C—l 3

30

Pomaderris phylicaefoUa sometimes occurs in this association. At a greater or lesser distance from the sea, heath is the common feature of these northern Auckland dunes, but it is generally rather on the very sandy soil covering the consolidated ancient hills than on the more recent, and in such a case the Morula is much richer. Heath of this character is almost identical with that of the ordinary clay lands, so far as species are concerned, and contains many species not given in my list of dune-plants. Ecologically it is more wind-swept, and various shrubs, normally erect, are flattened to the ground. (c.) Fixed Dunes. (a.) Shrubby and Grassy Dune. —Even now, short as the time is since the European first settled in New Zealand, it is not easy, or indeed possible, in the majority of cases to say what the vegetation of the fixed dunes was exactly like. The pasturing of stock, frequent burning of the vegetation, and the spread of introduced plants has, in most places, called into existence a plant association quite foreign to primitive New Zealand. But there are certain places, here and there, more or less undisturbed, where probably even yet all the original native species remain, though in a much different percentage from what they were in the primeval vegetation. Heath, as above described, would probably be the important association of many places. Arundo conspicua and Phormium tenax would play an important part. In Southland certain species of piripiri (Acaena) are abundant, and there is a considerable quantity of the round silvery patches of Raoulia auslralis, and of another species allied to H.-Monroi ; also Gentiana saxosa, Gnaphalium trinerve, Poa caespitosa, and Pimelea Lyallii are common in certain places e.g., on the dunes facing Colac Bay. Bracken-fern (Pteridium esculent urn) is generally present in abundance, but how far heath of this character was a true association in primitive New Zealand is certainly not known. As for turf-making grasses, Danthonia pilosa, D. semiannularis. Dichelachne crinita, Microlaena slipoides, and Zoysia pungens in the north would be present. The most important part played by the primitive vegetation in the first instance, and aided afterwards by the introduced plants, was the formation of humus from their decay, the surface thus getting covered by degrees with sandy loam, itself alone an excellent fixiiig-agent which would render the dune stable unless disturbed. So much so is this the case that in certain places these stable dunes are successfully used for growing crops of oats or rape (see Photo No. 35). But, notwithstanding such a use, the greatest care has to be exercised, for if once the true sand is exposed there is a rapid reversion not merely to semi-stable sandhills, but to the wandering dune itself (see Photo No. 14). With burning and "stocking" the equilibrium between the species was upset, space was opened up for new plants, so that various grasses, leguminosae, &c. came in, and at the present time a turf, containing white clover, cocksfoot, Yorkshire fog, meadow poa, and even rye-grass, together with many worthless weeds, covers much of the ground of many dune-areas, interspersed w'th a varying quantity of the indigenous plants mentioned above, and others according to the geographical position of the dunes in question. (b.) Dune Forest. —-Dune forest is not a very frequent feature of the coast. It occurs most abundantly where the number of rainy days is excessive. Thus, on the dunes of Chatham Island lowforest consisting of the so-called akeake (Olearia Traversii) and the matipo (Suttonia chatliamiea) was originally common. Drifting sand has, however, destroyed this forest in many places. On the lee slope of a portion of the Mason Bay dunes, in Stewart Island, there is a quite luxuriant though low forest, consisting of the broadleaf (Griselinia littoralis) as the dominant tree, its thick leaves being in harmony with the dry soil. In some parts the southern rata (Melrosideros lucida) is abundant. its dark-green leaves contrasting with the much lighter green of the Hat forest-roof. The other trees are Uapanea Urvillci, Aristotelia racemosa, Nothopanax simplex, Carpodetus serratus, Pittosporum faseiculatum, Fuchsia excorticata, Pseudopanax crassifolium, and Dracophyllum longifolium. (For further details see Cockayne, 92, p. 18.) On the dunes of western Nelson, south of Westport, there is old forest consisting of the foresttrees of the locality, but I have no detailed notes. It comes quite close to the sea, from which it is separated by the sand-grass (Scirpus frondosus) dunes. Forest occurs occasionally in sheltered gullies on the, fixed dunes of both Wellington and western Auckland. The components ((insist in part of coastal trees, and in part of inland trees found frequently near the coast. The following are some of the members : Karaka (Corynocarpus laevigata). akeake (Dodonaea viscosa), pohutakawa (Metrosideros tomentosa —Auckland only), kawakawa (Macropiper excelsum), kowhai (Sophora tetraptera), ngaio (Myoporum laetum), puriri (Vitex lucens —Auckland only), titoki (Alectryon excelsum), rangiora (Br achy glottis repanda), kohutu (Pittosporum lenuifolium), kohekohe (Dysoxylum spectabile), and mahoe (MelicyPUs ramiflorus). The forest of the Southland dunes has been already noted. Also in certain parts of that district, forest, but not necessarily of the dunes, has been buried by sand, the tops of the dead trees still projecting above the surface. (iii.) Hollows and Sand-plains. (a.) General. It has been shown how the advancing dune-ridge leaves in its wake level sandy ground, which continues to be lowered by the wind until moisture, rising from the quite adjacent water-table, forbids further removal of sand. It is obvious that such hollows are in themselves quite stable, while (heir moisture permits an altogether different class of plants to those of the dune proper to establish themselves. The final destiny of these level areas does not depend, however, upon their own plant covering, but on the stability of the adjacent dunes, and according to the behaviour of the latter so is the subsequent history of the vegetation. One of two things may happen — there may be an invasion of sand and a reversion to dune conditions, certain transient plant associations arising only to be destroyed ; or there may be a long stage of stability, in which case an evolutionary series of associations will succeed one another, culminating in a climax association ; but this final stage may be reached by different patfis.