Chapter Eight

Caves

INTRODUCTION

Caves¹ are simple natural ecosystems of great value for understanding ecological inter-relationships as well as for their own intrinsic interest. They have advantages over many other ecosystems in terms of potential for research, both theoretical and applied, in that the boundaries are discrete, and most of the species inhabiting them can be easily studied and manipulated in the cave or laboratory.

Caves are enormously varied: they range in length from a few metres to over 100 km in a cave network in the U.S.A., and in depth from a few metres up to 1,311 m in France. The largest chamber, indeed the largest enclosed space, natural or synthetic, in the world, is the Sarawak Chamber in Mt. Mulu National Park, Sarawak. It is nearly two kilometres in circumference, and the floor area is equivalent to 17 football fields. Caves may have small entrances opening into large chambers, or massive entrances behind which the cave is only penetrable for a short distance. Some caves have rivers flowing through them, and are called active, and others are more or less dry having been formed in the past, and these are called non-active. Caves of sorts were created during the Japanese occupation when tunnels were dug into hillsides. These have many of the same features as natural caves and are not without interest. For example, the tunnels at Kawangkoan near Tomohon, Minahasa, are home to both bats and swiftlets.

All of the large areas of karst limestone in Sulawesi (fig. 6.14) may contain caves but hardly anything is known about their abundance, size or ecosystems. To stimulate interest, an attempt has been made to make an preliminary inventory (table 8.1). The longest cave is the 11 km Salukan Kalang Cave in Maros which was penetrated and investigated for the first time during the preparation of this book (Anon. 1985b). It is the second longest cave known so far in Indonesia (R.V.T. Ko pers. comm.).

One of the earliest accounts of a Sulawesi cave was by James Brooke (later Rajah Brooke of Sarawak) who visited Mampu Cave, Bone, in 1840 hoping to confirm reports of the presence of ancient statues carved by animistic peoples. The statues turned out to be nothing more than fallen stalactites with various names arbitrarily bestowed on them. These 'statues' can still be seen and are said to be the members of the court of the Mampu who were turned to stone through a curse. This curse was set when the princess dropped a spool and asked her dog to pick it up because she was too lazy to do it herself. Brooke found the cave impressive enough:

N - Non-active cave, A - Active cave, S - Spring at or near cave entrance.

Maps: a - this book; b - Anon. 1985b, 1986; c - Rees n.d.; e - Anon. 1986; f - L. Clayton pers. comm.

After Gtover 1978; Anon. 1985; Anggawati 1986; Anon. 1986; Effendi n.d.; D. Owen pers. comm.; and EoS teams.

Figure 8.1. Habitats associated with limestone caves. A - soil and litter in dense forest; B - superficial underground compartment; C - evoluted soils; D - resurgence; E - underground stream; F - pools inside cave; G - sump; H - speleothems, clay, etc.; I - guano; J - river sink; K - river; L - soil and litter in dense forest; M - superficial underground compartment.

After Ko 1986

Mampu Cave is a production of nature, and the various halls and passages exhibit the multitude of beautiful forms with which nature adorns her works; pillars, and shafts, and fretwork, many of the most dazzling white, adorn the roofs or support them, and the ceaseless progress of the work is still going forward and presenting all figures in gradual formation. The top of the cave, here and there fallen in, gives gleams of the most picturesque light, whilst trees and creepers, growing from the fallen masses, shoot up to the level above, and add a charm to the scene. — MUNDY 1848

Caves have potential as tourist areas. For example, part of the Bantimurung Nature Reserve has also been designated a Tourist Park because of its beauty, its geological interest as a classic example of tower karst, its prehistoric interest and its caves. A survey by a member of the Federation of Indonesian Speleological Activities,² however, regarded the known caves as too small, too short and too poor in stalactites and stalagmites to hold any real touristic value (Anggawati 1986). Due at least in part to this, the Ujung Pandang Forestry Office has recommended the area south of Bantimurung-Patunuang as an extension to the Tourist Park. This has some longer caves (up to 375 in) with attractive cave formations. In the future the Bantimurung-Patunuang area may be proposed for National Park status (Anon. 1985a). The newly discovered Salukan Kalang Cave could certainly become an important object of specialist tourism.

The characteristic features of a cave are its definable limits, its enclosed nature, low light levels, and comparative stability of climatic factors such as temperature, relative humidity and air flow (Bullock 1966). Variation in these characteristics between caves creates a surprisingly wide range of habitats which determine the type and number of animals that can inhabit a cave. Habitats associated with caves include soil and litter in limestone forest, the superficial underground compartment,³ cave streams, sump zone, and cave floor habitats (fig. 8.1).

CAVE FORMATION

Cave formation is a subject full of special terminologies and opposing theories (Jennings 1971) but until the geology of the various karst areas in Sulawesi is known in more detail, only the simplest of outlines is given below.

