Plant Succession on the Oreti River Sand Dunes
[Read before the Otago Institute, 13th June, 1933; received by Editor, 20th June, 1933; issued separately, September, 1934.]
The area of sand dunes considered here lies as a fringe along the eastern bank of the Oreti River, near where it runs into the New River Estuary at Invercargill (Map, Fig. 1). It is swept by the frequent south-west and westerly winds. The climate is prevailingly cool and damp, with many overcast days. The special problem attacked in this paper is that of plant succession on these dunes, the presence of wind-excavated dune-hollows in various stages of development and of stabilisation presenting a useful opportunity for such a study. These hollows all run in an easterly direction, more or less at right angles to the river bank, and are initiated by breaches made in the dune front in flood time. The account here given is based upon data gathered during a number of visits paid by both authors to the locality.
We wish to thank Dr H. H. Allan, of the Plant Research Station, Palmerston North, for identifying a number of the species here dealt with. This paper is based upon a thesis which was presented by the first-named author as part requirement for the degree of M.A. in the University of New Zealand. Both authors are, however, responsible for the paper in its present form.
Notes on the Environmental Factors Operating in the Area Dealt With.
For details concerning the rainfall and its distribution throughout the year, the number of rainy days, the range of temperature, cloudiness, etc., reference should be made to the Meteorological Department's records for Invercargill. In brief, it can be stated that all of these factors are conducive to a more or less uniformly low rate of evaporation from the soil surface and from the transpiring surface of the plants. There will, of course, be periods of dry, clear weather of varying duration, during which, at least in the summer, the temperature of the surface layer of the sand dunes will be high and the humidity of the air in contact with the sand surface will be low, with the result that at such times, in spite of the fact that the winds are mainly cool, moisture-laden ocean winds, evaporation will be greatly increased.
The average number of days per year on which winds blow from the southward and westward is 170. The coastal sand dunes show a strongly marked tendency to move in an easterly direction, and the drifting of the sand and the formation of new hollows in the line of dunes plays an important part in determining the nature of the plant covering.
In the case of the youngest hollows, the soil is practically pure sand which is of a fine texture and is readily blown by the wind. The incoming of humus depends altogether upon the development of a plant covering upon this soil. Thus the pioneer species have special edaphic conditions to face. As to the soil moisture, it is a well-known fact that on coastal sand dunes, in spite of the absence of humus with its water-holding abilities, and also even under conditions of strong surface evaporátion, the layers of sand below the uppermost loose dry stratum are moist and firm, owing to the capillary movement of water from below. One of the most important points to be considered in the development of a plant covering upon the floor of a new hollow is the loose dry shifting nature of the surface layer, since the seedlings of the pioneer species will have to get their roots down quickly through this to the moisture below, and the plants will at all times have to cope with the burying power of the wind-driven sand. When once these pioneer species have become well established, however, they will tend to consolidate the sand surface and to provide humus and so to act as seed beds, so that additional species will come in and in turn play their part in the succession. Even in the case of well-turfed hollows there will, however, be a tendency for sand to be blown on to the floor from the less closely covered sides.
The dunes considered in this paper have been very little, if at all, interfered with by man or by grazing animals. We have observed traces of rabbits, but to so small an extent that it can confidently be stated that rabbits have had no effect upon the nature of the plant succession.
During such times as the river is in flood, breaches are occasionally made in the dune front by the undercutting effect of the river, followed by the falling in of the sides of the breach so made. Such a breach is the initial stage in the formation of a new dune hollow (Fig. 2), its subsequent enlargement being due to the blowing out of the loose dry sand. The power of the wind to erode and shift the sand is clearly evidenced by the blown out hollow shown in Fig. 5.
The Dune Vegetation with Special Reference to Succession.
As has been stated above, all stages in the erosion of dune hollows and their subsequent stabilisation by a plant covering are to be seen in this area, up to the stage at which the large, wide, mature hollows are completely covered by a dense association of turf-forming species with Poa caespitosa and Scirpus nodosus tussocks, and with a Podocarpus totara shrubbery pushing its way down the hollow from the eastern upper end (Fig. 6). The hollows are separated from one another by flat ridges of varying width and of a uniform height of about 10–20 feet. These elevated portions of the dune area are covered closely with the Poa tussocks, but the plant covering between the tussocks is for the most part less close, and comprises fewer species, than in the hollows. However, the covering on these ridges is a comparatively stable one. The succession described below
refers only to what takes place on the floor and sides of the hollows. On the drier ridges the succession will be of a somewhat different nature, and will be determined largely by the spread of species from the hollows.
