An Archaeological Investigationof Late Archaic Cerros de Trincheras Sites

in Chihuahua, Mexico

Results of the 1998 Investigations

Robert J. Hard and John R. Roney

with contributions byKaren Adams, Susan Fish, Gayle Fritz, Kevin Hanselka, Bruce Moses,

Jennifer Nisengard, Lee Nordt, Kari Schmidt, Steven Shackley, Bradley Vierra,and Jos E. Zapata

Report toConsejo de Arqueologa

Instituto Nacional de Antropologa e Historia

Center for Archaeological ResearchThe University of Texas at San Antonio

Special Report, No. 251999

1

23

45

orig ina lsur fa ce

colluvium

be rmwall

b e dro c k

fine sedim en t

con struction fill

colluvium

Hard and R

oneyA

n Archaeological Investigation of the Late A

rchaic Cerros de Trincheras Sites in Chihuahua, M

exicoSR25

1999

REPORTto

Consejo de ArqueologaInstituto Nacional de Antropologa e Historia

An Archaeological Investigationof Late Archaic Cerros de Trincheras Sites

in Chihuahua, MexicoResults of the 1998 Investigations

Robert J. Hard Center for Archaeological Research

The University of Texas at San AntonioSan Antonio, Texas 78249-0658

U.S.A.

and

John R. RoneyBureau of Land Management

United States Department of the Interior435 Montao, N.E.

Albuquerque, New Mexico 87107U.S.A.

with contributions byKaren Adams, Susan Fish, Gayle Fritz, Kevin Hanselka, Bruce Moses,

Lee Nordt, Kari Schmidt, Steven Shackley, Bradley Vierra, and Jos Zapata

Translated by Jos Zapata

April 26, 1999

Center for Archaeological ResearchSpecial Report, No. 25

iContentsFigures ................................................................................................................................................................. iiTables .................................................................................................................................................................. iiAcknowledgments ............................................................................................................................................. iiiIntroduction ......................................................................................................................................................... 1Excavations.......................................................................................................................................................... 3Additional Field Studies ...................................................................................................................................... 8Analytical Results .............................................................................................................................................. 18Summary and Conclusions ................................................................................................................................ 28Appendix 1: Cerro Juanaquea Feature Descriptions ....................................................................................... 33Appendix 2: Cerro los Torres Feature Descriptions (94-287) ........................................................................... 53Appendix 3: Cerro Vidal Feature Descriptions (95-392) .................................................................................. 57Appendix 4: Modern Vegetation Found in the Janos Area, September 46, 1998 ........................................... 59References Cited ................................................................................................................................................ 61

ii

Figures1. Aerial photograph of Cerro Juanaquea, with the Rio Casas Grandes floodplain in foreground. .................. 12. Five cerro de trincheras sites photographed with aerial imagery. ................................................................... 23. Cerro Juanaquea with 1997 and 1998 excavations. ...................................................................................... 44. Cerro los Torres with 1998 excavation. .......................................................................................................... 55. Cerro Vidal with 1998 excavation. .................................................................................................................. 66. Idealized terrace construction stratigraphy. .................................................................................................... 77. Cerro Juanaquea projectile points. ................................................................................................................ 98. Geomorphic map of the Rio Casas Grandes and Rio San Pedro. .................................................................. 129. Geomorphic cross-section of the Rio Casas Grande. .................................................................................... 1510. Geomorphic cross-section of the Rio San Pedro. ........................................................................................ 1511. Locational of rock art panels on Cerro Juanaquea. ................................................................................... 1912. Obsidian sources. ......................................................................................................................................... 2113. Plan view of T97. ........................................................................................................................................ 3314. Cross-section of T97 ................................................................................................................................... 3415. Unit 1 and 2 profile, T97. ............................................................................................................................ 3516. Plan view of T163. ...................................................................................................................................... 3917. Cross-section of T163 ................................................................................................................................. 4018. Unit 1 and 2 profiles, T163. ........................................................................................................................ 4019. Plan view of T387. ...................................................................................................................................... 4720. Cross-section of T387. ................................................................................................................................ 4821. Unit 1 and 2 profiles, T387. ........................................................................................................................ 4822. Plan view of T1, Cerro los Torres. .............................................................................................................. 5323. Cross-section of T1, Cerro los Torres. ........................................................................................................ 5424. Unit 1 and 2 profiles, Cerro los Torres. ....................................................................................................... 5525. Rock-Ring Features on north end of Cerro Vidal. ...................................................................................... 56

Tables1. Shovel Test Results ......................................................................................................................................... 32. 1998 Surface Collections ................................................................................................................................. 83. Geomorphic Radiocarbon Dates from Samples ............................................................................................ 144. Summary of AMS Radiocarbon Dating, 1997 and 1998 .............................................................................. 205. Summary of 1997 Faunal Analysis ............................................................................................................... 226. Number and Percentage of Identified Faunal Remains from 1997 Excavations .......................................... 227. Summary of Faunal Remains from Terrace 163, 1998 Excavations ............................................................ 238. Charred Reproductive Parts ........................................................................................................................... 259. Wood Charcoal .............................................................................................................................................. 2510. 1997 Pollen Samples ................................................................................................................................... 2611. Percentages of Pollen Grains by Taxa ......................................................................................................... 27

iii

Acknowledgments

We conducted the field work reported here between June 1 and July 3, 1998, with the permission of INAHsCounsel on Archaeology (Instituto Nacional de Antropologa e Historia, PA/21/98/No. 3143) and with theconcurrence of the Municipalities of Casas Grandes and Janos, and the Ejidos of Casas Grandes, Hidalgo, andJanos. The study was funded by the National Science Foundation (SBR-97086210 and SBR-9809839). Wecould not have conducted this work without the support of these organizations and greatly appreciate the assis-tance their personnel have provided us. We greatly appreciate the assistance of the many individuals who haveassisted us and we apologize for being unable to include everyone.

We want to thank Ing. Joaquin Garca Barcena, President of INAHs Counsel on Archaeology (INAH-Mxico);Antrop. Jos Luis Perea Gonzlez, Director of Centro INAH-Chihuahua; Lic. Arturo Pea Zazueta, Administra-tor of Customs, Paloma, Chihuahua; Sr. Nazario L. Prieto, Mayor of Janos; Sr. Lorenzo Barajas, President of theJanos Ejido; Sr. C. Alfonso Bustillos, Mayor of Casas Grandes; Sr. Espiridin Terrazas of Nuevo Casas Grandes;and Sr. Cruz Lara, President of the Hidalgo Ejido.In the field, our project was greatly aided by the efforts of Jorge Bencomo, Casimiro Lucero, and Peter Wall,residents of Colonia Oaxaca and Colonia la Virginia. We also greatly appreciated the assistance and patience ofSr. Gerardo Barajas, Mayors Secretary, Municipality of Janos; Sra. Geli Rubio and Sr. Saul Jaques, who pro-vided our living accommodations in Janos; and Sr. Filiberto Lopez, for the many instances that he assisted uswith repairs of various sorts. And for the attention, assistance, and the many delicious meals, we particularlywant to thank Sra. Angelina Madrid and members of her family.

We are extremely fortunate to have some of the finest colleagues in the world whose expertise, dedication, andpersonal qualities have made this project successful and enjoyable from the beginning. Thanks to: Dr. KarenAdams for her wonderful ethnobotanical analyses; Dr. Susan Fish for her detailed pollen work; Dr. Gayle Fritz forher careful analysis of our cheno-am samples; Dr. Lee Nordt for his unraveling of the geomorphology of thefloodplain and the trincheras; Dr. Steven Shackley for his thorough obsidian work; and Dr. Brad Vierra for hissophisticated lithic analysis.

The support of our home institutions has been essential to negotiate all the details necessary to conduct a projectof this scale: The Bureau of Land Management has been very supportive of this project. At UTSA we would liketo thank in particular Dwight Henderson, dean of the College of Social and Behavioral Sciences, for his support;Sherri Sunaz, administrative assistant at CAR, has performed hundreds of tasks related to the preparations forand administration of the project; Carol Hollingsworth, Kathi Kortz, and Cyndi Orth of Grants and Contractsand Mike Wright of CAR have expertly handled the many financial transactions; and Britt Bousman has servedasinterim and associate director of CAR and kept CAR running smoothly while Robert J. Hards attention andpresence was focused on this project. Bruce Moses has prepared all of the excellent figures and William Bishelalso has done a superb job preparing this document for publication. A UTSA Faculty Development Award wasalso instrumental in allowing Hard time to work on this report and other aspects of the project analysis andreporting.