Rainwater contains carbon dioxide absorbed from the atmosphere and is therefore slightly acid (p. 468). This weak acid dissolves calcium carbonate (the main constituent of limestone) and forms channels which, in time, achieve the dimensions of caves, often with a stream running through them.

Two of the commonest features within a cave are stalactites and stalagmites, which together with other cave decorations are known as speleothems. They are columns of calcium carbonate containing various impurities (the cause of the wide range of pale colours found) and are formed by the repeated evaporation of water containing calcium carbonate, leaving thin layers of mineral deposits (fig. 8.2).

Evaporation in caves is very slow because there is no solar radiation to excite the water molecules, air movement is absent or minimal, and the air is virtually saturated with water vapour. At a point 1 km into Salukan Kalang Cave, air temperature at about midday was only about 26°C, relative humidity was 95% (when outside it was about 65%), and there was no detectable wind. In such conditions, the daily evaporation from small pans is 0.044 mm (Dunne and Leopold 1983), and so a column of water only 1 mm deep would take 23 days to evaporate. This explains why speleothems take so long to grow. Precipitation rates are greatly affected by air movement and by impurities in the limestone but the figures above serve to illustrate the slow growth. It has been suggested that small stalactites commonly grow in length by only 0.2 mm per year.

Figure 8.2. The formation of stalactites and stalagmites, a - water evaporates from drips precipitating calcium carbonate and impurities thereby forming a stalactite; b - water evaporates also precipitating calcium carbonate but forming a squatter stalagmite; c - where water drips quickly no stalactite is formed; d - stalactite and stalagmite eventually join to form a single column; e - where the drips from a stalactite fall (or fell) into a river, no stalagmite is formed.

There are many types of speleothems in addition to the well-known stalactites and stalagmites. Water dripping onto the cave floor may form a splattered calcite formation, similar to the shapes formed momentarily when a raindrop falls into thick mud; concentric spheres of calcite may form under special conditions; calcite may precipitate out of water as it flows over cave walls and rocks thereby forming 'frozen waterfalls', often of beautiful colours because of the impurities. In non-active caves eerily-shaped, fragile, hair-thin speleothems called helictites may project from the walls. Lastly, on floors once covered with shallow water, slow evaporation may result in the formation of coral-like spikes, fans, and glistening crystals. Another physical feature of caves which can sometimes be found is 'moon-milk', a soft whitish mass of carbonate minerals some of which are associated with certain species of bacteria (Poulson and White 1969). All speleothems act to increase the surface area of a cave and therefore also the living area available to the cave inhabitants.

Some growths on cave walls, loosely referred to as wall-fungus, are not of mineral origin but are colonies of actinobacteria which look like white or coloured pendants of lichen or fungus. It is possible that these organisms give some caves their characteristic musty smell (Williams and Holland 1967). Growths resembling encrusting lichens found on the wall of a Sumatran cave in 1983 (Whitten et al. 1984), have been identified as an Arthrobacter (Actinobacteria) in association with a Penicillium fungus (H. Reid pers. comm.).

The chemical composition of water seeping into a cave depends on the capacity of the percolating water to dissolve rock and sediment, the rate at which minerals dissolve, and the rate and nature of deposition of calcium carbonate by evaporation. Water seeping into Anak Takun Cave in tower karst in Peninsular Malaysia was collected from 63 sites and the results of analyses show considerable variation. Specific conductance varied from 135 to 1,399 pmho/cm (25°C) and potassium concentration from 0.10 to 26.01 ppm. This variation was determined by whether the water had come into contact with guano (table 8.2). The higher calcium hardness⁴ is probably largely due to the more acid water which passed through old guano dissolving more calcium carbonate (table 8.3) (Crowther 1981).

A unit of tower karst generally has an insufficient catchment area to support permanent streams above the water table. Thus, apart from seasonal or short-term inputs from diffuse-flow seepages, the caves within it are generally dry. As a result enlargement of existing caves may be minimal and they are instead subject to gradual infilling. It appears, however, that limestone may be dissolved below the floor deposits due to renewed aggressiveness (table 8.3) imparted to seepage water after passing through deposits of bat guano (Crowther 1981).

After Crowther 1981

* Aggressiveness is a measure of the amount of calcium carbonate a solution will dissolve.

After Crowther 1981

TEMPERATURE, HUMIDITY AND CARBON DIOXIDE

The insulating role of the walls and roofs of caves effectively buffer the relatively wide daily variations in temperature and humidity of the outside world. Conditions thus remain fairly stable day to day, but there are still seasonal changes which can greatly alter conditions in caves. For example, during a rainy period the humidity and amount of free water within a cave tends to increase.

Air movement is also buffered by the cave walls, but still occurs as air is drawn out of the cave during the day when the air outside is warmer and lighter. This air movement follows a regular pattern, but leaves pockets of stagnant air in deep caves where spiders can weave delicate and complex webs, and preserves pockets of high humidity. Under such stable conditions the minute disturbance of air caused by the approach of predators or prey can be detected.

