Chapter Eight
INTRODUCTION
Lowland forests vary quite considerably in composition (p. 198) and three types are discussed individually below because they are so strikingly different. They are: heath forest/padang vegetation, ironwood forest and forest on limestone. The first two types were investigated by a team from CRES and are discussed in detail below.
HEATH FOREST/PADANG VEGETATION
The only places where extensive areas of heath forest and padang vegetation can be found in Sumatra are on Bangka and Belitung Islands, but small areas also exist in parts of eastern Sumatra, the Lingga and Natuna Islands and Jemaja Island in the Anambas group (Frey-Wijssling 1933b; Kartawinata 1978b; van Steenis 1932, pers. comm.; Valeton 1908; Ver-stappen 1960). Heath forest is known by botanists as 'kerangas forest' from an Iban (Sarawak) word which refers to forested land which, if cleared, would not grow rice. It has nothing to do with 'rengas' trees (e.g., Gluta, Melanorrhoea), although these may occasionally be found within the forest. Padang vegetation is often associated with heath forest and although its status as a completely natural vegetation type is still open to some debate (p. 263), it is certainly a long-lasting and relatively stable secondary growth forming after heath forest is cut or burned.
Soil
Heath forest usually grows on soil derived from siliceous parent materials which are inherently poor in bases and commonly coarsely textured and free draining (Burham 1975). The soils in intact heath forest or under bushes in padangs have a covering of brownish-black, half-decomposed organic material. In the open padangs the surface is generally pure white sand about 0.5-5 cm thick above a darker layer (Burham 1975). In the ecological literature the soils of both heath forests and padangs are called white-sand soils.
White-sand soils usually originate from ancient eroded sandstone beaches which were stranded due either to uplift of land or fall of sea level. If conditions are suitable for soil development, a hard iron-rich layer (podzol pan) forms a metre or so below the soil surface (Burham 1975). One of the first descriptions of a tropical podzol was by Hardon (1937) from a padang soil on Bangka Island (table 8.1). The hard iron pan can cause temporary waterlogging of the soil after heavy rain before the water is able to drain away.
The 'A-horizon' soils have a very high percentage of sand and only about 1% of clay (Hardon 1937). This leads to high porosity, rapid leaching, and low ion retention properties so that white-sand soils are probably among the most nutrient-poor soils in the world (Mohr and van Baren 1954). Analyses of soils from a Sarawak heath forest are also given by Proctor et al. (1983) and Richards (1941).
Drainage Water
The water draining from heath forest is generally a black water (see p. 171). A water sample taken from a small stream flowing out of the heath forest visited by the CRES team was very acid (pH 4-4.5) and had low nutrient levels (2 ppm nitrogen, 1 ppm potassium and 1 ppm phosphorus). These results are characteristic of heath forest water (Whitmore 1984). A study of white-sand soils in the Rio Negro region of Amazonia concluded that, apart from iron, the concentration of nutrients in the river water draining from them was the same as the region's rain water. This suggests that there are no nutrients being weathered from the parent material (Ungemach 1969). The ecological aspects of black waters in peatswamp areas described on pages 176-177, are all relevant to drainage water for heath forests and padang vegetation.
From Hardon 1937
Vegetation
Heath Forest. No detailed survey of heath forest ever seems to have been conducted on Bangka or Belitung Island and therefore much of the following is taken from studies conducted in Sarawak and Brunei. Heath forest can be strikingly different in its floral components and structure from usual lowland dipterocarp forest and it can itself show considerable variation (Brunig 1973; C.G.G.J, van Steenis, pers. comm.). In normal lowland forest, fresh, green-coloured vegetation evenly and loosely fills the space from the floor to the multi-layered canopy, whereas in heath forest a low, uniform, single-storey canopy seems to be formed predominantly by crowns of large saplings and small 'pole' trees (Ashton 1971). These sometimes form such dense vegetation that walking can be difficult. For example, in East Kalimantan heath forest, 454-750 trees of 10 cm diameter at breast height and over were found per ha (Riswan 1981, 1982). Brown and reddish colours are common in the foliage of the upper part of the canopy (Ashton 1971). Many of the tree species have thick leaves (p. 286). The medium leaf thickness of 12 tree species measured by the CRES team was 0.25 mm (range 0.10-0.45 mm).
Trees of large girth or with buttresses, and large woody climbers are very rare. Palms are also found and the CRES team found several giant mountain rattans Plectocomia elongata, a species which, as the name suggests, is normally found on mountains, where the soils are also poor (Whitmore 1977) (p. 281). Thin, small climbers may be quite common, as may epiphytes, several of which have associations with ants or catch small insects. The ground flora is sparse and frequently consists of mosses and liverworts. The number of tree species is relatively low. Proctor et al. (1983) found 123 species in a 1 ha plot in Sarawak compared with 214 species in a 1 ha plot in nearby dipterocarp forest. However, a total of 849 species from 428 genera are reported from the heath forests of Sarawak and Brunei (Brunig 1973).
