Chapter Seven
INTRODUCTION
The physical and biological differences between the hot, luxuriant lowlands of Sulawesi and its cold, stressed mountain tops are very striking and parallel some of the changes found when travelling from equatorial to temperate and arctic regions. As such, mountains are stimulating subjects for study and can be instructive in understanding the physical limitations to growth and reproduction in both plants and animals.
Most of Sulawesi lies above 500 m and about 20% of the total land area is above 1,000 m. The highest blocks of land, those over 2,000 m, lie in Central and northern South Sulawesi (fig. 7.1). With lowland forest all but disappearing due to the many pressures upon it, mountains are important refuges and sources of genetic diversity. An understanding of what does or can live at higher altitudes is important to regional planning.
CLIMATE
Temperature
To spend a night on a mountain top is to learn some fundamental physical principles. Mountains do not warm up or cool down in the same way as the lowlands primarily because there is a shorter, and therefore less dense, column of air above them. Infra-red (warming) radiation from the sun excites gas molecules in the air causing them to collide with each other and thus to give off heat. Where the air is less dense, fewer collisions occur and hence less heat is generated. There is less water vapour and particulate matter in mountain air and this results in there being less reflection and absorption than at sea level even though the actual intensity of radiation is greater.
Radiation of heat from the ground occurs both by day and night, but at night there is no compensating radiation from the sun. Heat is lost quickly from high altitudes at night causing the daily temperature range there to be as much as 15°C to 20°C (fig. 7.2). As the surfaces of plants, soil or rocks cool down at night so the air around them also cools. This, being heavier than the warmer air, will become progressively colder if there is no slope down which it can flow, for example in hollows and valleys. Frosts are most likely to occur in flat hollows, called frost pockets, on clear, calm nights when long-wave radiation is lost to the skies. The fastest rate of cooling occurs on non-conductive surfaces such as dead plant material and dry soil rather than conductive material such as living plants, wet soil, and water.
Figure 7.1. Land over 2,000 m (in black) and the major mountains.
Figure 7.2. Drop in temperature from the afternoon to early morning at sea level (Ujung Pandang) and at three altitudes on Mt. Rantemario.
Data from an EoS team
The rate of temperature change with altitude (the lapse rate) is generally accepted to be about 0.6 °C/100 m but this depends on factors such as cloud cover, time of day, and amount of water vapour in the air. Readings of minimum temperature from Mt. Rantemario, Sulawesi's highest mountain (3,450 m), taken during an EoS expedition demonstrate the decrease in minimum temperature with increasing altitude, and from these few readings the rate of decrease appears to be 0.7°C/100 m (fig. 7.3). It is quite likely that frost would occasionally form on the summit of Mt. Rantemario and a few of the other highest peaks, and during the cooler periods of the Quaternary permanent snow would probably have been present.
Relative Humidity
The percentage saturation of a mass of air increases as its temperature falls. The dew point (the temperature at which condensation occurs and clouds or drops of dew form) at different altitudes depends therefore on the temperature and the initial moisture content of the air. The forests of higher altitudes experience a very high relative humidity, particularly at night when the temperature falls, and the dew point is frequently passed so that water condenses on the leaves. Relative humidity has been recorded over 60 days at 1,700 m on Mt. Nokilalaki and results ranged from 86% to 100% (mean average 94%) at dawn, 68% to 98% (mean average 87%) in the early afternoon, and 86% to 100% (mean average 94%) after dark (Musser 1982). During dry spells at higher altitudes above the main cloud layer (above about 2,500 m), however, the relative humidity can reach 20% during the day, leading to wider daily extremes.
Figure 7.3. Minimum temperature readings on successive nights during the ascent of Mt. Rantemario.
Data from an EoS team
Clouds
Clouds originate where ascending air reaches its dew point and where there are the necessary dust or other particles for the water vapour to condense upon. Once water droplets have been formed they tend to act as 'seeds' and the droplets may grow as they bounce around in the clouds. During the wettest months, the slopes and peaks of high mountains can be enveloped in clouds for days on end. During the drier months, however, when the air is not saturated with water vapour, it is common for a belt of clouds to form around a mountain, often at about 2,000 m. As experienced mountain climbers know, a low, even continuous layer of cloud at 2,000 m does not always mean that clouds are present at 3,000 m or above.
Rainfall
Although rainfall is generally higher on the side of a mountain facing the prevailing wind (fig. 7.4), there appear to be no guiding principles relating altitude to rainfall except that rainfall on mountain slopes up to 1,500 m will generally be greater than on the surrounding lowlands. At the cloud level, however, rainfall measurements are not especially useful for ecological studies because much of the water used by plants is in the form of water droplets in the clouds which adhere to the surfaces they touch. Studies on Mt. Pangrango (3,025 m) in West Java, showed that the summit received more frequent rain than the slopes but this amounted to a lower total rainfall. There is, however, great variation between years: the summit of Mt. Pangrango has in one year received only 13 mm of rain between July and September but 460 mm during the same period in another year (van Steenis 1972). When the EoS team reached the summit of Mt. Rantemario in October 1985 a note written by climbers the previous July appeared almost pristine despite being open to the elements for the intervening months.
Figure 7.4. Rainfall on mountain slopes is generally greater in the lowlands and on the side facing the prevailing wind.
After van Steenis 1972
Ultra-violet Radiation
It has been suggested that ultra-violet radiation on tropical mountains is probably more intense than on any other mountainous region on earth (Lee and Lowry 1980). This is because the amount of ozone, which absorbs ultra-violet radiation in the stratosphere, is appreciably less near the equator than elsewhere, and the atmosphere at low altitudes is more turbid and dense and thus more capable of absorbing or reflecting this radiation. Climbers must therefore protect their skins or rapidly suffer the effects of sunburn. One possible effect of high levels of ultra-violet radiation is to cause cell mutation, so it is possible that the rate of cell mutation on these mountains is particularly high, and this may be one reason why the levels of endemism are relatively high on mountains.
SOILS
The nature of mountain soils changes with increasing altitude, becoming more acid and poorer in mineral nutrients, and this is particularly the case where acid peat is present. Peat accumulates in wetter places, in the cloud belt, or upper montane zone for example, where decomposition processes are generally slower. Ridges, knolls and summits only receive water from the atmosphere and so their soils are continually leached because soil and water in the interflow (p. 265) cannot be received from above. The soil in these places is therefore drier and more nutrient-poor than soil in valleys or the lower slopes (Burnham 1984). Differences in bedrock composition and climate are the major factors influencing soil formation up a mountain, although steepness of slope and openness of the vegetation cover are also important. Low temperatures slow down the processes of soil formation for various reasons. For example, reduced evapotranspiration causes less movement of water through the soil, chemical reactions occur 2-3 times slower with every 10°C drop in temperature, and the reduced abundance of soil organism's means that biological processes affecting soil formation are also slower. With increasing height the soil is less well formed, and the roots of the plants growing in it are shallower. The differences in the soil obviously have their effects on the vegetation but whether the soil is more important than direct climatic effects and the changes in communities of disperser animals is not known.
Soils were collected on EoS expeditions and results of the analyses (table 7.1) illustrate some of the principles described above.1 Mountain soils are generally deficient in calcium (<1.8 meq/100g) (Whitten et al. 1984) but this does not seem to be the case for most of the soils examined in Sulawesi. The high potassium (and relatively high cation exchange capacity) on the western (more open) slope of the summit of Mt. Rantemario may be a result of fires set by people. The soil collected from Mt. Kara was on lava that was a product of an eruption in 1983, and although the low concentrations of cations are what would be expected, the percentage of carbon seems surprisingly high.
From Burhanuddin 1986; and EoS teams
VEGETATION
Structure
The forest on mountain slopes below 1,000 m is very similar to other lowland forests although on small mountains changes with altitude are evident (fig. 7.5). With increasing altitude, however, the trees become rather shorter and less massive, and epiphytes such as orchids become more common. This is called lower montane forest. Further up, the tree canopy becomes more uniform, the trees are even shorter, squat and gnarled, the leaves are small and relatively thick, and moss abounds. This is called upper montane forest. Beyond this, if the mountain is tall enough, is the subalpine forest of yet smaller trees with smaller leaves, whose branches bear epiphytic lichens in profusion but virtually no orchids (fig. 7.6; table 7.2). Valleys and other depressions in the subalpine zone are generally devoid of trees, and it appears that no trees have adapted to high altitude, water-logged conditions. Here and elsewhere in this zone there is, instead a covering of shrubs, colourful herbs and tough grasses. The woolly hairs of some subalpine plants have been variously attributed to the ability to protect the plants against high temperatures (Lee and Lowry 1980), intense ultraviolet radiation (Mani 1980) and frost/freezing (Smith 1970). Maybe they play a role in all three. For plants with leaf rosettes on the ground such as the silverweed Potenlilla (Rosa.), however, the wooly hairs are concentrated on the lower surface and petioles only. This would make sense as frost protection but not against ultra-violet radiation or high temperatures. Other plants provide their growing buds and perennial parts with thermal insulation by retaining old leaves, and by having tufted branches and persistent scales. Such protection is necessary because the 'cost' to a subalpine plant of replacing a leaf is much greater than for plants growing in the more hospitable environments lower down the mountain or in the lowlands. Adaptations of leaves to intense ultraviolet radiation are thick cuticles, wax deposits, and high concentrations of red, protective leaf pigments called anthocyanins in young leaves (Lee and Lowry 1980).