The constant high humidity in the deeper parts of caves appears to have led many cave invertebrates to become morphologically similar to aquatic arthropods. Many have lost cuticle pigments and wings, while some have developed a thinner cuticle, a larger and more slender body than their relatives outside the cave, adaptations to their feet allowing them to walk on wet surfaces and also a lower metabolic rate (Barr 1968; Howarth 1983). In these deeper areas, the concentration of carbon dioxide increases if there is no inflow of air except from the cave mouth; this was detected in Salukan Kalang Cave. It has been suggested that the lower metabolic rate of some cave invertebrates may be a physiological response to this high carbon dioxide concentration (Howarth 1983).

CHARACTERISTIC ANIMALS AND FOOD CHAINS

Cave animals can be divided into three ecological groups:

• troglobites or obligate cave species unable to survive outside the cave environment;

• troglophiles or facultative species that live and reproduce in caves but that are also found in similar dark, humid microhabitats outside the cave; and

• trogloxenes or species that regularly enter caves for refuge but normally return to the outside environment to feed.

In addition, some other species wander into caves accidentally but cannot survive there (Howarth 1983). It is now realized that some small 'cave-adapted' animals also live in superficial underground compartments even in locations distant from caves.

It used to be thought that there were no troglobites anywhere in Southeast Asia (Leefmans 1930), but exploration of caves in Mt. Mulu National Park, Sarawak, have confirmed the existence of at least 27 species there (together with about 70 species of troglophile) (Holthuis 1979; Roth 1980; Peck 1981; Chapman 1984). A troglobitic crab and a prawn are known from the Mt. Sewu region of Central Java (Holthuis 1984; Ko 1986). Small, white atyid prawns with much reduced eyes found in the deep Salukan Kalang Cave by a APS⁵-EoS team have recently been examined and found to be not just a new species but also the first troglobitic atyid known from Indonesia (L. B. Holthuis pers. comm.).

All cave dwellers are dependent for food on material brought into the cave from outside. Some animals feed on plant roots attached to the cave roof, wood and other material washed in during floods (if the cave has a river running through it), or the organic matter percolating through from the surface. The major providers of food, however, are bats and swiftlets⁶ that roost and breed in the cave but feed outside. Bats and swiftlets supply food in a number of ways:

• during their lives they produce faeces, collectively known as 'guano', which has nutritive value to the various animals that feed on it (coprophages), and is also a source of nourishment to fungi and bacteria;

• as live animals they are hosts to many parasites, both internal and external, and provide food for predators;

• they moult hair and feathers and shed pieces of skin;

• they produce progeny which may be susceptible to different predators and parasites;

• when the bats and swiftlets die, their bodies form a source of food for various corpse-feeding organisms (necrophages).

Almost all animals provide food for others in these five ways, but within the cave ecosystem, in the absence of green plants, these are the only major sources of food. The bats themselves roost on the roof and walls and form the primary basis of one community; their faeces and dead bodies fall to the cave floor and form the basis of another. Thus there is a distinct division of the animals into a roof community and a floor community (Bullock 1966).

EFFECTS OF DARKNESS

The cave environment can be divided into three zones according to the degree of darkness and other physical conditions:

• the twilight zone near the cave entrance where light and temperature vary, and in which a large and varied fauna can be found;

• the middle zone of complete darkness but variable temperature, in which a number of common species live some of which make sorties to the outside; and

• the dark zone of almost constant temperature and complete darkness⁷ in which obligate, cave-adapted species are found (Poulson and White 1969). Since light is essential for photosynthesis, green plants are not found in the dark parts of caves. Plant roots can penetrate fissures leading to a cave and these can commonly be seen attached to or hanging from the cave roof. The most important effect of this virtual total exclusion of green plants is to make all cave dwellers dependent on material brought in from the outside and to exclude all animals that feed directly on the above-ground parts of green plants.

Most animals have a clearly-defined daily cycle of activity, being most active at night (nocturnal), during the day (diurnal), or around dawn and dusk (crepuscular). Such cycles are obviously associated with daylight and darkness and thus might be thought to be absent in a cave. There are, however, certain events which may impose a daily rhythm on cave inhabitants. The most important of these is the departure and subsequent return of the bats and swiftlets. In their absence, food is not available for the free-living ectoparasites in the roosts, and there is a halt to the rain of fresh faeces from the roof (Bullock 1966). In addition, air draughts probably have a daily rhythm caused by the heating and cooling of air outside the cave. Therefore, although data are lacking, it is probable that a daily rhythm does exist in a cave.

