Deceptive Appearances: Environment, Demography, and Agricultural Change

The hamlet cluster on which this study is based is called Huta Ginjang, 'top village.' It lies some 3 degrees north of the Equator, perched about two miles up the north side of Pusuk Buhit, 'Mount Navel,' a dormant volcano that connects Samosir 'island' (actually a vast peninsula in Lake Toba) to the body of Sumatra (see Map 2). The nearly 600 inhabitants of the village share a single spring for all water they use domestically.

Generally the climate of Sumatra is characterized as being of "equatorial monsoon" type. Samosir has the most pronounced dry season of all Sumatra, since it lies in a highland crater, from which the surrounding landmass slopes down to the encircling seas (see Map 1). In the dry season, the winds blow in, often with incredible and repeated gusts, from the southwest and from the west. They may start as early as April and continue well into October. Beginning in July, the wind often blows constantly, day and night, at some 50 miles an hour, with gusts up to 80 miles an hour.

At times one gazes out over the parched expanse of the northern half of Samosir 'island' and imagines that if only its clayey whiteness were a dark tree-green, the misty clouds that hover around the lake at the edges of the crater would somehow yield a bounteous wet rain rather than the mistlike drizzle that never penetrates the ground more than an eighth of an inch. After a three-week blow, the wind lets up and gives way to scorching hot, clear days with only a rare puff of cloud in sight. Then, just as it seems that enough of these rare puffs are appearing to give rise to a real thunderhead or two, the wind starts up again, and what had seemed to be clouds reveal themselves to have been apparitions dissipating into the mist-bearing collar that sits, apparently immobile, hugging the upper edges of the lake crater in uncanny defiance of the wind. The sky turns white, and another three-week blow begins. So it goes, alternately, from May through September, primarily between June and August.

The seasonal oscillation of rainfall recorded between 1922 and 1928 by the Dutch administrators at Pangururan, the town at the point where the mountain and island are joined, can be seen in Table 1.1. Whereas in the dry season, during the stretches of overcast gale winds, the sun often does not break through the cloud cover, much of the "rainy season" is actually sunny for most of the morning and afternoon, except for an hour's downpour toward the evening. Parts of the rainy season are, then, much hotter than parts of the dry season.

From its vantage point halfway up the mountain, Huta Ginjang overlooks the north end of Samosir and, to the west, the raised, stream-fed bowl of Sagala Valley (invisible from the lake or from Samosir Peninsula), that has one bottleneck egress lying between the mountain and the lake-crater wall that encircles both. In 1933, J. C. Vergouwen referred to this area as the "out-of-the-way mountain territory of Sagala." Numerous other isolated, well-watered valleys nestle in the crevices of the vast crater enclosing the lake. What makes for Sagala's apparent isolation is that it is not a thoroughfare and is divided from the lake by a long-solidified shoulder of the mountain. To the south, it is cut off from its proper extension, Limbong Valley, by another shoulder of Pusuk Buhit. (Two of the grandsons of the first man, Si Raja Batak, were named Limbong Mulana and Sagala Raja.)

Hundreds of tourists each year pass within 5 kilometers of the village on "round-the-island" diesel steamer tours and do not see it. Its existence is signaled by what appears to be a tiny clump of trees in the vast, seemingly barren and uninhabitable expanse of brush and savannah that cloaks the mountain. A keen eye may note the darker brown of cultivated areas of level stone terraces, an occasional puff of smoke from some distant shoulder (the sign of a raging savannah fire), or the deeper green that sets off ripening rice in scattered, minimally terraced patches high on the mountain. But even the ubiquitous terraces of the mountain's middle and lower reaches are normally hidden from sight by brush and Imperata grass cover, at least until a boat reaches the canal at Pangururan and passes through the land bridge that connects the mountain to the Samosir peninsula. There, the terraces covering the lower half (as high as the eye can see) are in almost constant use, and at seasons when most are planted, the mountainside resembles a wall of stone.

Water and Settlement Patterns

Sianjurmula-mula, the purported site of the original Batak hamlet and home of the first man, Si Raja Batak, is at the southern end of Sagala Valley, at the foot of the spur connecting the mountain to the high plateau that extends west from the lip of the crater. It was on our way to visit it that my wife, Hedy, and I happened on the village of Huta Ginjang.

