Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
2. Two types of swidden systems
2.1. The ''established'' swiddeners
2.2. The''pioneer'' swiddeners
As pointed out earlier, there are two basic types of integral swidden systems in Northern Thailand today. They correspond closely to Conklin's two basic types of "pioneer" and "established'' swiddening (Conklin 1957, p. 3; Walker 1975, pp. 5-6). In Thai language they are called rai luen loi and rai muun wian respectively. The two adaptations vary in detail from place to place and from group to group, but general features of each type can be discerned and will be discussed in this chapter. Those who practice the "pioneer" type include Hmong, Lisu, Lahu, Akha, and Yao. All are SinoTibetan-speaking peoples, most of whom have migrated into North Thailand in the past one hundred years or so. The "established" swiddeners include Karen, Khmu, Htin, and Lawa. All except the Karen are Mon-Khmer groups who have probably been in Southeast Asia for millennia. The Karen, although generally classed as Sino-Tibetan, have lived in Northern Thailand for probably several hundred years, perhaps much longer, and have lived in nearby Burma for many centuries.
The two swidden adaptations today relate heavily to the perceptions (norms, models, ideals, etc.) of their operators as to how swiddening should be practiced. Thus, although the''pioneer" adaptation is one that ''under ideal conditions [involves the clearing of] substantial portions of climax forest each year" (Walker 1975, pp. 5-6), it nevertheless involves conceptions and practices that are attempted even though there is little swiddenable climax forest remaining in Northern Thailand today. To understand the reasons for and logic of the ''pioneer" adaptation, then, it will also be necessary to analyse it in relation to the farming of primary forest where the system was institutionalized. The practice has involved considerable village movement to follow primary forest. The ''established'' swiddeners use a rather different form of swiddening. Their villages are located in relatively fixed territories where they farm secondary forest in a "rotational'' manner. Both adaptations have undergone considerable change over time and indeed are still changing. The practice of pioneer swiddening, for example, has changed with changing environmental conditions (namely, the disappearance of primary forest). This chapter will discuss the two adaptations, both as they are today and, where evidence is available, in terms of the changes they have undergone. Because of differences in the two types of swidden systems, and because of related changes that have occurred in both society and environment in relation to these systems, it will be seen that any consideration of "development'' of swidden agriculture in Northern Thailand will have to take cognizance of these differences.
2.1. The ''established'' swiddeners
2.1.1. Principal natural resources
2.1.2. Principal resources produced
2.1.3. The transformation process
2.1.4. Ecological and socio-cultural constraints
The "established" swiddeners in Northern Thailand number about 200,000 of a total integral swiddener population of about 290,000. Of these, the Karen alone form the vast majority, numbering about 1 62,000, or 81 per cent of the established swiddeners and 56 per cent of all integral swiddeners (TRC 1974). These figures include only hilldwelling populations. Additionally, some of the smaller groups are undergoing acculturation. Many of the Lawa, for example, now practice permanent-field farming, speak Thai, and are considered by Thai authorities virtually to be Thai already. The discussion to follow, then, will draw heavily on the Karen agricultural system, although general features of the system can be applied to other "established" swiddeners as well.
It is possible that the Karen, probably unlike the other established swiddeners, have only relatively recently adopted established, secondary forest rotational swiddening in Northern Thailand. There is some evidence that before the mid-nineteenth century many Karen were still rather mobile and preferred to farm virgin or highly regenerated forest.5 If this is true, there must have been a gradual shift in the direction of established swiddening. By around the turn of the century perhaps over half the Karen villages had made this shift (cf. Hallet 1890, pp. 36-37; Graham 1924, p.176). Today, established swiddening is well institutionalized in virtually all Karen swidden villages, although one researcher suspects that there are still several nomadic Karen villages left "in the deep forests of the western borderland of Mae Sarieng district" (lijima 1970).
Whatever the case, it is clear that the Karen, as well as other established swiddeners in Northern Thailand, are now fully committed to the established, secondary forest pattern-that is, their population densities and the degree to which farming practices have been institutionalized prevent their being able to pursue a mobile pattern such as that practiced by pioneer swiddeners. The reasons for this should become increasingly clear as the discussion here progresses. The discussion will cover the following topics: (1) principal natural resources in the system, (2) principal resources produced, (3) the transformation process from (1) to (2) (i.e., the "swidden year"), and (4) general ecological and socio-cultural characteristics of and constraints on the overall system.