The departure of the bats is an event which may continue for two hours after the first bat flies out of the cave. The different timing and pattern of flight activity of bat species in a single cave have been interpreted in different ways, such as avoidance of competition for food, avoidance of predators, and optimizing energy budgets (Fenton et al. 1977; Erkert 1982). It should be stressed, however, that bats are not out of their caves from dusk to dawn, nor are they are necessarily roosting or asleep while in the cave. Instead, even if not disturbed by voices or carelessly-aimed flashlights, individuals or groups can often be seen flying around the caves and there appears to be a constant high-pitched chatter during the day.

In total darkness, a cave-dweller becomes reliant on senses other than sight to detect food or enemies. This is not peculiar to cave animals, however, because many nocturnal and cryptic animals found only above ground and animals living in soil also depend almost exclusively on hearing, smell and touch (Bullock 1966). Some, for instance, have very long appendages, such as the legs of scutigerid centipedes and the antennae of cave crickets (fig. 8.3), although antennae can also function as chemoreceptors and may be sensitive to relative humidity (Howarth 1983). Diurnal animals can live in a cave provided their other senses are sufficiently acute, and sight may even be useful to animals ranging into the twilight areas, and also to those that range outside the caves. The presence or absence of eyes in troglobites is immaterial, since they are restricted to the dark zone. In fact, the lack of evolutionary selection to maintain good sight permits deleterious variation to appear resulting in the total blindness of some species (Holthuis 1979; Roth 1980; Peck 1981). The relative scarcity of food in the deepest parts of caves has led to the few animals that live there being able to withstand long periods of starvation, to gorge when food is available and to store a large amount of fat (Howarth 1983).

Figure 8.3. A cave cricket Diestrammena gravelyi showing its extremely long antennae.

Echo-location

Many bats and swiftlets (p. 553) have developed the ability to echo-locate; that is, a sound is produced and the echoes which reflect back from solid objects are interpreted to give a 'picture' of the surroundings. Echo-location in bats has evolved in at least two ways, but in all cases it is characterized by high frequency sounds (10-200 kHz) which are produced in the larynx or speech-box, mostly far above the threshold of human hearing (15-20 kHz), and by the reception of echoes in complex and often large ears. The principle can be appreciated if you stand close to and facing a wall and then speak; then compare the sound of your voice with the sound when you turn around and speak into the middle of the room. The two sounds are quite different. This ability to sense the proximity of large objects can become quite well-developed in blind people. Rain disturbs echo-locating bats because very humid air absorbs high frequency sounds, and the raindrops confuse the echos received by the bats (p. 426).

Mouse-eared bats (Vespertilionidae)⁸ (fig. 8.4) use a predominantly frequency-modulating (FM) system; that is, the frequency of the sound they emit through their mouth varies and is given in very short pulses. When cruising⁹ in the open a pulse is emitted and some time is spent listening for echoes. When closing in on a flying insect, however, pulses are emitted rapidly so that the exact locations of the prey can be determined. The system used by a flying mouse-eared bat is sufficiently refined for it to detect objects less than 1 mm across.

Figure 8.4. Representatives of the four most common families of insectivorous bats, a - Miniopterus schreibersii (Vespertilionidae), b - Hipposideros diadema (Hipposideridae), c - Rhinolophus arcuatus (Rhinolophidae), d - Megaderma spasma (Megadermatidae).

After Payne et al. 1985

Horseshoe and leaf-nosed bats (Rhinolophidae and Hipposideridae) use mainly a single rather than a variable frequency and each species uses a characteristic frequency. Instead of emitting the sound through their mouth like mouse-eared bats, these bats keep their mouth shut and emit the sounds through their nostrils which are positioned half a wavelength apart to give a stereo impression when the echos are received. The peculiar 'horseshoe' around the nostrils has the function of a megaphone, causing the sound to be emitted in a concentrated beam.

Figure 8.5. Lacewing (Neuroptera) and its ant-lion larva. Lacewings have 'ears' on their wings that can detect the approach of a calling bat.

It used to be thought that bats using echo-location had no difficulty catching their insect prey, but it now appears that some moths can detect bats from 40 m away and before the bat has detected the moth, using 'ears' on their chests, abdomens or mouths. Lacewings, the adults of ant-lions (fig. 8.5), have 'ears' on their wings and those that have been artificially deafened have a 40% greater chance of being caught by a bat than those that can still 'hear' (Fenton 1983). Some moths have developed the ability to utter clicks that confuse the bat, while others have a variety of behavioural responses which make it difficult for the bats to predict their flight pattern. Some bats in their turn do not keep their echo-locating system 'switched-on' continuously so as to give as little warning as possible to the moths, while others emit frequencies above the hearing threshold of moths (Fenton and Fullard 1981; Fenton 1983).

