sign in sign up
<<
Home , Great Basin, Mesoamerica

PERSPECTIVE

The diffusion of to the southwestern and its impact

William L. Merrilla, Robert J. Hardb,1, Jonathan B. Mabryc, Gayle J. Fritzd, Karen R. Adamse, John R. Roneyf, and A. C. MacWilliamsg aDepartment of Anthropology, National Museum of Natural History, Smithsonian Institution, P.O. Box 37102, , DC 20013-7012; bDepartment of Anthropology, One UTSA Circle, University of at San Antonio, San Antonio, TX 78249; cHistoric Preservation Office, of Tucson, P.O. Box 27210, Tucson, AZ 85726; dDepartment of Anthropology, Campus Box 1114, One Brookings Drive, Washington University, St. Louis, MO 63130; eCrow Canyon Archaeological Center, 23390 Road K, Cortez, CO 81321; fColinas Cultural Resource Consulting, 6100 4th Street, Private Mailbox #300, Albuquerque, NM 87107; and gDepartment of , 2500 University Drive Northwest, University of Calgary, Calgary, Alberta, T2N 1N4

Edited by Linda S. Cordell, University of , Boulder, CO, and approved October 30, 2009 (received for review June 22, 2009)

Our understanding of the initial period of in the southwestern United States has been transformed by recent discoveries that establish the presence of maize there by 2100 cal. B.C. (calibrated calendrical years before the Christian era) and document the processes by which it was integrated into local foraging economies. Here we review archaeological, paleoecological, linguistic, and genetic data to evaluate the hypothesis that Proto-Uto-Aztecan (PUA) farmers migrating from a homeland in intro- duced maize agriculture to the . We conclude that this hypothesis is untenable and that the available data indicate instead a homeland for the PUA, the breakup of this speech community into northern and southern divisions Ϸ6900 cal. B.C. and the dispersal of maize agriculture from Mesoamerica to the US Southwest via group-to-group diffusion across a Southern Uto- Aztecan linguistic continuum.

early agriculture ͉ migration ͉ US Southwest ͉ Mesoamerica ͉ Uto-Aztecan

he dispersal of maize agricul- led to the diversification of PUA into dis- maize ( mays spp. mays) was intro- ture from Mesoamerica to the tinct dialects and eventually the distinct duced to the southwestern United States southwestern United States is a that comprise the Uto-Aztecan no later than 2100 cal. B.C. (calibrated perennial topic of renewed in- (UA) family, distributed at the calendrical years before the Christian era) Tterest as part of a broad, interdiscipli- of European contact from the Great (Table 1 and Table S2) (25, 26). The ar- nary effort to understand the develop- Basin of western to Cen- rival of maize in this region was preceded ment of maize agriculture in the New tral America (see Table S1) (15–17). by a long period during which its wild an- World (1–3). At present, there are two The credibility of Bellwood’s proposal cestor, Balsas teosinte (Zea mays ssp. par- competing models of the mechanisms was bolstered by the strong endorsement viglumis), native to southwestern , of the Uto-Aztecanist linguist Jane Hill, through which this dispersal took place. was transformed through interven- who in a series of publications presented The first proposes that maize agriculture linguistic, cultural, and human genetic tion into a productive grain crop (3, 27– diffused northward as a complex com- data that she interpreted as supporting his 29). The oldest known maize macrofossils, prised of the cultigen itself and the perspective (9, 18–21). Largely as a result recovered from Guila´Naquitz Cave in the knowledge associated with its cultivation of her work, the Bellwood–Hill hypothesis southern Mexican of , date to and use, transmitted from one foraging is now regarded as offering a plausible 4280 cal. B.C. (30, 31). For the purposes band to another without major popula- scenario for the origins of maize agricul- of this essay, we will assume that the ini- tion movements (4–6). The second ar- ture in the US Southwest (22–24), but it tial dispersal of maize from Mesoamerica gues that maize agriculture was intro- has not yet received the critical evaluation to the southwestern United States began duced to the southwestern United States that the evidence currently available al- Ϸ4300 cal. B.C. through the long-distance migration of lows. Here we review this evidence, con- The earliest evidence for maize in the Mesoamerican farmers (7–9). cluding that it strongly supports the alter- southwestern United States comes from Over the years, most scholars have con- native view that maize agriculture spread three open sites and two rockshelters in sidered group-to-group diffusion more from Mesoamerica to the Southwest via and western (Table likely, but since 2000 the long-distance group-to-group diffusion. We propose that 1, Fig. 1) (32–37). Ranging in elevation migration model has gained adherents Uto-Aztecans played a crucial role in this from Ϸ700 to 2200 m, these sites are no- due primarily to perspectives developed by diffusion but more than a millennium af- table for the diversity of their settings. the archaeologist Peter Bellwood regard- ter the PUA speech community began Clearwater and Las Capas are located ing the expansion of agriculture and lan- dividing into subgroups and migrating out along the middle Santa Cruz in the guage families in diverse of the of the homeland that its members shared world (8, 10–14). Bellwood suggested that at that point in their history. We offer new data to argue that, at the time of the maize agriculture expanded from its area Author contributions: W.L.M., R.J.H., J.B.M., G.J.F., K.R.A., of origin in Mesoamerica to the US initial breakup of this speech community, J.R.R., and A.C.M. designed research; W.L.M., R.J.H., J.B.M., its members were foragers located in the G.J.F., K.R.A., J.R.R., and A.C.M. performed research; Southwest as a consequence of the migra- Great Basin rather than farmers in Me- W.L.M., R.J.H., J.B.M., G.J.F., K.R.A., J.R.R., and A.C.M. ana- tion of farmers who were members of the soamerica and that this breakup began lyzed data; and W.L.M. and R.J.H. wrote the paper. Proto-Uto-Aztecan (PUA) speech com- Ϸ6900 B.C., significantly earlier than pre- The authors declare no conflict of interest. munity. He proposed that this migration viously proposed. This article is a PNAS Direct Submission. Ϸ began 3500 B.C. and involved the 1To whom correspondence should be addressed. Email: spread of PUA speakers across the region The Introduction of Maize Agriculture to the [email protected]. extending between Mesoamerica and the Southwestern United States. Multiple AMS This article contains supporting information online at www. US Southwest. This population expansion radiocarbon dates firmly establish that pnas.org/cgi/content/full/0906075106/DCSupplemental.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0906075106 PNAS ͉ December 15, 2009 ͉ vol. 106 ͉ no. 50 ͉ 21019–21026 Downloaded by guest on September 26, 2021 Table 1. The earliest maize sites in the US Southwest Calibrated date (2 ␴ and Site Location Elevation Type median) (refs. 25, 26)

Old Corn , western New Mexico 34°30’ 1,910m Open 2460–2060 B.C. (2260 B.C.) McEuen Cave Gila Mountains, eastern Arizona 33°14’ 1,340m Rockshelter 2270–1940 B.C. (2080 B.C.) Clearwater Tucson Basin, southern Arizona 32°13’ 720m Open 2200–1960 B.C. (2080 B.C.) Las Capas Tucson Basin, southern Arizona 32°20’ 680m Open 2200–1940 B.C. (2050 B.C.) Three Fir Shelter Colorado Plateau, northeastern Arizona 36°35’ 2,240m Rockshelter 2470–1540 B.C. (1990 B.C.)

