Athena Review 2,1: Early plant domestication in Mesoamerica

free trial issue subscribe back issues

Athena Review Vol.2, no.1

Early plant domestication in Mesoamerica

Mesoamerica (fig.1) is a cultural super-region extending from northern Mexico to Central America, whose diverse environmental zones (arid; mountainous; tropical lowlands) each became ancestral areas for major crops ranging from squash, corn, and beans to cotton and chocolate. Regionally distinct types of food production emerged as the subsistence base of a wide range of early civilizations, and the direct precursors of some of today's most important crop plants. Much of their origin, however, still remains a mystery.

[Fig.1: Environmental zones and major sites of early plant cultivation in Mesoamerica (after McClung 1992 and MacNeish 1992).]

Domestication versus Cultivation: A domesticated plant, strictly defined, is one whose reproductive success depends on human intervention. It is often quite difficult to apply this term accurately to plant taxa represented in the archaeological record. Archaeobotanical specimens (ie., charred cobs of corn or maize) are recognized as possible early cultigens because they possess attributes that fall in between those of known wild progenitors and those of modern domesticates. In some instances, anatomical features may be present that suggest that the species would do poorly on its own in the wild, and, therefore, are unlikely to have been the result of natural selection.

For example, inefficient seed dispersal mechanisms (as in the maize ear), while greatly diminishing a plant's chances for successful reproduction in the wild, are always advantageous for seed-harvesting humans, due to more retention of seeds in the pod. In a great many instances, however, the deteriorated and fragmentary plant remains recovered from archaeological contexts lack the definitive anatomical clues allowing determination of just how far along the road to "true" domestication the specimen may have been.

Wild-Food Production: In contrast to collection (also called foraging or gathering), wild-food production means behavior intended to alter the abundance and distribution of plant resources. The term, while somewhat misleading if we take "production" in the conventional sense of increasing yield, remains a useful concept for energy investment in plant resource extraction, involving a group of strategies designed to reduce collection time. This is mainly done by reducing travelling time to the source areas in various ways. One is the protection of known stands of exploitable plants by "tending," to assure a continous availability at known, convenient locales. Other intensive gathering or "tending" methods include weeding, discouraging predators, pot-irrigation, and limited harvesting to ensure reproduction.

Another obvious time-saving solution and precursor to domestication is the direct transplantation of desirable resources closer to a camp or settlement. Probable candidates for transplantation are mainly small trees or bushy plants with a perennial growth habit, whose produce (sap, fiber and wood as well as edible portions) is readily harvested and processed for consumption. In Mesoamerica this fits such plants as avocado, cacao, capulín, ciruela, cotton, mesquite, papaya, ramón, zapote blanco, and zapote negro, among others. Other easily transported, edible species include cacti (ie., prickly pear), maguey, and fruit-bearing palms such as coyol. Many, though not all, of these species can be propagated with cuttings.

To document early transplantation, it must be demonstrated that a species occurs in prehistoric sites outside of its natural habitat. Good evidence for this practice exists from some occupation strata in the dry cave sites of the Tehuacán Valley .from at least 3500 BC and perhaps much earlier (with dates for wild avocado proposed for ca.7000-6500 BC). Archaeologists agree that the remains of avocado, ciruela, coyol, zapote blanco and zapote negro represent exotics introduced by humans into the Tehuacán Valley. All of these species are native to considerably more mesic (wetter) environments, and their establishment in the semiarid Tehuacán area would have required some irrigation (McClung 1992; Pearsall 1995; C. Smith 1987).

The Tehuacán cave deposits are extraordinary for their preservation, richness and depth. More usual is that the practices of tending and transplantation may be archaeologically undetectable. Transplantation does not necessarily involve the crossing of biotic zones. More often, traces of plants we suppose to have been exploited regularly by prehistoric peoples simply do not turn up in the excavated material for one reason or another. The ramón seed, for example, a ubiquitous and valuable resource available to occupants of the tropical lowland forests, generally is not found in archaeological deposits, and is not likely to be found owing to its soft, highly perishable seed coat (Fedick 1995). Thus, we are unable to track its history as a cultigen. The archaeobotanical record for Mesoamerica is full of such huge gaps regarding when and how the dispersal of economically useful plants were effected via human cultural practices.