Whereas the bats mentioned above catch flying insects, false vampires (Megadermatidae) feed by picking lizards, frogs and small rodents off the ground, insects off leaves, or fish from the surface of water. They have also been known to eat other bats (Medway 1967). To avoid swamping the echos from their prey they 'whisper' their sounds which are FM like those of mouse-eared bats. False vampires also sometimes hunt like owls and locate their prey solely by homing in on sounds made by the prey itself. This is similar to a frog-eating bat which has been studied in Panama and which can differentiate between the calls of edible and poisonous frogs and between the calls of small frogs and frogs that are too big to capture. Its efficiency at catching frogs has probably led to adaptations in the frogs' calls so that the males still call to attract females but in such a way as to reduce their chances of being caught (Tuttle and Ryan 1981).

The only fruit bats to echo-locate, the cave-dwelling rousette bats Rousettus, use a low-frequency (1.5-5.5 kHz) tongue-click like swiftlets, which is audible to humans and reminiscent of a wooden rattle (Yalden and Morris 1975; Fenton and Fullard 1981). For this bat and the swiftlets, the echos enable them to detect large objects, or rock walls such that they are able to navigate, nest and breed within a totally dark cave but the system is not sufficiently accurate to enable them to catch insects at night (Medway 1969). Rousettus celebensis, at least, appears to learn the route in and out of its cave and does not always use its echo-locating ability. These bats were caught by EoS teams when they collided with the team members in the caves.

It might be thought that the different activity periods of bats and swiftlets would represent temporal partitioning of a common food resource. In fact, swiftlets feed mainly on small wasp-like insects (Hymenoptera) (Medway 1962; Hails and Amiruddin 1981), whereas insectivorous bats concentrate on various moths and beetles (Yalden and Morris 1975; Gaisler 1979; Fenton 1983). Bats do interact ecologically, however, with nocturnal birds (Fenton and Fleming 1976).

Roof Community

The roof community includes bats and swiftlets as well as all those animals that feed on or parasitise them. Over half of the insectivorous bat species and three or four of the fruit bat species probably use caves as permanent or temporary roosts. The large flying foxes Pteropus spp. rarely roost in caves although this is not unknown (Stager and Hall 1983).

Cave-roosting bat species differ in their preference for certain conditions. Some, such as the dawn fruit bat Eonycteris spelaea are found in chambers near to the cave mouth. Some have wings that allow them to maneuver in tight spaces such that they are found roosting in narrow crevices or 'chimneys' (Goodwin 1979). Others, such as the long-fingered bats Miniopterus tend to be found in the dark zone. It is quite common to find at least two species of Miniopterus living in the same cave system or even the same chamber. The species differ little in size and in total darkness they must use olfactory and vocal cues, tactile behaviour and physiological responses to distinguish their own species.

EoS teams visited various caves in the course of collecting information for this book and one, the winding, ascending cave of Konangan, north of Kotamobagu, Bolaang Mongondow, was of particular interest because a nursery of common long-fingered bats M. schreibersii was found. Over a thousand naked pink bats hung so close to each other on a projection from the cave roof that the rock beneath could not be seen. Nursery colonies such as this are known in relatively few species of bats and in Sulawesi they are formed only by long-fingered bats. Pregnant females congregate once each year to give birth at 'traditional' locations after a 4.5-5 month gestation period (Medway 1971), and the nurseries are not necessarily used only by the bats that normally inhabit a particular cave, but may also be used by pregnant females from other caves in the area. In India a nursery of M. schreibersii is known to service an area of 15,000 km² (Hill and Smith 1984). The dense cluster of babies huddle together and this increases the thermoregulatory potential of the group. The mothers roost away from the babies visiting them only to suckle. In M. schreibersii communal care has even extended to the mothers nursing the young indiscriminately. The young may start to fly only a month after birth by which time they are almost the same size as their parents (Krishna and Dominic 1983), but they may continue suckling for a couple of months after this.¹⁰

Little is known about bat predators, but pythons are sometimes seen in or around the cave entrance (Sarasin and Sarasin 1905). The bat hawk Macheiramphus alcinus was recorded for the first time on Sulawesi in 1981 near Lake Lindu; it was previously known from Africa south of the Sahara, Madagascar, the Malay Peninsula, Sumatra, Borneo and southeast New Guinea. It is rarely seen, possibly because it is only active around dawn and dusk, but is quite distinctive with the white throat contrasting with the rest of the plumage which is black, a short crest, long and pointed wings, and a square-cut tail. It is about the size of a male peregrine falcon Falco peregrinus (Bartels 1952; Eccles et al. 1969; Klapste 1982). Owls may occasionally take bats, and this is one possible reason why bats are not very active during the times of full moon. Birds of prey such as owls characteristically evacuate pellets of undigested food such as hair and bones through their mouth. This is often done from a regular roosting or nesting site. Pellets were collected by an EoS team from the floor of the furthest chamber in Mampu Cave and were probably produced by a barn owl Tyto rosenbergi. Careful dissection of the dense pellets revealed the skulls of some of the owl's prey: rice-field rats Rattus argentiventer, shrews Crocidura sp., but only one bat—the relatively large diadem leaf-nosed bat Hipposideros diadema (Boeadi pers. comm.). Of additional interest was part of the leg and claw of what can only have been a young owl, it is known from studies elsewhere that when food supply is limited, the stronger of two young owls in a nest may kill and eat the weaker.

