Art. XVII.—A. Sketch of the Plant Geography of the Waimakariri River Basin, considered chiefly from an Œcological Point of View.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 32, 1899, Page 95

Art. XVII.—A. Sketch of the Plant Geography of the Waimakariri River Basin, considered chiefly from an Œcological Point of View. By L. Cockayne. [Read before the Philosophical Institute of Canterbury, 2nd August, 1899. Plates X.-XIII. Part I.: Introduction. Regarding New Zealand plant geography, nearly the whole that has appeared up to the present time, especially in the “Transactions of the New Zealand Institute,” has been of a floristic character. This, although of great importance to science in general, and of especial interest, moreover, to New Zealand botanists, does not meet the present-day demand by biologists for more minute details regarding the environment of plants and the adaptations to such that they have assumed. Scattered through the writings just referred to are here and there facts incidentally mentioned which bear on my subject. Sir James Hector, in 1869,* “On the Geographical Botany of New Zealand,” by Sir James Hector (Trans. N.Z. Inst., vol. i.). selecting a portion of Otago, arranged the plants into zones, determined by altitude, and divided, thiese into subdivisions according to the prevalence of certain plants. He also illustrated the paper by means of an excellent sectional map, showing at a glance the nature of the plant-covering in that part of New Zealand. Haast,† Report of the Geological Survey of the Province of Canterbury, Christchurch, 1864, p. 23. in 1864, pointed out the important fact that the climate of central Canterbury was of a continental rather than of an insular character. Various authors have divided the plants into communities, of which those adopted from Watson‡ “Topographical Botany,” 1883. are not to be commended—littoral, rupestral, and the like; indeed, when we examine the meaning of these terms as given by Kirk§ “On the Botany of the Thames Goldfleld” (Trans. N.Z. Inst., vol. ii., p. 89). it can readily be seen that from an œcological point of view they have little value. Thus, littoral plants are described as those of the seashore, whether growing on sandy or muddy beaches, in salt meadows, or on seacliffs—stations truly which may offer very different life-conditions. Certain papers, such as on the naturalised plants of Port Nicholson,∥ T. Kirk, Trans. N.Z. Inst., vol. x., p. 362. on the fertilisation of plants,¶ G. M. Thomson, Trans. N.Z. Inst., vol. xiii., p. 291. on the displacement

placement of native by introduced plants,* T. Kirk, “The Displacement of Species in New Zealand” (Trans. N.Z. Inst., vol. xxviii., p. 1). and others to be referred to later on, are of great œcological interest. But of all works on New Zealand vegetation the one which most concerns us here is that of Diels, entitled “Vegetations-Biologie von Neu-Seeland,” which appeared originally in Engler's Botan. Jahrbuch for 1896. Although the author had never set foot in New Zealand, yet, with the aid of a considerable number of herbarium specimens, and of whatever literature was available, he wrote an œcological account of the vegetation of New Zealand and its adjacent islands. He also touched at some length on the origin of the component parts of the flora. Part of the data concerning the plants of the montane, subalpine, and alpine regions was, in response to a request from Dr. A. Engler, furnished by me. Now, since some of my statements may not have been altogether correct, I may be allowed for my own credit's sake to explain that I distinctly stated in a letter to Mr. Diels that I could not guarantee their scientific accuracy, since they were written from memory for the most part, and not from observations taken note-book in hand from the plants in their habitats. Nor had I any idea of the scope of the proposed work, or I might have been more cautious still. Be that as it may, the work appeared, and, when the conditions under which it was written are considered, it is indeed a work of no small merit. That it should contain errors goes without saying; the only marvel is that it does not contain many more. At any rate, it marks a distinct epoch in New Zealand botany, and now it remains for local botanists to supersede it with something more full and more accurate. To fulfil this want in some small measure will be attempted in this present work. That a field botanist in a distant colony can endeavour to solve any of the very difficult problems which plant œcology presents is hardly to be expected. The very fact of being at active work in the field hinders the close study necessary before approaching and while engaged in such work; also, the almost entire lack of recent literature makes critical work quite out of the question. These higher problems—e.g., the verifying or disproving hypotheses by accurate physiological experiments; the collecting multitudes of facts from the records of many writers; considering such, and deducing conclusions therefrom; or the examining material collected from all over the world to establish some point—such work and the like can only be attempted by men of great natural ability, special knowledge, and vast learning, with the most modern appliances to hand, and access to the literature of any particular subject. But such specialists, at home in their laboratories

tories, cannot observe for themselves the actual environment of a plant in a distant land, or the manifold details connected with its life-conditions and the responding adaptations of the organism to these. Such can only be learnt by studying the plants in nature, and at all seasons. Experiments also with living plants can be more easily carried out at home than abroad, where, in many instances, it is very difficult to keep them alive, and still more so to grow them in a natural manner. This latter branch will form some small part of my subject, but the former—viz., the environment of the wild plants and their more evident adaptations—will constitute the greater part of what is to follow, so that the work as a whole will savour more of the open air than of the botanical laboratory or the study. I have selected a small portion of the South Island for treatment rather than the Island in its entirety, since the time is hardly ripe for a work of such an extensive nature, and a smaller area can be treated much more thoroughly, and yet at the same time be eminently typical of the whole Island. Of all regions to be found in this Island none seems to offer itself as better or even as well suited for such an investigation as that under consideration. For, in the first place, a portion at any rate of the mountain region is easily examined, thanks to a coach-road, an excellent coach service, and the accommodation offered by two hotels. Other parts, too—the Canterbury Plains, the coastal region, and most of the foothills—are easy to examine. Also, it is the only region where meteorological records at so high an elevation as 641 m. have been kept, and which, in combination with those of Hokitika, afford some idea of the heavy western rainfall. On the other hand, the records of Christehurch and Lincoln tell us something of the weather of the Canterbury Plains and of the sand-dune region. The district also presents an admirable example of two distinct climatic regions, the one extremely wet, the other dry. There are also lower mountain, subalpine, and alpine zones, and many very characteristic plant-formations depending upon peculiarities of soil. From what is written above it must not be imagined that the district as a whole is easy of access. On the contrary, a considerable portion is quite uninhabited. The valley of the River Poulter, for a distance of thirty miles or more, does not contain an inhabitant or even a hut, and the main sources of the Waimakariri, as also most of the country to the north of that river, is practically uninhabited. The mountains also in that part of the district being forest-clad up to a height of from some 900 m. to 1,200 m. makes the approach to the subalpine and alpine regions difficult. One trouble in the way of exploring such country is the taking of a sufficient food-supply to meet the danger of being cut off from the settled regions

by floods, for, since the stony river-beds are the sole roads, such an occurrence is by no means unlikely. I have especially called attention to these details as an excuse for the very imperfect examination that I have made of many parts of the district. One has often to leave interesting ground through bad weather, or this same weather may put a stop to all work. But if some of the region is uninhabited, this is all the more interesting to the botanist, since it has led to many places being still unused as pasturage for sheep, and where in consequence, the original plant-covering of the soil remains undisturbed. Good examples of this occur at Walker's Pass, Goat Pass, the sources of the Poulter, and many places near the main sources of the Waimakariri. Mount Hikurangi, in the North Island, is also in the same condition, according to Mr. James Adams, who writes:* “On the Botany of Hikurangi Mountain,” by James Adams, B.A. (Trans. N.Z. Inst., vol. xxx., 1898, p. 416). “Neither cow, nor horse, nor sheep, nor pig has ever desecrated the summit of the mountain, or disturbed there the designs of nature in the manner of the growth of plants.” All the observations recorded in this and to be recorded in the succeeding papers have been made by myself, unless the contrary be distinctly stated; they are the results of many botanical excursions taken during the past twelve years; also I reside and have an experimental garden† This will be referred to as “Tarata Garden.” in the coastal region, situated partially on the older sand-dunes. Regarding the photographs taken and to be taken of the various plant-formations, &c, I propose to place a set, numbered in accordance with the references in this work, in the Museum at Christchurch, New Zealand, since only a limited number can be published; and also dried specimens of all special forms of plants treated of. For the identification of species I am in the main responsible; other identifications will be specially noted. Of course, in a flora not yet thoroughly understood errors of identifieation must occur; others will creep in through the necessarily rapid examination of many localities; both these sources of error are at the present time unavoidable, but I hope they will be minimised by the deposition of doubtful specimens in the museum, as stated above. There are also in the New Zealand as in all floras a very considerable number of plant forms which some botanists consider as good and others as bad species. Now, it is essential for cecologieal work that every form treated of shall have a name of some kind, whether specific or varietal matters little. In many instances the structure of certain organs of a plant—for example, the leaves of a so-called variety which reproduces itself “true” from seed—differs altogether from that

of the type species to which it is referred. Other species vary simply through environment, and when such varieties are grown side by side under the same conditions they become indistinguishable. In such cases œcology can assist systematic work and check its results. Of late, botanists in New Zealand have been much more ready than formerly to admit closely related forms as good species, which certainly seems a step in the right direction.* As to what is meant by the term “species,” Romanes has written at considerable length: see “Darwin and after Darwin,” vol. ii., London, 1895, chapters viii. and ix., especially pp. 202–208 and 229–235. Only the angiosperms, gymnosperms, and pteridophytes are here treated of With these, many lower plants are, of course, associated in the plant formations as component and often most important constituents, but an account of these can only be undertaken by specialists. The country to be treated of includes all the land drained by the River Waimakariri and its tributaries, with the exception of those portions of Big Ben and the Malvern Hills which form a part of its river basin. To the following I beg to express my most hearty thanks Mr. D. Petrie, M.A., to whose splendid collection of Otago plants I have had full access; Professors A. F. P. Schimper, K. Goebel, and A. Dendy, for assistance regarding literature; Captain F. W. Hutton, F.R.S., who most kindly corrected the geological details; Sir James Hector, F.R.S., and Sir John Hall, K.C.M.G., for assistance regarding metebrqlogy; Messrs. T. W. Adams, John Deans, W. Cloudesley, J. Rountree, S. Weetman, E. G. Staveley, manager of the Loan and Mercantile Company, A. L. Taylor, of the Christchurch Botanic Garden, and Mr. T. Douglas, manager of Mount White Station, who have rendered me much valuable aid. Topography. With regard to the main features of the region under consideration, we may make a primary distinction into lowland, table-land, and mountain regions. Commencing with the lowland region, we find, near the mouth of the river, extensive sand-dunes, which terminate at the wide brackish lagoon formed by the river as it empties itself into the sea. Formerly the river also flowed into the sea more to the south, one channel, at any rate, going down the present bed of the Avon.† See map facing p. 396, in Haast's “Geology of Canterbury and Westland,” Christchurch, 1879. The sand-dunes consist of two varieties—the recent, which are in a very unstable condition, constantly moved by the frequent high winds, and having the yellow-leaved Desmoschœnus spiralis as the leading plant; and the older, which are

