The Maya milpa: fire and the legacy of living soil

Authors


Abstract

Swidden is an agroforestry system in which woody vegetation is regenerated after a period of annual cropping. Associated with most forested areas of the tropical world, swidden is often blamed for deforestation but it also plays a role in forest conservation. Here, we examine the contemporary milpa, a type of swidden agriculture common to Latin America and historically used by the Maya people of the lowlands of southern Mexico and northern Central America; we focus on one group in particular, the Lakandon people of Chiapas. One element of milpa agriculture that receives a considerable amount of criticism is the burning of cut vegetation after clearing. Fire can have negative effects on ecosystems but swidden cultivators are often sophisticated managers of fire. Among the benefits of fire use in this setting is its contribution to nutrient flow and to long-term soil fertility in the form of biochar, charcoal produced by low-temperature pyrolysis in agriculture. When properly managed, the milpa cycle can result in long-term carbon sequestration and an increasingly fertile anthrosol (soil that has been greatly modified by long-term human activity) and enriched woodland vegetation.

Swidden, also known as “shifting cultivation”, is probably the oldest form of farming in the Americas (Barker 2006) and is also perhaps the most maligned and misunderstood. Swidden is often blamed by conservationists for the deforestation that affects many tropical areas. However, because swidden is an agroforestry system that regenerates woody vegetation after a period of annual cropping, these systems are often associated with the most densely forested areas of the neotropics and in many cases play a role in their conservation as woodlands. The element of swidden systems that receives the most criticism is the use of fire to burn the plant debris that remains after forest clearing.

In a nutshell:

  • Swidden, an ancient form of horticulture, rotates crops and woodlands; milpa, as practiced by ancient and modern Maya, shapes woodland ecology

  • Succession is carefully managed to contribute to soil fertility and biodiversity

  • Low-temperature burning sequesters carbon as persistent biochar

  • Traditional Maya milpa management supports forest ecosystem services

The term “swidden” derives from the Old Norse “svithinn”, meaning “clearing in the forest to be burned”. We prefer this term to alternative names for this kind of agroforestry, such as “slash-and-burn” or “shifting cultivation” that invoke an image of a primitive scorched earth operation, carried out by shady cultivators who wander randomly through the forest. There is no doubt that swidden fires have a potentially negative effect on soils and on forests – high temperatures can destroy soil ecology, reduce organic matter, damage woody vegetation, and select for fire-resistant species in the woodland seed bank. Fires that escape from the swidden clearings can cause widespread ecosystem destruction. Burning also has major implications for regional and global climate through the emission of a variety of greenhouse gases (Wilken 1987; Cochrane 2003). In fact, swidden cultivators are often careful and skilled users of fire who seek multiple objectives in the burning process, among them the minimization of some of these negative effects (Hecht and Posey 1989). Burning of newly felled vegetation in tall forest areas is often necessary to clear the land for planting. Fire can also make an important contribution to long-term soil fertility through the addition of biochar that has been produced by low-temperature burning (Glaser et al. 2002; Xu et al. 2012). In fact, we argue that the Maya swidden cycle, when properly managed, results in enriched anthropogenic soil and represents a lasting expression of agroecological skill and creativity (Wilken 1987; Peterson et al. 2001).

The Maya milpa: a resource management tool

We define milpa as an open-field polyculture centered on maize (Zea mays) that rotates with woodland vegetation in a cycle of around 10 to 25 years. Although found at varying levels of intensity and productivity throughout Mesoamerica, in its classic form milpa – a Spanish term roughly meaning “maize field” – can be characterized as successional agroforestry and, as such, is a form of horticulture rather than agriculture, since it involves intensive individual plant management, embedded in a diverse forest environment. Milpa is more than a system of cultivation. By rotating annual crops with tropical secondary forest in a successional cycle (Figure 1), milpa moves beyond successful food production and becomes the central axis of a resource management system that upgrades woodlands with species useful to humans, accelerates succession, and constructs an anthrosol (soil that has been heavily modified by human activities) of ever-increasing fertility (Ford and Nigh 2009). The limits to milpa productivity are not population growth or the shortening of the fallow period, as is often assumed, but rather labor availability and skill (Nigh 2008).