When once the above-mentioned, practically mature stage in the succession in the old hollows has been reached (Fig. 6), it might appear to be unlikely that the wind would erode it again. The hollow is now wide and relatively shallow, so that the wind passes over it, rather than being drawn through it as through a funnel as it is in the case of the younger, narrower hollows. However, all the hollows in this area, mature as well as young, are very narrow at their lower end where they approach the river bank, this lowermost end being blocked by a more or less massive barrier of consolidated sand, covered by vegetation, which has accumulated there owing to a combination of wind and river action. This is well seen in the hollow shown in Fig. 4, where already two new breaches have been made in this barrier. Even in the case then of the oldest and widest hollows it is conceivable that by the making of a breach in this way wind erosion of the close turf and tussock covering could begin again at the lower end of the hollow and could steadily proceed up it. For example, this was found to be taking place in the hollow shown in Fig. 5, which is intermediate in size and age to those shown in Figs. 4 and 6, whose upper eastern end had been completely blown out, and whose floor was for the most part strongly eroded.
(b) The Pioneer Species.
The small densely growing cushion-former Raoulia australis var. albosericea was found to be the first comer on the floor of new breaches made in the dune front.* For example, in the case of the very young breach shown in Fig. 2 a few small individuals of this species had already made their appearance on the uneven bare floor. On another similar but slightly older breach no plants of any kind were observed on the floor on the occasion of our first visit to the locality, but some seven months later we were able to find a few seedlings of both the Raoulia and of Colobanthus Muelleri. That this Raoulia plays a conspicuous part in the early stages of colonisation is shown also by the fact that it is commonly to be found in isolated cushions (never large, but of various sizes up to 8in in diameter) on either a loose sandy surface such as the bare sides of a hollow, or on a firm, bare, wind-swept surface. It was plentifully present, for instance, on portions of the bare, hard floor of the highly eroded hollow shown in Fig. 5. Fig. 3 is a quadrat made at a selected place in this last-named hollow, and shows the patch forming individuals of the Raoulia drawn to their relative sizes. The only other species present here was Colobanthus Muelleri,
[Footnote] * With regard to this plant Dr Allan states: “It is identical with the coastal plant of Wellington, and very close indeed to the Volcanic Plateau plant that Colenso described as R. albosericea (T.N.Z.I.: 20, 1887, p. 195). I am inclined to place them all together as a compound variety, using Colenso's name.”
represented by a few scattered individuals. This latter species is a very small tuft plant, less than one inch high, which does not form patches, and is too small to play a part of any significance in consolidating the sand surface. As is referred to in detail below, Raoulia australis var. albosericea possesses morphological characters which obviously fit it for withstanding the severe conditions of its station on the dunes. The part that it plays in preparing the way for other early colonising species is probably insignificant, except when, as we found only rarely to be the case, it is present in considerable quantity (Fig. 3), and then it will certainly help to hold the drifting sand. Its cushions are comparatively small, and are moreover so dense that wind-driven seeds of other species will rarely find a lodgment on them. As a matter of fact, most of the cushions examined by us did not show the presence of other species on them. Not a few seedlings of this Raoulia of different ages were observed by us in several bare localities, so that it is clear that germination and establishment takes place fairly freely.
Passing from the consideration of this species, there is no doubt that the following early comers play an important part in the formation of a close plant covering on bare sand, viz., Hydrocotyle americana var. heteromeria, Raoulia australis var. apice-nigra, and Acaena microphylla var. pauciglochidiata, and, to a somewhat less degree, Epilobium nerterioides var. minimum. These are all quick growing, copiously branching, low creeping herbs, rooting at the nodes, whose mat growth-form is well fitted to consolidate the sand in their immediate neighbourhood. These species are true pioneers, albeit for the most part on such parts of the floor of young hollows as are firmer than those on which Raoulia australis var. albosericea can gain a footing. From such places they extend outwards into looser sand, commonly two or more of them growing intermixed, so giving rise to a plant covering which is the initial stage in the formation of a turf. These species, besides serving to anchor the surface layers and consolidate them, will provide a certain amount of humus by their own progressive decay, and will catch wind-driven seeds of other species, and so act as a seed bed for the latter.
(c) Other Incoming Patch-forming Species.