stable for the most part, and bear in consequence a more abundant vegetation than the first named. Passing in a westerly direction from the dunes, the Canterbury Plain is entered upon, through which the river flows in a fairly straight course over a wide stony bed, margined occasionally with wet ground, in which situations Arundo conspicua, Phormium tenax, and often Cordyline australis are plentiful, and become a pleasing feature in the landscape On the shingly river-bed, in its firmer portions, various species of Raoulia flourish, forming large silvery, moss-like patches. On either side of the river are stretches of often extremely stony ground, looking in some places almost as if it had been paved, and having for a plant-covering a low-growing and very characteristic vegetation, amongst which the dwarf, leafless, shrubby Carmichaelia nana is conspicuous. Interspersed with, and of much greater extent than, this stony ground are larger or smaller tracts of land suitable for cultivation, and varying from very light sandy and stony soil to rich dark-coloured loam of very great depth, the whole forming, indeed, one of the finest farming districts in New Zealand. The richest part of this land was for the most part originally swamp, occupied by Phormium, but such is now almost altogether reclaimed, and yields immense crops of cereals, potatoes, and the like, especially if the season be not too wet. These swamps and their immediate vicinity were most likely at an earlier date occupied by forests. This subject will receive further consideration when the pine forest plant-formation is treated of; here it need only be mentioned that a small portion of the primeval forest still remains in very fair preservation, having been well cared for by its owner, Mr. John Deans, of Riccarton, and affording a most valuable record of the former arboreal vegetation. At about fifteen miles from the sea, in the neighbourhood of Courtney, some interesting low sandhills, mounds, or ridges are met with, having Isolepis nodosa on the sunny and Phormium tenax on the south-west side, with Sophora prostrata near their bases. From the sea to the mountains the plains rise gradually at an average of 10.9 m. per mile, until at near the base of the latter—Mount Torlesse—they reach a height of 450 m. (For full particulars as to the fall of the Canterbury Plains, see Haast, loc. cit., p. 403.) At a distance of at from six to seven miles from where the river leaves the plain occurs the lower gorge; here the river has cut for itself a passage right through an isolated hill—Gorge Hill—standing towards the middle of that part of the plain, a most remarkable phenomenon when we consider that the river appears to have had the whole of the plain at its disposal for a bed. An account and probable explanation

of this is given by Captain Hutton* Trans. N.Z. Inst., vol. xvi., p. 449. in a paper entitled “On the Lower Gorge of the Waimakariri.” For our purposes the chief interest lies in the fact of a station for plantlife so different from the uniformity of the plain being provided by the rocky walls of this gorge, and where, indeed, quite a different vegetation is encountered. Amongst the plants found there are a number which, so far as I know, rarely occur at any distance from the sea, and this seems to strongly favour Captain Hutton's theory, the crucial point of which is that the sea at one time came up to this spot. Of these plants Linum monogynum, Parsonsia rosea, Chenopodium triandrum, and Angelica geniculata† have not observed any of these species growing in the upper gorge. Mr. T. W. Adams tells me that L. monogynum was formerly common in the river-bed of the Hawkins, at 183 m. above sea-level. may be especially cited. Regarding this latter Kirk writes,‡ “The Students' Flora of New Zealand,” Wellington, 1899. “I have never seen this plant far from the sea, and doubt its occurrence inland.” Passing through a deep rock-bound gorge between the Torlesse and Puketeraki Ranges, six miles in length, according to Haast (loc. cit., p. 244), the table-land is entered. On both sides of this gorge—the upper gorge—hills of some considerable size arise, covered with mixed forest in places, having Fagus solandri as its leading tree. This portion of the river has hitherto been most difficult to examine, but quite recently a road has been made for railway purposes, so the gorge on its south side is now easily reached. It is to be feared that this railway-construction will lead to the destruction of much of the vegetation, since already several fires have destroyed large areas of forest. The upper plains of the Waimakariri form an extensive table-land, in which the bed of the river itself occupies, of course, the lowest part. The average height of the whole is, according to Haast, 630 m. (loc. cit., p. 214), and the extent sixteen miles long by eleven miles wide. From its centre rise several high peaks, which are quite isolated—Mount St. Bernard, Mount Sugar-loaf, and Broken Hill. It is drained by various tributaries of the main river, which have cut for themselves deep beds and formed many extensive river-terraces, which afford in places a station for an interesting plant-formation of xerophilous shrubs, of which Veronica cupressoides§ Henslow, “Origin of Plant Structures,” London, 1895, p. 108, speaks of this shrub as growing at great elevations on the mountains, whereas it is essentially a plant of the lower mountain and lower subalpine regions, growing on river-terraces. is one of the most remarkable. The table-land is surrounded on all sides by lofty mountains, which are cleft by the Waimakariri and its tributaries, whose river-beds lead right into their fastnesses, so enabling the distant summits

to be reached, which but for such natural roads would be very inaccessible. To the northward are Esk Head, the Candle sticks, and the Snow-cap Range, giving rise to. proceeding from east to west, the Rivers Esk, Poulter, and Hawdon; to the eastward is the Puketeraki Range; to the southward the Mount Torlesse Range; to the westward the Craigieburn Mountains, giving rise to the Broken River and its tributaries Leaving the table-land and going in a westerly direction, the upper valley of the Waimakariri, twenty-two miles in length, is entered, its bottom occupied almost entirely by the wide stony bed of the river, and with mountains rising steeply on either side. On the north is a portion of the Dividing-range, giving rise to the River Bealey, River Crow, and the north branch of the River Waimakariri; on the south is the Black Range, giving rise to Bruce's Creek, Broad Creek, and several other unnamed creeks. At its western extremity the valley narrows, and turns quite suddenly to the south, soon rising above the forest line. Here it is traversed by the main branch of the Waimakariri, the White River, which rises in a large glacier situated at the head of the valley. This valley is three or four miles long; the mountains are steep and rugged, and the river a foaming torrent. In height the mountains vary from about 2,400 m. to 1,350 m., Mounts Rolleston, Armstrong, Davie, and Greenlaw in the Dividing-range, Mounts Franklin, Hunt, and McCrae in the Snow-cap Range, Mount Enys in the Craigieburn Mountains, and the main peaks of Mount Torlesse, Mount Binser, and the Puketeraki Mountains, being among the most lofty. Unlike the Westland rivers, those of the Waimakariri rarely become impassable gorges. It is usually fairly easy to follow them from mouth to source. The Dry Creek from Mount Torlesse, the Broken River, River Esk, and River Poulter in the table-land, the Craigieburn River, Andrew's Creek, and the River Minchin, above Lake Minchin, present exceptions, but in no instance, the last-named river excepted, do we find anything approaching the Otira Gorge, to quote a familiar example. Geology. The mountains both of the eastern and western plantregions belong geologically to Haast's “Mount Torlesse Formation,” which contains rocks of various ages, from the Carboniferous to the Lower Jurassic (loc. cit., pp. 266–280), consisting of sandstones, slates, and shales of various kinds. These sandstones are most easily split by the weather, and such excessive weathering, more particularly in the eastern region, has given rise to débris -fields, locally called shingle-slips, of very great dimensions, whose appearance, &c., will be

described more fully when treating of their very peculiar and amazing plants. The surface of the table-land consists of fluviatile and glacial deposits, covered often with a very scanty alluvial soil, below the general level of which the rivers meander in several streams over wide stony beds, with huge terraces, sometimes four in number, on either side. The Trelissick basin has been examined in a thorough manner by Captain Hutton,* Trans. N.Z. Inst., vol. xix., p. 415. and so the relation between the soil and its plant-covering can be there better studied than in any other part of the district. The sections shown in pl. xxv. are of great importance for this purpose. The Trelissick basin consists of undulating country, rising in places to a height of 900 m., and formed of rocks belonging to—the Pareora system (Miocene), a series of blue clays, shales, and sandstones; the Oamaru system (Oligocene), consisting of coralline limestone, underlain by volcanic grits and tuffs; the Waipara system (Cretaceous), formed of argillaceous limestone and calcareous sandstone underlain by marl. In addition to these are to be found some brown-coal beds towards the northern extremity of Mount Torlesse, at the base of the Craigiebun Mountains, in the Craigieburn Creek, and elsewhere. In the Esk Valley there is also a small tract of limestone country, which is described by Haast (loc. cit., p. 151) as consisting of “younger outliers of greensand and calcareous limestone, which have been broken through and covered with dolerites, to the hard nature of which they doubtless owe their preservation during the great Ice age.” Leaving the mountain region and coming to the Canter bury Plain, it is seen to be formed of stony débris brought down by the rivers from the neighbouring mountains. Its flat monotony is slightly relieved at its upper end by several low hills named respectively Racecourse Hill, Little Racecourse Hill, Gorge Hill, View Hill, and Burnt Hill. Of these the two latter owe their importance to the fact of the rocks consisting of basalt and dolerites; Gorge Hill belongs, on the other hand, to the Mount Torlesse formation, already mentioned; while Racecourse and Little Racecourse Hills consist of shingle. At about four miles from the sea the sand-dunes are encountered—at first ancient dunes, which are separated from the more recent by stretches of low-lying swampy land. These sandhills are formed by the sand blowing inland from the seashore during the frequent and persistent east winds, while the heavy north-west and south-west (if dry) gales blow back again to the shore large quantities of sand, often changing more or less the aspect of the scene.

Meteorology. I have purposely kept back the publication of a number of details shown in the various meteorological tables, which have been issued from time to time by the Colonial Museum and Survey Department under the direction of Sir James Hector, F.R.S., until treating specifically of the plant-formations of each region. Of these tables we need only consider those of Hokitika, Bealey, Christchurch, and Lincoln. These are, of course, so far as they go, of considerable value for a consideration of the effect of the climate of those stations on the plants in their immediate vicinity. Hokitika, however, is not in the district under consideration, and its statistics are only quoted so that some inference can be drawn therefrom as to the rainfall on the dividing range. Lincoln also is not in this district, and, although only a few miles distant from Christchurch, it has, as pointed out by Mr. Meeson,* “On the Rainfall of New Zealand,” Trans. N.Z. Inst., vol. xxiii., p. 546. a rainfall of 2 in. or 3 in. greater than the latter locality. As for Christchurch, observations taken in a town, with shelter from houses and smoke, do not give any good criterion from which to estimate the temperature of the surrounding country. The details from Bealey, on account of its almost subalpine situation, at no great distance from glaciers of considerable size, receiving as it does a large portion, at any rate, of the great western rainfall, are by far the most valuable for plant œcology. From the above considerations it is abundantly evident that our data as to the climate of the region under consideration are not very satisfactory. According to Schimper† “Pflanzen-Geographie auf Physiologische Grundlage,” Jena, 1898, p. 190. there should be given for each month of the year the mean maximum and the mean minimum temperature, the rainfall and the number of rainy days, the mean maximum amount of moisture in the air and the mean minimum, the hours of sunshine, the force of the wind, and the evaporation. But should the meteorological records of any station be of the greatest accuracy and voluminousness, they would only furnish a most general idea of the climatic influences to which a plant is subjected. The side of a gully without sunshine at all during winter—to quote an extreme but quite common case—which even in midsummer receives but a scanty supply compared with the opposite side, and which during the slight frosts of early April remains frozen hard all day, presents an altogether different plant-station to the sunny side. Particular instances of this will be cited, and the differences in vegetation presented by two such sides will be seen to be quite remarkable.

