Chapter Six
INTRODUCTION
The composition and ecology of lowland forests are extremely variable due to a variety of physical factors. The more obvious forest types on distinctive soils or in dry areas are described below, together with any noteworthy aspects of their animal life.
PEATSWAMP FOREST
Formation and Location
Peat is a type of organic soil which is at least 50 cm deep and which contains more than 65% organic matter (that is, it loses more than 65% of its dry mass when burned). Peat deposits can be of two forms although only the second is found in Sulawesi:
• Ombrogenous peat: the most common type in Southeast Asia which has its surface above the surrounding land. Plants that grow on this peat utilize nutrients only from within the plants themselves, the peat or directly from the rain; there are no nutrients entering the system from the mineral soil below the peat or from rain water flowing into it. This type of peat is usually found near the coast behind mangrove forest and can be extremely deep (about 20 m). The peat and its drainage water are very acid (pH <5.0) and poor in all nutrients (oligotrophic) particularly in calcium. Outside Sulawesi, vast areas of ombrogenous peatswamps occur in South and West Kalimantan and eastern Sumatra.
• Topogenous peat: this peat occurs in depressions. Plants growing on this peat can extract their nutrients from the mineral subsoil, river water, plant remains and rain. Topogenous peat can be found behind coastal sand bars and at other sites where free drainage is hindered. The peat is usually found in a relatively thin layer (less than 4 m).
Peat accumulation in topogenous swamps is relatively slow and the high acidity (low pH), characteristic of ombrogenous swamps, is less pronounced. Instead, the peat and drainage water are only slightly acid (pH 5.0) and nutrients are relatively abundant (mesotrophic).
The only major area of peatswamp in Sulawesi is Aopa Swamp, 100 km west of Kendari, and it forms much of the northern portion of the Rawa Aopa-Watumohae National Park. The major attractions of this park are fishing and watching the traditional hunting of deer by horsemen using lassos1 (Anon. 1985). Some years ago there were plans to mine the peat of Aopa Swamp and to burn it for electricity generation but companies have been far more attracted by the massive volumes of peat available in Borneo and Sumatra. Aopa Swamp is all that remains of an ancient lake which is reaching the final stages of its life as an aquatic ecosystem. In contrast to the extensive ombrogenous peat swamps which arose because of the inability of decomposer organisms to break down organic material in saline conditions, the Aopa peat was formed because of the continuous, water-logged conditions which are unfavourable, but not anathema, to decomposer organisms. A contour map shows how the swamp is almost entirely surrounded by high ground (fig. 6.1).
Aopa Swamp covers a maximum area of about 31,400 ha, and this is surrounded on all sides by grasslands, cultivation or transmigration settlements. The area of the swamp varies through the year depending on rainfall. It is at its maximum between May and September, falling to 20,000 ha between November and December, about 15,000 ha in January, rising to 28,500 ha between March and May. Its depth varies between 1 m and 2 m depending on location and time of year. The rivers draining the swamp, particularly in the western half, are dystrophic, that is, the water is shallow, nutrient-poor and dark-coloured because of suspended plant colloids. The abundant organic matter is decomposed by microorganisms, most of which require oxygen, and as a result the concentration of dissolved oxygen is low (Anon. 1983). Small areas of coastal peatswamp also exist such as in the west of Muna Island (Anon. 1982) but these have so far received no attention from ecologists.
Vegetation
The extent to which the vegetation of Aopa Swamp is known is thanks to a small team which in 1978 made preliminary botanical observations there2. Aopa Swamp is in fact two swamps partially separated from each other by a ridge extending southwards from Mt. Makaleleo. The swamp to the east is relatively disturbed and has only a few small trees of she-oak Casuarina (Casu.) scattered all over the area beneath which grow sedges, such as Scleria (Cype.). The next most abundant tree is a Eugenia (Myrt.) which has aerial roots produced from the trunk up to 2 m above the ground. Other plants include sago palms Metroxylon sagu, a large tree palm Pholidocarpus, occasional Corypha tree palms (p. 481), 5 m-tall pandans, at least two species of climbing rattan, and epiphytic Lecanopteris ant ferns (p. 509). Epiphytes in general, however, are uncommon.
Figure 6.1. The location of the enclosed Aopa Swamp with land 400-1,000 m (medium grey), and 1,000+ (dark grey).
When approaching the western swamp from the road which crosses the Aopa River, one passes expanses of open water. The most spectacular plants at the edges are lotus lilies Nelumbo nucifera (Nelu.) (p. 281). Before reaching the forest, there is a floating mat of vegetation dominated by a tall papyrus-type sedge Cyperus sp., and in the drier parts the even taller grass (>2m) Saccharum spontaneum (Jacobs 1979; Susanto 1984). The floating mat of vegetation is probably extending its area by growing over the water surface; one floating runner, for example, was 7.5 m long.
The forest in the western swamp is more luxuriant than in the east. The trees grow to 35 m tall, but no conifers or she-oaks were found by the team in 1978. Large buttresses are common but aerial roots are rare perhaps because the substrate is more stable than in the eastern part. Strangling fig trees are a distinctive component. There are many palms: the trunkless Licuala, the tree palms Livistona, sugar palm Arenga, the spiny Oncosperma, rattans, but no sago. Climbers are common as are creepers such as the small pandan Freycinetia and the superficially similar Pothos (Arac.). Epiphytes are comparatively rare but the herbs of the undergrowth are varied and interesting (Jacobs 1979). Other trees recorded are Hopea gregaria (Dipt.), Baeckia frutescens (Myrt.), Saccopetalum horsfieldii (Anno.), Planchonia valida (Lecy.), Diospyros malabarica (Eben.), Artocarpus teysmannii (Mora.) and Calophyllum soulattri (Gutt.) (fig. 6.2) (Anon. 1978; Susanto 1984). This last plant is particularly easy to identify because its leaves, in common with other members of its genus, have numerous straight veins, all running parallel to each other on both sides of the midrib.
Figure 6.2. Calophyllum soulattri. Scale bar indicates 1 cm.
After Soewanda n.d.a
The two most common trees in the peatswamp in west Muna are members of the teak family Verbenaceae, that is Geunsia paloensis and Premna foetida (Anon. 1982).
Fauna
Elsewhere in Southeast Asia it has been found that densities of mammals in peatswamps are generally lower than in most dryland lowland forests. Detailed studies of animals in Aopa Swamp have not been conducted but it appears that most of the larger mammals are present with the exception of the Sulawesi civet Macrogalidia musschenbroeckii (p. 532) which has never in fact been reported from anywhere in Southeast Sulawesi. Wallows surrounded by tracks of the lowland anoa Bubalus depressicornis are found in the forest although this animal is not particularly common. In addition to the indigenous fauna, wild water buffalo and deer are found, particularly in the grassy areas of the eastern swamp (Anon. 1984).
A single day's bird watching in Aopa Swamp resulted in 80 species being seen (Anon, n.d.), a relatively large total for Sulawesi, and many of these were associated with the water (p. 305).
Three large reptiles are reported from the swamp: the sailfin lizard Hydrosaurus amboinensis (p. 301), crocodile Crocodylus porosus (p. 302) and the reticulated python Python retirulatus which is hunted nowadays by some of the Balinese transmigrants for food.
FRESHWATER SWAMP FOREST
Physical Conditions
Freshwater swamp forest grows where there is occasional inundation of mineral-rich freshwater of pH 6 and above and where the water level fluctuates such that drying of the soil surface occurs periodically. The floods may be a result of heavy rain or of river water backing up in response to particularly high tides. The major difference from conditions of the peatswamp forest described above is that the peat layer is more or less absent although a thickness of a few centimetres may be found on the soil surface. Freshwater swamps are normally found on riverine alluvium3 but also occur on alluvium deposited in lakes such as in the southeast of Lake Lindu, the southeast of Lake Poso (Sarasin and Sarasin 1905), and the south of the Ranu lakes (p. 259) (L. Clayton pers. comm.).
The alluvial soils are more fertile than those on adjacent slopes and have great agricultural potential when drained. The soils are very variable but are generally young and therefore do not have clearly differentiated horizons. They comprise relatively fine particles and the general colour is grey turning mottled grey or orange-grey where occasional drying occurs as a result of gleying.4 Soil animals such as termites and earthworms, which normally take organic matter into the soil, are not tolerant of the waterlogged anaerobic conditions and so there is little mixing of soils.
