Human Selection and Domestication of Chia ( Salvia Hispanica L.)

Abstract

French

The plant species Salvia hispanica L. has a long history of plant-human interaction. This study defines the human-selected morphological characters associated with domestication of S. hispanica and presents a review of ethnobotanical and historical information pertaining to the human selection forces that resulted in changes in morphology. The compiled ethnobotanical information for S. hispanica is applied to the framework of the hypothetical ecological-evolutionary continuum of plant-human interactions, highlighting the contributions of ethnobotanical data to our understanding of plant domestication processes.

Keywords

domestication Mexico ethnobotany agriculture

Introduction

This paper examines traditional plant management techniques and direct human selection as forces effecting morphological changes in the crop plant Salvia hispanica L. from an ethnobotanical perspective. In pre-Columbian Mesoamerica, the species known commonly as “chia” was a major commodity and its seeds were valued for food, medicine, and oil (Cahill 2003). After Spanish contact and colonization, the level of cultivation of the species plummeted. Limited cultivation and traditional medicinal, culinary, and religious uses persist today; however, the origins of cultivation and the domestication process are unknown (Cahill 2003). The ecological-evolutionary-economic model of plant domestication developed by Harris and others is expanded upon by considering data from ethnobotany and the genetics of human selected traits for S. hispanica (Harris 1989; Hynes and Chase 1982; Rindos 1980; Weirsum 1997).

Ethnobotany has contributed considerably less than other scientific disciplines to models of plant domestication in part because ethnobotanical knowledge related to human selection is less abundant than knowledge of plant use. With respect to Salvia hispanica, knowledge of selection is also less abundant because typically only one or a few people in a community cultivate the species. Few studies of plant domestication include ethnobotanical data, and still fewer correlate management techniques with morphological variation in the crops. An exception is Nabhan's classic investigation of Hopi blue maize, in which he found a positive correlation between depth of planting and increased seed size (Nabhan 1989). Casas and Caballeros found that local management methods selected for certain phenotypes such as larger plants in the arborescent genus Leucaena (Fabaceae) (Casas et al. 1996). However, these studies focused on management and cultivation techniques within communities, and did not encompass the full range of any species in terms of geographic and morphologic diversity. Given the successful dispersal of useful plants, not all aspects of plant domestication processes can be presumed to take place in one location. Thus, at a given point in time and place the full continuum of the domestication process cannot be observed. Application of the more generalized Harris model to ethnobotanical data that covers the geographic, morphological, chronological, and cultural diversity of Salvia hispanica reveals the strength of ethnobotany as a tool in understanding domestication. A model of plant-people interactions for S. hispanica based on ethnobotanical data can be used to develop testable hypotheses about domestication that can be evaluated with genetic and other data for the species.

Salvia Hispanica as an Example of Domestication

For plant species, increased range of morphological and phenological diversity, suppression of natural mechanisms for dispersal and protection, gigantism, and other characteristics have been associated with domestication (De Wet and Harlan 1975; Harlan 1975; Hawkes 1983; Schwanitz 1967). Salvia hispanica, having a spectrum of domesticated varieties, is no exception. For the purpose of this study, varieties have been categorized loosely as “wild,” “primitive domesticates” (plants in cultivation that have been modified from wild-type morphology), “full domesticates” (plants that lack dispersal mechanisms and are dependent on human cultivation for survival), and “advanced full domesticates” (plants that exhibit human-selected traits and also lack dispersal mechanisms). This paper expands on published descriptions of the morphological details that distinguish wild and domesticated varieties of Salvia hispanica (Cahill 2004; Cahill and Provance 2002).

With each mature plant capable of producing a thousand seeds or more, the progeny of a single plant could supply a large portion of the next year's crop. This high fecundity combined with a short annual life cycle and a strong genetic component to inheritance for some domesticated traits made human selection in Salvia hispanica a powerful force in evolution of the crop (Cahill and Ehdaie 2005). These biological characteristics considered in conjunction with ethnobotanical data related to selection make S. hispanica an excellent candidate for the examination of plant-people interactions leading to domestication (see Table 1).

Table 1

Continuum of plant-people interactions for Salvia hispanica.¹

Materials and Methods

The morphological descriptions presented are based on a collection of 42 accessions categorized as wild or domesticated from the known range of the species (Cahill 2004). Recording of observations took place in common garden and common greenhouse experiments in Riverside, California, designed to eliminate effects of plasticity and to gauge genetic differences (Cahill and Ehdaie 2005; Cahill and Provance 2002). Herbarium voucher specimens were deposited at the herbarium of the University of California, Riverside (UCR)1 and are referenced in this study whenever possible. Field work was conducted during the winter of 1998–1999 in Pachali, Sacatepequez, Guatemala (lat 14°38′ N, long 90°38′ W, elev. 2200 m), in the Tarascan village of Capacuaro, Michoacan, Mexico (lat 19°33′ N, long 102°03′ W, elev. 2300 m) and in Santa Lucia, a rural region near Buena Vista, Santa Rosa, Guatemala (lat 14°28′ N, long 90°14′ W, elev. 2120 m). Field work was also conducted at the end of 1999 in the Coran village of Mesa de Nayar, Nayarit, Mexico (lat 22°13′ N, long 104°41′ W, elev. 1800 m) and in Acatic, Jalisco (lat 20°46′ N, long 102°53′ W, elev. 1720 m). The information presented is based on field interviews and participant observation when possible. In each location, a list was compiled of individuals who reportedly had knowledge of Salvia hispanica. Plants were collected and used during informant interviews as described by Cahill (2003). In each location, about twenty informants who volunteered information were re-interviewed using a more extensive set of questions (Cahill 2003). While many informants had extensive knowledge of uses of the plant, typically only one or two individuals had knowledge related to collection or management of wild plants or cultivation of the species in western Mexico. Information for those informants has been provided by Cahill (2003) and they are referred to by name below.