The length of time which the snow lies on the ground and its depth, the shelter or exposure experienced by plants, earth and water temperatures, amount of radiation from differently coloured soils or leaves, these and many more particulars which could be suggested are not to be learned from existing meteorological reports, which treat, of course, only of strictly atmospheric phenomena. My own notes furnish a few details on some of these or similar matters, and will be quoted in due course, but they are, unfortunately, both meagre and disconnected. Notwithstanding what has been said above, the New Zealand meteorological reports furnish us with some valuable information, much of which is given in an admirable condensed form by Sir J. Hector.* “Climate, Temperature, and Rainfall,” New Zealand Official Yearbook, 1894, p. 423. Firstly, as to Rainfall.—The average yearly rainfall from 1864 to 1893 inclusive is—Christchurch, 25.10 in.; Hokitika, 119.91 in. The greatest rainfall for any part of the Waimakariri district which I can find recorded is Bealey, 1878, 155.891 in. in 207 days, while Hokitika for the same year registered 154.446 in. in 259 days, and Christchurch during that year the lowest recorded up till 1893—viz., 13.540in. in 104 days; Taking Lincoln for the years 1890, 1891, 1892, 1893, the rainfall and the number of days on which rain fell was—14.836in., 104 days; 20.575 in., 98 days; 27.883in., 124 days; 22.05in., 115 days; At Bealey, for the eleven years previous to 1879 the rainfall and rainy days was 103.767 in., 174 days; at Hokitika during the thirteen, years previous to 1879 it was 119.047in., 193 days; and at Christchurch during the fifteen years previous to 1879 it was 24.907 in., 117 days. The following extract from Hector's paper (loc. cit., p. 428) is of interest:— Review of the Proportions of Rain in New Zealand.—Percentage of Rainfall for Winter, Spring, Summer, and Autumn respectively. Hokitika 24 28 28 20 Bealey 22 28 31 18 Cbristchurch 31 21 25 23 From the above statistics it may be seen that there falls nearly five times as much rain on the western side of that part of the Island under consideration as falls on the eastern, and that Bealey is well within the region of the western rains. There are also almost twice as many days on which rain falls at Hokitika as at Christchurch, while, if the average amount of rain be considered which must fall upon each of these days, the difference as to climate will appear still more marked. The proportions of rain as quoted for each season show a remarkable equality, still it is plainly evident that the driest

seasons in the west are—(1) Autumn; (2) winter; while springs and summer are almost equally wet. On the east winter is considerably the wettest season, and the remainder are much alike, spring and then autumn being a little drier than summer. Taking the months (loc. cit., p. 428), on the east (Christchurch) the driest is September, 1.161 in., and the wettest June and July, 3.189 in. and 2.449 in. respectively; and on the west (Bealey) the driest month is March, 3.921 in., and the wettest is October, 15.501 in., closely followed by December, 14.087 in., a considerable contrast to the same month (1.622 in.) in the east. Secondly, as to Temperature.—The average yearly maximum temperature for Christchurch is 88.16° Fahr., the average minimum 25.16° Fahr.; and for Bealey 78.08°. Fahr., and 12.38° Fahr. The average temperatures at Bealey are—Winter, 37.40° Fahr.; spring, 46.04° Fahr.; summer, 54.86° Fahr.; autumn, 48.56° Fahr.; for the year, 46.76° Fahr.: and at Christchurch they are—Winter, 43.52° Fahr.; spring, 53.24° Fahr.; summer, 61.52° Fahr.; autumn, 53.60° Fahr.; for the year, 52.88° Fahr. The average daily range of temperature is of interest; unfortunately, I can only quote that of Christchurch:— Deg. (Fahr.). Deg. (Fahr.). January 18.36 July 16.56 February 16.56 August 16.02 March 17.46 September 16.20 April 17.10 October 18.54 May 16.38 November 19.08 June 14.94 December 17.10 Read in conjunction with the above the following mean monthly temperatures at the same station for the years preceding 1877 are important, though the maximum and minimum temperatures for the same period would be more valuable:— Deg. (Fahr.). Deg. (Fahr.). January 62.1 July 42.7 February 61.2 August 44.0 March 58.0 September 48.6 April 53.0 October 53.1 May 47.8 November 56.9 June 43.3 December 62.2 At Bealey during the same period the averages were:— Deg. (Fahr.) Deg. (Fahr.) January 56.7 July 35.8 February 56.3 August 37.1 March 53.8 September 41.6 April 48.6 October 46.0 May 42.6 November 50.4 June 37.1 December 54.9 The following table shows the maximum and minimum temperatures at Bealey for a course of years:—

Year. Jan. Feb. March. April. May. June. July. August. Sept. Oct. Nov. Dec. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Mix. Min. Max. Min. 1871 79.2 75.7 72.6 62 58 53.8 4.4 53.5 53 63 68 73.5 73 1872 82 74 74 73 59 26 55 12 52 13.8 49 14.4 60 22 64 24 77 30.5 78 36 1873 79 34 80.5 36 84 35.5 68 26 58 25 56 16.6 48 20 53 20 56 27.3 72 14.8 69 30.6 78 37 1874 81 33 84 31 73 32 64 27 57 23 52 7.6 50 23.7 53 21.4 54 24 65 28 79 30 74 31 1875 82 38.4 74 40 70 29.4 68 29 58.5 26 58.8 18 52 19 50 20 57 25 69 26 69 31 84 39 1876 81 34 83 36 74 31 69 28 64 22 51 4.8 47.6 16 50 29 55 26 72 30 78 33 77 39 This may be compared with the temperatures of the coastal region as shown in the following table, giving the maxinmim and minimum temperatures at Christchurch for the same series of years: Year. Jan. Feb March. April. May. Jane. July. August. Sept. Oct. Nov. Dec. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. Max. Min. 1871 86.9 76.8 82.4 71 74.1 53.1 24.9 65.1 64.5 65.5 72.4 78.8 82.3 1872 95.7 94.1 81.8 72.2 68.5 32 64.9 21.5 58.6 27.5 58.2 23.7 74.2 30.2 74.1 30.3 86.8 39.9 92.3 41.8 1873 85.8 42.2 75.3 43.4 80.3 41.4 63.2 30.6 68 29.8 68.2 258 55 28.3 66 27.5 62 30.9 83 32.5 78.5 39.5 83 43 1874 88.8 41.5 81.8 40.1 88 39.6 81.6 30.6 64 30.3 59.5 26.9 62 26.1 63 28.5 69.2 28.1 70.3 33.7 80 44.1 82 43.6 1875 85.2 41.4 80 41.6 80.3 35.8 71.7 36.7 65 32.3 60.4 28.3 62.4 28.5 58.3 29.8 68 31 72 31.8 77.5 32.8 84.8 43.7 1876 82.5 40.7 82 48.5 76.4 39.2 72 36.5 77.5 26.5 68.4 24.6 59 23.3 68 27.6 70.8 29.1 77 32 78 36.7 79.4 39.2

This last winter—1899—has been of exceptional severity, and is referred to again (see below in “Acclimatisation”); here I only quote a few temperatures: Sand-dunes one mile from the sea, in Tarata Garden, the thermometer, placed on the ground during June and July, fell several times below 20° Fahr., and on the 26th July registered 15.8° Fahr.; at Hororata—altitude 240 m.—on the same night the minimum temperature registered by a thermometer at Sir John Hall's, placed against the wall of a house facing east and at a height of 44 cm. from the ground, was 6° Fahr. Had the temperature been taken on the grass it would probably have registered 0°Fahr. The following are some of a few scattered observations which I have made from time to time in the Waimakariri district: At Springfield, altitude 387 m., thermometer on grass on 17th June, 1897, at 8.45 p.m., registered 20.75° Fahr., and at 9.30 p.m. 19.4° Fahr.; on 18th June, at 7.45 a.m., in same position, it registered 19.4° Fahr. On 18th June, 1897, temperature on ground (shingle), at 12.30 p.m., with wind blowing from S.W., near summit of Porter's Pass, 944m. altitude, was 39.2° Fahr. At Castle Hill, altitude 723 m., during night of above day, minimum was 10.4° Fahr.; on 19th June, during from 8.30 a.m. to 12.30 p.m., maximum was 48.2° Fahr., and minimum 15.8° Fahr.; at 12.30 p.m. it was 40° Fahr. On Longspur Hill, at 1,010m., facing N.E., at 2.20 p.m. on same day on ground, it was, in sunshine, 37.4° Fahr., and under 5 cm. of snow 31.45° Fahr. At 1,220 m. on Longspur Hill, Craigieburn Mountains, a thermometer left at head of a small shingle-slip facing south during nights of 19th and 20th June, 1897, showed minimum temperature of 17.6° Fahr., almost exactly the same degree of cold being registered below at altitude of 723 m. The following temperatures were recorded on Arthur's Pass during December and January, 1897–98, the thermometer on the ground in the midst of clumps of Celmisia armstrongii. Maximum. Deg. (Fahr.). Manimum. Deg. (Fahr.). Dec. 19 59 46.4 " 20 67 41 " 21 68 44.6 " 22 68 28.4 " 23 33.8 " 24 89 6 36.5 " 25 105.8 44.6 " 26 34 Dec. 27 26.6 " 29 94 42.8 " 30 92 45 " 31 67.5 36.5 Jan. 1 69.8 46.4 " 2 75.5 42.8 " 3 73.4 50 Thirdly, as to Moisture in the Atmosphere.—Saturation = 100. The average per month at Christchurch, Bealey, and Hokitika previous to 1877 was as follows:—

January. February. March. April. May. June. July. August. September October. November December. Christchurch 73 75 77 77 82 86 83 79 78 72 73 73 Bealey 71 71 75 80 81 86 89 84 84 78 78 76 Hokitika 81 85 85 86 86 88 89 87 84 86 83 85 Fourthly, as to Direction and Force of Wind.—The average daily force in miles for the year and the maximum velocity in miles in any twenty-four hours at Christchurch and Bealey were as follows:— 1868. 1869. 1870. 1874. 1875. 1876. 1877. 1878. 1879. Christchurch 146–490 137–420 133–519 140–1319 116–691 164–566 167–554 142–656 Bealey 97.6–509 90–581 133–750 136–395 142–399 174–654 191–552 232* For ten months only.-665 I do not think that these figures are very reliable; at any rate, they do not bring home to us the often great violence of the wind. Sir J. Hector, in a note to the Meteorological. Report for 1873, writes as follows: “But the local modifications of both force and direction of the wind, due to the influence of mountains and gorges, affect the results even more seriously than the disturbances referred to by the author—Dr. J. Hann. The only reliable observations available in any part of New Zealand are those taken from the motion of the clouds.” The next table, showing the frequency of the various winds, is of great œcological importance. It must be borne in mind that the north, west, and north-west winds are rain-bringing winds in the western part of the region and dry hot winds in the eastern, especially on the Canterbury Plains. The south-west wind is a rain-bringing wind in the east—often snow on the mountains—but the rain comes in great gusts, and is very cold; when unaccompanied by rain it exercises a great check upon vegetable growth, and will, by causing excessive transpiration, blacken young growth of leaves and buds just after the manner of frost. It also often blows for three days at a time, occurring usually after a nor'-wester, and lowering the temperature many degrees in a few hours. The north-east and east winds are not nearly so violent as the south-west and north-west, but they are very steady, also raw and cold, and have a considerable mechanical influence on plant-structure as well as promoting transpiration. The Christchurch records only give an idea of the climate near the sea; more inland