Ultimately, the existence of this village may be attributed to the fortuitous occurrence of a spring on the northwest flank of the mountain. Water is carried to the houses throughout the year by women and girls, who imitate the carrying of buckets from the time they learn to walk. Villagers bathe, wash clothes, and get water for household use at a divided bathing place with three pipes below a small holding tank. Two pipes go into the women's compartment and one into the men's, since some women and girls need water for domestic purposes while others are bathing or washing clothes. No water is piped for household use to any of the hamlets because many are afraid that the flow would not be sufficient to meet the needs of washing, etc., and they therefore refuse to invest in pipe. In the dry season the downstream bed of the brook is parched for months, because all the "excess" water is diverted for irrigation.

The area on the southern flank of the mountain contrasts with Huta Ginjang on its north. It was never even settled in the traditional walled-hamlet pattern, because of the lack of year-round springs. Some 200 scattered houses built there in the recent past have rectangular concrete water tanks with a capacity of 2,000 to 3,000 gallons of water fed by gutters that run around the edge of the roofs.

Until this century, Sumatra's coastal lowlands had much lower population densities than the interior highlands. In 1824, on a journey to what is now the site of the regional capital, Tarutung, the missionaries Burton and Ward, after scaling the escarpment on the Indian Ocean side of Sumatra, were astounded to find a long, broad valley of wet-rice terraces between grass-covered hills.

Why were population densities greater in the highlands? In addition to Batak fear that the spirits thought to cause tropical diseases are more prevalent in the lowland areas and to their dislike of the heat and humidity of those areas, it may be that over the course of many centuries the denizens of the coastal areas grew averse to their exposure to the levying of taxes and demands of fealty by coastal/riverine patrols emanating from Srivijaya and other maritime powers that controlled the sea-lanes. A trend to move upland, away from the coast, may have developed. Lower temperatures and less rain in the mountains would also have favored the deforestation so necessary to high population densities, because forest would not retake clearings as quickly as under moister conditions. In particular, the pronounced dry season of the Samosir region would have inhibited regrowth of secondary forest, and lake fish would have helped assure settlers a reliable source of protein. At the same time, the perennial springs could be exploited for irrigating rice.

Dutch data on Samosir population, based on the counts of village headmen who were expected to collect and pay taxes for those under their jurisdiction, might be underreported. In 1913 there were 78,169 Samosirese (40,187 males and 37,982 females) in 2,236 hamlets (Middendorp 1913:13). By 1931, the number had risen to 96,267. The chief administrator (Dutch, Controleur) at the time remarked that "emigration is hardly taking place" (van Bemmelen 1931:13-14). Although much out-migration has since occurred, the Indonesian government reported the 1980 population as 126,696. These and other data are shown in terms of density in Table 1.2.

The density of over 500 persons per square mile far exceeds the densities reported for forest shifting cultivators (Conklin [1957:10] reports a range of 65-91 for the Hanunóo; Freeman [1970:133], a range of 52-65 for the Iban; Dove [1985:381], 30 persons per square mile for the Kantu'; Hanks [1972:57], an average of 31 persons per square mile based on six cases). However, from the area of fallow terraces one sees in Samosir, the population may have been even higher at times in the precolonial past. This would explain the exaggerated claim, often made by educated, city Batak, that Samosir, or their lake-region homeland generally, has been deserted by all but the very old and the very young. The notion of near-total abandonment of the Toba homeland has a strong and pernicious hold, perhaps because quite a number of young men become emigrants for varying lengths of time. Yet, as will be shown, in spite of adverse climatic and other environmental factors, the people maintain a firm foothold and finance a variety of means of more or less successful out-migration in agriculture, trade, and education.

Illusions of Timelessness in the Landscape

Most accounts of Batak agriculture highlight the feats of engineering of their irrigation systems, but mention of the equally impressive, rain-dependent, dry-field terracing, so much in evidence in the Samosir area, is rare. All around Mount Pusuk Buhit and in the valleys on the west, north, and east sides of the lake, as on the east side of Samosir, the earth on which terraces were built appears to be more than half stone, much of it boulders. A tremendous amount of effort and skill was required to transform the terrain into the humanized landscape it now presents to the eye.

Contemporary villagers remain expert rock-wall builders. Steepness of slope never appears to prevent terracing, either for irrigated fields or for dry. In long-fallow areas, walls are completely dismantled when a field is reworked. They are also built around some fields next to cattle paths, and in the past walls were built between certain clan areas. Running up the western side of Mount Pusuk Buhit, from the point at which the Sagala and Limbong Valleys come closest to joining, is a wall of boulders, surrounded on both sides by deep drainage ditches. Its purpose was said to have been not to keep cattle in but to prevent attacks by horse-mounted riders!

It is easy to imagine that, except for terracing, walls, and irrigation ditches, the landscape of the Toba Lake crater is the product of traditional agriculture in conjunction with natural causes and to assume that where the earth appears scarred by erosion, the terrain has been exploited in a prodigal manner because it is "so deforested." But the appearance of parts of the area has other causes, including mining of sulfur, of stone for road-building, and even of earth for production of china.