On the floor of the hollow shown in Fig 4 there is no bare sand. The floor is covered with a dense turf on which the large tussock grass Poa caespitosa, and, in damper parts, the sedge Scirpus nodosus also, is established. The plant covering of this hollow may be taken as typical of that of all the hollows in this area of similar age. The above-mentioned four creeping species are all abundantly present, and can be regarded as constituting the main element in the turf. Small isolated plants of Colobanthus Muelleri are common in the turf, and there are also occasional small cushions of Raoulia australis var. albosericea. Other creeping species which are also more or less commonly interspersed in the turf are Helichrysum filicaule and Mentha Cunninghamii, and, in addition, some or other of the species referred to below as playing a more or less important part in the hollows.
Three patch-forming species of Gunnera must next be mentioned, viz., G. monoica var. albocarpa, G. arenaria, and G. Hamiltoni. The first-named is to be found not infrequently intermixed with the other turf formers, although in damp places it may become dominant. G. arenaria is much less commonly present, being restricted to a few localities, although it is abundant on the Sandy Point Domain on the opposite (western) side of the Oreti River. G. Hamiltoni (Fig. 8) is known to be confined to the shores of Foveaux Strait. It is more robust in habit than any of the other New Zealand members of the genus, and in these dune hollows forms very dense and extensive patches, usually to the exclusion of all other plant species, sending out fast-growing stolons which quickly consolidate bare sand (Fig. 7). It forms a conspicuous feature in the plant covering of several of the hollows. As with all Gunnera species, it needs a certain degree of dampness in the substratum, and is to be found especially on the more shaded western sides of the dune hollows, from whence it spreads on to and across the floor. On account of its rapid growth and dense robust patch-forming habit it is a most efficient stabiliser of a sandy surface, resisting the eroding effect of wind more than any of the other turf-forming plants. It is, however, apparently not a true pioneer species.
Two species which in these hollows have adopted the character of patch-formers are Pimelea Lyallii and Gentiana saxosa. The former is typically a semi-prostrate, sub-shrubby, much-branched plant which is commonly present along the coast of Southland. It is fairly common in dune hollows, and comes in at a comparatively early stage in the consolidation of the hollow. Both it and the Gentian are able to establish themselves on the unstable sides of hollows, and in this station are of especial importance. By its quick growth Pimelea keeps the ends of its straggling branches above accumulating sand, and by the copious development from them of adventitious roots it consolidates the sand around itself. Such a plant finally appears above the surface as a low-growing flat patch which may be up to 3 feet in diameter, thus assuming the role of a turf former. Gentiana saxosa is a perennial littoral species, very abundant around the Foveaux Strait coastline, and, like the Pimelea, commonly present in dune hollows. On a substratum other than sand it forms small loose clumps up to 6 inches in height. In the dune hollows it can keep the tips of its branches above accumulating sand, and, owing to the copious branching of its straggling stems under the sand, it forms a patch of rather open nature with its small, stunted, fleshy leaves lying on the surface of the sand. The individual plants of the Gentian are in some places so numerous that an almost pure turf of this species is formed.
Two other species not infrequently present in the turf of the hollows are Plantago Raoulii and Geranium sessiliflorum (Fig. 9), the latter being the commoner. Although these occur usually only as isolated individuals, they undoubtedly aid in compacting the surface layers.
(d) The Climax Association.
The dune area, which is here considered, along its eastern boundary is covered by an open shrubbery of Podocarpus totara and P. spicatus, among which are old large trees of the same two species. The lianes Rubus schmidelioides, Fuchsia Colensoi, and Muehlenbeckia australis form a conspicuous feature in this shrubbery. This represents an altered Podocarpus forest which is in process of being overwhelmed by the moving sand dunes driven eastward by the prevailing winds. The whole of the district of Otatara, lying between the Invercargill Estuary and the lowermost reaches of the Oreti River (See Map, Fig. 1), is sandy in nature and represents an old and very extensive dune area. The major portion of this area carries
a heavy mixed forest of Dacrydium cupressinum, Podocarpus ferrugineus, P. spicatus, and P. dacrydioides, with a varied shrubbery in which Myrtus pedunculatus and Suttonia divaricata are dominant. This will be then the climax association to develop, under the prevailing climatic conditions, on this old sand dune formation. In the seaward direction it passes into the present fringe of moving dunes, as has been stated, through the more open Podocarpus totara- P. spicatus type of forest, which has become more or less altered to a Podocarpus shrubbery.