Figure 7.5. Changes in relative frequency of dominant tree species with altitude on Mt. Tangkoko (1,109 m).
After Anon. 1980
Where the slopes in the sub-alpine zone are treeless, the cause is probably hunters having set fires in order to make hunting anoa easier (p. 494).
Figure 7.6. Cross-sections through five forest types to show the decrease in tree height and simplification in structure with an increase in altitude, a - lowland forest, b - lower montane forest (lower levels), c - lower montane forest (upper levels), d - upper montane forest, e - subalpine forest.
Adapted from Robbins and Wyatt-Smith 1964
Two transects (200 m x 10 m), one at 3,100 m in upper montane forest and the other at 2,100 m in lower montane forest, were enumerated on the slopes of Mt. Rantemario (table 7.3). Trees in the upper montane forest clearly have thicker trunks for their height compared with trees in the lower montane forest: average height in the former was 6.5 m whereas, for trees of the same trunk diameter range (15 cm to 20 cm), the average height in lower montane forest was 11.5 m or nearly twice as tall.
Trees in exposed, windy sites tend to have a greater wood density than species characteristic of more sheltered habitats. The denser wood would to some extent counteract the damaging effects of wind on exposed sites, but since investment in wood must be made at the expense of investment in leaves, shoots and trunk thickness, this may be a reason for the dwarf stature of upper montane forests. This shorter forest with dense wood probably grows slower than forests at lower altitudes, and together with its greater structural stability would result in the rate of gap creation being less than in the lowlands (p. 360). This must have significant effects on forest dynamics and composition (Lawton 1984). Morphological differences within a species of tree in the upper parts of lower montane forest have been shown to be related to a gradient of wind stress, which reaches a maximum on ridge crests. For individuals of a given height, trunk girth and branch thickness, increases with proximity to a ridge crest (fig. 7.8). It would seem, therefore, that gnarled trees may be a direct result of wind stress (Lawton 1982), but factors such as periodic drought and nutrient stress might be involved.
1 - the leaf-size classes refer to a classification of leaves devised by Raunkier (1934) and modified by Webb (1959). The definitions are: mesophyll: 4,500-18,225 mm2, notophyll - 2,025-4,500 mm2, microphyll - 225-2,025 mm2, nanophyll: less than 225 mm2. An approximate measure of leaf area is 2/3 (width x length). See figure 7.7.
After Grubb 1977; Whitmore 1984
Figure 7.7. Hypothetical leaves of 4,500 mm2, 2,025 mm2 and 225 mm2 to assist assessment of leaf-size class.
As can be seen from table 7.2, average leaf size changes with altitude, but so also does leaf thickness. The thickness of ten leaves of each of the four most common plants around the summit of Mt. Rantemario were measured and the median thickness was found to be 0.45 mm (range 0.30-0.85). The equivalent figures for a sample of leaves from lowland forest (Bogani Nani Wartabone National Park) were 0.2 mm (range 0.1-0.3). The thicker leaves, or pachyphylls, characteristic of montane habitats (Grubb 1977; Tanner and Kapos 1982), have a thick palisade layer (the long, vertical photosynthetic cells) and frequently a strengthened, protective hypodermis (the layer below the epidermis). The thick hypodermis may be an adaptation to resist undesirable penetration by fungi, bacteria, moss and lichens, some of which grow on leaves where they are called epiphylls (p. 358). The high humidity provides ideal conditions for the growth of these plants and they probably represent a greater threat to leaf life in the frequently cloud-swathed upper montane and subalpine forests than does insect damage, the major factor in the lowlands (Grubb 1977).
Data from an EoS team
Figure 7.8. Changes in height of trees in lower montane forest in relation to average wind velocity on slopes and exposed ridge crests.
After Lawton 1982
Lower down the mountain, the growth of epiphylls is not so luxuriant and different adaptations seem to have evolved to cope with a constant film of water on the leaf. Under such conditions, transpiration is very slow because water cannot evaporate from the openings or stomata on the leaf (Leigh 1975). In addition water on leaves can leach nutrients from leaf tissues and can also reflect sunlight reducing the rate of photosynthesis and hence growth. It would therefore be reasonable to expect plants to have adaptations to ensure that water drips off, rather than sits on, their leaves.
One common leaf feature noticed last century is the long, narrow leaf tip which, even within a species, is longer and narrower in wetter micro-habitats: leaves possessing long tips shed water faster than blunter leaves. Leaves would not need such 'drip-tips' if they could be held vertically but they obviously need to be orientated in such a way as to make the most of the little sunlight received in the understorey.2 The optimum position for photosynthesis would be horizontal but then the water would not drain away. Assuming the majority of light comes from directly overhead (the shortest distance to the 'outside'), 100% of the light is reflected if the wet blade of the leaf is tilted down to 40° from the horizontal yet only 3% is reflected when tilted at 30° from the horizontal. Thus orientation between 30° and 40° would be optimal. If the leaf tip is narrow and downward-pointing, however, the leaf can be held more horizontal and yet still lose its water (Lightbody 1985). Leaf orientations are not rigidly fixed and in fact can vary through the day, being steeper at night when precipitation is greatest and evaporation least, and more horizontal during the day to intercept sunlight except when rain falls when the leaves tilt further downwards (Dean and Smith 1978).
The high winds, the rocky, nutrient-poor soils, and relatively low rainfall (at least seasonally) cause the plants of summits, and other exposed locations, a certain amount of stress. One means by which a plant can cope with such a poor, drying environment is by developing a relatively extensive root system to support a relatively small above-ground component. The ratio of dry weight of below-ground parts to above-ground parts (often called the root: shoot ratio) within a given species of plant or among communities of plants tends to increase with altitude. In lower montane forest in Papua New Guinea, the ratio was about 0.1 (Edwards 1982a), meaning that the trunks, branches and leaves were together ten times as heavy as the roots. In harsh habitats, such as exposed locations in the subalpine zone, the ratio is generally considerably higher. This was examined on Mt. Rantemario by the EoS for two plants: Potentilla leuconata (Rosa.) and Hedyotis sp. (Rubi.). Roots were carefully excavated from the soil and later dried with the above-ground parts. The results demonstrated that although there was considerable variation, the median ratio of the eleven plants was 0.7 (table 7.4).
Zonation
The altitude at which the various forest types occur is determined largely by temperature and cloud level. Various schemes have been devised to differentiate between the forest types3 and a useful scheme for Sulawesi would be:
|
lowland and hill forest |
0 - 1,500 m |
|
lower montane forest |
1,500 - 2,400 m |
|
upper montane forest |
2,400 - 3,000 m |
|
subalpine forest |
3,000 m + |
Such zones are compressed on small mountains or on mountains on the periphery of large mountainous areas. This is known as the 'Massenerhe-bung' (literally 'mass uplift') effect (Grubb 1971), and should be borne in mind when reading about zones on mountains in, for example, Peninsular Malaysia (highest peak 2,200 m) or New Guinea (highest peak 5,031 m), where zones will be lower and narrower, and higher and broader respectively. In Sulawesi itself the effect is hard to observe because most of the vegetation on the smaller and outlying mountains has been felled. Zone compression may be thought to occur when dwarfed, moss forest is found at the top of even quite small mountains such as Mt. Tangkoko, but the apparent similarity of these forests to true upper montane forest is due to similar physical conditions; steep slopes, high winds and a relatively low cloud level (p. 492). The botanical composition of moss forest on peaks of 1,000 m or so is completely different from that of moss forest on high mountains of 3,000 m+. For example, the dominant trees in the elfin moss forest on Mt. Tangkoko are Rapanea sp. (Myrs.), Phaleria capitata (Thym.), and Saurauia lepidocalyse (Acti.) (Anon. 1980) normally found in lowland forest.
