Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Diminishing biotic diversity and increasing vulnerability
Chang is not only responsible for initiating the Green Revolution in rice. Through his work of genetic conservation at the IRRI and as Director of its Germplasm Center, he is also responsible for exerting exceptional efforts to lessen some of the most seriously detrimental consequences of the rapid spread of HYVs of rice. Rice is a self-pollinating plant that readily pollinates across contiguous fields, thus continuously producing new varieties. Over millennia, as rice spread throughout South-East Asia, it developed enormous local variability. In Indonesia alone, it is estimated that there were well over 8,000 traditional cultivars of rice (Bernsten, Siwi and Beachell, 1982: 8). The greater part of this variety (and the even greater range found elsewhere in South-East Asia) was displaced in less than a decade by the new HYVs. From a historical perspective, the dissemination of the new varieties of rice at the end of the 1960s and the beginning of the 1970s was indeed explosive.
Many researchers (Barker, Herdt and Rose, 1985; Dalrymple, 1986; Hargrove, Cabanilla and Coffman, 1985; and Hargrove, Coffman and Cabanilla, 1979, among others) have noted the rapidity with which the new HYVs were adopted by farmers and also by breeders in national breeding centres, where first-release varieties were used as genetic material for developing new, local high-yielding types. Thus, for example, the semi-dwarf variety IR8 was released in 1966. By the late 1960s, it is estimated that approximately 25 per cent of Asia's rice land was planted with IR8 or similar semi-dwarf varieties and about 40 per cent by 1984 (Hargrove, Cabanilla and Coffman, 1985: 3). By 1982, the IRRI estimated that just one variety, IR36, was grown on 11 million hectares of rice land, making it the most widely grown strain of any crop at any time in world history ( IRRI, 1982).
The spread of the new HYVs in Indonesia was certainly as rapid as elsewhere in South-East Asia and possibly more far-reaching in its effects on other rice varieties. Basically, there were four phases to the dissemination process of new strains in Indonesia's rice intensification programme (J. J. Fox, 1991). The first phase from 1967 to 1974 involved three important IRRI varieties derived maternally from the Indonesian variety, Petal These varieties were IR5, IR8 and C4 (the last was developed at the University of the Philippines). Others were bred from these. Two closely related 'sister' varieties, Pelita I and 2, were bred in Indonesia from IR5, IR22 and IR24 (the last two were derived from IR8). By the 1974-5 wet-rice season, approximately 53 per cent of the country's irrigated rice consisted of four varieties: IR5, IR8, C4 and Pelita. Pelita varieties alone accounted for 26 per cent of all irrigated rice. In seven years, Indonesia's rice production increased by over 34 per cent, but it was at this stage in the mid-1970s that the rice programme suffered its first major set-back, a severe outbreak of a hitherto minor pest, the brown plant-hopper (BPH: Nilaparvata lugens Stal.).
This outbreak and a succession of similar BPH infestations that continued for several years through to the late 1970s required a shift to the use of 'pest-resistant' seed varieties in an attempt to protect against planthopper devastation. The second phase of Indonesia's intensification programme involved a search for new pest-resistant varieties. The IRRI was able to release several new varieties of rice-again involving further breeding of Peta-that were resistant to the BPH. Six of these-IR26 and IR30, released in 1975 and IR24, IR28, IR32 and IR34, in 1976-were quickly adopted to replace earlier nonresistant varieties. However, the sequential release of these resistant varieties, all with similar parentage, provided the conditions whereby rapidly developing populations of the BPH could quickly adapt to feed upon them. Their 'resistance' was quickly overcome and Indonesia continued to experience severe BPH outbreaks. Only with the release of IR36 in 1977-also derived from Peta but with a far more complex array of genetic resistance-were the BPH outbreaks of the 1970s overcome. By this time, the 1979-80 wet season, 67 per cent of all irrigated fields were planted with HYVs, and in many areas, this planting was dominated by one particular variety, IR36.