The calendrical dates are calibrated to the maximum 2 ␴ range with the median (25, 26). Uncalibrated radiocarbon dates and references are in Table S2.

basin and range country of the upper So- The near contemporaneity of the appear- aged Ϸ1.3 km per year, using the noran . McEuen Cave is found in a ance of maize in and the US Ϸ2,800-km highway distance between deep canyon in the eastern Gila River Southwest, as well as the genetic differ- and Tucson as a proxy for valley, in the mountainous transition zone ences between the early maize recovered the actual distance covered. At this pace, between the southern basin and range in the two areas, suggest that maize did it would have proceeded an average of country and the Colorado Plateau. The not diffuse to the Southwest through one degree of latitude (120 km) every 92 Old Corn site is in an upland valley on northeastern Mexico (40). Mexico’s Pa- years, allowing adaptations to local condi- the southeastern edge of the Colorado cific coast, the , tions to take place over numerous genera- Plateau, and the Three Fir Shelter is situ- and the eastern and western flanks of this tions of maize. ated on the central Colorado Plateau, cordillera have been proposed as likely near the north rim of Mesa. The routes (41, 42), but the absence of addi- The Proto-Uto-Aztecans as Mesoamerican fact that maize appears almost simulta- tional data precludes identifying both the Maize Farmers. Bellwood proposed that neously in these upland and lowland sites point(s) of origin of the early southwest- migrating Mesoamerican farming popula- suggests that the ancient landrace(s) were ern maize and the route(s) of its tions could have diffused maize agricul- tolerant of a wide range of environmental dispersal. ture northward because the distribution of settings. Given the presumed adaptability of Uto-Aztecan languages between Me- Fifteen radiocarbon dates on maize early maize, any or all of these routes are soamerica and the US Southwest con- macrofossils recovered from these five possibilities. Even if adaptations to the formed to his assumption that the spread sites cluster Ϸ2100 cal. B.C. (Table S2). diverse ecological settings represented of language families over such great dis- The earliest dates for maize at Old along these alternative routes were re- tances often was the result of the expan- Corn, McEuen Cave, and Clearwater quired, ample time would have been avail- sion of agriculture (8, 11). Hill developed are consistent with dates derived from able for them. If we assume that maize this proposal into a more detailed sce- other associated materials at these sites, diffused northward over the course of nario (9, 18–21). She argued that during as well as those for the oldest maize re- Ϸ2,200 years, between 4300 and 2100 cal the early period of maize agriculture, the covered from Las Capas and Three Fir B.C., its rate of diffusion would have aver- PUA speech community most likely was Shelter. Two other sites, Los Pozos in the Tucson Basin and Bat Cave in the Mogollon , also have yielded maize with very early dates, but the reli- ability of these dates has been ques- tioned (4, 33). The presence of maize in about half of the features at the Old Corn site, including roasting and storage pits, and the existence of a few pre-2100 cal. B.C. maize dates from Old Corn and McEuen Cave (34, 35) raises the possibility that maize arrived in the southwestern United States before 2100 cal. B.C. The paucity of archaeological data for the area between Mesoamerica and the US Southwest for the 4300–2100 cal. B.C. time period renders any scenario for the northward diffusion of maize agriculture highly speculative. The only early maize sites that have been excavated in these are the Ocampo Caves, located in the highlands of the northeastern Mexi- can state of Tamaulipas, where the oldest maize dates to Ϸ2400 cal. B.C. (38, 39). Fig. 1. Principal early maize sites and proposed Proto-Uto-Aztecan homeland before breakup.

21020 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0906075106 Merrill et al. Downloaded by guest on September 26, 2021 located on the northwestern edge of Me- to attest to the phonologically expected soamerican populations. Haplogroup A soamerica, in the region of the modern form, raising in the latter case the pros- also is present in all SUA populations, Mexican states of , , Du- pect that they are loan words from SUA which display a cline in frequency from rango, , and Aguascalientes, and or other languages. Hill attributes the ab- low (5–7%) among the Upper Piman that at least some segments of this com- sence of a corpus of unequivocal NUA Akimel O’odham and Tohono O’odham munity of maize farmers migrated north- cognates in part to gaps in our knowledge in the US Southwest, to intermediate (30– ward to gain access to new arable lands. of many NUA languages and SUA lan- 34%) among the Tarahumara, Cora, and Following Bellwood, she identified the guages as well (9). However, in the case of in northwestern and central principal factor motivating this migration Hopi, a well-documented NUA language Mexico and to intermediate or high (38– as population pressure resulting from associated with the NUA tradition- 63%) among different Nahua populations rapid population growth that she assumed ally most reliant on maize agriculture, not in central Mexico (44, 47–49). had followed their shift from foragers to only would we expect a larger of LeBlanc suggests that this cline could farmers. She suggested that the presence cognates but the meanings associated with reflect the dilution of haplogroup A in of other early farming on the these cognates should also resemble more northward-migrating UA farming popula- east and of the PUA homeland closely those of the proposed SUA tions as the result of intermarriage with blocked PUA expansion in these direc- cognates. ‘‘non-A-bearing forager women’’ (24). tions, so the PUA maize farmers migrated With regard to the PUA homeland, However, comparable results would be northward, displacing or assimilating the neither Bellwood nor Hill offer data that expected if SUA populations with a low local populations of foragers they encoun- would sustain a compelling argument for incidence of haplogroup A had migrated tered. The vanguard of this northward the location of the PUA speech commu- southward from the migration was comprised of the ancestors nity in Mesoamerica. Hill implied that and intermarried with women from of speakers of what would become the such data exist, for example, in the form non-UA populations with higher haplo- northern branch of the UA language fam- of ‘‘evidence for the presence of Uto- group A frequencies. In addition, the fre- ily. In Hill’s scenario, these Proto- Aztecan cultural practices in Mesoamerica quency distributions of the other haplo- Northern Uto-Aztecans were farmers at an early date’’, but this evidence relates groups (B, C, and D) determined for when they departed Mesoamerica and only to her claim that ‘‘Aztecan people NUA populations are more similar to arrived in the Southwest, but the majority were important in Mesoamerica through those of neighboring, non-UA populations later shifted back to foraging because the the Classic period’’, a position that refers in the western United States than they are wealth of wild resources in the areas to only one UA subdivision millennia af- to those of any SUA populations, just as where they lived made foraging a more ter maize agriculture diffused out of Me- the Nahua haplogroup frequency distribu- productive strategy than agriculture or soamerica (9). She also proposed that cer- tions most closely resemble those of other because the environmental conditions in tain genetic evidence supports a Mesoamerican populations (47, 48, 50). these areas precluded agriculture. Mesoamerican homeland for the PUA, This pattern is best explained by millennia The viability of the Bellwood–Hill hy- stating that ‘‘members of both northern of gene flow across adjacent cultural and pothesis hinges on establishing that the and southern branches of Uto-Aztecan. . . linguistic boundaries, a conclusion rein- members of the PUA speech community seem to carry a clear Mesoamerican sig- forced by the fact that detailed haplotype cultivated maize before the breakup of nature in the form of a rare mutation at analyses have failed to detect evidence of this community and that they were lo- the Albumin locus, ‘Albumin Mexico’’’ a UA migration from Mesoamerica to the cated in Mesoamerica, where maize agri- (19). However, Albumin Mexico Southwest (44, 50). culture initially developed. Hill makes the (AL*Mexico) is extremely rare among The mtDNA data indicate marked dif- case that the PUA were maize farmers NUA populations while showing a high ferences in female lineages between the primarily on the basis of her claim that level of incidence among individuals affili- US Southwest and Mesoamerica, but, as terms related to maize cultivation can be ated with certain Yuman-speaking societ- Hill and others have argued (19, 44), they reconstructed to the PUA period (9, 19– ies in or near the southwestern United would not preclude a PUA homeland in 21). She identified twelve terms as com- States and also the SUA language group Mesoamerica if only UA males had mi- prising ‘‘the Uto-Aztecan maize complex’’, known as Tepiman, which includes the grated northward. A 2008 study of the compiling in each case sets of terms from Tepehuan and Piman languages (22, 44). demographic impact of the introduction modern UA languages that she regarded This pattern suggests that the AL*Mexico of maize agriculture to the US South- as being related to one another and upon mutation occurred first among Yumans, west reports that recent research on which a reconstructed PUA form could who transferred it to Tepiman speakers Y-chromosome variation ‘‘provides no be proposed. She characterized these during a period of interaction at some support for close genetic connections be- twelve sets as ‘‘cognate sets,’’ that is, sets point during the first or early second mil- tween southwestern and Mesoamerican of terms that derive from the same proto- lennium A.D. (45). The mutation would Uto-Aztecan populations’’ (51), but the form, but in her analysis she described then have been diffused southward by associated data are not presented. An many of these terms as potential, rather Tepehuan speakers who, at the time of evaluation of this male-only migration than confirmed, cognates or as phonologi- European contact, were distributed from hypothesis requires data on the distribu- cally similar forms, i.e., ‘‘resemblants,’’ to the northern margins tion of Y-chromosome haplotypes among rather than cognates (9, 20). of Mesoamerica (46) and from there to both NUA and SUA populations as well In the final analysis, Hill presents no other Mesoamerican populations. as populations affiliated with the various Northern Uto-Aztecan (NUA) terms that It should be noted that the available non-UA language groups in both Me- (i) are unquestionably cognate with mitochondrial DNA (mtDNA) evidence is soamerica and the US Southwest. In the maize-related terms in Southern Uto- not consistent with the proposal of a Me- absence of these data, we simply note that Aztecan (SUA) languages and (ii) also soamerican homeland for the PUA. Mod- a 2,000-km migration of male farmers share comparable maize-related meanings ern mtDNA data from individuals affili- from Mesoamerica to the US Southwest with them (43). Her proposed NUA cog- ated with NUA societies in the US seems improbable, as is the prospect that nates either show the phonologically ex- Southwest, the Great Basin, and Califor- small groups of UA males would have not pected form but lack the expected mean- nia indicate the absence of haplogroup A, been absorbed into local, non-UA speech ing or have the expected meaning but fail which occurs with high frequency in Me- communities along the way.