There are other reasons why the distribution pattern of a particular species of montane plant does not necessarily correlate with altitude. For example, the physical conditions experienced at high altitudes can occur in frost pockets (p. 489), beside mountain streams, on bare soil and near waterfalls. High-altitude plants in these relatively low locations may not flower or may have infertile flowers. To maintain a population, therefore, they are dependent on seeds or spores being dispersed from above. A similar but opposite mechanism occurs up the mountain whereby sterile or non-flowering plants are found higher above the species' normal range, for example, around volcanic fumaroles. Thus, for a given species there is a zone of permanent occurrence which supplies seeds or spores to adjacent zones of temporary occurrence. For this reason, a mountain may lack a species of plant even though at the same altitude on a higher mountain (with a zone of permanent occurrence) that plant may be present (fig. 7.9) (van Steenis 1972).
From O. Maessen pers. comm.
Figure 7.9. Zone of permanent occurrence with zones of temporary occurrence above and below it, showing the distribution of a hypothetical species (shaded areas).
Alter van Steenis 1972
Characteristic Plants
Whereas lowland forests are not dominated by any one family of trees (p. 386), the lower montane forests are characterized by the large numbers of oaks Lithocarpus and chestnuts Castanopsis (Faga.). This has been reported, for example, from Donggala and Gorontalo between 800 m and 1,200 m where the Castanopsis is in constant association with Phyllocladus, Agathis and Eugenia (Steup 1931). Only four species of Lithocarpus and two of Castanopsis are known to be present in Sulawesi (compared with 50 and 21 respectively in Borneo) (Soepadmo 1972) and these can be identified quite easily from their fruits (fig. 7.10).
Oaks and chestnuts have large, relatively heavy fruit which are poorly fitted for long-distance dispersal (Soepadmo 1972) (p. 379). Pigs and anoa eat the seeds and those which pass through the gut without being chewed or otherwise damaged may remain viable. Squirrels certainly predate on the seeds but they also take fruit out of the parent crown and store and hide some of them in the ground or in other 'safe' places (Becker et al. 1985). A proportion are forgotten and subsequently germinate.
Figure 7.10. Fruits of Sulawesi oaks, a - Castanopsis acuminatissima, b - C. buruana, c - Lithocarpus havilandii, d - L. glutinosus, e - L. elegans, f - L. celebicum. Scale bar indicates 1 cm.
Drawn from specimens in the Bogor Herbarium
Others trees common in the lower and upper montane forests are members of the Coniferae (pines and related trees) such as Podocarpus (Podo.), Dacrycarpus (Podo.), Dacrydium (Podo.), Phyllocladus (Phyl.) (fig. 7.11) and the magnificent and commercially important Agathis (Arau.).
Agathis is a genus of tall trees, commonly known as 'kauri' after their Maori name in New Zealand. The trees found in Sulawesi are A. dammara which is also known from the Philippines, Moluccas and parts of Borneo (Whitmore and Page 1980). Although found in lowland areas, they are dealt with in this chapter because the stocks of lowland specimens have been seriously depleted through general forest clearance and over-exploitation for the resin, the tapping of which has led to the death of many trees. It is the resin, known in the world of commerce as Manila copal, which has given Agathis much of its commercial interest because it can be tapped from the bark on a sustainable basis. The resin is used in spirit varnishes, lacquers and in making linoleum, but in villages it has been used as fuel for torches (Burkill 1966). The wood itself is finely-grained, pale and uniform and commands a price above most other hardwoods. The value of the wood has encouraged foresters to establish Agathis plantations with varying degrees of success, and in Java yearly increments of 30 m3 timber/ha have been achieved. It can grow up to 45-50 m in height with a diameter of over 1.5 m. The bark is composed of round scales which flake off, and the leaves are broad, flat blades (fig. 7.12).
Figure 7.11. Some conifers of Sulawesi montane forests, a - Phyllocladus hypophyllus male, b - P. hypophyllus female, c - Dacrycarpus steupi, d - D. imbricata, e - Dacrydium nidulum juvenile leaves left, adult leaves right, f -Podocarpus neriifolius. Scale bars indicate 1 cm.
After Wasscher 1941; de Laubenfels 1969; Keng 1972, 1978
Agathis is recorded from all over Sulawesi except the southeast peninsula and the region between Parigi and Toli-Toli. The former absence may be a result of a climate which is too seasonal, whereas the latter may be an artefact due to inadequate collecting. It is affected by climate, such that only in the central block of Sulawesi is it found in the lowlands; on the four peninsulas it is known only from the mountains. A general survey of distribution published in 1940 indicates that it is found in discrete but extensive pockets of forest where it is usually the largest tree present, although young trees are abundant in the undergrowth and lower canopy (van der Vlies 1940). Many of these pockets are as yet too isolated for logging companies to consider felling them. It has been suggested that the occurrence of Agathis in the forests of Mt. Kinabalu, Sabah, is correlated with the presence of podzol soils (Askew 1964).
Figure 7.12. Agathis dammara. Male branch with pollen cones (left), female seed cone (right). Scale bars indicate 1 cm.
From Keng 1972
The price of resin was very high in the first few decades of this century, and in the rush to harvest the resin many trees were killed. Attempts made to control the harvest met with little success (van der Vlies 1940). Agathis is not a strong pioneer species: forest dominated by Agathis on the limestone Mt. Malindu between the lakes of Matano and Towuti was felled in 1969 but five years later no seedlings or saplings were to be found (Whitmore 1977).
Rattans occur in lower and upper montane forest and some species are excellent indicators of altitude. Until they are collected properly, however, the genuine relationships will remain obscure (J. Dransfield pers. comm.).
Tree ferns4 are a common component of montane (particularly lower montane) areas in both disturbed and primary forests. Tree fern trunks can grow vertically one metre in 15 years but, like palm trees, some years elapse between the first growth of a young tree fern and the start of the verdeal growth of its trunk (Tanner 1983). The rough trunks evidently provide an unusual niche since it was noted on Mt. Roroka Timbu that a species of the fern Lindsaea (fig. 7.13) was found exclusively on tree ferns (van Balgooy and Tantra 1986). At present twenty species of tree ferns are known from Sulawesi, almost all of which are found on mountains (table 7.5). The species are distinguished by the form of the reproductive parts, and by the shape of the fine brown scales found covering the young, coiled leaves and the bases of older leaves.
Figure 7.13. Lindsaea pulchella an epiphytic fern of tree ferns and trees in montane forests.
After Kramer 1971
One genus of epiphytic ferns Lecanopteris, found in lower montane forest, has developed a large, multi-chambered, spiky rhizome in which ants live. Sulawesi appears to be the centre of evolution for these plants with more species found here than anywhere else. An expedition to Mt. Rantemario in 1969 discovered a new species (fig. 7.14) which was inhabited by Crematogaster ants. These enter the chambers through holes formed when a frond dies, falls off, and leaves a raised leaf base with a soft, vulnerable centre which breaks down rapidly, providing access to the inside (Jermy and Walker 1975). The relationship between different ants and another species of Lecanopteris in Sarawak has been described (Janzen 1974), and the principles involved have broader application. Apart from the obvious benefit of a secure shelter, the ants probably also receive food in the form of highly nutritious spores, or young or aborted sporangia. The latter are full of globules which could be oil bodies (Holttum 1954), but this has yet to be confirmed. The plant probably benefits from organic and mineral material deposited in the chambers in the form of dead bodies, faeces and discarded food. The fern certainly grows roots into the chambers but whether these absorb anything more than water has yet to be proven. It is possible also that the ants defend the fern from animals which may seek to eat the rhizome or fronds, but this too, has yet to be proven (Jermy and Walker 1975).
Pitcher plants Nepenthes (Nepe.) are also found in montane habitats as they are in other nutrient-poor sites (p. 450), where their unusual means of obtaining nutrients gives them a competitive advantage. The nutrient concentrations in pitcher leaves from montane forests were compared with the concentrations in leaves of nearby plants that were not able to catch insects. The results showed that pitchers had three times as much phosphorus and twice as much potassium as the other leaves (Grubb 1974).
Characteristic of the upper montane forest are members of the Ericaceae such as the large and colourfully flowered Rhododendron, bilberries Vaccinium, and wintergreen Gaultheria. There are 24 species of Rhododendron known from Sulawesi of which 19 are endemic. The bilberries Vaccinium are represented by 16 species of which 13 are endemic. The leaves, flowers and large black berries of Gaultheria taste strongly of winter-green oil (methyl-salicylate) a substance which is often used to alleviate the symptoms of rheumatism. Only two species are known and both are endemic, G. celebica with white, pink or red flowers, and G. viridiflora with greenish-white, red-based flowers. A lesser-known genus Diplycosia with smaller berries than Gaultheria has 17 species all of which are endemic and some of which also smell slightly of wintergreen. Rhododendron has very small seeds each of which has a small tail or wing at both ends which is presumably an adaptation for dispersal by wind. The fleshy fruits of the other species are eaten by birds (particularly thrushes and white-eyes), and mammals (Sleumer 1966-67).