2.1.1. Principal natural resources
Like all swiddeners, Karen are able to meet their needs by interrupting a stable ecosystem and temporarily turning it into a pulsing one, using the power generated in the pulse to produce for man. This contrasts with permanent-field agricultural systems established on prairies or grasslands which use the same seasonal restart mechanisms as are found in the natural ecosystem (Odum 1971, pp. 11 2-11 5). In many ways, however, swiddening is "imitative" of the tropical forest ecosystem and more ecologically conservative than permanent-field farming (see Geertz 1963; Reed 1965). Typically, tropical forests contain a larger amount of nutrients in their living biomass than in their relatively impoverished soils. The swiddener taps this resource by cutting the forest, allowing the debris to dry out, and burning it to release nutrients into the soil from the ash produced. Crops are then grown utilizing these nutrients, but the swiddener must sooner or later move to a new field site as the supply of nutrients diminishes. If weeds are a problem, he may have to find a new site even sooner.
The established swiddeners in Northern Thailand, since they were there before other swiddeners moved into the area, have generally staked out the best forests in which to operate this process. These are primarily the better stands of Tropical Evergreen Forest, which have the greatest biomass, the greatest species diversity, and the greatest moisture content of all Northern Thailand forests (Ogawa et al. 1961).
As population has expanded, however, portions of Hill Evergreen Forest have also been taken under cultivation. Very moist pockets (Moist Tropical Evergreen Forest) are not taken under cultivation, however, but serve as game and water reserves and sources of useful forest products. Farmed conservationally, these forests provide the Karen with rapid regeneration of secondary forest so that sites may be farmed relatively frequently for small periods at a time. Since it is a secondary forest system, regeneration does not reach the same proportions of biomass and nutrient availability as highly regenerated forest, but this is compensated for by being able to use the land again after relatively short fallows.
In Northern Thailand, Tropical Evergreen Forest is commonly found on northern or eastern slopes. At higher altitudes (around 1,000 m elevation) it tends to blend into Hill Evergreen Forest, to which it is closely related (Kuchler and Sawyer 1967, p. 318). It is a richer forest than would normally be expected at this latitude, apparently because of locally favourable soil moisture (Ogawa et al. 1961, p. 11 0). In a mature state it is a triple layer forest (not counting ground cover) and contains the tallest trees. Characteristic species among the largest trees include Dipterocarpus costatus and Dipterocarpus alatus, but no one species dominates, except perhaps D. costatus in places (Kuchler and Sawyer 1967, pp. 316-317).
2.1.2. Principal resources produced
Swidden systems commonly produce an incredible variety of agricultural products. Hundreds of different species may be grown. The complementarily of swidden plant assemblages is ecologically sound, is valuable in getting the most use from the swidden with the least effort, promotes forest regeneration, and offers the swiddener a large degree of market independence by meeting the vast majority of his nutritional and other needs. Hinton (1970, p. 4), for example, is of the opinion that the Karen swiddener can provide for almost all his essential needs except salt and iron.
Rice is the main crop produced by secondary forest swiddeners in Northern Thailand. Also grown in the rice swiddens are numerous other crops. Still others are customarily grown in household gardens or in small separate plots. Among the main swidden crops (other than rice) mustard greens and corn are probably the most important. Karen and Lawa swiddeners harvest mustard greens from May through July-it is an essential part of their diet in this early period of the growing season (Hinton 1970, p. 13; Kunstadter 1970, p. 73). Corn is harvested in August, just in time to substitute for rice as the staple (for families whose supplies have run out) until some early rice can be harvested in October. It can also be preserved by smoking (Hinton 1970, p.13; Kunstadter 1970, p. 72). Other important main swidden crops include cotton, sorghum, peppers, beans, various roots and tubers (cassava, etc.), squashes, yams, melons, cucumbers, millet, elusine, sesame, and various herbs and spices, including garlic and onions (Hinton 1970, pp.12 - 13; Kunstadter 1970, pp. 74 75). Crops are harvested as they mature, most of the vine plants (melons, pumpkins, etc.) maturing at about the same time as the rice, in early November, with only two crops-chills and sometimes cotton- continuing to be harvested for a few months after the main rice harvest (Hinton 1970, p.13; Kunstadter 1970, p. 71). Separate garden plots contain longer-living plants, such as sugar cane, bananas, ginger, and bamboo (Kunstadter 1970, p. 94), as well as tobacco. Household gardens contain a variety of plants, such as eggplants and tomatoes, and fruit trees such as jackfruit, mango, pomelo, lime, and guava (Hinton 1970, pp.12-13; Kunstadter 1970, pp. 75, 94). Hunting and fishing also contribute to the diet, and gathering of wild forest products is quite important. Bamboo shoots and other wild vegetables are gathered extensively and are an important part of the diet, whereas other forest products, such as orchids, are gathered for sale (Hinton 1970, p. 2; Kunstadter 1970, p.104).