Finally we want to recognize the members of our team who bore the brunt of the heat, the hill, and the hardwork; but always with good cheer: Kevin Hanselka, Bruce Moses, Jennifer Nisengard, Gerry Raymond, RudiRoney, Kari Schmidt, Cindy Tennis, Jos E. Zapata, and Bridget Zavala. In addition we were aided by thewonderful assistance and company of Kim Rydel (Hard) and Christopher Hard. The assistance of Jos E.Zapata, who has also served as expert liaison and translator for the project has also been invaluable on manyfronts, large and small.

To each and everyone we are very grateful.

Robert J. Hard and John R. Roney

iv

1IntroductionCerro Juanaquea is a large cerro de trincheras lo-cated in northwestern Chihuahua, in the municipio ofJanos. The site was built over 3000 years ago on thesummit and slopes of a 140 meter high basalt hill whichoverlooks the floodplain of the Rio Casas Grandesand its major tributary, the Rio San Pedro. Large con-structed terraces cover an area of about 8 hectares,with over 8 kilometers of terrace wall and 108 stonecircles (see Figure 1). A large number of stone arti-facts are found with these features, and excavationshave revealed rich midden deposits with abundantbone and carbonized plant material (Hard and Roney1998a).

In our 1998 informe (reporting results of our 1997field season) we described test excavations in threeterraces and four stone circles, as well as mappingactivities, collection of surface materials, geomorphicinvestigations of the Rio Casas Grandes floodplain,documentation of rock art, and reconnaissance at othersimilar sites.

This informe summarizes the investigations under-taken at Cerro Juanaquea and other related sites un-der the oficio No. C.A. 40136/560 (22 de mayo de1998) authorized by the Consejo de Antropologa. Theproject was funded in 1997 by National Science Foun-dation (NSF) SBR-97086210 and in 1998 by NSFSBR-9809839. Our primary objective during 1998 wasto recover charcoal for radiocarbon dating andmacrobotanical analysis, as well as to expand thesample of faunal materials. Our strategy was to select10 terraces on Cerro Juanaquea for test excavation.Descriptions of these excavations are in Appendix 1.We continued geomorphological investigations offloodplain deposits, which were begun in 1997, as wellas in-field analysis of ground stone and limited sur-face collection. With assistance from the INAH CentroRegional in Chihuahua we also arranged to have large-scale aerial photographs taken of Cerro Juanaqueaand four other nearby cerros de trincheras, as well asa part of the floodplain near Cerro Juanaquea. Ana-lytical activities continued, and we are able to reporttwo additional radiocarbon dates on materials exca-vated in 1997, as well as preliminary results of the

Figure 1. Aerial photograph of Cerro Juanaquea, with the Rio Casas Grandes floodplain in fore-ground. View toward northwest.

2C e rro Ju a na qu ea

C e rro e l C a nelo

C e rro la C ruz

C e rro los To rre s

C e rro V id a l

Ri o

C

a

a

sr

s

nG

a

d

e

s

R i o

S

a n

a

t M

a

r i a

L a g . G u z m n

L a g . S a n t a

M a r i a

R i o

S

a n

P

oe

d r

M N

U N I T E D S T A T E S

M E X I C O

Figure 2. Five cerro de trincheras sites photographed with aerial imagery.

3geomorphological study, pollen analysis, and lithicanalysis.

In 1998 we expanded our project to include limitedwork at two other cerro de trincheras in the Rio CasasGrandes drainage, Cerro los Torres and Cerro Vidal(see Figure 2). Both of these sites were originally reg-istered and mapped during Minnis and Whalensproject in 1996. Cerro los Torres is located about 20km north of Nuevo Casas Grandes. The site consistsof over 2 km of terrace walls and perhaps 20 rockrings built on an 85 m high isolated hill overlookingthe floodplain. Although some pottery and a few ar-row points have been found on Cerro los Torres, italso has dart points and other artifactswhich are reminiscent of those at CerroJuanaquea. The terraces and stonecircles are also similar to the features onCerro Juanaquea. For these reasons wesuspect that the site was originally builtand used during the Late Archaic period.This site and our test excavations thereare described more fully in Appendix 2.

Cerro Vidal is a third cerro de trincheraslocated adjacent to the Rio PiedrasVerdes 6 kilometers south of ColoniaJuarez in the municipio of Casas GrandesViejo. This site is located on an 120 mhigh hill. It includes approximately 2.3 km of terracewall and about 40 rock rings. No pottery has beenfound, and although there are arrow points on the site,its overall similarity to Cerro Juanaquea in terms ofartifacts and form of its features suggests that it alsodates to Late Archaic times. Owing to newly acquireddata, this last suggestion has been reconsidered. Thissite and our test excavations there are described inAppendix 3.

Excavations

Shovel Tests

In 1998 we excavated nine small shovel tests at CerroJuanaquea, each less than .5 m x .5 m in size. Theseunits were excavated in 10 cm increments and the ex-

cavated soil was screened through 1/8-inch meshscreen. Artifacts and bone from these units were tabu-lated, but not curated. Instead they were put back intothe holes as they were backfilled. The purpose of theseunits was to determine whether or not more extensivetesting was likely to be productive in particular ter-races. Positive results in these small test units led usto place larger test units in T387, T413, and R234a.Several other terraces appeared promising on the ba-sis of surface observations but were not selected formore extensive testing because the small test pitsyielded negative results (T74, T195, T213, T217,T261, and T273). The shovel tests and their resultsare summarized in Table 1.

Test Units

Ten terraces at Cerro Juanaquea, one terrace Cerrolos Torres, and one terrace at Cerro Vidal were se-lected for more intensive testing. Figures 3, 4, and 5show maps of each site and the locations of our exca-vations. Appendices 1, 2, and 3 provide detailed de-scriptions of the individual terraces which were testedas well as the results of these excavations. Our pri-mary objective in these test units was to obtain char-coal for radiocarbon dating and macrobotanicalanalysis. In addition, we hoped to expand the sampleof faunal material and to expose stratigraphic profileswhich would help us better understand the formationprocesses on Cerro Juanaquea. The terraces were se-lected with several criteria in mind. First, results ofthe 1997 excavations suggested that bone and char-coal are not well preserved near the surface, so weattempted to select terraces which had deposits a meter

Table 1. Shovel Test Results

Feature Depth ResultsT 74 30 cm 6 lithicsT 195 80 cm 17 lithics, 1 boneT 213 40 cm 11 lithics, 3 boneT 217 50 cm 19 lithics, 9 boneT 261 40 cm 1 lithic, 1 boneT 273 40 cm Numerous lithics and bone fragmentsT 387 52 cm 28 lithics, 11 bone, 14 charcoal

T 413 50 cmA few lithics and bones, significant charcoal

R 234a 40 cm 16 lithics, 11 bone

4Figure 3. Cerro Juanaquea with 1997 and 1998 excavations.

R1a

6

1 0

R28

1 26

1 63

167

R222

2 34R2 34 a

R250

2 73

2 90

2 97

9 7

537

3 87

4 13

?

M N

0 25 50 75 100

M N

0 25 50 75 100

5Figure 4. Cerro los Torres with 1998 excavation.

M N

M N

0 25 50 75 10 0

0 25 50 75 10 0

199 8 excava tion sw a ll/te rra cebedroc k le dgerubblec a irnou tc roptra il

T1

T1

6Figure 5. Cerro Vidal with 1998 excavation.

T20

MN

MN

0 25 50 75 10 0

0 25 50 75 10 0

w a ll/te rra c eb ed ro c k le d g eru b b lec a irno utc ro ptra il

7or more in depth. In addition we looked for ashy de-posits, bone exposed on the surface, and artifact asso-ciations suggesting intensive use. Finally, at CerroJuanaquea we attempted to select terraces from vari-ous parts of the site, so that we could more confidentlydefine the span of occupation. It is possible that dif-ferent parts of the site were occupied at different times,and in order to investigate this possibility chronomet-ric dates from different areas are required.

We were successful in recovering fragments of charredmaize from both Cerro los Torres and Cerro Vidal.These samples will allow us to begin dating othercerros de trincheras. At Cerro Juanaquea eight of theten terraces tested yielded material suitable for AMSradiocarbon dating. Additional lithic artifacts, bone,and/or botanical materials were also recovered dur-ing these excavations and provide the basis for ongo-ing analyses (see below). We did not find any structuralremains, post holes, storage pits, hearths, or other simi-lar features in 1998. The profiles exposed in these testunits have also given us a better understanding of for-mation processes on the cerros de trincheras.