the north-west wind is much more frequent. Often a north-west wind blows in the eastern lower mountain region, while in the lowland maritime region the east wind blows steadily. N. N.E. E. S.E. S. S.W. W. N.W. Calm. Bealey—    1869 0 39 1 23 2 32 0 177 91    1870 1 37 1 22 1 25 0 223 55    1871 0 34 0 22 1 21 1 220 66    1872 0 26 1 33 0 22 0 208 76    1873 0 45 7 63 0 35 0 164 51    1874 1 31 10 56 1 22 1 194 48    1875 12 15 18 37 19 16 19 181 53    1876 66 25 21 61 17 15 17 106 45    1877 64 23 6 35 12 15 12 120 81    1878 103 23 9 23 26 26 26 90 50    1879 57 30 10 50 8 12 8 121 67 Christchurch—    1869 4 48 113 16 9 127 14 23 11    1870 9 43 95 22 23 135 11 21 6    1871 7 105 68 9 21 124 12 14 5    1872 10 90 78 21 15 105 7 30 10    1873 5 97 51 39 21 105 1 39 7    1874 1 121 31 29 11 130 1 41 0    1875 3 96 45 36 8 133 9 35 0    1876 7 101 41 16 9 149 10 32 13    1877 4 128 28 13 7 149 11 25 0    1878 3 117 33 6 4 138 18 46 0    1879 3 94 70 17 5 144 11 21 0 To the heading “Calm” the following note is appended: “These returns refer to the particular time of observation, and not to the whole twenty-four hours, and only show that no direction was recorded for the wind on that number of days.” It will be plainly seen, from reference to the above table, that the prevailing wind of the western region is north-west, and that the prevailing winds in the extreme east are, in order of frequency, south-west, north-east, and east, all more or less cold winds, as before mentioned. Any account of the climate of the Canterbury Plains would be incomplete without special reference being made to the north-west wind, since it influences vegetation to such an extent, even in the most shady and sheltered positions, as to entirely prevent the growth of certain hygrophytes—e.g., Hymenophyllums, Todea superba, &c., and they can only be cultivated in Wardian cases, and the like. Sweeping through the river-gorges, it bursts with great fury upon the plain—a hot dry wind, its progress marked by clouds of sand and silt rising out of the bed of the Waimakariri and the other great rivers. In the distance clouds can be seen over the mountains, indicating the rain-storm that is raging there; but over

the plain is a clear blue sky, while a hot sun strikes down. On plants the leaves hang flaccid, in orchards the trees are stripped of their fruit; everywhere the surface of the ground becomes dry as dust. On the dunes clouds of sand are blown back to the sea, sandhills are bodily removed, and the rope-like stems, several metres in length, of Desmoschœnus laid bare. In openings in the mountains, such as Porter's Pass, the fury of the storm is something to encounter. Mr. J. Rountree, for many years driver of the West Coast mail-coach, tells me that the coach is sometimes forced to halt near the summit of the pass, unable to proceed except with great risk of being overturned, and that small stones are meanwhile carried with great force through the air. When the north-west wind rages on the mountains, it is frequently impossible to stand upright on a ridge exposed to its full blast. Where such a wind blows even heavy seeds need no special provision for their distribution. Fifthly, as to Radiation.—Hector writes (l.c., p. 429), “The effect of the prevalence of clouded sky is best illustrated by the average difference between the readings of the black-bulb maximum thermometer in the sun and of the minimum thermometer exposed to the night-sky, and for this purpose two stations on either side of the Southern Alps may be selected:— Christchurch. Hokitika. Insulation. Radiation. Difference. Insulation. Radiation. Difference. Summer 131.72 44.78 86.94 84.02 48.38 35.64 Autum 111.92 37.94 73.98 73.04 41.72 31.32 Winter 91.22 28.04 68.18 61.70 33.44 28.26 Spring 124.25 34.34 90.18 75.02 39.56 35.46 Extremes 158.00 14.54 143.46 97.64 21.92 75.42 Acclimatisation. The behaviour of introduced exotic plants with regard to hardiness, rapidity of growth, variation in form both externally and internally, or of change with regard to reproduction and the like under their changed climatic conditions, is of great value for furnishing details regarding climate. Information of this kind would be an interesting contribution to plant geography in general, and the subject is well worthy of extended treatment. Here I can only deal most briefly with this matter, confining my remarks for the most part to the question of hardiness, and selecting first of all a few common plants.

The potato is sometimes cut by frost, even so late as November, within one mile of the sea; it is grown with success at times at so cold a station as the upper roadman's hut, Bealey Valley, altitude 762 m. The French bean is grown, at any rate, up to an altitude of 400 m., but at both Castle Hill and Bealey it is killed by frost. The vegetable-marrow has been grown successfully both at Mount White and Bealey. Pinus insignis is the common shelter-tree all over the plains; it is quite hardy at Castle Hill, but of much slower growth than in warmer localities. The same remarks also apply to Cupressus macrocarpa. The grapevine, unless when trained against a wall, will not bear fruit in the open; under this treatment it will ripen its fruit even in the central portion of the plains. Ulex europæus is a weed of grass-paddocks, sand-dunes, and river-beds, frequently growing to a height of 4 m. or more, and forming impenetrable thickets. At an altitude of 600 m. it ceases to become a weed, the climate perhaps being too severe for the seedlings with their different leaf-structure to the adult, and, although growing well enough when planted, it does not reach any great dimensions. Hypochæris radicata ascends to the grassy alpine meadows, occurring even in the more stony ground. The laurel regularly ripens its fruit at Craigieburn in the garden of the homestead, altitude 605 m. The gooseberry has naturalised itself abundantly in the Castle Hill Fagus forest, in clearings, and bears abundance of small fruit. Eucalyptus globulus* This exceptional winter, 1899 (vide ante), has killed the leaves, at any rate, of nearly every blue-gum tree in the middle and upper portions of the Canterbury Plains, some of which must have been more than thirty years of age (see Appendix, where are also some observations regarding the laurel). grows well at 500 m. altitude, at the base of Mount Torlesse. Speaking generally, plants of the cold temperate zone will thrive anywhere in the lower mountain and lowland regions, while species belonging to the Mediterranean and Californian floras are usually hardy up to 400 m. Australian and South African plants, those of the higher regions excepted, are not usually hardy even near the sea. The Cape of Good Hope silver-tree, hardy in hilly stations near Dunedin, will not endure even the mildest winter of the Canterbury Plain. Pelargoniums and Cape heaths require the protection of a wall near New Brighton, although quite hardy at Summer, only a few miles distant, where they are sheltered by the Port Hills from the south-west wind. Although not nearly so favourable a locality as many others in this Island—e.g., Dunedin—exotic alpine plants can be grown even near the sea-coast as easily as in England,

judging from accounts of their culture in the Gardeners' Chronicle and the Garden; indeed, such are often much more amenable to cultivation than are those indigenous to our Alps, many of which can only be grown with the greatest difficulty. Silene acaulis, Dianthus alpinus, D. neglectus, Papaver alpinum, Viola calcarata, Gentiana verna, G. acaulis, Leontopodium alpinum, Anemone sulfurea, A. vernalis, Androsace lactea, Primula farinosa, P. longiflora, P. auricula, Arnica montana, Aster alpinus, Saxifraga aizoides, S. aizoon, S. exarata, Draba aizoides, Campanula pusilla, to quote some examples from the Swiss alpine flora, all thrive or have thriven in the Tarata Garden. Many other alpines—Himalayan, Andean, Siberian, or North American—could also be quoted. The successful acclimatisation of any plant belonging to exotic temperate regions in the lowland or lower mountain region seems to depend less upon temperature and rainfall than upon the nature of the soil, and shelter from the prevailing winds. In certain parts of the Canterbury Plain—e.g., near the lower Waimakariri Gorge—the force of the north-west wind has been sufficient to blow away from a newly sown wheat- field both surface-soil and seed. It is in many instances rather New Zealand plants themselves which are difficult to acclimatise than exotics. Numbers of North Island forest-trees, some from the neighbouring Port Hills, and others from Stewart Island, far to the south, are damaged or killed outright by the cold even near the sea.* L. Cockayne: “An Inquiry into the Seedling Forms of New Zealand Phanerogams” (Trans. N.Z. Inst., vol. xxxi., p. 359). In fact, the winter climate of the lowland region, or perhaps of the Canterbury Plain as a whole, is more severe than any other similarly situated region in the South Island, nor is the maritime region much milder than sheltered inland stations at 300 m. altitude. Phenology. I am not aware that phenological records have ever been published regarding any part of New Zealand.† Since writing the above I have found an interesting account of the blooming of New Zealand orchids in the Thames district, for which see Trans. N.Z. Inst., vol. xvi.: “On the Botany of the Thames Goldfield,” by J. Adams, B.A. Such are of extreme value as a measure of climate. Bailey‡ L. H. Bailey: “The Principles of Fruit-growing,” New York, 1897, p. 127. writes, “Such records are more accurate measures of seasonal climates than instrumental measurements are. Some day—referring to the United States—” the country will have charts of

isophenal lines as well as of its isotherms.” For these records to be of value they should be kept for a series of years at different stations. The leafing and defoliation of common deciduous trees, the flowering of the same varieties of fruit-trees, the blooming of Cordyline australis, Arundo conspicua, and, Phormium tenax; these and many more equally easily observed periodic phenomena would form an excellent record. I have kept, but unfortunately too irregularly to be of much use, a record for several years of the blooming of many of the plants in my garden. A few extracts may be of interest:— Chionodoxa sardensis bloomed on the 21st August, 1895; 10th August, 1896; 6th August, 1897; 19th August, 1899. Narcissus nanus bloomed on the 13th August, 1894; 27th August, 1895; 13th August, 1896; 25th August, 1899. Narcissus jonquilla bloomed on the 28th August, 1895; 5th September, 1896. Anemone nemorosa bloomed on the 5th September, 1896; 19th September, 1895. Veronica formosa bloomed on the 11th October, 1893; 6th October, 1894; 14th October, 1899. Clematis montana bloomed on the 17th October, 1893; 25th October, 1895; 28th October, 1899. Aquilegia thalictrifolia bloomed on the 3rd November, 1894; 3rd November, 1895; 29th October, 1899. Potentilla pennsylvanica bloomed on the 20th December, 1893; 22nd December, 1895; 25th December, 1899. Kniphofia macoweni bloomed on the 5th January, 1893; 3rd January, 1895; 10th January, 1899. Of much greater value is a comparison of the blooming, &c., of the same plants at different elevations and in different regions. The following may be accepted as fairly correct:— Strawberries are ripe near the seaboard at the beginning of December; at the Bealey (640 m. altitude), about the first week in February. The double daffodil blooms in the Christchurch district during the first and second weeks of September, and at Castle Hill (723 m. altitude) towards the middle of November or a little earlier. Fagus cliffortioides comes into new leaf-growth in Tarata Garden, New Brighton, towards middle of October, and at Castle Hill towards first or second week in November. Ranunculus lyallii blooms in cultivation near Christchurch from the beginning of October, on Arthur's Pass (911 m. altitude) at end of November and beginning