Many of these features of heath forest vegetation have parallels in mountain and peatswamp vegetation, which also grow on poor, acid soils (pp. 171 and 281). The similarities with peatswamp forests are particularly striking as many of the tree species are the same. In figure 5.3 (p. 173), the four types of peatswamp forest in Sarawak have, from tallest to shortest respectively, 27%, 55%, 54%, and 70% of their species also recorded from heath forest (Whitmore 1984).
Various aspects of heath forest suggest it is restricted in its productivity by the low nutrient content of the soil. First, heath forest generally has a lower biomass than lowland forests on usual latosols. For example, Proctor et al. (1983) calculated the above-ground biomass of heath forest in Sarawak to be 470 t/ha (dry weight) compared with 650 t/ha in nearby dipterocarp forest. These are probably both above average for their type. The basal area of trees in an East Kalimantan heath forest was 6.4-16.9m2/ha (Riswan 1981, 1982) compared with 25-30 m2/ha in a normal lowland forest. The heath forest examined by a CRES team at Padang Kekurai, Bangka, had a basal area of 13.5 m2/ha for trees of 15 cm diameter and over. Second, plants with supplementary means of mineral nutrition are common (e.g., myrmecophytes and insectivorous plants) (Huxley 1978; Janzen 1974b). Third, heath forest is easily degraded into padang if burned or abandoned after brief cultivation; that is, heath forest does not seem to regenerate. This may be because the nutrient level of the soil has fallen below some level critical for full regeneration. While the heath forest is intact, nutrients from the detritus are recycled into the vegetation efficiently by the plants' roots.
The small litterfall (leaves and small wood) in the plot of heath forest examined by Proctor et al. (1983) was relatively low in its concentration of nitrogen and this may be due to the very low soil pH which could limit the mineralisation of the organic nitrogen. They suggest that the small leaves typical of heath forest trees are likely to be an adaptation to low nitrogen levels but the mechanism of this adaptation is not explained. The leaf thickening had earlier been explained as an adaptation to periodic drought as well as to low levels of soil nutrients (Brunig 1970; Whitmore 1984). Experiments by Peace and MacDonald (1981) have, however, shown that the heath forest species they examined had no special ability to avoid or resist desiccation. However, if features of heath forest such as the relatively even canopy and the small, often slanting and reflective leaves are seen as adaptations to reduce water loss, then the roots would not have to cope with such large volumes of water containing potentially toxic hydrogen ions (very acid soils have high concentrations of hydrogen ions) and phenolic compounds. Phenolic compounds can significantly inhibit the uptake of ions by certain crops (Glass 1973, 1974), and occur in high concentrations in heath forest leaves.
Padang Vegetation. Padang is a shrubby vegetation in which the tallest trees usually reach only about 5 m, but trees reaching 25 m are not unknown. Plants from padangs on Bangka and Belitung have been collected by several botanists (van Steenis 1932). In the padang (Padang Kekurai) examined by the CRES team, the dominant small trees were Baeckia frutescens and Melaleuca cajuputih, but Calophyllum sp., Garcinia sp. and Syzygium sp. were quite common. The most notable absences were of trees normally common as colonisers of open ground such as Macaranga spp., Mallotus spp. and Melastoma spp. Bare patches of white sand were frequent and these were often bordered by the sedges Fimbristylus sp., Rhyncospora sp., Xyris microcephala, and Xyris sp.; the lily Dianella and the bush Rhodomyrtus tomentosa were also locally common. Pitcher plants Nepenthes sp. were abundant and Valeton (1908) also records the presence in the damper parts of Bangka padangs of the small insectivorous plant Drosera burmanii. The climbing epiphyte Dischidia sp. covered the lower trunks of some of the larger trees. This plant has yellowish leaves shaped like wrinkled, inverted saucers. When the plant is pulled off a tree and the underside examined, numerous ants are found sheltering beneath the leaves, which also cover the plant's roots. This mutualistic association is similar to that found in Myrmecodia (fig. 8.1) and Hydnophytum (Huxley 1978;Janzen 1974b) (both of which are also found in heath forests), with the plant providing protection and the ants supplying nutrients in the form of waste food, faeces and dead ants.
Figure 8.1. Myrmecodia tuberosa.
The leaf thickness of 10 plant species were measured and the median value was 0.30 mm (range 0.25-0.70 mm), thicker than leaves in normal lowland forest, indicating an impoverished soil (p. 286).
