The GenusNothofagus has been known in temperate lands bordering the South Pacific since the mid-nineteenth century, but it is only relatively recently that botanists have become aware that an equal or larger number of species also occurs in montane situations in New Guinea and New Caledonia (Van Steenis 1953, 1954). The latter are regarded as a distinct group within the genus—‘brassii group’—and are distinguished from the two temperate groups—‘menziesii’ and ‘fusca’—by differences in pollen form, and in anatomy and external morphology. Some authors (Baumann-Bodenheim 1953, Hjelmquist 1953) suggest that all or some of the brassii group should be regarded as a separate genus.

Fossil evidence (Couper, 1961) has shown, firstly, that species of the brassii group formerly occurred in all higher latitude land areas within and beyond^* the present range of Nothofagus, becoming extinct there during the Pleistocene glaciations; secondly, that Nothofagus appears to have reached New Guinea, presumably from the south and conceivably overland, only late in the Tertiary in the wake of Pliocene cooling. Nothing is known of the history of Nothofagus in New Caledonia, but if the genus migrated there at a similarly late stage then it must have traversed considerable stretches of ocean from New Zealand or Australia. Another possibility is that Nothofagus has been in New Caledonia since the page 2 Cretaceous when the genus first appeared in the higher latitude areas and when there is at least the possibility of a land link between New Zealand and New Caledonia.

Up to the present time five species of Nothofagus have been described from New Caledonia (Baumann-Bodenheim, 1953). On a recent visit I was able to collect and observe three species and I found them to be surprisingly different in appearance from those of New Zealand. The most striking differences were in the leaves, which were relatively large in size, thick and coriaceous in texture and, at the adult stage, revolute at the margins and recurved at the tips.

In Nothofagus sp. (probably undescribed) adult leaves were about 3 X 1 inches; Nothofagus balansae 3 1/2 X 1 inches; and Nothofagus codanandra 8 1/2 X 2 inches. Juvenile leaves were noticeably different in the last two species, being quite flat and larger than the adult. In Nothofagus codanandra juvenile leaves were seen which measured 10 X 3 1/2 inches! It does not seem, however, that leaf size can be used as one of the distinguishing features of the brassii group as several of the New Guinea species within the group have quite small leaves and some South American species outside it have large leaves.

Other notable differences from the New Zealand species were the attachment of the stipules near their mid-points (peltate) rather than at one end and the large bi-lobed cupules at the fruiting stage each containing three large, markedly flattened and winged seeds. Examination of the bud scale scars in the three species showed that each leaf lasts for two or three years. This is in agreement with Nothofagus menziesii in New Zealand, but in the three other New Zealand species, all in the ‘fusca group’, the leaves last for only one year. Other species of the ‘fusca group’ in South America and Tasmania are deciduous.

Ecologically, one of the New Caledonian species of Nothofagus, N. balansae, where it occurred in a non-serpentine area, displayed many interesting similarities with the species in New Zealand and elsewhere.

The area referred to was near the centre of the island just south of the Col des Roussettes, where the main road from the west to the east coast crosses the island. The terrain was steep and comprised a series of ridges rising gradually to an average height of about 2,000 feet. Milling was in progress in the area and for this reason, as well as periodic fires, the vegetation nearest the page 3

Fig 1: Some Nothofagus species of New Caledonia (A-E) and New Zealand (F-J) × 3/5. A-C. N. codanandra (A-juvenile leaf, B-adult leaf, C-twig with cupule containing 3 seeds); D. N. sp. (adult leaf); E. N. balansae (adult leaf); F. N. fusca; G. [gap — reason: unclear]cata; H. N. solandri var. cliffortioides; I. N. menziesii; J. N. solandri var. Photo: M. D. King

page 4 main road was in a very disturbed condition. The predominant vegetation type here was an open woodland of Niaouli (Melaleuca leucadendron), although the original vegetation was probably rain forest of which only scattered regenerating patches remained. Near a ridge crest small patches of Nothofagus balansae occurred alongside a milling road. The individual plants were mostly young and slender, although in places there were groups of trees with trunks up to about one foot in diameter. Nothofagus seedlings were locally abundant on clay banks edging the road.

With increasing distance from the main road the Melaleuca woodland diminished and the area of rain forest increased. The Melaleuca became restricted to scattered patches within the rain forest, often leading from a timber road to a ridge crest. On one occasion I walked from a campsite at about 1,700 feet up a ridge which levelled off at about 2,000 feet. The lower parts of the ridge were in rain forest, but this gave way at higher levels to Melaleuca woodland showing obvious signs of a recent undergrowth fire. The Melaleuca was extensively developed on one side of the ridge, over the crest and for a short distance down the other side, where it formed a sharp boundary with rain forest. Prominent along this boundary were a number of small trees of Nothofagus balansae.

Continuing further along the ridge the Melaleuca gave way to an almost pure ridge crest stand of Nothofagus at a point where the ridge began to slope down into a saddle. The trees were young and of two size classes — on the ridge crest no more than five or six inches in diameter and just to one side of the ridge mostly about 10 inches in diameter.