of December, and near the Waimakariri glaciers (1,219 m. altitude) from the middle of January to the middle of February. Clematis australis blooms in Tarata Garden during middle of October; at Bealey at end of November or beginning of December, and on Arthur's Pass at beginning of January. Celmisia bellidioides bloomed in. Tarata Garden 2nd October, 1895, and on Arthur's Pass at end of December, 1897. As a general rule New Zealand alpine and subalpine plants bloom in cultivation near sea-level from a month to a month and a half earlier than in the mountains at an elevation of from 750 m. to 1,000m. The time at which these alpine plants bloom in their natural habitats is probably dependent not on the temperature of the air, but upon the time when the snow melts, and it seems an hereditary habit to bloom at a fixed time after such melting, each species having its own limit of time, some longer some shorter, but dependent in large measure, I think, on the preceding year's climate having been sufficient to produce the incipient bud up to the requisite state of development; while the blooming of the same species in cultivation exposed to a mild winter and spring, and with no covering of snow, must depend upon heredity varied by the climate of the current year. Some of the alpine plants are so precocious as to put forth their blooms even before the snow is quite away—e.g., Caltha novæ-zelandiæ. Buchanan writes,* “On the Alpine Flora of New Zealand” (Trans. N.Z. Inst., vol. xiv., 1881, p. 343). “The intense heat of the sun at high altitudes is, no doubt, an important element in hastening growth, but the chief cause must be ascribed in many cases to the advanced stage at which the plants have arrived before the melting of the snows in spring has uncovered them. Large plants such as Ranunculus buchanani were found 8 in. to 10 in. high, with the leaves and flower-buds fully formed.” The phenomenon of Euphrasia partly developing its blooms while in the freezing-chamber at Lyttelton seems also a case in point.† L. Cockayne: “On the Freezing of New Zealand Alpine Plants” (Trans. N.Z. Inst., vol. xxx., p. 439). Christ‡ H. Christ: “La Flore de la Suisse et ses origines.” Édition Françise, Bale, Genève, Lyon, 1883, p. 376. describes similar phenomena in the Swiss Alps thus: “Avant que la surface du sol, comprimée par le poids de la neige, ait commencé à se réveiller de sa torpeur, avant même que les bourgeons jaunâtres des herbes aient commencé à pousser on voit s'ouvrir tout près de la neige, et parfois dans la neige même toute une sérée de

fleurs des plus fines et des plus délicates, dignes au plus haut degré d éveiller, notre attention et notre sympathie.” Climatic Regions and Plant-formations. It is well known that many plants are accustomed to grow in company with other plants, and that such communities, or plant-formations as they are often called, occur again and again in different parts of a district without their constituents varying to any great degree. Thus a forest having Fagus cliffortioides as its dominant tree may be expected to contain certain other definite plants as undergrowth, while on the trunks of the trees themselves the same mosses and lichens will usually be found. If a portion of such forest be examined in any part of a particular district, noting carefully its plant-members and their relative frequency, it will be found that such a forest will be typical of all others of that class, and that if any marked change occur it will be in a region of considerable difference in climate, soil, or altitude. Again, a stony river-bed will carry certain plants, and such a station, subject to the provisos indicated above, will have the same vegetation, no matter in what part of a district it may occur. Into groups such as these may the flora of a region be divided, and so the relations of each group of plants as a whole with regard to moisture, soil, light, heat, air, and animals be conveniently studied, and the resultant life-forms of the component species noted and considered. From the foregoing it must not be imagined that a plant-formation is something definite and invariable; on the contrary, hard-and-fast distinctions cannot be made. Transitions occur between most; certain plants enter into a formation in one place which are entirely absent in another, or a plant may have crept in which does not belong to the formation at all. Thus in the Fagus formation near Arthur's Pass occurs a solitary example of Olearia lacunosa, a plant really belonging to the upper limits of the Westland forest, or sometimes forming a small percentage of the subalpine scrub of that region. Nevertheless, this classifying the plants of any district into communities which resemble one another so much in their adaptations to their environment that they usually grow together is well fitted for the object aimed at—viz., a study of such adaptations—while also from the floristic point of view it is a concise method of mapping the vegetation. One great difficulty is to properly estimate the effect of the various external factors on plant-life, and statements must necessarily be often based on very uncertain and incomplete observations, especially so indeed in a “new country,” where even the most general details as to temperature, composition of soil, rainfall, &c., are not accurately known. Two factors

determine plant-formations—climate and the nature of the soil, the former affecting wide areas usually, and the latter being often very limited in its operation. This has recently been termed by Schimper “edaphic” (edaphisch) (loc. cit., p. 5), and he characterizes the formations as climatic and edaphic respectively. In what follows I shall use the term “formation”* An account of the meaning of the term “plant-formation” as used by different authors, and an admirable exposition of the aim of æcological plant geography, may be seen in “Natural Science,” vol. xiv., p. 109, in an article by Robert Smith, B.Sc., entitled “On the Study of Plant Associations.” I have purposely avoided this latter title, and kept to the older one of “plant-formation,” since Schimper uses the term—“Die Genossenschaften”—to indicate the more intimate relations between lianes, epiphytes, saprophytes, parasites, and the plants with which they are associated. to include the more or less well-marked smaller communities, while the larger will be called “plant-regions,” and will, of course, include numerous plant-formations. By Kerner† Kerner and Oliver: “The Natural History of Plants,” London, 1895, p. 896. the term “formation” is restricted to those communities which are intermingled and exhibit a kind of stratification. A forest with undergrowth of shrubs and a carpet of smaller plants on the ground would be an example; to such Drude‡ “Manuel de Geographie Botanique,” par Dr. O. Drude (French translation and revised edition), Paris, 1897, p. 207. gives the name of “formations étagées.” The Climatic Regions. The moisture-laden north-west wind, hurrying over the Dividing-range, deposits there and in the immediate vicinity large quantities of water. Whatever moisture it still contains falls, as it furiously drives along, upon the eastern spurs and slopes of the Dividing-range, the rain reaching a greater or shorter distance according to the quantity of moisture originally in the air and the velocity of the wind. In very rare cases during a north-west storm a slight shower is experienced at the eastern seaboard; more often the summits of the Mount Torlesse Range experience a more or less heavy downpour or only a trifling shower, while under such circumstances the table-land up to an elevation of 900 m. gets no rain at all, but only a drying tempest. Still more often it is only the actual Dividing-range for a mile or two to the east which gets the rain, and sometimes, indeed, none falls on the eastern side, it being confined to the ridge and mountain-tops. As an example of this, Bealey does not get nearly so much wet weather as the country at the source of the Waimakariri. I have seen rain fall when camped at the mouth of the River

Crow, six miles from Bealey, which place experienced fine weather the while, and at the same time the downpour on the high ranges added a considerable increase of water to the river. It is quite a common sight to see a heavy storm in progress up the Waimakariri while the sun shines brightly at Bealey. On Arthur's Pass I have observed heavy rain, while down the Bealey Valley, half a mile away, it was quite fine. This great difference between the eastern and western rainfall (see above, “Meteorology”) has written its mark most unmistakably upon the vegetation. A line can be drawn which certain plants do not cross usually, and marking what I should take to be the average eastern limit of the northwestern rain, and which separates, though not sharply, the district into two great climatic regions—-viz., the western and the eastern. Throughout this and other papers dealing with the same subject I purposely refrain from speaking dogmatically as to the range of plants. The distribution of New Zealand plants is still too imperfectly known to allow definite statements to be made, and most of such that have been published up to the present are not of much scientific value. Ranunculus lyallii, a common western subalpine plant, and one which would not be thought to exist on the dry eastern hills, is found in an isolated station near Lake Lyndon. Metrosideros lucida, another western plant, occurs on dry rocks at Broken Hill.* Another station in this district is the forest near Patterson's Creek, Mount Torlesse, whence I have received specimens collected by Mrs. J. G. L. Scott. Gelmisia lyallii, a most typical eastern plant, with, according to Diels, the structure of a steppe grass (loc. cit., pp. 265 and 268), has been collected in the neighbourhood of Cook River, Westland, by Mr. Wilson, formerly District Surveyor, Hokitika. These extreme cases show that it may be possible to find any eastern plant in the western region, and vice versd, in one or more isolated stations, and the same reasoning applies to north and south limits and altitudinal range for New Zealand plant geography generally. But this should not hinder us from designating certain plants as “eastern,” “western,” “northern,” “southern,” “alpine,” “lowland,” and the like with perfect propriety. The other climatic divisions of these two regions depend upon altitude, and may be classed as “lowland,” “lower mountain,” “subalpine,” and “alpine.” All four occur in the eastern region, but only the two latter to any extent in the western, so far as the Waimakariri system is concerned. The Trelissick basin, a part of the eastern lower mountain region, almost forms a distinct climatic region, owing to its peculiar position with regard to the

mountain-ranges which bound it and set down both the north-west and the south-west rains. Thus it often rains at Lake Pearson when the higher ground of the Trelissick basin is rainless; and, again, it may rain on the Canterbury Plains and on Mount Torlesse and yet the region under consideration be dry. Perhaps the Trelissick basin bears more resemblance æcologically to Central Otago than to any other region in New Zealand.* The plants of Central Otago are distinguished in Petrie's list by the mark C (Trans. N.Z. Inst., vol. xxviii., 1895, p. 540). The Eastern Climatic Plant-region.—The æcological character of this region is essentially xerophilous.† Plants of the pine forest and certain marsh plants form an exception to this; also, the New Zealand Fagi are classified by Schimper as tropophytes, on account, I suppose, of the conspiouous young-leaf growth during early spring, and the often deciduous habit of the most exposed branches. This is not to be wondered at when the small rainfall and constant drying winds in conjunction with the usually stony soil is considered. To be sure, there is much swampy ground, but such to many of its inhabitants is physiologically dry. Schimper writes (loc. cit., p. 4), “In der That ist ein sehr nasses Substrat für die Pflanze vollkommen trocken, wenn sie ihm kein Wasser zu entnehmen vermag, während ein Boden, der uns vollkommen trocken erscheint, manche genügsame Pflanze hinreichend mit Wasser versorgt. Es muss zwischen physikalischer und physiologischer Trockenheit, bezw. Feuchtigkeit unterschieden werden; letztere allein kommt für das Pflanzenleben, also auch für die Pflanzengeographie in Betracht.” Diels, in a long list embracing plants of varying stations (loc. cit., pp. 216, 217) with regard to moisture, classifies them all as hygrophytes, evidently more from their station than from their structure. Phormium tenax, e.g., usually a denizen of swamps, occurs also on dry river-terraces, on sand-dunes, and even on dry faces of rock—upper Waimakariri Gorge. Leptocarpus simplex, a curious restiaeous plant lives equally well either in brackish marshes or on sand-dunes. In cultivation in my garden, on an almost perpendicular rockery, exposed to the full rays of the sun and to the north-west wind, Epilobium macropus, whose usual habitat is cold running water in subalpine streams, thrives amazingly, yet receiving exactly the same treatment as Paronychia argentea‡ Nicholson for this plant recommends a light dry soil: “The Illustrated Dictionary of Gardening,” London, vol. iii., p. 25. of the Mediterranean region, side by side with the former, and in excellent health. Claytonia australasica grows both in shallow running water and on shingle-slips.