The agricultural potential of these swamp soils has meant that fresh-water-swamp forest have suffered greatly from human activities (p. 97).
Vegetation
The vegetation of freshwater swamps varies according to the wide variation in its soils and the proximity to free water. Few if any plants are restricted to freshwater swamps and these forests are relatively poor in species. The fringes of the swamp near open water are grassy but, where the ground is firmer, palms and pandans can be common. Behind these, a forest very similar to normal dryland lowland forest is generally found. The relative instability of the soil, by virtue of its periodic inundation with water, is probably one reason that supportive structures associated with trunks such as long, winding buttresses and stilt roots are common, and certain species have accessory breathing structures such as pneumatophores (p. 121).
Only four areas of freshwater swamp appear to have been examined in Sulawesi:
• a patch of coastal freshwater-swamp forest on the west coast of Sulawesi across the hills from Donggala and Palu was found to contain Barringtonia racemosa (Lecy.), the rare Quassia indica (Sima.) (fig. 6.3), Terminalia copelandii (Comb.), Polyalthia lateriflora (Anno.) (with pneumatophores like Sonneratia), Elaeocarpus littoralis (Elae.) and Horsfieldia irya (Myri.) (Meijer 1984);
• an EoS team examined a freshwater-swamp forest adjacent to the transmigration settlement of Tinading, Baulan, Toli-Toli, and at least 90% of the trees were Terminalia copelandii;
• the freshwater-swamp forest in the southeast of Lake Lindu comprised, in 1934, almost pure stands of Nauclea sp.5 (Rubi.) trees and tall grasses (Bloembergen 1940), but it is not clear whether these still exist;
• freshwater-swamp forest near the shores of Lake Ranu in Morowali National Park contains the trees Palaquium (Sapo.), Manilkara (Sapo.), Mimusops elengi (Sapo.), Calophyllum soulattri (Gutt.), Parinari corymbosa (Rosa.) and Haptolobus celebicus (Burs.) (Anon. 1980).
Pitcher plants Nepenthes (Nepe.) appear to be ubiquitous in freshwater-swamp forest. They are generally climbing or scrambling plants with the climbing stem arising from a rosette of young leaves. A single plant can bear two or three distinct forms of pitchers; the pitchers on the ground generally have a rounded base, whereas those somewhat higher up are more slender and have two frilly vertical wings on their front. The uppermost pitchers are longer and more funnel-shaped and their wings are reduced to prominent ribs. The variety of forms has led to confusion in identification, and the number of species on Sulawesi is not yet known since new species continue to be recognised (Kurata 1983; Turnbull and Middleton 1984). The lip of the pitcher is ribbed and often toothed around the inner edge where nectar glands are situated. The lid, which also has nectar glands, projects above the pitcher but does not prevent rainwater from entering. The nectar and perhaps the colour of the pitchers attracts insects many of which slip on the pitcher lip and fall into the liquid below.6 The inner wall of the pitcher is divided into two zones: the upper half is waxy, very smooth, and more or less impossible for insects to climb up, and the lower zone which has numerous glands that secrete digestive enzymes. These enzymes digest unwary insects that fall into the deadly brew having been attracted by the nectar and the blood red colour. The fluid in the pitcher has a pH of 3 to 6 and this favours the action of the digestive enzymes, and the plant then absorbs the available nutrients.
Figure 6.3. Quassia indica. Scale bar indicates 1 cm.
After Nooteboom 1962
It is well known that insects are lured into the pitchers, but it is less well known that the water inside a pitcher supports a community of winged-insect larvae and other organisms that are resistant to the digestive enzymes (fig. 6.4). Most of these animals feed directly on the drowned insects or feed indirectly on them by being predators on bacteria and other microorganisms that 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 1979a, b), but on the other hand they enhance the plant's digestive process by breaking down the accumulating detritus of dead animals into nitrogenous compounds (Kitching and Schofield 1986).
Figure 6.4. A pitcher plant showing the animal community within it.
After Kitching and Schofield 1986
The community of animals in a pitcher is relatively poor in species, because of the specialized nature of their habitat, and as such is amenable to study. Most of the animals exploit the detritus: larvae of biting midges and mosquitoes filter out and feed on fine particulate and suspended material, while mites feed on the very finest material and fungal strands. These animals are protected from many potential predators, but not from all. Predatory larvae of syrphid hoverflies crawl among the detritus to catch fly larvae but are generally seen at the water surface with their posterior breathing tubes poking just out of water (Kitching and Schofield 1986). Root maggot flies lay eggs in the unopened pitchers but by the time the larvae are ready to feed the first prey will be available (fig. 6.5) (Kevan unpubl.).
Figure 6.5. Members of the pitcher plant community. a - adult and larva ceratopogonid biting midge, b - adult and larval syrphid hoverfly, c - adult and larval anthomyiid root maggot fly.
The food webs in pitcher plants in the seasonal Bogani Nani Wartabone National Park are relatively simple compared with those found in Peninsula Malaysia and in Morowali National Park (L. Clayton pers. comm.). This agrees with the hypothesis that complex food webs tend to occur only where the environment is relatively stable and predictable (Kitching and Schofield 1986; Kitching in press).
Around Lake Tempe, in those areas that are little disturbed by people, a form of swamp forest is dominated by Gluta renghas (Anac.) (Anon. 1977), which has a pinkish-brown, warty, puckered fruit. The seed can apparently be eaten after roasting (Gluta is in the same family as the cashew Anacardium occidentale). Otherwise its parts should be regarded as poisonous and it is unwise even to shelter under these trees during rain, since the resin washed from the leaves and branches can cause severe itching. The trees can usually be recognized by the black stains of dried sap on the bark, dippled scaly bark, and spirally-arranged leaves (Burkill 1966; Ding Hou 1978).
RIVERINE FOREST
Vegetation
Riverine forest is found along rivers, influenced by both the wet conditions and by the new sediment deposited by rivers on the inside of meanders. The only location in Sulawesi where riverine forest has been studied in any detail is in the Sopu valley northeast of Lake Lindu and Mt. Nokilalaki (van Balgooy and Tantra 1986). The forest here was dominated by the distinctive Eucalyptus deglupta (Myrt.) (fig. 6.6), some of which reached 60 m in height with a 1.5 m girth but in some areas (see below) it grows even taller.7 A 20 m x 100 m transect in the Sopu valley contained 115 trees of at least 10 cm girth belonging to 47 species, 38 genera and 28 families, but very few of these are restricted to riverine forest. Some trees such as the sugar palm Arenga pinnata and the pandan Pandanus sarasinorum, are common along the main stream but not the tributaries, but others such as Pigafetta filaris (Palm.) and Erythrina (Legu.) were common along the tributaries but not the main stream. Other large trees in the Sopu riverine forest were Duabanga moluccana (Sonn.) (Fig. 6.7), Ficus (Mora.) and Elmerillia ovalis (Magn.). One of the smaller trees Saurauia oligolepis (Acti.) is of special interest because its inflorescences grow from the base of the tree, but, descend into the soil and the white flowers are finally seen up to two metres away from the tree. This phenomenon of geocarpy—the pushing of fruit into the soil by the inflorescence—was observed in a number of species present in this forest (van Balgooy and Tantra 1986), and is also known in riverine species of figs (S. C. Chin pers. comm.). One possible advantage of geocarpy is that a buried fruit is less likely to be swept away by water when a river is in flood, although dispersal of fruits away from the parent plant is the more usual strategy (p. 380).
Riverine forests in Lore Lindu National Park are reported to contain Octomeles sumatrana, Eucalyptus deglupta and Duabanga moluccana (Wirawan 1981), whereas Dracontomelum dao (Anac.) and the red-barked Pometia pinnata (Sapi.) are common along both the Tumpah River near Toraut (Whitmore and Sidiyasa 1986) and the Boliohuto River in the west of Bogani Nani Wartabone National Park (EoS team).