Padang has been regarded as degraded heath forest caused by fire or felling (Janzen 1974a; Whitmore 1984), and as a natural ecosystem in its own right growing on extremely impoverished soils (A.J.G.H. Kostermans, pers. comm.; Specht and Womersley 1979). The former is certainly true but it does not necessarily follow from that that the latter is false. Padang is very slow growing (Riswan 1982). Janzen (1974a) mentions 1-2 m of plant growth in 30 years on the padang at Bako National Park, Sarawak, and it seems unable to regenerate back into heath forest (p. 377). Decomposition of dead plant material (burned stumps, etc.) is also very slow and it is possible that a padang formed many decades (or even centuries) ago might show no signs of its heath-forest past. Kartawinata (1978) reports having found charcoal 50 cm below the soil surface in a heath forest in East Kalimantan. This may indicate that heath forest regenerated from a padang vegetation formed by fire but the depth of the charcoal shows that the fire was clearly a very long time ago.
From the work of Brunig, Specht and Womersley (1979) counted 83 plant species that had been recorded from padangs in Sarawak and Brunei, and 48 of these were not found in heath forest.
Fauna
Animals in heath forest and padang vegetation face very similar problems to those in peatswamp forest. Poor soils cause low productivity which causes plants to defend their edible parts with toxic or digestion-inhibiting compounds to make themselves unattractive to herbivorous animals (Janzen 1974a). Not only that, but the nutritive value of the vegetation can be less on white-sand soils than on normal soil. Hardon (1937) analysed the chemical composition of leaves of two species of plant, one set from padangs on Bangka Island and another from specimens growing at the Botanic Gardens in Bogor on an andesitic laterite soil. The results are shown in table 8.2. Low nutrient levels were also found in leaves of heath forest trees by Pearce and MacDonald (1981).
From Hardon 1937
In view of the above it is hardly surprising that animal communities are considerably reduced in heath forest and padang vegetation. Wallace (1869) describes a visit to the port of Muntok on Bangka in 1861 and writes, "A few walks into the country showed me that it was very hilly and full of granite and lateritic rocks with a dry and stunted vegetation; and I could find very few insects". Janzen (1974a) describes his intensive, but barely successful, attempt to find vertebrates and invertebrates using baits and direct observations in the heath forests/padangs of Bako National Park, Sarawak, and writes that "the absence of bird calls was deafening". In the same area, Harrison (1965) estimated there was only one mammal per 2 ha and one bird per 0.4 ha. The results of collecting frogs, lizards and snakes in three forest types is shown in table 8.3. Heath forest had less than half the number of species found in the other forest types (Lloyd et al. 1968). Similarly, in a comparison of a number of forest types at Mulu, Sarawak, heath forest had the smallest number of dung beetles (Scarabaeidae) (Hanski 1983).
From what was written about the effects of nutrient-poor soils on the chemical defences of tree leaves in peatswamp forest (p. 175), it would be reasonable to expect similar strategies to reduce the impact of herbivores in heath forest. Only two studies are known that have compared the levels of phenolic compounds in leaves between a forest growing on acid, nutrient-poor white-sand (heath forest) soil and a forest growing on more usual soil. In the first study, McKey showed that there was approximately twice the concentration of phenolic compounds in the leaves of trees growing on the white-sand soils. Monkeys living in forest on usual soils ate a wide range of leaves, whereas those living in the forest on white-sand soils avoided almost all the leaves expect the most nutritious ones where the benefit of eating a good source of food was greater than the cost of ingesting defence compounds (McKey 1978; McKey et al. 1978). Although no sweeping generalisations should be made from results of a single study, there is reason to suppose that the acute lack of animals in heath forest in Sumatra and adjacent regions is at least partly due to higher than normal concentrations of toxic phenolic and other defence compounds. The second study found that the mean value of total phenols in oven-dry heath forest leaf litter was 2.49 mg/100 g, compared with 1.68 mg/100 g in nearby dipterocarp forest (Proctor et al. 1983). It might be expected, then, that consumption of the heath forest leaves would be relatively low. In fact, the percent of leaf area missing from fallen leaves (22% with no damage, 50% with 20% damage, 15% with 20%-40% damage) was similar to that found in the other three forest types examined (Proctor et al. 1983). This does not, however, necessarily negate the hypotheses of Janzen (1974a) because although leaves may fall after the same approximate degree of destruction, the rate of destruction of heath forest leaves may be much lower.
After Lloyd et al. 1968
Leaf samples from padang and heath forest were collected by the CRES team but the results of the chemical analyses are not yet available.
Fauna of Pitcher Plants. Pitcher plants provide an excellent example of an ecological microcosm. It is well known that insects are lured into the pitchers where they slip on the smooth, waxy lip and drown in the pool of water held in the bottom of the pitcher. Glands inside the pitcher secrete digestive enzymes and so the nutrients from the corpses are available to the plant.