Bats are hosts to parasites, some internal, and many external which bite their host to suck blood. Some, such as the spider-like wingless nycteribiid bat flies¹¹ (Marshall 1971), live almost their entire lives on bats, while others such as streblid bat flies, bed bugs (Cimicidae) and chigger mites (Trombiculidae) spend only part of their life cycle on bats (fig. 8.6). It has been suggested that the tropical bed bug Cimex hemipterus may have begun its association with humans when they used caves as habitual shelters.

Figure 8.6. Common parasites of bats.

After Fenton 1983

The parasitic insects on bats are diverse taxonomically—six families from four orders are represented—but they show considerable convergence in many characteristics. They are generally flattened, some vertically and some horizontally, to ease their movement between the bat hairs. Most of the insects have tough but expandable 'skin' which allows for the consumption of large meals of blood from the host. The skin often bears backward-pointing spines which lessen the chance of being dislodged by a scratching bat. They also have well-developed grasping claws. Wings have been lost in species belonging to all but one of the insect families concerned. The loss of wings and general lack of light in caves have led, not surprisingly, to the loss or reduction of eyes in some of the species (Marshall 1971).

Most species of parasitic insects are found on only one or two closely related host species (table 8.4). For example, each of 28 species of wingless bat flies were found on only a single bat species. Within a species of host, however, not all individuals are necessarily crawling with these parasites. A detailed study of one nycteribiid bat fly on its major host showed that male bats were more often infested than females. Of the males, most were not infested at all and of those that were, most had only one individual parasite. The reason for this low population is not fully understood but it may be because of the relatively low density of bats in the abundant roosting sites (Marshall 1971). There is some evidence to suggest that animals on low-protein diets carry larger populations of ectoparasites (Nelson 1984), and this may be due to the host being less able to scratch the ectoparasites out of its fur.

It is frequently stated that bats carry rabies. This is certainly true in Central and South America but there is only a single report of rabies in a bat from Asia—a dog-faced fruit bat Cynopterus brachyotis in Thailand (Hill and Smith 1984).

After Marshall 1980

Swiftlets

Three species of swiftlets are known from Sulawesi, all of which inhabit caves, though not necessarily exclusively (table 8.5). The moss-nest and white-rumped swiftlets emit rattle sounds indicating that they are able to echo-locate, but the white-bellied swiftlet does not. This last species is found, not surprisingly, relatively close to the cave mouth, and once the birds roost at nightfall they do not move again until dawn. All swiftlets use saliva produced by exceptionally large salivary glands to cement the nest material together, and to anchor the nest firmly on to vertical or overhanging surfaces. The species of swiftlet that uses only saliva to build its nest, a delicacy sought after for Chinese cookery, is not known from Sulawesi.

FLOOR COMMUNITY

The organic matter on the floor of most dry caves is composed largely of material formed from waste products and bodies of animals. Samples of this guano were taken from three caves, and the analyses showed that although the composition varied (table 8.6), there was generally a low level of carbon, a moderate to high concentration of nitrogen, a very low carbonrnitrogen ratio, and an extremely high level of phosphorus. Not surprisingly, local villagers are extracting the guano to sell as fertilizer.

On the cave floor the coprophages and necrophages predominate. It is often difficult to distinguish between them, because whereas a few animals are exclusively necrophagous, many of the coprophages will include dead bats or swiftlets in their diet. The majority of cave-dwelling bats are insect-eating and the faeces they produce are hard and dry and readily exploited by coprophages such as woodlice, caterpillars of Tinea moths which carry a cocoon around with them, flies and beetles, although the primary decomposers are bacteria. The faeces produced by the few species of fruit-eating bats that roost in caves, however, are soft and rich in carbohydrates and not generally utilized by coprophages. In this case, cockroaches ingest the faeces and the general coprophages feed in turn on the faeces of the cockroaches (Doyle 1969). Cockroach density can exceed 100/m² (Ko 1986). Fungi (and some bacteria) digest these faeces and some coprophages such as crickets and small Psocoptera flies also exploit this food resource (McClure et al. 1967).

The bacteria found in caves are in no way unique, but are a selection of the species found outside the cave. It is believed that they may produce antibiotics that exclude fungi and moulds from the foods on which they live. The clayey floor deposits on which the bacteria grow may be eaten by small invertebrates which may in fact be dependent on them. This would explain why certain cave invertebrates are difficult if not impossible to rear away from cave deposits (Poulson and White 1969).

Figure 8.7. The measurement of standard wing length.

After King et al. 1975

* See figure 8.7.