Endemic species shown in bold type, ? = doubtful record needing confirmation.
After Holttum 1963
Figure 7.14. Lecanopteris spinosa, an ant fern discovered on Mt. Rantemario. Scale bar indicates 1 cm.
After Jermy and Walker 1975
One of the most obvious characteristics of the wetter parts of the upper montane forest is the enormous quantity of moss covering the ground and adorning every twig, branch and bough up to 2-3 m above the ground. In the upper parts of lower montane forest garlands of moss, typically Aerobryum, are frequently seen hanging from branches. Above the cloud zone the most common epiphyte is the beard 'moss' Usnea. This is in fact a lichen, a composite plant composed of a complex spongy framework of fungal mycelium strands within which sit algal cells. The algae photosyn-thesize and the fungus feeds on a proportion of the food produced. The algae can in fact lead a completely independent life but not so the fungus and so new generations have to re-establish the relationship. Thus Usnea and other lichens are given a name of their own although they are composed of two separate organisms. Interestingly, Usnea has been used as a medicine since classical Greek times. People living near mountains on Sulawesi use it for a number of purposes, but primarily exploit its astringent properties for curing intestinal troubles (Burkill 1966). It has been found also that the usnic acid found in the tissues of Usnea has antibiotic properties against tuberculosis (Fitting et al. 1954).
In the subalpine zone the myrtles Leptospermum and Decaspermum predominate, although it seems that they are never found together. In the open areas, many beautiful herbs can be found, many genera of which are typical of temperate regions, as well as grasses, sedges and rushes. Many subalpine plants form rosettes of leaves just above the ground. These rosettes are formed by the distance between successive leaves being extremely short, in order to reduce water loss and to keep the plant near the warmer ground. Only the flower stalk rises any distance above the ground, to facilitate pollination and seed dispersal. The under-surface of the leaves in some species with rosette leaves are covered with dense, silky, white hairs (p. 495).
Selected Mountains
Mt. Rantemario (3,450 m). The southwest slopes of this mountain have been converted to agricultural land up to 1,650 m where the lower montane forest begins. Elements of upper montane forest are found from about 2,150 m but oaks, the jagged-leaved Phyllocladus hypophyllus (Phyl.), the yew Taxus sumatrana (Taxa.),5 and the conifer Podocarpus steupi of the lower montane forest are still dominant at that altitude. At 2,650 m the forest is clearly upper montane with Rhododendron bushes and some Phyllocladus predominating. The first subalpine herbs are found here, notably the deep-purple, closed flowers of the gentian Gentiana laterifolia (Gent.), and the large, bright-yellow flowers of Hypericum leschenaultii (Hype.) described as 'the jewel' of the mountain flora of java (van Steenis 1972) and equally beautiful in Sulawesi. This plant is one of the rare examples of a wild plant being used directly as as a garden plant without modification through selective breeding. Also common in the undergrowth is the erect, dark-green giant moss Dawsonia and the small magnolia Drimys piperata (Magn.) with its attractive hanging white flowers (Buwalda 1949).
By 3,200 m the small trees comprise mainly Rhododendron, Decaspermum and Hedyotis sp. (Rubi.) with tightly stacked, thick leaves. Beneath these are found small shrubby Gaultheria, Styphelia suaveolens (Epac.) with small pink berries, and colourful flowered herbs such as the daisy Keysseria, a small ginger Alpinia, the little violet Viola kjellbergii, and two silver-leaved Potentilla (Rosa.): P. leuconata with large yellow flowers, and P. parvus with small yellow flowers. Along a little river at about 3,200 m a distinctive community of plants thrives including a ragwort Senecio (Comp.), a Swertia (Gent.), and clumps of the small spreading Epilobium prostratum (Onag.) with leaves less than 10 mm long but with small pale purple flowers atop a 8 cm flower stalk. Several grasses such as Poa, Monostachys, and Agrostis and sedges were also found as were the erect stems of the club moss Lycopodium.
In places where the subalpine forest opens into glades and where the bedrock of biotite schist is more or less visible, various species of 'cushion'-forming plants are found such as Centrolepis philippinensis (Cent.) (Ding Hou 1957), Monostachys oreoboloides (Gram.), Eriocaulon celebicum (Erio.), and Oreobolus ambiguus (Cype.). The shoots of these plants are densely packed; this is an adaptation against strong winds, keeping transpiration to a minimum and the temperature inside the cushion to a maximum. When these plants are relatively young they form single round cushions, but with age the centre (the oldest part) tends to die and a ring is formed as the plant grows outwards. Some rings found during the EoS expedition measured over one metre across although the rings were not complete. These rings are still single plants which have extended vegetatively and must be of considerable age. Most rings were, however, only 15-30 cm in diameter. Analyses of soil taken from inside these rings showed high nitrogen and phosphorus levels (0.23 and 48.6 mg/1 respectively) relative to outside the ring (p. 494).
The plant collection, made by a brief botanical expedition to this mountain in 1981, demonstrated how poorly known the Sulawesi flora is. The collection comprised 57 high-altitude species, and more than 10% of these were new to science while others were previously known from just one or two specimens (Smith 1981; van Steenis and Veldkamp 1984).
Mt. Lompobatang (2,871 m). This mountain, whose name literally means 'swollen belly' (van Zijll de Jong 1934), had its first recorded ascent in 1840 when James Brooke, later Rajah Brooke of Sarawak, reached the summit (Mundy 1848). Sited as it is in a sea of densely-populated agricultural land east of Ujung Pandang, it is not surprising that the lower slopes of this mountain have been utilized for firewood and felled for farms and plantations. Reports from 1905 (Sarasin and Sarasin 1905) reported the encroachment of farmland and in 1933 plantations of exotic pines were established above 1,650 m. In the lower montane forest that remains, various species of the native conifer Podocarpus as well as the native maple Acer caesium can be found (fig. 7.15) (Meijer 1983).6 It is one of the few trees in relatively unseasonal areas to loose its leaves regularly and when its branches are bare it generally flowers. The layer of recently fallen leaves is one clue to its presence, another is the buzzing of bees and wasps in the canopy visiting the flowers for honey (Bloembergen 1948).
The upper montane forest is dominated by twisted Leptospermum, but towards the first summit (Mt. Asumtatumpang or Buluasumpolong) this is found together with Vaccinium, Styphelia and Rhododendron. Above 2,500 m the only trees found are oaks and Leptospermum. In 1931 the tree trunks were all charred, the result of a fire 13 years earlier, and the situation now is similar to that described for Mt. Rantemario above. The Leptospermum seems to regenerate better than the oaks perhaps because the relatively large acorns are a favoured food of anoas and pigs, or perhaps because the Leptospermum is better able to shoot from its base. The summit of Mt. Lompobatang has blackberries Rubus (Rosa.), Hypericum, bilberries Vaccinium and many herbs and grasses.
Interesting orchids also used to be present (Bouman-Houtman 1926), but their current status is unknown.
Mt. Tambusisi (2,422 m). In the lower montane zone of this mountain in Morowali National Park, from about 1,300-1,700 m, the forest has an average height of 15 m with occasional 25 m-high Agathis emerging. The undergrowth has more palms than found in the lower forests. In the upper montane forest, between 1,700 m and 2,500 m, tall shrubs up to 4 m high are found, mainly Vaccinium. The epiphytes are numerous. Nearer the top, between the patches of bare rock, the shrubs are barely 1 m tall and were interspersed with herbs including the sticky-leaved, insect-catching Drosera peltata (Lack n.d.). This plant appears to have a wide altitudinal range, having been collected between 800 m and 3,225 m (van Steenis 1953).
Figure 7.15. Acer caesium branch and fruit. Scale bar indicates 1 cm.
After Whitmore 1972
In the valleys and in depressions in the ridge, elfin forest about 5 m tall is found. The gnarled and twisted trees have trunks of less than 6 cm in diameter, yet are surrounded by a 10-15 cm layer of brown and orange moss and liverworts. These also cover the ground to a thickness of 0.5 m or more. In this, as well as all the other montane areas, pitcher plants were common and at least eight species have been collected (A. Lack pers. comm.).