Besides these many other instances could be quoted of bog plants or those of very wet stations growing on dry ground.* See also Goebel, “Pflanzenbiologische, Schilderungen,” zweiter teil, Marburg, 1891, pp. 46, 47, where he quotes certain plants which are found in Germany in bogs, but in the arctic regions on dry hills; and species of Espeletia, in the Venezuelan Andes, grow both in marshes and on rocks. The stony character of the ground, with a gravelly subsoil at a greater or lesser depth (see Plate XIII.), has been mentioned when treating of the topography of this region. On the mountains clayey loam is often found, varying much in its percentage of clay; also very stiff clay is met with, but always underlain by rocky débris, and so subject to rapid desiccation. In addition to these are characteristic xerophilous stations, such as rocks, stony river-beds, salt meadows, sandhills, and shingle-slips. The rivers, as pointed out before, flow in beds often deep down below the surrounding country, much of which consists of steep slopes, so that every facility is provided for natural drainage. To support a luxuriant vegetation in such a region would require a moist atmosphere and a considerable rainfall. On the contrary, the rainfall on the plains near the sea is little more than 20 in., increasing, of course, as the hills are neared; on the alpine heights and at the north and west boundaries it is much greater still, but possibly decreasing to below that of Christchurch, in the Trelissick basin. In addition to this the wind is nearly always blowing, sometimes a violent, hot, dry nor'-wester, sometimes a cold, dry son'-wester, at other times a steady, drying, east or north-east wind. The sun, too, is hot in summer, and cloudless days prevail. In winter clear frostless days with warm sunshine are common. Excessive weather, too, is not unknown; two periods of drought have occurred during the past ten years. In 1895 the snow lay near the sea for more than a week, and skating on ice was possible in that neighbourhood, while on the table-land and the mountains the snow lay for many weeks longer than its average period. This winter of 1899 has been specially referred to in dealing with the meteorology, but other periods of cold almost as great have occurred during the past twenty years. Low-lying hollows in the alpine region are filled with water one month and dry the next. The stones on shingle-slips become so hot that it is unpleasant to touch them with the hand, while the night of that same day may witness a frost. In such a station the plants are often frozen quite hard before being protected by their winter covering of snow, and while in this state are exposed to drying winds. Some plants, again, grow in stations where the winter snow can never protect

them, on, e.g., the final rocky precipices of a mountain peak. These examples may give some idea of the climate that the plants of this region must endure in many instances. To the above conditions of soil and climate the plants have responded in various ways. The details of such adaptations will be entered into more fully when dealing with the constituents of the formations; here it is only necessary to point out a few of the more common adaptations assumed by plants frequently enough to affect the general plant physiognomy of this region. Where constant and often furious winds blow it is necessary that the plants should protect themselves in some way or another from being broken or torn to pieces. But such winds, besides acting in this mechanical way, also indirectly exercise a most powerful physiological effect upon plant-life by causing excessive transpiration. And how enormous must such transpiration be during a hot nor'-wester! That plants even when frozen can be killed by excessive transpiration is pointed out by Schimper (loc. cit., p. 45). Nor is such transpiration confined to leafy plants; even deciduous trees in the cold winter of Michigan, U.S.A., may be damaged through excessive transpiration through the bark (Bailey, loc. cit., pp. 13–18). Bearing these facts in mind, the cold high winds of winter must also have a powerful influence upon plants. Many of the commonest forms of our plants are adaptations against the effect of wind, and serve at the same time both to resist its violence and to reduce transpiration in many cases to its smallest limit. Reduction of transpiration is also controlled by other means, to be treated of in due course. One very common form which meets the end in view is a round ball-like growth of the entire plant, and is to be seen in great perfection in numbers of shrubby plants—e.g., Veronica traversii, Hymenanthera alpina, T. Kirk (var.), and Plagianthus divaricatus, this latter growing in salt meadows, a station physiologically dry. In the cases cited the very close-growing, divaricating, often intertwining branches shelter very considerably both themselves and their leaves, presenting a surface which the wind cannot damage, and so ball-like are many—e.g., Veronica odora—that they look as if trimmed into that shape purposely by a gardener; indeed, few things are more strange than to see a clump of these perfectly globe-shaped bushes with their varnished green leaves, the whole looking like some portion of a cultivated shrubbery, away up a distant river-bed, where man may never have set foot before. Even in sheltered places in cultivation these plants still maintain this habit, which is evidently quite hereditary. The well-known tussock form of certain grasses and sedges serves the same purpose, and is a most distinguishing form of the entire

region. Such close growth of branches is carried to its extreme limit in Raoulia eximia (Plates XI., XII.), an inhabitant of rocks facing north and north-west in the alpine regions, whose branches are so close that they form a compact coral-like mass, upon which one can sit down without affecting its shape in the slightest degree. Excessive transpiration is also combated in many plants by reduction of their leaf-surface, the most obvious method of which is by reducing the size and number of the leaves. This is carried to such a pitch that some plants have few or no leaves at all. Among the most common are the Carmichaelias of this region, whose assimilating functions, &c., are performed entirely by the stems, in which a most excessive stereome tissue produces great inflexibility and wind-resisting power. Another very common method of avoiding rather than combating the wind is low or prostrate growth, a very common feature in the alpine region, and seen in the lowland and lower mountain region amongst the Carmichaelias, with stiff, erect, leafless branches 2 cm. or 3 cm. high, which form large patches on the stony ground, or in the trailing, almost leafless Muhlenbeckia ephedrioides, or, most of all, in the moss-like river-bed Raoulias—a most efficacious method, since it is well known that the force of the wind is much stronger at some distance from than close to the ground. Plants covered with dense masses or mats of hair are frequent, by which not only transpiration but extreme temperatures are guarded against. Such plants are especially characteristic of the alpine or subalpine region. Compare, e. g., Craspedia alpina of the shingle-slips, quite snow-white with its thick covering of woolly hairs, with its much less hairy relative C. fimbriata, var., of lowland swamps, or with its thin-leaved almost hygrophilous relative of the west. Numerous grasses have rolled leaves. Other plants have the leaves incurved or recurved, all methods for reducing leaf-surface. But I will enter into no more details now, except to point out that Diels was much struck with the extreme xerophilous character of many plants, which he considered out of all proportion to any severity of climate they have now to endure, for, according to him, at the present time the driest regions of New Zealand are less arid and possess a more equable climate than middle Europe—“dass selbst die trockensten Stricbe Neuseelands unter minder excessivem Klima und seltener Dürren leiden als Mitteleuropa” (loc. cit., p. 247); and so he considered Carmichaelia, Hymenanthera, Corokia, and some others to be descendants of a forest flora which had been forced to retreat northwards during a rising of the land which led to the

formation of a dry easterly steppe region, where the survivors of the forest had become modified, and assumed the structure and physiognomy of desert plants. I shall refer to this again when treating of these plants, but I think the observations recorded in this paper show a climate in the east quite sufficient even at the present day to account for structure so xerophilous as that of Carmichaelia nana. That these adaptations are really for the purpose of resisting drought, and that they have been evoked by a dry climate, has been more or less proved in certain instances. Goebel was the first to show that Veronica cupressoides, when cultivated in moist air, produced true leaves, with stomata on both surfaces, and which were strictly of an extreme hygrophilous type.* Pflanzenbiologische Schilderungen Erster Teil, Marburg, 1889, pp. 19, 20. At about the time of the publication of this experiment Mr. R. Brown was engaged in similar experiments here. He cultivated pot plants of Raoulia tenuicaulis, Ozothamnus microphyllus, and Veronica armstrongii in his greenhouse, the glass of which was not shaded, with the result that all produced leaves differing much from the wild plants.† Regarding Brown's work, see also Cockayne (loc. cit., p. 360). I have repeated Goebel's experiment, and find that Veronica cupressoides is so unstable that it will produce these reversion leaves in six weeks' time when cultivated under a bell glass and kept constantly moist. Under the same conditions all the other whipcord Veronicas—V. lycopodioides, V. hectori, &c.—will equally rapidly produce reversion leaves. Nor is it always necessary to give such moist treatment as the above. Rooted cuttings placed in a fairly sheltered spot in a flower border will revert so far as the young growth is concerned, though such reversion will be soon succeeded by normal growth. In my shade-house this spring both V. tetragona and Ozothamnus microphyllus are rapidly developing true leaves. Other plants with adaptations against drought of other kinds will also under similar treatment change their leaf-form. Olearia cymbifolia, planted in the open border, will often produce a growth of flat leaves, instead of the normal leaf with strongly revolute margins. This change seems to have a tendency to remain more or less permanent, a plant growing on the extreme summit of a rockery in my garden having leaves not nearly so revolute as the type. Another plant in the shade-house for nearly twelve months has now every leaf flat. On the other hand, plants growing in cultivation in a dry part of the garden show no change. The similar behaviour of Carmichaelia (any leafless species), Raoulia eximia, and R. bryoides could also be cited.

Although the artificially produced leaves mentioned above are reversions to the seedling form,* Cockayne., loc. cit., p. 357. and therefore hereditary, all the same they are in such experiments as these evoked directly by the environment, and they show how xerophytes may be changed into hygrophytes by excessive moisture in the air. Soon after these hygrophilous conditions are removed a plant with reversion leaves will develope ordinary shoots once more. The plant of Veronica cupressoides used for the experiment quoted above, although still in a fairly moist atmosphere (that of a greenhouse), is producing normal growth. Whether in cases such as these, where both the reversion leaves and the ordinary leaves, or cladodes, are hereditary, continual treatment under artificial conditions would keep the plant permanently changed, or whether in a moist climate such as that of Westport or Greymouth the reversion form would endure under natural conditions, can only be ascertained by experiment.† From what is said above it seems evident that the two cases (Olearia cymbifolia and Veronica lycopodioides) mentioned by Captain Hutton in “Darwinism and Lamarckism” (New York and London, 1899, pp. 217, 218) are both examples of one and the same phenomenon, and not—the former an example of an acquired habit which had not become hereditary, and the latter an example of an acquired habit which had become hereditary. Generally speaking, the plant-life of the whole of this region is scanty. It is only in certain places where shelter, richer and deeper soil, or moisture favour growth that anything like luxuriance may be seen. To the casual observer a hillside in the lower mountain region would seem clad merely with dry, brown tussock-grasses, but careful search will reveal a number of lowly plants growing in the shelter that such tussocks afford. The same remarks apply to the higher regions, where many of the plants are small, low-growing, often moss-like, and by no means numerous on the expanse of clayey or stony dry ground. Since the conditions of soil and climate do not permit a luxuriant plant-covering over considerable areas, it seems hard to understand how introduced plants can have spread so abundantly. Even in 1869 Mr. J. F. Armstrong recorded 178 for Canterbury, and this list has been since extended in the “Students' Flora of New Zealand.” Now, the land on which such especially flourish is that which has been reclaimed by man, the rich drained swamp lands and their environs, or even the more stony ground ploughed and made suitable for plant-life. In such virgin soil, inhabited formerly by plants which would not require quite the same food-material, introduced plants would flourish and have flourished amazingly. On ground untouched