In the case of the largest and most stable dune hollow examined by us (Fig. 6), scattered shrubs of P. totara, were found to be present over its easternmost upper parts, and this species is clearly spreading down into the hollow. Single young shrubs of the totara have also become established in some of the less mature hollows or on the tussock-covered elevated areas between them. No doubt the next stage in the succession would be that a closer shrubbery of this species would develop and would cover the whole dune area. However, this is present only here and there in the area, since the dune hollows, at longer or shorter intervals of time, are blown out. The large hollow shown in Fig. 5 is a good example of this, the greater part of the floor having been denuded right down to the firmly compacted substratum. The sand has travelled eastward, and is in process of burying the trees on the outskirts of the Podocarpus forest. The turf over which the sand is advancing consists mainly of Acaena microphylla, Fuchsia Colensoi, and the introduced Canadian thistle. This flat turf represents the floor of the original Podocarpus forest covered by a small depth of blown sand and consolidated, and the Fuchsia (usually a liane) is thus another example of a species which under these conditions can change its growth-form and become a turf-former. The largest hollow (Fig. 6) has also at some earlier period had its upper end blown out, but has become consolidated again, and carries a uniform plant covering, as previously noted, in which Poa caespitosa, intermixed with Scirpus nodosus, is dominant throughout.
Briefly, then, the sequence in the development of a plant covering on these dunes begins with the establishment of an open association of certain small mat and patch-forming species, aided in places by others which, although not typically patch or mat-formers, under the conditions here prevailing, adopted this growth form. This turf becomes more and more closed owing to the incoming of additional small creeping or rosette species. The tussock grass Poa caespitosa next becomes dominant on this turf. Finally a shrubbery of Podocarpus totara develops. Judging from what can be observed in places behind the sand dunes where this shrubbery has become dense, the tussock grass tends there to disappear and the composition of the turf to change, the liane Fuchsia Colensoi, for example, becoming a member of the turf, and some of the other turf species disappearing. A Podocarpus totara-P. spicatus forest here develops. Further back
still, where there is no action of the wind upon the floor, and where the substratum is damper and there is a greater accumulation of humus, a mixed rain forest develops.
Notes on the Autecology of Certain of the Dune Species.
Miss Pegg (4) has published an account of the sand dune plants of New Brighton, Canterbury, devoting her attention to the autecology of the species. Amongst others she considers the following—which also play a part in the Oreti River dune hollows—Poa caespitosa, Scirpus nodosus, Epilobium nerterioides, Gunnera arenaria, and Pimelea arenaria. The last-named species is represented on the Southland dunes by P. Lyallii She concludes that the species of the moist hollows “for the most part are strong mesophytes,” specifically referring to Epilobium nerterioides and E. Billardierianum, among others, as coming under this category.
In a paper on the mat plants of the Cass River bed, Canterbury, Foweraker (3) deals at length with several species of Raoulia (including R. australis) and gives figures of the leaf and stem anatomy of certain of them.
Cockayne (1, pp. 23, 31) emphasises the difference between moist and dry dune hollows. He states that the species to be found commonly in the moist hollows “are merely species of other wet or moist stations without the dune area” (p. 23), and, amongst others, mentions Epilobium nerterioides and E. Billardierianum. Speaking of the Epilobium and of certain other “moisture-loving” plants of these damp hollows, he remarks that although the substratum is usually wet and cool, there are periods in the summer when the surface temperature is high and the surface layers are dry, and expresses surprise that such species can tolerate these extreme conditions. In the true dry hollow, on the other hand, there is always a surface layer of sand which is liable to drift. Cockayne refers to the fact that in Southland Raoulia australis and Geranium sessiliflorum are commonly present in dry hollows (p. 32). Dealing with grassy fixed dunes (p. 30), he states that in Southland certain species of Acaena and of Raoulia are abundant, and in some cases also Gentiana saxosa, Poa caespitosa, and Pimelea Lyallii are present along with others.
The Oreti River dune hollows considered in the present paper come properly under the category of dry hollows, although in the mature stage, in which the hollow is extensive and closely covered with turf and tussock, parts of the floor may be more or less moist. In most of these hollows the sand surface, where not covered by vegetation, is loose and liable to be blown, and there is a total absence of halophytic species, which, as Cockayne notes (2, p. 94),
are usually to be found in the true “damp” hollows. On the other hand, the general nature of the plant covering of the Oreti River hollows described above, except for the absence of halophytes, corresponds with that of moist rather than with that of dry dune hollows of other parts of New Zealand. The abundant presence of such species as Epilobium nerterioides, Hydrocotyle americana, and the species of Gunnera, may be especially emphasized with respect to this. The reason no doubt is to be found in the prevailingly cool and damp climate of Southland. Evaporation from the sand will be for the most part low, with the result that the surface mantle of loose dry sand will be as a general rule shallower than in the case of dunes in drier parts of New Zealand. This is clearly of importance to the plant life both on account of the fact that the first root of seedlings, and new roots formed from creeping stems, must get down through this mantle before they become anchored and begin to absorb, and also because the loose sand is liable to be blown and so has burying power. Transpiration from the subaerial parts of the dune species also will be, for the most part, low. It is undoubtedly for reasons of this nature, then, that certain “moisture-loving” species are able to take part in the succession on the Oreti River dune-hollows. However, there still remains the fact to be considered that these species will have to endure occasional periods, especially during the summer, of more severe dessication of their subaerial parts.