Mt. Nokilalaki (2,280 m). Mt. Nokilalaki can be seen from Palu but it is generally climbed from Lake Lindu (Bloembergen 1941). Agathis (Arau.) used to be a characteristic tree of the lower montane forests but many have now been felled. Formerly they were 40 m to 50 m tall with diameters in excess of 1 m. At about 1,600 m trees also include other conifers such as Dacrydium and Podocarpus, magnolias Aromadendron, Mastixia (Corn.), the maple Acer caesium (Acer.), Elaeocarpus (Tili.), all four species of oak Lithocarpus (Faga.) known so far from Sulawesi (p. 504), Calophyllum (Gutt.), and the endemic Macadamia hildebrandii7 (Prot.) (fig. 7.16), but the most common tree is the chestnut Castanopsis acuminatissima (Faga.). Higher up, above 2,000 m, Agathis is absent and the most common trees are Tristania (Myrt.), Lithocarpus and Castanopsis which are progressively smaller and more covered with moss with increasing altitude. The actual summit was cleared, as was the case for almost all Indonesia's major mountains, when the Triangulation Service set up concrete pillars in the early decades of this century for surveying and mapping—the Mt. Nokilalaki pillar was set up in 1919.8 In the 200 m below the summit the most common tree is the conifer Dacrydium which here grows to 3-5 m tall, with Rhododendron, Ardisia, and Psychotria beneath it. Below these are the giant mosses Spiridens and Dawsonia, peat moss Sphagnum, pitcher plants, scrubby Saurauia, and many ferns. A collection of these in 1976 yielded nine species previously unknown from Sulawesi (Meijer 1983). Among the other plants present is a climbing pandan Freycinetia, blackberry Rubus (Rosa.), the small magnolia Drimyspiperata, and Trachymene (Umbe.) (Bloembergen 1941).
Mt Roroka Timbu (2,450 m). The lower montane forest is dominated by the oaks. For example, Castanopsis acuminatissima is first found at 1,500 m and was extremely abundant up to about 1,700 m. In October the abundant acorns are apparently collected by local people and are roasted like peanuts. The tallest trees were Lithocarpus glutinosus, L. elegans, figs and an occasional Macadamia sp. In the understorey the dominant trees are species of Eugenia and a species of climbing bamboo Racemobambos (Dransfield 1983; van Balgooy and Tantra 1986).
Between 1,900 m and 2,000 m the vegetation changes quite abruptly with Agathis becoming dominant up to 2,250 m. These are the straightest and tallest trees (about 40 m) and beneath them grow other conifers such as Dacrycarpus, Dacrydium and Phyllocladus. Among the understorey trees are giant pandans reaching 25 m in height with stilt roots 10 m long. The dense crowns of the Agathis and Pandanus make the forest floor very dark, and small trees, shrubs and herbs were only found near small streams above which the canopy is not so thick. The only common climbers are pitcher plants, and a new species of small climbing pandan Freycinetia micrura (Stone 1983). The branches and trunks of trees are covered with abundant epiphytes.
Above 2,000 m the forest becomes lower and more crooked, the canopy is more open, and hence the undergrowth is denser. By 2,250 m Agathis has been left behind and the dominant trees are various conifers and Lithocarpus havilandii. The understorey is rich in Myrtaceae (Leptospermum and Eugenia), and Rhododendron and Vaccinium are found growing epiphytically and on the ground. At the summit the woody plants are almost all Leptospermum, Rhododendron, Dacrycarpus, Phyllocladus and Podocarpus, and all are covered with thick cushions of moss. Various herbs grow in the moss such as pitcher plants and orchids (van Balgooy and Tantra 1986).
Figure 7.16. Macadamia hildebrandii. Scale bar indicates 1 cm.
After Soewanda n.d.
Pollination and Dispersal
Studies of flowering, fruiting and new leaf production have not been made on Sulawesi mountains but relevant information is available from Java. Most species have peaks of flowering relating to climate, but there are generally at least a small number of individuals from any one species flowering at any one time. Some plants appear to be indifferent to climatic patterns, others flower most in the dry season, while others appear to concentrate flowering in the wettest season; these differences may be due to the availability of different agents of pollination.
The pollination of montane plants has been little studied and simple observations of insects or other animals visiting flowers do not necessarily constitute observations of pollination. Also, plants which look as though they are adapted for pollination by a certain group of animals or by wind, may in fact be self-pollinated, even though they are visited by beetles, flies and other insects. On Mt. Rantemario one obvious self-pollinating flower is the purple gentian Gentiana laterifolia whose flower never fully opens. The plants of montane vegetation are visited by a range of animals such as:
• bats on bananas in lower montane forest;
• moths (noctuids and sphingids) on whitish or greenish long-tubed flowers with nocturnal scent;
• butterflies on colourful trumpet-shaped flowers such as the orange form of Impatiens platypetala (Bals.) endemic to Sulawesi;
• carpenter bees (up to about 1,400 m) and bumble bees (above this) on Vaccinium, Gaultheria, Hypericum and many other plants;
• flies and beetles on flowers such as certain orchids and aroids with a smell resembling decaying flesh;
• wasps on figs and, perhaps;
• birds such as flowerpeckers on long-tubed red flowers, but observations from Sulawesi mountains do not seem to have been recorded.
There are four main ways in which plants disperse their fruit and seeds (or, in the case of ferns, mosses, fungi, etc., their spores):
• in water;
• by wind;
• by terrestrial animals onto which seeds stick, or which eat seeds but do not destroy them, or which remove seeds for later ingestion but then forget them; or
• by sticking to, or being eaten and not destroyed by, flying birds or bats. An interesting example of a Sulawesi plant with a double adaptation for dispersal is the herb Triplostegia glandulifera (Dips.) (fig. 7.17) whose fruit has both sticky hairs and hooks (van Steenis 1951; van Steenis and Veldkamp 1984).
Dispersal by water or mammals are clearly of no significance in explaining the regional distribution of high-altitude plants (p. 62) because water only flows downstream to unfavourable habitats, the distances between high mountain tops are too great for terrestrial mammals to cross before the seed is either excreted or rubbed off, and bats are scarce at high altitudes. Wind would seem a likely candidate, therefore, for the dispersal of the minute, dustlike seeds of orchids, fern spores, and the light, parachute-like, plumed seeds of daisies, but the expected wide distributions of such species is not always found to be the case (van Steenis 1972).
Many of the shrubs and small trees in the upper montane and subalpine forests produce berries which are an important source of food for birds and mammals. Many of the seeds in these juicy fruits pass through bird intestines without harm. Since the time taken by a seed to pass through the intestine is little more than an hour, however, the possibility of transferring seeds to mountain tops more than about 30 km away is remote and, in addition, birds of mountain tops do not generally fly far. Sticky or adherent seeds and fruits may attach themselves to feathers or legs but, before a long flight, birds will generally preen themselves thoroughly to ensure maximum flying efficiency, and by so doing dislodge any attached seeds. A further complication to understanding dispersal is the observation that communities of plants having a variety of dispersal mechanisms have exactly overlapping ranges on different mountains as though means of dispersal were immaterial (van Steenis 1972).
Figure 7.17. Triplostegia glandulifera, a plant with two means of animal dispersal. Scale bar indicates 1 cm.
After van Steenis 1951
As stated previously (p. 62), the lowering of forest zones during the Pleistocene would have made chance dispersal easier if only because more 'islands' of suitable habitat would have become available. But even then, large gaps between these habitats would have existed and it is clear that the means by which plants or groups of plants effected their dispersal is not yet fully understood.
Biomass and Productivity
The trends in biomass and productivity change from lowland to upper montane forests are as follows (Grubb 1974):
• biomass decreases proportionately less than height, resulting in shorter, stockier trees in the upper montane forest;
• production of woody parts declines from about 3-6 t/ha/yr to about 1 t/ha/yr;
• production of litter, particularly leaf litter, decreases proportionately much less than biomass or production of woody parts;
• biomass of leaves decreases proportionately much less than total biomass;
• mean life span of leaves increases only slightly;
• the leaves become thicker and harder and the leaf area per gram decreases from up to 130 cm2/gm to about 80 cm2/gm.
Thus, with increasing altitude, total production decreases but plants invest a greater proportion of their production in making leaves although these more 'expensive' leaves may last no longer than the 'cheap' leaves of the lowlands (p. 370).
Only two studies of litter production have been conducted on Indonesian mountains and both of those were on Java. The annual rate of total litter production in lower montane forest was between 6,000 kg/ha/yr and 7,500 kg/ha/yr, about 75% of which was leaves (Yamada 1976; Bruijn-zeel 1984), which is not so different from results from lowland forest in Peninsular Malaysia (Lim 1978). Leaves fall throughout the year but the main peak is during the dry season. Branch fall is greatest during storms which occur most often during the wetter seasons.
Mineral Cycling
There are two main classes of mineral cycling in a forest. First, there is the rapid cycling of minerals in the fallen leaves and twigs, and in the rain falling through the forest canopy to the floor. Second, there is the much slower cycling of minerals held in the tissues of large woody parts of trees.