Merrill et al. PNAS ͉ December 15, 2009 ͉ vol. 106 ͉ no. 50 ͉ 21021 Downloaded by guest on September 26, 2021 A Great Basin Homeland for the Uto-Aztecans. (mainly , P. edulis, and P. distance migration would be marked dis- The Bellwood–Hill proposal of a Me- cembroides) in the PUA homeland. Before continuities in the archaeological record soamerican homeland for the Uto-Aztec- expanding across the Great Basin during of the southwestern United States be- ans challenges the longstanding view that the middle , pinyon populations tween the pre- and postmaize periods. A the PUA speech community was based reached the northern limits of their distri- potential example of such a discontinuity somewhere in the western United States bution along the Basin’s southern and is the appearance of ceramics in early or northern Mexico at the time of its ini- southwestern edges, extending from there maize sites in the Tucson Basin, in the tial breakup (9, 52). Fowler identified this to southern Mexico, where pine nuts have form of low-fired fragments of human broad area as the most likely PUA terri- been recovered from archaeological con- figurines and what appear to be small tory based on reconstructions of PUA texts before 7000 cal. B.C. (54–61). If the serving vessels. Dated to Ϸ2100 cal. B.C., terms for plants and animals and the as- PUA homeland had been located within they provide the earliest record of fired sumption that the biological taxa labeled this area, a PUA term for ‘‘pinyon’’ would ceramics in the US Southwest (2), but a with related terms, or cognates, in differ- be expected. Mesoamerican origin is doubtful. The ent, and especially the more widely sepa- We propose that the PUA speech com- oldest known human figurine in Me- rated, UA languages would have been munity before its breakup was located soamerica, recovered in the Valley of present in this ancestral homeland (52). north of the northern extent of the distri- Mexico and assigned an earlier date of She offered likely PUA forms for a vari- bution of pinyon at the time and also east 2300 B.C. (69, 70), was fashioned from a ety of plants (long-needled pine, prickly of the Sierra range, probably in single piece of clay whereas those of the pear, , cane, and a kind of grass) the west central Great Basin (Fig. 1). The Tucson Basin were created from multiple and animals (fly, , louse, flea, , determination of these latter limits is pieces of clay assembled in an applique´ turtle/tortoise, a kind of , small bird, based on the absence of a PUA term for approach (71). Ceramic vessels appear in screech , horned owl, , hawk, ‘‘oak’’ and the presence of oaks to the far southwestern Mexico Ϸ2000 cal. B.C., heron, sandhill crane, , and a kind west, northwest, east, and south of this about the same time as in the Tucson Ba- of woodpecker), noting that these taxa portion of the Great Basin (52, 53, 62– sin, but they are quite distinct in form and would be found ‘‘in a mixed woodland/ 64). There are no other areas that corre- size from the Tucson Basin examples and grassland setting, in proximity to montane spond to the ecological zone identified by also were fired at higher temperatures (72, forests’’, but not in hot desert or tropical Fowler and accepted by Hill for the PUA 73). In any case, ceramics are absent from settings. She concluded that the south- homeland in which the taxa for which all other known early farming sites in the western United States and northwestern PUA terms can be reconstructed were US Southwest outside of the Tucson Mexico corresponded to the area where present and pinyons and oaks were Basin, suggesting that their production this particular ensemble of taxa was most absent. was a highly localized, not imported, likely to be found Ϸ3000 B.C., the date This set of circumstances existed in the development. she adopted for when the PUA speech Great Basin during the early Holocene, It has been suggested that Gypsum- community was intact but dispersed in a between 9700–6900 cal. B.C. [10,000– style, contracting stem points were intro- dialect chain across Ϸ1,500 km of this 8,000 radiocarbon years before the present duced to the US Southwest from Me- region (52). (14C yr b.p.)], a period during which the soamerica, where similar points are found In support of the alternative view of a PUA speech community probably was in association with maize during the Abe- Mesoamerican homeland for the Uto- intact. We attribute its breakup to a sig- jas phase (4200–2900 cal. B.C.) in the Te- Aztecans, Hill pointed out that Fowler’s nificant decrease in effective moisture that huaca´n Valley (7, 31, 74, 75). Nonetheless, diagnostic taxa also are found in central began in the Great Basin Ϸ6900 cal. B.C. various contracting-stem point traditions Mexico, and she explained the inability to (65). We propose that this climate change existed in this region of western North reconstruct PUA terms for tropical spe- motivated the PUA bands to migrate to America for millennia before the intro- cies by the loss of these terms in the lan- the south and west, into areas watered by duction of maize, as well as being wide- guages of the Uto-Aztecans who migrated runoff from the range. In spread in areas lacking evidence of early northward into temperate zones (9). How- our scenario, the ancestors of the NUA maize agriculture, and Gypsum-style ever, our examination of the Uto-Aztecan remained in this region at least until points themselves may predate the ap- biological lexicon us to conclude Ϸ5500 cal. B.C. (6600 14C yr b.p.), when pearance of maize in the Southwest (4, 6, that the PUA homeland was located in pinyon populations began spreading 37, 76). In fact, no items of material cul- the north rather than in Mesoamerica and northward (56, 60, 66). We suggest that, ture that would indicate the likely pres- also that the breakup of the PUA speech before this date, the bands ancestral to ence of Mesoamerican migrants, such as community began Ϸ6900 cal. B.C., much the SUA split off and moved southeast- the distinctive stone bowls found in early earlier than proposed by Fowler, Hill, or ward toward the , crossing farming settlements in central and south- anyone else. Our perspective is based on a into the Gila River drainage of Arizona, ern Mexico (77), have been recovered consideration of both the biological terms where the onset of the increasingly arid from the early southwestern maize sites. that can be reconstructed for PUA and conditions may have been delayed until If in-migrating Mesoamerican farmers those that cannot, including, in the latter Ϸ6400 cal. B.C. (7500 14C yr b.p.) (67, 68). had introduced agriculture into the US case especially, the term for ‘‘pinyon’’ or Southwest, a more diverse suite of crops ‘‘.’’ Other Considerations Relevant to the Long- might be expected at the earliest maize Both Fowler and Hill noted that the Distance Migration Hypothesis. The ab- sites there. By 3540 cal. B.C., maize, pepo NUA languages share cognates for ‘‘pin- sence of convincing evidence that the squash ( pepo), and bottle gourd yon’’, reconstructed for Proto-Northern PUA were Mesoamerican farmers renders (Lagenaria siceraria) were being cultivated Uto-Aztecan as *t¯ipat, whereas the few the Bellwood–Hill hypothesis untenable, together in central and southern Mexico terms for ‘‘pinyon’’ that have been re- but it does not automatically preclude the (39), but they apparently entered the US corded in the SUA languages are not cog- possibility that maize agriculture was in- Southwest as separate introductions over nate either with the NUA terms or with troduced to the US Southwest by in- the course of nearly two millennia. Pepo one another (19, 21, 52, 53). In our view, migrating, proto-Mesoamerican farmers squash is reported at McEuen Cave in the absence of a PUA term for ‘‘pinyon’’ who were not Uto-Aztecans. The most southeastern Arizona by 2000 cal. B.C. reflects the absence of pinyon species compelling evidence for such a long- (34, 37), but the data required to evaluate