After Hartert 1896; Medway 1966, 1975; Wells 1975

Figure 8.8. Tineid moth, the caterpillars of which carry cocoons around with them and feed on guano.

The floor community includes many predators such as long-legged scutigerid centipedes, assassin bugs Bagauda, and medium to large spiders which feed on the coprophages and the small Tinea moths. Some of these predators live on the walls and wait for wandering coprophages to come to them, or only venture to the ground when hungry. Small predators may also form part of the diet of larger visiting predators such as shrews Crocidura.

Tineid moths are dull-coloured with a wingspan less than 30 mm across and with furry scales on the wings (fig. 8.8). Of the many species known, most are associated with guano although only twenty are recognized as being strictly cave-dwelling.¹² The larvae, which carry cocoons around them, do not eat plants but bat and bird guano as well as dead bats, birds and invertebrates. Where more than one species of tineid moth is found in a cave it is likely that they exploit different types of food (Robinson 1980). For example, it is important to remember that the composition and appearance of the guano produced by the different bats and swiftlets differ markedly, and these different types are generally found in different locations according to the main roosting sites of the bats and birds. In this way, competition between species is to some extent avoided.

Data from EoS teams; L. Clayton pers. comm.

Tail-less whip scorpions (Amblypygi) are common in caves and look dangerous and nightmarish, but the mouthparts do not contain poison glands. The pedipalps (appendages just behind the mouth), end in a simple claw but they look formidable, with long spikes on their inner surface. They are used for seizing and holding prey in a similar fashion to the legs of a praying mantis. The front pair of legs can be mistaken for antennae (which whip scorpions, spiders and related creatures do not possess), for they are long and thin and are usually held in front of the animal. One of these whip scorpions is Stygophrynus (fig. 8.9) which is known from the floor and walls of Mampu Cave where they possibly hunt the large Rhaphidophora crickets (Leefmans 1930).

An examination of the 'stomach' of these crickets revealed pollen and other plant tissue, bits of small insects such as wasps and silverfish (Thysanoptera), and other animal tissue such as bat hair. Some of this could have been eaten if the crickets had ventured outside the cave to feed, but the contents could equally well have come from the bat dung on which the crickets are commonly seen (Leefmans 1930).

Fifty years ago invertebrates were extremely abundant in Mampu Cave. The upper layer of the guano consisted almost entirely of living insects such as click beetles (Elateridae), ground beetles (Carabidae) (fig. 8.10), scarab beetles (Scarabaeidae) (larvae of these live in the guano as do the caterpillars of tineid moths). Also present were at least three species of cockroaches as well as crickets, earwigs, centipedes, millipedes, and even some ants (Leefmans 1930). A French speleological team that visited Mampu Cave in 1986 reported an abundant invertebrate fauna, most of which are dependent on the guano produced by the thousands of bats that roost there.

Members of Operation Drake made a detailed survey of Peda Cave, near Taronggo in Morowali National Park (fig. 8.11). The most conspicuous invertebrates in the dark zone were the very large Rhaphidophora crickets with bodies 50 mm long and antennae 90 mm long, whip scorpions and large hairy spiders. The crickets appeared to consume virtually all the organic debris available and even ate through a wax candle placed in the cave for surveying (Rees n.d.). The cave moth Tinea mirrophlhalma was found (M. Brendell pers. comm.) which was described only recently from specimens collected in the Philippines. This species is noteworthy in that it has reduced eyes, the first tineid to have lost some of its sense of sight. This appears to go against the conventional cavers' wisdom that guano-dependent invertebrates tend to retain the morphology of their terrestrial relatives, and non-guano invertebrates are those that tend to change.

Figure 8.9. Whip scorpion Stygophrynus from Mampu Cave, Bone.

After Leefmans 1930

The various relationships described above may be drawn as a food web (fig. 8.12). As the understanding of a cave ecosystem grows, so its food web becomes more and more complicated. As food webs become more complicated (and it should be remembered that caves probably represent Sulawesi's simplest ecosystem), there is good reason to produce generalized representations or models. The animals at the base of a food web are relatively abundant while those at the top are relatively few in number, with a progressive decrease between the two extremes. This 'pyramid of numbers' is found in ecosystems all over the world and provides a useful means of comparing communities. To construct the pyramid, species are grouped together according to their food habits. Thus all the plants are called primary producers, herbivores are called primary consumers, predators on herbivores are called secondary consumers, predators on secondary consumers are called tertiary consumers, etc. In general, a large number of primary producers support a smaller number of primary consumers which support an even smaller number of secondary consumers supporting one or two tertiary consumers (fig. 8.13a). Variations of the pyramid shape occur when, for instance, a single tree is considered (fig. 8.13b). Even an inverted pyramid can be formed if one considers a single animal, such as a bat, or a plant, which carries a large number of parasites, which are themselves parasitized by an even larger number of hyper-parasites (fig. 8.13c).