Figure 6 is an idealized stratigraphic profile which re-flects our current understanding of the terrace deposits.Each individual terrace exhibits its own distinctive char-acteristics, but all conform to this general pattern. Thedeepest stratum in Figure 6 is bedrock. Resting on bed-rock we sometimes recognize traces of the original soilwhich was present on the hillslope before constructionof the terraces (zone 1). When present, these original

surface deposits can be recognized by the presence ofcalcium carbonate in the sediment and encrusting therock. Typically the sediments are very dark grayish brownto black, sandy clay loam, containing about 50 percentbasalt pebbles and cobbles with a blocky structure. Abovebedrock and original soil is the constructed cobble bermwhich defines the terrace wall (zone 2). This feature iscomposed of larger cobbles (520 cm) resting againstone another, with relatively little soil (2550 percent)filling the spaces between the rocks. Behind the terracewall is the construction fill consisting of a brown sandyclay loam and up to about 25 percent pebbles and somecobbles placed behind the berm deposit which forms theterrace platform (zone 3). There is usually more soil inthis deposit and the rocks are smaller, allowing it to bedistinguished confidently from the terrace wall itself. Inmost cases the deeper portions of these deposits are verysoft, becoming increasingly compact toward the surface.We hypothesize that during original construction of theterrace many voids remained between the cobbles, andthat subsequently through several different mechanismssoil has made its way down to fill these voids. Above thecobble and soil terrace platform deposit we sometimesfind a layer of very dark brown, fine sandy clay loamwith some pebbles which is 5 to 20 cm thick (zone 4).Sometimes distinct horizontal zones of compaction canbe recognized in this stratum. We believe that this is alayer of fine sediment which the prehistoric inhabitantsof Cerro Juanaquea placed on the terrace platform toprovide an actual use surface. In the experimental ter-race which we constructed these sediments worked theirway down into the platform itself, filling voids between

1

23

45

origina lsur face

colluvium

bermwall

be d rock

fine sediment

construction fill

colluvium

Figure 6. Idealized terrace construction stratigraphy.

8cobbles. If this happened in prehistoric times, itwould have been necessary to periodically add newsoil to the surface. Above this layer we often find afew centimeters of fine sandy loam which are alsodark brown, although slightly lighter in color (zone5). These appear to be post-occupational deposits,probably derived from material transported to thehill by aeolian processes, then redeposited as collu-vium on the terrace surfaces leaving multiple finelaminae.

Backfilling

A combined total of 44 shovel tests, units, andbackhoe trenches were excavated (see Geomor-phological Studies and Appendices 1, 2, and 3).With the exception of the backhoe trenches, be-fore initiating these excavations, an area within closeproximity of the shovel test or unit was selected forscreening of excavated soils. The surface area was thencovered with plastic sheeting in order to isolate thescreened soils, and subsequently facilitate the back-filling process. The last two days of field work wereutilized for this process. Prior to backfilling, a coinand/or aluminum can was placed the test unit, whichwas then covered over with plastic sheeting. The unitswere then backfilled with the excavated backdirt. Theprocess was photodocumented and the photographsare on file at the Center for Archaeological Research,University of Texas at San Antonio (CAR-UTSA).Backhoe trenches were all backfilled as well.

Additional Field Studies

Surface Collections

During the 1998 field season we continued to collectartifacts from the surface of the three sites. As in pastyears the location of each artifact was carefully plot-ted on a large-scale maps of the sites. Each artifact isnumbered individually and entered into the curatedcollection of artifacts and samples. Table 2 summa-rizes the items which were collected in this way fromeach of the three sites.

Digital images of all of the projectile points and othersmall chipped stone artifacts were captured by scan-ning them directly using a digital scanner (Houk andMoses 1998). Preliminary drawings of many of theseartifacts have been made using CorelDraw7, based onthe scanned image (Figure 7).

Ground Stone

The large numbers of manos and metates present onthe surface of Cerro Juanaquea offer an importantsource of information about the kinds of food pro-cessed, the extent of food processing, and the loca-tions of residential activity. During the 1997 fieldseason we conducted a detailed analysis of a surfacesample of ground stone tools both from terraces thatwere tested and with transect sampling units. Duringthe 1998 season we continued to collect surface groundstone data by making detailed observations on allground stone tools found on the surface of each exca-vated terrace.

During the surface mapping of each terrace and priorto initiating excavations, the crew placed pin flags byall ground stone tools, as well as other significant ar-tifacts, and marked their location on the map. An iden-tifying number was assigned to each ground stone tool.Later one or two team members would return to theterrace and walk over the surface a second time to

Cerro Juanaquea

Cerro los Torres

Cerro Vidal

Projectile Points 24 7Biface 5 1Cruciform 3Drill 1 1Notched Flake 1Utilized Flake 1Core/Split Cobble 2Stone Bowl 5 1Stone Pipe 1Mano 2 1Metate 2Pestle(?) 1Sherd 1Obsidian Nodule 1

Table 2. 1998 Surface Collections

9Figure 7. Cerro Juanaquea projectile points. K and L are a cruciform and distal end of a bone awl.

10

search for additional ground stone tools prior to re-cording the tools as the dense rocks on the terrace wallsmade the ground stone tools difficult to see. Each toolor tool fragment would then be closely examined andits attributes recorded. For the manos the followingattributes were recorded in the field: feature, itemnumber, material, number of ground surfaces, degreeof wear, the presence of edge shaping, the presence ofpecking, the presence of striations, striation direction,plan shape, cross-section shape, length, width, thick-ness, plan shape, and the surface depositional context(on terrace surface, talus slope, etc.). For the metateswe recorded the same information as the manos andthese additional attributes: metate type; the dimensionsof the hole in the bottom of the metate, if present; thelength and width of the grinding surface; and the depthof wear. In addition to this recording, we drew andphotographed samples of various metate and manoforms and we collected a few of them.

In 1998 we recorded a total of 17 whole manos andseven mano fragments from the surface of nine ter-races and one rock ring. Each terrace had betweenone and five manos present. The combined results of1997 and 1998 indicate a total of 80 whole manoshave been recorded from 32 features. Eighty percentof these were manufactured from the local basalt withthe balance being made from rhyolite and unknownmaterials. There are at least two general groups ofmanos. One tends to be more circular or oval in planand convex to spheroid in cross-section. This shapewas likely used with the heavily used, large basinmetates. The second form is flat in cross-section andwas likely to have been used with the small grindingslab metates. The mean area of all manos is 141 cm2and the mean length is 13 cm, indicating these aresmall compared to those found in later pueblo sitesyet they are somewhat larger than typical Archaicmanos. Recent research by Hard and others has sug-gested that mano size is related to agricultural depen-dence. The manos at Cerro Juanaquea are roughlycomparable in size to those from a number of pithouseperiod occupations in the Southwest that exhibit amoderate level of agricultural dependence (Hard etal. 1996). For example they are similar in size to manosused during the Dona Ana phase (ca. A.D. 11001200)in the southern Jornada area of southern New Mexicoand west Texas and those used during the Basketmaker

II period in the Black Mesa region of northern Ari-zona.

In 1998 we recorded 65 metate or metate fragmentson the surface of eight terraces and one rock ring fea-ture. Metate and metate fragments are more commonthan manos, with as many as 20 metate pieces occur-ring on one terrace. Combining the 1997 and 1998metates and considering only whole metates, 55 wholemetates have been recorded on 22 terraces and onerock ring, with as many as seven occurring on a singlefeature. All of the whole metates were made out oflocal basalt, although a few rhyolite metate fragmentsare present. Fifty-five percent of these were basinforms, 44 percent were slab forms and 1 percent werea combination slab and basin. The basin metates arelarge, with overall dimensions averaging 60 cm x 47cm x 17 cm. The deeply worn basins have steeplyangled sides that form almost a V-shape with deepwear that averages 8.2 cm (n=30) which indicates in-tensive food grinding activity over many generations.The slab metates are much smaller, averaging 32 cmx 29 cm x 12 cm (n=24), tend to be only lightly worn,and show no signs of resharpening or shaping. Clearlythe slab metates were used far less intensively thanthe basin metates, probably for grinding a differentfood. It appears that there are at least two types ofmetate-mano sets. The spheroid to convex manos wereapparently were used on the heavily worn, large basinmetates while the flat manos were used on the smaller,lightly used slab metates. However, at this time we donot understand the specific functions of each type ofset.

