Introduction

The vegetation of highland New Guinea holds a fascination for New Zealand botanists, for it contains many clues to the ecology and relationships of our own plants Therefore, although my reason for visiting New Guinea was to extend my study of timber line to a region without well-defined seasons, my impressions of other aspects of botany may be of some interest. I will confine my remarks to the vegetation above 2,000–2,500 m. Very broadly, this is the upper limit of permanent settlement, and the level at which extensive grasslands resulting from fire and shifting cultivation give way to largely undisturbed forest. It is the usual level at which the remnants of the mid-altitude forests, dominated by oaks, give way to forests dominated by podocarps and beeches, although there are places where such forests occur at very much lower altitudes. It is the altitude at which ‘New Zealand’ genera become conspicuous elements of the vegetation, although many genera common to the two countries occur also at lower altitudes, and some, such as Agathis, Dysoxylum. and Beilschmiedia, do not occur at high altitudes.

Coniferous-Dicotylous Forest

Most of the high altitude forest which I saw in New Guinea, like much New Zealand forest, can be designated ‘coniferous-dictotylous’, a term which indicates that the conifers are usually equalled and often exceeded in abundance by evergreen hardwood trees (Fig. 1). Familiar genera include Elaeocarpus, Carpodetus, Quintinia, Phyllocladus, and in lower storeys Uncinia, Elatostema, Freycinetia, Macropiper, Libertia and many ferns. Among the podocarps, Podocarpus imbricatus, P. cinctus and P. compactus are very similar to our P. dacrydioides, and are likewise now included in the new genus Dacrycarpus. Podocarpus pilgeri and its relatives are related to the P. totara group, but have much larger leaves and page 114

Figure 1: Forest at 2,900 m. on Mt Albert Edward, with an overstory of Podocarpus compactus and Papuacedrus papuanus.

strobili. Dacrydium nidulum seems close to D. cupressinum, but apparently it is local and I did not see it. Taxonomic revision could make other New Guinea and New Zealand plants congeneric; potential examples are Papuacedrus and Libocedrus, Rapanea and Myrsine, Euodea and Melicope. However, despite the marked similarities, the upland forests of New Guinea are far richer in species, genera and even families.

In the coniferous-dicotylous forests of New Zealand apparent anomalies in the regeneration patterns of the conifers have aroused frequent comment. In many ‘climax’ stands lacking any discernable history of disturbance, the conifers are represented mostly by large trees, and there is a deficit in the young trees which would not be expected in a population maintaining itself. This inadequate regeneration has sometimes been attributed to a climatic deterioration, culminating about 1600–1800 A.D. It would be useful to examine the regeneration patterns of related species in undisturbed New Guinea forests, to find whether similar anomalies occur in a different climatic region.

Primary and fire-induced successions leading to stands of conifers also occur in both countries. On Mt Albert Edward, Papuacedrus papuanus and Dacrycarpus compactus are regenerating freely in fired margins (Fig. 2), but scarcely at all within closed forest. On Mt Wilhelm. where there is said to have been a ban on burning for some years, D. compactus seedlings, other than very small ones, are quite hard to find. The seedlings of Podocarpus pilgeri, on the other hand, seem to be extraordinarily shade tolerant, and survive almost completely enshrouded in a mesh of bryophytes on the forest floor. page 115

Figure 2: Papuacedrus regenerating in a burnt forest margin.

Nothofagus Forest

New Guinea beeches occur both as extensive forests, like those on the slopes of Mt Giluwe, and as isolated stands on the crests of ridges (Fig. 3). The latter, from a distance, are as unmistakably Nothofagus as the similar ridge-top stands in Wellington province. The species belong to the ‘brassii’ group, and the ones I saw were large trees, with bark patterns resembling that of Nothofagus menziesii. The leaves are entire or very finely serrate, but vary in size and shape between species, from small, oval leaves like those of N. solandri, to large leaves more like those of N. pumilio. Reproductive organs also show considerable diversity, especially the cupule, which varies from a woody structure 2 cm. long in N. brassii to an almost abortive structure in several species.