Figure 8.10. Carabid ground beetle, one of the animals that feed on guano.

Figure 8.11. Elevation (above) and plan (below) of Peda Cave, Morowali. Crosses indicate guano deposits; the dotted line is the line surveyed for the elevation; cross sections are illustratedat A, B, C and D.

After Rees n.d.

Figure 8.12. Simplified food web of a cave ecosystem.

Figure 8.13. Pyramids of numbers: a - with a large number of primary producers, b - with a single primary producer, c - the case of parasites and hyperparasites.

After Phillipson 1966

The information contained in a pyramid of numbers permits us to state the number of herbivores supported by a certain number of plants and so on. For comparisons between ecosystems, however, a better approach is to use biomass rather than numbers, and to show this as a 'pyramid of biomass' which usually has a similar shape to the pyramid of numbers (Phillipson 1966).

Differences within and Between Caves

It is clear from the descriptions above that although at first sight a cave appears to be a fairly uniform habitat, this is certainly not the case. One of the chief factors causing variation in the cave habitat is the distribution of bats, the producers of guano. Cave maps (figs. 8.14-8.16) show that the occupation of a cave roof by bats is very patchy, and the distributions of guano and invertebrates on the floor reflect this. Mampu Cave comprises nine major chambers all of which, except for Chamber 9, have either swiftlets or bats roosting in them. The total bat population is in excess of 8,000 but these are divided into distinct colonies consisting of between 100 and 4,000 bats. Physical factors also vary from place to place in the cave; for example, during a rainy period, standing water will accumulate in one part of a cave but not in another, and some parts are subject to air movement, while others are not.

Conditions between caves clearly differ far more than conditions within a cave, or even within a single area of limestone, and when comparing caves of broadly similar conditions, the differences are striking. The invertebrates in various Maros caves and in Mampu Cave show considerable differences (table 8.7) as do the species of bats (table 8.8).

Effects of Disturbance

Little is known about the detailed effects of disturbance on caves. At one extreme, it is clear that opening up a cave by mining the limestone and allowing in the sunlight will utterly destroy the specialized cave communities. There is no such thing as regeneration of cave life without rebuilding the cave. Instead, a succession of plants from the limestone flora (p. 474) will colonize the rocks where light newly penetrates. At the other extreme, moderate extraction of old guano need not have catastrophic effects since cave organisms utilize only relatively fresh guano as food. Even so, the resilience of the cave fauna to disturbance is unknown and this sort of exploitation would obviously be better undertaken after an ecological survey and assessment of the possible impacts of different collection techniques and sites.

Bats are sensitive to disturbance and, when caves are visited during daylight hours, strong lights and loud noises should be avoided in the darker chambers. Catching bats requires skill and practice and should not be attempted except with good reason. Scientific collecting of bats should be conducted in moderation and other studies should be planned to cause as little disturbance as possible. Some of the bats that pollinate flowers of commercial fruit trees roost in caves, and if they abandoned a disturbed cave and left the area this could cause considerable financial loss. Some caves are being considered as tourist sites (Anon. 1985a; Anggawati 1986) and unless this form of development is carefully controlled it could have significant negative impacts. For example, Permona Cave near Tentena, Poso, is a National Cultural Reserve, and suffers considerable disturbance from visitors and consequently not many bats are found (L. Clayton pers. comm.).

Figure 8.14. Sketch plan of Mampu Cave, Bone. Dotted lines - approximate dimensions, grey lines - roof openings to the outside, star - 'King of Cave' shrine, arrows - point up slopes, numbers - indicate approximate numbers of roosting bats.

Figure 8.15. Elevation of the central part of Tadula Cave, Kuku, Poso. Note that the vertical scale is exaggerated. Bats: a - 500 Hipposideros cervinus, b - 30 Rousettus amplexicaudatus, c - 1,000 Hipposideros diadema.

After L. Clayton pers. comm.

Figure 8.16. Sketch plan of Permona Cave, Tentena, Poso, a National Cultural Reserve, a - 20 Emballonura monticola, b - 30 Rhinolophus arcuatus and E. monticola.

After L. Clayton pers. comm.

Limestone mining can cause a variety of disturbances, even if it is not immediately adjacent to a cave. For example, the vibrations caused by blasting necessary to break up the rock can cause shock waves that break stalactites and stalagmites, or cause thin cave roofs to collapse. Even so, there are still quite a number of bat species roosting in caves near the three Tonasa limestone and cement factories¹³ but it is not known what species were present before quarrying began. Limestone is vital for construction projects and for the agricultural productivity of the marginal soils of most transmigration settlements, but to avoid unnecessary negative impacts of quarrying, it is important that the blasting be conducted with care and informed understanding.

After L. Deharveng pers. comm.

<

After Hill 1983; Boeadi pers. comm.; H. Dekker pers. comm.