Aerial Photography and Mapping

In 1998 we contracted with Cooper Aerial Photogra-phy, SA for low-level aerial photography (nominalscale: 1:8000) at five of the cerro de trincheras sitesin the Rio Casas drainage. The five sites were CerroJuanaquea, El Canelo, Cerro los Torres, Cerro LaCruz, and Cerro Vidal (see Figure 2). Before the pho-tography we established between five and ten groundcontrol points at each site. The control points weremarked on the ground and their vertical and horizon-tal relationships were measured using a total stationor electronic distance measuring (EDM) equipment.

11

Livestock and thunderstorms destroyed five of the sixcontrol points at Cerro la Cruz before the photogra-phy was taken. However, at the other four sites theground control points are visible in the aerial photo-graphs and development of stereographic models cantake place without any further fieldwork. The aerialphotography was taken in July 1998.

Cooper Aerial Survey Co. developed an orthophotoand topographic map from the 1:8000 scale aerialphotographs of Cerro Juanaquea. The orthophotoconsists of a 1:1000 scale mylar print and bluelinepaper copies covering an area approximately 900 m x1000 m in size, centered on Cerro Juanaquea. Wealso obtained a digital CD-ROM copy of theorthophoto (100 mm ground resolution, 500 dpi at1:1000). Many details of the cultural features are vis-ible in these documents. The topographic map alsocovers an area 900 m x 1000 m in size. The mylarversion of this document was printed at a scale of1:2000 with one meter contour intervals, while thedigital copy includes versions with both one-meter andfive-meter contour intervals in AutoCad DWG andASCII point file formats.

We have produced an accurate base map of the entiresite by combining these two data sets (the orthophotoand the topographic map) with the tape-and-compassmap prepared by Roney during Minnis and Whalens19951996 project. The tape-and-compass map wasscanned, and the orthophoto and topographic map wereimported directly into the computer drafting programCorelDraw7. The scale of these three images wasmanipulated until all were identical. We then super-imposed the tape-and-compass map over theorthophoto and adjusted the locations of the culturalfeatures to correspond to their locations as they ap-pear on the orthophoto. A similar process was thencompleted using first the topographic map by itself,and then the orthophoto and topographic map usedtogether. The final result is a computer file which con-sists of high resolution, rectified aerial imagery, topo-graphic contours, and cultural features. These data setsare built on different layers, so we can work with anyof the three along or with any combination oforthophotography, topographic map, and map of cul-tural features. In future years we hope to fine-tune

this map through ground-truthing, and to develop com-parable information for some of the other sites.

Riverine Geomorphological Investigations

A significant component of our research has involvedexploring the geomorphology of the riverine flood-plains below the site in order to answer two questions:First, are there buried Late Archaic villages in the al-luvium? Second, was the Rio Casas Grandes aggrad-ing during the Late Archaic period? At the time of itsoccupation, Cerro Juanaquea probably would havebeen part of a land-use system that included numer-ous sites in diverse ecological settings. Certainly themajor drainages would have been intensively utilized.However, the archaeological survey conducted by P.Minnis and M. Whalen (1994) failed to locate a singleArchaic site along the Rio Casas Grandes, despite thepresence of numerous Viejo and Medio period sites.It is likely however, that Late Archaic sites were bur-ied under riverine alluvium and are no longer visibleon the modern surface.

In southern and southeastern Arizona large Late Ar-chaic farming villages formed along the principaldrainages that were particularly suitable for farmingdue to climatic induced changes in stream behavior(Huckell 1996). Prior to 4000 B.P. increased rainfalland warmer temperatures brought about rapid downcutting and erosion of stream channels. By 3000 B.P.this trend had reversed itself and more modern condi-tions with reduced rainfall and cooler temperaturesprevailed. These conditions resulted in slowly aggrad-ing floodplains. B. Huckell (1996) has proposed thatan aggrading floodplain environment was particularlyconducive to farming as overbank deposits providedexcellent agricultural fields and slightly cooler tem-peratures would have enhanced the effective moistureavailable to plants. Were the conditions that werefound in southern Arizona also present in northernChihuahua?

In order to answer these two questions Lee Nordt, as-sistant professor of geology at Baylor University, ex-amined the floodplains below Cerro Juanaquea tosearch for buried Late Archaic sites and to reconstructthe alluvial history of the both the Rio Casas Grandes

12

and Rio San Pedro as they converge only one km north-west of Cerro Juanaquea. Aerial photographs, maps,and field observations were used to construct a geo-morphic map and valley cross-sections. A total of 18backhoe trenches were excavated in 1997, and in 199816 new trenches were opened and seven of the previ-ous years trenches were reopened to examine the al-luvial stratigraphy. All trench locations were firstapproved by either the land owner or ejido president,as appropriate. The backhoe trenches were also fencedto prevent harm to livestock. The modern surface ofeach trench location was carefully inspected to avoidany surface archaeological remains.

Nordt made detailed stratigraphic descriptions of most ofthese trenches and collected soil samples for humate C14dating. Sediment to be used for pollen analysis was alsoextracted from these humate samples forpaleoenvironmental data. The discussion below incorpo-rates the analysis of the 1997 investigations (Nordt 1998)and preliminary findings from the 1998 data as well as 16radiocarbon dates. Six landforms were identified: theHolocene floodplain alluvium, the Holocene flood terracealluvium (Jorge terrace), the Late Pleistocene flood ter-race alluvium (Oaxaca terrace), a Pleistocene alluvial fan,a bedrock/fan alluvial complex, and bedrock (Figure 8).

The bedrock/fan alluvial complex emanates from a valleynorth of the town of Janos and is Pleistocene in age. The

Figure 8. Geomorphic map of the Rio Casas Grandes and Rio San Pedro.

yyyyyyyyyyyy

yyyyyyyyyyyy

yyyyyyyyyyyy

yyyyyyyyyyyy

yyyyyyyyyyyy

yyyyyyyyyyyy

yyyyyyyyyyyy

yyyyy

yyyyy

yyyyy

yyyyy

yyyyy

yyyyy

yy

yy

yy

yy

yy

yy

yy

yy

yy

yy

yy

yyyyyyyy

yyyyyyyy

yyyyyyyy

yyyyyyyy

yyyyyyyy

yyy

yyy

yy

yy

yy

yyy

yyy

yyy

0 .5 1 2

km

yy

y

Holocene floodplain alluvium

Holocene flood terrace alluvium

Late Pleistocene alluvium

Abandoned late Pleistocene valley

Pleistocene fan alluvium

Bedrock/fan alluvium complex

Bedrock

Cerro Juanaquea

backhoe locality

Janos

p

on

mlr

kq

ji

h ag f

e h

d

b

Rio Casas

Grande

Rio San Pedro

a

Jorge terrace

old San Pedro

Oaxaca terrace

MN

13

distal end of this fan is present on the west side of the RioCasas Grandes valley. The second fan is also Pleistocenein age, although portions have a Holocene alluvial veneer.This alluvial fan is present to the east and west of the RioCasas Grandes valley (see Figure 8).

Three alluvial landforms have been identified. TheLate Pleistocene alluvial terrace is the highest andoldest landform of the Rio Casas Grandes and RioSan Pedro and has been named the Oaxaca terraceand is present on both sides of the Old Rio San Pedro.The Holocene flood terrace is known as the Jorgeterrace and it is below the Oaxaca terrace on the westside of the Rio San Pedro as well as on the west andeast sides of the Rio Casas Grandes. The Holocenefloodplain is the modern alluvial floodplain of the RioCasas Grandes and it is below the Jorge terrace.

Five paleosols have been identified buried under theJorge terrace and within the Old San Pedro Valley aswell as a buried peat layer. Paleosols form during pe-riods of stability when flooding events are infrequentenough to allow the surface to become vegetated andan A soil horizon to form. As such they are one of thebest contexts to locate buried archaeological sites. Thedeposits, from oldest to youngest are: the peat layer,the Antonio paleosol, the Pedro paleosol, theTrincheras paleosol, the Janos paleosol, and the Viejopaleosol. Table 3 lists the radiocarbon dates for eachdeposit and Figures 9 and 10 show the cross-sections.