Figure 1.

Fire-stimulated regeneration: the lowland Maya milpa cycle. (Courtesy of A Ford/ISBER/MesoAmerican Research Center, UCSB).

Milpa is the best known agroecological system in the Maya culture area and has been well described in the academic literature (Bernsten and Herdt 1977; Nations and Nigh 1980; Gómez-Pompa 1987; Hernandez Xolocotzi et al. 1995; Steinberg 1998; Nigh 2008; Diemont and Martin 2009; Isakson 2009; Terán and Rasmussen 2009). Milpa, in one form or another, has been practiced for millennia, in all ecological settings within the Maya region and is still a central activity in traditional Maya management of environmental sustainability. The ingenuity of milpa lies in its potential for adaptation to the extremely varied local ecosystems and cultures of the region. Milpa systems are found from sea level to highlands above 2000 m; it can be integrated with home gardens (infield, referring to crops grown around the family house site; Caballero 1992; Corzo and Schwartz 2008) or may be situated some kilometers from the family dwelling (outfield; Pyburn 1998; Netting 1977), and is capable of degrees of intensity and productivity that respond to environmental constraints, market demands, and the household domestic cycle, varying according to the number of mouths to feed and the available family labor (Netting 1993; Anderson and Anderson 2011). The milpa system also harbors an enormous agrobiodiversity, heritage of an ancient agrarian civilization, and still serves as both family garden and a nursery for agroecological experimentation (Alvarez Buylla 1989; De Clerck and Negreros Castillo 2000).

As practiced by traditional farmers, milpa is devoted primarily to maize, with intercrops selected from over a hundred species domesticated in pre-Columbian times and complemented today by additional species from all over the world (Terán and Rasmussen 1995). The Maya milpa entails a rotation of annual crops with a series of managed and enriched intermediate stages of short-term perennial shrubs and trees, culminating in the re-establishment of mature closed forest on the once-cultivated parcels of land (Nations and Nigh 1980; Hernández Xolocotzi et al. 1995; Nigh 2008; Terán and Rasmussen 2009). The integration of the milpa cycle into neotropical woodland ecology transformed successional processes, leading to the formation of the contemporary tropical forest of the Maya lowlands as a garden, where more than 95% of the dominant tree species have utility for humans (Levy Tacher et al. 2002; Campbell et al. 2006).

Early cultivators employed and expanded small, natural woodland clearings, observing and eventually intervening in the processes of ecological succession to mold the environment for their own benefit. Many agroforestry practices originated with the inhabitants of the Maya lowlands during the Archaic (8000 to 2000 BCE), the precursors of what would become elaborate forms of agro-forestry practiced by indigenous peoples throughout the neotropics (see Alcorn 1990; Peters 2000; Toledo et al. 2008). These activities left their imprint on the woodland environment long after some of these territories had been abandoned (Campbell et al. 2006; Ross 2008).

Intensively managed woodland patches have probably been important in shaping the structure and composition of the Maya forest. An example is the pet kot, a form of forest modification practiced by the Yukatek Maya. The pet kot is a tall, managed tree garden, created in favorable ecological niches (Gómez-Pompa et al. 1987). Pet kot were developed on sites of 19 000–24 000 m2, where particularly useful plants were cultivated, forming a kind of nursery within a protected forest ecosystem. Many of these species were also common to local home gardens (such as those of the genera Brosimum, Spondias, Pithecellobium, Malmea, Bursera, Sabal, and others; Gómez-Pompa et al. 1987, 1990).