Cotton and its dispersal: A prime example of this incomplete state of knowledge on plant origins is cotton (Gossypium spp.), which has both Old and New World species. We do not know when or where prehistoric Mesoamericans first began to appreciate the utility of the cotton plant, even though something is known about the complex distribution of wild populations in the New World. As far as has been determined, wild cotton occurs in distinctive and restricted littoral (shoreline) or related habitats. It must have taken considerable time for the species to have dispersed so widely throughout tropical and subtropical Mesoamerica, and to have developed the long-fibered seeds for which it is so highly valued. The seeds of the littoral wild cottons, perennials which do not propagate vegetatively, have very sparse fibers. The oldest known archaeological samples of Mesoamerican cotton are from the Tehuacán Valley, dated 3500-2300 BC; these represent a fully domesticated form, Gossypium hirsutum (Smith and Stephens 1971).

[Fig.2: Map of New World cotton species (after Fryxell 1979).]

Cotton may possibly have been introduced to Mesoamerica as an already domesticated form from the southern hemisphere where, in coastal Ecuador and Peru, cultivated cotton of another species (G. barbadense) has been documented for much earlier periods. Yet most views favor the independent domestication of G. hirsutum in Mesoamerica (Fryxell 1978; Heiser 1990; Pickersgill 1989). Confirmation of this judgment is needed from the archaeological record.

Adding to the general uncertainty concerning the earliest attempts to manage wild-food production is that mere collection behaviors may have side effects which contribute to the domestication process. Transportation of seed-bearing fruits to a settlement may enable a species to spread beyond its original "natural" range. Newly established populations may develop characteristics different from those of parents if environmental conditions (ie., altitude, soil conditions, precipitation) foster such adaptations, or if "unconscious selection" operates because collectors habitually harvest specimens with particular traits such as large size, ease of detachability from the stem, or lack of spines. The end result is to promote speciation, even when no food production techniques were practiced.

Straightforward interpretations of archaeobotanical assemblages are rarely possible. Social factors including group size, division of labor, and degree of sedentism, as well as technological factors, all must be considered in reconstructions of subsistence related behaviors. Wild-food production could fit within most social organizations, but presumes at least a single season's residency at one place. Fully mobile groups obtain no subsistence advantage by investing time in propagating or maintaining resource stands which are visited infrequently. The Coxcatlán Cave data from the Tehuacán Valley show a predominance of fruits gathered from perennials (trees, agave, and cacti) in the earlier levels of ca. 7000-3500 BC (MacNeish 1992; Pearsall 1995). Wild grasses and annuals were certainly available but appear not to have been high priority foodstuffs. Likewise, no grasses at all were reported for the Archaic period (8900-6670 BC) occupations at Guilá Naquitz Cave in the Oaxaca Valley near the town of Mitla (C. E. Smith, 1986). Kent Flannery (1986) and colleagues, who excavated Guilá Naquitz, emphasize the selectivity of the occupants with respect to number of plants exploited compared with the great variety available. Archaeologist Brian Hayden (1995) warns that, since sites like Coxcatlán Cave and Guilá Naquitz "were peripheral to the main valley ecosystem[s] and may have been visited only infrequently," the botanical remains recovered may not represent the full range of resources utilized. This idea warrants serious consideration. Nonetheless, there is no reason to believe that, in a moderately bountiful environment, people would have had no food preferences and simply consumed indiscriminantly whatever lay in their path. The choice of easily harvested, relatively large food packages that required minimal processing prior to consumption fits behavioral expectations for band-level foraging groups with a fairly high degree of residential mobility.

Cultivation: The line drawn between wild-food production and cultivation can be a thin one that rests mostly on assumptions regarding the effort made by prehistoric peoples to manipulate the growth of plants. According to Harris, cultivation, as the second level of energy investment, involves systematic soil preparation and planting as well as tending and harvesting. Initially, cultivation can be thought of as an extension of the collection-time reduction principle, where the major goal is ensuring availability. However, as more of the time saved by not collecting is rerouted into field and crop management, it will be seen that such attention often improves the health and yield of the subject plants. Cultivation, therefore, ultimately comes to represent a commitment to improving the product itself and its productivity, in addition to availability.

Depending on the genetic malleability of the species, selection for desirable traits will become a conscious and prominent aspect of the process. In the archaeological record, the plant taxa identified as the earliest cultigens are not the same sort of plants that were the focus of wild-food production. In general, they are annuals or short-lived perennials that readily colonize disturbed habitats, such as human occupation sites. It also happens that most of the early Mesoamerican cultigens represent much smaller food packages than, for example, collected tree fruits, and require extra time to process for consumption. The energy trade-off only makes sense where there is at least part-time sedentism and a group sufficiently large for some members to devote themselves to time-consuming tasks of processing small food items.