In composition, the beech stands resemble the mixed beech forests at low altitudes in western Nelson, rather than the pure page 116

Figure 3: Nothofagus forest at 2,800 m. in the Chimbu district.

stands of beech which occur at high altitudes in New Zealand. That is, the beeches grow with podocarps and other hardwood trees, and the lower storeys are well developed. In the stands which I saw, a slender climbing bamboo (Nastus) was very abundant, making a point of resemblance with Chusquea-containing beech forests in Chile.

New Guinea beeches apparently have a well-marked annual flush, during the relatively dry season between June and August, and growth rings of a tree of N. grandis felled on Mt Kaindi were shown to be annual by C-14 dating (H. S. Jansen pers. comm). The upper limits of Nothofagus, which lie below 3,000 m., may well he determined by susceptibility of the young growth to frost, which can occur at all seasons, especially in localities where cold air ponds at night. In contrast, genera which do not ascend as high as Nothofagus in New Zealand have relatively higher limits in New Guinea. Thus, on Mt Wilhelm, Dacrycarpus, Cyathea, Quintinia and Rapanea, as well as Bulbophyllum and other orchids are present at the upper limits of tree growth at about 4,000 m.

Beeches have markedly discontinuous distributions in both New Guinea and New Zealand. Botanists have tended to interpret these as smaller scale manifestations of the massive discontinuities which page 117

Figure 4: The alpine region of Mt Wilhelm (4,509 m.)

characterise the genus as a whole, and to assign a relict status to isolated beech stands. This view is partly based on the supposed inability of Nothofagus seeds to be dispersed over long distances, although in New Zealand there are isolated stands which can only have become established from seed carried by wind for at least 4 km., or by water over similar distances. It has been suggested, and supporting evidence is coming from pollen analyses, that the present distribution in the South Island is largely related to the return of Nothofagus to areas where it was destroyed by Pleistocene giaciation (Moar, 1971). In the centre of the North Island, volcanism may have played a similar role (Nicholls, 1963). The influence of Pleistocene climatic fluctuations and volcanism on the distribution of Nothofagus in New Guinea is still not known.

Vegetation of Open Areas

There are several mountains in Papua and New Guinea which are high enough to support a true alpine belt, although none carry permament snow (Fig. 4). Vegetation of alpine character also extends far below tree limit where forest is inhibited by nocturnal ponding of cold air, poor drainage, and fire. It includes extensive tussock grasslands which, in the subalpine belt, are dotted with Cyathea tree ferns. There are also swamps and bogs dominated by sedges, Gleichenia or cushion plants, stony herbfields of both xeric and mesic aspect, and shrub-dominated communities, which are especially developed as fire-influenced ecotones around the margins of montane and subalpine forests. These ecotones or transitions are described by Wade and McVean (1969) in their detailed account page 118

Figure 5: Grassland at 3,300 m. on Mt Albert Edward. The large tussocks are Danthonia archboldii.

of the vegetation of Mt Wilhelm, and also by Gillison (1970), who has studied the successions which are taking place.

It is in this vegetation that the New Zealand botanist feels most at home, if only because the plants are small enough to be readily examined and identified. However few, if any, of these plants have exclusive links with New Zealand; they have equally close relatives on the mountains of Tasmania and south-eastern Australia, or further a field. Familiar names include Acaena anserinifolia, Hierochloe redolens, Carpha alpina, Styphelia suaveolens (syn. Cyathodes colensoi), Coprosma (C. divergens is very similar to C. pseudocuneata), Oreobolus, Oreomyrrhis, Parahebe, Gnaphalium (e.g. species resembling G. traversii and G. involucratum (Drury, 1972), Haloragis, Olearia, Gaultheria, Uncinia, Drapetes and Astelia. Most of the danthonias are obviously referable to Notodanthonia, but one, D. archboldii (Fig. 5), looks like a snow tussock although it is not a true Chionochloa (H. E. Connor, pers. comm.). Some of the plants of austral affinities do not occur in New Zealand, e.g. Drimys, Trochocarpa and Tetramolopium.

For a few groups. New Guinea is the mid-point between Asian and New Zealand parts of the total range. Examples are Coriaria and also Anaphalis if this genus is considered to include Gnaphalium trinerve and its relatives (see Drury, 1970). Many species in both countries belong to genera which occur under cool climates throughout the world, such as Viola, Schoenus, Carex, Festuca, Poa and Deschampsia (D. klossii being the predominant tussock in New Guinea). Closer acquaintance with these could show whether the page 119 nearest affinities lie in a particular direction. Thus, the commoner buttercups in New Guinea seem to belong to the Australasian Ranunculus lappaceus group, whereas the gentians are tiny versions of Eurasian gentians, with blue, trumpet-shaped flowers. There are also shrubby senecios, which most resemble S. hectori among our species, although in some details they are closest to the Australian genus Bedfordia (D. G. Drury, pers. comm.).