Lower quantities of minerals cycle through leaves in montane forests than in lowland forests, and fewer still through woody parts. This is a consequence of lower production of woody parts compared with leaves, and the concentration of nutrients in leaves of montane plants being about half that of their counterparts in lowland forest. The concentrations of minerals in the montane leaf litter are proportionately even less than in lowland litter because montane plants absorb into their permanent tissue about half of their leaf minerals before shedding them, whereas lowland plants absorb only about one quarter of their leaf minerals.
Assuming results from New Guinea to be broadly applicable to the forests of Sulawesi, the major external input of minerals is from the rain (Edwards 1982b) (fig. 7.18) which has been contaminated by ash from fires (Ungemach 1969) and volcanic eruptions. Not all the rain falling over a forest reaches the floor because it evaporates, is intercepted by epiphytes and other leaves, or is absorbed by bark, root mats, etc. When the rain is light, little may reach the forest floor, and even in the heaviest rain only about 25% is intercepted (Edwards 1982b). The rain that does complete its journey to the ground will have also picked up minerals from animal droppings, plant exudates, and humus around epiphyte roots and will be far more mineral-rich than rain falling in a clearing.
Figure 7.18. Mineral cycling in a lower montane forest. Figures in boxes are in kg/ha. Complete arrows represent major pathways of mineral transport, and figures alongside them are in kg/ha/yr. Dashed arrows are presumed pathways that have not been quantified.
From Edwards 1982b
As indicated above, leaf litter does not have the same mineral content as the living leaves because of absorption by the plant. Some minerals, such as calcium and magnesium, do not seem to be resorbed while others such as nitrogen and phosphorus are. It is possible that minerals that are absorbed may be relatively scarce and limiting in the soil.
Volcanoes
Volcanic eruptions are major geological and biological events. Not only does a great deal of subterranean material come to the earth's surface, but also many plants and animals die as a result of the intense heat of the lava, and the smothering by ash. The major volcanic product in Sulawesi is ash and this can interfere with living leaves by adhering to them and preventing photosynthesis. A layer of ash just 1 mm thick can reduce light penetration to the leaf by 90% or more. These are short term effects, however, and rain soon washes off the dust.
Around active craters or sulphur vents the soil is usually rocky, very pervious, sterile, acid, and lacking in organic matter. For example, soil collected from near a fumarole above Lake Moat, Bolaang Mongondow, had a pH of 3.6. The atmosphere is generally a choking mixture of sulphur dioxide (S02), and hydrogen sulphide (H2S) mixed with smaller quantities of chlorine (Cl2), carbon monoxide (CO), and nitrous oxide (NO). In addition, the ground is frequently heated from beneath. These are conditions which are hardly favourable to plants, but it is amazing how tolerant some can be. Those closest to the sulphur vents tend to be dwarfed, prostrate, and to grow extremely slowly. The closest plants are often Vaccinium, Rhododendron and ferns. Lichens may be found closer still and some blue-green algae are able to live in hot, muddy, sulphurous pools (Kullberg 1982).
It may take centuries before new lava flows have been sufficiently weathered to support closed forest but, if there is sufficient moisture, the first plants can be found virtually as soon as the lava cools down. For example, an EoS team examined the lava that erupted from Mt. Api on Siau Island two years after the eruption in 1983. The only plants that could be found were ferns and some patches of lichens.
The vegetation on volcanic ash develops relatively quickly. For example, the island volcano of Ruang (fig. 1.4), erupted dramatically in 1874, and 11 years after that the ash slopes were covered in sedges and leguminous herbs (Hickson 1889). The rate of colonization is obviously affected by the proximity of a source of seeds and spores and also of disperser animals (p. 55). The warm, volcanic ash of Ruang had, in 1885, already been marked by abundant footprints of maleo birds Macrocephalon maleo (p. 155) (Hickson 1889) which may have brought seeds on their feathers or in their guts from the neighbouring island of Tahulandang. Other colonizers of ash slopes include grasses, Vaccinium, Gaultheria and Rhododendron, but the fern Gleichenia (Glei.)9 (fig. 7.19) is common where landslips have occurred. An interesting and conspicuous plant found on Mt. Awu, on Sangihe Island and on volcanoes in Minahasa is Gunnera macrophylla (Halo.) which has coarse and strongly-veined roundish leaves measuring up to 50 cm in diameter. At the base of the leaf stems are three warts, immediately below which a root emerges. These warts contain colonies of the blue-green algae Nostoc which presumably help in nitrogen fixation. This would be distinctly advantageous in nutrient-poor, young volcanic soils. It is believed locally that the flowering of the Gunnera on Sangihe means that an eruption of Mt. Awu is imminent. The plant produces large amounts of yellow pollen and it may be by association that the relationship is held to exist (van Steenis 1972).
Figure 7.19. Gleichenia truncata. Scale bar indicates 1 cm.
After Holttum 1959
A comparison between the number of tree species and the percentage of ash in the topsoil over an area north of Mt. Batuangus shows that, as expected, fewer trees are found on soils containing high quantities of ash (fig. 7.20). Due east of the Batuangus crater the ash content is very high, but some ash blew into the neighbouring Tangkoko crater and even over the crater wall to the north of Tangkoko. The very high quantity of ash half way along the coast is due to a separate centre of volcanic activity. This correlation indicates that the last eruption in 1839 must have killed large numbers of trees where most of the ash fell and the number of species is still increasing in the impoverished areas (Anon. 1980). Parts of the cone of Batuangus are still bare but others are sparsely forested with groves of Casuarina (Casu.). This same tree is also present in the parts of the Tangkoko crater worst affected by the Batuangus ash (Anon. 1980).
Figure 7.20. Distribution of ash and tree species on the slopes of Mt. Tangkoko (left) and Mt. Batuangus (right). For ash: darkest shading indicates at least 80% ash in topsoil, lightest shading 40%-44%, intermediate shadings in increments of 5%. For tree species: darkest shading indicates at least 45 species per 0.5 ha, lightest shading 5-9 species, intermediate shadings in increments of 5 species.
After Anon. 1980
ANIMALS AND THEIR ZONATION
Most tropical animals are unaccustomed and unadapted to cold temperatures. This is because the majority of animals, including insects, amphibians and reptiles, draw their body warmth from their surroundings. Some of these will raise their temperatures by basking in the sun and as long as nighttime temperatures do not reduce body temperatures below a critical minimum, then these animals can live successfully in relatively cold climates. No lizards or amphibians seem to have been recorded on mountain tops in Sulawesi but this may in part be because they have not been specifically sought. The cold experienced even on summits would not totally exclude the possibility of these animals living there: for example, a skink found in sub-alpine vegetation in south-east Australia had a critical minimum temperature of -1.2°C (below the freezing point of the animal's blood) and was able to move around below a cover of snow (Spellerberg 1972).
A number of animals are restricted to high mountains, but even on Mt. Tangkoko large animals of lowland forest show preferences for certain altitude ranges (fig. 7.21).
Invertebrates
The number of invertebrates declines with altitude. On Mt. Mulu, Sarawak, for example, the decrease in soil macro invertebrates was accounted for largely by the gradual reduction in the abundance of ants and termites. Biomass, however, reached a maximum in the peaty soils of the taller parts of upper montane forest where beetle larvae took over from termites as the major detrivores. At higher altitudes, centipedes and spiders took over from ants as the major predators (Collins 1980a). Land crabs and detritivore noctuid moth larvae were found on the top of '1,440', one of the peaks in Bogani Nani Wartabone National Park, indicating a distinctive detritivore community. Land crabs have also been reported from 1,350 m in the Quarles Mts., South Sulawesi (Coomans de Ruiter 1950).10 Land crabs found in alluvial lowland forest in Sarawak were found to have vegetable, mineral and insect remains in their guts (Collins 1980b). Ten pitfall traps11 were set at 2,000 m and 3,250 m on Mt. Rantemario and were cleared after two days and nights. The ones at 2,000 m contained a variety of tiger beetles, ground beetles, a few spiders and springtails (fig. 7.22) whereas those at 3,200 m caught mainly large numbers of melolonthine scarab beetles, some ground beetles, and one spider. Earthworms were absent above 2,100 m on Mt. Mulu and dousing 1 m2 of ground with weak formalin12 resulted in no earthworms being found at 3,250 m on Mt. Rantemario. At 2,200 m, however, 32 of at least 10 cm length were caught with this same method and on the very summit of Mt. Rantemario, large earthworms were found under every sizeable rock and these must be significant detrivores in the subalpine zone. Huge 50 cm long earthworms Amyntas sp. have been reported from 1,350 m in the Quarles Mts. (Coomans de Ruiter 1950). Spoonfuls of honey were placed on the ground at 3,250 m on Mt. Rantemario and not a single creature was seen to visit them, a dramatic contrast to the rapid attention honey attracts in lowland forest.