It is less well known, however, that the water inside a pitcher supports a community of insect larvae and other organisms which are resistant to the digestive enzymes. Most of these animals feed directly on the drowned insects or feed indirectly on them by being predators on bacteria and other microorganisms which have themselves fed on the corpses. Since many of these species spend only the early part of their life cycle inside pitchers, they represent an export of nutrients which would otherwise have been used by the plant (Beaver 1979c, d).
Thienemann (1932, 1934) documented the fauna of pitcher plants in Sumatra, but the only study to consider the ecology of these unlikely communities is by Beaver, who worked on Penang Island. In the three species of pitcher plants examined (Nepenthes albomarginata, N. ampullaria and N. gracilis) he found 25 species of insect and three species of arachnid. The predominant family was Culicidae or true mosquitoes, but eight other families of flies (order Diptera) were found as well as representatives of the Hymenoptera (bees, wasps, ants), Lepidoptera (butterflies and moths), Araneae (spiders) and Acari (mites). Bacteria, protozoans, rotifers, nematode worms, oligochaete worms and crustaceans are also known from pitcher plants but these occur most frequently in old pitchers which have been deserted by insects and arachnids. These old pitchers no longer attract prey and so the input of nutrients is low and population levels of these microorganisms can never become high. Microorganisms are probably infrequent in younger, attractive pitchers because of the filter-feeding activities of some mosquito larvae. A negative correlation between protozoa and mosquito larvae in such habitats in North America was shown by Addicott (1974).
Figure 8.2. A vertical section of a pitcher showing a thomasiid spider waiting for its prey. The round marks below it are glands on the inside of the pitcher wall from which digestive enzymes are secreted.
The fauna of pitcher plants can be divided into three groups using terms coined by Thienemann (1932) according to the frequency of use of pitchers. Nepenthebiont species normally live only in pitchers and these account for 66% of the pitcher fauna found in Peninsular Malaysia and Singapore (Beaver 1979c). A few species are considered nepenthephils because they are common in pitchers but are frequently found in other habitats, and other species are called nepenthexene. This last group occur in pitcher plants but are far more often found elsewhere. One such nepenthexene mosquito, Aedes albopictus, is the only medically important insect to have been found in pitchers.
The fauna of pitchers occupy a range of feeding and spatial niches. The drowned animals in the pitcher fluid are fed upon directly by the larvae of certain Diptera, the larger species of which generally only tolerate a single individual because of aggressive competition for limited food supply. Mosquito larvae generally feed by filtering the microorganisms from the pitcher fluid, but some will also browse over the surface of the corpses, ingesting loosened particles. At the very bottom of the pitcher, finely divided detritus from the remains of animals and plants accumulate and this is fed upon by the larvae of biting midges (Ceratopogonidae) and by mites. The pitcher fluid is also inhabited by insect predators such as a predatory mosquito larva which preys on other mosquito larvae; because of their large size, only one is found per pitcher. Above the fluid, a thomasiid spider (fig. 8.2) waits, and the predatory larva of a mycetophilid fly spins a web, to intercept insects falling from the pitcher rim as well as adult insects emerging from the pitcher after developing in the fluid (fig. 8.3). Finally, small encyrtid and diapriid parasitic wasps lay eggs in the pupae of phorid flies (fig. 8.4). A simplified food web for the community in two species of pitcher plants is shown in figure 8.5.
Figure 8.3. A horizontal cross-section of a pitcher showing the web of a mycetophilid fly larva.
Figure 8.4. A phorid fly, larvae of which are parasitised by small wasps in pitchers. Phorids are peculiar flies in that most species are virtually wingless and most are less than 1 mm long.
After Mani 1972
Figure 8.5. A simplified food web of the community living in a pitcher, showing the functional groups.
After Beaver 1979d. pers. comm.
Distribution of Heath Forest on Bangka and Belitung Islands
It has been suggested or implied that the natural vegetation for the majority of Bangka and Belitung Islands would have been heath forest (Frankena and Roos 1981). This may not be the case. Berkhout (1895) and de Leeuw (1936) described the forests and general situation of Bangka Island but make no mention of a forest like heath forest, nor of species that would be especially expected from heath forest or padang. Instead they describe the conventional dipterocarp forests that today clothe only the hills. Descriptions of birds (Berkhout 1895), fishes (de Beaufort 1939) and snakes (Westermann 1942) from Bangka and Belitung Islands make no mention of an unusual forest type nor give any hint that the fauna was poor in species. Heath forests are notoriously poor in species (see p. 258). A.J.G.H. Kostermans (pers. comm.) spent six months collecting botanical specimens primarily in the south of Bangka Island and found areas of padang but no heath forests nor brown or blackwater streams, as would be expected to flow out of heath forests (p. 254). However, when flying from Pangkal Pinang to Jakarta it is possible to see blackwater rivers disgorging into the sea on the east-facing shore of Bangka. In addition, van Steenis (1932) has suggested that padang vegetation on Bangka Island is restricted to parts of the north and northeast and he makes no mention of heath forest.
