Quaternary International 404 (2016) 137e155 Contents lists available at ScienceDirect Quaternary International journal homepage: www.elsevier.com/locate/quaint
n area, highland Late ninth millennium B.P. use of Zea mays L. at Cubila Ecuador, revealed by ancient starches
n-Jime
nez a, *, Ana M. Guachamín-Tello a, Martha E. Romero-Bastidas a, Jaime R. Paga Angelo R. Constantine-Castro b a
n, Instituto Nacional de Patrimonio Cultural, Ave. Colo
n Oe1-93 y Ave 10 de Agosto, Quito, EC 170129, Ecuador Laboratorio de Investigacio
gicos y Antropolo
gicos, Facultad de Ingeniería en Ciencias de la Tierra, Escuela Superior Polit
Centro de Estudios Arqueolo ecnica del Litoral, Vía Perimetral Km. 30.5, Campus Gustavo Galindo, Guayaquil, EC 090150, Ecuador b a r t i c l e i n f o a b s t r a c t Article history: Available online 20 October 2015 Today, maize is acknowledged as a plant with a great culinary and industrial versatility. It also has a deep relationship with the native cultures of the Americas and is still a vital food source for hundreds millions of people worldwide. By means of starch grain extraction from ancient lithic artifacts used more than 8000 years ago, here we report what is so far the oldest documented occurrence of maize in highland South America. This study places maize, together with other important economic plants, in the southern Ecuadorian Andes during a period coinciding with the initial stage of maize diversification and long distance expansion after its domestication in southwestern Mexico. These results allow us to unravel an early episode of human innovation previously unknown for South America which is related to the first steps toward the full re-shaping of human subsistence strategies in the continent. © 2015 Elsevier Ltd and INQUA. All rights reserved. Keywords: Maize Domestication Dispersals Andes South America Starch grains 1. Introduction Research on the very early dispersal and use of domesticated plants in South America has traditionally been focused on coastal or closer inland areas of the western neotropical lowlands (Piperno and Pearsall, 1998; Pearsall, 2008; Grobman et al., 2011; Piperno, 2011). Late Pleistocene and early Holocene archaeological sites in those regions have revealed that some important crops such as manioc (Manihot esculenta C.), sweet potato (Ipomoea batatas), squash (Cucurbita spp.), and maize (Zea mays) were domesticated, or introduced, into northern and northwestern South America at sometime around 10,000 to 7500 cal. BP (Piperno et al., 2000; Piperno and Stothert, 2003; Pearsall, 2008; Piperno and Dillehay, 2008). Regarding maize, it has been postulated that shortly after its initial domestication in southwestern Mexico by around 10,000 to 9000 BP (Matsuoka et al., 2002), and its later acquisition and use by humans in the mid elevation tropical deciduous forests of that same area by 8970e8600 cal. BP (Piperno et al., 2009), this crop was rapidly dispersed to the south, reaching southern Central America * Corresponding author.
n-Jime
nez), ana.guachamin@ E-mail addresses: jpaganpr@yahoo.com (J.R. Paga inpc.gob.ec (A.M. Guachamín-Tello), martha.romero@inpc.gob.ec (M.E. RomeroBastidas), renattoconstantine@hotmail.com (A.R. Constantine-Castro). http://dx.doi.org/10.1016/j.quaint.2015.08.025 1040-6182/© 2015 Elsevier Ltd and INQUA. All rights reserved. between 8014 and 7700 cal. BP (Dickau et al., 2007; Piperno, 2011). According to recently gathered microbotanical data (Aceituno and Loaiza, 2014), humans introduced maize into the mid-elevation and very humid premontane forests of the Colombian Andes in South America, as indicated by maize starch grains recovered in various handstone tools from El Jazmín and La Pochola sites in contexts that were dated to ranges between ~8000 and 7820 and ~7700e7600 cal. BP, respectively. Concomitantly, maize also reached the seasonally dry forests of the coastal lowlands of Ecuador (Las Vegas-OGSE-80, Santa Elena Province) at ~8053e7818 cal. BP (Fig. 1a), which is an earlier date than those
and Colombia (Piperno, 2011). After the obtained from Panama early entry of maize into northern South America between ~8100 and 7800 cal. BP, its dispersal and importance was increasingly evident and consistent through space and time, especially in Colombia (Piperno, 2011), Ecuador (Pearsall et al., 2004; Pearsall, 2008; Zarrillo et al., 2008), and Perú (Perry et al., 2006; Grobman et al., 2011). However, little attention has been paid to understanding early phytocultural dynamics in areas such as highland South America during the Late Pleistocene/Early Holocene transition, despite the presence of some important micro-regions through that interface that contain numerous Paleoindian, HuntereGatherers, or Preceramic archaeological sites (Temme, 2005;
nchez, 2011; Constantine, 2013). Thus, Stothert and Sa 138
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga Fig. 1. Selected Ecuadorian preceramic sites and area of recent findings. (A) Approximate location of known preceramic areas and sites in Ecuador. (B) Distribution of preceramic localities within the Cubil
an area and periphery. The big dark circle in Fig. 1B represents the approximate area where previous studies identified another 23 preceramic-like localities (known as Cubil
an area). interpretations regarding early plant domestication and dispersals into northwestern South America are mainly drawn from the archaeobotanical data from coastal Ecuador and mid-elevation areas of Colombia (Piperno and Stothert, 2003; Stothert et al., 2003; Piperno, 2011; Aceituno and Loaiza, 2014). This situation subtly promotes the idea that the lowland or mid-elevations areas of northwestern South America were the first regions, and perhaps the only ones, directly implicated in these early hemispheric scale events. This article report the results of recent microbotanical studies
n site #2 (Cu-S2) in highland Ecuador (Fig. 1a and b), from Cubila where we recovered maize starches from milling and scrapping lithic tools associated with contexts dated to 8078 e 7959 cal. BP. This is the oldest evidence of maize in South American highlands.
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga Starch grains from other low to mid-elevation plants such as manioc (tentative identification), chili pepper (Capsicum sp., tentative identification), wild yam (Dioscoreaceae) and wild Calathea sp. were also identified. We first briefly describe the geographical setting and the chronological context of the findings. Then, we provide detailed data on artifact management, sample extractions, testing of potential contaminants (modern starches) during the analysis, and on comparative criteria for starch grain identification for which we used more than 40 modern maize landraces aimed at securing ancient maize's starch identification. Archaeobotanical results are presented and discussed in order to suggest potential uses of maize cobs and kernels at this early archaeological context. Finally, results are used to frame the first hypotheses regarding early food plant dispersals and inter-regional interaction among highland and lowland South America. Therefore, based on the evidence thus far obtained, it is proposed that some
n area in Ecuador were engaged in early inhabitants of the Cubila the manipulation and cultivation of maize and other useful plants in the highlands or in the neighboring midlands. 2. Context of the findings: micro-regional setting, site stratigraphy, chronological dataset and sample provenance As part of an ongoing nationwide research program on the paleoethnobotany of Ecuador's ancestral cultures, archaeological prospection and test excavations were conducted at four previously unknown highland archaeological localities within the
n area (Temme, 2005) in On ~ a, Provincia de Azuay, southern Cubila
n area cover about ~52 km2 between Loja and Ecuador. The Cubila Azuay Provinces where previous studies identified 23 other archaeological localities, characterizing two of them as lithic workshops or camp sites of hunter-gatherer societies that 139 occupied the region by ~12,700 to 9900 cal. BP (Temme, 2005). Site Cu-S2 is located within the Upper Jubones river watershed, an irregular mountainous area located between 2400 and 3170 m above sea level (asl). The topography consists of narrow and sometimes partially flat ridges, and small elongated valleys surrounded by hills and old fluvial terraces. At present, Cu-S2 is in the very humid montane forest that falls between two contrasting ecological zones (Ecuador República del, 2011): the very dry tropical forest ~5 km to the northwest, and the very humid premontane forest ~13 km to the southeast.
ramo herbs such as Calamagrostis intermedia P
aramo and sub-pa dominate the open vegetation of Cu-S2 and its surroundings,
ramo shrubs are also common. although p
aramo and sub-pa Discrete patches of primary humid montane forest are located on the east and west side of steep slopes which serves to protect them from constant and sometimes strong winds that come from the Amazon Basin. However, by around 8000 cal. BP palaecological records for southern Ecuadorian p
aramo show a marked increase in charcoal influx that has been related to human-caused fires intentionally set to clear lands, to drier and warmer climatic conditions, and to an expansion of the forest to higher altitudes when
n area (Jantz compared to current conditions registered at Cubila and Behling, 2012; Villota et al., 2012). This would explain the modern patchy distributional pattern of humid montane forests remnants within protective slopes at nearby altitudes. Cultivation practices or human settlements have not been recorded in recent times for the study area and there are no ethnographic and historical data which could place any kind of plant production activities, present or in the recorded past, at Cu-S2 or its periphery. However, some small river banks and fluvial terraces found 2e4 km west of Cu-S2, at elevations between 2600 and 3000 masl located at the bottom of those protective slopes, have potentially good soils Fig. 2. Planview of Cu-S2 showing test units excavated during 2013 and 2014 field seasons, and artifact horizontal distribution among identified strata.