21022 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0906075106 Merrill et al. Downloaded by guest on September 26, 2021 this report remain unpublished. This cu- Unfortunately, no archaeological data for Aztecans at European contact. We sug- curbit is not documented elsewhere in the this interval are available for this zone, gest that the bands in this southern van- region for another 800 years, appearing but even if such migration took place, it is guard were the first Uto-Aztecans to ca. 1200 cal. B.C. in the Mogollon High- doubtful that, once the farmers reached acquire maize, which then diffused north- lands of New Mexico, where the earliest the region of higher effective moisture, ward from one UA-speaking band to evidence for bottle gourd, dated ca. 300 their populations would have increased so another. B.C., also occurs (4, 78). Other Me- rapidly that they would have been forced The arrival of maize in the southwest- soamerican domesticates—including Mexi- to seek new arable lands much farther ern United States Ϸ2100 cal. B.C. (3700 can grain (Amaranthus cruen- north. There is no reason to assume that 14C yr b.p.) occurred near the midpoint of tus), the common ( the rate of population growth in this area the wettest interval of the late Holocene vulgaris), and (Gossypium hirsu- would have differed significantly from that (3900 B.C.–A.D. 50) (5000–2000 14Cyr tum)—appear in the US Southwest be- further south in Mesoamerica, where, de- b.p.) (2, 68, 83, 84, 86, 90, 91, 93). The tween Ϸ800 B.C. and A.D. 300 (Table spite the cultivation of maize and other subsistence and settlement strategies of S3). Although research may alter domesticates, population densities re- the Archaic residents of the region at the the dates for the introduction of these mained low even during wetter periods time provided the cultural contexts into crops, the evidence currently available (57, 72, 90). In addition, the area of which maize was introduced, but the pau- indicates not only that they arrived in the greater effective moisture in the 3900– city of excavated premaize Archaic sites US Southwest at different but that 2100 cal. B.C. period extended at least limits our ability to construct models of each became integrated into local econo- 900 km south of the Tucson Basin, to these strategies (4, 33, 79, 94–96). Floral mies in distinct ways in different parts of near the northern borders of Me- and faunal remains recovered from a few the region (79). soamerica (90). It is difficult to conceive of these sites suggest a general pattern of Bellwood and Hill (9, 11) proposed that that the more southerly portions of this exploitation of a wide range of resourc- the integration of maize cultivation into vast area would not have included suffi- es—, cacti, wild grasses, a variety of Mesoamerican subsistence systems led to cient arable land to accommodate the other herbaceous plants, and diverse ani- population growth and the need to in- needs of any farming populations that mal species—with an emphasis on a com- crease food production, which encouraged might have migrated there. paratively small subset of these (33). In farming populations to expand into new keeping with behavioral ecological models agricultural areas. However, such demo- An Alternative Scenario. Our interpretation of subsistence strategies in unpredictable graphic factors appear to have been of of the data currently available on early environments like the southwestern little significance in the northward dis- maize agriculture in the US Southwest United States, it is assumed that this pat- persal of maize agriculture. Estimates of provides the basis for the following sce- tern reflected a combination of consider- population size and densities offered for nario. We propose that soon after the ations regarding resource availability, pre- different areas of Mesoamerica are quite beginning of the drier middle Holocene dictability, and productivity, with Archaic low throughout the terminal Late Archaic Ϸ6900 cal. B.C., the foraging bands that foragers focusing their subsistence activi- (2500–1600 B.C.) (72), during the same comprised the PUA speech community ties on species with higher energetic re- period that maize agriculture was already migrated from the west central Great Ba- turn rates when these were available and being adopted in the US Southwest. For sin to the southwestern edge of the Great shifting to resources with lower return example, even though maize had been Basin, with the bands ancestral to the rates when they were not (3, 34, 89, cultivated in the 2,000 km2 of the Valley SUA then migrating to moister areas east 97–99). of Oaxaca for at least 2,000 years, the to- of the Colorado River. Decreasing effec- Although many of the same or similar tal Late Archaic population at any given tive moisture in subsequent centuries wild plant and animal species were found time is estimated at 75–150 people, repre- probably motivated these SUA bands to throughout the area where the earliest senting a maximum density of 0.075 per- abandon the lowlands of this region in maize sites are located, some high priority sons/km2 (57). In addition, except in favor of better-watered middle Holocene resources were more abundant in or re- lowland areas where swidden cultivation refuges, which may have included the Col- stricted to certain areas, for example, pin- may have been practiced and soils were orado Plateau and the northern Sierra yon pine in the Colorado Plateau uplands nutrient-poor, there is no indication that Madre Occidental. Climatological factors and saguaro cactus and mesquite in the early Mesoamerican farmers migrated may also have encouraged some SUA lowlands. The lowland over any great distances for lack of arable bands to continue migrating southward. and grasslands also were charac- land (80, 81). Evidence of northward- The interval of decreased effective mois- terized by a greater species diversity and migrating populations also is absent in the ture in the southwestern United States density than the uplands, and in some late Middle Archaic archaeological record and northwestern Mexico between 6400 areas like southeastern Arizona, uplands of northwestern Mexico (6, 82). and 3900 cal. B.C. (7500–5000 14Cyrb.p.) and lowlands occurred in relative close The migration of Mesoamerica farmers corresponded to a period of increased proximity (97, 100, 101). Intraregional would more likely have been motivated by effective moisture in some areas of central differences in the distribution, density, climatological factors. During the 4300– Mexico (68, 83–86, 88–92). and seasonal availability of taxa of dietary 2100 cal. B.C. period when maize agricul- Linguistic evidence, mainly cognate significance obviously would have affected ture would have been diffusing northward, densities among the various modern SUA settlement strategies, with the result that a the climate of central Mexico languages, supports the view that the SUA wide array of specific strategies emerged. was characterized by decreased effective speech community did not remain intact The range of settlement strategies usu- moisture, whereas after 3900 cal. B.C., the for a long period but rather began sepa- ally is conceptualized in terms of differing southwestern United States and north- rating into distinct bands, with some re- degrees and forms of mobility. At one western Mexico experienced increased maining in the US Southwest and adja- end of the continuum is residential mobil- effective moisture (2, 68, 83–89). Farmers cent areas of northwestern Mexico while ity, in which hunter–gatherer bands living in the transitional zone between others moved south (15, 16). The latter moved from one resource zone to an- these two regions at this time may have probably included ancestors of the speak- other, often in a seasonal round. At the been drawn northward by the relatively ers of Cora, Huichol, and the Aztecan other is logistical mobility, in which resi- greater effective moisture available there. languages, the southernmost of the Uto- dential base camps were used for months