2.1.3. The transformation process
Site selection begins in February, in the dry season. The Karen know their territories well, having farmed the sites over and over in the past, so the use of indicator plants is minimal and site readiness is judged by regeneration and re-establishment of woody, broadleaf vegetation. Clearing is not as difficult a task as primary forest clearing, where large buttressed trees would have to be felled by arduous and dangerous axe-work. Instead, secondary forest clearing can usually be done with the bush knife alone and women as well as men can participate. Although the task is considered one of the nastiest types of work of the swidden year, it requires much less labour and less time than primary forest clearing and can usually be completed in several weeks (cf. Freeman 1970, p. 245; Miles 1967, p. 97; Conklin 1957, p.150). Burning takes place around the end of March or beginning of April, after which planting commences in time for the coming of the rainy season. From June through mid-September, weeding of the swiddens is a nearly continuous task, after which the rice plants are tall enough to keep ahead of the weeds and weeding requirements are greatly reduced. Crops are then guarded until harvest in late October and early November, immediately followed by threshing, winnowing, and transporting of crops to house sites until well into December.
The Karen secondary forest system in Northern Thailand is under a good deal of pressure from increasing population density (see Hinton 1975, p. 21), and most Karen families cannot meet year-round subsistence needs without supplementary resources. Rice supplies often run out before harvest, when corn or swidden foods other than rice provide some relief. This is apt to occur around September, and it is at this time, when weeding requirements also begin to slack off, that many Karen men look for wage labour while weeding is continued by the women (Hinton 1969, p. 47). Other periods in the swidden cycle that allow some time off for wage labour occur in late December and in January. Wage labour has become extremely important to the Karen. A survey conducted in 1967 found 30 per cent of a large sample of Karen households engaged in wage labour to make up rice deficits, but, in a later survey carried out in one area, it was discovered that an average of one Karen per household found it necessary to engage in wage labour for the Forest Department alone (computed from TRC 1975, pp. 2, 61) ! Wage labour is periodically available with the Forest Department, in the mines, and from lowland Thai wet-rice farmers, other Karen, and other hill groups, particularly the Hmong (see below ).
Although wage labour seems by far the most important supplementary activity for the Karen, they also engage in other activities. A few Karen attempt to raise cattle and other domestic animals for sale to lowlanders. Some have arrangements with lowlanders to take care of water buffaloes and oxen during slack periods in the wet-rice year. Some raise small cash crops (such as Chili peppers). Some gather and sell the leaves of Dipterocarpus tuberculatus (used for roofing and packaging )-a dry-season activity-while others sell other forest products such as lacquer or stick-lac. An increasingly popular source of cash is the sale of handicrafts for the tourist industry. A principal supplement to the swidden rice is wet-rice in the terraces, a few of which can be found in most Karen villages. Most land feasible for conversion to terraces, however, has already been converted; thus the use of terracing as an increased supplement to or substitute for swiddening is limited.
The raising of cash crops among Karen swiddeners has generally not been popular, despite experimental efforts in this direction. There are structural reasons in the secondary forest swidden system for this difficulty with cash crops, as will be discussed in the next section. A very small number of Karen villages have participated in opium-growing, and there has also recently been increasing interest in expanding production of corn which can be sold for export to Japan after having been shucked and cut from the cob. A few families in one village known to the author tried to expand the production of corn in 1975 by planting it in year-old swiddens which would normally have been fallowed. There may be some danger in this practice, as will be seen below.
