that regular tidal (rather than seasonal) flooding is the most important factor, rather than a difference in salt tolerance between trees and smaller plants (Janzen 1985; Corlett 1986). Exactly how this flooding acts against the establishment of herbs has yet to be determined. Living in saline conditions is clearly costly in terms of energy, and this is supported by the observation that mangroves were killed after only a single spraying with defoliants in Vietnam, whereas nearby terrestrial trees had to be sprayed several times to achieve the same results (Janzen 1985).
Most trees of the mangrove forest have developed peculiar root systems to allow for gaseous exchange above a water-logged and anoxic soil (Mann 1982) (fig. 2.13). Such 'breathing roots' are known as 'pneumatophores'. The stilt roots of Rhizophora may also be effective in preventing the growth of seedlings too close to a growing tree. These stilt roots are generally unbranched but secondary or tertiary branching can occur due to damage of the primary root tip by scolytid beetles (Docters van Leeuwen 1911) or by boring isopod crustaceans (Ribi 1981, 1982; Whitten et al. 1984) (fig. 2.14).
The roots of Sonneratia and Avicennia are similar in gross structure and consist of a horizontal cable root held in place by anchor roots growing vertically downwards. The pneumatophores grow upwards from the cable roots and, small nutritive roots grow horizontal from these. As mud is deposited on the forest floor, so new nutritive roots are produced higher up the pneumatophores. In Bruguiera the cable root loops in and out of the soil and the exposed 'knee-roots' act as pneumatophores. Ceriops does not have special root adaptations but its bark has many large openings, or lenticels, to assist in gas exchange. If oil drifts into a mangrove forest the lenticels on the exposed parts of the pneumatophores become clogged and this is the primary reason that trees so afflicted will die. The dark oil also causes the water temperature to rise and the concentrations of dissolved oxygen to fall (Mathias 1977; Lugo et al. 1978; Getter et al. 1984). Mangrove trees that survive exhibit signs of chronic stress such as reduced productivity and gradual leaf loss (Lugo et al. 1978; Saenger et al. 1981). These effects can be quite local, however, as is evidence by the relatively small patches of dead mangrove trees around the natural oil seeps on the shore near Kabali, southwest of Luwuk.
The feathery flowers of Sonneratia are superficially similar to those of the Myrtaceae (such as rose apples Eugenia spp., and eucalypts Eucalyptus) and in common with many of those, are pollinated by bats which may fly up to 40 km from their inland roost when Sonneratia is in flower (Start and Marshall 1975). The flowers of Rhizophoraceae have a range of mechanisms by which they effect pollination. For example, the anthers of Bruguiera open explosively when the flowers are visited by sunbirds Nectarinia (in the large flowered such as B. gymnorrhiza), or butterflies and other insects (in the smaller-flowered species such as B. paruiflora) in search of nectar. Ceriops tagal also has explosive anthers, triggered largely by moths. Rhizophora flowers are largely wind-pollinated but bees may also be involved (Ding Hou 1958; Tomlinson et al. 1979). Sunbirds can sometimes be seen visiting Rhizophora trees but this is largely to lick the sweet, sticky exudate from leaf buds or young flowers which have been slightly damaged by insects. Since the sunbirds also eat insects the birds help to reduce the damage to Rhizophora by scale insect Coccidae (Christensen and Wium-Anderson 1977; Primack and Tomlinson 1978; Wium-Anderson and Christensen 1978; Wium-Anderson 1981).
Figure 2.13. Different types of roots in mangrove trees.
Figure 2.14. Branching of Rhizophora mucronata roots caused by small scolytid beetles burrowing into the pith of the root tip.
After Doctors van Leeuwen 1911
Observations of fruiting and flowering of mangrove trees in Australia and Thailand showed that there was activity in every month but that most species flowered during the dry season, and dropped ripe fruits during periods of peak rainfall. This pattern is very similar to that often found in dry lowland forests (p. 366) and is probably related to insect abundance. The production of new leaves was depressed when fruit and flower production were maximal. The time it took for a flower to form a ripe fruit varied both between and within species but was about 1-2 months for Nypa, 2-6 months for Avicennia and about 15 months for Ceriops (Chris-tensen and Wium-Anderson, 1977; Wium-Anderson and Christensen 1978; Wium-Anderson 1981; Duke et al. 1984).
A few species of mangrove trees have evolved an unusual, though not unique, form of reproduction. Generally speaking, fruit develops on a plant and, when it is ripe or fully developed, the fruit or the seed inside it is then dispersed; the seed germinates when, or if, it comes to rest in suitable conditions. In most of the Rhizophoraceae such as Rhizophora and Bruguiera, however, the fruits ripen and then, before leaving the parent tree, the seeds germinate inside the fruit, possibly absorbing food from the tree. The hypocotyl (embryonic root) of the seedling pierces the wall of the fruit and then grows downwards. The cotyledons (first leaves) remain inside the fruit. Eventually, in Rhizophora mucronata for example, the root may reach a length of 45 cm. The seedling then drops off by separating it self from the cotyledon tube, the scar of which forms a ring around the top of the fallen seedling, and the small leaf-bud can be seen above this scar (fig. 2.15). Bruguiera behaves similarly, but the break occurs at the stalk of the fruit. This form of behaviour presumably allows the rapid establishment of the young plant. These types of fruit are described in more detail elsewhere (MacNae 1968).
Figure 2.15. The propagule of Rhizophora mucronata showing the root (often mistaken for part of the fruit) and the top of the seedling that detaches itself from the parent plant.
Zonation
Mangrove tree species tend to grow in zones or belts (figs. 2.16, 2.17 and 2.18). On gently sloping accreting shores (where sediment is being actively deposited), the forest nearest the sea is dominated by Avicennia and Sonneratia, the latter usually growing on deep mud rich in organic matter (Troll and Dragendorf 1931). On firm clay sediment A. marina is more common whereas on softer muds A. alba predominates (Ding Hou 1958). Behind these zones Bruguiera cylindrica can form almost pure stands on firm clays which are only rarely inundated by the tide. Further inland B. cylindrica becomes mixed with Rhizophora apiculata, R. mucronata, B. panriflora and Xylocarpus granatum (the canopy of which can reach 35-40 m). The mangrove forest furthest from the sea is often a pure stand of B. gymnorrhiza. Seedlings and saplings of this species are tolerant of shade but only under larger trees of other species; they are unable to grow under the canopy of their parents. This is presumably due to some chemical interaction. The boundary zone between mangrove forest and inland forest is marked by the occurrence of Lumnitzera racemosa,4 Xylocarpus moluccensis, Intsia bijuga (fig. 2.19), Ficus retusa, rattans, pandans, the stemless palm Nypa fruticans, and the tall spiny-trunked palm Oncosperma tigillaria. Where the mangrove forest has been opened, the most common undergrowth plant is the fern Acrostichum aureurn. Relatively steep-sided creeks, bays and lagoons are generally fringed with Rhizophora trees.
Figure 2.16. Zones of mangrove forest observed in part of Malangke.
After Anon. 1980a
Figure 2.17. Changing abundance of adults (above) and juveniles (below) of four tree species along a transect inland from the seaward edge of mangrove forest at Lainea, Kendari. Vertical bars represent 10 trees or seedlings/10 m2.
After EoS team
Recognizable zones may arise for two different reasons where neighbouring vegetation associations have little or no floristic affinity despite growing in the same environmental