Collection from Wild Populations (Gathering)

The descriptions of human manipulation of wild populations of Salvia hispanica presented here roughly correlate to ‘pre-domestication’ components of the domestication process continuum (e.g., Harris 1989; Rindos 1984). Wetlands and areas along streams above 1800 m in pine or oak forest of Mesoamerica constitute ideal habitat for wild populations of S. hispanica. In Santa Lucia, Guatemala, a farmer called Lazaro del Cid and his family explained how the seeds of S. hispanica (Cahill and Castillo 2948, UCR), known locally as chan, were collected from the wild and mixed with seeds collected from wild Salvia tiliaefolia Vahl. The farmer stores the seed mixture in a glass jar, and combines portions of the seed mixture with water to form a medicinal beverage. Collection of small quantities of seed with no or minimal modification of environment is sustainable. In Santa Lucia, the vigorous wild population grows along mountain streams and has only wild type morphological characters, including short height (less than 1 m), striated stem pigmentation of anthocycanin, pubescent calyxes, open calyxes, small black seeds (less than 12.0 mg/100 seeds), short inflorescence (less than 15 cm), leaves and light blue corollas that are smaller than the domesticated varieties, and bushy growth form.

Similarly, in the Tarascan villiage of Capacuaro, informants identified local wild populations of Salvia hispanica as cueruni. The term cueruniquarani meaning “to gather chia” in Tarasco was also frequently mentioned and has been reported by others (Gilberti 1983; Sandoval 1989). Because Tarascan use is limited to medicinal use of small quantities of seeds, again there was no observable morphological variation accompanying this pattern of use, nor were there any apparent effects on population size.

In other wild populations, humans practice selective collection. For example, the informant Silvino Lobatos Estrada of Mesa de Nayar described in detail the technique for collection of seeds from the local wild population of Salvia hispanica (Cahill 2979, UCR). The pine forest of the region contains a habitat, known locally as “the swamp,” where this dense wild population flourishes. As with other wild populations, dark-seeded plants are present; however, in Mesa de Nayar a patchwork pattern exists with white-seeded wild plants, resulting in a heterogeneous population with respect to this character. According to Silvino Lobatos Estrada, selective collection of dark-seeded plants proceeds under the premise that white seeds do not have the “strength” of darker ones. Whether the “strength” pertains to taste or some property of the culinary flour made from the seeds is not clear. Since the flour forms the main ingredient in an atole beverage consumed during a religious ceremony, the rationale behind the ‘strength’ of dark seeds may have origins in superstition or religious belief. Though seasonal, culinary usage requires collection of large quantities of dark seeds. This may have the effect of selecting for an increase in the frequency of white-seeded plants in the population.

Informants described the process by which white seeds are excluded from collection. First, the seed color of an individual plant is determined by shaking a small quantity of seed out of the calyxes. If the plant is white-seeded the seeds are dropped to the ground. This shaking and dropping assists dispersal, spreading seeds further from the parent plant, resulting in decreased seedling competition for white-seeded plants. Apart from collection of seeds, informants clearly indicated that no other techniques such as broadcasting, thinning, burning, or weeding were employed.

I had assumed that white seeds were intentionally selected for because certain fully domesticated varieties produced in the valley of Mexico consist exclusively of white-seeded plants, suggesting direct human selection for white seeds. Modest natural selection may occur in wild populations as white seeds contrast with the dark soil, making white seeds more susceptible to predation. However, human selection is undoubtedly more intense. The white-seeded phenotype is rare in wild populations, with no other reported white-seeded plants from wild populations other than Mesa de Nayar. Since geography, altitude, and flowering time form reproductive barriers between wild types and domesticated varieties, gene flow from domesticated to wild populations seems an unlikely origin of white-seeded plants. The rise of white-seeded plants in wild populations, ancestral to domesticated varieties, was likely a more complex process as exemplified by the Mesa de Nayar population. There, approximately 60 percent of plants are white-seeded, a trait known to be recessive and controlled by a single gene (Cahill and Provance 2002). This relatively high percentage may be the result of the generations of collection of dark seeds that left more white-seeded plants to disperse, artificially increasing fecundity for white-seeded plants. A unique approach to further research would involve a synthesis of population genetics and ethnobotany, beginning with an extensive survey of population size, frequency of white-seeded plants, and intensity of human collection.

In the case of wild populations of Salvia hispanica, ethnobotanical evidence indicates only small quantities of plant material are collected. No evidence for or information about management of wild populations other than that based on seed collection, and no modern examples of reliance on the species as a primary or secondary source of nourishment, have emerged. This may account for the absence of ethnobotanical accounts of management techniques such as broadcasting, burning, or weeding that have been reported for other gathered wild species (see Table 1). The examples from S. hispanica provide documentation of direct selective forces on wild populations and the resultant morphological changes, offering important insights into beginnings of the domestication process.