In dealing with the autecology of the dune-hollow species then, the factors of the environment to be regarded as of especial significance are the loose dry nature of the surface layer and the occasionally-strong evaporating power of the air. Our notes will therefore concern both the growth form of the species and also their ability to conserve water. Details should also have been included with respect to the seedling plants, if such had been available. It is obvious that the seedling stage is a very critical one in the establishment of a plant on a surface which is dry and liable to move. These details should concern not only the anatomical characters of the seedling, but also the season of the year at which germination takes place, the depth below the surface at which the germinating seed lies, and the rate of growth of the first root in getting down to the moist substratum. Without having actual observations on these points to bring forward, we surmise that in these Oreti River dune-hollows, under the climatic conditions prevailing, the establishment of the seedlings will not be so precarious, during at least the autumn, winter, and spring months, as in dry dune-hollows in other parts of New Zealand.
There are, as is well known, a number of leaf characters which are usually regarded as efficient in the retardation of transpiration from the leaves, or in the holding in the leaf of a considerable reservoir of water. Among the chief of these are a thick cuticle, sunken stomata, a tomentum, compactness of the inner tissues with
small air spaces, and an unusually large development of these inner leaf tissues with a correspondingly increased ability to hold water. It is, of course, very rarely that these characters are all present together in any one species, and external conditions which are adequately met in one species by a certain combination of leaf protecting characters may be met in another species by other characters. The dune species described below differ from one another considerably in the character of their leaves and other exposed parts.
The following anatomical and growth-form details relate, of course, to the species as they were actually growing in the dune-hollows described in this paper. We have selected for description those species which play an especially significant part in the succession.
(a) Raoulia australis var. albosericea.
In the youngest seedlings found it was evident that the first root penetrates down through the upper loose sand layer before the first foliage leaves have fully developed. Both the cotyledons and first foliage leaves are strongly tomentose, and correspond closely in structure with those formed later.
Growth of the above-ground part and development of the small compact cushion is slow, but root development is fast and extends far horizontally just below the loose surface layer. One plant measured had a cushion 2½ inches in diameter, with straggling wiry roots up to 3½ feet long. Neither the stem and branch system nor the roots are as efficient in sand-binding as are those of the true creepers with their more open and more branched mat form.
It is known that there are several distinct varieties of R. australis. Foweraker (3, p. 14) states that the species, as it occurs on lower river terraces in the Cass River bed, forms a flat, more or less open, quickly growing mat with straggling branches.
On the dunes the small cushions are frequently to be seen more or less buried by the sand. It is well-equipped for withstanding drying. The form and structure of the leaf is closely similar in several of the species of Raoulia examined by Foweraker (3). In these species there is a more or less well developed palisade tissue under both surfaces of the leaf (3, Fig. 3). The usual spongy mesophyll is replaced by a water storage tissue of large cells. Both surfaces of the leaf are densely covered with tomentum, the hairs having a characteristic form. The lamina is somewhat folded inward along both sides of the midrib on the adaxial side of the leaf and the leaves more or less closely imbricate over each other. R. australis var. albosericea possesses all of these characters (Fig. 10); in addition, the stomata on the outer, abaxial surface of the leaf are noticeably
sunken below the surface, and the cuticle is heavier on that surface than on the other. The roots are especially wiry and tough, this being due to the heavy groups of fibre which develop in the pericycle.
This species can be regarded then as well fitted, morphologically, to act as the first comer on the bare floor of a hollow. As has been mentioned above, however, it usually seems to play no part in preparing the way for the next comers, so that the succession may be said to begin properly with those species next to be described.
(b) Hydrocotyle americana var. heteromeria.
A creeping, open, mat-forming species, with the rhizome close beneath the surface. Roots are borne freely at each node. The leaf petioles are long, and by their elongation the lamina is kept above the surface of accumulating sand. The rhizomes grow rapidly and branch, and together with the petioles and roots are efficient in binding
the sand. When this species is accompanied by other creeping species, as is commonly the case, the anchoring of the sand is very effective.