The third phase of Indonesia's intensification programme, through to the mid-1980s, saw the continuing importance of IR36, and also the introduction of locally bred rice varieties that were well-adapted to local conditions. These varieties bore Indonesian names such as Cisadane, Cimandiri, Cipunegara, Sadang and Krueng Aceh. Two of these, Cisadane and Krueng Aceh, surpassed IR36 both in taste and yield potential. Together, these three important and popular varieties-Cisadane, Krueng Aceh and IR36-carried Indonesia to its targeted goal of self-sufficiency in 1984. Between 1968 and 1984, production had grown from 11.38 to 25.93 million tonnes per hectare of milled rice, an increase of over 127 per cent. Shortly after achieving this goal, Indonesia began to experience new outbreaks of the BPH on its widely planted new varieties of rice, such as Cisadane and Krueng Aceh. Separate outbreaks had begun earlier in Sumatra on IR42, a sister variety of IR36; IR36 resistance to the BPH, however, continued to hold.
Indonesia's response to these new, and potentially more threatening, outbreaks involved, as in the past, the release of a new resistant variety of rice, IR64. More importantly, this response was also directed to some of the ecological causes of the problem. By beginning to take account of the ecological factors required to sustain production, this new response marked the next phase in Indonesia's intensification programme, recognizing that these factors meant rethinking the original assumptions that pervaded the 'packet' of technical recommendations disseminated as part of the Green Revolution.
Facing the problems of sustainability
New Genetic Material
The crucial technology of the Green Revolution consisted of important new genetic material: short, stiff-strawed, photoperiod-insensitive rice varieties that tillered profusely and produced heavy panicles of grain. These varieties were highly responsive to nitrogen fertilizers but, unlike taller varieties which tend to fall over as their stalks lengthen, a high proportion of the uptake of nitrogen in the new HYVs went into the production of grain. Because of this capacity to utilize nitrogen, an essential requirement of the technology was increased application of nitrogen fertilizers. Other fertilizers were critical but nitrogen fertilizers were essential.
The 'appropriate' level of application is in fact a complex question and one that has continued to confront farmers, policy makers and scientists alike. The solution varies, depending on whether concern is with affordability, with improved average yields or with maximum yields. Indonesia's national strategy called for maximum yields and thus, from the outset of the programme, it committed itself to a substantial fertilizer subsidy.
Efficiency is a further issue of importance (De Datta, 1987). Since studies have shown that in applying urea on to flooded rice fields, there is an average recovery of only 30 per cent of the fertilizer nitrogen by the rice crop, much research and experimentation has been directed to improving the efficiency of fertilizer management. More generally, the effects of high levels of fertilizer use on the environment are also important. The production of methane from flooded fields, and the more pressing concern with nitrate and phosphorus contamination of rivers, aquifers and surface water are questions to be addressed.
Fertilizer Subsidy
Indonesia's fertilizer subsidy began as a subsidy for nitrogen; but by the end of the 1970s, it had become a general subsidy for all fertilizers used for rice (urea, ammonium sulphate, triple superphosphate and potassium chloride). This subsidy provided rice farmers with all their fertilizer requirements, regardless of their different costs, at a single, relatively low farm-gate price. Generous subsidies, coupled with the government's own high fertilizer recommendations, prompted farmers to adopt application levels that were some of the highest in all of South-East Asia. By the time Indonesia achieved self-sufficiency in the mid-1980s, the cost of this fertilizer subsidy had ballooned to approximately one-third of the entire development budget. This increasing subsidy was itself unsustainable and forced a re-evaluation. In the late 1980s, while still continuing to expand its production, Indonesia began to reduce the levels of its fertilizer subsidy, to create a differential price structure for nitrogen and non-nitrogen fertilizers, and to reduce its official recommendations for the application of fertilizers, especially triple superphosphate. Having urged farmers over the decades to increase their use of fertilizers, the government found it a daunting task to persuade farmers to begin immediately to reduce their application levels.
Pest Control
The third element of the technology of the Green Revolution involved the use of modern pesticides to prevent crop loss and ensure high production. In retrospect, it is possible to see the profound effect of pesticide use on shaping the Green Revolution and development of its essential seed component. Initial introduction and rapid adoption of a small number of closely related HYVs reduced genetic diversity. Genetic uniformity invariably sets the stage for widening crop vulnerability to disease and destruction by insects. The heavy and routine use of pesticides, high applications of nitrogen fertilizers, the closer spacing of plants, and a continuous monocrop culture involving double or triple cropping-all these practices in combination-also increased vulnerability. As the risks of crop loss increased, the use of pesticides escalated. Farmers throughout Indonesia were instructed in the use of pesticides which were often represented as 'medicine' for growing plants. They were advised to conduct spraying at regular intervals during the growing cycle. Pesticides were subsidized by the government at levels that often exceeded that for fertilizers and were included in the various formal packets of the intensification programme. Periodic outbreaks of BPH infestations appeared to confirm for farmers the need for regular pesticide use and, on irrigated rice in particular, this rose steadily.
