These are most abundant in tree crowns, although some also occur at lower, shadier levels. Sun epiphytes are more numerous and diverse than shade epiphytes and can be grouped into several growth forms.

Mat Epiphytes

These epiphytes form mats or patches mostly on inclined or horizontal branches, and they comprise three orchids and one fern which range throughout the country. Mat epiphytes may establish directly on bare bark, particularly if it is rough and fissured, but may also avail themselves of moss cushions. The fern is Pyrrosia serpens (Fig. 38), which belongs to a genus of epiphytes centred in tropical Asia. Our species is also found in Australia and the islands of Polynesia. Pyrrosia often establishes directly on bare bark and has slender, freely branching rhizomes which form a complete network over sunny branches. The leaves are simple and smooth-margined, varying from almost round to long and narrow. They have a fleshy texture and a dense felt of buff coloured hairs page 70

Figure 39 Tall kahikatea (Dacrycarpus dacrydioides) with many asteliad nests. South of Kaitaia, northern North Island.
Photo: B. V. Sneddon.

page 71

Figure 40 (opposite) Asteliad nests with the pendent epiphytic fern Asplenium polyodon below them.
Photo: M. D. King.

page 72beneath, which presumably restrict loss of water. Pyrrosia also grows on the trunks of trees in the open and on rocks. It can be quite abundant on introduced trees, particularly Cupressus macrocarpa.

The epiphytic orchids⁴² which form mats or patches belong, or are closely related, to large tropical genera and can be regarded as outliers, reduced in both leaf and flower size. They all have specialised roots which, as well as serving for attachment, also efficiently absorb and store water in a special outer layer of dead cells known as the velamen.

Drymoanthus adversus, often attached to quite smooth bark, is unlike the other species in that it has a short stem, which does not grow along the bark surface. The roots arising at the base of the tuft of leaves are particularly conspicuous as they spread out 'like the rays of a spider's web' for a considerable distance, often encountering the roots of other plants of the same species. Drymoanthus includes our species, another in east Australia and a third in New Caledonia, but there is some doubt as to whether it should be separated from the Asian and Australian genus Sarcochilus. Our two species of Bulbophyllum, although small, form quite dense patches with their branching rhizomes. In common with their many tropical relatives, each of their rhizome segments swells at the end into a water-storing 'pseudobulb' with a single leaf arising from the top. The stalk bearing the small flower or flowers arises from below the pseudobulb. Bulbophyllum pygmaeum is the smaller species with leaves about a centimetre long but it forms larger patches than those of B. tuberculatum. The leaves of B. tuberculatum are several times longer, but this species is less frequently seen and is not known further south than the north coast of the South Island.

Nest Epiphytes

Much more evident to the casual observer are the massive nest epiphytes perched high in tree crowns (Figs. 26, 39, 40). In New Zealand there are three long and narrow-leaved species belonging to two closely related genera of the lily family — Collospermum hastatum, C. microspermum and Astelia solandri. Two other species of Collospermum are found outside New Zealand, one each in Fiji and Samoa; both are epiphytes. In the much larger and more widespread Astelia found mostly in the southern hemisphere, only a few species are consistently epiphytic. The others are found in New Zealand in a variety of habitats — coastal cliffs, swamps, forest floors, alpine tussock grassland and a few reduced page 73turf-forming species in alpine bogs.

All three nest epiphytes usually establish among mosses and lichens in branch forks or on inclined or horizontal branches. As their stems, completely hidden by the leaf clusters, are short and more or less erect the plants are fixed in position; although when they branch to form additional leaf clusters the resulting massive clumps of foliage may be metres in diameter. The 'nests' are attached to their supports by extensive root systems and as the old roots and leaves die and decay, considerable depths of dark spongy soil are built up. The likely eventual fate of these large soil and plant masses is to fall to the ground. Heavy rain absorbed by the soil greatly increases the weight of the mass and if the rain is accompanied by wind, complete branches may crash to the ground under the weight of epiphytes.

Astelia solandri is more shade tolerant than the collospermums and so is often found below them in the lower crowns and on the upper trunks of trees. The silvery green leaves are in three ranks and are 1–2 m long, but only 2-3 cm wide. Their bases are tightly folded, forming a narrow ridge at the back. Astelia solandri is found in lowland forests throughout the North Island, near the northern coast and down the western side of the South Island to about 44° S.