Finally, there are northern genera which do not extend to New Zealand, including Vaccinium, Rhododendron, Eurya (Theaceae) Anthoxanthum, Lactuca, and Eriocaulon.

Basis of New Guinea — New Zealand Relationships

Southern conifers and beeches have always been a mainstay of the botanical protagonists of continental drift. Their absence from oceanic islands (to which Araucaria excelsa on Norfolk Island and Podocarpus pallidus on Tonga may be exceptions) suggests that long distance dispersal does not take place, and the dioecious habit of most podocarps and the relatively heavy, non-succulent propagules of the other genera can be cited as biological reasons for this. These views seem vindicated by recent developments in geology. According to Jones (1971), Antarctica and Australia began to drift apart in the mid-Eocene, the Tasman Sea began to dilate in the late Eocene and the Northland (New Zealand) — New Caledonia — Owen Stanley (New Guinea) stratigraphic belt may have been disrupted by the formation of the Coral Sea only in the late Tertiary. If this is correct, it follows that the only barriers to exchange of plants between New Zealand and New Guinea during the early Tertiary were climatic. Among the genera now common to the two countries, Couper (1960) lists Dicksonia, Phyllocladus, Dacrydium, Podocarpus, Ascarina, Nothofagus, Elytranthe and Dysoxylum as being represented in pollen assemblages from the New Zealand Eocene. However, much more needs to be known of the palaeobotany, especially of New Guinea. For instance, when were various genera exchanged, and was it via Australia or a land bridge of which New Caledonia is a remnant?

While substantially continuous land connections probably explain the distribution of forest plants of known or assumed antiquity, the same does not necessarily hold for the flora of the high mountains. No direct evidence can be expected from palaeobotany, for mountain plants are seldom fossil in other than young, temporary deposits. Also, geologists are rather insistent that the mountains of both New Guinea and New Zealand did not attain their present elevation until the late Pliocene. Since a number of our mountain plants occur at low altitudes on very poor soils where they do not have to page 120 compete with forest, Tertiary precursors probably occurred in similar habitats (Wardle, 1968), but this is unlikely in New Guinea.

Mt Kinabalu in Borneo also supports high mountain plants although it is regarded as geologically young. Van Steenis (1964, 1967) nevertheless proposes that these plants have a long history in the area, if not on Kinabalu itself, then on neighbouring mountains which have since become worn down. In the 1964 paper, he invokes land bridges to account for this florula, which includes both Asian and Australian elements. The alternative is to account for the New Guinea high mountain plants substantially in terms of existing geography, though admitting the possibility of an occasional peak supporting high mountain plants where there are none now, especially during the cold phases of the Pleistocene when altitudinal zones were depressed.

Dispersal over long distances or via ‘stepping stones’ need not be invoked for high mountain plants which have their closest relatives in the forests below. Thus, the shrubby Drimys brassii which occurs above tree limit doubtless shares a common origin with the numerous species of Drimys in the upland forests, although the broader distribution of the genus is austral. The same applies to the small subalpine vacciniums and rhododendrons, for the New Guinea highlands are an important centre of diversity for these predominantly Eurasian genera.

Speciation in the Two Floras

The truly alpine species in New Guinea are not especially numerous, and none of the genera have shown evolutionary explosion like that of Hebe and Celmisia in New Zealand. Nevertheless, there are some interesting evolutionary problems. For instance, have the Ranunculus lappaceus-like buttercups, such as R. basilobatus, given rise to ones of less usual appearance, such as the high-alpine R. sarawagedicus, or a species on Mt Albert Edward with entire leaves resembling those of Plantago major?