2.1.4. Ecological and socio-cultural constraints
The nature, amount, and distribution of labour and their relationship to environmental factors-particularly seasonality-in agriculture are crucial to understanding the potentialities of and constraints on any particular agricultural system. This is especially true of swidden agriculture. In humid equatorial zones with no pronounced dry season (or with two ''less wet" seasons and two periods of heavier rain per year) conditions differ substantially from monsoon tropical areas, where wet season and dry season are more pronounced, as they are in Northern Thailand and in many other areas of the world. In such areas, timing is very important. For example, felling and clearing must be accomplished in time for the debris to dry out before the burn, and burning must be done at just the right time. If it is carried out too soon, an incomplete burn may result, and crucial nutrients will be lost and competing flora not destroyed; if too late, the rains may come and prevent an effective burn. But more pervasive aspects are also involved. In a sense, by clearing a new field, if weeds are a problem in old fields, the swiddener trades off increased weeding labour in the growing season for clearing labour in the dry season. If he uses his fields for several years in a row, the increasing difficulty with weeds will have to be weighed against the difficulty of clearing a new field, while bearing in mind that the clearing labour occurs in a season when he has more time for it. The difficulty of the type of vegetation to be cleared is one factor (Watters 1971, p. 39). Another factor is who does which labour, since traditionally weeding is a woman's job and felling (and most clearing) is a man's job. Finally, all these considerations must be weighed against whatever other labour is needed or desired, its seasonal timing requirements, if any, and its conflict with the swidden cycle requirements. Concomitantly, the type of swiddening practices and the crop assemblages grown will affect both the weeding and clearing labour needed. All these in turn will also be influenced by such other environmental factors as soil, moisture, temperature, etc.7
Except for a small amount of experimentation (the corn example mentioned above) the Karen secondary forest swiddener does not farm his field more than one year in a row. The Karen state that they would not be able to keep up with the weeds if they did this and that yields would be ridiculously low. Weeding is already an onerous task in which both men and women must participate. This may be due partly to the type of forest they swidden (regrowth is rapid) and partly to the low fallow time, which for most Karen today probably averages about six years (certainly well less than ten).8 The fact that rice is the principal crop is also important, since it must compete for the same niche with many of the early weeds. In Latin America corn can be cultivated in secondary forest plots for several years in succession without serious weed competition, after which plots are fallowed due to nutrient rundown (see Cowgill 1961, p. 49, passim; Watters 1971, pp. 40 - 41 ). (This is discussed more broadly in the next chapter.)
Because they use their fields for only one year at a time, and because of other conservational practices (such as fire-guarding, pollarding, minimum tillage, watershed protection, and preservation of natural vegetation above swiddens), forest regeneration in Karen areas is rapid (Nakano 1978, pp. 424-427). It is possible that the weeding requirements also serve as a check against that spectre of high-population density which nearly every researcher of swidden agriculture has warned against. The Karen know that they will not be able to make a living from a forest plot if they cut it more often than every four years, and thus it is most unlikely that they would reduce the swidden fallow period beyond a certain point unless they became quite desperate. (Appropriate development should, of course, aim to reduce the cause of desperation.) Instead, the Karen have increasingly turned to other sources of income. Birth control is also practiced (Hinton 1975, p. 9), and outmigration of both individuals and families is occurring (Grandstaff 1976, pp. 227-231; Kunstadter 1970, p. 55). It is also most unlikely that resident Karen would attempt to increase yields by increasing the size of their swidden plots, since it is excessive weeding labour that already limits the size of Karen swiddens (Hinton 1970, pp.18-19). The Karen themselves are suffering, the forest less so. There is no evidence that Karen forest areas are turning to grassland, and erosion and water run-off are minimal (Zinke et al.1970; Sabhasri 1970). Kunstadter (1967, p. 652) points out that the Lawa system, for example, has been stable for at least 150 years, ''using the same [swidden] field areas." The way established swiddening is practiced in Northern Thailand contradicts the assertion of Nye and Greenland (1960, p.130) that it is the desire to remain in place rather than moving on that promotes environmental degradation by swiddeners.
The practice of following rice with corn in the second year, however, could conceivably be detrimental if it were widespread. Since it would require weeding at the same time as the rice, it seems unlikely that such a practice would be widely adopted. If the field is reburned before planting in the second year, shoots and suckers could be killed which are important to regeneration or to the maintenance of species composition. Any widespread changeover from rice to corn could have drastic effects on the entire system, but this seems unlikely since the Karen seem to place a good deal of value on growing their own swidden rice. The special characteristics of corn swiddening are not well understood, even in areas where it is widely practised.9 Experimentation and testing would be most valuable in Northern Thailand, especially if the marketing situation continues to improve.