Primitive Domesticates

Archaeological Evidence

The seed crop Salvia hispanica has taken a simple path to domestication, making it an excellent model for examining the plant domestication process. Unfortunately, little archaeobotanical evidence of chia has been found (Gonzalez Quintero 1981, 1986; Hard and Roney 1998). S. hispanica is a herbaceous annual with soft oily seeds, so there are no portions of the plant that preserve particularly well, making it difficult to draw any conclusions regarding domestication from this evidence. As a result, the time of earliest cultivation and domestication of chia are unknown.

Historical Evidence

Economic historians have suggested Salvia hispanica as a staple food was as important as maize, and in some areas more important (Harvey 1991; Hebert et al. 1995; Hunziker 1952; Perm and Carrasco 1974; Rojas Rabiela 1988). The codices of sixteenth-century Mexico indicate that large areas of agricultural land were devoted exclusively to chia cultivation. For example, 21 of the 38 Aztec provincial states gave chia in annual tribute and separate records from independent states such as Huexotzinco also list chia as tribute (Berdan and Anawalt 1996; Perm and Carrasco 1974). Seeds from wild populations formed a portion of this tribute, but the quantities recorded in tribute lists would have necessitated cultivation. Human transportation of chia seeds and the resulting dispersal to new habitats was perhaps the greatest ecological effect of cultivation. Sandoval reviewed the sixteenth-century geography of S. hispanica cultivation and concluded it was widely grown until the later part of the sixteenth century, at which time it was displaced by European crops (Sandoval 1989). The origins of cultivation for this species remain unclear, though several domesticated varieties with incremental degrees of human selected traits still survive.

From 1575–1577, Fray Bernardino de Sahagún compiled the Florentine Codex, working with Nahua informants in central Mexico who retained memories of life prior to arrival of the Spanish. Sahagún described a method of planting chia that inadvertently selects for larger seeds. The description and illustrations of chia, subsequently recognized as Salvia hispanica by numerous botanists, describes how sown chia seeds are “…covered over with the soil, just thinly, just smoothed over with the foot.” (Sahagún 1950–1982, book 10, chapter 20, page 75). Smaller seeds do not contain enough stored resources to reach the surface and are selected against. Though unintentional, increased seed mass represents the first quantitative trait to be affected by human activities associated with the species. Seed mass samples collected from common garden and common greenhouse plants of different accessions showed seed mass for individual plants and within accessions did not vary significantly (Cahill and Ehdaie 2005). These results also indicate a significant difference in seed mass between wild and domesticated accessions. Since each plant is capable of producing thousands of seeds at maturity, the small increase in seed size translates into markedly greater yields. Apart from the essentially wild material collected in Pachali (see below), all other primitive domesticated varieties in the current germplasm collection exhibit increased seed size.

Ethnobotanical Evidence

Ethnographic examples of plant manipulation are presented that correlate to certain morphological characteristics. The examples begin with the examination of the most morphologically primitive domesticated varieties and the associated management techniques which have shed light on the origins of cultivation. In Pachali, Department of Sacatepequez, Guatemala, local farmers managed the growth of wild Salvia hispanica in fields of other crops, including beans and maize (Cahill and Castillo 2970, UCR). After the bean harvest, plants of S. hispanica were left standing to mature while all other species had been cleared. The sporadic chia plants standing in stark contrast to the neatly harvested and cleared field emphasize the value of the species and the effort put into its cultivation. When asked if the seeds were planted, the response was an emphatic no, and informants went on to explain that seedlings just appeared in the same place year after year and were intentionally spared when weeding. Plants do not spread within fields from year to year and do not appear on margins or outside of the fields as weeds. Small quantities of seed are collected in January and stored for use in medicinal beverages throughout the year. This management by tolerance of volunteers has preserved many pre-Columbian primitive domesticated varieties that formerly may have been grown as primary crops. The material from Pachali has retained wild morphological characters, and is indistinguishable from other Central American wild types when grown in a common garden. Wild populations exist in the vicinity of Pachali, suggesting that human or natural dispersal from a wild population gave rise to these plants found in cultivation. With no human-selected traits and no actual sowing of seeds or tillage, the Pachali material represents a very early stage of domestication where plants are cultivated but no apparent modifications from wild type morphology are observed. The presence of Salvia hispanica in this cultivated setting seems largely dependent on the ecological effect of selective weeding in the fields, without which S. hispanica would likely be out-competed and eventually disappear from the fields. Management via cultivation continues to play an important role in preserving varieties.

Planting of Salvia hispanica seeds in cultivated settings traditionally proceeds by broadcasting seeds into the wind without soil preparation. Hernández Gómez reported this technique in the settlements of Olinalá and Temalacacingo in the Mexican state of Guerrero (Hernández Gómez 1989, 1994). Kelly reported the identical technique in Ayotitlán, Jalisco (Kelly 1944). Generally, broadcasting increases population size and density, resulting in greater productivity in comparison to natural dispersal. Since later-germinating seeds are out-competed in these dense cultivated plots, domesticated varieties have developed rapid and uniform germination. Comparative observations made in germinators at a constant 30°C showed wild types do not germinate uniformly and rapidly, but sporadically over a 10-day period. Wild material from Mittlenam, Santa Lucia, Sinaloa, Mexico, from the collection of the late Howard Scott Gentry best exemplified this sporadic and delayed germination. In contrast, all primitive and fully domesticated varieties germinated within two days. Several published studies have described germination of 90 to 100 percent within 48 hours with no difference between white and dark seeds, but these studies do not indicate whether the seeds originated from domesticated, cultivated, or wild plants (Labouriau and Agudo 1987; Naqvi et al. 1992).