Climate variability and especially periods of severe drought have historically been a part of the region's ecology and may have been determining factors during certain periods of Maya civilization, for example the so-called Classic Collapse (600–900 CE) (Gill 2000; Peterson and Haug 2005; Kennett et al. 2012; Medina-Elizalde and Rohling 2012). During these periods, areas such as the pet kot may have provided important refugia for useful woodland species, allowing subsequent afforestation during wetter periods to be skewed toward these species. Such agroforestry practices are a key element in the Maya adaptation to long-term climate variability (Ford and Nigh in press).

These are examples of practices that probably arose during the “long transition”, as Archaic mixed mobile horticulturalists occupied the lowland forest in the 5000 year-long “climatic optimum” (a period of warm, moist, and stable climate before 4000 years ago; Smith 1998; Burroughs 2005). Eventually, a true agrarian society emerged from this ecological matrix, one that showed an increasing dependence on settled horticulture (Barker 2006). This horticultural evolution resulted in a domesticated landscape that transformed the forest into a cultural feature, cultivated by the Maya (Fedick 2010). The creation of the Maya forest garden is the result of an accumulated investment in and intensification of the milpa cycle, and forms part of a dynamic socioecological system (Ford and Nigh 2009).

Having evolved within the historical ecological context of the tropics, the milpa forest garden cultivation cycle supplied households with food and other botanicals at every successional stage. The stages of the milpa cycle recapitulate the ecological stages of vegetation succession that is characteristic of tropical forest dynamics (Kellman and Tackaberry 1997; Finegan 2004; Nigh 2008). The economically valuable plants that proliferate in the Maya “feral” forest today (Campbell et al. 2006) owe their dominance to human selection over millennia in the context of the milpa cycle (Atran 1993).

The evolution of the tropical woodlands under the milpa system is a central aspect of the natural history of the Maya culture area. The Maya forest is as much a vestige of Maya civilization as are the ancient stone temples of Palenque or Tikal and their surrounding small structures, which are the remains of house foundations and agricultural terraces. The neotropical woodlands in which ancient Maya civilization developed have been greatly influenced by the human societies it has nurtured to the present day.

Archaeologists and paleoecologists have sometimes assumed that the evidence of milpa agriculture in paleoecological data implies widespread deforestation (Mueller et al. 2010; Kennett et al. 2012), but agroecologists understand that this is not the usual outcome and recognize the importance of traditional systems of cutting and burning for managing local biodiversity (Gliessman et al. 1981; Ford 2008). The Maya forest, as with other wooded areas of Mesoamerica, has been structured by the milpa cycle, the fundamental agroforestry management tool of both ancient and contemporary Maya (Finegan 2004; Campbell et al. 2006).

Many contemporary examples of milpa are associated with degraded ecosystems (Lawrence et al. 2007; Daniels et al. 2008; Schmook 2010). These are usually forms of de-intensified milpa agriculture. As family labor is diverted to off-farm wage work, the tasks required to maintain the long-term sustainability of the milpa system, including intergenerational transmission of knowledge, are increasingly neglected and labor-saving tools (eg herbicides) are introduced (McGee 2002).

The milpa forest garden cycle – Lakandon Maya agroforestry

We can gain some insight into the relationship between traditional Maya milpa agriculture and the tropical rain forest by examining a recent ethnographic example of the Lakandon Maya of Chiapas, Mexico (Figure 2). Historically, the Maya have employed a wide variety of farming systems, but the example of the Lakandon milpa, practiced almost universally by the Lakandon until the end of the past century, illustrates the skillful cultural engagement with the neotropical woodland environment that is characteristic of many indigenous Mesoamerican adaptive management systems (Wilken 1971, 1987; Alcorn 1990; Toledo et al. 2003; Nigh 2008). An understanding of Lakandon land-use management reveals the flexibility of the milpa forest garden system.

Figure 2.

A typical Lakandon milpa. Maize stalks are doubled over to allow ripened ears to dry while sweet potato and squash provide ground cover.

