other localities where the natural vegetation would appear to be manuka heath, the heath soils do not, to the eye, differ significantly from the soils of the Lonneker's Bush totara/stunted matai forest type, assuming loss of topsoils through wind erosion. It might be stressed here, however, that the greater part of existing manuka heath is excluded from this category. It is considered to be induced heath and its origin will be further discussed below.

On the basis of all evidence to hand, the history of the vegetation of the Waiau Valley might be outlined as follows.

At some comparatively recent date podocarp forests characterised by the frequent dominance of matai occupied the valley lowlands from Merrivale, in the south, to Lake Manapouri in the north, from West Woodlaw and Island Bush in the east to Motu Bush in the west. Outlying stands were developed well to the north, west and south of these limits within the then rimu forests of coastal districts and within rimu/beech forests near Lake Te Anau; and the Waiau Valley matai forests were linked, through Island Bush, with extensive matai/totara/kahikatea forests developed on the Southland Plains.

Consequent on some regional climatic change these matai forests then entered into a long period of stagnation and decline. There was no longer any effective regeneration of the physiognomic dominants, particularly, at first, on the poorer sites and soils, and all subsequent growth of trees already established was poor, leaving the residual stands with an excessive proportion of stunted, defective or malformed stems. As the stands opened up through death, decay and windfall, the former secondary species of the forest understories, characteristically broad-leaf and fuchsia, assumed dominance though wherever a seed source was available silver beech and mountain beech invaded the stands. This could only happen, and has only happened on high ground or on riparian sites near streams and rivers draining from high ground. As conditions became more markedly adverse the secondary scrub forests, in turn, opened up and died out in patches with incoming of such atypical forest species as cabbage trees and kowhai, and with local development of pockets of grassland within the main forest mass. The final result has been the development of a mosaic of forest and grassland with some manuka heath developed on the poorest or eroded soils. The remnant pockets of forest are restricted, in general, to the deeper, moister and more fertile soils and each and every remnant pocket is distinct in composition and in condition reflecting all the many local site type differences evident in the ancient forest and reflecting all possible stages in forest stagnation and decline.

Kahikatea, in contrast to matai or, more accurately, to a greater extent that matai, survived through facultative adaption to swamp forest conditions; but all the very old and very large kahikatea are to be found in the derelict hill forest stands in association with matai. During the period of the matai forest climatic optimum kahikatea was mainly a species of the hill slopes, not a typical swamp tree as it is to-day. True totara, like matai, possessed no such plasticity and, over the greater part of its former range, has died out, occasionally leaving, as evidence of its one time presence, fallen logs of durable heartwood; or, in other instances, it has lost its identity in hybrid swarms with Hall's totara, a species far better adapted to the new climates.

The final decay and breaking up of the forest was undoubtedly accelerated by fire. Thus, at an altitude of 3,000 ft. on the western slopes of the Takitimu Mts.,

and in the midst of a waste of rock rubble, scree and bed-rock gullies, the soil under a remnant patch of tussock grassland showed ten inches of a dark, crumb structured grassland soil over two inches of fine, slightly weathered rock fragments which overlay, in turn, a shallow bed of charcoals. Below the charcoals there was a truncated fossil soil, strongly leached with traces of iron pan and marked root channels. The climate must, at one time, have been such as to support forest growth on these exposed high altitude sites. This forest was destroyed by fire, presumably following deterioration in the climate and when the forest was no longer in a condition to re-establish. Following the fire the site was occupied, after a short period of accelerated erosion, by tussock grassland which persisted for a sufficiently long period of time for formation of an appreciable depth of true grassland soil on top of the charcoals and erosion detritus until destroyed, in turn, on initiation of the present erosion cycle. To-day, in respect to this particular site, the climate would appear too severe to permit re-establishment of any type of indigenous forest. It is possibly too severe to permit even the re-establishment of indigenous tussock grassland.

At lower altitudes, in the same general locality, over-grazing or over-burning of the tussock grassland leads to invasion of the pasture by manuka heath. On these sites the climate would seem, to-day, to be favourable to the development of mountain beech forest and, where this is so, the invasion of the damaged grassland by manuka might be interpreted as a first stage in the development of beech forest, a development which has been long delayed by the entrenched sub-climax tussock grassland which occupied the site following disintegration of the original matai forest and in the absence of a beech seed source.