BPH INFESTATIONS
Indonesia's use of pesticides was in line with that of other countries in Asia. Despite increasing applications, in the late 1960s and through the 1970s, BPH outbreaks occurred not just in Indonesia but throughout South-East Asia, India and China. Why a minor pest of rice that had previously merited no more than a footnote in the standard textbooks should suddenly become the major pest threat to rice in all of Asia became the subject of intensive investigation. Gradually, in the second half of the 1970s and early 1980s, the phenomenon of 'resurgence' began to be understood. The answer lay in understanding the complex balance of predator-prey relationships that had developed and persisted over centuries in the rice-growing environments of Asia. As the evidence began to accumulate, it became clear that the BPH was an extremely vulnerable rice pest. IRRI researchers identified well over 100 predators or parasites that preyed upon the BPH. Its great strength was its ability to breed and spread rapidly. This breeding capacity gave it enormous powers of resurgence, which none of its predators possessed.
The BPH is a fast-breeding invader pest. It has a short generation time, high fertility, enormous tolerance of crowding and tremendous mobility, thanks to its complex development cycle that produces winged progeny every other generation. Lodged in the stalks of rice, the eggs of the BPH survive, sheltered from insecticide spraying; and, as the larvae of these hoppers emerge, they suck the juices of the rice plant. In large numbers, they are capable of sucking a healthy green crop of rice to a withered, 'burnt' brown in the course of a day or two. At severe infestation levels, the spread of the BPH is rapid and relentless, resulting in widespread devastation known as 'hopperburn'.
This pest presents a double danger. It is also the carrier of a virus that causes a disease technically known as 'grassy stunt', which manifests itself in what superficially resembles a normal verdant crop but, in fact, the rice plants do not form grains. The result is an empty harvest. Often, this disease appears in the crop that follows a heavy infestation of the BPH. Another hopper-the green planthopper-is the carrier of a more serious virus known as tungro. Together, the two planthoppers, if not kept in balance, offer formidable threats to rice cultivation.
The introduction and increased use of broad-spectrum insecticides that accompanied the dissemination of high-yielding rice-seed varieties altered predator-prey relationships in sawah environments. Certain insecticides were more damaging to the parasites and predators of the BPH-useful spiders, beetles and dragonflies-than they were to their intended target. In effect, insecticides actually cleared the way for the BPH by destroying its natural enemies, allowing it enormous scope to burst forth, expand and multiply. Sublethal doses of insecticides may even have stimulated female reproduction. Retrospective interpretation of the evidence from the 1970s, based on IRRI research findings, indicates quite clearly that the BPH outbreaks that disrupted Indonesia's intensification programme from 1974 to 1979 were insecticide-induced (Heinrichs and Mochida, 1984: Kenmore et al., 1984). By the mid-1980s, instead of continuing to emphasize the value of pesticides, the IRRI began to insist on the 'judicious' use of such chemicals, to support an increasing variety of research on biological controls, and to recommend the newly rediscovered strategies of integrated pest management (IPM). (4)
By 1979 in Indonesia, with the introduction of reliable resistant varieties such as IR36, the need for insecticides should have receded. Instead, pesticide usage increased dramatically. Subsidies on pesticides, whose real costs were not borne by farmers themselves, tended to promote indiscriminate and injudicious use: and excessive application had the effect of accelerating natural selection by BPH populations to allow them to feed on previously resistant varieties of rice. To a certain extent, it can be said that Indonesia, by its pesticide subsidy, proceeded to finance its subsequent BPH outbreak in the mid-1980s.
Faced with an outbreak situation potentially more threatening than any in the 1970s, President Suharto adopted the advice of his national scientific advisers. On 5 November 1986, he personally issued a presidential decree (INPRES, 3/1986) embodying a series of ecological measures to overcome the spread of the BPH. The most prominent of these measures was a ban-for use on rice-on 57 varieties of insecticides (including the entire spectrum of organo-phosphates) implicated as resurgence-causing agents. Only certain carbamates were allowed to be used in severe outbreak conditions. The decree also recommended the use of a new, highly effective juvenile hormone that prevented the development of the BPH to reproductive maturity. The presidential ban was immediately implemented with remarkable effectiveness for the 1986 7 rainy season.