by cultivation mainly those species have firmly established themselves which are specially fitted by their structure to the environment. Thus, to choose two out of many examples, Verbascum thapsus, a most tomentose plant, is abundantly naturalised on many river-terraces (Kowai River and River Porter, e.g.), and Lupinus arboreus* L. arboreus is furnished with exceedingly succulent stems, made up of abundant water tissue. Its leaves also are hairy on the under surface, and they have the power of reducing the leaf-surface considerably through the two upper sides of the lamina of each leaflet folding together. The leaves exposed to the sun may thus be seen closed up, and those in the shade open wide. on the sand-dunes. Had the numerous weeds arrived before the ground was prepared for them by man they would have met with a much more stubborn resistance in the undrained Phormium swamps, &c. Statements such as that by Wallace,† “On Darwinism,” London, 1889, p. 28. of white-clover wiping out Phormium, should be accepted with great caution. Such swamps even now possess few introduced plants, and remain virtually intact.‡ Every student of New Zealand plant geography should read carefully Mr. Cheeseman's very excellent paper, “The Naturalised Plants of the Auckland Provincial District” (Trans. N.Z. Inst., vol. xv., 1882, p. 268). He has gone into the subject of the spread of weeds at some length, and come to some very just and well-considered conclusions. This eastern climatic region may be conveniently divided into the following subregions:— (1.) The Lowland Region: This embraces the country drained by the Waimakariri, on the Canterbury Plain, and extends from the ocean to the foot of the ranges. It varies in altitude from sea-level to 457 m. Here (excepting towards its western boundary) snow rarely lies for more than twenty-four hours, usually melting as it falls. The rainfall varies on an average from less than 25 in. on the sand-dunes to more than 40 in. at 457 m. This last estimate is a mere guess; but it is a fact that that altitude gets rain from both north-west and south-west which never reaches the sea-coast, and that rainy days must also be much more abundant. The north-west and south-west winds have been already treated of. The former is much more frequent in the western portion of this region, and often blows there with great violence, while a steady east wind is being experienced at the coast and for some miles inland. The rain from the south-west coming in squalls often does not wet the ground on the sheltered side of large plants, so the smaller ones growing there remain quite dry. Frosts may occur during any month except December, January, or February. After rain the ground in most places dries very rapidly. The whole of the region is exposed to the full blasts of the winds, excepting

such small tracts as are sheltered by river-terraces. The only other shelter is from the few plantations of exotic trees—mostly Pinus insignis and Eucalyptus globulus—and gorse hedges which have been planted by the settlers. The vegetation may be divided primarily into the following plant-formations:— Name of Plant-formation. Most Characteristic Plant. 1. The moving sand-dune Desmoschoenus spiralis. 2. The fixed sand-dune Discaria toumatou. 3. The salt meadow Plagianthus divaricatus. 4. The flowing water Myriophyllum elatinoides. 5. The brackish water Zostera nana. 6. The lowland swamp Phormium tenax. 7. The lowland forest Podocarpus dacrydioides. 8. The grassy meadow (* This formation will be subdivided into several minor formations, each with a different characteristic plant, depending in large measure upon the depth of the soil. Large stretches of very stony, arid, barren land are occupied by Leptospermum scoparium, especially towards the centre of the Canterbury Plain, on the north bank of the river.) 9. The stony river-bed Raoulia tenuicaulis. Besides the above formations, several minor divisions will be treated of. It is not always easy to select the most characteristic plant; cases frequently occur in which various plants could be chosen with equal propriety. (2.) The Lower Mountain Region: This varies in altitude from 457 m. to 761m., and includes the foot-hills of Mount Torlesse and Mount Puketeraki (in part);the whole of the table-land, including the Trelissick basin, where it ascends to over 900 m., also the valleys of the Esk, Poulter, and Upper Waimakariri to about 761 m. The climate of this subregion may be gathered from what has gone before, and varies from extremely dry in the Trelissick basin and vicinity of the Mount White homestead to very wet with many rainy days and a considerable amount of fog and cloud at its junction with the western climatic region. The ground is mostly extremely stony, consisting of a great depth of stones, sometimes quite without soil, and repeating on the river-terrace slopes the shingle-slips of the higher levels in miniature. With the exception of the Trelissick basin the whole of this subregion is wetter than the lowland subregion, but the loss of water in the soil, owing to the great depth and slope of the stony ground, is much more rapid. Swampy ground is of frequent occurrence, and streams are abundant. Before joining a main river the tributaries usually stretch out into huge fans, with remains of older fans bounding them. On such a fan coming

from Mount St. Bernard into the Winding Creek are a most surprising number of forms of two species of Veronica. Nume-rous grass-fires, bush-fires, a certain amount of cultivation, and the pasturage of many sheep have in some places changed the character of the vegetation, especially so far as percentage of components is concerned, and by destroying the tussock shelter must have almost or entirely eradicated some of the more lowly plants. The following are the most important plant-formations:— Name of Formation. Most Characteristic Plant. 1. The mixed Fagus forest Fagus solandri. 2. The rook Veronica raoulii. 3. The river-bed Epilobium melanocaulon. 4. The shingle-slip (* The shingle-slips have to be subdivided into limestone and sandstone, each bearing a somewhat different set of plants.) 5. The grassy meadow Tussock of Festuca duriuscula.† Several subdivisions will here also be necessary. Perhaps Aciphylla colensoi should have been given as the principal plant; it is certainly most characteristic of the whole region in many ways, and it gives a peculiar and special physiognomy to the landscape where growing in abundance. 6. The terrace scrub Veronica traversii. 7. The marsh Carex gaudichaudiana. 8. The lake Isoëtes alpina. 9. The running stream Montia fontana. Besides the above formations, other smaller ones will be treated of. Some of the lowland and subalpine formations also extend into this region. (3.) The Subalpine Region: This region is of great interest, and contains many remarkable plants. It is here that the effect of varied stations on plant-form can be especially well studied. For example, it is hard to believe that Epilobium chloræfolium, of a very shady station among shrubs is the same species as the form growing in the open. All the mountainous country between an altitude of 761m. and 1,220 m. found in the eastern climatic region belongs to this subregion. Snow lies on an average for about one month during winter, but for a longer period towards the upper portions of the region, and, of course, in times of a heavy snowfall, for considerably longer still. Frosts may occur during any month in the year. The south-west wind often brings, rain. Not uncommonly is the soil of fair depth, in that case consisting of yellow, rather sandy, clayey loam. Heavy winds are frequent, but on the whole there are more calm days than in the lowlands. The surface of the ground is usually very steep,

except in river-beds and their vicinity and on the mountain passes. Streams are often abundant. As an example of the great permeability of the soil with regard to water, I may mention that the water of a waterfall of considerable volume descending from the eastern slopes of Mount Torlesse sinks completely underground at a few metres below the fall, while at a few metres above there is also no sign of water, except during heavy rain. In this case the water-channel is very steep. Kaka Creek, at the base of Broken Hill, behaves in a similar manner; indeed, for the water of a river suddenly to disappear through its shingle is of quite common, occurrence in all the South Island mountains. The chief plant-formations of this region are:— Name of Plant-formation. Most Characteristic Plant. 1. The Fagus forest Fagus, cliffortioides. 2. The subalpine scrub Dracophyllum uniflorum. 3. The subalpine meadow Celmisia lyallii. 4. The river-bed Epilobium polyclonum.. 5. The running stream Epilobium macropus. 6. The marsh and wet ground (* These will include several minor formations, differing in great measure upon the amount of stagnation of the water.) 7. The rock († The plants of wet rocks, of course, differ from those of dry rocks, so here some minor distinctions will have to be drawn.) The Fagus forest descends into the montane region, where it often occurs on the shady side of river-terraces sheltered from the north-west. The shingle-slip to be mentioned in the next subregion also frequently descends into this. Several smaller formations will also be considered. (4.) The Alpine Region: This extremely interesting and important region extends from 1,220 m. to 2,134m. It includes all that portion of the mountains which lies above the winter snow-line, consequently its plants undergo pressure by the snow and exclusion from light unknown to those of lower regions. The rainfall is certainly greater than that of the subalpine zone, and the wind on the exposed portions of the mountains much stronger. Frost will be more severe than in the subalpine region, but no statistics exist on this head as yet with regard to New Zealand. According to Schimper (loc. cit., p. 727), the air-temperature decreases at the rate of 0.58° C. for each 100 m. This would give a difference of 5.2° C. or 9.3° Fahr. between the temperature of the highest and lowest parts of the region, so far as the air was

concerned, or a difference of 14.6° between the middle of the Trelissick basin (609 m.) and the summit of Mount Enys (2,134m.). This excess of cold is met by the protecting action of the snowy mantle which covers all the alpine region except certain precipitous rocks near the very summits. Even these unprotected spots have their phanerogamic vegetation* Similar examples are given by Christ (loc. cit., p. 295) from Switzerland, where on the Col de St. Théodule, which has an average temperature for the year of -5.59° C., a minimum of -21.4° C., and a maximum of 15.1° C., with an altitude of 3,333 m. above the sea, thirteen species of flowering-plants grow.; on one such in the Craigieburn Mountains, at an altitude of 1,980 m., I have noted Cardamine enysii and Veronica epacridea growing in profusion, in places during January the rocks being white with blossoms of the former. Summer snowstorms are frequent, especially during south-west wind, but sometimes also coming from the north-west. The rainfall varies according to position with regard to north-west rain; thus it is certainly greater on the summit of the Craigieburn Mountains than on Mount Torlesse. This heavier rainfall and greater frequency of rainy days has caused in favourably situated stations a richness of vegetation approximating to the western region, and even some of the western plants make their appearance-Senecio scorzonerioides, e.g.. The sky is frequently cloudless, and insolation stronger than in any other region. There is no comparison at times between this direct action of the sun and the temperature of the air. It may be most delightfully warm on one side of a gully, sitting in the sunshine, and on the other, at hardly the distance of two metres away, bitterly cold, ground and plants being frozen hard. Temperatures such as these have been measured in various parts of the world. Hooker† Himalayan Journals, vol. ii., London, 1854, p. 410. writes, “This effect (solar radiation) is much increased with the elevation; at 10,000 ft., in December, at 9 a.m., I saw the mercury mount to 132° with a difference of + 94°, while the temperature of shaded snow hard by was 22°.” The moisture in the atmosphere at times when the mountains are clothed with cloud and mist must be very considerable. On the other hand, during the clear weather the rarified atmosphere must be very dry. It is a beautiful and interesting sight from an elevation of 1,200 m. on Mount Torlesse, the sky overhead clear, to look down upon a great level stretch of white dense cloud marking the Canterbury Plain. The soil and subsoil are much the same as in the subalpine region, but streams are less frequent, and towards the summits altogether wanting. Moist ground with permanent lagoons sometimes occurs. Many of the slopes consist of shingle-slips, sometimes made

up of quite loose stones, at other times of firmer consistency, and with a greater or less amount of finely triturated particles of sandstone, yellow in colour, and perhaps containing small amounts of clay. From these shingle-slips rocks not yet disintegrated crop out in places, sometimes several metres in height and sometimes at the level of the shingle, presenting the driest of physically dry plant-stations. The plant-life consists in large measure of cushion and patch plants, bearing a most striking resemblance to many Andean plants, so far as œcological characters are concerned—e.g. Maja compacta, of New Granada, looks like Heliophyllum colensoi; Verbena minima, of Peru and Bolivia, like Raoulia grandiflora; or Loricaria ferruginea like Veronica lycopodioides. Many of the subalpine plants ascend to this region; some of the lowland and lower mountain are also found here. Others seem to be confined to stations where snow lies for a long period every year, and if they are found elsewhere it is most unusual-e.g., Luzula pumila, Celmisia viscosa. The principal plant-formations are:— Name of Plant-formation. Most Characteristic Plant. 1. Alpine meadow Celmisia laricifolia. 2. Rock Raoulia eximia. 3. Shingle-slip Ligusticum carnosulum. The soil of the alpine meadow is often very patchy, there being in many places only narrow strips between shingle-slips, or there may be oases of good soil here and there on the surface, and on the margins of these deserts of shingle. The shingle-slip plant-formation is an edaphic formation of the very greatest interest, and may be divided into the unstable or shingle-slip proper and the more stable shingle-slip, which latter contains a much larger variety and number of plants. Some of the shingle-slip plants are, so far as I know, never found in any other station—e.g., Ranunculus haastii, Stellaria roughii, Ligusticum carnosulum, Craspedia alpina, Lobelia roughii, Cotula atrata, Cotula dendyi, sp. nov., MSS., mihi, Notothlaspi rosulatum, Epilobium pycnostachyon, and Poa sclerophylla. The Western Climatic Plant-region.—This region, so far as it concerns the country here treated of, is an outlying easterly portion of the great western plant region, which extends on the west side of the great Dividing-range right from the north to the south of the South Island of New Zealand, and whose flora has only been examined in a few places. The small portion treated of here, properly speaking, only occupies the