It can be seen, then, that a principal constraint on secondary forest swiddening (of rice) is the amount and duration of labour needed for weeding. It will be seen that this is the opposite situation to primary forest swiddening, where opium cultivation has been practiced in addition to growing subsistence rice. The weeding requirements of the Karen prevent their using a field for more than one year in a row, but it is these same weeding requirements which help to ensure the viability of the system, since fields are not exhausted by consecutive re-use, nor are regeneration and species composition detrimentally affected by consecutive reburning. Viewed from the other direction, the labour requirements of the system also limit what other activities the swiddeners can engage in to help support themselves, since only at particular times can the swiddener free himself from the operation of the system. Periodic wage labour, then, can be seen as an activity compatible with agricultural system operation, whereas the growing of cash crops, if their scheduling interfered with the swidden cycle, would not be compatible (see Keen 1972, p. 28).
Socio-cultural aspects are also important, however, in analysing constraints on secondary forest swidden agriculture in Northern Thailand. Characteristics of particular cultural groups are likely to be extremely important, but the attempt will be made here to discuss some general aspects that ought to be more broadly applicable. To begin with, some general characteristics exist that apply to most, if not all, integral swiddeners in Northern Thailand. Economically, the integral swiddener tends to be more subsistence-oriented than his lowland neighbours. In lowland society a sizeable number of people are not directly engaged in farming. Their food comes from the farmer, who in turn receives goods and services from them. In swidden society each household is capable of meeting most of its own needs. Very few, if any, full-time specialists exist who are not engaged in swiddening. Part-time specializations are not uncommon, however.
The requirements of the swiddener and how he is able to meet them affect the social and political organization of swidden society. In permanent-field society, mechanisms must exist to order the exchange of goods and services above the village level. To assist in the functioning of such a system one finds a hierarchical political organization, a code of laws, taxes, soldiers, policemen, religious orders, engineers, etc. Because the swiddener is much more able to meet his own needs, his need for institutions of this scope, beyond the village level, is minimal. His village may have a headman, but the headman's powers are usually persuasive rather than coercive. The headman is often also the religious specialist. In any case, such roles do not relieve the individuals of the necessity to swidden their own fields like everyone else. Both role differentiation and class differentiation are less complex and less fixed.
Basically, in swidden society, no household can tell any other household what to do. Within the household, individuals are still extremely free from having to take orders. Although men usually hold the position of household head, many tasks are divided by norms into men's work and women's work; it is thus not up to the household head to decide who shall do what. Many other tasks not prescribed by norm are done when and if the individual himself decides to do them. Whatever coercive power the household head may have over subordinate (almost always younger) members of the household is offset by the desire and ability of the subordinate members to get out and establish their own households as quickly as possible. It is not uncommon to find even elderly widows or widowers living alone in their own houses if they can still support themselves.
Any development assistance designed to aid swiddeners will first have to come to terms with these basic organizational characteristics of swidden society. For example, the sort of development work in which an agricultural assistance officer passes his recommendations on to the headman and expects the headman to get the villagers organized to carry out his plan is sure to fail, nor will the assistance officer have much better luck if he tries to do it himself. It will be seen in the next chapter that reorganization of swiddeners in order to make swiddening more organized and efficient has only occurred in cases where a good deal of coercion was involved, or else in cases where the swiddeners were so badly off (usually through no fault of their own) that they had no other options but to comply. This is not the sort of development advocated here, nor can proponents of such schemes truthfully state that their proposals are certainly for the swiddeners' own good.
A second aspect of swidden society relates to economic conditions by which class formation and ascendancy of elites can occur, so that swidden societies "evolve" in the direction of more rigidly structured forms of agricultural society (similar to "feudal" society in Europe). There are mechanisms in swidden society that generally work to prevent this from occurring. Swidden land is neither owned nor rented, and accumulation of other forms of wealth tends to dissipate by ritual requirements (feasting, etc.) whereby wealth is traded for prestige. Labour is the true valuable item, but households which have accumulated labour tend to lose it eventually as children break off and establish their own households, and so on. The economic advantages to be gained from being a member of such a household are offset by the distaste for having someone else trying to control one's time. Nevertheless, there are signs of increasing stratification in some Karen villages. The ownership of wet-rice terraces becomes economically advantageous when swidden labour is excessive and rice is in short supply. Increased communications with lowland society have also made available a variety of modern items (such as transistor radios), the desire for which makes subordination to others increasingly attractive in return for money to buy these items. In several Karen villages (and probably many) a few Karen starting from a position of comparative advantage have managed to acquire nearly all the terraces in the villages and now employ other Karen to work them in exchange for a portion of the harvest. In at least one village, terrace owners have expanded into opium production, employing poorer Karen during key labour periods in September, December, and January; in the same manner many Karen are hired by Hmong and other opium-growers for the same purpose.