In addition, it is reported from 50 years ago that the Dumoga valley, now the site of a major irrigation scheme (p. 629), was often under water and much of this area may once have been extensive riverine forest. The forest even then was in a disturbed or secondary state but the trees mentioned include Octomeles sumatrana (Dati.), Koordersiodendron pinnatum (Anac.), and Eucalyptus deglupta (Myrt.) (Steup 1933).
Eucalyptus deglupta dominates almost all riverine forests which are more or less undisturbed. It has the distinction of being the first-described species of Eucalyptus, and it is now know from Mindanao, Sulawesi, Seram, New Guinea, and New Britain, but it is grown as a timber and reforestation tree throughout the tropics. In Sulawesi it has been collected in all provinces except Southeast Sulawesi (Davidson 1977; Cossalter 1980) but it is almost certainly found there. Efforts have been made by foresters since the 1970s to make full use of existing E. deglupta genetic resources in order to assist in the domestication and improvement of the species. Evaluation of early experiments show that E. deglupta from Sulawesi are among the best for reforestation of wet lowland sites (Davidson 1977).
Figure 6.6. Eucalyptus deglupta. Scale bar indicates 1 cm.
After Soewanda n.d.b.
Eucalyptus deglupta is the only one of its huge genus (about 500 species) that is regularly found in a rain forest. It has the thinnest bark (3 mm) of all eucalypts and is, as a consequence, very sensitive to fire which is of course extremely rare in the moist habitats it inhabits (Pryor 1976). Stands of E. deglupta are generally dense and comprise individuals of similar size. Mature stands at Tobou-Tobou on the Tomini River, south of Lake Poso, have 30-40 trees per hectare with the average tree reaching 65-70 m, 40-45 m of which is clear bole, and a diameter at breast height of 115 cm. The tallest trees reach 75 m to 80 m, beneath which a dense lower stratum of forest has developed. Occasionally, however, mixed-age stands are found, such as at the confluence of the Sopu and Lindu Rivers, where the river-banks are unstable, and a mosaic of different-aged stands from seedlings to giant trees can (used to be) be found on the banks of both rivers. Similar stands can also be found near Sedoa, Wuasa and Torire on the upper part of the Lariang River. It is found in small groups at the base of steep mountain slopes, such as along the Tumpah River in the area used by Project Wallace in Bogani Nani Wartabone National Park, whereas it is found only occasionally on the banks of the five broad rivers flowing into the north of the Gulf of Bone (Cossalter 1980).
Figure 6.7. Duabanga moluccana. Scale bar indicates 1 cm.
After Soewanda n.d.b
The flowering of Eucalyptus deglupta occurs at the peak of the wet season when vegetative growth also reaches its maximum. The feathery flowers are pollinated by beetles, flies (syrphids and callophorids) and bees. Twelve weeks later the buds have developed into mature fruit or capsules but it is two or three more weeks, usually at the end of the wet season, before the fruits open and the seeds are dropped onto the newly-emerged gravel or silt along the riverbed which has been swept clear of the vegetation unable to withstand high river flows. The diminishing river flow strands seeds on these banks and the seeds are therefore given the best possible start (Cos-salter 1980).
Eucalypt seeds are dispersed predominantly by wind although for E. deglupta dispersal by water is probably also important. In fact, the distance a eucalypt seed falls from its parent tree can be calculated from the equation D = VwH/Vt where 'D' is the distance a wind8 of velocity 'Vw' will blow a seed falling with a terminal velocity 'Vt' from a height 'H'. Of 15 species of Eucalyptus used in experiments in Australia, E. deglupta had the lightest seeds (4,210 seeds/g), the lowest terminal velocity (2.1 m/s), and the greatest dispersal distance: (52.9 m) if released from 40 m with a wind of 10 km/h. At the other extreme, E. globolus had only 71 seeds/g, which travelled at a terminal velocity of 5.54 m/s and dispersed only 20 m under the same conditions (Cremer 1977).
Fauna
Animals of the riverine forest would include certain frogs9 such as Occidozyga spp, near small rocky rivers, and Rana arathooni and R. leytensis on the banks of somewhat broader rivers. The last two species can include both fish and large crickets in their diets (J. Dring pers. comm.). No birds are restricted to riverine forest although a number of species of kingfishers might be seen along the riverbanks more commonly then in the forest away from the river. The shy endemic rail Aramidopsis plateni is found in disturbed riverine forest near small rivers (Coomans de Ruiter 1946).
FOREST ON ULTRABASIC SOILS
Soils
Ultramafic rocks are dense, igneous in origin and are composed of magnesium and ferric minerals (hence the term 'ma' 'fic'). Ultrabasic rocks are those that contain less than 45% silica. Most ultramafic rocks are ultrabasic, and both are rich in iron, magnesium calcium, aluminum and heavy metals (Parry n.d.). In the following discussion only the term ultra-basic is used.
The soils that develop on ultrabasic rocks are notoriously infertile due to combinations of the following factors: high levels of exchangeable magnesium and a skewed calcium: magnesium ratio, a deficiency of calcium, nitrogen, phosphorus, potassium, molybdenum and zinc, and toxic concentrations of heavy metals such as nickel, cobalt and chromium. The infertility of these soils is well known to indigenous people and their distribution in many areas is more or less delineated by the boundary between cultivated and uncultivated soils (Parry n.d.). Examples of this are the area south of the Loa River near Kolonodale, and parts of Morowali National Park on the other side of Tomori Bay (D. Holmes pers. comm.). It is also of interest that the five-armed island of Padamarang, off the west coast of Southeast Sulawesi, comprises ultrabasic rocks and has little water, and as a consequence of both, the island has virtually no inhabitants. The low agricultural potential of ultrabasic soils has resulted in few studies being made of the soils, a singular disadvantage when plans for transmigration and other land development are considered.
Although patches of ultrabasic rocks occur in Sumatra, Borneo, Halma-hera, Timor, Sumba and Irian Jaya, the most extensive, both in Indonesia and in the world, are in the east of Central and South Sulawesi and in Southeast Sulawesi where they cover some 8,000 km2 (fig. 6.8). These rocks are largely composed of peridotite and serpentinite minerals. The residual soils are highly weathered, well-drained, friable when moist, and of a reddish colour. The concentrations of metals are generally high, exhibiting a considerable range, but the amounts available to plants are relatively small and appear to bear no significant relation to the total content (table 6.1). Alluvial and colluvial soils10 also have high concentrations of heavy metals and magnesium although these are lower than for residual soils (table 6.2).
Little is known about the action of chromium in the soil except that it is a non-essential element for plants and relatively inert. Some authors regard it as being severely toxic, particularly in the form of chromates, but others regard it as only moderately toxic. Concentrations of nickel in ultrabasic soils can be ten or more times greater than in other soils. As indicated above, however, the amount of nickel available to a plant will be much less, but even as little as 3 µg/g of soil can be toxic to certain plants. Other plants, however, are able to tolerate much higher concentrations and to concentrate the metal in their tissues such that 5%-25% of their mineral (ash) content is nickel (Lee et al. 1977). Samples of the shrub Psychotria dourrei (Rubi.) from New Caledonia have revealed the highest known concentration of nickel in plants—2.2% by weight of its dry leaves (Baker et al. 1985). The ingestion of such plants by humans can lead to heavy metal poisoning. These accumulator plants represent one end of a spectrum of responses to high metal concentrations, the other extreme being plants that actively exclude the metal. Some other plants may act as indicators; that is, there is a correlation between metal levels in the soil and concentrations in the plants. Some other plants, however, may act as an accumulator, indicator and excluder across a range of soil metal concentrations. Accumulators tend to be those species that are confined to soils with high metal concentrations, and excluders tend to be those with both metal tolerant and metal-intolerant traces (Baker 1981).
Figure 6.8. Areas of ultramafic (ultrabasic) soils (in black) on Sulawesi.
After Parry n.d.