The Lakandon are one of the smallest Maya groups (currently, a population of over 600 individuals, divided among three communities), but also the one with the longest history of continuous occupation of the Chiapas rain forest (Palka 2005). They possess a detailed knowledge of the tropical lowland environment and, until the end of the 20th century, widely practiced a diverse and intensive agroforestry swidden based on maize and a variety of other crops (Wilken 1971; Nations and Nigh 1980).

The Maya subsistence strategy involved multiple land uses, in which several ecological zones were (and still are) managed and exploited concurrently. In addition to milpa and home gardens around the Lakandon house site, regenerating forest derived from previous milpa cycles, mature woodlands, and aquatic and semi-aquatic ecosystems typical of the area provided various resources. Lakandon men traditionally dedicated the greater part of their daily effort to milpa farming, in addition to hunting and gathering other forest resources. Women and children helped during periods of high labor demand, such as during harvest, as is common in other parts of Mesoamerica. Such dedication to milpa work led to levels of diversification and productivity rarely seen for this agroecosystem in recent times. This provides us with a unique contemporary example of what Wilken (1971) called the “high-performance milpa”, a form of intensive milpa agriculture likely to have been far more commonly practiced in densely occupied ancient Mesoamerica.

Maya farmers chose cultivation sites surrounded by mature forest to maintain a source of tree seeds for succession. The result of this practice combined with intensive daily selection and weeding of the cropping area resulted in a careful control of the soil seed bank, oriented toward accelerating and directing ecological succession and achieving rapid tall forest regeneration (Nigh 2008; Levy Tacher and Aguirre Rivera 2005). These regenerating, reforesting parcels of land following milpa cultivation are known as jurupche (Yukatek Maya) or acahual (Spanish); they develop anthropogenic vegetation formations of enormous diversity, with both the plant and animal resources directly managed by the Lakandon (Nations and Nigh 1980; Durán Fernández 1999; Levy Tacher et al. 2002; Nigh 2008; Diemont and Martin 2009). Careful weed management extended the useful life of the field for annual crop production, allowing 4–8 years of high-yield, continuous cropping, in contrast to the less intensive milpas that exist in this region today that, under conventional weeding or herbicide treatment, can be planted for a maximum of 3 consecutive years before being overwhelmed by herb and shrub competition.

Ecologists no longer subscribe to “equilibrium” models of succession, in which woodlands were believed to return spontaneously to a “climax” state after a disturbance (eg wildfire, blow-down, clearing for cultivation). However, the processes of tropical forest succession are known to develop as a series of stages, usually leading to some form of closed forest. This succession is driven by groups of dominant woody species defined by morphological, phenological, biochemical, or trophic responses (functional groups) at each stage (Chazdon 2008). The Lakandon recognized and named these stages and the associated functional groups of tree species (Nigh 2008). The species composition of forests derived from traditional Lakandon milpa are much more diverse than those that derive from contemporary milpa practices. Thus, forest recovery is hastened under this traditional management system (Levy Tacher and Aguirre Rivera 2005).

In one management practice, Lakandon farmers dispersed seeds of the balsa tree (Ochroma pyramidale) in order to create thick stands of this early successional, fast-growing, short-lived canopy species (Levy Tacher and Golicher 2004). Ochroma has been used by generations of Lakandon farmers to accelerate forest regeneration, replenish soil organic matter, and improve weed control. For example, common bracken (Pteridium aquilinum), which is a pyrogenic species, has become a serious invasive weed problem in the region, as traditional control methods (such as planting Ochroma; Figure 3) have been abandoned (Schneider 2004; Vester et al. 2007; Turner and Sabloff 2012). At least a dozen other tree species were also managed for their beneficial effects on soil fertility and succession (Levy Tacher 2000; Diemont et al. 2006; Roman Dañobeytia 2012). In this way, the Maya obtained a selection of species of value to humans during the process of forest regrowth.

Figure 3.

A stand of balsa (Ochroma pyramidale) deliberately planted by Lakandon forest gardeners to aid in the control of invasive, fire-resistant bracken (Pteridium aquilinum).