The Mexican agronomist Rulfo provided invaluable accounts of traditional cultivation techniques for Salvia hispanica, including preparation of soils by “roughing” (Rulfo 1937). Roughing represents the first significant increase in human energy input into a system of cultivation. Disruption and loosening of the soil surface provides an environment where more seeds are hidden, escaping predation and facilitating establishment of seedling roots. While Rulfo described roughing in conjunction with broadcasting, other traditional cultivation methods build on roughing by covering seeds with soil.

The morphologically most primitive of these large-seeded domesticates has a restricted range of cultivation centered near Merida in the Yucatan peninsula of Mexico. In addition to larger seeds, the gene for white seed coat is present in the gene pool. Apart from these two characters, the plants resemble typical wild ecotypes. The presence of the gene for white seed coat in otherwise morphologically primitive domesticated varieties suggests early human selection for this trait. In the Florentine Codex, Sahagún described a variety of Salvia hispanica known as iztac chia ‘white chia’, illustrated with white seeds (see Figure 1) (Sahagún 1950–1982). Certainly, white seeds should not be considered a compulsory character among the large-seeded primitive domesticates, but early selection for the trait seems likely.

Figure 1

Illustrations of three Nahuatl varieties of Salvia hispanica from the sixteenth-century Florentine Codex: iztac chian ‘white chia’ (top); mitzic chian, a type of black chia (middle); and chian tzotzol ‘wrinkled chia’ or ‘chia that is ground in water’ (bottom).

Material cultivated in Temalacacingo, Guerrero, Mexico, does not produce white-seeded plants, but human selection for increased branching results in an increased number of inflorescences per plant. It is also associated with large seeds in this primitive domesticated variety. In his agronomic assessment trials, Hernández Gómez found it to be more productive than many of the fully domesticated varieties exhibiting greater numbers of human selected traits (Hernández Gómez 1994). However, this variety fully disperses seeds, making it difficult to harvest without affecting yield. In a common garden experiment at the University of California, Riverside, plants of this variety became so heavy with seeds the main stems broke, so plants were propped up. In Temalacacingo, chia is reportedly still cultivated in tlacololes, plots that have been recently slashed and burned, associated with maize (Hernández Gómez 1994). The management techniques of burning, broadcasting, and covering seeds require more labor than the simple tolerance for plants observed in Pachali. The degree to which these management techniques are directed specifically at chia rather than the associated crops deserves more investigation, since cultivated forms of Salvia hispanica grow predominantly in mixed-crop settings today. In Tlalhuica in 1564, chia reportedly grew on the same irrigated plots of land with beans and maize, sown between the maize (Rojas Rabiela 1988). In November of 1999, Cora informant Reyes Camarena of Mesa de Nayar reported seeing mixed-crop cultivation of Salvia hispanica with calabash (Cucurbita sp.) in the Sierra San Andras and Santa Katarina. The erect stems of S. hispanica provide runners for the gourd vines to climb. Whether directed at S. hispanica or not, the management regime in such mixed-crop settings has produced quantitative morphological variation.

From an ecological perspective, competition introduced by mixed cropping could account for some quantitative morphological changes. Increased competition from crops of greater stature like maize could result in the increased height seen in certain advanced fully domesticated varieties. The increased branching observed in material from Temalacacingo could be the result of competition from crops, decreased competition due to weeding, or a particular management technique. Despite morphological changes that enhance productivity in comparison to wild Salvia hispanica, primitive domesticated varieties have largely been displaced by fully domesticated varieties with a greater number of human selected traits. This process of displacement would have begun after the appearance of full domesticates in pre-Columbian times. As a result, primitive domesticated varieties are rarely encountered today and their survival seems bleak, given their dependence on traditionally cultivated plots of land and the general shift towards agricultural modernization in Mesoamerica.

Full Domesticates

Historical Evidence

Francisco Hernández, writing from 1571–1577, described how “[t]his plant chia grows wherever it seeds, principally in cultivated, irrigated, and watery places” (Hernández 1959:207–210). This and other sixteenth-century accounts stress the importance of water for successful cultivation of chia. Over 9000 ha of land in the early 1500s were cultivated under the traditional chinampa raised field “floating garden” system (Armillas 1971). In the chinampas, chia faced an extended growing season not affected by drought (Coe 1964; Rojas Rabiela and Sanders 1985). Unique agricultural techniques practiced in the chinampas may have worked as selective forces upon morphology. Sandoval believes chia was sown directly and was not grown in almacigos (seedling nurseries) or in chapines (muck cubes) (Sandoval 1989). The agricultural techniques applied to chia in the chinampas and possible resulting morphological variation remain unresolved issues. While chia constituted one of the major crops of the chinampas, after Spanish conquest chia cultivation in chinampas ceased (Harvey 1991; Weaver 1981). In contrast, Dr. Juan Jimenez-Osornio reported occasional cultivation of Salvia hispanica in the modern chinampas system.2 The history of chia cultivation in the wet soils of the chinampas is a reflection of the dependence of ancestral wild ecotypes on wet soils. This wild heritage translates well to the modified aquatic system of the chinampas. The ecological effect of decreased or lacking water stress combined with the extended growing season may have allowed for selection of traits associated with fully domesticated chia varieties.