The leaf is glabrous, the cuticle not specially thickened, and the stomata (which are present on both surfaces) lie flush with the surface. In these respects, then, the leaf could be described as of a mesophytic type. In its internal structure, however, it is not typically mesophytic (Fig. 11). The palisade tissue is three layers
of cells thick, and will hold a considerable amount of water, and the spongy mesophyll is compact with only small air spaces, so that air percolation through the mesophyll will be slow. It would seem that it is to these internal leaf characters that this usually moisture-loving species owes its ability to withstand dessication.
The rhizome has a relatively wide cortex which stores starch so abundantly that a white deposit of it soon collects on the bottom of a watch glass in which sections are placed. No doubt the cortex can also hold a supply of water.
(c) Epilobium nerterioides var. minimum.
A freely branching, quickly growing creeper, whose small sessile leaves lie flat on the sand surface. The stems are surface growing and copiously rooted. The rapid elongation of the branches ensures the continued life of the plant when the older parts are buried by the drifting sand.
The leaf has the same glabrous nature, unthickened cuticle, and unprotected stomata, as has that of the Hydrocotyle, and also, on the other hand, a correspondingly well developed palisade tissue and compact spongy mesophyll. Raphide sacs and mucilage cells are abundant, especially in the spongy mesophyll, the mucilage cells undoubtedly being effective in holding water. We cannot agree with Miss Pegg (4, p. 167) in regarding the leaf of this species as “strongly mesophytic.”
(d) Raoulia australis var. apice-nigra.
As in the other variety of this species described above, the seedlings show a rapid development of the root system. The mature growth form is an open mat, the creeping stems being usually buried, the short lateral leafy branches growing up erect to the surface, growth being fast enough to keep the leafy parts above the accumulating sand. The leaves are densely tomentose, but less imbricating than in the var. albosericea.
The leaf anatomy corresponds very closely with that of the other variety. The stem is strengthened with a continuous zone of cortical fibre, and patches of pericyclic fibre are also well developed. The root has the same remarkable development of pericyclic fibre as in the other variety. Thus both root and stem are well protected from drying and from mechanical injury by moving sand.
(e) Pimelea Lyallii.
This shrubby species usually occurs in inland hill country, and has a low sub-erect habit. On the Oreti River dunes its growth form is similar to that of P. arenaria, the flexible straggling stems and main branches being frequently buried to a depth of many feet, especially when growing on the sides of the dune hollows. The leafy parts of the plant rise less above the surface and are more mat-like in P. Lyallii than in the other species, but the power of rapid stem elongation under conditions of accumulating sand is noteworthy in both.
The toughness and flexibility of the stems is due largely to the remarkable development of fibre intermixed with the soft bast, the bast forming long wedge-shaped patches separated by the more compressible tissue of the medullary rays. There is also a well-developed periderm. The leaves are covered on their abaxial surface with long silky hairs, these being also present on the younger branchlets and terminal buds. This tomentum undoubtedly plays an important part in retarding transpiration and in protecting the buds from mechanical injury by blowing sand.
The mature leaf of P. Lyallii has a thick cuticle on both surfaces. The stomata are of a xerophytic type and are much sunk below the surface (Fig. 12). They are confined to the adaxial surface, although
Fig. 2.—Bleach in the dune front: the imtial stage in the development of a dune hollow.
Fig. 3.—Quadrat (10 feet square) at a selected place on the bare consolidated floor of the wind-eroded hollow shown in Fig. 5, showing unusually abundant distribution of cushions of Raoulia australis var. albosericea. No other plant species present except a few scattered individuals of Colobunthus Muellcri.
in P. arenaria they are present equally on both surfaces. The epidermal cells on both surfaces show the very peculiar gelatinisation of their inner walls which is characteristic of most of the genera in the family (5, p. 716), a feature which accords with the fact that the family as a whole is a dry habitat one. These mucilaginous cell walls will hold water very retentively. Both the spongy mesophyll and the palisade tissue are of a rather loose and open nature with plenty of air spaces, as Miss Pegg found also in P. arenaria (4, p. 164). On the whole, however, the leaf must be regarded as efficiently protected from excessive drying.
(f) Geranium sessiliflorum.
This species, like the last named, apparently is common on sand dunes only in Otago and Southland (2, p. 91). In its typical form there is a multicipital crown to the tap root lying at or just above the soil surface. In its dune form this develops as an elongated, much branched stem, bearing adventitious roots, and buried right up to the branch apices (Fig. 9). The plant is clearly able to keep pace with the accumulation of sand around it, and keeps its leaves above the surface. The stem and branches are stout and woody, and densely clothed in their upper parts with the bases of the old petioles and their large scaly stipules. In their lower parts they are worn clean of the leaf bases by the moving sand, but are further protected by a thick periderm. The deep growing woody tap root also has a well-developed periderm. In the dune form the leaves are only about one-third the diameter of those of the type.