actual ridge of the Dividingrange, as far as the upper limit of the subalpine Fagus formation. This latter formation, however, contains many western plants, especially in its upper portion (e.g., Panax colensoi, P. lineare, &c.), and it forms a distinct natural barrier which shuts off the eastern from the western plants. Its much greater extent and its different constituents separate it from the easternFagus formation, and so, although strictly neither eastern nor western in character, it seems more correct to include it in the western plant region. Some western plants also occur at its eastern limit, such as Gunnera dentata, Hymenophyllum armstrongii, and Carmichaelia grandiflora, and this, too, would form a reason for placing it as I have done. The climate of this western region has been indicated already from the meteorological reports of Hokitika and Bealey, but the rainfall must be very much greater than at either of these places; indeed, I should not be at all surprised to learn that it amounted on an average to 200 in. yearly, or even more. The number of rainy days must be very much greater than on the Westland lowland. During a stay of six weeks on the summit of Arthur's Pass (900 m.) in the months of December and January, 1897–98, I had an opportunity of learning some little about the climate. During that time it rained on more than half the days, the rain sometimes lasting for two days and a half at a time. There was one heavy thunderstorm. My tent, situated on a usually quite dry spot, was not unfrequently filled with water to a depth of 15 cm. The wind blew north-west during the whole of the six weeks, with the exception of a few hours, when a south-west wind gave a slight sprinkle of snow. Once, too, it also snowed from the north-west. The wind blew often with enormous force. Had our tents not been sheltered by trees they would have been frequently levelled to the ground. This excessive wind has written its mark on the vegetation (Plate X.), where the tussocks of Danthonia raoulii are all bent to the south-east. Often when not raining on the pass it was doing so on the adjacent heights, and when raining on the pass was fine in the Otira Gorge (W.). During the fine weather it was extremely hot, so much so as to make climbing an exertion. For a few days before our arrival it had been fine, and then the ground and all the vegetation was so dry that one had to be most careful when lighting a fire not to set the forest in a blaze. A creek near the camp was perfectly dry, and so were many watercourses; yet the rain previous to these few fine days had been excessive for weeks! Snow lay (December) all over the hills above a height of 1,200 m., and in the river beds and hollows in the mountains would be quite 15.m deep in many places. As an example of how deep snow may be

in the New Zealand Alps, Haast says* Report on Head-waters of River Rakaia, Christchurch, 1886, p. 30. that in October he had found by actual measurement 48ft. of snow on Browning's Pass, and further, on page 31, he writes, “It is difficult to describe my astonishment when I looked down a frightful gorge with nearly vertical rocky walls about 800 ft. high. It was on the snow which had filled up this precipice that we had ascended five weeks previously. In estimating the altitude of the slope where we thus travelled up from the gully at 300 ft. I think I rather over-than under-estimated it. This would leave 500ft. as the depth of snow in the gorge.” At the sources of the Bealey, the Poulter, and Nigger Creek every year great masses of snow must accumulate, and also the extensive subalpine meadow flora just at the terminal face of the Waimakariri Glacier must experience a tremendous pressure from drifted snow. By the end of January (to continue the account of observations on Arthur's Pass) a great deal of snow had melted, the mountains were almost bare, but still there remained a great depth at the head of the Bealey River, under the precipices of Mount Rolleston. The thermometric records, which were not kept very regularly, may be seen in the portion of this paper relating to meteorology; they show the great variations of temperature which occur even in the middle of summer. During the time of excessive north-west wind and rain described above the eastern climatic plant-region was experiencing a great drought, whose effects would be heightened by the constant hot drying winds, and serious bush-fires raged in many parts of the Canterbury District, especially near Oxford. The south-west wind seems to exert a great influence on the distribution of the vegetation in this region. The West-land subalpine scrub never occurs where the full force of this cold wind can strike it; in such situations only Fagus formation exists. Thus the Dividing-range has on its eastern a vegetation of an entirely different physiognomy to that on its western face, and in some cases these two œcologically different formations occur within a few metres of one another. It is only in sheltered gorges or protected within the forest that the peculiar Westland subalpine plants can live. At the eastern base of Walker's Pass, in a moist sheltered hollow into which two waterfalls descend, and protected completely from the south-west by a rocky wall, growSeneciolatifolius, a Westland hygrophyte. In sheltered places up Punch-bowl Creek, Bealey Valley, is Dracophyllum traversii, and the solitary plant of Olearia lacunosa before mentioned may be observed from the coach-road, sheltered in the forest near the summit of Arthur's Pass.

In December, 1898, while on an excursion with Profèssor K. Goebel, we had an opportunity of observing the vegetation of this region under conditions of dryness, for during the fortnight we were in Westland only a few quite local showers occurred. In the usually extremely moist forest near Greymouth the filmy fern Trichomanes reniforme was generally found with its fronds quite withered up and apparently dead. So dry was the subalpine scrub on Arthur's Pass that some of a survey party with whom we were camped, climbing Mount Rolleston, set fire to a considerable portion. Much of the usually swampy ground was quite dry, and the plants which generally were surrounded by water were growing on extremely dry ground. Outwardly, except as shown by the presence of certain hygrophytes, there appeared nothing to indicate an extremely moist climate. From this short account of the climate it may be seen that the plants of this western region have to endure considerable extremes so far as wet and drought are concerned, and with this as one consequence amongst other factors the vegetation is essentially xerophilous. A few hygrophytes and tropo phytes are also met with, but under exceptional circumstances. Thus in a climate having an abnormally great rainfall, and an atmosphere usually charged with much moisture, we meet with plants whose structure seems more suitable for deserts. Such are sclerophytes with tomentose leaves and strong stems (Senecio viridis); herbs with leaves whose internal morphology calls to mind the steppe grasses—Diels, l.c., p. 265 (Celmisia armstrongii); shrubs with much-reduced leaves (Pittosporum rigidum); plants whose stems, furnished with coriaceous imbricating leaves, form dense cushions (Donatia novæ-zelandiæ); tufted plants formed of hard rosettes of incurved, imbricating, needle-shaped leaves, the whole pressed close to the ground (Celmisia sessiliflora), &c. All the same, the plants as a whole, when compared with those of the eastern climatic region, are not of quite so extreme a xerophilous structure; nothing is here found to approach in that particular the vegetable-sheep or the coral-broom (Corallospartium) of the eastern ranges. Many plants have considerable leaf-development, and a much greater luxuriance of growth generally is apparent. The alpine and subalpine meadows are like meadows, and not mere patches of plants surrounded by stony wastes. In the east the subalpine scrub occurs only here and there, but here it forms a great belt at the limit of the upper forest.* Cockayne, “On the Burning of Subalpine Scrub” (Trans. N.Z. Inst., vol. xxxi., p. 400). The never-ending war which is waged for the supremacy between

tree-forms and grassy meadows, described by Schimper (loc. cit., p.176) has here, in the lower regions, resulted in a complete victory for the trees, which cover with a dense black mass the basal portions of the mountains. In the east, on the contrary, the grasses have vanquished, and tussock meadows rule supreme. The western may be divided into similar subregions to the eastern region, with much the same limitations as regards altitude that have been proposed for the latter. They are, however, only two in number, if we omit for the present certain grassy river-flats (since no part of this region descends much below 762 m.)-viz., the subalpine region and the alpine region:— (1.) The Subalpine Region: This occupies all the eastern slopes of the Dividing-range below a height of 1,220 m., and has its lower limit at the line of the average north-west rainfall. It includes the following plant-formations:— Name of Plant-formation. Most Characteristic Plant. 1. Fagus forest Fagus cliffortioides. 2. Subalpine scrub Phyllocladus alpinus. 3. Marsh Donatia novæ-zelandiæ. 4. Wet meadow Celmisia petiolata. 5. Rock Helichrysum grandiceps. 6. River-bed Raoulia haastii. 7. Running stream Montia fontana. 8. Meadow Ranunculus lyallii. Nearly all these formations will be subdivided into minor formations, &c., and there are some others which will receive attention. (2.) The Alpiné Region: Here the snow certainly lies longer and the snowfall is greater than in the eastern alpine region. It extends from the upper limit of the subalpine region to the mountain summits, if below the snow-line, or to the glaciers and the perpetual snow. According to Haast the glacier at the head of the White River descends to 1,300 m. with its terminal face, and that up the Crow River to 1,372 m. (loc. cit., pp. 147, 148). Many of the plants are the same as those of the eastern alpine region, but these usually when truly eastern are few in numbers-e.g., the very common Celmisia spectabilis of the east is here rare and local. The plant-formations will be given the same names as those of the east, for the most part, and so need not be enumerated until treated of in detail. At various places on the dividing-range are low passes, in height from 908 m. to about 1,200 m. In such places the ground is usually flat, well watered, and in places swampy.

Huge rocks and masses of coarse débris abound, and generally there are mountain tarns and waterholes. Here the subalpine becomes mixed with some of the alpine vegetation; the plants grow in profusion and with great luxuriance, forming when in bloom, from the end of November till the end of January or middle of February, beautiful natural flower-gardens. Appendix. The following notes on the effect of the severe winter of 1899 on plants, more especially on native trees and shrubs at Hororata, were most kindly furnished by Sir John Hall:- (1.) List of New Zealand Trees and Shrubs damaged by the Frosts of June and July, 1899. Olearia forsteri in nearly every instance was completely killed; it has suffered perhaps more than any other native tree, even when completely sheltered. Height in some instances, 7 ft. O. traversii very badly damaged; in most cases killed out-right; in other cases the lower parts were unhurt. Veronica salicifolia (plant 5 ft. high) completely killed. Pittosporum eugenioides and P. tenuifolium in most cases were considerably damaged, especially towards the tops. A hedge of P. tenuifolium which had been kept closely clipped suffered very severely, every leaf being browned, but it may possibly recover. Drimys axillaris suffered badly in some cases. Panax arboreum has suffered very badly, and most likely in many cases has been killed outright. Clianthus puniceus killed outright. Griselinia littoralis for most part undamaged, but many leaves have dropped off, apparently through the effect of the frost. Podocarpus totara slightly browned in places, but virtually quite unhurt. Senecio greyii considerably damaged. (2.) New Zealand Plants growing at Hororata which were not damaged. Olearia avicenniæfolia. " illicifolia. " furfuracea. " insignis. Gaya lyallii. Senecio compacta. " monroi.

Cassinia fulvida. Hoheria angustifolia. Sophora microphylla. Veronica macroura. (3.) Exotic Plants which were undamaged. Quercus ilex (Mediterranean). Eucalyptus gunnii (Tasmanian). In many instances quite small plants fully exposed were undamaged. Choisya ternata (Mexican). Cistus ladaniferus, in three varieties (Mediterranean). (4.) A fine laurel hedge is very much damaged. Almost every leaf on the west side is killed, while on the other side the hedge is still quite green. The undamaged side was not exposed to the wind, while the damaged portion received a slight westerly breeze.