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga 140 Fig. 3. South stratigraphic profile of test unit 1A at Cu-S2, and approximate projection of the studied artifacts and and warmer conditions suitable for the cultivation of several high altitude edible plants such as potato (Solanum tuberosum), melloco (Ullucus tuberosus), maize, and some legumes. Other production practices, such as sporadic cattle grazing, have been confirmed for
ramo area, although owners of these communal lands this sub-pa abandoned this practice several years ago. Surface collection data from the Cubilan area revealed lithic debitage and formalized and expedient flaked tools (e.g., scrapers, projectile points, unifaces and bifaces) produced by bipolar and free-hand flaking. These artifacts, when not related to other archaeological features such as ceramic fragments and ground stone tools, are typically ascribed to Ecuadorian preceramic techanchez, 2011; nologies of lithic production (Stothert and S
Constantine, 2013). 14 C range of dates. At Cu-S2, no pre-Columbian ceramics or modern artifacts were found. Two 2  2 m test units excavated during 2013 at this locality produced 10,922 artifacts, mostly siliceous lithic debitage (n ¼ 10,192). The ten lithic tools selected for this analysis (Appendix A) come from test units 1 and 2 (Fig. 2): seven formalized scrapers, two unmodified flakes with marginal use-wear consistent with scraping activities, and one lightly utilized rudimentary milling stone base fragment. These tools were recovered from stratum IIa and underlying stratum II (strata numbered IeII, top to bottom, Fig. 3). Stratum Ia and stratum I did not yield archaeological tools, though lithic debitage of siliceous origin (chert, chalcedony) were recovered. Thus, no lithic artifacts from these strata were included in this study. Additionally, five 14C dates were obtained from the 2013 test units (Table 1). Table 1
n area, Ecuador. Chronological dataset for Cu-26, Cu-27, and Cu-S2 (this study), Cubila Site Radiocarbon Lab. Code Provenance Uncal. Date Dated material 2 s (BP) and relative area Reference Cu-27 Cu-27 Cu-26 Cu-26 Ki-1640 Ki-1642 Ki-1859 Ki-1860 No No No No 10,500 ± 130 10,330 ± 170 9100 ± 120 9160 ± 100 charcoal charcoal charcoal charcoal 12,683e11,946 (0.993306) 12,560e11,393 (0.993192) 10,521e9856 (0.969113) 10,571e10,125 (0.9482) Temme Temme Temme Temme Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Cu-S2 Beta-362877 Beta-362876 UGAMS-19143 UGAMS-19134 Beta-364214 Beta-364213 Beta-362878 UGAMS-19130 UGAMS-19140 UGAMS-19141 UGAMS-19136 UGAMS-19137 UGAMS-19139 UGAMS-19138 UGAMS-19135 UGAMS-19132 UGAMS-19131 Test Test Test Test Test Test Test Test Test Test Test Test Test Test Test Test Test >43500 13,430 ± 60 8780 ± 30 8760 ± 30 8360 ± 40 7260 ± 40 7210 ± 40 4340 ± 25 3930 ± 25 3900 ± 30 3860 ± 25 3850 ± 30 3740 ± 25 3720 ± 25 3630 ± 25 3510 ± 25 3410 ± 25 charred remains organic matter sediments charred wood charred wood charred wood charred wood charred wood charred wood charred wood charred wood charred wood charred wood charred wood charred wood charred wood charred wood charred wood NA 16,309e15,878 (1.0) 9832e9560 (0.917446) 9795e9550 (0.966459) 9463e9233 (0.925786) 8078e7959 (0.714983) 8057e7927 (0.938595) 4893e4828 (0.8428) 4419e4233 (0.982686) 4411e4218 (0.828138) 4299e4140 (0.845255) 4299e4086 (0.941832) 4101e3961 (0.861194) 4091e3904 (0.982607) 3982e3826 (0.979809) 3833e3678 (0.885409) 3695e3555 (0.950189) This This This This This This This This This This This This This This This This This data data data data Unit 1, Stratum II Unit 1, Stratum II Unit 2a, Stratum IIa-II Unit 2a, Stratum IIa-II Unit 1, Stratum IIa Unit 1, Stratum IIa Unit 2, Stratum IIa Unit 1a, Stratum I Unit 3, Stratum I Unit 3, Stratum I Unit 3, Stratum I-IIa Unit 1a, Stratum I-IIa (ex-situ, sieves) Unit 3, Stratum I Unit 3, Stratum I Unit, 2a Stratum I Unit 1a, Stratum I Unit 1a, Stratum I (2005) (2005) (2005) (2005) study study study study study study study study study study study study study study study study study Table notes: All dates in this table were calibrated using Calib Radiocarbon Calibration Program, version 7.0.1 and Southern Hemisphere 13 radiocarbon calibration curve (SHCal1314C). Other dates quoted in the article were all calibrated using Intcal13 with the program mentioned above.
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga During 2014, two 2  2 m test units and one 2  1 m test trench were excavated at Cu-S2 (Fig. 2) in order to collect more organic samples for 14C dating and for better defining vertically segregated occupational layers at the site. Twelve additional dates were obtained (Table 1). Combining artifact and chronological data gathered at Cu-S2 in 2013 and 2014, the following stratigraphic scenario emerges. Four stratigraphic units were identified: stratum Ia (O horizon), stratum I (or A1 horizon), stratum IIa (A2/AB horizon), and stratum II (B horizon). Stratum III (C horizon) was identified in four of the five excavation test units, though it was not excavated. From these stratigraphic units, at least six different occupational layers were identified after finely excavating levels within strata at 1 cm intervals for detecting subtle horizontal distributional changes in artifact and/or organic (e.g., charred wood fragments) assemblages. Stratum Ia is a very dark brown to black layer of clayey silt of 8e10 cm thick with a high content of modern organic material such as grass (Calamagrostis intermedia) and shrub roots, and worms. Bioturbation processes related to these agents are evident in this stratigraphic unit that contained scattered lithic debitage. No formal or expedient lithic tools were identified among the recovered artifacts. The underlying stratum I is also a very dark brown to black clayey silt layer 10e12 cm thick. However, unlike stratum Ia, the organic content of this unit is well integrated into the soil because of ancient decomposition processes that took place in it. Modern roots and worm activity were noted in the upper section of this stratum. Lithic artifacts (debitage and core fragments) were recovered at three different occupational layers within stratum I. The uppermost layer revealed very few small lithic debitage in association with tiny charred fragments of wood. Five dates that range between 4101 and 3555 cal. BP (Table 1) are directly related to this layer as well as with the bottom of stratum Ia. Two more layers vertically segregated along the middle and lower sections of stratum I revealed chert or chalcedony debitage with cortex, including a few exhausted small multidirectional core fragments. No formal or expedient tools were recovered. From these two lowermost layers of stratum I, five additional dates were obtained that range between 4893 and 4086 cal. BP (Table 1). Together, all ten 14C dates obtained from the three occupational layers in stratum I are chronologically related to middle and late Formative agroceramic manifestations previously documented in the Ecuadorian highlands, such as those recorded at Pirincay and ~ ar Province, Challuabamba in Azuay Province, Cerro Narrío in Can Alausí in Chimborazo Province, Cotocollao in Pichincha Province
n area), and Catamayo B at (all the above, to the north of the Cubila Loja Province (to the south of Cubil
an area) (Zeidler, 2008). From
n area was a middle to the above, it seems plausible that the Cubila late Formative period locale for the periodic procurement of stone raw materials (chert, white chalcedony) that were initially processed at this highland locality. Below stratum I there is a transitional (AB or A2 horizon) layer labeled as stratum IIa. This silty clay unit is a buried AB horizon of 9e10 cm thick with oscillating colors between very dark grayish brown to yellowish brown. It is considered a transitional unit because some features from the upper stratum (e.g., silt, decomposed organic matter) precipitated into the upper section of this layer, though the matrix is a plastic clay more similar to the lower stratum II (B horizon). The middle section of stratum IIa in test units 1 and 2 revealed a single occupational layer that was sampled for 14C dating, resulting in two statistically overlapping dates between 8078 and 7959 cal. BP (Table 1). Abundant chert and chalcedony debitage were recovered, as well as some expedient and formal scrapers. Under stratum IIa we documented stratum II (B horizon), which is a 40e50 cm thick strong brown plastic-residual clay unit. This horizon was subdivided into “Bb” and “B”. Bb is a subhorizon at the 141 uppermost section of B horizon, where archaeological artifacts were detected at two apparently segregated layers. An upper layer containing scattered lithic debitage is about 4 cm beneath the contact zone between strata IIa and II. A single date associated with this apparently brief occupational event ranges between 9463 and 9233 cal. BP Additionally, two other dates that range between 9832 and 9550 cal. BP (Table 1) are associated with the lowest and first occupational layer identified in stratum II, between Bb and B horizons. This layer revealed very dense lithic debitage including flakes, chips, and core fragments. Below subhorizon Bb, there is the B horizon containing unmodified chert and chalcedony blocks (some of them as long as 25e30 cm) that are associated with the geological formation of this stratigraphic unit. We documented isolated and conglomerated layers of these stone blocks at different profiles across the Cubilan area, outside any known or registered archaeological locality, thus confirming the true geologic nature of these siliceous stones within B horizon. However, small to medium size flakes and blocks with signs of damage similar to those produced by initial flaking of cores were detected in this horizon in test units 1, 1a, 2 and 3. These materials were not horizontally aggregated or contextually related to each other. However, various lithic materials which could be true archaeological tools were selected for analysis from this horizon. Two dates obtained by the extraction of organic content from this matrix at two points near the middle section of the horizon were: 16,309e15878 cal. BP, and >43,500 uncal. BP (Table 1). The later date was rejected by the radiocarbon lab due to insufficient organic matter. Finally, the surface of a C horizon was reached. It is a mixed unit of residual clay and unconsolidated parental material (andesite, rhyolite and pyroclasts) (Litherland et al., 1993) that notably underwent different alteration processes in our study area (e.g., decomposition, remineralization). This late Miocene-Pliocene horizon is archaeologically sterile, although some unmodified siliceous blocks were identified near its surface. Four of the studied tools came from stratum IIa dated to 8078e7959 cal. BP, while the other six were from the underlying stratum II. All these tools were next to, or below the previously mentioned range of dates, thus confirming