Merrill et al. PNAS ͉ December 15, 2009 ͉ vol. 106 ͉ no. 50 ͉ 21023 Downloaded by guest on September 26, 2021 at a time, with some members traveling of a range of settlement strategies associ- tracts suitable for -table and over- for short periods to acquire more distant ated with successful maize agriculture, but bank flood agriculture were available resources. Other strategies fall between the low visibility of many of these strate- but unused in other sections of the local these two extremes, combining seasonal gies in the archaeological record may have river valleys. residential mobility between resource limited our appreciation of their diversity LeBlanc suggests that these sites may zones with logistical mobility within these and distribution within the region. have been constructed by Mesoamerican zones (37, 102). The Tucson Basin sites clearly docu- migrants who were entering northwestern The arrival of maize in the southwest- ment the shift from a foraging to a mixed Mexico at the time that they were occu- ern United States coincided with a pe- foraging–horticultural economy, followed pied (119). However, similar constructions riod of increased effective moisture that by increasing reliance on maize and the do not appear in Mesoamerica until cen- presumably would have promoted abun- emergence of agricultural intensification turies later (120), and if the builders had dance in local wild resources and pro- in the form of irrigation. Features inter- originated in Mesoamerica, ceramics, vided a reliable subsistence base for the preted as irrigation canals have been which were widespread in Mesoamerica hunter–gatherers there (2, 83, 84, 89, dated to the period between 1750 and by 1600–1400 B.C., should be present 103). Berry has argued that successful 1500 cal. B.C., and networks of canals (72). We suspect that the residents of hunter–gatherers in the region would have been identified for the period begin- these sites migrated from relatively mod- not have been attracted to maize agri- ning at Ϸ1250 cal. B.C. (2, 37, 109–111). est distances to exploit the agricultural culture because its adoption would have This evidence predates that for Me- potential of the local river valleys, possibly disrupted their preexisting subsistence soamerica, where irrigation was estab- from the northwest, north, or east where strategies (104). He concluded that lished between 1200 and 800 B.C. (37, 79, Archaic sites are numerous (121–123). maize more likely would have been in- 112, 113). Many of the Late Archaic projectile troduced to the region by migrating The construction and maintenance of points found at these sites are most simi- farmers from the south because they irrigation systems implies the development lar in style to projectile points found to were already committed to its cultiva- of new forms of social organization in the the northeast in southern New Mexico tion. However, the damp alluvial locales Tucson Basin during the first millennium and far western Texas, and the identified most favorable for farming were already following the introduction of maize (37). sources for the stone tools recov- being used intensively by foragers, and The emergence of more complex labor ered from the site are located to the those bands that followed a more logisti- organization at this time also is suggested northeast, east, and southeast (124). cal mobility strategy could have inte- by the complex terrace constructions at grated maize agriculture into their econ- Cerro Juanaquen˜a, located in the Río Ca- Discussion omies with few adjustments, particularly sas Grandes Valley of Chihuahua, Mexico Our conclusion that maize agriculture if they focused during the summer - (Fig. 1) (114–117). Occupied between spread from Mesoamerica to the US fall season on wild resources available in 1300 and 1100 cal. B.C., this 10-hectare Southwest via group-to-group diffusion areas suitable for growing maize (4, 5, site includes Ϸ8 km of berm walls and is based on two fundamental consider- 79, 105). 550 residential terraces on the top and ations. First, other than maize itself, the Such a preadaptation for maize agricul- sides of a 140 m-high hill. We estimate early agricultural sites in the Southwest ture has been proposed for early farmers that the construction of these features have yielded no evidence of a Me- in the Tucson Basin, where the exploita- entailed Ϸ30 person-years of labor, an soamerican presence that would be ex- tion of diverse, concentrated wild re- investment comparable with that required pected if maize cultivation had been in- sources on the floodplain and adjacent to build a 550-room stone pueblo like troduced by in-migrating Mesoamerican areas of the middle Santa Cruz Valley those constructed in the southwestern farmers. Second, the earliest agricultural may have been associated with reduced United States during the first millennium components of these sites document residential mobility during much of the A.D. (115). Although we do not regard continuity with longstanding foraging year (37, 97, 101, 106, 107). Nonetheless, Cerro Juanaquen˜a and the Tucson Basin traditions and these and subsequent some of the earliest, directly dated maize sites as representative of all early agricul- components are consistent with demo- in the entire southwestern United States tural sites, they do illustrate the dramatic graphic and behavioral ecological mod- was recovered from the Old Corn site on impact that maize agriculture had in some els of what would be expected in this the Colorado Plateau, an area character- areas of the region. region when an introduced cultigen like ized by low species diversity and a short Cerro Juanaquen˜a forms part of one maize was integrated into a broad- growing season (35). This fact suggests of three clusters totaling ten terraced spectrum foraging subsistence strategy that certain features of maize—for exam- hilltop sites located along 85 km of the (51, 99, 111). ple, its storability and potential for higher Río Casas Grandes floodplain. We in- The scenario we propose here fits with yields than many wild, small-seeded terpret these clusters as fortified, defen- archaeological evidence that the desert plants—made it attractive to Middle Ar- sive sites because they display several lowlands of the Arizona– border- chaic hunter-gatherers in diverse ecologi- attributes expected of such sites: evi- lands, which appear to have been largely cal settings, who adjusted their settlement dence of planning and rapid, coordi- depopulated during the generally arid and subsistence strategies to take advan- nated construction, perimeter berms, middle Holocene (6400–3900 cal. B.C.), tage of the benefits that it offered (89, high frequencies of projectile points, were reoccupied from both north and 108). Yet the apparently short duration of 360-degree views, and intersite visibility, south at the onset of wetter conditions at the Old Corn occupation (2150–2050 cal. which would have facilitated line-of- Ϸ3900 cal. B.C., creating what we suspect B.C.) may indicate that the initial cultiva- sight signaling (118). Although identify- was a SUA linguistic–cultural continuum tion of maize in such upland areas was ing precisely who their enemies were is across which maize could have diffused less successful than in the Tucson Basin, impossible, they may have been defend- into the US Southwest (2, 6). It also is where a record of the continuous use of ing themselves and their food stores consistent with other, mainly linguistic, maize exists from the early third millen- against the raids of hunter–gatherers in data relevant to understanding the role of nium B.C. onward. Alternatively, relatively the area. We doubt that they were con- the Uto-Aztecans in the spread of maize short-term occupations similar to that in- structed because of competition among agriculture. The NUA and SUA branches dicated for Old Corn may have been one the farmers themselves as extensive are rather distinct, suggesting an early sep-

21024 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0906075106 Merrill et al. Downloaded by guest on September 26, 2021 aration, but patterns of cognate-sharing short-distance migrations of local farming and the US Southwest. We anticipate indicate that interaction among speakers populations, which frequently repositioned that research currently under way will of distinct UA dialects and languages con- themselves over the landscape to take ad- soon produce new insights and more re- tinued to take place, especially within vantage of damp floodplains and other fined chronologies and that additional each of these two main branches (15, 16, conditions propitious for cultivating maize. excavations, AMS dating, and genetic 125). By envisioning the northward diffu- We also would not argue that the dis- comparisons of archaeological samples of sion of maize as involving primarily inter- persal of maize agriculture and language early maize will provide the empirical action among SUA bands, our model of- families never cooccurred in Me- foundation upon which increasingly so- fers a parsimonious accounting for the soamerica. An example of this cooccur- phisticated models can be constructed. high level of cognate density for maize- rence perhaps can be seen in the case of However, extensive archaeological and related terms in these languages and the the Mayan language family and, more rel- paleoecological research in the area be- near-total absence of possible cognates for evant to the issues at hand, in the north- tween Mesoamerica and the US South- these terms in NUA languages (9, 43, ward migration of proto-Huastecan west, especially for the Archaic period, is 126–128). Mayan speakers from the postulated required to evaluate many components of Although we reject the hypothesis that Proto-Mayan homeland in highland Gua- these models. Although noting this need maize agriculture was introduced to the temala to northeastern Mexico, a distance has become a cliche´, the importance of filling this gap in our knowledge cannot US Southwest through the long-distance of Ϸ1,100 km (17, 129). That this migra- be overemphasized. migration of Mesoamerican farmers, Uto- tion took place after the Proto-Mayans Aztecans or otherwise, population move- received maize is indicated by the fact ACKNOWLEDGMENTS. We are grateful for the ments form a crucial component of our that reflexes of the postulated Proto- logistical support provided by the Mexico-North scenario for the diffusion of maize from Mayan term for ‘‘maize’’ are found in 29 Research Network and the University of Texas at San Antonio and for the research conducted by Mesoamerica to the Southwest. We are of the 30 modern (128). Desert Archaeology, Inc. We thank the Instituto convinced that the southward migration of The near ubiquity of shared cognates for Nacional de Antropología e Historia for permission SUA bands facilitated the diffusion of ‘‘maize’’ in these languages contrasts to conduct archaeological field research in Mexico. Funding for this research was provided by National maize agriculture from Mesoamerica to sharply with the lack of shared cognates Science Foundation Grants SBR-9708610, the US Southwest, and that maize agricul- for ‘‘maize’’ throughout the Uto-Aztecan SBR-9809839, SBR-0219185, and SBR-0220292, the ture was spread across much of the south- language family. National Geographic Society, the National Endow- western United States and adjacent areas Research conducted over the last de- ment for the , the Smithsonian’s Na- tional Museum of Natural History, the University of of northwestern Mexico through both cade has transformed our understanding Texas at San Antonio, the Arizona Department of group-to-group diffusion and relatively of early maize agriculture in Mesoamerica Transportation, and the City of Tucson.