Cultivated Plants: Several species of the squash family (Cucurbitaceae; genus Cucurbita) are among the earliest known cultigens. Cucurbita pepo, the pumpkin squash (fig.5), has been dated to 8000-7000 BC at Guilá Naquitz and at Ocampo Cave in Tamaulipas to about 7000-6500 BC (Heiser 1989; B. Smith 1997). The specimens were identified as cultivated forms on the basis of morphological criteria and, as the Guilá Naquitz material has been dated directly, C. pepo is confirmed as the oldest known cultigen for Mesoamerica (see AR, I,3). C. pepo, along with the other domesticated cucurbits known from archaeological contexts (C. argyrosperma, C. ficifolia, C. lundelliana and C. moschata), appear to have arisen from different wild species. Although none of the wild cucurbit progenitors has been identified with certainty, it is believed that the centers of cultivated species diversity are the areas of origin. The ancestral squashes would have been cultivated principally for their nutritious edible seeds, since their scant bitter flesh would have required repeated boilings to eliminate the unpalatable cucurbitacins (Nee 1990). Selection for larger, fleshier, non-bitter fruits could have been accomplished relatively rapidly given the growth habit of these species and their capability of self-pollination. Another member of the Cucurbitaceae family, Lagenaria siceraria, the calabash or bottle gourd, is thought to have been brought under cultivation for its usefulness as a container, owing to its unusually thick rind, rather than for food. The bottle gourd also occurs very early in several Mesoamerican archaeological deposits (Tamaulipas ca. 7000 BC, Tehuacán ca. 5050 BC, Guilá Naquitz ca. 7000 BC). It is usually encountered as rind fragments and it is not known exactly when Lagenaria became domesticated.

[Fig.3: Seed of an early cultivated squash, Cucurbita pepo (after B. Smith 1997).]

The evidence for bean (Phaseolus spp.; Fabaceae) cultivation is not of such antiquity as that for squash, although ultimately beans became the high protein staple of the Mesoamerican diet. It is sensible to presume that beans were cultivated quite early given that they have many of the same characteristics that made the squashes such willing domesticates. Wild beans had a significant presence in the older levels at Guilá Naquitz of ca. 8900-6500 BC (Kaplan 1986). The kinds of remains suggest that in Preceramic times, beans had been harvested for their tender shoots or pods. Cultivation and selection for certain seed characteristics probably followed the introduction of ceramic technology and attendant cooking techniques that enhanced the seeds' desirability as food. The transition from noncultivated to cultivated forms is not clear in the archaeological record, but recent biochemical investigations strongly indicate that the common bean, Phaseolus vulgaris, evolved independently in Mesoamerica from the common wild bean, with a likely center of domestication in Jalisco in West-Central Mexico (Gepts 1990; Singh et al. 1991).

In the tropical lowlands, starch in the diet is more likely to have come from roots or tubers than from cereals. Unfortunately, the archaeological record for lowland Mesoamerica has not provided generous amounts of data with regard to possible root crops. The case for manioc (Manihot esculenta; Euphorbiaceae) cultivation is the strongest, especially since some phytogeographers favor a Mexican origin for the species (Hawkes 1989; C.E. Smith 1987). Variants with large underground storage organs would survive best in areas with a marked dry season where such reserves would have been adaptive. But, few sites have yielded direct evidence for manioc consumption. Root crops are rarely found as carbonized remains and desiccated specimens are extremely hard to identify. The starchy pulp that constitutes the comestible portion has hardly any distinctive morphological characteristics. M. esculenta pollen has been identified in freshwater cores from northern Belize. On the basis of the pollen data, the researchers estimated a date of ca. 3400 BC for the introduction of manioc. Since this date roughly coincides with that for the appearance of maize as well as indications of local deforestation, they conclude that the manioc was cultivated (Pohl et al. 1996). Manioc has been identified for Formative Period Cuello, also in Belize (Fedick 1995). These identifications are open to doubt, and the major arguments for manioc cultivation still rely on inferences based on availability, the weedy tendencies of the plant, and the technical simplicity of cultivating it. We do not know whether sweet and/or bitter varieties were consumed, though the latter demands a great deal more processing prior to consumption. The claim for manioc at Tehuacán is extremely shaky, as is the claim for yam (Xanthosoma sp.; Araceae) at Tikal, and no evidence has yet come to light for the prehistoric cultivation of jícama (Pachyrrhizus erosus; Leguminoseae).