Vegetation
Nowhere in the tropics with the exception of New Caledonia (Jaffre 1976) has forest on ultrabasic soils been examined in detail; a consequence, in part, of its low agricultural potential and the virtual absence of marketable timber. It would be expected that a fairly high proportion of the plants growing on these soils would be confined to them, having become adapted to the exceptional soil conditions. These would be of two types: those that have evolved on the islands of ultrabasic soil, and those that were once widely distributed but are now found only on ultrabasic soils. Some of the plants confined to ultrabasic soils may actually require high concentrations of magnesium, nickel or even chromium (Proctor and Woodell 1972). In a 500 m x 10 m transect through an ultrabasic area on the western shore of Lake Ranu in Morowali National Park the trees of 15 cm diameter or more at breast height had a basal area equivalent to 31 m2/ha which, with 348 trees/ha, shows an average basal area of 0.09 m2/tree which is only slightly less than trees in other lowland forests (p. 350). A histogram of tree heights in the 0.5 ha transect shows how few understorey (<10 m) trees were present and that more than half the trees were between 10 m and 20 m in height (fig. 6.9) (L. Clayton pers. comm.). A 2 m x 30 m transect in this forest was enumerated and its profile shows that the trees are relatively short and scrubby compared with trees on, for example, alluvium (fig. 6.10). Most of the species were either Ficus or members of the Sapotaceae. Given the high concentrations of metals and low inherent fertility, it is not surprising that the vegetation on ultrabasic soils tends to be relatively low and scrubby, and as such it is easily distinguishable on satellite images or air photographs.
From Parry n. d.
The ultrabasic forest around Soroako has a great deal of local iron wood Metrosideros (Myrt.), some Agathis (Arau.), much Calophyllum (Gutt.), various Burseraceae and Sapotaceae and at least two dipterocarps (Vatica and Hopea celebica). This forest has a relatively regular, low canopy and the only emergents are the Agathis but most of these have now been felled. The dominant family reaching the upper canopy is the Myrtaceae such as Eugenia, Kjellbergiodendron and Metrosideros. The nutmegs (Myri.) such as Horsfieldia, Gymnacranthera and Knema (including the endemic K. celebica with forked [emarginate] tips to its leaves) are common below the tallest trees (de Wilde 1981; Meijer 1984; van Balgooy and Tantra 1986). A conspicuous tree is Deplanchea bancana (Bign.) which has dense heads of yellow flowers (which attract large numbers of sunbirds), and long, oblong pods of winged seeds (van Balgooy and Tantra 1986).
From Parry n.d.
Figure 6.9. Histogram of tree heights along a 10 m x 500 m transect of forest on ultrabasic soil.
After L. Clayton pers. comm.
At 700 m on Mt. Konde, west of Soroako, the forest on ultrabasic soils has a regular closed canopy reaching 30-35 m above the ground. Few trees have trunks more than 60 cm in diameter and few have buttresses. The tree crowns are compact and the leaves are generally dark and leathery. No single family predominates and the most common large trees are Eugenia (Myrt.), Ficus (Mora.), Kjellbergiodendron (Myrt.), Lithocarpus (Faga.) and Santiria (Burs.). A 0.2 ha plot contained 234 trees of 10 cm diameter or more at breast height belonging to 36 genera and 27 families. It is not possible to calculate an index of diversity to compare with other sites because some of trees have not yet been identified to species.
Around the south shore of Lake Matano, to the east of Soroako, the ultrabasic vegetation comes right to the water's edge. It is dominated by just one family, the Myrtaceae, and most of the forest trees are less than 15 m tall. The main species are Metrosideros petiolata (fig. 6.11), with cream-coloured flowers and opposite leaves, and Xanthostemon confertijlorum, with bright red flowers and alternate leaves. Fruit measuring 8 cm x 15 cm found floating in the water belong to a lakeside Kjellbergiodendron with a gnarled trunk and spreading branches, which is probably not the same as K. celebicum found in inland forests. The spongy flesh around the single seed is apparently eaten by bats. The tallest trees in the community are the relative of the she-oak Gymnostoma sumatrana (Casu.), Planchonella (Sapo.), and the endemic Terminalia supitiana (Comb.). The most conspicuous among the smaller is an Elaeocarpus (Elae.) with yellowish-green fruit and white flowers. Epiphytic on the trees are Drynaria and Lecanopteris ferns and Hydnophytum (fig. 6.12), and the similar Mynnecodia (Rubi.), all of which are served by ants bringing organic frass into the enlarged, chambered stem (van Balgooy and Tantra 1986). Such ant-plant interactions appear to be relatively common in low-productivity vegetation. This is possibly because the organic frass dropping into or around epiphytes from the tree canopies is inadequate or of too low a quality for most epiphytes to grow. In such locations plants in which ants can deposit large quantities of insect (mainly ant) parts will be relatively successful (Janzen 1979). Although not yet investigated for Sulawesi, there is generally a preponderance of plants on infertile soils such as ultrabasics that have seeds adapted for dispersal by ants; this may be due to ants being particularly common in such forests or to some additional factor.
Figure 6.10. Profile of 2 m x 30 m of forest on ultra-basic soils in Morowali National Park. Horizontal and vertical scales are the same.
After L. Clayton pers. comm.
Figure 6.11. Metrosideros petiolata. Scale bar indicates 1 cm.
After Soewanda n.d.a
Analyses of leaf protein, fibre, ash, carbohydrate and major minerals from four species of trees growing nearby, but on ultrabasic soils, were compared with results from analyses of leaves from four different tree species growing on alluvial soils in Morowali National Park. No significant differences were found, but when analyses were made comparing the composition of leaves from four species found on both ultrabasic and alluvial soils it was found that protein concentrations were consistently lower in leaves from the ultrabasic sites (medians of 6.00% and 8.25%), and fibre concentrations were generally higher (medians of 23.0% and 19.7%) (L. Clayton pers. comm.).
Figure 6.12. Hydnophytum with the enlarged stem cut away to show the chambers inside.
One area in which an ecological understanding of ultrabasic vegetation and its succession processes could be of significant benefit is in the regreening of industrial sites on ultrabasic soils. The major examples of this are the P.T. INCO mining and hydropower sites near Soroako south of Lake Matano where revegetation of the bare areas caused by these operations has been generally slow. Many of the plants used in regreening programs have reflected a certain degree of tunnel vision among the proponents: plants such as Hibiscus and pines, neither of which is found naturally on ultrabasic soils, are unlikely to be much value.
In flat areas where the top soil is cleared the tree small Alphitonia (Rham.) appears to be one of the best colonizers, but on steeper areas revegetation is very slow although eventually the stiff grass Miscanthus sinensis (Gram.) forms a dense cover, sometimes with the yellow-flowered Scaevola oppositifolia (Good.) (fig. 6.13). Also there are fast growing trees on exposed soils in the mining area such as Alphitonia, Macaranga gigantea (Euph.), Homalanthus (Euph.), Callicarpa (Verb.) and Trema amboinensis (Urti.), and these pioneers could protect young trees of more mature forest growing beneath them. Baeckia frutescens (Myrt.) with Gleichenia ferns together form an effective shrub cover. Many garden plants do not thrive on ultrabasic soils without expensive dressing, but suitable plants can be found in forest on ultrabasic soils that have been or are soon to be logged. One such plant is the endemic gardenia Gardenia celebica (Rubi.) (Meijer 1984 pers. comm.).
Fauna
There is very little known about animals on ultrabasic soils in Sulawesi or anywhere else in the world (Proctor and Woodell 1968), but the little information which does exist indicates some certain peculiarities. For example, a butterfly and its larvae are restricted to ultrabasic soils near San Francisco Bay, California. The plant used as food by the caterpillar is not restricted to such soils, however, nor has any chemical difference between the plants in the butterfly area and those outside it been detected. In the laboratory the larvae will eat plants from either soil type with no ill effects. The butterfly distribution may, therefore, be a result of microclimate differences caused by different vegetation structure (Johnson et al. 1968).
Differences in vegetation have also led to the occurrence of a large population of a gopher11 in another area of California. Its principal food comprises the corms of a widespread plant that grows abundantly on ultrabasic soils, possibly because of reduced competition for light. The corms in question have three to four times as much magnesium as calcium in their tissue but how the animals are adapted to this is not known (Proctor and Whitten 1971).
The reports of experienced field workers who have visited ultrabasic areas in Sulawesi agree that the density of vertebrate animals is, in general, very low. Whether this is due to low productivity, high levels of defence compounds in the leaves, low concentrations of major nutrients in leaves (Jaffre 1976), or toxic concentrations of metals in the plant parts is not yet known (p. 458). An understanding of the mechanisms are important if appropriate conservation policies are to be made for Sulawesi.