It will be suggested in the final chapter that development among swidden agriculturalists need not follow this pattern of increasing social stratification with economic differentiation between rich and poor (nor need it be conducted as a patronizing or coercive effort of one group of people telling another group of people what to do, whether or not for their own good). It may be that increasing stratification is representative of a former typical stage in the transition of swidden agriculture to permanent-field agriculture. But swidden societies exist in the present and are as much a part of that present as any other group. They are not isolated, as if on another planet, nor do they now have to proceed through any series of evolutionary "stages" in duplication of what other societies may or may not have undergone in the past. A modern society ought to be able to accommodate swidden values and lifestyles. Indeed, if we can set aside our preoccupation with the obvious material differences. their values may be a good deal closer to the most modern societal conceptions than we have allowed ourselves to believe.
More specific environmental and socio-cultural constraints of established, secondary forest swiddening in Northern Thailand will be discussed below after a discussion of the ''pioneer'' swiddening adaptation, in contrast to which these contraints will be better appreciated.
2.2. The''pioneer'' swiddeners
2.2.1. Principal natural resources
2.2.2. Principal resources produced
2.2.3. The transformation process
2.2.4. Ecological and socio-cultural constraints
The''pioneer" swiddeners of Northern Thailand account for about 44 per cent of the integral swiddener population. Most have arrived in Northern Thailand within the present century. Around the turn of the century, and perhaps as early as the mid-nineteenth century, isolated villages already existed in border areas, but it is probable that the great increase in pioneer swiddener population since that time has occurred more as a result of immigration from China, Burma, and Laos than from natural population growth. It is probable that these groups adopted the growing of opium as a significant cash crop around the third quarter of the nineteenth century, and that opium then facilitated more rapid expansion (see Grandstaff 1979; 1976, pp. 164-183; Dessaint 1972, p.11; Bradley 1975, pp.1516; cf. Geddes 1976). Opium helped to pay for the long moves necessary when primary forest in a contiguous area had already been farmed. The long move involved costs of transportation, costs due to having to leave behind household materials, food costs until a crop could be harvested in the new area, and costs involved in securing unhindered access to the new forests. In the process, choices were involved at each step, and those who would not or could not make the longer moves remained behind and eventually had to adopt alternative farming practices. Some even became wet-rice farmers (see Mickey 1947 on Hmong in Kweichow). Reasons for not continuing to follow primary forest swiddening might have existed in lack of cash or other support, or in lack of information about suitable new areas. As a result, the cultural groups that are represented in Northern Thailand pioneer swiddening are spread throughout northern Southeast Asia and southern China, and kinship links exist among some peoples separated by great geographical distances. With the disappearance of suitable primary forest, this great expansion is now at an end.
The Hmong (Meg, Miao) have been generally the most successful of the pioneer, primary forest swiddeners in pursuing the primary forest adaptation. There are very few suitable areas that they have not farmed at one time or another. In both Thailand and Laos they are by far the largest pioneer group. They can be discussed here, then, as the prototype of the pioneer primary forest swidden adaptation. As in the previous section, discussion here will cover: (1) principal natural resources in the system, (2) principal resources produced, (3) the transformation process from (1) to (2), and 14) general ecological and sociocultural characteristics of and constraints on the overall system, to include the contrasting of these with the characteristics and constraints of established, secondary forest swiddening.
2.2.1. Principal natural resources
The Hmong, like other pioneer groups, prefer to farm primary forest or as highly regenerated forest as they can find. Binney (1968, p. 124) states that the Hmong prefer to farm primary forest because of the better yields they can expect for the labour. Overall, as will be seen below, not only will primary forest allow better yields for the labour in the first year, but in succeeding years on the same plot yields can remain high, often with even less labour. The increased labour spent in initial felling is made up for by the reduced weeding requirements and the possibility of using the plot for several years in a row.
The type of forest preferred for swiddening rice is Tropical Evergreen, as with the secondary forest swiddeners. However, since many of these forests had already been cultivated by secondary forest swiddeners and were thus not available as primary forest, Hill Evergreen Forest was more often cultivated by the pioneers. Another limiting factor was the increasing necessity to cultivate opium in order to retain the increasing mobility that the Hmong needed as the best forest spots became smaller and farther between, making movements both longer and more frequent (see Grandstaff 1976; 1979). The opium poppy will not produce opium in Northern Thailand below about 1,000 m elevation. Above this altitude, Hill Evergreen Forest is much more common than Tropical Evergreen. Since the Hmong were primarily subsistence rice growers (Binney 1968, p. 73; Bernatzik 1947, p. 347) but also needed opium, sites above 1,000 m, usually in Hill Evergeen Forest, where both rice and opium could be grown, became the mode.