Many obstacles present themselves to archaeologists wishing to document the early practice of cultivation, not the least of which is the lack of straightforward evidence. The most convincing proof of cultivation is, of course, intentional selection for desired traits. Often, however, archaeobotanists have only the most limited kinds of morphological indicators with which to make such determinations, such as an increase in seed size over time. As so many of the cultigens were initially ruderals (ie, growing in trash middens), it is very possible that they began to acquire such characteristics as an increase in fruit size simply because they were doing so well in an amenable environment. For example, the highly nitrogenous soils associated with human habitation sites can allow plants to achieve larger than "normal" size without any intentional selection for large size on the part of incipient cultivators (ie., incidental domestication).

The case for purposeful cultivation is made more plausible by combining archaeobotanical and palaeoenvironmental information with an understanding of the social context in which cultivation behaviors are believed to have occurred. As noted above, time taken from collection to devote to cultivation presupposes certain social organizational features that make cultivation an efficient and attractive strategy. Still, the practice of cultivation covers a great range of energy input levels, from marginal "back-door" gardening to a high degree of dependence on the products of cultivation. Many of the products of cultivation are storeable resources. Evidence of storage, in the form of food caches or specialized facilities, can help us gauge the relative importance of cultivation as a subsistence strategy (see also Great Basin article, this issue).

Agriculture: In Harris' "scale" of people-plant interactions, agriculture represents the highest level of energy investment. Unlike incipient cultivation, where hunting and gathering remain important, once the stage of agriculture is reached, the group is solidly dependent on food production for the greatest proportion of its subsistence needs. According to Harris, what differentiates cultivation from agriculture is not precisely the kinds of behavior involved, but rather the scale at which they are practiced. Agriculture is just highly intensive cultivation that, on the whole, leaves little time for other subsistence methods. Its practice presupposes the prior existence of suitable plants for intensive cultivation, or domesticates. So, whereas the activity of selection is the primary clue to the beginnings of plant cultivation, it is the goal of increased productivity of species with known desired qualities that defines agricultural practice.

For reasons of scale, evidence for agriculture is highly visible in the archaeological record and is unlikely to be misinterpreted. From the standpoint of the botanical remains, the inventory of food items will be dominated by one or a few cultigens. In Mesoamerica, only one domesticated plant, maize (Zea mays subsp. mays; Gramineae) ever became the object of large-scale, intensive agriculture. Often readily detectable, as well, are related modifications to the landscape from clearance, ground levelling via terracing or other means, and irrigation or other water management schemes (ie., diversion, drainage, and containment). Perhaps most significant is our knowledge that sizeable sedentary populations could not be sustained in many environments without large-scale food production.

Maize and its origins: Despite the ease with which the practice of agriculture can be identified archaeologically, the transition to agriculture remains one of the most poorly understood phenomena in Mesoamerican prehistory. This is largely because the evolutionary history of maize needs unravelling. For agriculture to be successful, the cultivated crop must possess certain attributes in order to justify the enormous efforts expended in land tillage, planting, tending and harvesting. Above all, the yield of usable produce per plant must be quite high. This criterion is met by maize by virtue of its remarkable ear and is the reason maize became the most important crop by far throughout Mesoamerica. The question is, how and when did the fruitful maize cob develop?

Referring to these mysterious origins, Kent Flannery (1986) has called maize the "most controversial and most enigmatic of any major cultivated plant." All specimens of primitive maize known from archaeological sites are domesticated forms. Maize is not included in the taxa considered above under "wild-food production," nor does it appear on the list of the earliest cultigens. This is because it simply wasn't there. Identifying the wild progenitor(s?) of maize means reconstructing genetic events with little material corroboration, and so has been the subject of much speculation and debate for at least 100 years. Opinions are numerous, but currently the three leading hypotheses are: the Wild Maize Hypothesis, the Orthodox Teosinte Theory, and the Catastrophic Sexual Transmutation Theory (as shown in Benz and Iltis 1992).

The Wild Maize Hypothesis claims that there must have once been a wild race of maize, now extinct, from which the cultivated form evolved. This theory explains the development of the cob by asserting that the hypothetical ancestor was a naturally occurring pod-corn with vertical rows of kernels (polystichy). The problem with this theory is that no fossilized examples of the putative ancestral maize has ever been found. Prior to domestication this wild maize must have been exploited to some extent by prehistoric people and, therefore, should have found its way into the archaeological record.