A brief survey of birds in different areas around Soroako found that the least number of species was found in vegetation on ultrabasic soils although to some extent this may have been due to difficulties of observation. Five species, there of which were endemic to Sulawesi, were found only in ultrabasic areas. It also appeared that some species of flycatchers and starlings had a preference for ultrabasic areas but, again, this may be an artefact of the brevity of the survey (Holmes and Wood 1980).
Figure 6.13. Scaevola oppositifolia. Scale bar indicates 1 cm.
After Leenhouts 1957
Five pitfall traps baited with shrimp paste were set 10 m apart in both ultrabasic and alluvial forests in Morowali National Park to compare the invertebrate fauna. Each trap was left for 20 hours before collecting the animals caught in the water at the bottom of the trap. One of the major differences was that many more small fruit flies were caught in the alluvial area than in the ultrabasic area, possibly due to a greater abundance of fruit in the more productive alluvial forest. Ants were visibly more conspicuous in the ultrabasic forest, reflected somewhat in the pitfall trapping, and this may be explained by the large number of plants possessing mutu-alistic associations with ants. The greater abundance of ants was also demonstrated by comparing the times it took ants to find prices of shrimp paste in ultrabasic and alluvial forest. In the former it took just 2.9 minutes for the paste to be found whereas in the latter forest it took over 8 minutes (L. Clayton pers. comm.).
FOREST ON LIMESTONE
Physical Conditions
The limestone areas of Sulawesi (fig. 6.14) comprise rocks of various origins. There are Quaternary-Tertiary reef limestones found over much of Banggai, Togian, Muna, Buton and the coast north of Palu, Tertiary limestones formed in Eocene to mid-Miocene times in southwest and northeast Sulawesi, and Cretaceous limestone found near Lake Matano and on Buton (Sukamto 1975).
Most but not all limestone areas are described by geomorphologists as karst landscapes that have arisen from the abnormally high solubility of the bedrock (Jennings 1971). Calcium carbonate, or calcite (the major constituent of limestone), is not particularly soluble in pure water but is more soluble in weak acid. Carbonic acid arising from the solution in rainwater of carbon dioxide from the air and, to a lesser extent, humic acids arising from the soil as a result of vegetation decay, are thus effective in dissolving the rock. The process is rather complex with reversible reactions and ionic dissociations but in its simplest form can be summarized thus:
The karst landscapes of Sulawesi are of two major forms typical of the humid tropics: conical hill karst, such as found in the north of Bone, on Buton and Muna (fig. 6.15) and tower karst such as the famous hills around Maros and Tonasa (fig. 6.16). As in many categorizations of the natural world, intermediates are found, but in general the two forms are distinguishable.
The hills east of Maros and Tonasa (fig. 6.17) comprise about 300 km2 of Eocene and Miocene coral limestone. These overly the limestones of former backreefs and lagoons (fig. 6.18). The limestones were uplifted and exposed briefly to erosion in the early Miocene before being covered by the products of volcanic eruptions from a later period in the Miocene. Since then the volcanic rocks have eroded away and the exposed limestones have been dissolved by intense rainfall (up to 400 mm per day) and dissected by rivers (Balazs 1973).
The Maros hills are mostly between 150 m and 300 m high but the tallest reaches 575 m. The area receives an annual rainfall of about 3,500 mm which causes a great deal of surface runoff. The morphology of the hills was probably caused by rivers which arise in the east and flow to the west through caves which the rivers deepened and widened. The roofs of the caves eventually collapsed forming steep-sided valleys. The steepness of the valleys is emphasized further by undercutting caused by the erosive action of meandering rivers, some distance beneath the surface of the plain, which tend to flow around the hill bases rather than across the alluvial plain. The undercutting forms rock shelters some 2 m to 3 m high, tens or hundreds of metres in length, and usually 1 m or 2 m deep although some passages extend far into the hills (Jennings 1976; MacDonald 1976).
Figure 6.14. Limestone areas (in black) of Sulawesi.
Figure 6.15. Diagrammatic representation of conical hill karst in north Bone between the Maryosi and Macao Rivers.
After Sunartadirdja and Lehmann 1960
Whereas tower karst may have arisen from conical hill karst, not all conical hills necessarily progress to towers. The processes by which conical hills are formed are less clear, but they are possibly formed by streams flowing between blocks of limestone. The process is quite rapid; even the coral reefs raised during the Quaternary to form much of Muna Island have conical hills (fig. 6.19) and various factors have combined such that these hills are better developed than those in other limestone areas. In adjacent areas of raised coral lagoons, sink holes or dolines up to 150 m across can be found. Dolines also occur in the Maros tower karst, but these are not as wide nor as frequent (Balazs 1973; Anon. 1986). These are more typical of karst in temperate regions and look like pock marks on an otherwise relatively flat surface. In some areas of the north Bone conical-hill karst, such as the Macao River, steep canyons have formed where a previous subterranean river has dissolved away the roof of its course until it collapsed (Verstappen 1957).
Figure 6.16. Diagrammatic representation of lower karst near Maros and Tonasa showing the position of the Bantimurung waterfall.
After Sunartadirdja and Lehmann 1960
Figure 6.17. Geological situation of the Maros/Tonasa hills.
After MacDonald 1976
Figure 6.18. Cross-section through the hills east of Maros.
After MacDonald 1976
Soils
Soils on limestone, not surprisingly, are often richer in bases particularly calcium and magnesium, with a higher cation exchange capacity than soils in similar situations on different parent materials. On moderate slopes and hollows, clay-rich, leached brownish-red latosols are formed (Burnham 1984). On crests and shelves an acid, humus-rich, peat-like soil can sometimes develop on top of the limestone. On steep slopes and craggy hill tops the soil is very shallow. The composition and depth of soil can be variable depending on the purity of the parent material and the topography (Crowther 1982, 1984). The soils of the conical hills of Kambara on Muna Island were rendzinas,12 5-25 cm deep, well-drained and with a pH of 7-7.5 (Anon. 1982). Those examined by an EoS team at Hanga-Hanga waterfall, Luwuk, had high concentrations of calcium and magnesium (82 and 44 meq/100 g soil respectively).
Hydrology
Rain penetrating any soil over permeable rock passes downwards through an unsaturated (or vadose) zone, where rock pores are only temporarily filled with water, into the saturated (or phreatic) zone. The upper surface of the phreatic zone is called the watertable. The watertable is more or less parallel to the land surface and the groundwater flows according to the slope of the watertable. Springs are encountered where the watertable intersects the surface. There is an intermediate zone in which the watertable rises and falls. Thus the same opening that act as spring in the rainy season act as drainage openings in the dry season (Jennings 1971).
The actual nature of water circulation in karst areas is not fully understood and hypotheses vie for acceptance. An understanding is of great importance to the development of karst regions because very little water is found near the surface due to the extreme porosity of the rock. For people living in karst areas the supply of water can be critical and the famines in the Mt. Kidul area of Yogyakarta in the early 1970s were a direct result of bad land management in a critical karst area. Also, water issuing from karst springs is extremely pure and less variable in flow compared with rivers outside karst areas. The phreatic zone thus represents a significant reservoir of groundwater recharge for domestic, agricultural or industrial supplies although significant groundwater recharge may only occur after heavy rainfall (Crowther 1984).
Figure 6.19. View east over the west coast of Muna Island to show conical hill karst.
After Anon. 1982
It is extremely difficult to determine the exact drainage area of an underground river because topographical and geological catchments are often not identical. Even so, the mean annual runoff from the 300 km2 Maros karst has been estimated as 35-40 1/s/km2, which removes in solution about 80 m3/km2/yr of limestone (Balazs 1973).
Despite being relatively stable in its flow, the discharge of karst springs in the Maros area varies according to seasonal patterns of rainfall. Some cave rivers often flow during the rainy season and are dry for the rest of the year, whereas others with a larger catchment are permanent. In several areas around the Maros karst, villagers have dammed cave springs or rivers to create reservoirs so that outflow can be controlled and availability increased. Such impounded water is used mainly for domestic purposes. Most of the irrigation water in the Maros region originates from the karst region and, loaded with calcium and magnesium cations it may serve the additional role of raising the pH of acid soils.