In the mid-sixteenth century, Sahagún described “iztac chian” or white chia as “…that which is uprooted, which can be rubbed in the hands. I uproot chia. I rub chia in my hands.” (Sahagún, 1950–1982). This technique of hand threshing signals the development of the second major increase in human energy input associated with Salvia hispanica. This laborious threshing technique also requires harvesting and drying of the whole plant, a departure from seed selection of wild and primitive domesticated varieties, which proceeds by shaking seeds off the unharvested mature plants into textiles or baskets (Castello Yturbide 1972; McVaugh 1943). In the tenth book of the Florentine Codex, Sahagún described the world of the Aztec markets; a seller of wrinkled chia describes how he sold “unthreshed” chia plants in markets, how “he rubs it between his hands to clean it” (Sahagún 1950–1982). In all wild types and primitive domesticated varieties, seeds fall from the plant with just a brush or strong wind. The transport of harvested inflorescences to Aztec markets suggests that non-shattering morphological forms already existed by the 1500s. In addition, linguistic evidence of cultural knowledge of non-shattering domesticated forms appears throughout Mesoamerica. Examples include the reported seventeenth-century use of the term ti her, ti bak aaq, which means “to thresh chia seeds” in the Mayan language Cakchiquel (Acuña 1983), and the more recent use of qui tu thohui inchoní, which means “to thresh chia” in the indigenous language Matlatzinca (Basalenque 1975; Sandoval 1989).

Morphology of Full Domesticates

The group of full domesticates is defined by their closed calyxes, larger seeds, variability in calyx pubescence, and gigantism. Closed calyxes prevent seed dispersal and effectively prevent survival of fully domesticated varieties outside of human cultivation. This character meets the criterion for “full domestication” as described by Harlan (1975) or “agricultural domestication” as described by others. None of the herbarium specimens of wild populations or wild material sampled for this study exhibit any degree of closure of calyxes. To date, the trait has been observed only in the fully domesticated varieties of Salvia hispanica and has not been reported in wild populations or related species of Salvia. Like other qualitative traits in the species, closure of calyxes could have developed as the result of conscious human selection.

Fully domesticated varieties exhibited seed mass ranging from 15.0–16.5 mg/100 seeds, a large increase in comparison to wild types, but only a slight increase (though statistically significant) over primitive domesticated varieties. Furthermore, successful selection for increased seed mass has a strong genetic component demonstrating the potential for strong, rapid, and effective human selection (Cahill and Ehdaie 2005). However, no correlation between seed size and white color was observed among varieties.

Since dried calyx trichomes can induce sneezing, decreased calyx pubescence relieves the hand thresher from incessant irritation. The popular Mexican varieties chia poblana from Cuatepec, Puebla, and others grown in Sonora typify fully domesticated varieties with decreased density of calyx pubescence. Fully domesticated varieties from Central America, including those now cultivated in Jutiapa, Guatemala, San Salvador, El Salvador, and throughout Nicaragua, characteristically display an increase in the density and length of calyx pubescence in comparison to wild populations. Material from Nicaragua exhibits these traits to the greatest degree. For Salvia hispanica, an increase in the length and density of calyx pubescence is associated with adaptation to a more humid environment; in the absence of enhanced pubescence, precipitation causes the mature dry calyxes to soak up water, which renders the enclosed mature nutlets susceptible to hydration. Seeds release a polysaccharide mucilage upon hydration, causing the ripe inflorescence to develop a sticky coating that hardens upon drying, destroying the crop. The dry winters of Mexico do not present a problem, but cultivation in more humid areas outside the natural range of the species, such as Nicaragua, necessitates protection from precipitation. An additional novel adaptation to a humid environment found in Nicaraguan domesticates is lack of determinacy in timing of flowering and ripening of seeds. In the common garden experiment mentioned above, individual plants flowered continuously from December through May and retain leaves, suggesting a selective trend away from annualism.

A compact inflorescence, defined by small space between glomeruli and number of flowers per axil, appears in all fully domesticated varieties. Wild populations of Salvia hispanica have great variation with respect to these traits. As a general rule, however, the higher the elevation the more compact the inflorescence becomes.

Some degree of gigantism occurred in all fully domesticated varieties examined in a common garden setting (see Table 2).

Table 2

Results of common garden and green house experiments in Riverside, California: quantitative variation in Salvia hispanica.

Advanced Full Domesticates

While the fully domesticated varieties described above are rare, a second group of fully domesticated varieties are more common. They exhibit an even greater number of human-selected traits. These advanced fully domesticated varieties are more uniform with respect to morphology and genetic diversity and can be found in cultivation throughout Mexico (Cahill 2003). Three principal morphological characteristics separate this group from the other fully domesticated varieties described above: increased inflorescence length, apical dominance, and a further increase in plant height. An increase in apical inflorescence length from about 6 cm to 40 cm represents the first quantitative characteristic that presumably resulted from direct human selection. Longer inflorescence would increase yield, but the character always appears in conjunction with apical dominance so that only the inflorescence closest to the apex has increased length, and secondary branches and inflorescences remain normal or slightly stunted. There are clearly genetic factors controlling the quantitative traits of apical dominance and plant height observed in common garden and common greenhouse settings with well spaced plants. Hernández Gómez has shown competition from close spacing of plants can also effect these traits to some extent (Hernández Gómez 1989).