INTEGRATED PEST MANAGEMENT
More importantly, for the long-term prospects of the Green Revolution, the decree embodied a commitment to a national policy of IPM that replaced regular calendar spraying with a range of biological and cultural controls, plus judicious spraying only after identifiable thresholds of insects were exceeded. With the support of the Food and Agriculture Organization (FAO), Indonesia also committed itself to the training of 2.5 million farmers in the techniques of IPM. Furthermore within two years of the decree, Indonesia reduced and then eliminated all subsidies for pesticides.
From an international ecological standpoint, Suharto's decision in November 1986 is of fundamental importance. Having the previous year been awarded recognition for the success of Indonesia's intensification programme, Suharto took the decision to give priority to biological, as opposed to largely chemical, means of pest management. Indonesia is the first country in the world to take such a major step to protect its primary crop, and its continuing success from 1987 to 1990 in increasing production should have demonstrated to the world that IPM methods are viable and effective.
Rice production and sustainability in the 1990s
In assessing Indonesia's achievements in its rice intensification programme, it is insufficient to call attention only to remarkable yield improvements and spectacular production increases or simply to emphasize the enormous developments that have occurred in agricultural infrastructure and supporting rural institutions. It is also important to point to the ways in which Indonesia has modified the original technology of the Green Revolution. As it entered the 1990s, Indonesia continued to develop its national programme of IPM and was endeavouring in stages, without altering its commitment to higher rice production, to reduce its overuse of fertilizers and the subsidies that promoted such actions. Both of these initiatives, begun in the late 1980s, mark a dramatic change in direction from the policies of earlier intensification efforts.
However, in one vital respect-its genetic base-Indonesia's rice production is more precariously poised than in the early 1970s. In the early 1990s, a smaller set of HYVs accounts for a larger percentage of Indonesia's total rice harvest than at any time in the past. In the 1989-90 rainy season, four IRRI varieties of rice-IR64, IR36, IR42 and IR46-plus two Indonesian-bred varieties-Cisadane and Krueng Aceh-accounted for all sawah rice planted in Indonesia. Three of these varieties-Cisadane, IR64 and IR36-made up over 55 per cent of all irrigated rice. In Java, Indonesia's major riceproducing island, this proportion is even more pronounced: in the 1989-90 rainy season, the top three varieties accounted for 79 per cent of sawah rice planting. All these highly productive varieties are closely related, share the same semi-dwarfing gene and possess the Cina cytoplasm (Chang, Chapter 9; J. J. Fox, 1991; Hargrove, Cabanilla and Coffman, 1985; Hargrove, Coffman and Cabanilla, 1979).
Figure 9.3 provides, in a single consolidated pedigree, a graphic representation of all the major rice varieties (shown in dark outlined boxes) with Cina cytoplasm that have been planted in Indonesia. The figure includes the early 'improved varieties' that preceded the IRRI varieties that were introduced into Indonesia in the late 1960s. Of all the varieties shown on this pedigree, only a few are still planted in the early 1990s. (Indonesia's six principal varieties in 1989-90 are shown in shaded boxes.) Thus, as Indonesia's production has increased, it has come to rely on an ever more slender genetic base. The potential risks from bacterial, viral or pest attack are too great to allow this situation to continue through the 1990s. The varietal basis of Indonesia's rice production must be diversified to provide security for its production.
Indonesia also has special responsibilities in regard to the distinctive javanica-rice varieties that are now found almost exclusively on the islands of Bali and Java. Throughout these two islands, however, these varieties have been substantially displaced by improved HYVs. Yet there are still some villages that, for cultural as well as ecological reasons, cultivate javanica-rice on a viable basis. These villages could be considered 'living gene banks' and should be allowed-indeed encouraged-to continue their cultivation. In West Java, there exist groups of villagers known collectively as kesepuhan who continue to cultivate traditional rice varieties (not just javanica-rice) as part of an ancestral cult (Adimihardja, 1983, 1989). The genetic as well as the cultural importance of these practices should be recognized and be given scope to continue.