This species on the dunes thus forms small patches, the branches of the stem effectively anchoring the sand around the plant. As it sometimes occurs in considerable numbers on the floor of hollows, it can be regarded as playing an important part there.
The leaf lamina and petiole carries a certain amount of silky tomentum, and, more important still, the branch apices are closely covered by the scaly stipules of the leaves, and are thus well protected from dessication and mechanical injury.
The leaf anatomy is closely similar to that of the Hydrocotyle and Epilobium described above.
(g) Gunnera Hamiltoni.
This species forms very dense, extensive patches. It has a stout creeping rhizome at the surface of the sand which branches copiously, bearing a clump of stout rigid leaves at each node (Fig. 8). The nodes are near enough together for the leaves to form a closely interlocking system, with the result that in the main patches the surface of the sand is nowhere visible and is completely protected from the wind. These patches are so dense that it is but rarely that any other species is to be seen growing in them. From the edge of the patch the rhizomes extend outwards very rapidly (Fig. 7).
The rhizome is up to 5 or 6 mms. thick. Starch is abundantly present in the very wide cortex and also in the pith, and no doubt these tissues will hold much water also. The roots are copiously provided with root hairs to such an extent that, when dug out, a cylinder of the moist sand firmly ensheaths each.
The leaf petiole is long so that the lamina is rarely buried by the sand. The lamina is thick and firm. The distinction between palisade and spongy mesophyll is not clearly marked, the cells of the former being scarcely at all elongated at right angles to the surface. There are, however, 3 or 4 layers under the abaxial epidermis, very compactly arranged, which represent the palisade. There is a considerable spongy mesophyll which gives the main thickness to the leaf. This also is compact, with but small air spaces. The stomata are on both surfaces and are not sunken. The cuticle on both surfaces is comparatively thin. Miss Pegg has described with a figure the leaf anatomy of G. arenaria (4, p. 168), this corresponding fairly closely with that of G. Hamiltoni. She speaks of the leaf as “strongly mesophytic.” This we cannot agree with, since the extensive and compact inner tissues will hold ample water, and air movements through them will be considerably retarded.
(h) Gentiana saxosa.
This is a herb of comparatively small size which can play the part of a true turf former. In some of the Oreti River dune hollows it occurs in large numbers, even on the sloping sides of hollows where the sand is loose and bare. Commonly all that is to be seen of the plant are the small oval leaves, ½ to 1 cm. long, lying flat on the surface. The numerous branches, buried in the sand, will tend to stabilise the surface layer.
The leaf is fleshy, and the cuticle, though not thick, is better developed than in most of the other dune species already described. The stomata lie flush with the surface. The palisade tissue is compact and from 4–5 cells deep. The spongy mesophyll is also specially thick, but is open in character with plenty of large air spaces. The leaf can clearly hold much water in its tissues, and this, together with the rather well developed cuticle, forms a combination of characters which probably adequately explains why the leaves never show, even on a hot, dry, autumn day, so far as we have observed, any signs of wilting.
An account is given in this paper of plant succession as it takes place on sand dunes under the climatic conditions obtaining in the extreme south of New Zealand.
Some of the most characteristic of New Zealand dune species are absent from the area described, e.g., Scirpus frondosus, Coprosma acerosa, and Calystegia soldanella, as also is the introduced marram grass. Moreover, Pimelea Lyallii takes the place of P. arenaria of other New Zealand dune localities. Gunnera Hamiltoni is altogether confined to this southern coast. Gentiana saxosa, abundant on the Southland coast, is only found elsewhere in New Zealand on the western coast of the South Island. Thus the plant covering of the Oreti River dunes shows strong floristic differences from that of most other New Zealand dune areas.
The dune hollows described above are of the dry (1, p. 32) rather than the moist type, but the turf which develops in them is somewhat similar to that of a typical moist dune hollow. This turf consists very largely of species which are not true dune species at all, but which are widely distributed outside dune areas. Some of these can be described as “moisture-loving” plants, and it is largely due to their presence that the turf of these dune hollows has a more or less similar facies to that of a true moist hollow. That these hollows come under the category of true dry hollows follows from two facts, firstly that the surface sand layer on the floor is dry and liable to be blown, and, secondly, that in none of them are halophytes present as they so commonly are on moist stable sand flats. In this latter respect the plant covering of the damp sandy flats which lie behind the dunes on the western side of the Oreti River is in strong contrast to that of the sandy hollows of the eastern side described in this paper, the turf covering of the former containing abundantly such halophytes as Selliera radicans, Samolus repens, Triglochin striata var. filifolium, and Lilaeopsis novae-zelandiae.