Advanced fully domesticated varieties traditionally grown in parts of Jalisco and the central valley of Mexico produce considerably more anthocyanin pigmentation on the stems and calyxes. In contrast to wild ecotypes with striated stem pigmentation, stems of these varieties have an unbroken purple pattern that extends to the calyxes. Some advanced fully domesticated varieties produce pure purple calyxes, while in others pigmentation only occurs on the outer half of each calyx. Observation of wild populations in the Sierra Madre Occidental indicate some populations have a few plants that express slight purple shading on stems and calyxes under stress conditions towards the end of the life cycle, but not nearly to the degree seen in advanced fully domesticated varieties with purple calyxes. This may be an example of aesthetic selection, but the reasons for selection for excess anthocyanin pigmentation remain a mystery. In Amaranthus, pigmented varieties had associations with pre-Columbian rituals. Given the religious and cultural uses of Salvia hispanica, selection for religious use remains a possibility, but no explicit accounts correlating any use with pigmented varieties have been uncovered (Cahill 2003).

A shift towards use as a nutritionally reliable subsistence food and as a seed oil accompanies cultivation of domesticated chia. Today, monocropping of the advanced domesticated varieties has almost completely displaced the traditional intercropping system of the primitive domesticated varieties and the non-advanced full domesticates. It has led to a loss of genetic diversity over a wide geographic area.

Numerous studies have implicated cultivation and agriculture as causative factors in the development of allelopathy (Einhellig 1996; Seigler 1996). The beneficial chemical and ecological characteristic of allelopathy has accompanied advanced full domestication. Preliminary experiments with aqueous root leachates of the morphologically advanced domesticates of Salvia hispanica applied to Echinochloa and Amaranthus seeds show statistically significant inhibition of germination of Echinochloa seeds, but not for wild or any of the other domesticated varieties tested (see Table 3). Further experimentation is needed, both in isolating chemical compounds involved and identifying common weeds associated with Salvia hispanica monocrop systems in Mexico.

Table 3

Allelopathic activity among wild and domesticated varieties of Salvia hispanica.

Mechanized Agriculture

In recent decades, mechanized agriculture has come to dominate production of morphologically advanced fully domesticated forms of Salvia hispanica. Mechanized agricultural techniques may act as selective forces further reinforcing the morphological features of domestication. Uniform morphology of numerous characters, particularly height and determinacy, is found in varieties grown under mechanized agriculture, although it is not completely certain that these traits are due to mechanized agriculture. In the common garden plots, each Mexican advanced domesticated variety flowered and matured at the same time and all of the inflorescences on each plant flowered and matured at the same time in early October and set seed about 2–3 weeks later, though the timing varied among the varieties. As plants began to flower, leaves became necrotic and abscised, leaving a stem and ripe inflorescence. The lack of leaves facilitates harvesting, particularly in mechanical settings. Wild populations in Mexico also drop leaves as flowering begins, but individual plants and inflorescences on the same plant do not synchronously flower.

Informant Jesus de Latorre Guiterrez of Acatic, Jalisco, described his method for mechanical planting of Salvia hispanica. He plows furrows approximately 8 cm apart. On the ridges between the furrows, the machinery punches holes about 2.5 cm in diameter, 5 cm deep, and 50 cm apart. He then drops several seeds in each hole which is left open. The mesic microenvironment in the holes facilitates germination and within a few days the seedlings emerge. With other planting techniques, hydrated seeds tend to dry out, inhibiting germination and increasing predation by birds and ants. Approximately 10–20 seeds are placed in each hole, but only 3–5 plants per hole survive to maturity. Seedling competition likely selects for larger seed mass and rapid germination, even though the seeds are not covered with soil as with many other traditional cultivation techniques.

Two distinct varieties of morphologically advanced full domesticated varieties are found in the Acatic region. The predominant variety, known locally as chia pinta, forms heterogeneous populations. The majority of plants have gray seeds with black marbling and lack stem pigmentation; a minority of plants may be white-seeded, and a few have the striated stem pigmentation typical of wild populations. Individuals of the second unnamed variety occur sporadically among the vast fields of chia pinta, but produce only dark seeds and solid purple pigmentation on stems and calyxes. This contrasts to the striated stem pigmentation of wild types and the absence of stem pigmentation found in nearly all advanced full domesticated varieties. Additionally, it flowers and seeds slightly later than chia pinta. This second variety is distinct and has bred true in three generations of experiments. Nabhan has stated, “many traditional farmers sow multi-line mixtures of the same crop as a kind of insurance” (Nabhan 1989:73). This may have initially been the case in Acatic, as farmers attempted to increase productivity. However, none of the informants were aware of multiple varieties in their fields, and insisted they planted only chia pinta. With mechanized planting and harvesting, nearly all of the traditional agricultural knowledge associated with the crop has vanished. Each farmer saves seed from year to year and does not buy seed or select individual plants for improved traits. The only selective forces acting today on advanced full domesticates seem to be those stemming from the use of mechanical sowing and harvesting equipment. Ethnobotanical knowledge associated with direct selection methods and traditional management techniques and how they might act as selection factors in advanced full domesticated varieties has been lost with the conversion to mechanized agriculture in the commercial production centers of Salvia hispanica in Acatic and Jutiapa.