In one instance, at least, the pre-European fires spread beyond the limits of the stagnating podocarp forest and burnt back the advancing beech forest. Some ten miles to the west of the charcoal bed described above, pioneer type, branchy “forest edge” trees were discovered at a distance approximately half a mile within the present forest margin. And, once again, charcoals were found in the subsoil. Judging from the age of these pioneer type silver beech and mountain beech, and from the age of the young stands lying between them and the forest edge, this fire occurred somewhere between 300 and 400 years ago, i.e., two to three centuries before European occupation of the country. Other fires have been dated, by ring counts on fire scarred trees, to 200, 150 and 120 years ago.

The Polynesians, Moa-hunter and Maori, were in occupation of the land for a period of at least 800 years prior to the advent of the European settler (Duff, 1950). They used fire but had no effective means of controlling fire and existing forests carry the mark of these ancient fires to this day; but, in so far as can be estimated, the ancient matai forests of the Waiau Valley must already have entered their period of decline and were probably in an early phase leading to the development of a mosiac of grassland and forest when Moa-hunter culture was at its peak. Man was an agent in forest destruction but the primary cause lay in forest instability. The Polynesians played little part in shaping the forests of high rainfall regions but the relict matai forests and beech forests of low rainfall regions to the east of the Southern Alps bear the imprint of a long period of cultural interference.

The present conditions of the Waiau Valley matai forest is, therefore, readily explicable in terms of that same basic hypothesis evoked in explanation of the

abnormalities displayed by other forests in Western Southland. The climatic changes which lead to the invasion of the coastal terrace bog forests by rimu, to the invasion of the rimu forests by silver beech, and to the invasion of the silver beech forests by mountain beech, lead also to the stagnation and desiccation of the matai forests, to their partial replacement by tussock grassland and manuka heath, and laid them open to destruction by fire. On the pragmatically sound principle of minimum complexity this might well be considered strong grounds for a firmer belief in the validity of the basic hypothesis.

But before abandoning the topic of the ancient Waiau Valley matai forests, two further related items might be mentioned.

Reference has already been made to red beech (Nothofagus fusca) as a species found in association with veteran matai in the forests to the east of Lake Te Anau. Throughout the mountain forests from Te Anau northwards red beech is a frequent forest dominant. To the south of Te Anau it occurs in lesser amounts near Lake Manapouri and in several of the north draining valleys of the Takitimu Mts. In none of these latter localities does red beech appear vigorous and, following any gross disturbance of the forest canopy, it is rapidly displaced by mountain beech or, on certain sites, by silver beech. In other words, it behaves in these stands as a species at the extreme limits of its climatic range; but there is, nevertheless, evidence to show that, in the past, these limits exended very much further to the south.

Near Lake Monowai two separate half-acre pockets of rather unhealthy appearing red beech have been found within silver beech/mountain beech stands; and, well to the south again, a single group of three trees have been found in silver beech/mountain beech forest in the valley of the Bryce Burn, a tributary of the Lillburn. All three of these red beech pockets lie in such a position away from streams draining major areas of red beech forest that they can only be considered true relics. And this supposition is strengthened by the discovery, in several separate localities to the south of Lake Monowai, of undoubted red beech × mountain beech hybrids remote from any possible red beech parent trees.

It would appear probable, therefore, that the period which saw the matai forests fully developed was also a period during which red beech extended well to the south, almost to Foveaux Strait. Since that time its range has contracted northwards leaving remnant pockets such as those in the Bryce Burn and near Lake Monowai, the last survivors, in many other instances, having been swamped out in hybrid swarms with mountain beech. The distributional peculiarities of red beech in the forests of the far south, and its behaviour in the southernmost red beech dominant stands, are explicable only in terms of regional climatic deterioration unfavourable to this species.

Finally, some brief mention should be made of the peculiar scrub forest, known as “The Wilderness,” which lies surrounded by tussock grassland and manuka heath to the east of Lake Manapouri. This scrub forest with bog pine the physiognomic species appears in every way closely related to true bog forest as developed on wet ground in districts with rainfall treble that experienced to the east of Manapouri. Its development in a district with a low actual and effective rainfall has long been puzzling; but if it be regarded as a relict bog forest surviving from the wetter matai forest era then its occurrence is understandable. Many minor types of forest must have been contained within the ancient matai