The species which play an important part in the earlier stages of the succession on the Oreti River dunes are mat or patch-forming plants which are able to bind the sand in their immediate neighbourhood, and which can by rapid growth keep their leaf system above the accumulating sand. It has been shown above that the subaerial parts of these species must be regarded as protected, to a greater or
less extent, from excessive drying. Some of them are typical mat-formers with creeping stems. In the case of certain others, however, viz., Gentiana saxosa, Pimelea Lyallii, and Fuchsia Colensoi, the last named of which occurs only in the shrubbery which is being overwhelmed by the moving dunes, and to a certain degree Geranium sessiliflorum, the patch growth form is not the usual one for the species, and must be regarded as an epharmonic form, a direct adaptation to sand dune conditions.
A fruitful line of study, if we had been able to undertake it, would be to compare dune-growing individuals with individuals growing under “mesophytic” conditions in the near neighbourhood, or with individuals which had been transplanted from the dunes into the garden, with respect to details in the growth form or in the stem and leaf anatomy. Such a comparison would help to a better understanding of the ecology of the species, by indicating which characters are under the direct control of the environment, and to what extent. An extensive study of this nature has been carried out by Miss Anna Starr for the dune-growing species of Indiana, U.S.A. (6), and she finds that species normally belonging to mesophytic localities, usually show when growing on dunes well marked modification in both leaf and stem anatomical characters in accordance with the new environment.
Another line of investigation, referred briefly to earlier in this paper, should concern the establishment of the seedlings of the true pioneer species. The significance of the behaviour of the seedling on bare sand surfaces becomes apparent when it is remembered that the seeds of not a few of the plant species growing in the immediate neighbourhood will be, without doubt, scattered freely on bare surfaces, but that it is only with respect to certain definite species that the seedlings are able to establish themselves.
Since the above account was written, the senior author has continued the study of the consolidation of dune flats in other coastal localities in Southland and Otago where the climatic conditions are similar to those of the Oreti River area.
West of Colac Bay, on the Foveaux Strait coast, there is an extensive dune area which is undergoing continuous wind erosion. Certain of the flats are dry and are swept more or less down to the shingly substratum. Here the first species to occupy the ground is Raoulia australis var. albosericea, followed by the same species of Epilobium, Hydrocotyle, and Colobanthus as play the role of pioneers in the dry dune hollows in the Oreti River area. The Raoulia forms large flat cushions up to several feet in diameter, and undoubtedly plays an important part, along with the others, in arresting sand movement. The seedlings of this species, in all stages of development, were found to be abundant in the driest season of the year (January). The more extensive flats further back from the shore dunes are distinctly of the moist type. Here there is a firm sandy floor on which the halophytes Lilaeopsis novaezelandiae and Triglochin striata var.
filifolium, together with the moisture-loving Hydrocotyle trifoliata, are the pioneers. These form a rapidly extending turf formation, which on older flats of the same type has become remarkably dense owing to the incoming of such non-halophytic, moisture-loving species as Pratia angulata, Mazus radicans, Nertera Balfouriana, Marchantia, etc. Where this dense native turf is better drained on slight elevations a few inches above the surrounding surface, clovers and certain other introduced species begin to play a part, thus indicating a further stage in the succession.
1. Cockayne, L. Report on the Dune Areas of New Zealand. Dept. of Lands. Wellington. 1911.
2. —— The Vegetation of New Zealand (Die Vegetation der Erde, Vol. XIV), 2nd Edition. W. Engelmann. Leipzig. 1928.
3. Foweraker, C. E. The Mat Plants, Cushion Plants, and Allied Forms of the Cass River Bed. Trans N.Z. Inst., 49, p. 1–45. 1917.
4. Pegg, Miss E. J. An Ecological Study of some New Zealand Sand-dune Plants. Trans. N.Z. Inst., 46, pp. 150–177. 1914.
5. Solereder. Systematic Anatomy of the Dicotyledons (Engl. Transl.), Vol. II, p. 716. Clarendon Press. 1908.
6. Starr, Miss Anna M. Comparative Anatomy of Dune Plants. Bot. Gazette, 54, pp. 265–305. Univ. of Chicago Press. 1912.