Conclusions

This investigation of Salvia hispanica demonstrates that a detailed ethnobotanical study of wild and cultivated populations over the entire range of a species can offer a more comprehensive understanding of the human selective forces that set into motion the domestication process than one restricted to a single community or region. While such studies can be more difficult to carry out, from a biological or morphological perspective the benefits outweigh the costs and the resulting data may encompass the entire continuum or stages of interactions leading to domestication.

An evolutionary process for Salvia hispanica governed primarily by human selection has emerged from this more comprehensive single species approach. Whether intentional or unintentional, each instance of human selection could be placed within the context of the ecosystem or agricultural management regimes of the Harris continuum (see Table 1). Not all of the categories of Harris's more generalized continuum could be related to the data for a single species. Data that fits in many of the cultivation and agriculture categories overlap, supporting the concept of a continuum rather than distinct steps. In contrast, clear distinctions in morphology, such as larger seed mass and non-shattering inflorescence, specifically correlate with changes in human agricultural practices, suggesting plant-human interaction for S. hispanica may be a stepwise process rather than a continuum. While the ethnobotanical data presented in this paper by no means offer a complete picture of the domestication process of S. hispanica, the information does include known instances of selection, providing much needed documentation of plant-human interactions.

Bye provided ethnobotanical evidence that human management of weedy/wild populations results in quantitative morphological variation; management resulting in qualitative morphological variation, such as seed color, appears to be less common (Bye 1981). Results from the analysis of Salvia hispanica data suggest selection of qualitative traits occurred prior to incorporation of the species into agriculture settings. The selection against and frequency of white-seeded plants in the wild population of Mesa de Nayar best support this conclusion. Unlike Bye's weedy species, more intentional or intense selection of quantitative traits may be required for S. hispanica. Information collected in Pachali supports this idea in that S. hispanica is not weedy within or at the edges of that agricultural setting and does not persist without attentive human care.

The ethnobotany of Salvia hispanica can be viewed as a hypothetical framework for the process of domestication that can be tested for all the domesticated seed crops in Lamiaceae. Through ethnobotanical studies of wild ancestors and primitive forms of related species, ethnobotany will continue to provide answers, as it has in this study, to questions related to the genesis of plant-people interactions.

Footnotes

Acknowledgments

I wish foremost to thank Dr. Arturo Gómez-Pompa for five valuable years of guidance and assistance during my graduate studies. I also thank J. Giles Waines, Eugene Anderson, and Blanca Hernández Bautista. I am indebted to the farmers and native people in the areas of study for their willingness and excitement in relating information.

1

All voucher herbarium specimens cited herein are deposited in the herbarium of the University of California, Riverside.

2

Interview with Dr. Juan Jimenez-Osornio professor at Universidad Autónoma de Yucatán, in Riverside, California, September 18, 2002.

References

Acuña

1983. Thesaurus verboru, vocabulario de la lengua cakchiquel el guatemalteca. México, D.F U.N.A.M. (Originally written 1647 by Fray Thomas de Coto).

Armillas

1971. Gardens on swamps: archeological research verifies historical data on Aztec land reclamation in the valley of Mexico. Science 174:653–661.

Basalenque

1975. Arte y vocabulario de la lengua matlatzinga vuelto a la Castellana. Biblioteca Enciclopédica del Estado de México. México, D.F.

Berdan

and Anawalt

P. R.

, editors. eds. 1996. Codex Mendoza. University of California Press, Ltd. London. (Originally composed ca. 1541–1542).

Bye

R. A.

1981. Quelites—ethnoecology of edible greens—past, present, and future. Journal of Ethnobiology 1:109–123.

Cahill

J. P.

2003. Ethnobotany of chia, Salvia hispanica L. Economic Botany 57:604–618.

Cahill

J. P.

2004. Genetic diversity among varieties of chia (Salvia hispanica L.). Genetic Resources and Crop Evolution 51:773–781.

Cahill

J. P.

and Ehdaie

. 2005. Variation and inheritance of seed mass in chia (Salvia hispanica L.). Genetic Resources and Crop Evolution 52:201–207.

Cahill

J. P.

and Provance

M. C.

. 2002. Genetics of qualitative traits in domesticated chia (Salvia hispanica L.). Journal of Heredity 93:52–55.

10.

Casas

A. M.

, Vazquez

, Viveros

J. L.

, and Caballero

. 1996. Plant management among the Nahua and the Mixtec in the Balsas River basin, Mexico: an ethnobotanical approach to the study of plant domestication. Human Ecology 24:455–478.

11.

Castello Yturbide

, editor. ed. 1972. El maque, lacas de Michoacán, Guerrero y Chiapas: artes de México. Artes de México. México, D.F.

12.

Coe

M. D.

1964. The chinampas of Mexico. Scientific American 211:90–98.

13.

De Wet

J. M. J.

and Harlan

J. R.

. 1975. Weeds and domesticates: evolution in the man-made habitat. Economic Botany 29:99–107.

14.

Einhellig

F. A.

1996. Interactions involving allelopathy in cropping systems. Agronomy Journal 88:886–893.

15.

Gilberti

1983. Diccionario de la lengua tarasco. Balsal Editores S.A. México, D.F.

16.