After several days of travelling on Bangka Island, the CRES team finally found a belt of heath forest in the north of the island about 10 km northeast of Belinyu, on the landward side of 'Padang Kerurai'1. It is suspected that heath forest, on Bangka Island at least, is restricted in general to the soils called Brown Yellow Podzolic/Podzol Association by Soepraptohardjo and Barus (1974), and in particular to where the podzols are more pronounced, such as the old marine terraces near the coast. Hardon (1937), however, described a podzol from a site more than 12 km from the sea. The CRES team also found a white-sand area near Kurau, 25 km south of Pangkal Pinang, at the site of a large transmigration site for fishermen. All forest/scrub had been cleared from at least 200 ha and only a few paper-bark Melaleuca trees and coarse sedges Gahnia grew on the poor soil. Janzen dedicated his paper about the ecology of white-sand soils to "those unfortunate tropical farmers (who) attempt to survive on white sand soils... because they appear unexploited" (Janzen 1974a).
In inland areas of Bangka Island which are also shown on maps as having Brown Yellow Podzolic/Podzol Association soils and being over old sandstone parent material (e.g., around Poding and Petaling), small rubber plantations and fruit orchards are common. Large areas are used for pepper and this suggests that the capability of the land is limited but not critical. In the descriptions of typical heath forest/padang soils, however, it is suggested that they are virtually unusable for permanent agriculture. This suggests that heath forest would not necessarily have covered these flat or undulating inland areas.
It is interesting that in the form of the Melayu language spoken on Bangka Island, the word 'ngarangas' (similar to kerangas) is used to describe the dead and dying trees around the tin tailings. It is possible that ngarangas used to refer to heath forest.
The present area of intact or partially disturbed heath forest is not known but it is probably much less than that estimated by FAO/Mac-Kinnon (1982). It is probably Sumatra's most endangered ecosystem, and a thorough survey should be undertaken as soon as possible.
IRONWOOD FOREST
Introduction
Ironwood forest in Sumatra is of special interest because of its extremely low diversity of tree species, being dominated by the ironwood Eusideroxylon zwageri. The ecological interest centres on its apparently stable state despite its species-poor composition.
The Tree
Ironwood is a member of the laurel family (Lauraceae) and is distributed through Sumatra, Kalimantan, and the southern Philippines. There used to be two members of the genus Eusideroxylon but the one called melagangai has recently been moved to its own genus Potoxylon by Kostermans (1979). Ironwood is a large tree (50 m tall and 2.20 m diameter at breast height) with a warm red-brown bark, large leaves and heavy (± 300 g) fruits. Its timber is economically very valuable because of its strength and durability — it can resist rotting for up to 40 years when in constant contact with wet soil, or a century in dryer conditions (K. Kartawinata, pers. comm.; Riswan 1982; Suselo 1981). It is used for bridges, piers, road foundations, floors and, of course, the popular 'sirap' roofing tiles.
Ironwood seems to be restricted to southern Sumatra, having been found in Jambi, South Sumatra, and Bangka and Belitung Islands (De Wit 1949; Greeser 1919; Koopman and Verhoef 1938; Soedibja 1952), but its range has been greatly reduced by the heavy demand for its timber. It has been shown that ironwood could be an important plantation species, and guidelines for its cultivation have been proposed (Koopman and Verhoef 1938). Commercial germination of seeds and growth of seedlings can be quite successful if conducted in moist, shady places such as beneath secondary growth with a closed canopy (Koopman and Verhoef 1938; Tuyt 1939). The main threat appears to be a certain amount of predation of the seeds by porcupines (p. 270).
A characteristic feature of ironwood is that, when felled, numerous coppice shoots grow from its base. If cut carefully these can be used for plantation stock, and Beekman (1949) found that these shoots grew faster than seedlings.