The intensification of the Maya milpa has been studied from the standpoint of shortening the fallow period, which usually resulted in declining soil fertility and decreased ecosystem diversity (eg Lawrence et al. 2007; Ochoa-Gaona et al. 2007; Parsons et al. 2009). Merely shortening the fallow period in swidden production is actually a type of de-intensification, a shift to a less demanding cultivation system that results in decreased productivity. In the high-performance milpa, shortening of the fallow period is achieved by increasing the effort applied to management of successional stages (Johnston 2003). The practices developed by the Lakandon Maya in times of low population density and isolation in the remote tropical forests of Chiapas and Guatemala, between the 16th and 19th centuries (Palka 2005), continued commonly until the end of the 20th century. Families lived where they worked, and most of the household's subsistence was acquired directly from the milpa and derived woodlands. A few products, such as tobacco, were occasionally traded for outside goods (such as steel tools, cooking oil, and other industrial products). The 21st century provides a very different context for the inhabitants of today's Lacandon rain forest (McGee 2002).

Fire, nutrient cycling, and soil ecology

Over the past decade, scientists have thoroughly revised the conventional view of the nature of soil in the neotropics. The former stereotype of “tropical soils” as nutrient-poor, highly weathered and leached, and inherently limited in productive potential has been questioned from a number of viewpoints (Sanchez and Logan 1992). First, land in the humid tropics is far more heterogeneous than the stereotype implies (Vitousek and Sanford 1986). For example, in the regions that supported the ancient Maya civilization of southern Mexico and Central America, not only do ecological and soil conditions vary widely (Dunning et al. 1998), but extensive areas of the Maya tropical lowlands are dominated by mollisol, “one of the world's most agriculturally important and naturally productive soils, particularly under conditions of rain-fed cultivation” (Fedick 1996).

Even in those areas where soil types conform more closely to the conventional image of tropical soils, such as the nutrient-depleted oxisol of upland Amazonia, much attention has focused on the pockets of nutrient-rich, dark, anthropogenic soil, which features a stable organic matter content; these soils, known as terra preta do indio or Amazonian dark earths (ADE), are much sought-after by contemporary farmers. Among the interesting properties of this soil is that it has considerable long-term carbon (C) sequestration potential (Sombroek et al. 2003; German 2004). Although the pre-Columbian anthropogenic origins of ADE are widely accepted (Woods and McCann 1999; Peterson et al. 2001), the processes by which these anomalous soil types were formed and have maintained their fertility over centuries remain uncertain, with several explanations having been suggested.

One of the first theories about ADE is the “midden model”, which states that such areas are essentially prehistoric garbage dumps, reflecting long-term human occupation; one study concluded that all seven dark earth sites investigated along the middle Amazon valley derived from ancient middens and were enriched with materials originating in the adjacent floodplain (várzea; Lima et al. 2002). The authors of that study propose a limited distribution for ADE in the Amazon and question whether such soil supported the dense permanent settlements others have proposed (Denevan 1996). Contemporary examples suggest that the midden model may explain the creation of some ADE. However, a survey of a large number of sites in the Tapajós and Arapiuns river basins of the lower Amazon revealed that while some sites were middens, most were not. These authors concluded that the most plausible explanation for the high organic matter and nutrient content of ADE sites is “long-term soil management practices (especially mulching and burning)” (McCann et al. 2001).

Despite this evidence, however, it is generally agreed that the most common type of agriculture in the region, swidden, could not be a source of ADE special characteristics. Balée (2010) summarizes this view: “Charcoal is produced by incomplete combustion (sometimes called a ‘cool burn’) of organic material and it is what is responsible for retaining SOM [soil organic matter] at high levels in ADE, which cannot result from slash-and-burn cultivation, a ‘hot burn’”. However, this reasoning does not take into account that swidden farmers possess many techniques for controlling burn temperature and other factors. Our data suggest that the swidden cycle of the Lakandon Maya contributed to black C and other characteristics of ADE (cf Rumpel et al. 2006).