Collospermum hastatum accompanies Astelia solandri through the North Island and to about 42° 30' in the South Island. Collospermum microspermum is restricted to the North Island and replaces C. hastatum in montane forests above about 300 m. C. hastatum has fan-like arrangements of black-based leaves (Fig. 41) which are somewhat shorter and much broader than those of Astelia solandri. The bases of the leaves are strongly rounded and enclose spaces or 'tanks' which become filled with water when it rains. Often sufficient water is contained by the tanks to provide a shower bath for the unwary when a fallen Collospermum nest is lifted and tilted. A species of mosquito has been described whose larvae always develop in the water stored by Collospermum hastatum⁴³ but this is a very modest fauna compared with the many insects and even frogs which inhabit the 'tanks' of the tropical American epiphytes of the Bromeliaceae (pineapple family). The water stored by C. hastatum is partly absorbed by roots which grow into the tanks, but the suggestion that it is also absorbed through the embedded multicellular bases of overlapping scales has not been confirmed. In some tropical bromeliads all of the stored water is absorbed through the bases of similar scales.

Collospermum microspermum is equally specialised but its leaves are as narrow as those of Astelia solandri and dark-brown rather than black at the base.

The collospermums are the only known tank epiphytes outside the family Bromeliaceae, but as they form soil very efficiently, they are also nest epiphytes.

Pendent Epiphytes

Four New Zealand wide pteridophyte species often grow as epiphytes with their roots or rhizomes embedded in the soil of epiphyte nests. Though they occur elsewhere as well it is in these sites that their growth is most vigorous and their pendulous stems or leaves attain their maximum length. Of the four, Lycopodium varium is the most impressive; its slender stems sometimes forming huge masses up to 1.5 m long, below asteliad nests (Fig. 42). The stems branch repeatedly by equal forkings or dichotomies, so that they form a dense but well balanced mass. It is almost constantly in motion as even the lightest breeze can set the tassels swaying. In their upper parts the stems are clothed by small spreading leaves, which grade into small, close-set scales enclosing the sporangia towards the branch tips. Lycopodium varium is restricted to New Zealand, but there are related species in tropical forests.

The ferns Asplenium polyodon (= A. falcatum) and A. flaccidum may have leaves of a metre or more in length below the epiphyte nests. A. polyodon, with its double-toothed, wedge-shaped leaflets is perhaps the most attractive of the New Zealand species of the genus (Fig. 40). A. flaccidum has an unusual stringlike appearance with long and narrow, deeply-toothed leaflets (Fig. 43). Dobbie aptly describes the leaves of this species as appearing to have been 'cut from a piece of pale-green leather'. Asplenium polyodon is found from India to Australia and the Pacific and A. flaccidum in Australia and some Pacific Islands. Both Aspleniums and Lycopodium varium extend into montane cloud forests, but there they depend from mossy trunks and branches.

Tmesipteris, a genus restricted to the south west Pacific, is sometimes referred to as a 'living fossil' as it is considered to be one of the most primitive genera of land plants. One of the highlights for botanical visitors to New Zealand is to see a living plant of this genus. Tmesipteris elongata subspecies robusta has been observed growing from Collospermum clumps at a number of localities through the North Island, but not yet page 75

Figure 42 (opposite left) Hanging tassels of Lycopodium varium below an asteliad nest.
Photo: M. D. King.

Figure 41 (opposite above right) Close view of a Collospermum hastatum leaf fan showing the dark leaf bases and roots between the decaying outer leaves. The arrow indicates where water is seeping from the reservoirs between the leaves. Photo: J. W. Dawson.

Figure 43 (opposite below right) Epiphytic plant of the fern Asplenium jlaccidum.
Photo: M. D. King.

page 76in the South Island. Its stems, with their small, simple leaves, are unusually long for a Tmesipteris and dichotomise freely. Other species of Tmesipteris rarely branch.

Three orchid species can also be included as pendent epiphytes. The two Earinas belong to a small genus with other species in New Caledonia and Polynesia, but this genus is considered to be closely related to the larger Epidendrum of tropical America. Both species have spreading rhizomes and can sometimes extend for several metres along branches. The stems bearing the leaves droop downwards and can be 30 cm or more long. The leaves are formed in two rows, more or less in one plane; those of E. mucronata are narrow, thin and quite grasslike while those of E. autumnalis are broader and thicker, in keeping with the more robust nature of the plant as a whole. Both species form terminal sprays of small flowers, E. mucronata in the spring and E. autumnalis in the autumn. The flower clusters of E. mucronata hang down and are yellowish orange, those of E. autumnalis turn upwards and are waxy white with a strong spicy perfume.

Our sole species of the large tropical genus Dendrobium (D. cunningh-amii) is the largest of New Zealand's epiphytic orchids. Its freely branching stems and narrow leaves form feathery drooping masses. The stems are polished, often bright yellow and very bamboo-like in appearance. The white, reddish-centred flowers are scattered and while modest by tropical standards are, at 2-2.5 cm in diameter, the largest among our epiphytic orchids.