In the saddle itself Nothofagus gave way to rain forest, but the long, level ridge on the far side supported a narrow belt of apparently undisturbed Nothofagus forest in which Nothofagus balansae appeared to be the only tree species. The trunks were mostly a foot or less in diameter, but there were also scattered larger trees with trunks two to three feet in diameter and standing about 40 to 50 feet high. Some of the latter were dead and there were also a number of large logs, some recently fallen, others much older. It almost seemed as if an older generation of trees were being replaced by a newer generation. No vascular epiphytes or lianes were noted and shrubs and sapling Nothofagus were rare. The ground was open and completely covered with a leaf litter entirely derived from the Nothofagus. Nothofagus seedlings were numerous, only a few inches high, and predominantly at the three leaf stage.

In slight hollows and level areas there was a layer of fine fibrous roots about an inch thick with leaf litter above and pure clay below. This root layer held a certain amount of dry, humus-like material. In slightly raised places the leaf litter lay directly on page 5 the clay. From the fallen trees it was obvious that this species at least is relatively shallowly rooting as is the case with Nothofagus in New Zealand. The upturned root systems were flat and plate-like with masses of pure clay still attached to them.

This pattern of Nothofagus forest occurring on poor skeletal ridge crest soils with rain forest on the deeper soils of the valleys and valley sides can also be seen in New Zealand, for example near Wellington, and in South Queensland (Nothofagus moorei). Robbins (1961) states that in some places in New Guinea Nothofagus dominates on the ridges with podocarp-broadleaf forest below.

The other locality where I saw Nothofagus (N. codanandra) in abundance was above about 2,000 feet in the Montagne des Sources area, which forms part of the large southern serpentine region. The Montagne des Sources was to a large extent clothed with short xeromorphic shrubby vegetation, but there were also considerable areas of forest, particularly at higher elevations. The main type of forest was very distinctive being fairly open and dominated by species of Araucaria, Agathis and Casuarina. It differed in many respects from the rain forests dominated by broad-leaved angiosperms. The Nothofagus forest occurred as apparently randomly scattered patches within the predominant forest type and differed from the Col des Roussettes forest in exhibiting no particular preference for ridge crests. Some of these patches were quite large and were readily discernable from a distance owing to the distinctly reddish colouration of the new vegetative growth. Unlike the ridge crest Nothofagus balansae forests at the Col des Roussettes, N. codanandra was here not the only canopy tree, although it was certainly the commonest. Several other tree species were present including scattered Araucarias.

Schmid (pers. comm.) has studied Nothofagus on serpentine mountains further to the north—Mt. Mou (Nothofagus baumannii) and the massifs of Me Maoya at about 4,500 feet and Boulinda at about 3,500 to 4,000 feet (at both localities, there was the same, probably undescribed species of Nothofagus). He reports that, as was the case at the Montagne des Sources, Nothofagus forest mostly occurred on the side slopes of ridges, on soils of medium depth, rather than along their crests. Schmid also notes that the Nothofagus at these localities was frequently intermixed with other species, mostly belonging to the genera Araucaria, Podocarpus and Dacrydium.

Smaller groups of Nothofagus were seen at localities further south in the serpentine region. On the road leading to the Montagne des Sources a group of young Nothofagus balansae was noted and scattered river margin trees of N. codanandra were seen at quite low elevations along the Riviere Bleu, which arises from the Montagne des Sources. In the Plaine des Lacs area Nothofagus page 6 balansae was seen again, at low elevation, as a forest margin species. The forest occupied a hillside valley system and appeared to be second growth comprising mostly broad-leaved angiosperm species. The third Nothofagus species seen by me, apparently also undescribed, was also growing here as a forest margin species, although I saw only one mature and one young tree. The leaves of this species differed from those of the other two in having shallow marginal teeth.

For further information on Nothofagus on serpentine see Hurlimann 1962.

Another interesting ecological similarity between the New Zealand and New Caledonian Nothofagus species lies in periodicity of flowering. In both countries species of the genus flower and set seed in abundance only at intervals of several years. Such heavy flowering apparently follows unusually warm preceding summers in New Zealand (Poole, 1948; Kirkland, 1961) and has been recorded throughout the country or regionally in 1935, 1938, 1944, 1948, 1951, 1956, 1959, 1965. In New Caledonia Corbasson (pers. comm.) reports heavy flowering in Nothofagus in 1950, 1957 and 1964. Nothing is known of the cause or causes of this apparently seven-year cycle.

In conclusion, from the ecological pattern displayed in New Caledonia and the other regions already referred to, it would appear that where the climate is suitable for rain forest, and where there is no very special edaphic factor such as serpentine parent rock, Nothofagus tends to be restricted to situations where the soil is poor, particularly along ridge crests. At higher altitudes and latitudes, where climates are less favourable or unfavourable for rain forest Nothofagus may then form continuous and much more extensive forests.

Acknowledgments

I should like to thank M. Schmid of the Laboratoire de Botanique, Centre Orstom, and M. Corbasson, Director of the Bureaux des Eaux et Forets, for their assistance to me in New Caledonia and also for reading the manuscript of this article and providing valuable additional information.