Another example of the difficulties of taking the landscape of this area to be comprehensible to the naked eye involves what may be termed "reclamation." D. van der Meulen writes that in 1915 or so, when he was an administrator in the area, Dutch engineers undertook to create additional lakeside fields: "The Batu Bongbong [Hindering Stone], the threshold in the Asahan River, just at its outlet from the lake [of which it is the sole drain], was flattened with dynamite, by which the lake surface was substantially lowered" (van der Meulen 1977:57).

For administration this meant ... a sudden aggravation of civil law. Wholesale quarrels over land rights [in newly exposed lakeside fields] arose and if the administrators delayed with their decisions then men fought it out. Bataks ... stood there generally up to their knees in the thick mud. We were carried out on crude sedan-chairs to a higher spot. (ibid.:55-56)

A further instance of introduced change was the Dutch program of reforestation, which has been continued by the Indonesian government. The reports of Dutch Controleurs invariably refer to the fact that very little of the primary forest is left on Samosir and discuss the measures taken to reforest the upper reaches of the peninsula. Middendorp (1913:8-9) noted that the "alang-alang [Imperata] fields which are set ablaze every year" are usually so treated "for obtaining pasture for livestock which find fodder in the new-fired grass, whereas the Bataks also have the conviction that many fires are beneficial ... keeping [animal pests] and insects in a standoff." But these just observations had little effect on later policy. Large areas of the upper reaches of the Samosir peninsula and the east side of the lake were planted in forest, and the burning of pastures was banned, to their long-term detriment. There was a purported need for forest cover to induce more rain. Apparently unheeded was a report by J. van Breda de Haan on rainfall in plantation areas of the East Coast of Sumatra (1898), "largely devoted to the presentation of data showing that there is no demonstrable correlation between deforestation and rainfall" (Bartlett 1957:191). Further justification for forest planting was that it would increase the supply of wood for construction of dwellings (Haibach 1927:11), but the forests that were planted were not of locally used hardwood trees. Instead of providing a local source of lumber, they have become government reserves by default, reducing the available pasture area without benefitting the inhabitants, since only those with government contracts are permitted to cut the trees.

Dutch Agricultural Initiatives

The broadly conceived Dutch plans for altering the landscape essentially arose because agriculture was a central concern of administration. Middendorp listed rice, corn, and sweet potatoes in decreasing order of importance for subsistence. Rice was grown in "dry fields on an incline," rainfed terraces, and irrigable fields. According to him, at that time, "agriculture, the main means of livelihood," was "not sufficient to provide the subsistence needs of the people" (Middendorp 1913:19). We can trace the vagaries of the supply and demand for staples in several other Controleurs' reports. The next one available noted that in 1927 it was "no longer possible to speak of a shortage of food during the past years. If the weather conditions do not bring on a bad harvest people have enough for their daily sustenance:"

In 1922 rice worth 13,000 guilders was still imported. This amount got smaller with the years and in the past year [1927] was still a little above 5000 guilders. But a similar amount was exported from the neighboring areas of rice surplus, Limbong and Sagala, to the East Coast. (Haibach 1927:21)

A few years later, van Bemmelen claimed that the rice crop had to be supplemented by a large import from Porsea (at the south end of the lake) and that "due to lack of rice, a great part of the population subsists on corn and sweet potatoes" (1931:15).

I will come back to the point regarding people "subsisting on corn and sweet potatoes" below and in Chapter 6, where we examine contemporary self-sufficiency in rice in Huta Ginjang. Here I want to make the point that even before they had a clear grasp of the agricultural system and its vicissitudes, the Dutch were actively trying to alter that system as a complement to their concern with subsistence by introducing commercial export crops, as well as by other means to be described in the next section. Seven years after the outset of their regime, Middendorp noted:

Peanuts, first introduced by Controleur Stap in 1907 or 1908 grow very well in Samosir, so that [already] a small export goes on from Onan Runggu to Balige....

Coffee is much planted ... and in 1911-1913 the amount rose to 200,961. ... I bid coffee adat ['custom'] to be complied with by the [appointed area heads] ... namely that each household have in the ground 100 trees, all hamlet headmen 150, all village headmen 250, all raja padua 300, and all Djaihutan 500. The upland fallow suitable lands were then taken and divided up with generally mutual agreement. (Middendorp 1913:22) This passage shows both indigenous aptitudes and the character that Dutch administration often assumed. So-called coffee adat was pure invention, as were all the ranks except "hamlet headman."