Regeneration of secondary vegetation following swidden cultivation has a positive effect on soil fertility, increasing nutrients and organic matter. The contribution of ash to nutrient flow in swidden soil is amply documented, at least since Nye and Greenland's (1960) classic study. Pest control also may be favored by burning crop residues (Cochrane 2003). In some geologies, such as the limestone base that underlies the forest throughout most of the Maya area, burning releases calcium – essential for crop production – from the parent rock (Faust 1998; Lawrence et al. 2007). In general, burning reduces weeds, releases soil nutrients, replenishes nitrogen, and adds phosphorus, potassium, magnesium, and manganese contained in the ash of the burned woody vegetation to the soil (Sanchez 1976; Wilken 1987; Ochoa-Gaona et al. 2007; Schmook 2010). Charcoal added to soil increases nutrient storage, which counters nutrient removal by leaching (Glaser et al. 2002; Lehmann et al. 2003).

If milpa is characterized as a rotation between cultivated fields and secondary forest, then a function of woody vegetation is to build soil fertility after cultivation through nutrient cycling. Deep-rooted trees retrieve nutrients that have been leached from the upper layers. Data suggest, for example, that Sapium lateriflorum, a successional tree managed by the Lakandon Maya, serves as a phosphorus pump, bringing this limiting nutrient from deep soil through leaf litter (Diemont et al. 2006). Leaf fall also provides raw organic matter to store nutrients and contributes to high biological activity and good soil structure, while decomposition of organic matter fuels the soil food web. The Maya plant and protect trees they believe enhance and accelerate post-agricultural succession (Levy Tacher and Golicher 2004). Another important role of woody vegetation is as a source of ligneous material that under conditions of low-temperature pyrolysis creates a form of black C known as biochar.

Biochar is a primary ingredient in all variants of ADE and is also known in other agricultural contexts as an important soil additive (Lehmann et al. 2003). Biochar's positive effect on agricultural soil appears to be in increasing the surface area over which biological activity can take place. Charcoal positively affects cation-exchange capacity, as well as nutrient and water storage capacities (Liang et al. 2006). The hard, polycyclic aromatic C particles that result from low-temperature pyrolysis (300–500° C) are chemically and biologically stable, persist for centuries, and confer long-term soil characteristics.

Indigenous weeding practices have important implications for other aspects of agroecosystem management, in particular the use of fire when compared to conventional milpa, as practiced by recent colonists. In conventional milpa, the entire field is weeded in one sustained effort, two or three times during the cultivation cycle. After the last such clearing of the field, weeds are allowed to proliferate, as the maize crop is fully formed and needs only to dry in place before harvesting. This means, however, that annual weeds grow and reseed prolifically, requiring a hot burn over the entire field to prepare for the next planting cycle. By contrast, in traditional Lakandon practice, small piles of weed or crop residues are burned occasionally throughout the year and the resulting ashes and charcoal are spread about; a hot burn over the entire field occurs only once in the 25-year milpa forest cycle, when the vegetation is felled to initiate cropping. Most of the weeds are not burned at all but are left in the field to decompose, thereby providing a continuous supply of organic matter.

The low-intensity burns result in incomplete combustion of vegetable material (Figure 4) and addition of pyrogenic charcoal, reminiscent of the “slash-and-char” method suggested by Lehmann et al. (2003) or the “sweep-and-char” analogy suggested by Winklerprins (2003). This critical difference in fire intensity between conventional, contemporary milpa (or slash-and-burn) systems and the traditional Maya system is not usually appreciated, which may be why some researchers believe that ADE are not usually formed under swidden agro-forestry (Denevan 2006).

Figure 4.

Partially burned vegetation provides a gradual but continuous input of biochar and organic matter to the milpa soil.