1. Staller JE, Tykot RH, Benz BF, eds (2006) Histories of 17. Campbell L (1997) American Indian Languages (Ox- 31. Smith BD (2005) Reassessing Coxcatlan Cave and the Maize (Elsevier, Amsterdam). ford Univ Press, Oxford). early history of domesticated plants in Mesoamerica. 2. Mabry JB (2005) in The Late Archaic across the Bor- 18. Hill JH (2002) in Examining the Farming/Language Proc Natl Acad Sci USA 102:9438–9445. derlands, ed Vierra BJ (Univ Texas Press, Austin, TX), Dispersal Hypothesis, eds Bellwood P, Renfrew C (Mc- 32. Smiley FE (1994) The agricultural transition in the pp 41–83. Donald Inst Archaeol Res, Cambridge, United - northern Southwest: Patterns in the current chrono- 3. Piperno DR (2006) in Behavioral and the Tran- dom), pp 331–340. metric data. Kiva 60:165–189. sition to Agriculture, eds Kennett DJ, Winterhalder B 19. Hill JH (2002) Toward a linguistic prehistory of the 33. Gregory DA, ed (1999) Excavations in the Santa Cruz (Univ Press, Berkeley, CA), pp 137–166. Southwest: ‘‘Azteco-Tanoan’’ and the arrival of maize River Floodplain (Center for Desert Archaeol, Tucson, 4. Wills WH (1988) Early Prehistoric Agriculture in the cultivation. J Anthropol Res 58:457–475. AZ). American Southwest (School of American Research 20. Hill JH (2004) in Memoirs of the Seventh International 34. Mabry JB (2005) in Subsistence and Resource Use Press, Santa Fe, NM). Conference on Linguistics in the Northwest (Trans- Strategies of Early Agricultural Communities in 5. Minnis PE (1992) in The Origins of Agriculture, eds lated from Spanish) eds Moru´a Leyva MdC, Ortiz Southern Arizona, ed Diehl MW (Center for Desert Cowan CW, Watson PJ (Smithsonian Institution Press, Ciscomani RM (Universidad de Sonora, Hermosillo), Archaeol, Tucson, AZ), pp 113–152. Washington, DC), pp 121–141. Vol 2, pp 59–81. 35. Huber EK, Van West CR, eds (2005) Fence Lake Project 6. Mabry JB, Carpenter JB, Sanchez G (2008) . In Archae- (Statistical Research, Tempe, AZ). ology Without Borders, eds Webster LD, McBrinn ME 21. Hill JH (2006) in Histories of Maize, eds Staller JE, Tykot RH, Benz BF (Elsevier, Amsterdam), pp 631–645. 36. Whittlesey SM, Hesse SJ, Foster MS eds (2009) Recur- (Univ Press of Colorado, Boulder, CO), pp 155–183. rent and the Making of Place (SWCA En- 7. Berry CF, Berry MS (1986) in Anthropology of the 22. Smith DG, et al. (2000) Implications of the distribution vironmental Consultants, Tucson, AZ). Desert West, eds Condie CJ, Fowler DD, University of of Albumin Naskapi and Albumin Mexico for New 37. Mabry JB, ed (2008) Las Capas (Center for Desert Anthropological Papers (Univ Utah Press, World prehistory. Am J Phys Anthropol 111:557–572. Archaeol, Tucson, AZ). Lake City, UT) Vol 110, pp 253–327. 23. Matson RG (2002) in Examining the Farming/Lan- 38. Smith BD (1997) Reconsidering the Ocampo Caves and 8. Bellwood P (1997) in Archaeology and Linguistics, eds guage Dispersal Hypothesis, eds Bellwood P, Renfrew the era of incipient cultivation in Mesoamerica. Lat McConvell P, Evans N (Oxford Univ Press, ), C (McDonald Inst for Archaeol Res, Cambridge, United Am Antiq 8:342–383. pp 123–134. Kingdom), pp 341–356. 39. Smith BD (2001) Documenting plant : 9. Hill JH (2001) Proto-Uto-Aztecan: A community of 24. LeBlanc SA (2008) in Archaeology Without Borders, The consilience of biological and archaeological ap- cultivators in central Mexico? Am Anthropol 103:913– eds Webster LD, McBrinn ME (Univ Press of Colorado, proaches. Proc Natl Acad Sci USA 98:1324–1326. 934. Boulder, CO), pp 107–142. 40. Jaenicke–Despre´s VR, Smith BD (2006) in Histories of 10. Bellwood P (1993) An archaeologist’s view of lan- 25. Bronk Ramsey C (2001) Development of the radiocarbon Maize, eds Staller JE, Tykot RH, Benz BF (Elsevier, guage macrofamily relationships. Bull Indo-Pacific calibration program OxCal. Radiocarbon 43:355–363. Prehistory Assoc 13:46–60. Amsterdam), pp 83–95. 26. Reimer PJ, et al. (2004) IntCal04 terrestrial radiocar- 41. Haury E (1962) in Courses Toward Urban Life, eds 11. Bellwood P (2001) Early agricultural population dias- bon age calibration, 0–26 cal kyr bp. Radiocarbon poras? Farming, languages, and genes. Annu Rev Braidwood RJ, Willey GR (Wenner–Gren Foundation 46:1029–1058. Anthropol 30:181–207. for Anthropol Res, New York), pp 106–131. 27. Benz BF, Long A (2000) Prehistoric maize evolution in 12. Bellwood P (2003) in Areal Diffusion and Genetic 42. Benz BF (1999) in Pacific in Prehistory, the Tehuacan Valley. Curr Anthropol 41:459–465. Inheritance, eds Aikhenvald AY, Dixon RMW (Oxford ed Blake M (Washington State Univ Press, Pullman), 28. Blake M (2006) in Histories of Maize, eds Staller JE, Univ Press, Oxford), pp 27–43. pp 25–38. Tykot RH, Benz BF (Elsevier, Amsterdam), pp 55–72. 13. Diamond J, Bellwood P (2003) Farmers and their lan- 43. Campbell L (2002) in Examining the Farming/Lan- guages: The first expansions. Science 300:597–603. 29. Piperno DR, Ranere AJ, Holst I, Iriarte J, Dickau R (2009) guage Dispersal Hypothesis, eds Bellwood P, Renfrew 14. Bellwood P (2005) First Farmers (Blackwell, Oxford). Starch grain and phytolith evidence for early ninth mil- C (McDonald Inst Archaeol Res, Cambridge, United 15. Miller WR (1984) The classification of the Uto-Aztecan lennium b.p. maize from the central Balsas River Valley, Kingdom), pp 49–63. languages based on lexical evidence. Int J Am Linguist Mexico. Proc Natl Acad Sci USA 106:5019–5024. 44. Malhi RS, et al. (2003) Native American mtDNA pre- 50:1–24. 30. Piperno DR, Flannery KV (2001) The earliest archaeo- history in the American Southwest. Am J Phys An- 16. Cortina–Borjas M, Valin˜as Coalla L (1989) Some re- logical maize (Zea mays L.) from highland Mexico: thropol 120:108–124. marks on Uto-Aztecan classification. Int J Am Linguist New accelerator mass spectometry dates and the im- 45. Shaul DL, Hill JH (1998) Tepimans, Yumans, and other 55:214–239. plications. Proc Natl Acad Sci USA 98:2101–2103. . Am Antiq 63:375–396.