The seasonal variation of karst water chemistry around Maros is unknown but even at low flow periods there are evident, albeit slight, daily fluctuations in pH, alkalinity and total hardness in the spring water emerging from Baharuddin cave. Organic debris was found trapped 3-4 m above the surface of the river in Salukan Kalang Cave during a dry period indicating that floods sometimes occur that fill virtually the entire cross-sectional area of some active cave passages. The large volume of water during the wet season and the shorter time water would spend in the cave would presumably dilute the concentrations of bicarbonate, magnesium and sodium.
Vegetation
The appearance of the vegetation around the hills of Bantimurung was described by Wallace as follows:
Such gorges, chasms, and precipices as here abound, I have nowhere seen in the Indonesian Ar chipelago. A sloping surface is scarcely anywhere to be found, huge walls and rugged masses of rock terminating all the mountains and inclosing the valleys. In many parts there are vertical or even overhanging precipices five or six hundred feet high, yet completely clothed with a tapestry of vegetation. Ferns, Pandanaceae, shrubs, creepers, and even forest trees, are mingled in an evergreen network, through the interstices of which appears the white limestone rock or the dark holes and chasms with which it abounds. These precipices are enabled to sustain such an amount of vegetation by their peculiar structure. Their surfaces are very irregular, broken into holes and fissures, with ledges overhanging the mouth of gloomy caverns; but from each projecting part have descended stalactites, often forming wild gothic tracery over the caves and receding hollows, and affording an admirable support to the roots of the shrubs, trees, and creepers, which luxuriate in the warm pure atmosphere and the gentle moisture which constantly exudes from the rocks. In places where the precipice offers smooth surfaces of solid rock, it remains quite bare, or only stained with lichens and with clumps of ferns that grow on the small ledges and in the minutest crevices.
Compared with forests on deeper soils, forests on limestone generally have few trees and tree species (Crowther 1982; Proctor et al. 1983a, b), although the total number of plant species present is probably not dissimilar. The basal area of trees on moderate slopes (<45%) can be quite similar to other forests and only on steeper slopes and rocky hilltops do soil conditions seriously affect tree growth. The shallow limestone soils may be able to support relatively high basal areas of trees because of the relatively fertile condition of the soils. A representative patch of forest on steep limestone 30 m x 2 m was enumerated along a gorge cut by the Tomasa River near Kuku, Poso, to produce a forest profile (fig. 6.20). It can be seen that in the area sampled there were few large trees, but this may at least in part be due to the steepness of the gorge (>100% or >45°). Only five of the 13 trees were buttressed (L. Clayton pers. comm.).
The relative paucity of tree species probably arises because some trees of lowland forest cannot tolerate the high calcium levels in the soil. The different tolerance of trees to calcium and the unique physical habitat makes the composition of these forests rather different from other lowland forests, giving rise to a specific community of trees on limestone. On steep limestone cliffs with bare rock faces, clefts and shelves, a distinctive herbaceous flora occurs. Many herbaceous plants are known to be endemic to the limestone hills of Peninsular Malaysia: for example, of 1,216 plants recorded as growing on them, 254 (21%) are found only on such hills and over half of these are endemic to the Peninsula (Chin 1977, 1979, 1983a, b). A species of grass Cymbopogon minutiflorus appears to be endemic to limestone areas of Central Sulawesi (Dransfield 1980), and many others may be found when intensive collecting is undertaken. It has been predicted that rare plants occur on the Maros hills although these are much younger geologically than the limestone hills of Peninsular Malaysia (van Steenis 1933).
Most of the information regarding limestone flora on Sulawesi concerns parts of Maros and Rantepao/Makale (Toraja) largely due to visits paid by an EoS team led by Dr. S. C. Chin of Universiti Malaya. Brief surveys showed that virtually all of the accessible parts of the Maros and Toraja limestone hills have been exploited to some degree, but accessibility has to be measured by the standards of the ingenious villagers who manage to reach rugged outcrops and steep gullies using ropes and bamboos in their search for firewood and other forest products such as rattan and fruits of Pangium edule (Flac.) (p. 412) and candlenut13 Aleurites moluccana (Euph.) many of which are planted (N. Wirawan pers. comm.). In addition, sap is collected from the sugar palm Arenga.14 Leaves of Garcinia are added to the fresh sap apparently to delay fermentation. When the sap is boiled down, these leaves are removed and candlenut seeds and other hard objects are added to stop the sugary brew from boiling over (Burkill 1966; S.C. Chin pers. comm.). From the evidence available it is assumed that all the hills within several kilometres of the surrounding villages have experienced some significant degree of disturbance.
The upper parts of the limestone hills of Maros are clothed in small trees growing 7-10 m tall out of cracks and small pockets of soil. The scrub vegetation likewise depends on the availability of these cracks and soil pockets. On the steepest cliffs the most common trees clinging to the rocks are figs Ficus (Mora.) the roots of which clamber over the rocks, in and out of crevices, for as far as 70 m. Such an extensive root network allows these plants to exploit wide areas for water and nutrients.
Figure 6.20. Profile diagram of 30 m x 2 m of forest on limestone from Kuku, Poso. Only trees of 15 cm diameter at breast height are shown. Horizontal and vertical scales are the same.
After L. Clayton pers. comm.
In undisturbed valleys, trees grow to have a diameter at breast height of up to 50 cm and common trees are Pangium edule (Flac.), Artocarpus (Mora.) and figs Ficus.
The composition and development of the vegetation of the Maros hills is determined not only by the concentrations of calcium, relative abundance of soil, and the peculiar biogeographical position of Sulawesi as a whole, but also the relatively long dry season (p. 22). Many of the plants in the area have clear adaptations to this climate—some are deciduous, losing their leaves in the dry season and remaining more or less dormant until the rains begin; some have fleshy stems in which they store water, some have very stiff, coriaceous leaves, and some have their leaves reduced to spines, both of which also reduce water loss. Some plants, such as the yams Dioscorea, have large storage organs (rhizomes, roots and corms), and some have the ability to lose most of their water (except that within their protoplasm which resists drying out), and then to revive on rewetting, a phenomenon known as 'poikilohydry' (Whitmore 1984).
When a severe dry season occurs the tolerance of these plants may be exceeded and they may be forced to survive as annual plants; that is, they must complete their life cycle during the wet season so that their seeds can germinate when the following drought ends. Among those plants that appear to be annual in at least parts of Maros are the single-leafed Monophylla (Gesn.) and the maidenhair fern Adiantum. Fleshy plants there include a cactus-like Euphorbia (Euph.), the red-and-green Kalanchoe pinnata (Cras.)and the dragon tree Dracaena (Lili.).15 Kalanchoe is an example of a plant that avoids severe water stress during droughts by closing its stomata during the day. This prevents normal photosynthesis from occurring and instead a pathway called CAM (Crassulacean Acid Metabolism) is adopted in which important photosynthetic processes can be carried out at night. In the relatively high limestone areas of Tana Toraja, such adaptations are also seen, especially on exposed rocky hill tops. Although the rainfall here is greater and more evenly distributed, and the humidity higher due to the clouds that frequently swathe the hills, there is also a greater intensity of sunlight and exposure to wind. The presence of plants with thick fleshy leaves such as Kalanchoe suggests that there is periodic water stress (Chin 1986).
The forest on limestone at 1,000 m on Mt. Wawonseru, west of Soroako, has an uneven canopy reaching a maximum of 40-45 m. This only occurs where the soil is relatively deep, however, and in places where the rocks are covered with thinner soil, only shrubs are found. No single family predominates although Lauraceae and Annonaceae are common. Interestingly, neither of these families is well represented on nearby Mt. Konde which is on ultrabasic soils, and none of the trees from Mt. Konde is important on this particular limestone hill. Among the largest trees on Mt. Wawonseru are Bischofia (Euph.), Eugenia (Myrt.), Podocarpus (Podo.) and Vernonia (Comp.). In the lower canopy there are many Polyalthia (Anno.) and Antidesma (Euph.) and below this the undergrowth is rather sparse. Orchids, wild pepper Piper (Pipe.) and a Rynchoglossum (Gesn.) are able to live on the bare rock (van Balgooy and Tantra 1986).