Indonesia's primary goal remains that of increasing its rice production to meet a growing demand from an increasing (and more affluent) population. This has been made all the more difficult in 1991 by an El Niño drought year that has seriously affected production. As a predominantly Muslim country, peak consumption periods associated with the feasts of Idul Fitri and Idul El Ad shift in accordance with the Islamic calendar. Beginning in the 1990s, these periods of peak consumption will occur before the principal harvest of the year. This is similar to the situation Indonesia faced in the late 1960s and early 1970s. As a consequence, in 1991, to provide a buffer stock as well as to make up a deficit in production, Indonesia has purchased strategic quantities of rice on the world market.
Despite two years that have shown no substantial growth, the potential for increasing rice production in Indonesia is still considerable. No yield plateau has yet been reached. Increased production in the 1990s is now more likely to occur in steady increments rather than in dramatic jumps. For this reason, it is probably best to cease talk of a Green Revolution. This label may have been appropriate to characterize the early results that followed the introduction of new genetic material, but this revolution has now been thoroughly 'domesticated'.
The stage has thus perhaps been set for the next revolution in rice agriculture which will again involve the introduction of radically new genetic material. Rice scientists at the IRRI now envision the development of a new 'ultra-high yielding variety' of rice. Such a rice strain is still an ideotype-a vision of a breeding target for the future. This rice would be up to 30 per cent more productive than present HYVs, with yields from 13 to 15 tonnes per hectare. Its architecture would be strikingly different with fewer tillers (1-6 tillers compared to present rice plants' 20-25 tillers), but all these tillers would ideally bear panicles. One IRRI breeder's optimistic hope is that such an ultra-high yielding variety could be available for distribution to farmers in 5-8 years (IRRI, 1991).
Were this to occur, this new ultra-high yielding rice would, it is expected, begin a new Green Revolution approximately 30 years after the first. A new cycle would begin, but perhaps this time with a better understanding of the ecological dimensions of rice cultivation. Undoubtedly, the new ultra-high yielding rice will bear some genetic relationship to the HYVs. It is reasonable to expect, however, that the genetic differences between this rice and present rice varieties will be considerable. This new rice will draw on a diverse store of genetic material now preserved in the gene banks at the IRRI and elsewhere. It is therefore likely that the new rice will owe a particular debt to Chang's efforts in preserving such germplasm for future generations.
A final comment is worth making in relation to the Indonesian experience over the past quarter century (1965-90). The simple availability of new technology does not automatically ensure its use, adaptation, and development. Indonesia's success in achieving high production and its gradual adaptation and development of productive rice technology is the result of firm political commitment to a particular development strategy. A reasonable question to ask is whether this same commitment will continue through the 1990s.
Professor Mubyarto of Gadjah Mada University underscored several of the points raised by the principal speakers, including the successful increase in production, its heavy use of water and chemical inputs and, because of the cost of these inputs, dependence on government support. He also stressed the inefficiency of nitrogen use, and then went on:
It is remarkable that the success of rice farmers, to help 'solve' the rice crisis of the 1960s over the region, has been quickly forgotten by urban rice consumers and most government people outside the rice-production system. The result is lack of support when rice farmers face problems, for example, declining terms of trade. In Java, during 1976 88, the rice terms of trade declined by l] per cent, the lowest being in the major riceproducing region of East Java, where the decline was 22 per cent.
While rice farmers have been 'penalized' by their own success, non-rice farmers, notably the rainfed subsistence farmers, have been the hardest hit. Their real incomes have been left behind and the income gap between the two groups of farmers forever widens.
In response to this and other questions, Chang noted that the very strong preference for rice as a food in South-East Asia itself created problems of resource allocation. The HYVs, in particular, are not well-suited to all environments, and the reasons for reluctance to adopt them, in Kalimantan for example, include acid soils and water, as well as the low price of rice, the cost of chemical inputs, and problems with pests. There are, however, some very good adaptations, and J. J. Fox noted one in a Timor valley using limited resources in a triple-cropping succession involving dry crops, irrigated local rice and high-yielding rice; this is a good example of flexible use of the new technology.
There was concern about the loss of biodiversity with the dominant use of new varieties. Over 6,800 Indonesian rice varieties are now preserved at the IRRI, and there are still many unknown varieties on the islands; however, their number is diminishing. The number that has been lost is unknown. Another problem is the effect of pesticides on fish cultivation in the wet-rice fields, formerly an important source of protein. Even in southern Peninsular Malaysia, where the Green Revolution has had limited impact, these fish are now often inedible because of high pesticide residues. In Java, the problem is made worse because pesticides are also heavily used in the upland vegetable gardens, and they contaminate the water running off into the rice lands.
