Deciphering the artificial evolution of domesticated plants within dynamic habitats

Plant domestication gave rise to one of the most significant subsistence transitions in the prehistory of humankind with fundamental implications for the emergence of complex societies. Moreover, understanding the coevolution between humans and the most recent common ancestors of domesticated plants is an essential component of maintaining genetic diversity that will assist with future attempts at crop improvement in the face of changing climatic conditions. The Chiou et al. paper (1) provides a deeper understanding of the domestication of the widely cultivated chili pepper (Capsicum annuum var. annuum), the environments from which the domesticated species arose, and the challenges facing scientists in a region where archaeobotanical data collection is rarely optimal.

In general, archaeobotanical research in Mesoamerica has been plagued by poor preservation, especially in regard to plant macroremains and, oftentimes, pollen, as well. Information about early crop development and agricultural techniques of the past largely has been derived from a handful of dry caves and other situations where pollen and plant macroremains were unusually well-preserved. In these special contexts, macroremains were abundant, intact, and relatively easy to identify. Well-known Mexican sites in this category include the Tehuacan Valley caves (2), Guila Naquitz (3), and the caves of Tamaulipas (4). Plant remains from these sites have dominated the conversation about plant domestication and early farming techniques in Mesoamerica because, for many years, these sites represented the main body of information about plants found in preceramic contexts. Unfortunately, scholars often conflate the earliest discovery of a crop plant at an archaeological site with its zone of initial domestication. This may be true in an approximate sense, but it is rarely, if ever, true in the exact sense. For example, even though the earliest macroremains of chili peppers came from the Tehuacan Valley, it is unlikely that the first domesticates emerged there because the habitat of that area would have been unsuitable in early Holocene times. The wild progenitor (C. annuum var. glabriusculum) of domesticated chili is a warmth-loving species and the Tehuacan area would have been too cold for it when domestication probably occurred (1).

This leads to another major point of discussion raised by Chiou and associates. Scholars discussing the origins of domesticated plants often neglect to take into account the fact that climate and habitats have changed dramatically since the Late Glacial Maximum (20,000 B.P.), the time period when most crops were domesticated. To address this incongruity, the authors created an ecological suitability model which coupled bioclimatic data through time with known human occupation sites in Mexico to enable a diachronic model that would indicate where human populations were found and where suitable habitats for the wild chili ancestors might co-occur. Their findings indicated that appropriate habitat for chili pepper origins may have been limited to southern Mexico, in particular, the southern coastal areas of the state of Guerrero or the Yucatan peninsula (1).

Chiou et al. also contend that the role of women has largely been overlooked in the domestication discussion writ large. This would have been especially true for chili peppers. In ethnobotanical studies of Mesoamerican indigenous people, it is a common observation that peppers are grown in house gardens or “huertas” which are generally the domain of women (5–8). Furthermore, women are largely in control of the culinary aspects of Mesoamerican culture, and this tendency coupled with the ready availability of chilis in household gardens would augur well for selection pressures leading to more productive, resilient, and flavorful cultivars.

The discovery of additional archaeological sites with excellent preservation combined with improved analytical techniques have led to a greater understanding of the domestication process, ancient agricultural approaches, and concomitant anthropogenic environmental changes. Notably, the San Andrés site near the Gulf of Mexico in Tabasco with its water-logged contexts provided well-preserved evidence of maize (Zea mays) agriculture in a sandy, lowland, estuarine setting dating back to Archaic times, circa 7,300 BP with subsequent findings of manioc (Manihot esculenta) circa 5,800 y ago, and chili peppers (C. annuum var. annuum) circa 3,200 y ago as the local environment changed to a marshy habitat (9, 10).

In recent times, additional methods involving different kinds of plant materials and extractive measures have now been developed that enable more comprehensive analyses of plant use activities of the past. For example, evidence involving starch grains, particularly from a genus-specific starch morphotype, revealed that domesticated C. annuum has been cultivated in several locations in the western hemisphere, ranging from the Bahamas to Peru, as early as 6,100 y ago (11). Another example of the use of starch grain analysis, this time combined with phytolith and pollen studies, occurred at Xihuatoxtla rockshelter in Guerrero, Mexico (12). This effective combination of research techniques focusing on degradation-resistant plant products enabled researchers to push back the time of maize cultivation to 8,700 cal. BP.