A peat layer, dating to ca. 9000 radiocarbon years B.P.was identified in the Old San Pedro valley about 2.3m below the modern surface. Peat is a deposit of plantremains which contains little mineral matter. It formsunder constantly water-saturated conditions and theplant remains may not be fully decomposed (Waters1992:31). Shortly after 9000 B.P. the Rio San Pedromigrated to the northeast about one km from the OldSan Pedro valley to its modern location.

The Antonio paleosol was identified in the Old SanPedro Valley about 150 cm below the surface and datesto about 4500 radiocarbon years B.P. The Pedropaleosol was located west of the modern Rio San Pedrochannel, about 3.3 m below the Jorge terrace and itdates to approximately 3600 B.P.

The Trincheras paleosol is the most significant withrespect to this project as it is dated to approximately3100 radiocarbon years B. P. based on two humateradiocarbon dates. This paleosol lies about 2 m belowthe modern surface of the Jorge terrace and probablywould have been the surface that was exposed duringthe occupation of Cerro Juanaquea. Attempts to lo-cate buried archaeological sites in this paleosol havethus far been unsuccessful although the remains of asingle, disturbed hearth, represented by 23 scatteredburned rock fragments and a few small (1 cm) piecesof charcoal. Unfortunately the charcoal contained toolittle carbon to successfully date.

The Janos paleosol dates to approximately 2000 B.P.,based on five radiocarbon dates. It is typically foundbetween about 75125 cm beneath the modern sur-face and was identified within the Jorge terrace andsomewhat deeper in the Old San Pedro Valley. TheViejo paleosol is found between about 170265 cmbelow the modern Rio Casas Grandes floodplain andit dates to approximately 1600 to 1500 B.P., based ontwo radiocarbon dates. This sequence of six databledeposits lying beneath the floodplain offers signifi-cant possibilities of locating buried archaeologicalsites contemporaneous with Cerro Juanaquea as wellas to reconstruct the changing form of the river as wecontinue our analysis and field work.

Trincheras Geomorphological Investigations

Lee Nordt also conducted geoarchaeological investi-gation of the constructed terrace features on CerroJuanaquea for the purposes of better understandingthe terrace formation processes including the identifi-cation of natural, anthropogenic, colluvial, and allu-vial deposits. It was also hoped that his investigationswould help us identify occupational surfaces andevaluate the possibility that floodplain deposits weretransported from the Rio Casas Grandes up to CerroJuanaquea to use for construction. In 1997 and 1998he made detailed stratigraphic descriptions of the ter-race profiles of the terraces and collected samples formicromorphological thin sections and bulk soil analy-ses. In 1997 Nordt described T6 and T167 and in 1998he described T10, T126, T163, R234a, T297, andT387. For comparative purposes he also made test

14

Tabl

e 3.

Geo

mor

phic

Rad

ioca

rbon

Dat

es fr

om S

ampl

es

Sam

ple

No.

aU

nit/P

aleo

sol

BH

TbD

epth

bel

ow

surfa

ce c

m

14C

agec

(yr B

P)

13Cd

()

Mat

eria

l

B-12

3405

mo

dern

1791

-101

210

70-28

.2A

D 16

70, 1

950

AD

16

50-1

690

AD

19

25-1

950

detri

tal c

harc

oal

B-11

0618

Vie

jo4

232-

265

470

50-25

AD

14

35A

D 14

20-1

455

tree

root

(re

jected

)

B-11

0619

mo

dern

624

1-25

159

070

-25

AD

14

00A

D 13

00-1

420

sedi

men

t hu

mat

e

B-11

0617

Vie

jo4

87-9

714

006

0-25

AD

650

AD

62

0-67

5so

il hu

mat

e

B-11

0620

Vie

jo10

168-

178

1520

60

-25

AD

560

AD

46

5-47

5 A

D 51

5-62

0so

il hu

mat

eB

-11

0621

Jan

os

1211

1-12

118

606

0-25

AD

145

AD

90

-240

soil

hum

ate

B-12

3413

Jan

os

911

4-12

419

106

0-19

.3A

D 1

00A

D 55

-160

soil

hum

ate

B-11

0616

Jan

os

1671

-81

1980

80

-25

AD

45

AD

45

-110

soil

hum

ate

B-12

3406

Jan

os

1818

8-19

822

108

0-25

.634

5, 3

10, 2

10 B

C 38

0-16

5 BC

hear

th c

harc

oal

B-12

3412

Jan

os

280

-90

2290

50

-17

380

BC39

5-36

0 BC

280-

250

BCso

il hu

mat

eB

-11

2286

Trin

cher

as2

201-

211

3140

100

-25

1410

BC

1505

-128

5 BC

so

il hu

mat

e

B-12

3409

Trin

cher

as2

282-

292

3330

50

-20

.516

15 B

C 16

70-1

525

BCse

dim

ent

hum

ate

B-12

3411

Pedr

o7(P

robe

)30

0-31

036

308

0-22

.219

65 B

C21

20-2

080

BC

2050

-189

0 BC

soil

hum

ate

B-12

3410

An

ton

io22

146-

156

4540

80

-21

.533

35 B

C33

60-3

090

BCso

il hu

mat

e

B-12

3408

Old

Ped

ro v

alle

y22

215-

225

8950

70

-25

8000

BC

8040

-796

0 BC

pe

at

B-11

2285

terr

ace

122

7-23

714

410

60-20

.3se

dim

ent

hum

ate

Ca

lend

ar Y

ears

e

Line

inte

rcep

t

O

ne-s

igm

a

a B

eta

Ana

lytic

bB

HT

- bac

khoe

tren

ch fi

eld

nos.

c 13

C co

rrec

ted

14C

age

d

-25

.0 v

alue

s are

est

imat

es

e cal

ibra

ted

from

Stu

iver

an

d Re

imer

(199

3)

15

excavations and stratigraphic descriptions of threeshovel test units into the natural hillslope of the hillimmediately south of Cerro Juanaquea. Thin sectionsamples have been submitted for T6, T167, and T163as well as the shovel tests. In addition bulk soil analy-ses that will yield particle size distribution, calciumcarbonate content, and organic carbon samples havebeen submitted for T163, T387, and the shovel tests.Preliminary results of the T6, and T167 geomorpho-logical work were presented in our previous informefor the 1997 work.

Thus far Nordts analyses confirm the general schematicformation sequence described earlier in the Test Unitsection. The micromorphological analysis found chertand quartz grains in the Cerro Juanaquea terraces thatare not indigenous to the hill. There are three possibili-ties that could explain their existence on the hill: 1) trans-ported by aeolian processes from the floodplain; 2) aproduct of stone tool manufacture; 3) sediments may havebeen transported from the floodplain up to the hill by theinhabitants to use in construction. Although we cannotbe certain, the first alternative is considered the mostplausible at this time as the grains are rounded which isconsistent with aeolian or fluvial contexts.

Figure 10. Geomorphic cross-section of the Rio San Pedro.

yyyyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyyyyyy

yyyyyyyyyyyyyy

yyyyyyyyyyyyyy

yyyyyyyyyyyyyy

yyyyyyyyyyyyyy

yyyyyyyyyyyyyy

yyyyy

yyyyy

yyyyy

yyyyy

yyyyy

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyy

yyyy yyyy

yyyyyyyyyy

yyyyy

yyyyy

yyyyy

0meters

1

2

3

4

5

old San Pedro valley

Oaxaca terrace

San Pedro

Jorge terraceJorge terrace

Oaxaca terrace

AAF

E F G

GF

E

E

GF

E

C

BD2

D2D2

D1

2210 802290 50

3630 803140 100

3330 50

14410 604540 808950 70

2000 50 100 150

meters

BHT b

BHT h

yy

loams

sands

gravels

peat

Viejo paleosol

Janos paleosol

Trincheras paleosol

Pedro paleosol

Antonio paleosolStratigraphic units A-G backhoe trench localityBHT x

Medio site

hearth

abandoned channel

Figure 9. Geomorphic cross-section of the Rio Casas Grande.

yyyyyyyy

yyyyyyyy

yyyyyyyy

yyyyyyyy

yyyyyyyy

yyyy

yyyy

yyyyyy

yyyyyy

yyyyyy

yyyyyyyyyyyyyyyyyyyyyy

yyyyyyyyyyyyyyyyyyyy

yyyyyyyyyyyyyyyyyyyy

yyyyyyyyyyyyyyyyyyyy

yyyyyyyyyyyyyyyyyyyy

yyyyyyyyyyyyyyyyyyyyyy

yyyyyyyyyyyyyyyyyyyyyy

yyyyyyyyyyyyyyyyyyyyyy

y

y

0meters

1

2

3

4

5

1980 801860 601910 691400 60 590 70

1520 60

210 70

G

G G

G

G

BHT o

BHT i

BHT h

F

EF

E

F

F

E

D

floodplain

Jorge terrace

CasasGrande

2000 50 100 150

meters

yy

y

loams

Stratigraphic units A-G

BHT x

sands

gravels

bedrock

Viejo paleosol

Janos paleosol

Trincheras paleosol

stratified

backhoe trench locality

Viejo site

hearth

16

Terrace Labor Costs and Function

The 486 terraces and 108 rock rings constructed atCerro Juanaquea are striking and continue to chal-lenge many existing notions concerning the Late Ar-chaic period. Such questions as community size,regional population levels, level of sedentism, role ofagriculture, social and political organization, level ofintergroup conflict, and scale of land modification arenow being reconsidered. One approach to understand-ing some of these issues is through the considerationof the labor needed to construct this site.