The Forest
Ironwood forest may have ironwood as virtually the only species of large tree present, or ironwood may be one of many species but clearly the dominant one. An example of the former was investigated by a team from CRES in Rimbo Kulim2 near the River Kahidupan, north of Muara Tembesi in Jambi. It was difficult to find a patch of unexploited forest but eventually sufficient was found to enumerate 0.5 ha (10 x 500 m) of undisturbed forest. A total of 84 trees of 15 cm diameter at breast height were found and 81 (96%) of these were ironwood (fig. 8.6). Two of the remainder were unidentified small trees, and the other was a large specimen of Koom-pasia sp. The Simpson Index of species diversity (p. 189) for this situation is only 0.07. As many as 33 ironwood trees of 20 cm diameter and over can be present per ha in Kalimantan (van der Laan 1925). The equivalent figure for the forest examined by the CRES team was 126, indicating a perhaps unusual richness.
When inside the forest at Rimbo Kulim it is not immediately obvious that the forest is species-poor because there is a wide range of tree sizes. It is quite unlike, for example, being in a mangrove forest where the vegetation within a zone is relatively homogenous. The distribution of trunk diameters for ironwood trees at Rimbo Kulim is shown in figure 8.7. The undergrowth consists of many ironwood seedlings as well as small palms (Pinanga sp. and Licuala sp.) and the tall palm Livistona sp.
The ironwood forest examined by Suselo (1981) was not far from Rimbo Kulim and was also dominated by ironwood, but ironwood was only one of 37 tree species in his plot of 0.2 ha; He found 65 trees of 10 cm diameter or more, but only 17 (26%) were ironwood (fig. 8.8). (A similar percentage was found in a nearby plot (Franken and Roos 1981). Iron-wood, however, formed the main canopy of the forest. In ironwood forest examined by Greeser (1919), Koompasia, Shorea or Intsia were the main emergents and ironwood was a subdominant species.
In other areas where ironwood grows, such as Bangka Island, it can occur in mixed dipterocarp forests as just an occasional species.
Soils and Topography
Witkamps (1925) considered that ironwood was generally restricted to sandy soils of Tertiary origins. The ironwood forest examined by Suselo (1981) was, however, on a clay-loam (for the top 20 cm) on top of clay. The soil analysis for three samples taken at Rimbo Kulim is summarised in figure 8.9, and reveals a sandy silt-loam (but close to a clay-loam) top soil with clay below. The soil is moderately acid with quite high carbon but a moderate C/N ratio. The soil contains very low concentrations of phosphorus, sodium and magnesium but a fairly high concentration of potassium. The total exchangeable bases are very low and the cation exchange capacity quite high.
Figure 8.6. Relative abundance of the four tree species (n= 84) at Rimbo Kulim, Muara Tembesi, Jambi. Ninety-six percent of the sample comprises ironwood Eusideroxylon zwageri.
Data collected by a CRES team
Figure 8.7. The distribution of different diameter classes among the ironwood trees at Rimbo Kulim, Muara Tembesi, Jambi.
Data collected by a CRES team
Figure 8.8. Relative abundance of tree species (n= 65) at the forest examined at Dusun Aro, Jambi. a- Eusideroxylon zwageri, b- Hydnocarpus polypetala, c- Aglaia sp., d- Dysoxylon sp., e- Diospyros sp., f- Ixonanthes icosandra, g- Knema sp., h- 23 other species.
Based on data from Suselo 1981
The distribution of ironwood forests shown in figure 1.11 (p. 21) coincides with areas of red-yellow pozolic soils on folded sedimentary rock shown on the map in BPPP/Soepraptohardjo et al. (1979). There are, however, areas with the same soil characteristics in the same climatic zone where ironwood is known not to dominate the forests. Various authors have commented on ironwood's preference for slightly undulating topography but the cause of any limitation has not been established.
Fauna
Nothing seems to have been written on the fauna of ironwood forest. The birds seen during the CRES visit to Rimbo Kulim seemed similar in diversity and number to those in normal dipterocarp forest. Two groups of the unnamed black-and-white subspecies of leaf monkey Presbytis melalophos were seen but were not observed to feed. Where ironwood is virtually the only tree species, food may be a problem for many animals. The mature leaves are thick (0.35 mm) and coarse, and the fruit are very tough, large and heavy such that almost no arboreal animal would be able to eat them. An exception might be Prevost's squirrel Callosciurus prevosti which was seen several times at Rimbo Kulim. A few fruit were taken from Rimbo Kulim and small holes were observed in the fruit skin, caused probably by small beetles. Investigation showed that the seed had been slightly damaged but apparently not enough to halt germination.
Figure 8.9. The structure of soils under two ironwood forests in Jambi. Rimbo Kulim soil: 1-1.3 cm, 2-5-10 cm, 3 - 10-15 cm (soil collected by CRES team). Dusun Aro soil: a - 1-5 cm, b - 20-30 cm.
Data from Suselo 1981
In the forest investigated by Suselo (1981), 14 of the 37 species found in the plot produce fruit which would be suitable for frugivorous primates such as gibbons (Whitten 1982e). In the forest around Rimbo Kulim, dark-handed gibbons Hylobates agilis were heard singing, presumably where ironwood was less dominant.