Figure 7.21. Changes in relative frequency with altitude of common birds and mammals on Mt. Tangkoko.
After Anon. 1980
Figure 7.22. Springtail.
Communities of animals change with altitude. For example, a disjunction in the distribution of beetle species was found up Mt. Mulu such that no beetle species found below 500 m (lowland forest) also occurred above 1,500 m (upper montane forest) (Hanski 1983). The highland beetles were thus a distinct guild of species and preliminary results from Lore Lindu and Bogani Nani Wartabone National Park suggests that the same general pattern is found in Sulawesi (J. Krikken pers. comm.).
Dawn-to-dusk captures of insects were conducted in Morowali National Park using very bright mercury vapour lamps set against a white sheet at sea level, 1,200 m and 2,000 m, but unfortunately results are not yet available.
The flying insect fauna on the peak of Mt. Rantemario was dominated by large sciarid/syrphid flies which were most abundant on the flowers of Decaspernium. No bumble bees or butterflies were seen. Insects on the relatively open peak were quite abundant, and one reason may have been the quantity of anoa dung which many insects use as a food resource for their larvae. Insects (bugs, flies and beetles) have also been reported from the near sterile summit of the active volcano Mt. Soputan, Minahasa, in December 1985 (C. Bennett pers. comm.) This phenomenon has been observed elsewhere and the term 'summit-seeking' has been given to the way they fly up mountains and sometimes congregate in large numbers. The basic factors underlying this behaviour are not known and a bewildering variety of explanations have been proposed (Mani 1980).
Birds
Although numerous species have quite a wide altitudinal range, others are found within quite narrow limits (table 7.6). Because of this, knowing one's altitude on a mountain allows one to predict which species are likely to be encountered or, conversely, hearing or seeing a number of bird species enables one to determine one's approximate altitude (Heinrich 1939).
The lowland zone, up to about 1,200 m, is characterized by large or conspicuous birds such as hornbills, malkohas Phaenicophaeus calorhynchus and coucals Centropus celebensis, maleo Macrocephalon maleo, bee-eaters Merops spp., pittas Pitta, and by the relative paucity of songbirds. The middle zone is characterized by various doves and pigeons, drongos Dicrurus hottentotus, and song birds such as the serin Serinus estherae (Coomans de Ruiter 1950; Schuchmann and Wolters 1982; Bishop and King 1986). The ranges of these birds extend down to the lowland zone and upwards, not to a common upper limit as in the lowland fauna, but rather extending to varying degrees into the upper zone. The doves of the middle zone are replaced at higher levels by the fruit dove Ptilinopus fischeri and the very rare dusky pigeon Cryptophaps poecilorrhoa, songbirds are far more common, and Sulawesi woodcocks Scolopax celebensis may be found.13 A few birds can be found in the subalpine zone but only one species, the non-endemic Island thrush Turdus poliocephalus is confined to this habitat where it would eat the abundant fruit and earthworms. The others are more at home below this relatively harsh zone. Strangely, perhaps, on an island with so many endemic species, only about half of the high montane birds are endemic, and the Island thrush is widely distributed in Indonesia.
There are also a fair number of birds which can be found from, more or less, sea level to upper montane forest. Some species including the rare endemic hawk-eagle Spizaetus lanceolatus and the racquet-tailed parrot Prioniturus platurus are strong fliers, but other wide-ranging montane species such as the little fulvous woodpecker Mulleripicus fulvus, the pink-breasted cuckoo-dove Macropygia amboinensis and the ground-dove Gallicolumba tristigmata are not known for their flying prowess.
After Heinnch 1939; Marshall 1978; Klapste 1982; Watling 1983
The abundance of most species of birds varies gradually with increasing altitude but there are also pairs or trios of species which replace one another relatively abruptly with altitude. The precise altitude would vary from place to place depending on exposure, slope and rainfall. One such pair is Fischer's fruit dove Ptilinopus fischeri and the dark-chinned fruit dove P. subgularis which meet at more or less the boundary between upper and lower montane forests but not necessarily at the boundary of vegetation types. A trio of species with different altitudinal ranges is found among the white-eyes, at least near Lore Lindu National Park. Thus Zosterops atrifrons with a black forehead is found in lowland and hilly forest,1 Z. montana in lower montane forest, and Lophozosterops squamiceps in upper montane and subalpine forest (K. D. Bishop pers. comm.). Such transitions may not be heeded by juvenile birds and if a bird is found at the 'wrong' altitude, it is generally found to be a juvenile or immature individual (Diamond 1973).15
As a result of the different altitude preferences, the proportion of birds occupying different niches changes with altitude such that:
• the proportion of tree-nesting insectivorous species increases;
• the proportion of tree-nesting frugivorous species decreases;
• the proportion of tree-nesting omnivorous species stays about the same;
• the proportion of predatory birds decreases; and
• the proportion of ground-living birds increases (Kikkawa and Williams 1971; K. D. Bishop pers. comm.).
Mammals
Birds and mammals generate their own heat and heat loss poses enormous problems for the smaller species because their surface area is very large compared with their weight, and their heat loss must be compensated for by consuming large quantities of high-energy food. It is therefore surprising to learn that the smallest Sulawesi mammals, the shrews, are found at high altitudes.
Shrews are almost entirely carnivorous and eat small insects, snails, earthworms, millipedes, centipedes and spiders; resources which are scattered, cannot be stored and whose abundance is often unpredictable. Shrews therefore have to spend a great deal of time searching for food, an activity which in itself uses up energy. It is not surprising, then, that shrews eat about their own body weight in food each day and are active almost the whole day and night although they rest briefly every few hours. If more than a few hours pass without a meal a shrew will starve to death, although larger ones such as Crocidura elongata (head and body 94 mm) can survive hunger better than smaller ones such as C. lea (head and body 60 mm). The response to food shortage also differs with size: smaller species increase and larger ones decrease their foraging activity (Hanski 1985). Shrews make underground burrows in which they build a nest such that during the cold nights the nest temperature is probably significantly higher than outside.
The little red tree mouse Margaretamys parvus (head and body length 10 cm) and the small rat Haeromys (head and body length only just over 7 cm), experience similar heating and eating problems but the diet of the former is primarily composed of insects, like the shrews, and of the latter of fruit, roots and seeds, the sources of which are often relatively large and storable.
The mountains of Sulawesi are home to a large number of rats which have been studied intensively on Mt. Nokilalaki in Central Sulawesi (Musser 1982). Among the species present were three species of shrew-rats, so called because of their relatively long snouts and superficial resemblance to shrews. Their head and body lengths are between 11 cm and 15 cm and they are known only from mountainous areas in Sulawesi and the Philippines.
The shrew-rats live in mossy upper montane forest mainly amongst the wet undergrowth of wild gingers, young rattans, ferns, and small shrubs. In clearings they live mainly among the various sedges. Runways, regular routes used by the shrew-rats, are frequently found along, or partly under, rotting trunks or boughs. These runways provide a certain amount of security from predatory owls and may be shared by other species; for example, one particular runway was used by all the shrew-rats, a ground squirrel Hyosciurus heinrichi, a rat Rattus hoffmanni and shrews. One shrew-rat Tateomys rhinogradoides on Mt. Nokilalaki had a very limited range. The mountain's summit is 2,280 m yet no specimens of this animal were caught below 2,210 m despite intensive trapping.
Melasmothrix naso was found to be diurnal and most were caught in the middle of the morning. This shrew-rat shares with other diurnal rats in Southeast Asia the characteristics of a chunky body, tail shorter than head and body combined, and a dark chestnut colour. The two Tateomys species are nocturnal and are grey or greyish-brown, a characteristic of nocturnal rats. These differences presumably relate to camouflage effectiveness.
The dietary range among the various rats in the upper montane forest was determined by food trials with captive animals and stomach analyses (table 7.7). Both Tateomys species seem to eat only earthworms whereas Melasmothrix naso eats mainly earthworms supplemented with fly larvae. Animals favoured by other rats living in the same habitat included moths and cicadas. The food resources are divided between the rats and shrew-rats by virtue of the animals' different activity periods, and by their different means of finding food: for example, T. rhinogradoides digs into soil, moss and rotten wood on the ground, whereas T. macrocercus climbs over mossy boulders and tree trunks seeking worms in the moss. None of the other rats in the moss forest eat earthworms; the squirrels eat Lithocarpus acorns, and the shrews eat insects. There are, however, other earthworm-eating rats at lower altitudes such as Maxomys hellwandi, Echiothrix leucura, Bunomys andreiusi and B. chrysocomus.
