under the arid conditions and to learn if after, germination plants would be properly established. Evidently those which fail to become established after such treatment, unless they have been too thicky sown, will almost certai.ily be useless for either the -broadcast or raked methods. Also, near the fence in plots 1, 2, and 13, I sowed some insignis pine, dropping two or three seeds into shallow holes made with a short pick . and then covering them with soil. As well as the sowings made . inside the enclosures, small areas on the depleted ground outside were sown by the broadcast and raked methods. These also are shown on the plans. ■

''" In'addition to the twelve ' plots three' Sowings; exactly similar to those described for the plots, were carried out within the experimental area planted with trees and fenced from rabbits, on the Sugarloaf, near Lowburn Ferry, on. the 20th May, 1920, 10th September, 1920, and

6th March, 1921 ; also sowings were made outside on the first two dates. Here a good deal of regeneration had taken place since the fencing, especially of blue-grass and of tall blue-tussock, so it might well be expected that other drought-resisting plants would be established from seed.

Then in the Eamscleugh experimental area, already fully described in this series .of articles,* a sowing of -Jq acre in extent was made in May, 1920, on the half-acre of depleted ground which had been fenced from rabbits a short time before (see loc. cit., p. 179, Fig. 2), and another sowing was made outside; the same mixture and methods were used as those for the Northbum plots. '

The trampling of the numerous sheep used in the first palatability experiment on the Earnscleugh .area had amply proved that greatly increased germination of lucerne (see Fig. 1), cocksfoot, and other plants was so induced. This fact I brought before the DirectorGeneral, and he readily agreed that an enclosure should .be made on the lower country of the Northburn Run to test the sheep-trampling method. . By taking advantage of an existing fence it was possible, for the sum allowed, to fence in a piece of ground acres in extent and to, subdivide it into three half-acre plots, to be known respectively as 13A, 13B, and 13c, the first-named being the most westerly. Early in September, 1920, these plots were treated as follows : Seed was sown broadcast on plot 13c, and then the ground harrowed with one stroke of the harrow. ' Plot 13B had 128 sheep driven twice round over the broadcast seed. The same sheep were kept on plot 13A for ' forty-four hours, seed . having previously been scattered over the ground and hay placed there for food. On the 5th March, 1921, part of plot 13c was harrowed with one stroke, then seed was sown broadcast and the ground- again harrowed similarly.

In addition to the experiments already detailed some sowings have been made in the open on a larger scale, using the harrow method. Near plot 13, on the 23rd September, 1920, about if acres was sown broadcast and then given one stroke of the harrow. A portion of this area was again sown on the 5th March, 1921, but this time it was harrowed both before and after sowing. On the nth December, 1920, another area, adjacent to the last, but on much steeper ground, was harrowed, sown broadcast, and again harrowed. This, obviously, was an exceptionally severe test, the sowing being made at perhaps the most unfavourable period of the year. On the nth March, 1921, an area of 1 acre on the depleted ground above the Earnscleugh experimental area was harrowed, sown broadcast, and again harrowed. In the experimental area itself a similar sowing of half an acre was made. CONCLUSION.

In the final part of this article details will be given regarding the species used in the experiments, the amount of seed and rate per acre sown of each species, the climate month by month for the period of the experiments up to the nth May, 1922, the results of the experiments up to that date, and some of the lessons which they appear, so far, to teach. ■ . ; •

* See this Journal, Vol. xx, Feb., 1920, pp.'82-94.

* For an account of the Earnscleugh experimental area, Clyde, its condition in February, 1920, and a list of its plants, see No. VIII of this series in this Journal, Vol. xxi, Oct., 1920, pp. 178, 180, and 187-88. f See No. IX of this series in this Journal, Vol. xxi, Dec., 1920, pp. 324-34 and No. X, ibid., Vol. xxii, Mar., 1921, pp. 148-57; In the last-named article there are certain details regarding the climate of depleted Central Otago, pp. 149 and 151. . .p; 1

*" Report on the Grass-denuded Lands of Central Otago,” Bull. No. 23 (n.s.), N.Z. Dept. Agri., 1912, p. 13. .. . ♦... • f For photograph of the race and area adjacent, see No. Ill of this series in this Journal, Vol. xix, Sept., 1919, p. 134. ■ - . ■ r

* I am carrying out a series of experiments with many indigenous grasses, one object of which is to ascertain their rate of growth and seasonal or monthly yield when cultivated in the lowlands in the neighbourhood of Wellington. This, although the conditions are far from being arid, should give important information for purposes of comparison. ... f For a general account of studies of this kind see L. Cockayne, The Importance of Plant Ecology with regard to Agriculture, N.Z. Jour. Sci. & Tech., March, 1918, pp. 70-74.

* A. W. Sampson, Plant Succession in Relation to Range Managemeht, U.S. Dept. Agri. Bull. No. 791, p. 74, 1919 ; also, F. E. Clements, Plant Indicators, pp. 310-15, Washington, 1920.

* See this Journal, Vol. xxi, Oct., 1920, pp. 176-82.