Parahebe albiflora below tree limit on Mt Wilhelm looks like a larger, woodier version of P. lyallii; is a plant of this type ancestral to the depressed, tomentose P. ciliata on the summit rocks of the mountain? In New Zealand, the species of Oreomyrrhis seem almost prototype umbellifers, whereas O. linearis has dense tufts of simple, linear leaves, and O. azorellacea (Fig. 6) is a cushion plant that I confidently identified as Gaimardia, until Dr. P. Stevens of the herbarium at Lae told me to look for the inflorescence, which is reduced to a single flower. Rhododendron saxifragoides is another remarkable cushion plant (Fig. 7). The cushions are about 20 cm. across, and studded with solitary flowers borne on slender erect peduncles several centimeters long. These flowers, like those of other page 121

Figure 6: Oreomyrrhis azorellacea at 3,350m. on Mt Albert Edward.

Figure 7: Rhododendron saxifragoides at 3,750 m, on Mt Giluwe.

high mountain rhododendrons in New Guinea, are tubular and red, apparently being adapted for pollination by birds.

I am aware of only two endemic genera in the alpine vegetation, both being monotypic. One is Detzneria tubata (Scrophulariaceae) which is a shrub with something of the vegetative habit of a Hebe, although its blue flowers are large and solitary. The other is a fern, Papuapteris linearis, which has strap-shaped woolly fronds very different from any to be seen in the nearly-congeneric Polystichum.

If the alpine zone of New Zealand supports a much richer and more varied flora than that of New Guinea, as befits the far greater extent and diversity of habitat, the reverse is true of lower altitudes. page 122 According to Good (1960) there are 9,000–9,250 species of angiosperms in New Guinea, and even if the 2,600 species of orchid are discounted, this is impressive by comparison with fewer than 1,800 angiosperm species in New Zealand. Probably, the contrast is due to differing effects of climatic fluctuations during the Quaternary era, although the need to tolerate winter cold may also limit diversity in New Zealand. While Quaternary studies in New Guinea have only begun, the extent of lowering of altitudinal zones in the highlands may be indicated by comparing the Snow Mountains of West Irian, which attain 5,000 m. in Mt Carstenz and support permanent snow on summits higher than 4,600 m., with mountains of Papua and New Guinea such as Giluwe (4,087 m.) and Albert Edward (3,960 m.), which supported Pleistocene glaciers (Fig. 8). According to Loeffler (1972), Pleistocene temperatures were about 5°–6°C lower than at present. Fluctuations of this order would have been easily accommodated by local movements of plant communities up and down slopes, but I am curious about what happened within 1,000 m. of sea-level — was the fabulously rich tropical flora even richer before the Quaternary era?

In New Zealand the cold phases of the Pleistocene had far-reaching consequences. Here, a lowering of altitudinal zones by 750–1,000 m. during glacial periods would have sufficed to drive the greater part of the forest flora to the north of the North Island. Nevertheless, it was not a simple matter of extinction of all but the most cold-tolerant elements, for the species which survived to form our present forests show a wide spectrum of temperature requirements. By and large, the species which survived the oscillations of climate were those which were able to re-establish quickly in new areas, and many, including the ‘brassii’ beeches, were unable to do this quickly enough to avoid extinction. Probably also, whereas the New Guinea situation tended to favour genetic drift and development of related species in similar habitats but different localities, the New Zealand situation tended to prevent this, and forest species became relatively uniform over wide areas, through the gene pool being constantly mixed and depleted. In the same way, when closely related species came into proximity, only the most competitive would have survived. Significantly, the ‘Malaysian’ element in New Zealand largely consists of genera with only one or two species, whereas in New Guinea the same genera include several or many species. Examples, using data from Good (1960), are Bulbophyllum (558 endemic species in New Guinea: 2 species, at least one endemic, in New Zealand), Dendrobium (619:1), Eugenia (180:1), Elaeocarpus (120:2), Dysoxylum (46:1), Schefflera (55:1) and Freycinetia (59:1). The larger New Zealand genera, such as Hebe (80 species), Celmisia (60) and Carex (73) tolerate cool or cold climates, and could have accommodated Pleistocene climatic changes by local migrations, which promoted evolution rather than extinction. page 123

Figure 8: A glacial valley on Mt Giluwe. (Photo by M. J. Coode)

Acknowledgments

I would again like to thank those who made my visit to New Guinea possible and profitable, especially Mr. J. S. Womersley, Chief of the Division of Botany at Lae, Professor D. N. Walker of the Australian National University, and the Royal Society of New Zealand for a grant from the Mappin Fund.