that any tool or starch grains from stratum IIa or lower, fall within, or are earlier than these calibrated dates. However, six tools that yielded maize starches from upper stratum II (14-02, 14-05, 14-06), and from middle and lower stratum II (14-04, 13-98 and 13-99), were all located near or below the earliest dated occupational layers that ranges between 9832 and 9233 cal. BP. The apparent relationship between maize starch grains and these earliest occupational layers at Cu-S2 has been rejected for purposes of this article because these dates clearly predate the estimate for the onset of maize domestication in upper Central America (Matsuoka et al., 2002; Piperno et al., 2009). This range of dates simply chronicles the older anthropogenic deposit which rest on the top of stratum II (horizon B). Maize starch grains recovered and artifacts recovered from below this first occupational layer, are assumed to be intrusive. Below the earliest occupational layer at Cu-S2, there are two older dates (16,309e15878 cal. BP, and >43,500 uncal. BP) that were discarded by us because their provenance is the internal matrix of stratum II (horizon B), a layer of residual clay formed in situ by both alluvial processes and bedrock decomposition. These two dates are spatially related to unmodified siliceous rocks and to a few lithic artifacts that are clearly unassociated. In this case, the rocks and lithic materials are aligned vertically and discontinuously, so it seems these are not part of any true anthropogenic layer. Until more geochemical and geophysical analyses are conducted to better understand this irregularity, we infer that artifacts with maize starches were vertically transported by unknown processes from the basal section of stratum IIa to the 142
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga buried horizons below when it was an active anthropogenic surface. Potential disturbance processes that are now undetectable, such as the digging of vertical holes to access subsurface chert or chalcedony blocks, could have been created when the surface of stratum IIa was anthropically active. In this case, because no marked differences in color or texture has been noted at the base of stratum IIa and underlying stratum II, intrusive holes or other nonanthropic events such as localized small landslides near the edge of the archaeological deposits, could have displaced some artifacts that were originally located in stratum IIa downward. 3. Materials and methods 3.1. Artifact management, sample extraction and starch grain recovery Artifacts selected for this study were isolated from the field and placed individually and unwashed in new plastic bags. Standard criteria were used for lithics descriptions (Ranere et al., 2009). Methods for sediment extraction from lithic tools (Piperno et al.,
n-Jime
nez, 2009), and starch separation from sediments (Paga 2007) were modified from others published elsewhere. Once in the lab, artifacts were carefully rinsed to gently remove any modern potential contaminant, and then were placed separately in new and sterile zip lock bags, adding enough water to cover each artifact. Each bagged artifact was then placed into an ultrasonic bath for 40 min in order to release sediments from their used margins or surfaces, pores and cracks. Released sediment or residues were placed in new and sterile 50 ml centrifuge vials, and then centrifuged at 3500 rpm  15 min. Water was decanted and sample vials were placed in a clean lab oven at 31  C for 60 h until sediments were completely dried. A CsCl solution with a specific gravity of 1.79 was added to each sample, sufficient to cover the solid sediments in the vials by 4 cm. Vials were agitated manually for 20 s, and then more CsCl was used to release sediments and residues from the walls and lids of the vials. Centrifugation proceeded later at 3000 rpm for 20 min, and all the solution with the supernatant was transferred into new and sterile centrifuge vials. A second round of CsCl flotation was executed with the original samples in their own vials in order to catch more residues that could remain trapped within the matrix. The resulting solution and supernatant was also placed in the previously used centrifuge vials where the solution of the first round flotation was stored. Sufficient bi-distilled water was added to each sample to break down the 1.79 density of the CsCl solution. In this way, starches and other residues would be forced to sink to the bottom of the vials. Centrifugation at 4000 rpm was executed for 20 min and then the undesirable supernatant was decanted. The previous step of adding more water to each vial was executed four more times for diluting the remaining salts of the solution while retaining the residues at the bottom of the vials. Once the solution with the residue samples was completely washed from inside each 50 ml vial they were transferred into various new and sterile 2.0 ml microcentrifuge tubes for each sample. Bi-distilled water was added to each microcentrifuge tube and then centrifuged at 8000 rpm for 10 min. After that, the water at the top of each tube was taken off and the remaining samples (at the bottom of the tubes) were allowed to dry in isolation at 30  C inside a clean lab oven. Throughout the washing steps described above, the discarded solution was repeatedly tested to check to be sure that we were not losing ancient starch grains; no starch grains were detected at any point during the process. Following Zarrillo (2012), a major concern of this study was to test if modern starches were potential contaminant agents of the archaeological samples. To assess this, we submitted all the materials used during the extraction and flotation process to close examination for starch grains. Unpowdered nitrile gloves, glass beakers, new disposable pipettes, new zip lock bags, new and sterile centrifuge vials, new microscope slides and cover slips, as well as solutions and reagents such as glycerol, unused CsCl, and bi-distilled water were all sampled separately looking for potential modern starch grain contamination, following standard microscope sample mounting procedures. This testing resulted negative to the presence of contaminant starches. As a further precaution, new microscope slides and cover slips were washed with bi-distilled water and allowed to dry in isolation at 80  C inside a clean lab oven (Zarrillo, 2012). These materials were then placed in a plastic microscope slide holder that was previously washed and tested for possible modern starch contamination. After conducting all the steps described above, microscope bidimensional analysis consisted first in placing sample drops on microscope slides (sometimes up to 4 preparations per sample), and then mixing it with glycerol for adding viscosity to the media, which allow controlled movement and rotation of the found starches. Cover slips were placed and later microscope examination began. We used an Olympus BX53optical microscope with polarized capacity and two different digital cameras: an Olympus DP26CU, and an Infinity 1-2CB with their respective software. Every slide was surveyed entirely with a 10X objective, starches found were photographed, and 12 morphometric variables were recorded for each of them with a 40X objective. Finally, their position was registered with coordinates for allowing further revisions, and microscope slides were stored in new cardboard slide holders. Lithic tools analyzed for this study, and samples extracted from them, have been identified with a unique code. All of the tools recovered and the pertinent documentation for the CU-S2 studies are permanently curated at the Instituto Nacional de Patrimonio Cultural (INPC) in Quito. Please note that Ecuador's Decree No. 2600 of 9 June 1978 created the INPC. The INPC regulates the performance of anthropological studies, including archaeology, in Ecuador. As such, the INPC is the lead Ecuadorian authority regulating the practice of archaeology and is charged by law for safeguarding the archaeological heritage of the nation. No
n for permits were required for the Laboratorio de Investigacio conducting the field and laboratory studies. However, the field and laboratory studies performed for this research strictly adhered to INPC's national regulatory framework such as the
n de Permisos de Investigacio
n Reglamento para la Concesio
gica Terrestre of 20 February 1992, in compliance with Arqueolo national and international academic, professional and ethical standards of the discipline. 3.2. Comparative criteria for starch grain identification Starch identification used 140 modern neotropical and Andean
nstarchy plant specimens from the Laboratorio de Investigacio
n-Jime
nez’ reference collections. INPC (LIINPC), Ecuador, and Paga Taxonomical ascription of archaeological starch grains consisted of their comparison to a reference collection stored at the LIINPC made up of 88 neotropical specimens that comprise 22 plant families, 27 genera, and 36 species which includes 39 different South American maize landraces, 7 different Phaseolus spp. varieties, and 2 different Canavalia spp. species, among other important
n-Jime
nez, 2015). economic and wild plants of Ecuador (Paga
n-Jime
nez’ reference collection of panAdditionally, we used Paga tropical starchy plants that include 77 specimens comprising 68 genera and 61 species, in which 6 different Circum-Caribbean native maize landraces are represented. Published literature on diagnostic criteria for starch grains was also consulted (Reichert, 1913; Piperno and Holst, 1998; Piperno et al., 2000, 2009; Pearsall et al., 2004; Perry, 2004; Holst et al., 2007; Perry et al., 2007; Piperno and Dillehay, 2008; Zarrillo et al., 2008; Horrocks and
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga 143 Fig. 4. Selected secure and tentatively identified maize starch grains recovered from lithic tools. (AeD, FeN), secure identifications of maize starch. (E, OeP), tentative identifications of maize starch. (AeP), starches with transverse (tf), radial (rf), “y”(yf), and asymmetric (af) fissures. (D, I, J), starches showing also radial striations (pressure damaging). (A, B, D, G), starches with oval shapes commonly found together with irregular (polygonal) shapes in hard kernel South American landraces (see also Appendix B). (K, M), polygonal starches with open hilum. (C, E, F, HeP), irregular and polygonal shapes usually predominant in hard kernel landraces. Bell-shape starch (F) is widely present in many Andean
nez, 2015). (C, HeP), starches with typical pressure facets (pf) found on hard and soft kernels of South American landraces. (E, F, H, I, landraces (see e.g., Appendix B(C)) (Pag
an-Jime J, L), partially to heavily damaged starches showing typical double-border (db) characteristic of the species. Starch provenances: AeC, G [tool 13-99]; DeF, HeJ [tool 13-98]; K [tool 14-03]; L [tool 14-07]; M [tool 14-01]; N [tool 14-05]; O [tool 14-02]; P [tool 14-06].