A final problem was stressed by Chang: that of the heavy dependence of Green Revolution farming on government support, whether in the form of materials or subsidies. He painted a picture of 'experts' and bureaucrats, gingerly making their way along the bunds in patent leather shoes, as an integral element in the new rural scene.
10. The problems of upland land management
Introduction
The expansion and intensification of upland
agriculture, 1850-1950
Upland agriculture, 1950-1990: Logging, roads,
markets and cash
The environmental consequences of upland
agricultural expansion: Sustainability and unsustainability
Attempted
solutions
What
is to be done?
BRYANT J. ALLEN
THE problems of upland land degradation in South-East Asia received wide publicity during the 1980s. Scientists and environmental lobby groups, sensitized to world-wide soil erosion, and later to the contributions of tropical land-use change to global climatic warming, have decried the rapid rates of land clearance and deforestation in a number of South-East Asian countries. Some present a picture of looming environmental disaster (for example, Donner, 1987; Eckholm, 1976). While these concerns must be taken seriously, the sometimes highly emotional arguments from the developed world have tended to oversimplify the causes, and also the long-term consequences, of land degradation and deforestation in the Tropics.
Poor cultivation practices almost always result in land degradation and consequent losses of productivity, and this is true in uplands and lowlands alike. In some areas, forest clearance followed by cultivation has resulted in severe environmental damage, but in others forest clearance occurred hundreds of years ago and, at least until recently, the land has been cultivated without serious damage. Moreover, some tracts severely degraded in the past have been rehabilitated with or without state or institutional help. The South-East Asian region provides a rich variety of examples where the causes and consequences of land degradation can be examined. They enable some general statements to be made, and some possible solutions to the problems to be examined.
Explanations and solutions need to be sought in context. Simply to review the situation in each country would be to miss the essential point about the primary cause of events in the uplands of the region. As shall be seen, degradation is widespread, and so too is intensification. In this latter respect, experiences in South-East Asia are common to most agricultural systems throughout the developing world. There is, as Ruthenberg (1980: 35766) shows, a very general tendency to move away from more permanent systems, from less intensive to more intensive practices, towards higher-yielding crops and towards greater use of 'support energy'. The reasons for intensification therefore underlie any discussion of land-use change. It is simplistic to attribute all intensification to population pressure, following Boserup (1965). Other explanations of a perceived need to increase production also have to be taken into account; humans do not live by subsistence alone. This has practical as well as theoretical importance, for even if the population problem is successfully solved, a sustainable future will not have been created if some other force is also driving agriculture to intensify.
Although it is necessary to discuss wider trends in regional agriculture, including those in the lowlands, this chapter concentrates on the problems of the uplands. Their definition as a class of land is, however, difficult. In South-East Asia, the term is often used rather loosely to refer to unirrigated land but, if water is available, it is possible to irrigate almost any land, even steeply sloping land, if someone is willing to pay the costs. Nor is altitude a useful criterion; if 'uplands' imply steeplands, or hill and mountain country, they may begin at sea level. Irrigable land of low relief may be found at over 2 000 metres above sea level. Similarly, slope is not a helpful classifier; some very steep land is found at low altitude, and some almost level land is found in the highest areas occupied by people, now or in the past.
Using Spencer's (1949: 28) definition, 'uplands' could be defined as containing a core of 'hilly to mountainous landscapes of steeply inclined surfaces and the table lands and plateaus Iying at higher elevations'. It might be added that the discussion concerns land which is not flood-irrigated, not the immediate coastal fringe, estuarine or alluvial plains and swampland, nor is it seasonally flooded. Broadly, this definition by exclusion is followed in this chapter. Uplanders and lowlanders distinguish themselves as different groups of people in several of the South-East Asian countries, the one class of persons having a generic name for the other. It would be desirable to take account of this perceived basis of classification also, but in practical terms it is not feasible.