Native Americans constructed the terraces by firstpiling the local basalt cobbles to form berms thatbowed out in the center and pinched in at the endsagainst the slope. The pocket between the apex of theterrace and the slope behind was then filled in withsmaller rocks and finally sediment to form almost levelsurfaces that average about 50 square meters in size.Many of the terraces form small groups when 25 ter-races connect end-to-end. In one case about 25 joinedterraces form a 400 meter long alignment along thenorthern, eastern, and southern site perimeter.

The scale of the site clearly indicates that it was notconstructed by the social unit expected for the period,a small foraging band. In fact, sites of this scale havebeen attributed to large agricultural societies with vary-ing levels of sociopolitical complexity. In order tobetter understand the labor and organization neededto construct a terrace we built one during our first fieldseason in 1997. During 1998 we had additional mapsand aerial photos made of the site and we have re-cently conducted further analysis of these data.

In 1997 we selected a hillside near Cerro Juanaqueathat also had a 20 percent slope, dense basalt cobblesand thin soil to construct a terrace. A team consistingof two local men and a member of our team, JosZapata, constructed a terrace that was similar to theprehistoric ones. Sr. Zapata kept a detailed record ofthe labor and materials used so we were able to esti-mate the total cost of constructing a terrace. They builtthe terrace in 65 person hours using about 30 cubicmeters of rock and about 4 cubic meters of dirt. Usingdetailed maps, cross-sections, and computer analysisof the terraces at Cerro Juanaquea we were able to

estimate that the total volume of rock and dirt used tobuild all of the terraces is about 31,500 cubic metersor about 58,000 metric tons. We estimate that the siterepresents about 30 person-years of labor, if modern40-hour work schedules were used.

These energetic estimates indicate that our workersmoved an average of 1 cubic meter of rock and dirtevery 1.9 person hours. This is 914 times faster thanother archaeological building experiments and muchfaster than archaeologists can dig! The reason theseterraces can be built relatively rapidly is that rockswere simply picked up, tossed, or sometimes rolledonly a few feet before being dropped into place. Un-like most traditional building projects, the terrace con-struction involves only slight transport costs and theirassembly consists of no more than dropping or toss-ing the rocks into place (Hard et al. 1999).

We can put this labor investment into perspectivethrough comparisons of labor costs for pithouse andpueblo architecture. This level of work appears to beroughly equivalent to about 135 large Anasazi-stylepithouses (5 m x 5 m x 1.5 m deep) with support postsand roof or a 600 room pueblo consisting of 200 liv-ing rooms and 400 storage rooms. Sites of this magni-tude were generally not constructed in NorthwestMexico or the American Southwest until late in thefirst millennium A.D., 15002000 years after CerroJuanaquea was constructed.

Were the Cerro Juanaquea terraces built for farm-ing? We evaluated the labor costs and potential pro-ductivity of farming these terraces to answer thisquestion. The total flat surface area behind the ter-races is estimated to be approximately 2.5 ha. There-fore it would have taken an astounding 4300 days toconstruct each hectare of planting surface. In contrast,prehispanic agricultural terraces in the Maya regionand modern Mexican terraces are estimated to haveconstruction costs of approximately 50850 persondays per hectare and the specialized Maya raised fieldscost between 9503000 person days per ha. to con-struct. Thus if Cerro Juanaquea terraces were con-structed for farming, their labor cost was more thanfive times the cost per ha expended by the Maya intheir agricultural terraces and roughly similar to theirextremely labor intensive raised fields. Such high ag-

17

ricultural labor costs are only usually invested whenthere are very large populations, such as those foundin complex civilizations. There is nothing to suggestLate Archaic populations even approached this levelof occupational intensity or engaged in such inten-sive agricultural practices. If the total of 2.5 ha of flatsurface found behind all 486 terraces was planted inmaize the harvest would only support about 4 adultsfor a year, assuming each adult could be supportedon approximately .6 ha per year, the farming landarea each Tarahumara Indian needs today (Hard etal. 1999).

It has been suggested that some cerros de trincherasin Arizona were built as garden plots to take advan-tage of the longer growing seasons that exist on thehilltops and upper slopes above the cold air inversionsof the drainages. However, the virtual absence of win-ter rain in the Chihuahuan Desert would make winterfarming an unlikely purpose of the terraces. Also nowater control features are found on the terraces. Itappears that constructing these terraces for farmingwould have been impractical at best. Certainly farm-ing would have been far more successful in the wideRio Casas Grandes floodplain below the site.

It appears that the Cerro Juanaquea terraces wereconstructed as house platforms. The terrace surfacesand excavations reveal ample evidence of householddebris including heavily worn, massive basin metates,manos, projectile points, chipped stone debris, ashysediments, charcoal, charred maize, dense burned andunburned bone, significant quantities of flakes, andother stone artifacts. While no houses have been de-fined, in 1997 we exposed two small postholes andcompact occupational surface in Terrace 6 and in 1998we defined an additional clearly defined occupationalsurface in Terrace 163. However, a domestic functionfor these platforms would not preclude families main-taining small household gardens as part of the resi-dential function.

We have not yet identified any public architecture atCerro Juanaquea although public architecture amongsemisedentary and early sedentary societies is quitecommon. Certainly there are no mounds, constructedpublic platforms, or above-ground ceremonial walledbuildings representing large-scale communal architec-

ture such as those associated with the Paquime,Hohokam, or Trincheras culture sites. Even if we doeventually identify communal architectural featuresit is unlikely that they will have been constructed atsignificantly higher levels of labor effort than the do-mestic terraces. Based on ethnographic analogy andgiven the relative ease that a terrace can be constructedwe would expect that the people who occupied CerroJuanaquea were largely egalitarian. Each family prob-ably constructed their own terrace platforms ratherthan specialized builders.

While family labor groups probably built the terraces,the location of at least some of the terraces wasplanned. The most obvious of these is the encirclingchain of 25 joined terraces that form a site perimeteron three sides. The construction efforts of many dif-ferent families would probably have been coordinatedto achieve this. This suggests that one or more indi-viduals were responsible for site planning and deter-mining where some terraces were to be built. Byapplying what we know of ethnographic situations,we can suggest this leader may have had only limitedauthority.

The evidence suggests the terraces were not con-structed as agricultural features but were constructedas house platforms. Why would Native Americanshave expended the labor to live on a hilltop? Not onlyare the construction costs substantial, but the dailyenergy costs of transporting water, wood, and food upto the hilltop village would not have been insignifi-cant. Certainly there are plenty of locations near theriver that were occupied in later centuries. The long-est-standing explanation for the function of cerro detrincheras, defense, is the explanation we are nowconsidering. Under this scenario people were moti-vated to construct house platforms on the hilltop be-cause raids from competing groups were common. Theterraces did not serve as defensive fortifications, butby merely aggregating on the hilltop the occupantsmaintained an effective defensive posture. The steephillslopes would be difficult to attack, particularly ifthe only weapons were thrusting spears or spears usedwith throwers or atlatls. Given the nature of warfarein egalitarian or nonhierarchical societies, such a de-fense would likely be successful, particularly sincesieges, which cut-off supplies to defenders, were not

18

conducted by egalitarian societies as the attackerswould not have the logistical support sieges require(Keely 1996). Strike and run was the preferred methodof war.