n-Jime
nez, 2012; Rechtman, 2009; Mickleburgh and Paga
nez, 2015), whereas damaging Musaubach et al., 2013; Pag
an-Jime patterns observed in the recovered starches were contrasted to our own published work or to previously published literature (Babot, 2003; Piperno et al., 2004; Lamb and Loy, 2005; Henry et al.,
n-Jime
nez, 2012). Secure and tenta2009; Mickleburgh and Paga tive identifications of starches in this study are based on diagnostic and/or distinctive features described elsewhere. Specific morphometric features used are shape, size, presence and location of the hilum within the granule, presence and appearance of fissures, presence and type of pressure facets, presence and appearance of lamellae, and in some cases the appearance and projection of the Maltese cross. Because this study presents a new chronological and regional assignment of maize, a brief discussion on the morphometric discrimination between starches from maize and other Poaceae is in order to clarify any potential identification bias. There is no suggestive published data of other wild grasses in the Americas attesting to true morphometrical similarities between maize starches and those produced by other Poaceae. On the contrary, when comparing starches from maize and other wild grasses, Musaubach et al. (2013), Pearsall et al. (2004), Piperno et al. (2009), 144
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga Fig. 5. Selected starch grains of other specimens recovered from lithic tools. (A), damaged starch of Dioscoreaceae showing diffused Maltese cross (tool 13-98). (B), damaged starch of Calathea spp. with lamellae (l) (tool 13-98). (C), possible manioc starch with “v” flexion lines (vf) and hilum (h) (tool 13-99). (D), possible chili pepper starch with longitudinal fissure (lf) (tool 14-04). (E), damaged starch of Phaseolus spp. showing smooth lamellae and an irregular longitudinal fissure (tool 14-05). Reichert (1913), and Zarrillo (2012) clearly show what is so far a unanimous scenario: wild grasses from Central and South America produce smaller and/or different starch shapes and metrics clearly discernible from those of maize. However, Holst et al. (2007) showed that one wild teosintle-like grass (Zea mays spp. mexicana) produces starches strikingly similar to those stored in some domestic maize kernels. In spite of this, these authors believe that starches from this wild grass can be discriminated from the ones produced by domestic maize. On that basis, Piperno et al. (2009) later published their results on the identification of starch grains from the earliest domestic maize found to date in the New World, which is a finding located in a geographic area of western Mexico that is the natural habitat of many non-Zea and at least one Zea species. Therefore, based on the above, South America (Ecuador) does not presently have, nor did it have in the remote past, any native Zea species growing in the wild at lower or higher altitudes. The only recurrent wild Poaceae in our study area is Calamagrostis intermedia, whose tiny (less than 5 mm) kernels do not produce starches comparable to the ones from maize. Another South American wild grass which produces only a few starch shapes closest to those of maize is Sorghastrum pellitum, according to Musaubach et al. (2013). This grass genus (and species such as S. nutans) has not been registered in our study area and is more closely related to drier midlands and lowlands. Following Zarrillo (2012), and Pearsall et al. (2004), who studied several Ecuadorian sites including some in the highland, Cenchrus brownii produces starches in its seeds that are similar, but distinguishable from maize. Ecuador has at least 5 species of this genus that can be distributed from 0 to ~3000 masl. This genus, which is easily recognizable in the field, is not known to occur in our study area. In synthesis, previous researchers have indicated that no other studied wild Central or South American grass produces starches morphometrically similar to those of maize. Regarding maize and other Poaceae, it is important to consider what we know about artifact use and use-wear. Nine of the 10 studied artifacts in our study are formal or expedient scrapers. Grasses studied to date, or potentially present in South America, cannot be processed with these artifacts, and maize is the only grass in the region with rigid cob/peduncle that could be scraped to release its green or mature kernels. With this in mind, our starch grain analysis has taken into account all the above in proposing confident identifications of ancient maize starches. 4. Results and discussion The identification at Cu-S2 (Table 2; Fig. 4), of 103 individual, plus ~68 clustered starch grains, all from maize were fully or tentatively identified using diagnostic criteria previously published for maize and other non Zea taxa (Piperno and Holst, 1998; Piperno et al., 2000, 2009; Pearsall et al., 2004; Perry et al., 2006; Dickau et al., 2007,
n-Jime
nez, 2007, 2012; Berman and 2012; Holst et al., 2007; Paga Pearsall, 2008; Zarrillo et al., 2008; Dorsey et al., 2009; Grobman
n-Jime
nez, 2012; Musaubach et al., 2011; Mickleburgh and Paga
n-Jime
nez and Rostain, 2014) and confirmed by et al., 2013; Paga 161 (þc. 68) 5 37 c. 68 78 25 2 2 2 2 1 13-99 13-98 3 1 14-04 14-05 14-06 14-02 formal. tool, scraper flake with retouched borders formal. tool, burin with use-wear signs (scraper type) in its longitudinal borders Unit 1, St. II formal. tool, scraper flake 20e30 cmbd with retouched borders Unit 1, St. II expedient tool, scraper flake 30e40 cmbd with use wear signs Unit 1, St. II expedient tool, scraper flake 45 cmbd with use wear signs Unit 1, St. II expedient tool, milling stone 50e60 cmbd base fragment Unit 1, St. II formal. tool, “bec” type scraper 60e70 cmbd Total 14-08 1 1 1 1 3 1 2 1 1 48 (þc.60) 2 11 1 1 2 1 2 7 22 10 3 5 3 1 25 c. 60 10 5 1 2 2 47 15 5 4 6 2 2 4 3 14-07 formal. tool, end-scraper 1 3 14-03 6 c. 8 3 3 1 8 (þc.8) 5 15 8 7 1 14-01 formal. tool, scraper flake with retouched borders formal. tool, scraper Unit 1, St. IIa 10e20 cmbd Unit 1, St. IIa 10e20 cmbd Unit 1, St. IIa 10e20 cmbd Unit 1, St. IIa 10e20 cmbd Unit 1, St. II 20e30 cmbd Sample no. cf.Capsicum Calathea Phaseolus Fabaceae cf. Fabaceae Dioscoreaceae cf. Manihot cf. Sagittaria cf. Zea Zea Cluster cf. N.I. N.I. Tuber Total spp. spp. spp. esculenta latifolium mays mays Zea mays Tool type Provenance Table 2 Starch grains recovered from tools at Cu-S2, with sample proveniences and associated 14 C dates
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga 145 our own modern maize reference collection of 39 maize landraces/
n-Jime
nez, 2015) (Table 3, specimens stored at the LIINPC (Paga Appendix B). Starches from other important potential food plants such as beans (Fabaceae, Phaseolus spp.), yams (Dioscoreaceae), calathea (Calathea spp.), manioc, Indian arrowhead (Sagittaria spp.), and chili pepper (cf. Capsicum spp.) (Fig. 5), either from midelevations or the lowland, were also fully or tentatively identified in the same dated contexts (Table 4), adding potential information on poorly understood interregional phytocultural interactions between the South American lowlands and the highlands. For a more detailed discussion on these additional identifications, see Appendixes C to G. Table 5 shows specific bi-dimensional and morphometric data on tentative and secure maize starch grain identifications. According to specialized literature (Reichert, 1913; Pearsall et al., 2004; Holst et al., 2007; Piperno et al., 2009; Zarrillo, 2012; Musaubach et al., 2013) and our own maize comparative speci
n-Jime
nez, 2015) (Fig. 6), the mean length of the mens (Paga recovered maize starches, as expected, exceeds any studied wild grass or other Zea starch mean size. Maize starches were recovered on all ten tools, but were more noticeable in those two recovered between 50 and 70 cmbd in test unit 1. Tool 13-98, a rudimentary grinding tool fragment, produced maize starches that show damage features highly consistent with pressure (grinding/pounding), such as pronounced radial fissures/striations, as well as light to heavy modifications on its common shapes, documented in the heaviest
nez, 2012). tools 13-98 and 13-99 (Mickleburgh and Pag
an-Jime Another clear pressure-damage indicator, noted previously in experimental maize starches, is the direct relationship that exists between kernel hardness and starch enlargement due to differential kernel resistance during intensive grinding (Mickleburgh and
n-Jime
nez, 2012). Neither maize starch of the tools discussed Paga above, nor those recovered on the other 8 scrapers, show size ranges that fit any experimentally documented starches produced during grinding (Fig. 6). However, they match nicely with some modern maize landraces with hard or soft kernels from Ecuadorian
n, Blanco Blandito Andes mid-to high elevations (e.g., Mishca/Sapo redondeado, Morochillo, Tusilla) that were not experimentally modified. Tusilla, a mid-elevation maize with hard and vitreous endosperm is the only studied Ecuadorian landrace that produces mostly irregular starches clearly comparable to those documented on all ten tools. Other modern maize landraces with partially hard to fully hard kernels from surrounding countries such as Peru and Colombia commonly produces irregular starches similar to the ~ o, ancient ones described above (landraces Sabanero, Amagacen ~ o, Cabuya, Güirua, Imbricado, Montan ~ a, Pira, Puya Chococen
n-Jime
nez, 2015). Because maize endosperm formaGrande) (Paga tion and amylose synthesis are both genetically controlled (Smith et al., 1995), the morphometric features of maize starches identified on the tools suggest that the maize processed and used in
n was one with hard endosperm (kernel) and genotypic Cubila characteristics comparable to the modern Tusilla landrace. Hard kernel maize was previously suggested as the earliest domesticated type of maize for Mexico (Piperno et al., 2009) on the basis of recovered ancient starch grains that are noticeably similar to the ones identified here. Eight of the analyzed tools are relatively small formalized or expedient scrapers. Although they contained starches with signs of damage caused by pressure, the damage occurs at a lower percent than in the larger grinding and scraping tools. Previous studies clearly indicate that when totally green or immature hard maize kernels are processed with these types of tools, the starches do not undergo great morphometric transformation. Conversely, when mature and partially dry maize kernels, or totally dry, hard and unsoaked kernels were processed experimentally, the starch
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga 146 Table 3 Modern and archaeological starch grain size ranges of maize specimens. Some modern maize landraces, and kernel variability, were submitted to different grinding processing. Zea mays Range of sizes in mm Mean sizes in mm No. measurements taken