Small Shrub Epiphytes

The two species of Pittosporum and one each of Senecio and Coprosma in this category are not usually more than a metre high when growing as epiphytes, but may attain small tree size on the ground. All are endemic to New Zealand.

Pittosporum cornifolium is found throughout the North Island and although it is quite a common plant, many people are unaware of its existence, perched as it is inconspicuously in tree crowns. The stems of this plant are spindly and often hang down below the branches. The leaves are thin but firm with prominent veins and the flowers are small and yellowish red. The round, woody seed capsules are a surprise. When they open they reveal a bright red lining and shiny black seeds embedded in sticky, bright yellow fluid.

Pittosporum kirkii has a more restricted range, as it is not found further south than the central North Island. It has a more erect growth habit with thicker stems and longer, thicker, almost fleshy leaves with obscure veins. The flowers are bright yellow and the capsules are unusually large (up to 4 cm long), flattened and pod-like. Kirk,⁴⁴ after whom the species is named, states that the 'valves contract in a curious manner when the

Figure 44 The small epiphytic shrub Brachyglottis kirkii growing from an asteliad nest.
Photo: B. V. Sneddon.

page break

Figure 45 (opposite) The large shrub epiphyte puka (Griselinia lucida) on a kahikatea. The crown of the puka is just below an asteliad nest and its main descending root is to the left of the tree trunk. Te Marua, southern North Island.
Photo: M. D. King.

page 79capsule bursts'. The capsule is apparently not so colourful as that of P. cornifolium, but is described as having an orange lining.

Brachyglottis (Senecio) kirkii is found in lowland forests throughout the North Island but has not been recorded from the South Island. Its growth form has been described as 'candelabra-like' (Fig. 44). The leaves are soft and somewhat fleshy and the flowers, up to 5 cm in diameter, pure white and crowded into dense heads.

The thick and shiny-leaved karamu (Coprosma lucida) is best known as a ground plant in shrubby, early forest regrowth on drier sites, but is also reasonably common as an epiphyte in asteliad nests. The species is found throughout the country but presumably is common as an epiphyte only within the range of nest epiphytes.

Large Shrub Epiphytes

As well as being larger than those of the preceding category, these also eventually send a root to the ground and so overcome the water supply and soil nutrient problem.

Puka (Griselinia lucida)⁴⁵ is the most notable in this category. Its large, dark green shining leaves usually contrast so strongly with the foliage of the supporting tree that it stands out even to the casual observer (Fig. 45). Puka is distributed in lowland forests throughout the North and South Islands, but is more common in the north. Its seedlings generally establish in asteliad nests situated at branch forks and its roots ramify through the humic soil. After a few years a strong root begins to grow down the trunk of the supporting tree towards the ground. This root and its branches are closely appressed to the bark of the trunk, and frequently grow into crevices and behind bark flakes. The root tips are white and smooth, but a short distance away from them the root surfaces are often densely clothed with short root hairs. Where the roots are in contact with the trunk, they are anchored by the root hairs and the union is sometimes so complete that when the roots are pulled away they either remove portions of bark or leave strips of their own tissue behind.

Generally, when the root tips reach the ground, one main vertical root enlarges greatly until it attains a diameter of 10 cm or more. This main root usually has a few major branches near the ground and the whole system has a very distinctive appearance resulting from the more or less continuous and pronounced longitudinal grooves and ridges of page 80the bark (Figs. 46, 47). In its upper parts the main root gives rise to slender, horizontal, girdling roots (Fig. 48), which often encircle the trunk of the supporting tree many times and so ensure that the puka will not be dislodged even by the strongest gale.

Two other species, Griselinia littoralis and Pseudopanax colensoi, although mostly terrestrial, can grow as epiphytes in the moist montane or higher latitude forests they favour. When growing as epiphytes they are generally beyond the altitudinal or latitudinal range of asteliad nests and

Figure 46 Distinctively fluted roots of puka (Griselinia lucida) on kohekohes. Waikanae, southern North Island.
Photo: M. D. King.

page 81so establish in the moss and lichen cushions of branch forks. Like the puka, they eventually send one or more roots to the ground.

Broadleaf (Griselinia littoralis) is the only other species of its genus in New Zealand. Its leaves are smaller than those of puka, yellowish green and symmetrical or only slightly asymmetrical at the base. In puka, however, the leaf base is very asymmetrical as the two parts of the leaf divided by the midrib are of quite different lengths. Broadleaf has been observed as an epiphyte on a variety of trees. Its descending roots are

Figure 47 (opposite) Roots of puka (Griselinia lucida) on tawa (Beilschmiedia tawa). The arrow indicates the stem of a white climbing rata (Metrosideros perforata). Paraparaumu, southern North Island.
Photo: M. D. King.