Preliminary results on Lakandon milpa fields show increasing black C with soil depth, suggesting accumulation of black C over time, as fields are reused in repeated cycles of fire, cultivation, and directed secondary succession (Table 1). Burned trunks and large branches from the original clearing are distributed throughout the field and shed charcoal continually (Figure 4), contributing to the highly enriched anthropogenic soil observed on Lakandon fields (cf Wilken 1987). High levels of organic matter and phosphorus, as well as other characteristics (Mendoza-Vega and Messing 2005), are similar to the terra preta of the Amazon (Woods and McCann 1999; Peterson et al. 2001; Glaser et al. 2001; Balée 2010). Properly managed, milpa soils can be more fertile after each cultivation cycle.

Table 1. Properties of Lakandon milpa soil by sample depthThumbnail image of

Fire, however, can also be a negative factor, and there is a distinct advantage to incorporating crop residues and other organic matter directly into the soil rather than as ash. The goal of fire management in the milpa must be a low-temperature burn that creates charcoal and conserves as much organic matter as possible.

Burning a Maya milpa swidden

Opening a new swidden in woodlands involves cutting the existing vegetation, allowing it to dry in situ, and then burning it. In the following description of the burning activity in Xocen, a Maya village in the Mexican State of Yucatan (Terán and Rasmussen 2009), we consider how Maya farmers control the burning temperature in milpa swiddens. In practice, a great deal of knowledge and skill are brought to bear on the implementation of these activities.

Burning ceremony

Terán and Rasmussen (2009) report that in Xocen a ceremony to request a good burn from the spirits of the Earth is now only rarely observed. The ritual serves as an occasion to plan the upcoming burning activity and to exchange knowledge about burning techniques and safety. The Maya farmers consider it to be a “good burn” when all the dried leaves, small branches, and trunks – what in English we might call brush or slash – are completely converted to ash. Larger woody debris is only charred and becomes a source of charcoal to further enrich the soil. Winds should be gentle but steady and stable, and should blow only from one direction during the burn. An unanticipated shift in wind direction or an unexpected rain shower can cause the burn to fail. If the fire temperature attained is not too hot, considerable charred vegetable remains are left that will gradually be incorporated into the swidden soil (Figure 4).

Timing

An ideal burn depends on the external conditions of wind, humidity, and temperature that facilitate or complicate control of the fire. The brush must be completely dry, yet if too much time has passed since the field was cleared, weeds will have returned and may need to be cut again. Thus, the Maya farmer chooses the timing of the burn carefully. Ephemerides (calendars of ritual and other activities) in the Classic Maya codices depict gods in the act of using a wooden drill to make fire, associated with certain dates in the Maya calendar; the context suggests that fire-making was considered sacred.

Burning the milpa is usually a fairly brief event, lasting 45 minutes to an hour for a 1-ha field. Time of day influences the quality of the burn. In the early morning, temperatures are cool and the fallen vegetation is more humid, which may impede the combustion of the brush. Later in the day, the brush is dry and warm and burns easily. However, the fire can burn out of control and jump to surrounding areas should the wind suddenly shift or increase in strength.

Firebreaks

Firebreaks – strips of land that have been cleared free of vegetation and debris, forming a perimeter around the clearing – are essential for fire control. Their characteristics vary greatly, according to specific local conditions; these are not discussed here.

Wind tending

Using fire to remove most of the brush and litter from an area of land while keeping the fire under control and preventing it from jumping the firebreak and burning adjacent fields or converting the entire biomass to ash requires considerable skill. Individuals known as “wind tenders” in the Yucatan (yum ik'ob in Yukatek) carry out the burns. One common practice is to burn against the wind. For example, if the dominant and strongest winds are from the south-to-southeast, the two or more wind tenders begin in the northwest corner and proceed along the firebreaks on opposite sides, spreading the fire out as much as possible. Under ideal conditions, two teams of wind tenders work with rustic tools fashioned from fresh branches taken from leguminous trees to spread the fire throughout the dried brush, keeping the fire burning in a controlled manner. The two teams start from a single point and spread the fire throughout the field by following opposite perimeters and meeting again in the far corner of the field. In some cases firing proceeds first against the wind and then back again with the wind, resulting in two smaller fires rather than one intense fire, thereby allowing for more moderate fire temperatures (Terán and Rasmussen 2009). Other firing patterns have been reported (Zizumbo and Sima 1988), while additional practices may include keeping barrels of water at the site and using wet branches to keep fires cool and under control (Hernández Xolocotzi et al. 1995).