n-Jime
nez, 2007) Modern, control samples from the Circum-Caribbean (Paga Mature, dry and hard kernels soaked for 24 h before grindingi a. Pollo (CIMMyTa Id#:3106) 2e28 13 ± 3.9 b. Early Caribbean (CIMMyT Id#:1347) 3e20 13 ± 3.6 c. Negrito de Colombia (CIMMyT Id#:3199) 5e20 12.3 ± 3.3 d. Cateto cristalino (CIMMyT Id#:4113) 3e18 10.3 ± 3.1 e. Chandelle (CIMMyT Id#:3879) 2e20 12.3 ± 3.2 f. Tu
son (CIMMyT Id#:5495) 1e18 12 ± 3.2 Modern, control samples from Ecuador (this study)ii Dry and hard kernels, soaked ~ o/Arizona (LIINPCb) g. Tuxpen 4e23 12.79 ± 4.62 h. Canguil puntiagudo (LIINPC) 3e26 15.47 ± 5.03 i. Blanco blandito puntiagudo (LIINPC) 5e20 12.96 ± 3.68 j. Blanco blandito redondeado (LIINPC) 10e26 17.1 ± 4.68 k. cf. Tusilla (LIINPC) 6e26 15.8 ± 5.14
n (LIINPC) l. cf. Mischca/Sapo 3e26 18.2 ± 5.93 m. Chulpi Ecu (LIINPC) 9e17 12.9 ± 1.99 n. Morochillo (LIINPC) 6e26 16.7 ± 5.2 o. Morocho (LIINPC) 4e24 13.3 ± 4.71 p. Racimo de uva (LIINPC) 4e20 12.3 ± 4.05 q. Sangay (LIINPC) 3e18 13.3 ± 3.5 r. Triunfo (LIINPC) 5e18 12.5 ± 2.7 s. Trueno (LIINPC) 6e20 13 ± 3.3 Modern, control samples after experimentationiii
nez, 2012) Green to dry kernels soaked, or not (marked with * below), for 1 h before grinding (Mickleburgh and Pag
an-Jime t. Nal-Tel,mature, dry and hard (CIMMyT Id#: 815) 11e41 21.8 ± 7.7 u. Pollo, mature, dry and hard (CIMMyT Id#: 3105) 10e38 23.2 ± 6.6 v. Pollo, semi-mature and partially hard (CIMMyT Id#: 3105) 7e34 20.8 ± 5.7 w. Pollo*, green and soft (produced after CIMMyT Id#: 3105) 5e25 12.1 ± 4.7 Archaeological starch grains from Cu-S2, all artifacts (this study)iv Zea mays 10e24.87 16.27 ± 2.89 cf. Zea mays 12.5e26.25 18.92 ± 3.21 116 101 107 107 89 109 20 20 20 20 20 20 20 20 20 20 20 20 20 60 60 60 60 78 25
n-Jime
nez, 2007) were gently ground for 15 s with a marble mortar and pestle to Table notes: (i) After soaking, kernels of control samples from the Circum-Caribbean (Paga avoid overly damaging the starch grains. (ii) Soaked kernels of control samples from Ecuador (this study) were opened and gently scraped with a scalpel. (iii) After soaking,
nez, 2012) were ground intensively for 5 min with a marble mortar and pestle in order to examine patterns of kernels of experiment samples (Mickleburgh and Pag
an-Jime damage due to pressure. (iv) Maize starch clusters are excluded. a CIMMyT ¼ Centro Internacional de Mejoramiento de Maíz y Trigo, Mexico. b
n, Instituto Nacional de Patrimonio Cultural, Ecuador. LIINPC ¼ Laboratorio de Investigacio Table 4 Size ranges of starch grains identified in this study compared to modern specimens* Mean size in mm Taxa (Identified Size range in mm archaeological (minemax. Length) of archaeological starches) of archaeological starches starches Taxa (Modern Total no. of specimen identified archaeological starches) starches Domesticates cf. Manihot esculenta Zea mays 17.50 17.50 10e24.87 16.27 ± 2.89 78 cf. Zea mays 12.5e26.25 18.92 ± 3.21 25 Cultivars or transitional to full domesticates cf. Capsicum 16.34e29.96 24.38 ± 7.14 spp. 3 Phaseolus spp. 2 27.70e31.88 29.79 ± 2.96 1 Fabaceae cf. Fabaceae 20.77e32.29 20.99e26.47 27.42 ± 5.96 23.73 ± 3.87 3 2 Wild Calathea spp. 22.50 22.50 1 Dioscoreaceae 38.75e49.5 44.13 ± 7.6 2 Manihot esculenta Size range in mm (minemax length) of modern specimens Mean size in mm of modern specimens 6.7e37.3 17.16 ± 8 References Total no. (modern measurements taken in modern specimens) specimens 20 Zea mays, see Fig. 6, Appendix B Capsicum annum annum C. baccatum baccatum C. annum aviculare Phaseolus vulgaris P. vulgaris, Boca Negra black P. lunatus, brown P. lunatus, mottled Above (Phaseolus spp.) Above (Phaseolus spp.) This study 18.7e42 13.5e41.7 2e6.1 8.1e45.5 12.3e26.9 29.3 25.1 3.5 24.4 ± 8.8 19.8 ± 4.8 25 25 50 20 20 Perry et al. (2007) Perry et al. (2007) Perry et al. (2007) This study This study 8.4e49 9.8e50.2 28.5 ± 10.5 30 ± 9.2 20 20 This study This study 20 This study Calathea spp. S.Domingo Ts C. allouia 3.5e28.3 13.6 ± 7.02 8e40 28 ± 8.6 D. D. D. D. 3.6e64 9e49 9e58 15e75 27.7 ± 16.6 20 24 ± 9.5 20 22 ± 10.8 20 38 ± 13 126 piperifolia spp.#4 spp.#5 altissima This study 126
nez Pag
an-Jime (2007) This study This study This study
nez Pag
an-Jime (2007) Table 4 (continued ) Mean size in mm Taxa (Identified Size range in mm archaeological (minemax. Length) of archaeological of archaeological starches starches) starches cf. Sagittaria latifolia 13.75e18 15.88 ± 3.0 Taxa (Modern Total no. of specimen identified archaeological starches) starches 2 Sagittaria latifolia Sagittaria lancifolia Size range in mm (minemax length) of modern specimens Mean size in mm of modern specimens 4.6e26.8 11e79 14.8 ± 5.8 54 ± 8 Total no. References measurements (modern taken in modern specimens) specimens 20 30 This study Mickleburgh and
n-Jime
nez Paga (2012) Table note: * Archaeological starch clusters are excluded. Table 5 General morphology and other surface features (%) of recovered maize starch grains. Tool no. 14-01 14-03 14-07 14-08 14-02 14-06 14-05 14-04 13-98 13-99 Shape Hilum Regular (spherical to regular oval) Irregular (irregular oval to polygonal) 12 33 88 67 100 100 100 100 100 75 66 56 25 34 44 Fissure variants* Tra 50 56 75 100 100 100 33 38 19 28 m eX Cr 25 22 Potential damaging vector Y 13 50 25 25 67 38 33 22 100 33 13 9 13 13 3 3 3 3 Rad Circ 13 22 11 25 44 50 raL dCr Press. 13 50 44 50 50 50 6 16 33 38 78 75 6 Heat 25 3 Nothing 50 56 25 100 50 100 67 62 22 22 Total no. of maize starches per sample** 8 9 4 4 2 1 3 8 32 32 Table note: *More than one fissure variant can be present in the same starch grain if rotated. **Secure and tentative identifications have been combined. Secure and tentative identifications of maize starch clusters have been excluded. Fissure variants legend: Tra ¼ transverse, m ¼ m shape, eX ¼ expanded “X” shape, Cr ¼ cross, Rad ¼ radial, Circ ¼ circle, raL ¼ ramified line, dCr ¼ deep cross. Fig. 6. Error bar plot for minimumemaximum, and mean length size of maize, comparing archaeological starches (secure and tentative identifications combined, this study) to 23 modern reference specimens. 148
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga undergoes significantly greater morphometric transformations during grinding than that observed in our archaeologically
n-Jime
nez, 2012). On recovered starches (Mickleburgh and Paga this basis, we suggest that the maize intentionally processed with the grinding and scraping tools at Cu-S2 was likely a hard kernel variety in a mature state, still conserving a soft (possibly pasty) interior matrix. 5. Final remarks In summary, we suggest that in Cu-S2 the grinding tool was used for intentionally grinding mature, though hard maize kernels, whereas scrapers were used for peeling and releasing mature (possibly pasty) kernels that were heavily attached to maize cobs. These processing methods could be related to transforming maize into wrapped masses (jumint'a or tamal), or for preparing energetic or alcoholic beverages derived from doughy masses. It is evident that kernels were the main focus of processing and use of maize in Cubil
an, contrary to hypotheses suggesting sugary stalks were the main reason for maize domestication and its early dispersal (Smalley and Blake, 2003). Other cultural factors surrounding the preparation and consumption of food or beverages could govern the forms of intentional use of the maize documented here. The processing and manipulation of maize derivatives in what has been preliminary interpreted as a lithic workshop site indicates that this location constituted a multipurpose setting for performing varied, but culturally interrelated social activities. Although some evidence suggests that the earliest domestic maize was originally adapted to the humid Mexican lowlands (Piperno et al., 2009; van Heerwaarden et al., 2011), our findings fit various genetic-derived hypotheses which indicate that the highlands constituted fundamental settings for early maize domestication and pathways for its dispersal (Matsuoka et al., 2002; Lia et al., 2006). Maize kernels were present at the highland locality of Cu-S2 and it seems that they were processed when they were already mature or partially fresh. Having maize kernels at hand, it is not surprising that the inhabitants of Cu-S2 could produce maize plants experimentally (Smith, 2001; Pearsall, 2009) on-site when climate was warmer and drier by ca. 8000 cal. BP according to palaeoecological data (Jantz and Behling, 2012; Villota et al., 2012), or in nearby lower physiographic strata within a radius of ~5e13 km. Looking at the available information on early maize dispersals into South America at the regional level, an interesting setting comes into view. Recently gathered starch grain, pollen and phytolith data has suggested that humans dispersed maize into the upper lowland (920 masl) and transitional tropical/premontane wet forest of northwest Colombia at some time within a date range of 8997e8277 cal. BP (Santos Vecino et al., 2014). The terminal end of this date range could indicate the earliest introduction of maize into South America, though this new data is supported by only a single date that has been used for dating a soil horizon of 13e40 cm thickness at Primavera II. Later, human groups dispersed maize into the mid-elevation (1650 masl) and very humid premontane forests of the Colombian Andes by 8000e7600 cal. BP (Aceituno and Loaiza, 2014), based on maize starch grains recovered on lithic tools. A few decades before, maize reached the seasonally dry forests of the coastal lowlands of Ecuador (Las Vegas-OGSE-80) at ~8053e7818 cal. BP (Piperno, 2011) as phytolith data has pointed out. Overall, this seemingly contrasting information, together with new data presented here for highland locality Cu-S2, clearly suggests that distinct maize landraces formerly adapted to different altitudinal strata and climatic conditions could have entered South America and dispersed almost simultaneously from the northwest towards other continental regions in a overlapping time period between ~8300/ 8100e8000 cal. BP. The identification of starches from other plants adapted to mid- and low elevations on some of the studied tools allows us to infer that the inhabitants of Cu-S2 operated within a geographical wide-spectrum range of exploitation. It is foreseeable that the
n made use of some of the nearby diverse inhabitants of Cubila and contrasting physiographic environments to grow plants like maize, or to grow and extract other vegetal resources from these zones. However, we cannot completely rule out the possibility that the identified plants may have been acquired through exchange networks that extended across different altitudinal environments. The significant density of ancient archaeological localities in the Ecuadorian lowlands and highlands has recently