Ecological Significance
There is a general and popular view that in tropical regions where climate and soils are favourable for plant growth, high species diversity in natural ecosystems is necessary for ecological stability. On page 79 it was shown that mangroves do not conform to this preconception, and it would seem that the same applies to certain ironwood forests. Indeed, the former ecological tenet of ecosystem complexity leading to stability has been called into question.
On page 221 the strategy of gregarious fruiting was described for the dominant, large dipterocarps which, at their canopy level at least, form species-poor stands. Janzen (1974a) suggests that a second type of strategy common in species which occur at relatively high densities, is the production of large and toxic seeds on a more or less continuous basis. Some mangrove species fall into this category and so, apparently, does ironwood. Koopman and Verhoef (1938) describe the flowering of ironwood as irregular but with a peak around the middle or end of the dry season. Ripe fruits are found approximately three months after flowering. The results of analyses of ironwood seeds collected by the CRES team are not yet available but the very large numbers of seeds found on the ground, and the very large numbers of seedlings under parent trees (Suselo [1981] found 920 saplings per ha) is indicative of very little seed predation. Porcupine damage of ironwood seed-beds has not been described in detail but it is conceivable that only parts of the seed were eaten and, like the monkeys described eating leaves from trees growing on white-sand soils (p. 259), the benefits of eating a highly nutritious item of food may exceed the costs of ingesting defence compounds. In any case, porcupine damage of seeds in natural situations does not seem to be important.
The CRES team also collected leaves, but again the analysis results are not yet available. However, their feel suggests a high concentration of fibre (lignin) and the timber itself must be defended very effectively against insect attack for it to be so exceptionally durable. The ecology of a leguminous tree called Mora was studied in Trinidad where it accounts for 40%-60% of trees in the forests where it is found. Most of what Janzen proposed as theory for such trees, as described above, fits for Mora. It suffers very little seed predation in comparison with other trees around it. Its major seed predator is a large rat-like rodent Oryzomys but the seeds are generally less than one-third eaten. Indeed, this damage is usually confined to the surface of the cotyledons and the seeds were still able to germinate (Rankin 1978). It is possible that Mora seeds, and those of ironwood, are protected by lectins, a group of glycoproteins which are sticky and act as digestion inhibitors. Lectins are very strongly selected against by rodents (Janzen 1981c) and this many be the reason porcupines do not do more damage to the abundant ironwood seed crop.
Species-poor ironwood forests would make ideal study areas to try to unravel the reasons why such abundant trees are not exploited or controlled by herbivores. Species-rich forests are incredibly complex and any ecological investigation confronts numerous confounding variables. Iron-wood forests have some of these variables controlled and if an understanding of the ironwood ecosystem can be achieved, then attempts to understand species-rich forests are more likely to succeed.
Unfortunately, licences continue to be granted for the exploitation of ironwood trees and so it seems that the time remaining to make a thorough investigation of the exceptional forests of Sumatra is severely limited.
FOREST ON LIMESTONE
Introduction
Sumatra has limestone areas although few exceed 10 square kilometres in size (fig. 8.10). Limestone forms a characteristic landscape called karst which occurs in two forms, cockpit (or labyrinth) karst and tower karst, both which are found in Sumatra. Cockpit karst such as at Lho'Nga, North Aceh, and in the hills near Payakumbuh, has a regular series of conical or hemispherical hills and hollows with moderately steep sides (30-40°). Tower karst such as is found in the river plants near Payakumbuh consists of low (± 300 m) isolated hills with precipitous sides (60-90°) and is often riddled with caves, separated by flat depressions (Jennings 1972; Ver-stappen 1960). The surface of the limestone may be either rocky or covered with intricately sculptured, razor-sharp pinnacles as at Lho'Nga.
Soils
On moderate limestone slopes and hollows, brownish-red latosols would be expected to be formed which are clay-rich and leached. As might be predicted from the parent material, the soils are often richer in bases, particularly calcium and magnesium, with a higher cation exchange capacity than soils in similar situations but on other parent materials. On limestone crests and shelves an acid, humus-rich, peat-like soil can develop directly on top of limestone rock, but this has not been reported (as far as could be determined) from Sumatra. On steep slopes, of course, soil cannot form except in very small quantities in fissures and cracks (Burham 1975). Soils over limestone can be very variable depending on the purity of the parent material and the topography (Crowther 1972a).