The expansion and intensification of upland agriculture, 1850-1950
The Antecedents of Modern Upland Settlement
Forest clearing for agriculture in the South-East Asian uplands began over 5,000 years ago. Pollen preserved in swamps in the high-elevation interior of Papua New Guinea provides evidence that substantial forest disturbance first occurred there between 5,000 and 4,000 years BP (Before Present) and was well established by 2,300 BP (Walker and Flenley, 1979). In Central Taiwan and Sumatra, the earliest forest disturbance is probably of a similar age (Hutterer, 1983a). While there is no direct archaeological or palynological evidence that the disturbance was for the purpose of agriculture, there are few other causes which can be sensibly ascribed. Agriculture is inferred from 5,000 years BP on Taiwan and from 3,000 years BP in Sulawesi. In South China and northern Thailand, the beginning of rice cultivation is placed at 3,000 years BP, and irrigated rice, from 1,000 years BP (Bellwood, 1978: 161).
It would probably be a mistake to assume that this very early agriculture was restricted only to a form of shifting cultivation on upland sites. There is evidence from a well-researched site in the central highlands of Papua New Guinea, at an altitude of 1 500 1 600 metres, that swamp drainage was being employed prior to 5,000 years BP, probably in conjunction with shifting cultivation on slopes around the swamp (Golson, 1981). From this same site comes indications of soil washing off the slopes around the swamp and sealing in some features of the drainage systems.
Elsewhere in the Papua New Guinea highlands, studies of the sediments in a number of small lakes provide evidence of sharply increased deposition from 300 years BP, when it is assumed that an expansion of agriculture occurred within the lake catchments. From its earliest beginnings, upland agriculture was the cause of accelerated soil erosion, yet much of the land around these sites is still occupied, and some of it intensively cultivated, today.
A Period of Revolutionary Change
Despite these early agricultural beginnings, the period of 'revolutionary' change in SouthEast Asian agriculture occurred between 1850 and 1950, sometimes, but not everywhere, preceded by a century or half-century of much lesser change. This applies to most of the uplands as well as the lowlands. In the space of about 100 years, millions of hectares of formerly seasonally inundated ground in the lowlands were converted into wet-rice fields; the forest land which surrounded them was developed into commercial plantations and permanent mixed-crop, dry-field cultivations by smallholders. Large areas of hill forest were converted into swiddens for upland rice (Dobby, 1955). The organized states of precolonial South-East Asia were based on irrigated rice, but the expansion of wet-rice areas as a result of massive public works, drainage and canal building under colonial administrations during the 100 years from 1850 was on a vastly larger scale. This transformation has been described as the 'first Green Revolution, (Barker and Herdt, 1979: 8).
Uhlig (1984) provides data which testify to this amazing transformation of the SouthEast Asian landscape. In Burma's Pegu province, wet-rice areas increased from 228 000 hectares in 1855 to 1.25 million hectares in 1880 and the national total from 1.6 million in 1860 to 5 million hectares in 1931. Wet-rice areas in Thailand grew from 1.5 million hectares in 1907 to 6.9 million hectares in 1927 and 8.6 million hectares in 1979. In the Philippines, the big expansion began only after the American colonization. There, wetrice cultivation increased from about 1.4 million hectares in 1931 to 2.6 million hectares in 1968. The massive changes brought about by the drainage and irrigation works by the Dutch in Indonesia are well-known (Boeke, 1953). However, Palte (1989: 39-40) provides graphic description of the earlier expansion of irrigated rice fields (sawah) on the Panarukan plain in Java between 1805 and 1845.
The relationship between population and food supply is a complex one and this is not the place to argue the various cases for cause and effect. Nevertheless, during 1850 1950, South-East Asian populations clearly increased beyond the capacity of the land then under cultivation to satisfy the demand for food production. The response of populations all over the region was to move on to previously unworked or little-worked ground. In Java, the majority of upland areas were not brought into production 'until the possibilities for irrigation were exhausted, (Palte, 1989: 40). The pattern of occupation was influenced by social and political factors, but was primarily a consequence of growing population pressure on the available land and water resources in the lowlands. Penetration into previously uncultivated uplands began in the 1820s, largely around the Yogyakarta plain where heavy taxes and labour duties were imposed, and a war disrupted farming on the plain.
The most important expansion into the uplands in Java took place between 1860 and 1925, when the forced cultivation system gave way to taxation, and roads, railways and tramways opened up new regions. Large numbers of people took the opportunity to escape onerous taxes and corvée labour by illegally occupying and cutting upland forests for agriculture. The colonial government contributed by clearing large areas of forest for coffee plantations, and then ceding unused portions to settlers.