nchez, 2011; Constantine, 2013) been recognized (Stothert and Sa raising the possibility that the early peopling of the Ecuadorian territory followed active and preferred circuits for mobility and intensive human interaction that included the exchange of goods, products and ideas. This socio-spatial ingredient, possibly developed since the PleistoceneeHolocene transition in north
pez, 2007), could be the western South America (Ranere and Lo driving force that later triggered early long distance dispersion of cultivars and domestic plants betwixt and between Central and South America, which had, as one of its main corridors, the Ecuadorian highlands. Acknowledgments This research was supported by a grant from Proyecto Prom
n Superior, Ciencia, eteo of the Secretaría Nacional de Educacio
n (SENESCYT, Ecuador) to Pag

nez, Tecnología e Innovacio an-Jime and by another grant from SENESCYT to the Instituto Nacional de Patrimonio Cultural, Ecuador. Thank you is due to archaeologists ~ a, and to Jeff Walker, Reniel Rodríguez-Ramos, Carlos Solís-Magan botanist James D. Ackerman for editing our Spanglish of previous versions of this paper. Thanks also to Jacqueline Carrillo and
nchez Arias for making Figs. 1 and 3 respectively. We Fausto Sa want to thank the anonymous reviewers who evaluated this work and made punctual and useful recommendations. Finally, ~ a (especially to Omar Vallejo, thank you to the people of On
n Calle, Sebastia
n Espinoza, Pedro Capa, Patricio Rodríguez Fabia and Jenny Cobos) in Provincia del Azuay, for supporting this research.
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga 149 Appendix A Studied lithic tools. (A-A1), expedient tool 13-98, milling stone base fragment, chalcedony. (B-B1), formalized tool 13-99 “bec” type scraper, chalcedony. (C-C1), formalized tool 1401, scraper flake with a retouched and concave active border, chalcedony. (D), formalized tool 14-02, burin flake with scraper wear signs in its longitudinal borders, chalcedony/ chert. (E-E1), formalized tool 14-03, scraper with a concave active border, chert. (F), expedient tool 14-04, scraper flake with use-wear microflaking, chalcedony/chert. (G), expedient tool 14-05, scraper chip with use-wear microflaking, quartzite type rock. (H), formalized tool 14-06, scraper flake with a retouched and concave active border, and a podlid fracture (fire damaging), chert. (I-I1), formalized tool 14-07, end-scraper with convex active border, chert. (J), formalized tool 14-08, scraper with retouched and concave active border, chert. Scale lines for each tool represent 5 cm. 150
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga Appendix B
n, Instituto Nacional de Patrimonio Cultural [LIINPC] reference collection). (A), Blanco blandito Starch grains from 4 modern Ecuadorian maize landraces (Laboratorio de Investigacio redondeado e white and soft (floury) kernels. These starches are mostly regular (spherical to oval) with some few polygonal starches. Hilum is open and commonly visible, and fissures (usually “T” shape and transverse) are observed in a few cases. A prominent double-border is frequent. (B), Morochillo e white, vitreous hard kernels. Starches are mostly regular and the hilum is sometimes open, but is commonly similar to the wide and dark depression registered at the hilum area on starches submitted to the toasting of maize
n e yellow and soft (floury) kernels. Starches kernels. Fissures of “Y”, “T” and transverse-line shapes are very common. A prominent double-border is very frequent. (C), Mishca/Sapo are mainly regular (oval), though bumpy surface is usual on bigger oval or lightly irregular starch grains. Fissures are recurrent and transversal or restricted lineal fissures are the most frequent ones, but there are other fissure shapes such as “T”, “Y”, and radial or stellate. Different from other maize starches, a prominent double-border is not common. (D), cf. Tusilla e yellow-orange, vitreous hard kernels. Different from other modern Ecuadorian maize starches from mid-to high elevations, these are the only ones in our reference collections that are mainly irregular in shape (oval, strongly undulated, and quadrangular to hexagonal). Hilum is evident, but it is sometimes partially closed or very smooth. Transverse, radial and restricted-transverse fissures are very common. Like in other maize starches, a prominent double-border is evident. In sum, standard characteristics of above
n-Jime
nez, 2015) show that shapes are usually spherical to polygonal and display multiple modern maize starches and others extracted from 35 studied landraces or varieties (Paga
n-Jime
nez, 2012). The hilum is commonly open, slightly irregular, and most common in soft pressure facets according to matrix (endosperm) hardness (Mickleburgh and Paga endosperm maize kernels. Lamellae are generally absent, and fissures with different patterning (“Y”, “T”, transversal, and sometimes radial) are common, especially in starches with dried or hard kernels. One of the main diagnostic features defined for maize starch grains is the presence of a double-border. Size is slightly variable among the different maize landraces although a general range of 2.5e29.6 mm, with a mean size of 14.38 ± 4.4 mm for 39 indigenous South American landraces has been documented from our reference collections.
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga 151 Appendix C Starch grains from 4 modern bean (Phaseolus spp.) specimens (LIINPC, reference collection). (A), Phaseolus lunatus e brown seed variety. These starches are mostly oblong to oval, with kidney-shapes commonly present. Hilum is not observable though it is usually centric. General size range is from 8.35 to 49.02 mm with a mean size of 28.51 (±10.51). Lamellae consist in prominent layers of concentric undulated rings. Fissures are mostly longitudinal lines (“lf” in figure above), straight or undulated, that generally posses thinner parallel or lightly diagonal smaller striations. No pressure facets have been registered. (B), Phaseolus lunatus e dotted seed variety. These starches are mostly oblong to oval, with kidneyshapes commonly present. Compound starches are commonly present. Hilum is not observable though it is usually centric. General size range is from 9.76 to 50.24 mm with a mean size of 30.01(±9.24). Lamellae consist in prominent layers of concentric undulated rings. Fissures are mostly longitudinal lines (straight or undulated) that generally posses thinner parallel or lightly diagonal smaller striations. No pressure facets have been registered. (C), Phaseolus vulgaris e black seed variety. These starches are mostly oblong to oval, with kidney-shapes commonly present. Compound grains are usual. Hilum is not observable though it is usually centric. General size range is from 8.11 to 45.48 mm with a mean size of 24.43(±8.82). Lamellae, although smoother than previous specimens, consist of the same pattern of concentric undulated rings. Fissures are mostly longitudinal and straight lines. No pressure facets have been registered. (D), Phaseolus vulgaris e boca negra, black variety. These starches are mostly oval to oblong, with kidney-shapes commonly present. Compound starches are fairly common. Hilum is not observable though it is usually centric. General size range is from 12.27 to 26.93 mm with a mean size of 19.82 (±4.77). Lamellae consist of prominent layers of concentric undulated rings. Fissures are mostly irregular, less pronounced than in the other specimens described above, and sometimes they consist of radial lines combined with thinner asymmetric striations. No pressure facets have been registered. Archaeological starches tentatively or securely identified as Fabaceae or Phaseolus spp. at Cu-S2 coincides in shape, size, lamellae and sometimes in fissure characteristics with previously described modern starches. However, pressure damage on archaeological starches made it impossible to rotate them for exploring and validating identifications to a lower taxonomic level. Table 4 shows that size ranges of archaeological Fabaceae/ Phaseolus starches produced highly comparable metric data for proposing confident identifications at the three levels here considered. 152
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga Appendix D
n-Jime
nez (2007) reference collections). (A), Dioscorea piperifolia starch grains are mainly single with Starch grains from 4 modern yam (Dioscoreaceae) specimens (LIINPC, and Paga oval and trasovate shapes, and lightly undulated margins. Proximal section is wider than distal. Hilum is mostly closed and eccentric. General size range is from 3.6 to 64 mm with a mean size of 27.7 (±16.6). Concentric rings are the main feature of lamellae. Fissures are not present and Maltese cross is mainly an eccentric “X” shape with lightly wavy arms. A small single pressure facet is common at the distal section. (B), Dioscorea spp. #4 produces single and compound starches with triangular shapes and obtuse angles. They are wider in the proximal section. Hilum is closed, difficult to see, and eccentric. General size range is from 9 to 49 mm with a mean size of 24 (±9.5). A combination of concentric circles and angular rings are the main lamellae characteristic. Fissures are not present and Maltese cross is commonly an eccentric “X” shape with curved arms. A small single pressure facet is common at the distal section. (C), Dioscorea spp. #5 possesses single and compound starches with variable triangular shapes and less frequent trasovate and polymorphs grains made up by two or more granules. Proximal section is wider than the distal. Hilum is closed and eccentric. General size ranges is from 9 to 58 mm with a mean size of 22 (±10.8). Combined concentric circles and angular rings form the prominent lamellae on these starches. Fissures are not present and Maltese cross is mainly an “X” shape with wavy arms. At the distal section one or two small pressure facets are common. (D), Dioscorea altissima (syn. D. samydea and D. maranonensis in Ecuador and Perú) produces variable triangular shapes, notably wider at the distal section. Hilum is closed and eccentric at the narrower proximal section. General size range is from 15 to 75 mm with a mean size of 38 (±13). Concentric rings with wide obtuse angles at the observable end are the key feature of the lamellae. Curved-line fissures can be frequent between margin and hilum at the proximal section, while the Maltese cross has an “X” shape with three curved arms and one heavy wavy arm. No pressure facets were noted in these starches. Starches from D. altissima possess many of the main characteristics registered in one of the archaeological Dioscorea specimens identified at Cu-S2 (see Fig. 5A).