Gonzalez Quintero

1981. Análisis polínicos en la porción austral de la cuenca de México II. Tlaltenango. Revista de INAH 3:12–16.

17.

Gonzalez Quintero

1986. Contribuciones al conocimiento arqueobotánica de la cuenca del valle de México. Instituto Nacional de Antropología e Historia. México, D.F.

18.

Hard

R. J.

and Roney

J. R.

. 1998. A massive terraced village complex in Chihuahua, Mexico, 3000 years before present. Science 279:1661–1664.

19.

Harlan

J. R.

1975. Crops and Man. American Society of Agronomy. Madison, Wisconsin.

20.

Harris

D. R.

1989. An evolutionary continuum of people-plant interaction. In Foraging and Farming: The Evolution of Plant Exploitation. eds Harris

D. R.

and Hillman

G. C.

, editors. 11–26.Unwin Hyman. London.

21.

Harvey

H. R.

1991. Land and Politics in the Valley of Mexico: A Two Thousand-Year Perspective. University of New Mexico Press. Albuquerque.

22.

Hawkes

J. G.

1983. The Diversity of Crop Plants. Harvard University Press. London.

23.

Hebert

J. R.

, Loste

B. M.

, Noguez

, Albro

S. R.

, and Albroll

T. C.

. 1995. Codice de Huexotzinco. Ediciones Multiarte. México, D.F. (Originally written 1531).

24.

Hernández

1959. Historia natural de Nueva España. 2 vols. Universidad Nacional de México. México, D.F. (Originally written 1571–1577).

25.

Hernández Gómez

J. A.

1989. Efecto de la fecha siembra, densidad de población y competencia de malezas en el rendimiento de chia (Salvia hispanica L.). Masters Thesis, Colegio de Postgraduados,. Universidad Autónoma Chapingo. México.

26.

Hernández Gómez

J. A.

1994. Chia (Salvia hispanica) antecedentes y perspectivas en México. ed Cuevas

, editor. Universidad Autónoma Chapingo. México.

27.

Hunziker

A. T.

1952. Los pseudocereales de la agricultura indígena de America. Acme Agency, Suipacha. Buenos Aires.

28.

Hynes

R. A.

and Chase

A. K.

. 1982. Plants, sites and domiculture: aboriginal influence upon plant communities in Cape York Peninsula. Archaeology in Oceania 17:38–50.

29.

Kelly

I. T.

1944. Worked gourds from Mexico. In. Notes on Middle American Archaeology and Ethnology. 113–118.AMS Press. New York.

30.

Labouriau

L. G.

and Agudo

. 1987. On the physiology of seed germination in Salvia hispanica L. I. temperature effects. Anais da Academia Brasileira de Ciencias 59:57–70.

31.

McVaugh

1943. Economic Plants of Interest to the Americas: Oiticica and Chia Oils. United States Department of Agriculture, Office of Foreign Agricultural Relations. Washington, D.C.

32.

Nabhan

G. P.

1989. Enduring Seeds: Native American Agriculture and Wild Plant Conservation. North Point Press. San Francisco.

33.

Naqvi

H. H.

, Ting

I. P.

, Brown

J. H.

, Kumamoto

, and Matsumura

. 1992. Chia: a potential oil crop for arid zones. New Industrial Crops and Products. eds Naqvi

H. H.

, Estilai

, and Ting

I. P.

, editors. 179–202.Association for the Advancement of Industrial Crops. Riverside, California.

34.

Perm

H. J.

and Carrasco

. 1974. Matricula de Huexotzinco. Akademische Druck. Graz, Austria. (Originally written 1560).

35.

Rindos

1980. Symbiosis, instability, and the origins and spread of agriculture: a new model. Current Anthropology 21:751–772.

36.

Rindos

1984. The Origins of Agriculture: An Evolutionary Perspective. Academic Press. New York.

37.

Rojas Rabiela

1988. Las siembras de ayer, la agricultura indígena del siglo XVI. Secretaría de Educación Pública. México, D.F.

38.

Rojas Rabiela

and Sanders

W. T.

, editors. 1985. Historia de la agricultura epoca prehispanica siglo XVI. Instituto Nacional de Antropología e Historia. México, D.F.

39.

Rulfo

J. M.

1937. La chia. Agricultura (México) 1:129–137.

40.

de Sahagún

Fray Bernardino

1950–1982. Florentine Codex: General History of the Things of New Spain. eds Anderson

A. J. O.

and Dibble

C. E.

, editors. 14 vols. Monographs of the School of American Research. University of Utah Press. Salt Lake City. Originally written 1575–1577 or 1578–1580.

41.

Sandoval

A. Z.

1989. Amarantos y chias. Un estudio etnohistórico. Escuela Nacional de Antropología e Historia. México, D.F.

42.

Schwanitz

1967. The Origin of Cultivated Plants. Harvard University Press. Cambridge, Massachusetts.

43.

Seigler

D. S.

1996. Chemistry and mechanisms of allelopathic interactions. Agronomy Journal 88:876–885.

44.

Weaver

M. P.

1981. The Aztecs, Maya, and Their Predecessors, Archaeology of Mesoamerica. 2nd ed. Academic Press. New York.

45.

Weirsum

K. F.

1997. From natural forest to tree crops: co-domestication of forest and tree species, an overview. Netherlands Journal of Agricultural Science 45:425–438.