Figure 48 (above) Puka (Griselinia lucida) showing girdling roots. Waikanae, southern North Island.
Photo: M. D. King.

page 82

Figure 49 (above) Mountain five-finger (Pseudopanax colensoi) on a kahikatea. National Park, central North Island.
Photo: J. W. Dawson.

Figure 50 (right) Descending roots and a few girdling roots of the epiphyte northern rata (Metrosideros robusta). Te Marua, near Wellington. Photo: M. D. King.

often more massive than those of puka, but they are not grooved. The species ranges throughout New Zealand including Stewart Island. Beyond New Zealand Griselinia is found only in Chile, where there are five species, at least some of which are epiphytes.

Mountain five-finger (Pseudopanax colensoi) also has a wide range, but is absent north of 36 °S and from Stewart Island. I have observed it growing as an epiphyte on kaikawaka or mountain cedar (Libocedrus bidwillii) on Mt. Taranaki (Egmont) and on kahikatea (Dacrycarpus dacry-dioides) on the volcanic plateau near Mt. Ruapehu (Fig. 49).

The descending roots of puka and mountain five-finger seem too slender in relation to their height to stand alone when the supporting trees die, but this may be possible for the more massive roots of the broadleaf.

Tree or Strangling Epiphytes

Northern rata (Metrosideros robusta)³⁷ is the most notable and common example here. It is found in lowland forest throughout the North Island and near the north-west coast of the South Island. It is much more frequent as an epiphyte than as a ground plant and it prefers the tall emergent conifers as supporting trees. The earlier stages of its life cycle are very similar to those of the puka. It usually establishes in asteliad nests, although young plants have been observed attached directly to rough bark. A distinctive feature of some small northern rata plants is the development of tuber-like swellings on the roots which, it has been suggested, may serve for water storage.⁴⁶ Eventually a root grows down the trunk to the ground giving off horizontal girdling roots at intervals (Fig. 50). Unlike puka this descending root does not remain relatively slender, but gradually enlarges to become a metre or more in diameter. It is often branched near the ground to form a tripod or tetrapod arrangement (Fig. 51). More complicated patterns develop where several branching roots descend from a northern rata crown to form complexes several metres in diameter. In some cases more than one rata may be involved, although this is not easy to determine.

With the development of such a massive root system, when the supporting tree eventually dies the northern rata is able to stand alone on its 'pseudo-trunk' (Figs. 52, 53). If the support was an emergent then the rata now replaces it in that role.

The northern rata and some tropical epiphytic trees of similar habit are often referred to as 'stranglers'. This implies that these epiphytes kill the supporting trees by compressing their trunks within a complete or partial network of roots. Popular writers on New Zealand plants have taken enthusiastically to this idea, describing the northern rata variously as a 'predatory gangster', 'forest bandit' or 'notorious strangler' which 'crushes', 'smothers', 'stifles', or 'squeezes' the supporting tree in an 'iron', 'deadly' or 'fatal' embrace.⁴⁷

Partly as a reaction to these verbal flights, some botanists in recent times have tended to take a contrary view.⁴⁸ They point out that the page 84

Figure 51 Mature northern rata (Metrosideros robusta) with a tripod based trunk-like root. The original supporting tree is no longer present.
Paraparaumu, southern North Island. Photo: M. D. King.

page break

Figure 52 (opposite) Mature northern rata (Metrosideros robusta) with a branched trunk-like root system. The original supporting tree is dead, but its trunk persists inside the northern rata roots. The broken top of the trunk is indicated with an arrow. Kaitoke, near Wellington, southern North Island.
Photo: M. D. King.

page 86light-demanding northern rata generally establishes in the well lit crowns of mature trees so that, by the time the rata is large enough to stand alone, the supporting tree might well have died of old age. It does seem, however, that the northern rata must have some deleterious effect on the supporting tree through partial overshading, root competition, and perhaps in cases where the supporting trunk becomes enlarged within a well developed rata root cage, through some restriction in the movement of water and nutrients.

Recently a distinctive new tree species of Metrosideros (M. bartlettii) has been described.⁴⁹ It is restricted to a few forest patches near North Cape and is similar in epiphytic habit to northern rata.

Southern rata (Metrosideros umbellata) is rare and localised in the North Island, but quite common in montane and higher latitude lowland forests in the west of the South Island. It is mostly terrestrial, but has been observed growing as a 'strangling' epiphyte in several places. Similar Metrosideros epiphytes are known in New Caledonia, Fiji and Hawaiʻi.