A number of other small mammal species were found near the summit of Mt. Nokilalaki: two long-nosed ground squirrels Hyosciurus, the small tree squirrel Prosriurillus abstrusus, and the very small squirrel P. murinus. The large red squirrel Rubrisciurus rubriventer was found up to 1,500 m but no higher. Various small insectivorous bats were seen occasionally, and the most common bat was the frugivorous tube-nosed bat Thoopterus nigt-escens.
Some of the rats found around the summit of Mt. Nokilalaki, such as Paruromys dominator, Rattus marmosurus, R. hoffmanni, and Maxomys musschenbroekii, are also found all the way down to sea level but others such as Bunomys chrysocomus, B. penitus,1 R. arcuatus, R. callitrichus, R. hamatus, R. facetus, Margaretamys parvus, the little Haeromys sp., Eropeplus (anus and three species of shrew-rats are found only in lower or upper montane forests.
A study of rats in Morowali National Park showed that fruit-eating climbing species were more abundant at higher than lower altitudes. Also none of the species trapped in lowland forest was found in the montane forests. Only two rats were caught in the upper montane forest, a new Maxomys and a new Bunomys, but only the latter was restricted to this forest type. After a period of intensive trapping it appeared, however, that just 10 m down the slope, it was replaced by its congener B. chrysocomus—a neat example of allopatric distribution of sister species (C. Watts pers. comm.).
(a) = Arboreal rats nesting in trees but occasionally caught on the ground.
(b) = Ground-nesting rats which seek food in understorey and canopy trees. Others = mainly ground dwellers but occasionally found above ground.
After Musser 1982
In contrast to the commensal rat species of towns and fields whose high reproductive output (p. 584) and relatively low inherent survivorship mark them as r-selected species, the montane rats of the much more stable forest environment appear to be K-selected species with much smaller litters of just one or two young (C. Watts pers. comm.).
The largest mammal of mountain forests is the mountain anoa Bubalus quarlesi.17 These are generally solitary but a herd of five ran past an expedition climbing Mt. Nokilalaki (Meijer 1983). Dung samples collected near the summit of Mt. Rantemario were analysed to determine the plant species.18 The results showed that, as expected, the anoas were browsing animals for which grassy plants are relatively unimportant (table 7.8). Most interesting was the significant quantity of moss in the dung. Moss fragments are very easy to identity and thus may have been overrepresented when fragments were chosen, but there are enough present to suggest that it was certainly not ingested accidentally. Anoa may eat moss as much for its water content as for its food value. The suggestion that anoa eat moss was also made from observations of moss damage on Mt. Roroka Timbu (van Balgooy and Tantra 1986).
Anoa dung has often been reported as being common along the logging road constructed up to 2,225 m on Mt. Roroka Timbu in the northeast of Lore Lindu National Park. On a recent visit, however, very few signs of anoa were found. This might have been a seasonal phenomenon but many snares and traps set for anoas were found and these quite correctly were destroyed since the anoa is protected by law (K.D. Bishop pers. comm.). A ranger, Alex Alisi, of the Forest Protection and Conservation Service in Central Sulawesi, lost his life on that mountain and it is possible that his death was somehow connected with the poaching of anoa in and around
From M.J.B. Green pers. comm.
Lore Lindu National Park. He was posthumously awarded one of the Kalpataru Servants of the Environment awards in 1985.
Also known from upper montane forest is the Sulawesi civet19 Macroga-lidia musschenbroekii, one of the world's least-known carnivores. The animal itself is rarely seen and its presence is more often indicated by tracks,20 isolated faeces, and scratches on trees up to 2.5 m above the ground. The Sulawesi civet is a master of acrobatic climbing and like some of its relatives, its hindfeet can be rotated to allow head-first descent of vertical surfaces (Wemmer et al. 1983). In the last five years positive signs or sightings of the civet have been made on Mt. Ambang (Minahasa), Lore Lindu National Park (Wemmer and Watling 1986), and Mt. Rantemario (EoS expedition). In the more than 100 years since it was first described for science, very few have been seen by scientists or collected for museums despite intensive searches. In the past, however, the Toraja people are said to have trained these civets for hunting, although they were unreliable (Anon. 1977).
Analysis of 47 faecal samples from Sulawesi civets showed that small mammals and fruit were common components of the diet although they may also take ground-living birds. Most of the small mammals were rats but the small nocturnal cuscus Strigocuscus celebensis is evidently taken occasionally and feathers were found in two of the samples. Local informants report that young piglets, farmyard chickens and bananas are also eaten. From the fruit remains it was clear that Sulawesi civets ate the pulp-covered seeds of the wild sugar palm Arenga sp. but only when the fruit was ripe (Wemmer and Watling 1986). This is not surprising because when still unripe the pulp contains water-soluble sodium oxalate which, if it comes into contact with the lips, mouth or other mucus-covered tissue, absorbs calcium from the mucus and forms water-insoluble, needle-sharp crystals of calcium oxalate which are extremely painful. During the ripening process, however, the sodium oxalate is broken down and the ripe fruit can be eaten safely (Whitten 1980). Observations suggest that the Sulawesi civet walks around a particular area of forest, not necessarily on obvious paths, taking 5-10 days to complete the circuit of its home range measuring about 150 ha (Wemmer and Watling 1986).
Macaques have been observed up to 2,000 m; for example they have been seen on the peak of Mt. Klabat (1,998 m) (Sarasin and Sarasin 1905), but they are probably relatively uncommon above 1,500 mm.
EFFECTS OF DISTURBANCE
Most people in Sulawesi live in climates where plants grow extremely quickly. It is important, therefore, that those people who climb mountains for recreation or study should remember just how long-lasting the effects of disturbance can be in a much less productive environment.
As mentioned earlier, montane meadows are almost certainly engineered by people in attempts to attract and thence to kill anoa (p. 531). When Mt. Rantemario was climbed in 1937 it was remarked that a fire had quite recently burned the vegetation around the summit (van Steenis 1937; Steup 1939). When the EoS team climbed to the summit in 1985, most of the dominant Decaspermum trees were only about 2-2.5 m tall but there were isolated trees of the same species reaching 4 m. These are probably those that escaped the worst affects of the fire 50 years earlier. Lowland forests are not easily ignited and none of their trees have any fire-resistant characters. The vegetation is generally wet or moist and there is no pile of easily-dried material that could act as tinder. In upper montane and subalpine forests, however, periods of drought do occur and the oil-rich leaves of Rhododendron, Vaccinium and Gaultheria, can catch alight even in the rain. In addition, people are able, through successive attempts to burn an area, to increase the abundance of grasses and sedges, the dead leaves of which can serve as potential tinder for the next fire (van Steenis 1972). Members of the Ericaceae, such as Vaccinium and Rhododendron, may be fire resistant to some extent for in places in the Latimojong Mts. and near Rantepao, nearly pure stands of the endemic, 3-4 m tall, yellow-orange flowered Rhododendron vanvuurenii can be found in fire-generated grassland (Sleumer 1966-67).
The succession of plants in montane areas up to 2,000 m altitude is quite similar to that lower down. Eupatorium inulifolium (Comp.) and Lantana camara (Verb.) quickly dominate abandoned fields, form a protective cover for the soil and for young tree seedlings. Where the area is subjected to repeated burnings, herbs such as Lobelia nicotianaefolia (Camp.) (Moeliono and Tuyn 1960) and the white-leaved 'edelweiss' Anaphalis longifolia (Comp.), and bracken fern Pteridium aquilinum (Polp.) increase in abundance. With further burning alang-alang grass Imperata cylindrica would become dominant. In moist places the wild sugarcane Saccharum spontaneum dominates. The course of succession above 2,000 m has not yet been investigated.
The regeneration of Costa Rican subalpine vegetation following a fire21 has been studied and the response of two shrubs, Vaccinium and a Hypericum, both of which have species on Sulawesi mountains, was monitored for three years after the fire. The fire killed the above-ground parts of the shrubs but suckers grew from their bases. The suckers grew less than an average of 50 cm in the three years. The soil surface exposed by the fire did not, as one would expect in lowland areas, become covered with fast growing pioneer species. Instead, much of the bare soil remained unvegetated. In other places liverworts and mosses started to grow. Very few seedlings of the surrounding plants were found. Three years after the fire, dead stems of plants were still standing and even those that had fallen to the ground showed no signs of decay. This extremely slow rate of decomposition is caused partly by the low temperatures but also by the absence (or extreme paucity) of most of the lowland decomposers such as ants, termites, and earthworms. This in turn may be due to the extremely wet condition of the soil, branches, and logs which never warm up to any great extent (Janzen 1973).
Regeneration of Costa Rican oak forest, which is similar to that found on Sulawesi mountains, was also found to be extremely slow. Repeated cutting would deplete the seed stocks irrevocably and it was concluded that montane forests are 'truly the tropics most fragile ecosystems' (Ewel 1980).