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga 153 Appendix E
n-Jime
nez (2007) reference collections). (A), starches from the rhizome of Calathea spp. (wild, S. Starch grains from two modern calathea (Calathea spp.) specimens (LIINPC, and Paga
chilas) are mostly triangular with some variants. A pronounced convex margin is common at the distal section with a narrower proximal section. Hilum is closed Domingo de los Tsa and eccentric, hardly observable. General size range is from 3.5 to 28.3 mm with a mean size of 13.6 (±7.02). Lamellae consist of regular and very smooth concentric rings. No fissures were noted and the Maltese cross is eccentric, mainly a cross shape with straight arms or wavy arms in a few occasions. No pressure facets were registered. (B), starches from the tubers of Calathea allouia (Puerto Rico). Tubers stores quite regular triangular starches with obtuse angles and lightly undulating margins. The proximal section is always narrower than the distal one when they are in centric position. Hilum is eccentric and sometimes open. General size range is from 8 to 40 mm with a mean size of 28 (±8.6). Lamellae begin with a single circle followed by regular concentric rings. Fissures will be common in the form of small and restricted transverse lines just over the hilum. Maltese cross with an eccentric “X” shape and straight arms is the most frequent, although centric cross shape with straight arms have also been noted. No pressure facets are present. Archaeological starch grain identified as Calathea spp. (see Fig. 5B) does not match the cultivar species starches of C. allouia. However, shape, lamellae, Maltese cross and size range are closer to the modern wild Calathea specimen described above. Improvement of this specific identification will require the acquisition of more wild Marantaceae and Calathea specimens from lowland and mid-elevations areas. Appendix F Modern starch grains of manioc (Manihot esculenta Crantz) from Ecuador (LIINPC reference collection). Among the diagnostic features of the starches stored in the tuberous roots of manioc, the bell-shape with two to four pressure facets is the most conspicuous. This starch type also possesses eccentric “Y” and stellate (or radial) fissures in the proximal section (see examples of “ds” or diagnostic starches above). Other common shapes registered for these starches are the oval with 5 or 6 marginal and restricted pressure facets, and a truncated-shape with a single pressure facet at the distal section, together with a wide “V” flexion line that departs from the hilum and ends at the distal section (see arrows above). General size range for all these starches is from 6.7 to 37.3 mm with a mean size of 17.5 (±8). Hilum will be either open or closed and is usually found at centric or lightly eccentric position. Maltese cross is mostly a centric cross shape with straight arms, although eccentric “X” or cross shapes with wavy arms are present. Regarding the truncate starch grain tentatively identified as manioc in this study (Fig. 5C), it should be stated that other plants such as sweet potato (Ipomoea batatas), cocoyam (Xanthosoma spp.), and some Andean tubers and tuberous roots such Ullucus tuberosus (pinky variety), Tropaeolum tuberosum, Hypochaeris sessiliflora, Pachyrhizus tuberosus and Dioscorea piperifolia also produces this type of starch. However, “V” fissure or “V” flexion line documented by us on the recovered starch is only present in sweet potato and manioc. If this feature is seen together with the central and depressed hilum here registered on the recovered starch, then we need to say that this combination of features has only been documented by us in manioc. Moreover, sweet potato produces very low quantities of this starch type when compared to manioc, though among manioc they fall between 17 and 20 mm and lamellae is not obvious, while in sweet potato they are not bigger than 14 mm and lamellae is prominent. In sum, tentative identification of manioc in this article is based on the recovery of a single starch grain almost identical to the truncated-shape specimen described above and because it fits the mean size described for the modern species starches (17.5 mm, Table 4). 154
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga Appendix G
n-Jime
nez (2007) reference collections). (AeB), Starches from archaeological Indian potato (cf. Sagittaria spp.), and modern Sagittaria latifolia, and S. lancifolia (LIINPC, and Paga archaeological starch grains from cf. Sagittaria spp. (Cu-S2, artifact 13-99). (C), modern starches of Sagittaria latifolia (Pastaza, Ecuador). (D), modern starches of Sagittaria lancifolia (Puerto Rico). Starches from the rhizome of Sagittaria latifolia (wild, Pastaza, Ecuador) are mostly oval with undulated margins, although trasovate shapes with a rounded distal end is less common. Hilum is typically closed and eccentric, or quite centric on a few occasions. General size range is from 4.6 to 26.8 mm with a mean size of 14.8 (±5.8). Lamellae are smooth and consist of regular concentric circles. Fissures are very few and the most frequent is a small transverse fissure over the hilum. Other fissures, much less common, has an “H” or “T” shape at the proximal section. Maltese cross is mainly an eccentric cross shape with lightly wavy arms. However, it is recurrent with an eccentric “X” shape also with wavy arms. Smooth, small and marginal single pressure facets were registered in just a few cases. Archaeological starch grains tentatively identified as Sagittaria spp. does not match
nez, 2015) but they possess general features that coincide closely with S. latifolia starches previously diagnostic and bigger starch grains documented for S. lancifolia (Pag
an-Jime described. Both archaeological starches are oval to transovate; they fall within the range size of S. latifolia (Table 4) and mean size is statistically the same. Maltese cross in archaeological starches is an “X” shape with wavy arms, and the proximal section is lightly narrower than the distal one, similar to many starch grains of S. latifolia. Both archaeological starches show signs of damaging by pressure and their surfaces are partially rough, different from the unaffected starches of S. latifolia here studied. So, this final observation was the main issue that hindered the ascription of theses starches to a secure taxonomic level. References Aceituno, J., Loaiza, N., 2014. Early and Middle Holocene evidence for plant use and cultivation in the Middle Cauca River Basin, Cordillera Central (Colombia). Quaternary Science Reviews 86, 49e62. Babot, M. del P., 2003. Starch grain damage as an indicator of food processing. In: Hart, D.M., Wallis, L.A. (Eds.), Phytolith and Starch Research in the AustralianPacific-Asian Regions: the State of the Art. Pandanus Books, Canberra, pp. 69e81. Berman, M.J., Pearsall, D.M., 2008. At the crossroads: starch grain and phytolith analyses in Lucayan prehistory. Latin American Antiquity 19, 182e204. Constantine, A., 2013. The early settlement of continental Ecuador: new evidence from preceramic sites in the tropical rain forest. Quaternary International 317, 112e117. Dickau, R., Bruno, M.C., Iriarte, J., Prümers, H., Jaimes, C., Holst, I., Mayle, F.E., 2012. Diversity of cultivars and other plant resources used at habitation sites in the Llanos de Mojos, Beni, Bolivia: evidence from macrobotanical remains, starch grains, and phytoliths. Journal of Archaeological Science 39, 357e370. Dickau, R., Ranere, A.J., Cooke, R.G., 2007. Starch grain evidence for the preceramic dispersals of maize and root crops into tropical dry and humid forests of Panama. Proceedings of the National Academy of Science of the United States of America 104, 3651e3656. Dorsey, S., Perry, L., Reinhard, K.J., Santoro, C.M., Teixeira-Santos, I., 2009. Impact of empire expansion on household diet: the Inka in northern Chile's Atacama desert. Plos One 4, 1e5.