The forest on limestone near the tip of the northeastern peninsula visited by another EoS team differed structurally from many lowland forests on unexceptional soils by having no palms in a 0.5 ha transect (there are between 80-100/ha at Toraut). The forest had a generally small stature having a basal area of only 19.8 m2/ha and a basal area per tree of 0.056 m2 (both values about half those obtained at Toraut) (p. 350).
The plants growing on limestone along the shores of Lake Matano are of mixed composition and also do not show dominance by any particular family among the trees, shrubs or herbs (de Vogel 1986).
An EoS team that visited Luwuk attempted to determine any differences there might be between insect damage on plants of different succession stages and on two different soil types. Insect damage was examined in the vegetation on steep limestone soils near Luwuk and in vegetation on richer soils near the edge of Bogani Nani Wartabone National Park (p. 373). According to prevailing theory, plants will make greater investments in leaf defences where the environmental conditions are relatively harsh. The steep, thin soils of the limestone area may be said to represent a harsher environment than the flat, alluvial soils at Bogani Nani Wartabone. The percentage total amount of leaf area removed by insects did indeed appear to be lower, and leaf toughness greater, in the limestone area than in Bogani Nani Wartabone, although concentrations of nitrogen and water showed no consistent pattern (S. Greenwood pers. comm.).
Effects of Disturbance
Limestone vegetation is destroyed in the process of limestone quarrying for the Tonasa cement factories (p. 563), but even the enormous quantity of limestone quarried (the factory supplies almost all the cement needs of eastern Indonesia), cannot be considered a threat to limestone-related ecosystems in Sulawesi because the area affected is very small compared with the total area of limestone. It should be stressed, however, that new quarries and extensions to the existing ones should be sited such that the environmental impacts are minimal. The secondary vegetation in these limestone areas is dominated by the introduced Eupatorium (Comp.) and Lantana (Verb.) the dense thickets of which probably hold back the succession process by some years. Trees found in the succession include Homalunthus (Euph.), Lagerstroemia (Lyth.), Pterospermum (Ster.), Kleinhovia (Ster.) and Villebrunea (Urti.). Given their aggressiveness in competing with Eupatorium and Lantana, some of these trees could reasonably be used in programs for the reclamation of weed-dominated lands on limestone soils.
Fauna
No vertebrates are restricted in their distribution to limestone hills but certain species of snails, which use calcium to form their shells, have been found to be restricted to very small areas of limestone elsewhere in Southeast Asia. From 108 species found on 28 limestone hills of different degrees of isolation in Peninsular Malaysia, 70 were known from single hills, and one hill had no less than seven species known only from its slopes. At the other extreme, however, one snail species from a genus that included some very restricted species, was found on 19 of the hills (Tweedie 1961). There are clearly major differences between species in their environmental tolerances.
The snails of limestone hills on Sulawesi have not been studied. All one needs to start such a study is a bucket of water. Leaf litter from the base of a limestone hill is stirred around in the bucket and after waiting for a few minutes it will be seen that the soil and much of the vegetation will sink while small snail shells, which usually have a bubble of air trapped inside them, float on the surface where they can be collected. Identification to species is a problem better left to a specialist and representative samples of unnamed but recognizably different species can be sent to a major museum.
One of the more spectacular butterflies known largely from Sulawesi limestone, presumably because its food plant is restricted to a limestone habitat, is the large swallowtail Graphium androcles16 (fig. 6.21) made famous in the following description of the butterflies of the Bantimurung waterfall by Alfred Wallace:
As this beautiful creature flies, the long white tails flicker like streamers, and when settled on the beach it carries them raised upwards, as if to preserve them from injury. It is scarce, even here, as I did not see more than a dozen specimens in all and had to follow many of them up and down the river's bank repeatedly before I succeeded in their capture. When the sun shone hottest around noon, the moist beach of the pool below the upper fall presented a beautiful sight, being dotted with groups of gay butterflies—orange, yellow, white, blue and green—which on being disturbed rose into the air by hundreds, forming clouds of variegated colours.
Other collectors followed Wallace and 25 years later in 1882 G. androcles could no longer be found although thousands of other species remained (Guillemard 1889). This may have been an effect of the seasons (p. 368), however, because 45 years later the butterfly was found to be numerous again as were other swallowtails and less spectacular butterflies (Leefmans 1927), but now the butterfly is more or less extinct around the waterfall and many others have fallen prey to commercial collectors (Anon. 1985). A far-sighted gentlemen in Bantimurung has managed to raise butterfly larvae to adults but his enthusiasm has waned because so many of the adults are caught (illegally) and sold near the reserve such that population levels in surrounding areas have been decreasing (Soetjipto et al. 1982).
Figure 6.21. The upper surface of a fore and hind wing of Graphium androcles.
After Haugum et al. 1980
MONSOON FOREST
For the purposes of this book, monsoon climates are defined as those areas falling within the climatic zones 'E' to 'H' in the Schmidt and Ferguson classification (p. 22) that are characterised by a long dry season.
Vegetation
It is a matter of debate and definition whether primary lowland monsoon forest actually still exists in Sulawesi or indeed anywhere else in Indonesia (van Steenis 1957; Whitmore 1984). Its original extent has been much reduced and the primary reason for this has been fire. During the dry season, monsoon forest trees and other plants are easily burned and repeated burning eventually results in a persistent grassland vegetation. None of the grasslands or savannas of Sulawesi are natural but are instead the result of human activities. In Sulawesi, as in Java, "fire for hunting, for pleasure, for pestering neighbours or neighbouring villages, by carelessness, for clearing land, for making land passable, for converting forest into pasture land, in short for innumerable purposes, has played havoc with the monsoon forest" (van Steenis and Schippers-Lammertse 1965).
Only one area of monsoon forest appears to have been studied in any detail, and detail, and that is the much-disturbed forest of Paboya Reserve east of Palu (Sidiyasa and Tantra 1984). The Reserve was originally established to safeguard the sandalwood trees Santalum album (Sant.) (fig. 6.22). The oil in the heartwood of this tree is extremely valuable and wild stands of the tree have suffered considerably in dry areas of Indonesia because of uncontrolled exploitation. There are relatively few sandalwood trees remaining in Paboya, partly because of illegal felling, and partly because of fire. Fires are set by cattle farmers who wish to encourage the growth of young grass. The fires kill the sandalwood seedlings (as well as Acacia fernesiana trees planted in a regreening effort some years ago), although sandalwood will shoot from its base when the above-ground parts are burned. In addition to sandalwood, the trees of the Paboya forest include Duabanga moluccana (Sonn.), Ficus spp. (Mora.), Canaga odorata (Anno.), Harpulia sp. (Sapi.), Alstonia angustifolia (Apoc.), Elatoslachys verrucosa (Sapi.) and Buchanalia arborescens (Anac.). Where the forest has been disturbed, the most common trees are Casuarina sumatrana (Casu.), Pittosporum ferrugineum (Pitt.), Mallotus philippensis (Euph.) and the introduced Cassia siamea (Legu.) (Sidiyasa and Tantra 1984).
The forest comprises a simple, lightly-closed community of trees containing a large proportion of deciduous or leaf-shedding species. It contains relatively few species found in the everwet lowland forest and one of the few emergents in certain areas is Salmalia malabarica (Bomb.)17 (fig. 6.23). One tree of monsoon forest which is also found in slightly wetter areas, particularly those near the coast, is the large legume Pericopsis mooniana (fig. 6.24) which has a hard, dark-red timber greatly sought after for making furniture (p. 679). Its pods are flat, up to 10 cm long and 4 cm wide and contain only two to four seeds.