However, accumulating direct archaeological evidencein support of the defense hypothesis is difficult. Wedo not yet have any riverine sites to compare with thehilltop site nor to do we have such direct indicators asburials with evidence of injuries. At the current timewe can speculate that these terraces were constructedas part of a defensive strategy while we continue togather data and examine the issue.

Rock Art

Some additional photographs were taken to providemore detailed documentation of rock art at CerroJuanaquea, Cerro los Torres, and Cerro Vidal. Loca-tions of the rock art panels at Cerro Juanaquea areindicated on Figure 11.

Collection of Modern Plants

One of our principal research objectives is to learnmore about relationships between the prehistoric in-habitants of Cerro Juanaquea and their natural en-vironment. An important first step in realizing thisgoal is documentation of modern vegetation in thevicinity of the site. In September, 1998 ethnobota-nist Dr. Karen Adams made an inventory of plantsgrowing in the Janos area, including collection of63 botanical specimens. The inventory focused onCerro Juanaquea and the adjacent floodplain, riverterraces, and modern agricultural fields. Otherplants were documented at Ojo la Palatoada, a wet-land located a few kilometers west of Janos. Alto-gether we documented 100 taxa from 37 differentfamilies. The specific plants are listed in Appendix4.

Temperature

One of the proposed uses of the terraces was as agri-cultural terraces that were designed to take advantageof warmer temperatures and an extended growing sea-son that may exist on hillsides above the cold air drain-age. In fact, temperature data from Tucson indicatesthat the slopes and top of Tumamoc Hill have a fivemonth longer growing season than the river valleybelow (Fish et al. 1984). In order to evaluate this pos-sibility we placed a transect of eight digital, battery-operated thermal cells in a transect from the Jorgeterrace of the Rio Casas Grandes up the west side ofCerro Juanaquea to the peak of the hill and partiallydown the east side of the hill. These battery-operatedthermometers were set to take a temperature readingevery six hours for one year. When we return in thesummer of 1999 we will download the digital datafrom the thermal cells into a laptop computer so wemay then examine the temperature differences, par-ticularly through the winter months, between the rivervalley and hillslope.

Analytical Results

Radiocarbon Dating

One of the principal objectives of our 1998 excava-tions was recovery of additional material for AMSradiocarbon dating, and this was a principal concernin selecting terraces for testing. We attempted to se-lect terraces from different areas of the site, and be-cause charcoal is better preserved in deeper terracedeposits, we attempted to select terraces with a meteror more of depth. Of the 10 terraces tested at CerroJuanaquea in 1998, eight yielded material suitablefor radiocarbon dating. At both Cerro Vidal and Cerrolos Torres we were also successful in obtaining frag-ments of carbonized maize which yielded radiocar-bon dates that are reported in Table 4.

Table 4 summarizes all of the dating results from 1997,and lists those samples which have been submittedfrom our 1998 season. All of the dated samples from1997 as well as six samples from 1998 (T10, T97,T126, T167, T290, T297) come from terraces built

19

on the summit and upper slopes of Cerro Juanaquea.Two samples (T387, T413) are from terraces near thebase of the hill. The samples from Cerro Vidal and

Cerro los Torres were both excavated from terraceson the upper slopes of these two sites.

?

??

M N

0 500 302010 40 50

ro ck a rt p a ne ls

Figure 11. Locational of rock art panels on Cerro Juanaquea.

20

Chipped Stone

Approximately 3200 pieces of debitage, 20 cores, and38 retouched tools were analyzed from the 1997 and1998 test excavations at Cerro Juanaquea during the1998 season at the field lab by Dr. Brad Vierra of LosAlamos National Laboratory. This analysis includedall of the chipped stone from 11 of the 17 features thathave been tested thus far. In addition, all of the chippedstone artifacts recovered from the 1998 test excava-tions at Cerro Vidal and Los Torres were also ana-lyzed. This material consists of about 240 pieces ofdebitage, two cores and two retouched tools fromCerro Vidal and 280 pieces of debitage, three cores,and three retouched tools from Los Torres.

Several lithic raw material types are represented bythe chipped-stone artifacts at Cerro Juanaquea. Themajority of the debitage is made of local rhyolite, withsome chert, chalcedony, and obsidian. In contrast, mostof the retouched tools were manufactured of chert andchalcedony with less rhyolite and obsidian. Overall,it appears that higher quality materials were gener-

ally selected for tool production and lower qualitymaterials for core reduction, with obsidian being rare.The presence of waterworn cortex on all but the ob-sidian indicates that these materials were obtainedfrom secondary river gravel sources. In contrast, ob-sidian appears to have primarily been obtained fromprimary sources since it exhibits nodular cortex with-out any waterworn rounding.

An in-field analysis was conducted of the gravels inthe Rio Casas Grandes to determine if the materialtypes represented in the Cerro Juanaquea samplewere similar to those available in these gravels. It ap-pears that rhyolite also dominates the local river grav-els. In addition, there are small amounts of chalcedonyand basalt, but no chert was observed. Chert may there-fore have been obtained from a more distant source.An x-ray flourescence analysis of 42 obsidian arti-facts revealed that they were obtained from severalsources (see below). Most of these items were identi-fied as being derived from an unkown source, withless from Lago Fredico, Antelope Wells, Sierra Fresnaland Los Jagueyes. With the exception of the Ante-lope Wells source in southwestern New Mexico, and

Table 4. Summary of AMS Radiocarbon Dating, 1997 and 1998

Provenience Material INSTARR Lab No.

Radiocarbon years BP

Dendrocalibrated 2 age ranges AD

T537 Maize cob 3986 289050 910 (1030) 1250T222 17 Maize cupules 3995 293050 940 (1120) 1270T6 Fouquieria splendens 10056 298070 1020 (1200, 1250) 1380T167 Maize cob 3983 298050 1020 (1200, 1250) 1380T222 Cucurbita 10039 298040 1020 (1200, 1250) 1380T222 Cucurbita 3985 331060 1430 (1530, 1570, 1600)

T10 Maize cupuleT97 2 Maize cupulesT126 Maize cupuleT163 Maize cupuleT290 2 Maize cupulesT297 2 Maize cupulesT387 Maize cupuleT413 Maize grain

T1 Maize cupule 10591 292055 BC 200 (100) 0 AD

T20 Maize cupule 10592 210040 1265 (1120) 930

Cerro Juanaquea (1997)

Cerro Juanaquea (1998)In progress

"

"

"

"

"

"

"

Cerro los Torres (1998)

Cerro Vidal (1998)

21

the undetermined source, the remainder are from ad-jacent areas of Chihuahua.

The stone-tool technology at Cerro Juanaquea ap-pears to emphasize core reduction activities. Thismostly consists of core flakes with few biface thin-ning flakes. Only 20 cores were present in the sample.They were reduced using both a split cobble and cobbleuniface technique. It appears that local river cobbleswere brought on to the site, and then fully reducedinto small cores. Many of these are single platform orcore fragments.

Thirty-eight retouched tools were analyzed in thesample. They mostly consist of bifaces (11), retouchedflakes (7), projectile points (6) and unifaces (6) withfewer notches, denticulates, an endscraper and a pos-sible cruciform. The bifaces and and projectile pointsexhibit evidence of breakage during manufacturing,with some of the points also having been broken dur-ing use. The projectile points exhibit a range in mor-phology including side-notched, stemmed, andcorner-notched forms. Two of the points have bev-eled edges. The unifaces have steep retouched edges

and tend to be manufactured on large thick flakeblanks.

In summary, the chipped-stone technology at CerroJuanaquea emphasizes the core reduction of localrhyolites that were presumably used as expedient flaketools. There is only limited evidence for the produc-tion of retouched tools from local chert, chalcedony,and nonlocal obsidians.

Obsidian Sourcing

Although most of the chipped stone on CerroJuanaquea is material which is available in the im-mediate vicinity, small amounts of obsidian and chertseem to be non-local stone. We do not yet know wherethe chert may have originated, but some progress hasbeen made in identification of the obsidian from CerroJuanaquea. Figure 12 shows the location of knownobsidian sources in northwestern Chihuahua. We sub-mitted 42 samples of obsidian from Cerro Juanaquea,three samples from Cerro los Torres, and three samplesfrom Cerro Vidal to Dr. Steve Shackley, (Archaeo-logical XRF Laboratory, Phoebe Hearst Museum of

Figure 12. Obsidian sources.