Merrill et al. PNAS ͉ December 15, 2009 ͉ vol. 106 ͉ no. 50 ͉ 21025 Downloaded by guest on September 26, 2021 46. Wilcox DR (1986) in Ripples in the Chichimec , eds 73. Clark JE, Gosser D (1995) in The Emergence of , 101. SK, Fish PR, Madsen J (1990) in Perspectives on Mathien FJ, McGuire RH (Southern Illinois Univ Press, eds Barnett WK, Hoopes JW (Smithsonian Inst Press, Southwestern Prehistory, eds Minnis PE, Redman CL Carbondale, IL), pp 135–153. Washington, DC), pp 209–222. (Westview Press, Boulder, CO), pp 76–91. 47. Lorenz JG, Smith DG (1996) Distribution of four 74. Carpenter JP, Sa´nchez G, Villalpando C ME (2005) in 102. Doleman WH (2005) in The Late Archaic across the founding mtDNA haplogroups among Native North The Late Archaic Across the Borderlands, ed Vierra BJ Borderlands, ed Vierra BJ (Univ Texas Press, Austin, Americans. Am J Phys Anthropol 101:307–323. (Univ Texas Press, Austin, TX), pp 13–40. TX), pp 113–140. 48. Eshleman JA, et al. (2004) Mitochondrial DNA and 75. MacNeish RS, Nelken–Terner A, Weitlaner-Johnson I 103. Castiglia PJ, Fawcett PJ (2006) Large Holocene lakes prehistoric settlements: Native migrations on the (1967) The Prehistory of the Tehuaca´n Valley (Univ and climate change in the Chihuahuan Desert. Geol- western edge of North America. Hum Biol 76:55–75. Texas Press, Austin, TX), Vol. 2. ogy 34:113–116. 49. Kaestle FA, Smith DG (2001) Ancient mitochondrial 76. Justice ND (2002) Stone Age and Arrow Points 104. Berry MS (1982) Time, , and Transition in Ana- DNA evidence for prehistoric population movement: of the Southwestern United States (Indiana Univ sazi Prehistory (Univ Utah Press, , UT). The Numic expansion. Am J Phys Anthropol 115:1–12. Press, Bloomington, IN). 105. Ogilvie MD (2005) in The Late Archaic across the 50. Kemp BM (2006) Mesoamerica and Southwest prehis- 77. Adams REW (2005) Prehistoric Mesoamerica (Univ Borderlands, ed Vierra BJ (Univ Texas Press, Austin, tory, and the entrance of into the : Oklahoma Press, Norman, OK), 3rd ed. TX), pp 84–112. Mitochondrial DNA evidence. PhD dissertation (Univ 78. Cutler HC, Whitaker TW (1961) History and distribu- 106. Matson RG (1991) The Origins of Southwestern Agri- California, Davis, CA). tion of the cultivated cucurbits in the Americas. Am culture (Univ Arizona Press, Tucson, AZ). 51. Kohler TA, Pier Glaude M, Bocquet-Appel J-P, Kemp Antiq 26:469–485. 107. Roth BJ (1995) Late Archaic occupation of the Upper BM (2008) The Neolithic demographic transition in 79. Doolittle WE, Mabry JB (2006) in Histories of Maize, Bajada: Excavations at AZ AA:12:84 (ASM). Kiva the US Southwest. Am Antiq 73:645–669. eds Staller JE, Tykot RH, Benz BF (Elsevier, Amster- 61:189–207. 52. Fowler CS (1983) Some lexical clues to Uto-Aztecan dam), pp 109–121. 108. Diehl MW, JA (2006) in Behavioral Ecology and prehistory. Int J Am Linguist 49:224–257. 80. Neff H, et al. (2006) Early Maya adaptive patterns: the Transition to Agriculture, eds Kennett DJ, Winter- 53. Hill JH (2008) Northern Uto-Aztecan and Kiowa- Mid–late Holocene paleoenvironmental evidence halder B (Univ California Press, Berkeley, CA), pp 63–86. Tanoan: Evidence of contact between the proto- from Pacific . Lat Am Antiq 17:287–315. 109. Craig DB, Hopkins M, Carpenter T (2009) Summary of languages? Int J Am Linguist 74:155–188. 81. Pohl MD, et al. (1996) Early agriculture in the Maya Archaeological Data Recovery for Pima County Re- 54. Van Devender TR, Spaulding WG (1979) Development lowlands. Lat Am Antiq 7:355–372. gional Wastewater Reclamation Department Plant of vegetation and climate in the southwestern United 82. MacWilliams AC, Hard RJ, Roney JR, Adams KR, Merrill Interconnect Project, Pima County, Arizona (North- States. Science 204:701–710. WL (2008) in Archaeology Without Borders, eds Web- land Research, Tempe, AZ). 55. Wells PV (1983) Paleobiogeography of montane is- ster LD, McBrinn ME (Univ Press of Colorado, Boulder, 110. Herr SA (2009) Las Capas. Archaeol Southwest 23:9–11. lands in the Great Basin since the last glaciopluvial. CO), pp 35–54. 111. Thiel JH, Mabry JB eds (2006) Rio Nuevo Archaeology Ecol Monogr 53:341–382. 83. Ely LL, Enzel Y, Baker VR, Cayan DR (1993) A 5000-year Program, 2000–2003 (Center for Desert Archaeol, record of extreme floods and climate change in the 56. Madsen DB (1986) in Anthropology of the Desert Tucson, AZ). southwestern United States. Science 262:410–412. West, eds Condie CJ, Fowler DD, University of Utah 112. Doolittle WE (1990) Canal Irrigation in Prehistoric 84. Metcalfe SE, Bimpson A, Courtice AJ, O’Hara SL, Taylor Anthropological Papers (Univ Utah, Salt Lake City, UT) Mexico (Univ Texas Press, Austin, TX). DM (1997) Climate change at the monsoon/westerly Vol 110, pp 21–41. 113. Caran SC, Neely JA (2006) Hydraulic engineering in boundary in northern Mexico. J Paleolimnol 17:155– 57. Flannery KV (1986) Guila´ Naquitz (Academic, New prehistoric Mexico. Sci Am 295:78–85. 171. York). 114. Hard RJ, Roney JR (1998) A massive terraced village 85. Buckler ES, IV, Pearsall DM, Holtsford TP (1998) Cli- 58. Malusa J (1992) Phylogeny and biogeography of the complex in Chihuahua, Mexico, 3000 years before mate, plant ecology, and central Mexican Archaic pinyon pines (Pinus subsect. Cembroides). Syst Bot present. Science 279:1661–1664. subsistence. Curr Anthropol 39:152–164. 17:42–66. 115. Hard RJ, Zapata JE, Moses BK, Roney JR (1999) Terrace 86. Ortega–Ramírez JR, Valiente–Banuet A, Urrutia–Fucu- 59. Beck C, Jones GT (1997) The terminal /early construction in northern Chihuahua, Mexico: 1150 B.C. gauchi J, Mortera–Gutie´rrez CA, Alvarado–Valdez G Holocene archaeology of the Great Basin. J World and modern experiments. J Field Archaeol 26:129–146. (1998) Paleoclimatic changes during the Late Pleisto- Prehist 11:161–236. 116. Roney JR, Hard RJ (2002) in Traditions, Transitions, cene—Holocene in Laguna Babícora, near the Chihua- 60. Lanner RM, Van Devender TR (1998) in Ecology and and Technologies, ed Schlanger S (Univ Colorado huan Desert, Me´xico. Can J Sci 35:1168–1179. Biogeography of Pinus, ed Richardson DM (Cambridge Press, Boulder, CO), pp 163–180. 