In some strongly seasonal areas, repeated burning has given rise to a savannah with relatively fire-resistant trees such as Morinda tinctoria (Rubi.) (also found in wetter areas), Acacia tomentosa (Legu.) Phyllanthus emblica (Euph.) (fig. 6.25), tamarind Tamarindus indicus (Legu.),18 teak Tectona grandis (Verb.),19 paperbark Melaleuca (Myrt.), Timonius sericeus (Rubi.), Garuga floribunda (Burs.) (fig. 6.26), and the two large fan palms Borassus flabellifer and Corypha elata (Steup 1936, 1939; Metzner 1981; van Balgooy and Tantra 1986). The two large palms are easy to confuse but Borassus has blue-green or greyish leaves, a smooth trunk with a diameter similar to that of a coconut palm and no large spines on the leaf stalks. Corypha is generally more massive, its leaves are nearly 2 m in diameter and twice as large as Borassus, and it has small fruits up to 5 cm in diameter compared with the massive 18 cm fruits of Borassus. Corypha can be identified from a distance by the lines formed by the persistent leaf bases spiralling up the trunk. Corypha elata has the second largest inflorescence of any flowering plant20 with the flowering branches projecting nearly 4.5 m above the top of the palm and setting hundreds of thousand of fruit. After fruiting, the tree dies in the same manner as the more familiar sago palm Metroxylon sagu.
Figure 6.22. Santalum album. Scale bar indicates 1 cm.
After Soewanda n.d.b
Figure 6.23. Salmalia malabarica. Scale bar indicates 1 cm.
After Soewanda n.d.
Figure 6.24. Pericopsis mooniana. Scale bar indicates 1 cm.
After Soewanda n.d.a
Figure 6.25. Phyllanthus emblica. Scale bar indicates 1 cm.
Alter Whitmore 1972
Figure 6.26. Garuga floribunda. Scale bar indicates 1 cm.
After Leenhouts 1956
The Palu valley is the only locality in Indonesia that has the 'H' rainfall type of the Schmidt and Ferguson classification (p. 22) and it is not known exactly what its original vegetation would have been. Even in its present, greatly disturbed, condition it has many characteristic plants indicative of water stress such as Acacia farnesiana (Legu.), the dragon tree Dracaena (Lili.), Capparis (Capp.) and introduced plants such as the succulent-leafed, red-and-green flowered shrub Kalanchoe pinnata (Cras.), brought from Africa a very long time ago (Backer 1951), the tamarind Tamarindus indica (Legu.), and the prickly pear cactus Opuntia nigricans (Cact.) brought from South America. This plant is also found at the southern end of the southwest peninsula.
When the Palu district was visited in 1902, large stands of Dracaena were conspicuous and only a few Opuntia nigricans were seen (Sarasin and Sarasin 1905), but in 1911 dense stands of the cactus were abundant, particularly in abandoned rice fields (Grubauer 1923). By the late 1920s O. nigricans had become the dominant plant in the eastern foothills of the Palu valley and bay (Steup 1929) and in the mid-1930s the plant had taken over the entire northern part of the valley (fig. 6.27). Then, in 1934, the cochineal scale insect or mealy bug Dactylopius tomentosus2 1 was introduced from Australia where it had been used against Opuntia22 (van der Goot 1940). By 1939 it had done its work and stands of the cactus were restricted to just a few localities and Leucaena glauca (Legu.) trees had taken over where the cactus had been killed (Bloembergen 1940). One would have expected the cactus and the scale insect to remain in small numbers, as is the case in Australia, but O. nigricans is now abundant once more, particularly around the new campus of Tadulako University, and it is possible that the scale insect has become extinct. If so, then this most unpleasant plant will increase until controlled, hopefully with a biological rather than a chemical agent. In parts of America Opuntia are exploited and domesticated for their fruit, and in dry periods the spines are burned off with flame sprayers which makes them acceptable as cattle food (C.E. Russell pers. comm.).
Figure 6.27. Distribution of rice fields and the cactus Opuntia nigricans in the Palu valley in 1935.
After Metzner 1981
The extensive grasslands of the Palu valley comprise species such as Cynodon dactylon (Gram.), but the grasses are mixed with a few small legumes such as the yellow-flowered Tephrosia. Sword grass Imperata cylindrica is rarely seen except in areas which have been disturbed relatively recently. Some of the shrubs have spiny branches or thorns such as the introduced Alternanthera (Amar.), but their predominance maybe an artefact caused by the presence of many goats and sheep which favour other elements of the vegetation (Meijer 1984). Another plant that seems not to be eaten by goats is the giant milkweed Calotropis gigantea (Ascl.) with distinctive large, white and wooly leaves, and this avoidance may be due to the copious white sap which has emetic properties (Burkill 1966).
Fauna
Nothing has been published specifically on the fauna of monsoon forests; it probably does not differ much from that found in other forest, except that the species would have to be able to withstand the long dry season, either by some behavioural or physiological adaptation of by moving away from the area.
One bird that favours drier areas is the Asian palm swift Cypsiurus batasiensis, which was first recorded from Sulawesi in 1978 (Escott and Holmes 1980). Its unusual nest is built in the centre fold of the upper surface of those leaves of large fan palms that are ageing, folded and hanging vertically. A delicate nest is built from feathers and feathery seeds (such as kapok Ceiba pentandra [Bomb.]) glued onto this vertical surface using saliva. The normal clutch is two eggs on which the adults sit in an almost vertical position. The diet of these birds can be investigated by examining the insect wings that are voided in the faeces dropped below the nest by the nestlings. Their food appears to comprise primarily of flying ants, termites and beetles caught by the adults about 10-15 m above the ground and which is brought to the nesting every 15-75 minutes through the day. The interval is shorter at the start than the middle of the day and this is related to the availability of suitable food (Hails and Turner 1984).
Certain species of tropical butterflies are represented by different adult forms in the wet and dry seasons, particularly in strongly seasonal areas. Common differences shown are that the wet season forms are active, more colourful and may have conspicuous 'eyespots' on their hindwings, whereas the dry season forms are very cryptic both in pattern and movement. It is assumed that these differences reflect changes in predation pressure (Brakefield and Larsen 1984; Janzen 1984). This has not yet been recorded from Sulawesi.
The Palu Valley: Past and Future
In the early 18th century the Palu valley was described as 'a blessed place' with fertile rice fields (Valentyn 1724), but two centuries later an agricultural engineer gave the opinion that "I also think that nowhere in the Archipelago has deforestation had such a fatal influence as in this place which otherwise, being richly endowed with water flowing down from the forest-covered mountains, could have been a little Egypt" (quoted Steup in 1929). By that time much of the former rice fields had been abandoned, partly because of problems of internal security and partly because the vast amounts of soil debris washed down from the deforested slopes had made the irrigation systems inoperable. Even in 1896 the south and southeastern slopes of the Palu valley were virtually devoid of trees up to 1,700 m.
Trials to determine appropriate means of reforesting the hills began over 60 years ago but the vast majority of the land to the east of Palu Bay and Palu River is still bare (Hadisumarno 1977). It is used as pasture for sheep and cattle but its carrying capacity is extremely low at about 5 hectares per head of cattle per year (Rauf 1982), at least in part because of the scarcity of small leguminous plants (Bismark et al. 1978). The general comments made about reforestation on the ultrabasic soils of Soroako apply here: the best species to choose are those that clearly excel in the natural or semi-natural setting. However, it should be noted that the area has been so degraded that it is necessary to find appropriate plants to enrich the soil before planting trees such as Leucaena farnesiana that have some social and agricultural uses.23 This is particularly important because resettlement schemes in the region have been devised numerous times over the last few decades to encourage farmers in the hills to abandon their inappropriate forms of land management but almost all have met with some degree of failure (Metzner 1981).
One course of action that might be followed is to aim to recreate a monsoon forest on the Palu hills. A similar scheme, to recreate tropical dry forest, has begun in Costa Rica where a sufficient level of knowledge of this forest type, its components and their interactions has been reached such that judicious forest and animal management can be used to extend the area of this very rare forest type (Janzen 1986). All is not lost for the monsoon forest of Sulawesi—the Paboya Reserve, so degraded and small, probably contains most of the plant species that would be needed in the initial stages of forest development and disperser animals could be encouraged to return or reintroduced if necessary. The sandalwood is by far the most important monsoon forest species and provision of seedlings could be assisted by tissue culture techniques developed for this species by Dr. L. Winata of Bogor Agricultural University.
If a plan to recreate the monsoon forest could be put into effect, it would probably pay for itself in the long run given the huge amount of soil currently lost from the hills, the low productivity of the land, the potential
harvesting of goods from the forest at all stages of its growth and the consequent improvement to the irrigation schemes in the lowlands. All that is required is start-up money and political will.