The appearance of coffee rust, which devastated much of Indonesia's coffee estates, also led to the abandonment of upland areas planted with coffee and to their occupation by subsistence farmers. Concerns over watershed protection, and widespread destruction of forest, resulted in severe government discouragement of upland occupation, but a third phase of expansion occurred from 1942 to 1950 when control over forest land was disrupted by the Japanese occupation and the national struggle for independence (Nibbering, 1991b: 25; Palte, 1989: 48-9).
A different picture can be assembled from Thailand, although the outcome is broadly similar. Between 1855 and 1930, Thai rice exports increased 25 times while, over the same period, the Thai population, dependent on rice as a staple, grew to 12 million. Rice yields per unit area declined, however, and the inevitable outcome was a rapid increase in the cultivated area. More than half of this occurred on the Central Plain, where canal construction was funded from stat! revenues (Johnson, 1975, quoted in Hirsch, 1990). From 1912 to World War 11, the rice trade suffered a number of reversals, between which growth was slow. Hirsch (1990) and Uhlig (1984) imply that it was the failure of the rice-export trade to grow, and a combination of pricing policies, marketing structures and ecological imperatives, which led to a sudden expansion in upland agriculture and an increase in the importance of upland crops relative to rice. Production increased, not as a result of successful intensification or crop diversification in the rice areas, but because land was available in the uplands and markets were available for the crops which could be grown there-maize, cassava and sugar in particular (Uhlig, 1984: 125).
Expansion before Markets
Elsewhere in the region, expansion of upland agriculture occurred during this period (1850-1950) in the absence of markets or the state. In the Papua New Guinea highlands, growth began more than a century before 1850. There is abundant evidence that existing agricultural systems have intensified and expanded during the last 300-400 years (see, for example, Strathern, 1982). The widespread adoption of sweet potato (Ipomoea batatas), a previously unknown or little-used crop, allowed agriculture to be extended to an altitude of 2800 metres, well above the economic altitudinal range of likely previous staples such as taro (Colocasia esculenta). In addition, land which could no longer support taro because of declining soil fertility was brought back into cultivation under sweet potato (Clarke, 1977). It has been argued that the driving force behind these changes was increasing population numbers, increasing social and political differentiation, and the need to produce more pigs, which are used as a means of indemnification in social relations and conflicts between groups (Allen and Crittenden, 1987; Modjeska, 1982). Similar arguments have been advanced to explain the very rapid expansion of precolonial upland agriculture in Hawaii (Kirch, 1984).
The Philippines present yet another story. Many upland parts of the country were occupied by tribal groups prior to Spanish colonization in the 1500s. The Digos-Padada valley in Davao province, Mindanao, for example, was occupied by four major tribal groups, with smaller groups of Muslims on the coast (Simkins and Wernstedt, 1971). When one of the tribal groups came into conflict with the Spanish government in the nineteenth century and was dispersed, Christians from elsewhere in the Philippines began to occupy the abandoned land. Fields cleared in lower hill-slope forest were quickly converted to grass, but more forest was cleared inland and population densities remained low. Migration into the area continued, however, and between 1903 and 1918 the population increased by 65 per cent, then doubled between 1918 and 1939. Most settlers were concentrated in the lower valley, leaving the interior largely in the hands of tribal groups. After World War 11, however, the settlement frontier moved inland rapidly and by 1965 there was almost no land in the valley that was unoccupied or uncultivated.
Although upland agriculture has expanded all over the South-East Asian region, and increasing population has everywhere been associated with this expansion, it cannot, in many cases, be said to be the only cause. The reasons are varied and complex and it is not possible to make many general statements about the process. In Thailand, and perhaps in Papua New Guinea, intensification and expansion occurred before critical population pressures arose, and then the need to increase production for social and political reasons was as important as population growth. Even the Indonesian case does not stand on the basis of population pressure alone. Government policies, land taxes, compulsory labour, war and an increased access to upland areas, all contributed to the expansion of upland agriculture in Java. Elsewhere, as in the Philippines, migrant populations at relatively low densities have caused considerable land degradation, but they have moved inland into uncultivated forest, rather than intensify production on existing fields. Population increases are relevant in all cases, but sometimes as a cause, and sometimes as an effect, of agricultural expansion and intensification.