The Orthodox Teosinte Theory says that maize was derived from the wild grass teosinte (Zea mays mexicana). In some classifications annual teosinte is treated as a separate species (Zea mexicana). The basis of this argument is that teosinte is the closest known relative of maize. Some reject teosinte as the direct ancestor because the very minute female inflorescences ("ears") of teosinte seem unlikely to have been transformed into the huge polystichous maize ears by means of human selection. A variant of the Teosinte Theory, which serves to reconcile it with the Wild Maize Hypothesis, is that maize is a hybrid of teosinte and a wild maize ancestor. In other versions teosinte and another related grass are the hybridizing pair (Mangelsdorf 1986).

[Fig.4: Single-rowed ear of Teosinte (after Mangelsdorf 1976).]

Recently, considerable support for teosinte as the direct ancestor of maize has accumulated as a result of DNA and isozyme (genetically variant enzyme) studies that provide measures of genetic distance much more reliable than judgments based on morphological characters. These investigations have revealed such a high degree of similarity between teosinte and maize with respect to the presence and frequency of specific molecules that a close phylogenetic relationship must be a matter of fact (Doebley 1990, 1992). In other species such a level of similarity might not be so remarkable, but maize is so exceptionally polymorphic for the traits studied (Hoisington 1992) that the calculations of genetic distance are very refined. Moreover, the molecular evidence has identified a particular race of teosinte (Z. mays subsp. parviglumis), a perennial grass, more similar to maize than any other race of teosinte. This finding supplies a probable geographic origin for maize domestication, as well as an immediate ancestor. The wild Z. mays subsp. parviglumis race that best matches the isozyme profile of maize has a limited distribution today in the central portion of the Balsas River drainage in northern Guerrero, eastern Michoacan and western Mexico states (Doebley 1990). Archaeologists in general are impressed by the statistical soundness of the perennial teosinte's claim as the progenitor of maize and the view is widely (though not unanimously) accepted in the discipline. The Orthodox Teosinte Theory, however, does not resolve the questions about the development of the maize cob. Botanists are harder to convince and continue proposing other evolutionary reconstructions.

The Catastrophic Sexual Transmutation Theory, first advanced by Hugh Iltis, offers an explanation for the development of the maize ear while completely accommodating the Teosinte Theory. Briefly, the theory holds that through a series of genetic mutations the staminate teosinte tassel spike (the male reproductive organ) was transformed into the unisexual distillate maize ear (the female organ). This contrasts with conventional thinking that has the maize ear developing from the teosinte ear. But Iltis claims that there is greater structural homology between the teosinte spike and the maize ear than between the two ears. Furthermore, the maize ear exhibits vestiges of its former male anatomy. The radical and rapid anatomical changes associated with transmutation provides explanations for the late development of maize as a cultigen and the failure to find "wild maize" in the fossil record. Such evolutionary adaptations may have been encouraged by volcanic activity typical of the Balsas River region in the recent past, and sexual transmutation may have occurred as late as 8000 years ago (Benz and Iltis 1992). The Catastrophic Sexual Transmutation Theory is a complex argument not uniformly embraced by maize researchers (Wilkes 1989).

How the maize ear developed is a critical question for the timing of maize domestication but, as of yet, no consensus exists. Numerous examples of primitive maize obtained from archaeological excavations suggest that cultivated maize diffused quite swiftly throughout Mesoamerica, especially if we accept a single center of domestication and a date of about 8000-6000 BC. Tehuacán in the central highlands of Mexico has yielded the oldest remains of cultivated maize in Mesoamerica, dated to 5050 BC (calibrated) by accelerator mass spectrometry dating of actual specimens. Racial diversification arose as a result of adaptations to different environmental conditions and human selection for traits related to productivity, taste and cooking properties (DeWet 1992). Within a few thousand years, maize had become a staple food crop throughout Mesoamerica, and was a main subsistence base of the Maya, Aztec, and related early civilizations.

[Fig.5: Aztec maize offering depicted in the early 16th century Codex Borbonicus.]

(Abbreviated from the article by Virginia M. Betz on pp.24-31 in Athena Review, Vol.2, no.1.)

Athena Review Image Archive™ | Paleoanthropology in the News | Guide to Archaeology on the Internet | Free issue | Back issues