Sorting out the complexities of how and where crops were domesticated and which species were the closest wild relatives can readily be achieved with molecular genetic comparisons of modern domesticates to potential wild ancestors. This approach can determine the genetically most similar living wild relatives of a crop and oftentimes help to locate the zones of origin. Molecular genetic techniques, along with some thorough field collecting, helped to resolve the mystery of the wild ancestor of maize (13). Since then, ancient DNA combined with other techniques have been successful in elucidating the early improvement of maize (14, 15). Likewise, the interpretive power of molecular genetics helped to identify of the most closely related living ancestor of domesticated chili pepper, C. annuum var. glabriusculum (16). Ancestor-descendant questions for other domesticates, e.g., squash, are being resolved in similar ways (17).

The application of molecular genetic approaches has not only been useful in detecting the distribution of ancient domesticates, but they are also proving to be useful in describing ancient habitats in various geographic settings (18). This technique is often referred to as eDNA or aeDNA for ancient environmental DNA. It involves the recovery of ancient plant genetic material from soil or sediments. This can be distinguished from ancient DNA approaches that usually involve the extraction of DNA directly from ancient plants or animals where the organism is still intact through an unusual circumstance such as the preservation of a wooly mammoth in permafrost (19).

In the aeDNA approach to data collection, the individual (or a leaf, for example) dies and as its cells lyse, they release their DNA into the surrounding environment. In some situations, the released DNA molecules wrap around clay or even sand particles. When this happens, the DNA molecules become adsorbed to the surfaces and thus are protected from the predation of bacteria and other microorganisms (20). Sediments from ancient contexts can be subjected to extraction procedures followed by next-generation sequencing and subsequent bioinformatics analysis. If all conditions are favorable, useful sequences of crop plants can be identified. For example, a chili pepper aeDNA sequence from a ceremonial bundle at the base of a Maya ballcourt was recovered from Late Formative (ca. 100 BCE) deposits at the Yaxnohcah site in southern Mexico (21). The aeDNA approach succeeded while other recovery techniques applied at the site were unable to detect the presence of chili peppers.

The discovery of additional archaeological sites with excellent preservation combined with improved analytical techniques have led to a greater understanding of the domestication process, ancient agricultural approaches, and concomitant anthropogenic environmental changes.

Similarly, the aeDNA technique can be useful in describing habitats of the past, as suggested in the Chiou et al. paper (1). An example of this occurred at the ancient Maya site of Tikal, Guatemala, where sampling of the main reservoirs for pollen provided limited information because of preservation issues, but the aeDNA results (22) did reveal the plant species growing next to the Temple and Palace reservoirs (Fig. 1). This demonstrates how the use of aeDNA in conjunction with archaeological excavations can help to obtain critical information when other proxies do not. No doubt, aeDNA is a highly useful addition to the paleoethnobotanical toolkit. However, there are studies (e.g., refs. 22 and 23) that have shown that there is significant taxononomic overlap in the recovery capabilities of the proxies mentioned herein, with each approach having strengths that complement each other when used together. Accordingly, it is most effective to use them all concurrently whenever possible. In conclusion, the future is bright for those seeking information about crop origins and ancient habitats because of the many different options currently available for in-depth analysis.

Fig. 1.

Bird’s eye view of the Palace Reservoir at the Late Classic (600 to 870 CE) Maya city of Tikal showing vegetation based on aeDNA results (22). The banks of the main reservoir were landscaped with small trees, flowering shrubs, short palms, and tropical grasses. At the far end of the reservoir (western side) was an ancestral grove with large trees surrounding a spring pool that served as a portal to the underworld. The large pyramid to the Left of the reservoir is Temple V and on the Right side is the Central Acropolis, residence of the rulers of Tikal. Artwork by Yaniv Korman, sponsored by a grant from the Dumbarton Oaks Garden Landscape Studies program awarded to D.L.L.