The soil beneath the forest on limestone examined by Proctor et al. (1983) in Mt. Mulu National Park, Sarawak, lay on a 25-30° slope and was extremely shallow (average 11 cm, range 0-55 cm), and highly organic (80% weight lost on ignition). Its base saturation was 32% compared with 1.6% in dipterocarp forest and 2.9% in heath forest. Similarly, its cation exchange capacity was 210 m equivalents/100 g, compared with 37 in dipterocarp forest and 110 in heath forest. Despite the high organic content of the soil on limestone, its pH was 6.1, the highest recorded from the forests on Mt. Mulu.
Figure 8.10. The distribution of limestone in Sumatra.
Adapted from Verstappen 1973
Soil collected by a CRES team from small ledges on the rock walls outside Kutabuluh Cave, Tanah Karo, was basic and had very high concentrations of calcium and magnesium (table 8.4).
Vegetation
Anderson (1965) identified a range of habitats influenced by limestone in Sarawak.
a) At the base of limestone hills the soils, although derived from other parent material, are influenced by the water running off the limestone and by eroded limestone fragments. The relatively base-rich, fertile soils support a number of characteristics tree species but cultivation often extends to the very base of the hill.
b) Characteristic species can also be found at the base of limestone hills where the base itself is also over limestone.
c) The steeper slopes can support an untidy forest of trees whose roots cling to the rough surface or penetrate the rocks to emerge in caves below. Where the slope is too steep, small herbs thrive, many of which are not found in any other habitat. Most of these plants are subject to repeated water stress because the small volume of meagre soil in which they grow does not have sufficient water capacity. As a result they exhibit 'poikilohydry' — the ability to lose most of their water, except that in their protoplasm which resists drying out, and then to revive on rewetting (Whitmore 1984).
d) Small scree slopes also support a characteristic flora.
e) Forest of the moderate slopes and summits of limestone hills grows in the peat-like humus described in the previous section, and has certain similarities in species with those found in heath forest and inner peatswamp forest (see Anderson [1965] for details).
This was one type of forest examined as part of the study by Proctor et al. (1983a, b). In their 1 ha plots they found only 74 tree species in forest on limestone, compared with 215 in dipterocarp forest and 122 in heath forest. Forest on limestone also has a relatively low density of trees, small total basal area and mean basal area per tree, low tree height and trunk diameters of 10-20 cm. Dipterocarpaceae was the most important plant family. The phenolic content of leaf litter from forest is not easily interpreted, possibly because leaf litter rather than leaves taken directly from the parent plant was examined.
Limestone hills are peculiar in their discontinuous distribution, their relatively base-rich soils, the high levels of calcium in the rock, and their very steep or sheer slopes, which can become periodically very dry. It is not surprising, then, that some plants are confined to them. For example, of 1, 216 plant species recorded on limestone hills in Peninsular Malaysia, 254 species (21%) are found only on limestone hills and 130 of these species are only found in Peninsular Malaysia (Chin 1977). Considering the greater extent of karst limestone in Sumatra, it would be safe to suggest that an even greater number of plants are endemic to Sumatran limestone. About 10% of the plant species found on, but not restricted to, limestone in Peninsular Malaysia are specialised in living in crevices (Henderson 1939). Peninsular Malaysian limestone hills also represent the only known habitat of a genus of small serdang palms Maxburretia (J. Dransfield, pers. comm.; Whitmore 1977, 1984), and similar plants should be looked for in any ecological surveys of limestone hills in Sumatra. For example, the CRES team visiting the Lho'Nga limestone in North Aceh found a species of pigeon cane palm Rhapis which was, until then, a genus unknown in the Malesian region.
Fauna
The only fauna restricted to limestone habitats are certain molluscs. This was first recognised in the last century but the only thorough account of the phenomenon for the Sunda Region appears to be by Tweedie (1961). He examined the mollusc fauna of 28 more-or-less isolated limestone hills in Peninsular Malaysia and found a total of 108 species. The most remarkable finding was that 70 (66%) of these species were restricted to just one limestone hill. One hill even had seven mollusc species not found on any of the other hills (fig. 8.11). These endemic species must be extremely intolerant of living away from their limestone environment. Strangely, other species from the same genera as the endemic species were found on a wide range of hills; one such was collected from no less than 19 of the 28 hills. Some of these wider-ranging species have indeed been found in areas between the limestone hills, but others must be present in those inhospitable (for a mollusc) intervening habitats in such low numbers that transfers between hills are extremely rare occurrences.
Figure 8.11. The distribution of endemic species of snails of different numbers of limestone hills. Thus, four hills each had a total of four endemic snail species, and five hills had only one endemic species.
Based on data from Tweedie 1961
Similar situations to the above doubtless exist for the isolated limestone outcrops in Sumatra. Obviously, those hills supporting more endemic species would be more worthy of protection from use as, for example, limestone quarries. Anyone interested in pursuing such a line of study should consult van Benthem Jutting (1959) as the first step.