Hill Evergreen Forest in Northern Thailand is a partial cloud or "mist" forest. The trees "rake" moisture from the moist, cool air, causing it to condense on the ground. Like Tropical Evergreen, it is a three-level forest, but the strata are not as distinct. It has neither the species diversity nor the biomass of Tropical Evergreen-the trees are generally neither as large nor as tall (see Ogawa et al. 1961; Kuchler and Sawyer 1967). Typically, the forest's larger trees consist of "many kinds of oaks and chestnuts (Quercus, Lithocarpus, and Castanopsis spp.)" (Smitinand 1962, p. 3). No one species can be said to dominate; in fact "few if any are common" (Kuchler and Sawyer 1967, p. 324). As a semi-tempe,ate forest type, Hill Evergreen Forest, although having a smaller biomass than Tropical Evergreen, has a much higher proportion of nutrients stored in the humus layer and in the soil, a factor that may help to explain why fields in such forest may be used for several years in a row if carefully weeded, even if they are not cut from primary forest. Once fields are exhausted, however, regeneration is likely to be very slow indeed and succession to grassland is a frequent characteristic. When this occurs, the grasslands themselves become a resource of sorts. The grassland succession usually starts with invasion of Imperata cylindrica, which is probably succeeded by other grasses (and finally shrubs and forest regrowth ) if undisturbed (see Seavoy 1975; Kowal 1966).10 But grasslands are frequently burned for a variety of purposes (see Conklin 1957a), principal reasons being to restart the Imperata as young shoots suitable for grazing livestock, to facilitate passage, to retain Imperata for various uses, etc. (Imperata has a variety of practical uses, chiefly as a roofing material.) The effect of frequent burning, however, is to perpetuate the Imperata. Other forest products (e.g., mushrooms, fern leaves, wild banana leaves, wild tea leaves) are used for food, medicinal purposes, poison, and rituals (Binney 1968, pp. 35-37; Walker 1970, p. 527). Hunting is often probably more important than gathering, the higher forests becoming even less suitable for gathering due to pioneer methods of land use. In earlier times, Hmong and other primary forest swiddeners made use of a variety of primary forest resources which are no longer available with the disappearance of primary forest in Northern Thailand. Hunting was an extremely important source of protein; the Hmong were formerly avid hunters who hunted everything, including even rhinoceroses and elephants (Bernatzik 1947, pp. 339-345). Game is still hunted where possible. Meat is consumed and sometimes skins and horns are sold. A wide variety of animals are hunted, including barking deer, bears, monkeys, wild pigs, squirrels, small jungle cats, lizards, and wild fowl (Walker 1970, pp. 523-524). Exotic birds are also captured for food and for sale.
In most "pioneer" swidden areas today, mature forest in the area has already been swiddened in primary forest fashion. Damage to forest, although significant, is probably not as great as might have occurred in that many swiddens were fallowed before "exhaustion'' in order to move on to the relatively few mature forest sites remaining. Concomitantly, villagers who lacked opportunities to move on probably began taking better care of their land. Nevertheless, the majority of such villages have already been forced to start a''second cycle" on land cut from low secondary forest or bush, and yields are reported to be generally unsatisfactory, forcing increased reliance on cash crops.
2.2.2. Principal resources produced
Pioneer swidden practices have always varied from group to group. Rice, opium, and corn have generally been the common elements, but, as good swidden land became increasingly scarce, a good deal of experimentation has taken place, further increasing variation. Some chief differences are to be found in cash crops which have been adopted (i.e., other than opium). Potatoes have been popular in some areas and fruit crops have also been widely used. The group of Lahu Nyi studied by Walker also maintained separate chili pepper swiddens. Usually in all swiddens, however, a group of "catch crops" are grown along with the principal crop (variety being a principal functional characteristic of swidden agriculture throughout the world). Catch crops include: millet, sorghum, corn, beans (many types), sesame, cassava, potatoes, taro, spinach, eggplants, melons, tomatoes, pumpkins, gourds, spices (ginger, lemon grass, garlic, etc.), pineapples, bananas, sugar cane, tobacco, yams (several types), pigeon peas, and sunflowers (Walker 1970, pp. 382-384, 392 - 394, 406-407; Binney 1968, pp.104A, B; Geddes 1976, pp. 166-167). Each type of swidden can be expected to contain a balance of cereals, legumes, root crops, and vegetables, in addition to the principal crop. Fruit trees are also maintained, including coconut, jackfruit, leche, mango, peach, and pomelo (e.g., Binney 1968, p.104B). Unlike the case of the established swiddeners, household gardens are not usually kept unless either the gardens or the village pigs are fenced in.