87. Polyak VJ, Asmerom Y (2001) Late Holocene climate Univ Press, Cambridge, United Kingdom), pp 171–182. 117. Hard RJ, Adams KR, Roney JR, Schmidt KM, Fritz GJ and cultural changes in the southwestern United 61. Koehler PA, Anderson RS, Spaulding WG (2005) De- (2008) in Case Studies in Environmental Archaeology, States. Science 294:148–151. velopment of vegetation in the central eds Reitz E, Scarry CM, Scudder SJ (Springer, New 88. Nordt L (2003) Late Quaternary fluvial landscape evo- of California during the Late Quaternary. Palaeo- York), pp 315–333. lution in desert grasslands of northern Chihuahua, geogr Palaeocl Palaeoecol 215:297–311. 118. Hard RJ, Roney JR (2007) in Trincheras Sites in Time, Mexico. Geol Soc Am Bull 115:596–606. 62. Byrne R, Edlund E, Mensing S (1991) in Proceedings of Space, and Society, eds Fish PR, Fish SK, Villalpando 89. Hard RJ, Roney JR (2005) in The Late Archaic across the the Symposium on Oak Woodlands and Hardwood Borderlands, ed Vierra BJ (Univ Texas Press, Austin, ME (Univ Arizona Press, Tucson, AZ), pp 11–52. Range Management, October 31-November 2, 1990, TX), pp 141–186. 119. LeBlanc SA (2002) in Examining the Farming/Lan- Davis, California, ed Standiford RB (US Dept Agricul- 90. Metcalfe SE, O’Hara SL, Caballero M, Davies SJ (2000) guage Dispersal Hypothesis, eds Bellwood P, Renfrew ture, Pacific Southwest Research Station, Berkeley, Records of Late Pleistocene–Holocene climatic change C (McDonald Inst Archaeol Res, Cambridge, United CA) General Technical Report PSW-126, pp 182–188. in Mexico—A review. Quaternary Sci Rev 19:699–721. Kingdom), pp 357–365. 63. Van Devender TR (1977) Holocene woodlands in the 91. Waters MR (1989) Late Quaternary lacustrine history 120. Flannery KV, Marcus J (2003) The origin of war: New 14 southwestern deserts. Science 198:189–192. and paloeclimatic signficance of pluvial Lake Cochise, C dates from ancient Mexico. Proc Natl Acad Sci USA 64. Munroe JS (2003) Holocene timberline and palaeocli- southeastern Arizona. Quaternary Res 32:1–11. 100:11801–11805. mate of the northern , northeastern 92. Metcalfe SE, Street–Perrott FA, Perrott RA, Harkness DD 121. Formby DE (1986) Pinto–Gypsum Complex projectile Utah, USA. The Holocene 13:175–185. (1991) Palaeolimnology of the Upper Lerma Basin, Mex- points from Arizona and New Mexico. Kiva 51:99–127. 65. Wigand PE, Rhode D (2002) in Great Basin Aquatic ico: A record of climatic change and anthropogenic 122. Phelps AL (1998) An inventory of prehistoric Native Systems History, eds Hershler R, Madsen DB, Currey DR, disturbance since 11600 yr BP. J Paleolimnol 5:197–218. American sites in northwestern Chihuahua. The Arti- Smithsonian Contributions to the Earth Sciences (Smith- 93. Krider PR (1998) Paleoclimatic significance of late fact 36:1–176. sonian Inst Press, Washington, DC) Vol. 33, pp 309–367. Quaternary lacutrine and alluvial stratigraphy, Ani- 123. Miller MR, Kenmotsu NA (2004) in The Prehistory of 66. Grayson DK (1993) The Desert’s Past (Smithsonian Inst mas Valley, New Mexico. Quaternary Res 50:283–289. Texas, ed Perttula TK (Texas A&M Univ Press, College Press, Washington, DC). 94. Huckell BB (1996) The Archaic prehistory of the North Station, TX), pp 205–265. 67. Menking KM, Anderson RY (2003) Contributions of La American Southwest. J World Prehist 10:305–373. 124. Vierra BJ (2005) in The Late Archaic across the Bor- Nin˜a and El Nin˜o to middle Holocene drought and late 95. Mabry JB, Doolittle WE (2008) in Archaeology With- derlands, ed Vierra BJ (University of Texas Press, Aus- Holocene moisture in the American Southwest. Geol- out Borders, eds Webster LD, McBrinn ME (Univ Press tin, TX), pp 187–218. ogy 31:937–940. of Colorado, Boulder, CO), pp 55–70. 125. Heath J (1977) Uto-Aztecan morphophonemics. Int 68. Mabry JB (1998) in Paleoindian and Archaic Sites in 96. Roth BJ, Freeman A (2008) The Middle Archaic Period J Am Linguist 43:27–36. Arizona, ed Mabry JB (Arizona State Historic Preser- and the transition to agriculture in the Sonoran 126. Fowler CS (1994) in Corn and Culture in the Prehistoric vation Office, Phoenix, AZ), pp 19–31. Desert of southern Arizona. Kiva 73:321–353. , eds Johannessen S, Hastorf CA (Westview 69. Niederberger C (1979) Early sedentary economy in the 97. Diehl MW (1997) in Rediscovering Darwin, eds Barton Press, Boulder, CO), pp 445–467. Basin of Mexico. Science 203:131–142. CM, Clark GA, Am Anthropol Assoc Archaeol Papers (Am 127. Valin˜as Coalla L (2000) in Nomadic and Sedentary 70. Scott S (2001) in Archaeology of Ancient Mexico and Anthropol Assoc, Arlington, VA) Vol. 7, pp 251–266. Peoples in Northern Mexico (Translated from Span- , eds Evans ST, Webster DL (Garland, 98. Piperno DR, Pearsall DM (1998) The Origins of Agricul- ish) eds Hers M-A, Mirafuentes JL, Soto MdlD, Valle- New York), pp 266–270. ture in the Lowland Neotropics (Academic, ). bueno M (Universidad Nacional Auto´nomadeMe´x- 71. Heidke JM (2006) in Rio Nuevo Archaeology Program, 99. Kennett DJ, Winterhalder B eds (2006) Behavioral ico, ), pp 175–205. 2000–2003, eds Thiel JH, Mabry JB (Center for Desert Ecology and the Transition to Agriculture (Univ Cal- 128. Brown CH (2006) in Histories of Maize, eds Staller JE, Archaeol, Tucson, AZ), pp 7.1–7.93. ifornia Press, Berkeley, CA). Tykot RH, Benz BF (Elsevier, Amsterdam), pp 647–663. 72. Clark JE, Cheetham D (2002) in The Archaeology of 100. Huckell BB (1995) Of Marshes and Maize. Univ Ari- 129. Kaufman T (1976) Archaeological and linguistic cor- Tribal Societies, ed Parkinson WA (Int Monogr in Pre- zona Anthropol Papers (Univ Arizona Press, Tucson, relations in Mayaland and associated areas of Meso- history, Ann Arbor, MI), pp 278–339. AZ) Vol. 59. America. World Archaeol 8:101–118.

21026 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0906075106 Merrill et al. Downloaded by guest on September 26, 2021