gico. Ministerio de Agricultura, Ganaderia, Ecuador República del, 2011. Mapa Ecolo Acuacultura y Pesca, Quito. Grobman, A., Bonavia, D., Dillehay, T.D., Piperno, D.R., Holst, I., 2011. Preceramic maize from Paredones and Huaca Prieta, Perú. Proceedings of the National Academy of Science of the United States of America 109, 1755e1759. Henry, A.G., Hudson, H.F., Piperno, D.R., 2009. Changes in starch grain morphologies from cooking. Journal of Archaeological Science 36, 915e922. Holst, I., Moreno, J.E., Piperno, D.R., 2007. Identification of teosinte, maize, and tripsacum in Mesoamerica by using pollen, starch grains, and phytoliths. Proceedings of the National Academy of Science of the United States of America 104, 17608e17613. Horrocks, M., Rechtman, R.B., 2009. Sweet potato (Ipomoea batatas) and bannana (Musa sp.) microfossils in deposits from the Kona Field System, Island of Hawaii. Journal of Archaeological Science 36, 1115e1126. Jantz, N., Behling, H., 2012. A Holocene environmental record reflecting vegetation,
ramo of Quimsacocha, southwestern climate, and fire variability at the Pa Ecuadorian Andes. Vegetation History and Archaeobotany 21, 169e185. Lamb, J., Loy, T., 2005. Seeing red: the use of Congo Red dye to identify cooked and damaged starch grains in archaeological residues. Journal of Archaeological Science 32, 1433e1440.
mara, J.A., Miante Alzogaray, A.M., Poggio, L., Lia, V.V., Confalonieri, V., Ratto, N., Ca Brown, T.A., 2006. Microsatellite typing of ancient maize: insights into the history of agriculture in southern South America. Proceedings of the Royal Society Biological Sciences 274, 545e554.
gico del Ecuador. CODIGEM, Litherland, M., Eguez, A., Zamora, A., 1993. Mapa Geolo BGS, Quito. Matsuoka, Y., Vigouroux, Y., Goodman, M.M., Sanchez, G.,J., Buckler, E., Doebley, J., 2002. A single domestication for maize shown by multilocus microsatellite genotyping. Proceedings of the National Academy of Science of the United States of America 99, 6080e6084.
n-Jim
enez et al. / Quaternary International 404 (2016) 137e155 J.R. Paga
n-Jime
nez, J.R., 2012. New insights into the consumption of Mickleburgh, H.L., Paga maize and other food plants in the pre-Columbian Caribbean from starch grains trapped in human dental calculus. Journal of Archaeological Science 39, 2468e2478. Musaubach, M.G., Plos, A., Babot, M. del P., 2013. Differentiation of archaeological maize (Zea mays L.) from native wild grasses based on starch grain morphology. Cases from the Central Pampas of Argentina. Journal of Archaeological Science 40, 1186e1193.
nez, J.R., 2007. Paris Monographs in American Archaeology 18, BAR InPag
an-Jime
nicas en el Puerto Rico ternational Series. De Antiguos Pueblos y Culturas Bota
lago Borincano y la Llegada de los Primeros Pobladores Indígena. El Archipie Agroceramistas. Archaeopress, Oxford.
nez, J.R., 2012. Early use of maize and other food crops among Early Pag
an-Jime Ceramic and later Neoindian traditions of northeastern Amazonia revealed by ancient starch grains from ceramic and lithic artifacts of the Chemin Saint-Louis archaeological site, French Guiana. Archaeology and Anthropology 17, 78e107.
nez, J.R., 2015. Almidones. Guía de Material Comparativo Moderno del Pag
an-Jime
pico. Aspha anicos en el Neotro Ecuador para los Estudios Paleoetnobot
Ediciones, Buenos Aires. http://dx.doi.org/10.13140/RG.2.1.4020.0803.
nez, J.R., Rostain, S., 2014. Uso de plantas econo
micas y rituales en dos Pag
an-Jime comunidades precolombinas de la Alta Amazonia ecuatoriana: Sangay (Huapula) y Colina Moravia (c. 400 a.C. e 1200 d.C.). In: Rostain, S. (Ed.), Proceedings
s of the Third International Meeting of Amazonian Archaeology. Instituto France de Estudios Andinos. Instituto Nacional de Patrimonio Cultural, Quito, pp. 313e322. Pearsall, D.M., 2008. Plant domestication and the shift to agriculture in the Andes. In: Silverman, H., Isbell, W. (Eds.), The Handbook of South American Archaeology. Springer, New York, pp. 105e120. Pearsall, D.M., 2009. Investigating the transition to agriculture. Current Anthropology 50, 609e613. Pearsall, D.M., Chandler-Ezell, K., Zeidler, J.A., 2004. Maize in ancient Ecuador: results of residue analysis of stone tools from the Real Alto site. Journal of Archaeological Science 31, 423e442. Perry, L., 2004. Starch granule size and the domestication of manioc (Manihot esculenta) and sweet potato (Ipomoea batatas). Economic Botany 56, 335e349. Perry, L., Dickau, R., Zarrillo, S., Holst, I., Pearsall, D.M., Piperno, D.R., Berman, M.J., Cooke, R.G., Rademaker, K., Ranere, A.J., Raymond, J.S., Sandweiss, D.H., Scaramelli, F., 2007. Starch fossils and the domestication and dispersal of chili peppers (Capsicum spp. L.) in the Americas. Science 315, 986e988. Perry, L., Sandweiss, D.H., Piperno, D.R., Rademaker, K., Malpass, M.A., Umire, A., de la Vera, P., 2006. Early maize agriculture and interzonal interaction in southern Perú. Nature 440, 76e79. Piperno, D.R., 2011. The origins of plant cultivation and domestication in the new world tropics. Current Anthropology 52, S453eS470. Piperno, D.R., Dillehay, T.D., 2008. Starch grains on human teeth reveal early broad crop diet in northern Peru. Proceedings of the National Academy of Science of the United States of America 105, 19622e19627. Piperno, D.R., Holst, I., 1998. The presence of starch grains on prehistoric stone tools from the humid neotropics: indications of early tuber use and agriculture in Panama. Journal of Archaeological Science 25, 765e776. Piperno, D.R., Pearsall, D.M., 1998. The Origins of Agriculture in the Lowland Neotropics. Academic Press, San Diego. Piperno, D.R., Ranere, A., Holst, I., Hansell, P., 2000. Starch grains reveal early root crop horticulture in the Panamanian tropical forest. Nature 407, 894e897. 155 Piperno, D.R., Ranere, A.J., Holst, I., Iriarte, J., Dickau, R., 2009. Starch grain and phytolith evidence for early ninth millennium B.P. maize from the Central Balsas River Valley, Mexico. Proceedings of the National Academy of Science of the United States of America 106, 5019e5024. Piperno, D.R., Stothert, K.E., 2003. Phytolith evidence for early Holocene Cucurbita domestication in southwest Ecuador. Science 299, 1054e1057. Piperno, D.R., Weiss, E., Holst, I., Nadel, D., 2004. Processing of wild cereal grains in the Upper Palaeolithic revealed by starch grain analysis. Nature 430, 670e673.
pez, C., 2007. Cultural diversity in Late Pleistocene/Early Holocene Ranere, A.J., Lo populations in northwest South America and lower Central America. International Journal of South American Archaeology 1, 25e31. Ranere, A.J., Piperno, D.R., Holst, I., Dickau, R., Iriarte, J., 2009. The cultural and chronological context of early Holocene maize and squash domestication in the Central Balsas River Valley, Mexico. Proceedings of the National Academy of Science of the United States of America 106, 5014e5018. Reichert, E.T., 1913. The Differentiation and Specificity of Starches in Relation to Genera, Species, etc. Carnegie Institution of Washington, Washington D.C. Santos Vecino, G., Monsalve Marín, C.A., Correa Salas, L.V., 2014. Alteration of tropical forest vegetation from the PleistoceneeHolocene transition and plant cultivation from the end of early Holocene through middle Holocene in Northwest Colombia. Quaternary International 363, 28e42. Smalley, J., Blake, M., 2003. Sweet beginnings: stalk sugar and the domestication of maize. Current Anthropology 44, 675e703. Smith, A.M., Denyer, K., Martin, C.R., 1995. What controls the amount and structure of starch in storage organs? Plant Physiology 107, 673e677. Smith, B.D., 2001. Low-level food production. Journal of Archaeological Research 9, 1e43. Stothert, K.E., Piperno, D.R., Andres, C.T., 2003. Terminal Pleistocene/Early Holocene human adaptation in coastal Ecuador: the Las Vegas evidence. Quaternary International 109e110, 23e43.
nchez, A., 2011. Culturas del Pleistoceno final y el Holoceno temStothert, K.E., Sa prano en el Ecuador. Boletín de Arqueología de la Pontificia Universidad
lica del Perú 15, 81e119. Cato
n: dos estaciones precer
Temme, M., 2005. Cubila amicas en el curso superior del río ~ a, Provincias de Loja y Azuay, Ecuador. Universidad Te
cnica San Felipe de On Particular de Loja, Loja. Van Heerwaarden, J., Doebley, J., Briggs, W.H., Glaubitz, J.C., Goodman, M.M., de Jesus Sanchez Gonzalez, J., Ross-Ibarra, J., 2011. Genetic signals of origin, spread, and introgression in a large sample of maize landraces. Proceedings of the National Academy of Science of the United States of America 108, 1088e1092.
n-Y
Villota, A., Leo anez, S., Behling, H., 2012. Vegetation and environmental dy
ramo of Jimbura region in the southeastern Ecuadorian Andes namics in the Pa during the late Quaternary. Journal of South American Earth Sciences 40, 85e93. Zarrillo, S., 2012. Human Adaptation, Food Production, and Cultural Interaction during the Formative Period in Highland Ecuador. PhD Dissertation. Department of Archaeology, University of Calgary, Calgary. Zarrillo, S., Pearsall, D.M., Raymond, J.S., Tisdale, M.A., Quon, D.J., 2008. Directly dated starch residues document early formative maize (Zea mays L.) in tropical Ecuador. Proceedings of the National Academy of Science of the United States of America 105, 5006e5011. Zeidler, J.A., 2008. The ecuadorian formative. In: Silverman, H., Isbell, W.H. (Eds.), Handbook of South American Archaeology. Springer, New York, pp. 459e488.