Among the domestic animals, pioneer swiddeners keep fowl (mainly chickens) and cattle, as do the established swiddeners, but sizeable numbers of pigs are probably the most important. Pigs are fed on the corn (and/or corn cobs and husks, banana leaves and trunks, etc.). Horses and oxen, used for transport, are also important, since pioneer villages tend to be more remote than those of the established swiddeners.
2.2.3. The transformation process
When primary forest was still available, felling and clearing began in January. Subsidiary clearing and debris-cutting activities continued well into March, with burning in April. Felling was difficult and dangerous work, the province of men only who worked in groups to bring down the big trees (see Freeman 1970). Weeding requirements were light, usually one weeding (in June, July, or August). Originally, opium may have been planted in July in the main rice swidden, along with the numerous other crops, and harvested in December, after the main rice crop (see Geddes 1976, p. 160; Bernatzik 1947, pp. 347 - 358). With the increasing use of less regenerated forest, however, weeding became increasingly difficult, and it is probably for this reason that varieties of opium that could be planted later and had a shorter growing period were increasingly favoured. They could be harvested in January or February since, in secondary forest, this time was no longer needed to fell the big trees. The pioneer swiddeners could also make better use of the increasingly scarce ''ideal'' sites by planting rice in the sites below 1,000 m and opium in the ones above, the staggered growing seasons allowing adequate labour for both sites. But there was still a problem in preserving the opium fields and keeping them free of competing vegetation during the monsoon until time for planting, around September. To do this, the production of corn, formerly planted in the main rice swiddens, was expanded and planted in opium swiddens instead, shading out competing vegetation and preventing erosion until opium could be planted and the corn harvested. Corn could be used to feed pigs, an important source of protein that would help to make up for the increasing scarcity of wild game. At any rate, the opium-corn-pigs complex had become an important part of Hmong economics by the 1960s (Geddes 1976) and is now also used by other pioneer groups. If corn is not grown in the opium swiddens during the rainy season, weeding requirements (done by hoeing), prior to opium-planting, are greatly increased.
As mature forest became scarce, however, weeding requirements increased substantially in both corn and rice swiddens. For a variety of reasons-the increasing size of the villages, increasingly frequent movement, lack of good rice-swidden sites, and increasingly small swiddenable areas-the Hmong resolved this labour conflict between weeding corn and weeding rice in favour of the corn swiddens where opium was to be planted. Money obtained from the sale of opium could then be used to buy rice, as rice yields fell in the neglected swiddens, or as some individual households abandoned the swiddening of the rice entirely.
By the late 1960s and early 1970s, however, the scramble for new land was virtually over as more and more pioneers despaired of finding better areas than those they already possessed.12 Some efforts are being made to adopt secondary forest rotational swiddening patterns (Keen 1972, pp.149-150), but, in general, fields are already too degraded by previous swidden-to-exhaustion practices, and population densities in many pioneer swidden areas are too great for such a system anyway. Instead, there has been a variety of responses, none of which has yet caught on as being clearly advantageous for the majority. These include cattle-raising, growing fruit-trees (especially peaches), a little wet-rice farming (usually in terraces bought from the Karen and others), cash-cropping of various field crops (especially corn and potatoes), tea and coffee (experimental, with assistance from development authorities), and the ubiquitous cultivation of opium, whose prices probably still offer the best general prospects despite increasing labour, decreasing yields, and the dangers of raising an illegal crop.
As these changes have occurred among primary swiddeners, technology has also changed, although perhaps not to the extent one might expect. The axe has been traded for the hoe, but fields are still fired in the dry season as the customary method of initial clearing. The use of fertilizer, commercial or by manuring, has not yet been adopted (as of 1975), even in opium fields. Soil is often thoroughly turned, but erosion in fields is probably still slight because most fields are at higher altitudes, where rainfall is in the form of frequent gentle rains rather than massive downpours.