Conclusions

Much skepticism about the potential of milpa as a sustainable agroecosystem stems from its minor role as a mode of production in contemporary society (Terán and Rasmussen 2009). In pre-Hispanic times, the milpa system would have been the primary form of agriculture in Mesoamerica and was supported by the aristocratic class, who recognized it as the basis for successful subsistence. Current milpa systems can vary greatly in form, productivity, and intensity, primarily as a result of increasing marginalization of small-scale farmers. While the basic milpa systems practiced today share all the characteristics described in historical records on Maya farming (Terán and Rasmussen 1995), there is a failure to distinguish among different milpa strategies that have evolved since the Spanish conquest; they are treated indiscriminately, in part due to the tendency to focus exclusively on the maize component of the system.

The extensive milpa production that we call marginal or contemporary, and which is common throughout Mesoamerica today, must be distinguished from the traditional high-performance milpa as it has been practiced in historical settings for millennia (Wilken 1971; Nations and Nigh 1980; Terán and Rasmussen 1995 Terán and Rasmussen 2009). The primary difference between the high-performance milpa and de-intensified, marginal forms lies in the degree of labor and skill the farming family dedicates to the milpa cycle. During the phase of annual cropping of maize, beans, squash, and associated crops, considerable labor is required to manage the complex multi-cropped field. Later stages in the milpa cycle also require skilled labor inputs if the full range of milpa resources is to be marshaled for household subsistence and trade. One of these skills is the management of fire. As one Amazonia researcher states: “Fire is a defining feature of swidden agriculture. For a number of reasons this complex form of land management and its key technology have been cast as villains in the drama of tropical development, a process that may well have obscured a more careful and perhaps more fruitful analysis of the roles that fire and succession might have had in the creation of soil fertility over longer stretches of time” (Hecht 2003).

The detailed traditional ecological knowledge driving the Maya milpa is of great value today as we face the urgent need for ecological restoration and the renewal of rural livelihoods. Unfortunately, the current policies of many governments, both national and local, do not reflect this need. Lakandon farmers, for example, receive government subsidies under an environmental services program to end burning activities and allow their forests to grow unmanaged. This policy is ill-conceived. Under traditional Lakandon practice, milpa swiddens sequester C for long periods of time; in ADE, black C was found to be stable over centuries (Glaser et al. 2003). Under current C credit programs, however, C is sequestered temporarily until the forest ages, degrades, and perhaps burns spontaneously in a hot burn, with stored C then being returned to the atmosphere. These C credit programs should be discontinued, since they do not result in net C sequestration and also result in the loss of valuable traditional knowledge.

In recent years, concern has been expressed about the fitness of our current model of industrial agriculture to meet the demands of sustainability, biodiversity conservation, and food security for all (eg Matson et al. 1997; Altieri and Toledo 2011). Doubts are also being expressed about the current trends in milpa practice (Lawrence et al. 2007; Schmook 2010). The agroecological memory developed over 10 000 years of Maya agrarian civilization – and especially the potential role of fire in increasing soil fertility – is worth some consideration.

Acknowledgements

Fieldwork informing this article was supported in part by the Centro de Investigaciones y Estudios Superiores en Antropología Social (CIESAS; Mexico); the State University of New York, College of Environmental Science and Forestry; the Ohio Agricultural Research and Development Center; the Fulbright Program; and NSF Grant OISE-0431230.

Ancillary