N e w M e xic oA rizonaTexa s

C h ihua huaSonora

S ie rraFre sna l

Lag o Barre a l

Los Jag ueye s

S ie rra la B re na

An te lo pe We lls

O jo Fred ric oC e rroJua na quea

R

i o C

a s

a s

G

r a

n

ed

s

R i o

S a

n t a

M

a r

i a

B r

R i o a v o

n d

iR o G r a

e

Ro

i

d

e

l

m

C a

r

e

n

M N

22

Anthropology, University of California, Berkeley) forsource analysis.

Shackley (1998) used an x-ray fluorescence spectrom-eter to quantify 10 trace elements in each sample (Ti,Mn, Fe, Zn, Th, Rb, Sr, Y, Zr and Nb). The resultswere statistically compared to trace element compo-sition from known obsidian sources using the com-puter program SPSS for Windows.

Of the 42 samples from Cerro Juanaquea, 21 (50 per-cent) are from a single unidentified source. The highfrequency of obsidian from this source may imply thatit is located nearby. Eight samples (19 percent) matchobsidian source samples from Lago Fredrico, a poorlydocumented secondary source of obsidian locatedabout 12 km east of Ascensin. Six of the CerroJuanaquea samples (14 percent) match the SierraFresnal source. Shackley believes that this obsidianoriginates in a series of coalesced volcanic domes atthe north end of the Sierra Fresnal, but that it has beenwidely distributed toward the north and east by ero-sion. Four Cerro Juanaquea obsidian samples (9 per-cent) come from the Antelope Wells source. Althoughthe primary source of this obsidian is located 90 kmnorthwest of Janos, this obsidian has also been widelydistributed by erosion and undoubtedly occurs closerto the site in alluvium. The remaining three samplesfrom Cerro Juanaquea match source obsidian fromLos Jageyes, a secondary source located along theRio Santa Maria approximately 10km northwest of Progresso. Theprimary source and distribution ofLos Jageyes are unknown.

The three obsidian samples fromCerro los Torres included two fromAntelope Wells and onefrom Lago Fredrico.The three samples fromCerro Vidal includedtwo from AntelopeWells and one from anunknown source closelyresembling that whichwas well represented atCerro Juanaquea.

Faunal Remains

Archaeological excavations at Cerro Juanaquea inthe summers of 1997 and 1998 produced a large col-lection of unmodified faunal remains and severalmodified fragments of bone. An extensive study ofthe faunal material recovered was undertaken in or-der to assess the composition of the assemblage andits potential for future research. The study is signifi-cant as it represents one of the first quantifiable stud-ies of faunal data from the Late Archaic period innorthern Mexico, as well as one of the first reportsfrom a cerros de trincheras site.

A total of 9,885 bones were recovered during the 1997excavations. Table 5 shows the total number of bonesanalyzed from both excavation and flotation proce-dures, the number of identified specimens (NISP) foreach site, the number of different taxa identified, thenumber of burned bones, and the percentage of theassemblage that showed signs of burning.

The majority of the fauna identified consists of jack-rabbit (Lepus spp.), cottontail rabbit (Sylvilagus spp.),and indeterminate leporids (see Table 6). Addition-ally, a small percent of the sample is artiodactyl, in-cluding pronghorn (Antilocapra americana), muledeer (Odocoileus hemionus), indeterminate deer, andindeterminate artiodactyl remains. The remaining

Sample Type No. Total Bones NISP No. of Taxa

No. burned Percent burned

Excavation 6852 1082 18 3481 51Flotation 3033 80 6 N/A N/A

Table 5. Summary of 1997 Faunal Analysis

Identification Number identified (NISP) Percentage of assemblageFish 39 3.4Reptiles 9 0.8Birds 5 0.4Small Rodents 78 6.7Leporids (rabbits and hares) 987 85Carnivores 4 0.3Artiodactyls 40 3.4

Table 6. Number and Percentage of IdentifiedFaunal Remains from 1997 Excavations

23

identified specimens are rodents including squirrels,pocket mice, deer mice, cotton rats, pocket gophers,and kangaroo rats; carnivores including coyote andbadger; birds including quail, perching birds, and in-determinate species, reptiles including terrestrialturtles and non-venomous snakes, and small, indeter-minate fish that are most likely minnows or suckersbased on the size of the vertebra recovered. Thesepercentages are presented in Table 6 and include theidentified material from excavated contexts and flo-tation procedures.

Preliminary analyses of the faunal remains recoveredfrom the 1998 excavations show patterns similar tothose reported above. Analysis of 486 faunal remainsfrom a single terrace (T163) show similarities in com-position, percentage, and appearance to faunal remainsanalyzed from features identified during the 1997 ex-cavations. The T163 fauna indicate that jackrabbits(Lepus spp.), cottontail rabbits (Sylvilagus spp.), largemammals (including pronghorn [Antilocapraamericana], mule deer [Odocoileus hemionus], andunidentified artiodactyls), and a variety of small mam-mals were important game species. Additionally, asingle turtle carapace fragment (Terrapene sp.) and asmall fish element have been identified from this ter-race. Although the 1998 analysis is preliminary, per-centages are clearly similar to the 1997 excavations(Table 7).

In addition the small faunal material from all 1998heavy fractions has been examined. This materialsugests a greater recovery in number and percentageof fish, small mammals, and reptiles compared withthe 1997 remains. This increase suggests that the con-

tribution of aquatic resources may be more signifi-cant than previously suspected, but that the materialmay be difficult to recover in excavation practices. Ifthis is the case, materials recovered from flotation pro-cedures may be especially important to overall assess-ments of subsistence practices at Cerro Juanaquea.

The faunal remains from Cerro Juanaquea includesimilar taxa as those recovered from other early agri-cultural sites in the southwestern United States andnorthwestern Mexico, possibly indicating a similar setof exploited resources throughout the area. However,a possible difference between Cerro Juanaquea andother sites occupied in the greater Southwest may bein the manner and frequency in which artiodactyls wereexploited since they are not abundant in the assem-blage. At Cerro Juanaquea the ratio to artiodactylsto leporids is lower than at most sites in the South-west or Northwest as they are generally the secondmost abundant taxa (Bayham 1982; Lang and Harris1984; Olsen 1982). For example, artiodactyls are farmore common at sites in southeastern and south-cen-tral Arizona than at Cerro Juanaquea (Bayham 1982;Greene and Mathews 1976; Huckell 1995; Schmidt etal. 1998; Szuter and Bayham 1989).

Of the 1,242 identified specimens at CerroJuanaquea, only 43 are artiodactyl. Twenty-four ofthe artiodactyl remains are indeterminate, a designa-tion based largely on small tooth fragments, 13 areindeterminate deer, one is mule deer and five arepronghorn. There are several possible explanations forthe low artiodactyl relative abundance at CerroJuanaquea in comparison to other sites in the greaterSouthwest.

First, large game, forwhatever reason(s), maynot have been available tothe inhabitants of CerroJuanaquea. But, the factthat the ecological zonescharacteristic of this geo-graphic area were similarto other areas that ad-equately supported artio-dactyl populations rendersthis explanation insuffi-

Identification Number identified (NISP) Percentage of assemblageFish 1 1.25Box Turtle 1 1.25Pocket Mouse 1 1.25Chipmunk 1 1.25Indeterminate Canid 1 1.25Cottontail rabbit 13 16Jackrabbit 59 74Pronghorn 1 1.25Indeterminate Artiodactyl 2 2.5Unidentified 406

Table 7. Summary of Faunal Remains from Terrace 163, 1998 Excavations

24

cient. It is also possible that local artiodactyl popula-tions were impacted by human predation, althoughevidence of resource depression cannot presently beevaluated. Another possible explanation for the lackof artiodactyls is that the inhabitants chose not to usethese resources in favor of small game, culturally notselecting for the use of large game. This alternativeexplanation is unlikely, as it entails people foregoingresources which provide much higher post-encounternet caloric returns than those provided by smaller game(Broughton, 1994a, 1994b; Szuter and Bayham 1989).The final explanation for the near absence of artio-dactyl remains is that these resources were procuredand processed at a location away from the site proper.

Some researchers (Binford 1978; Kent 1991; Schmidt1999; Szuter and Bayham 1989) have discussed special-ized activity areas located at some distance from primaryareas of occupation. These small sites or caves appear tohave been used specifically as animal processing areas,and often leave little in the way of tools or debitage. Ifthe inhabitants of Cerro Juanaquea used similar typesof activity areas, we would not expect abundant artio-dactyl remains at the site. Ultimately, other lines of evi-dence (geomorphological and paleoclimatological), andlarger sample sizes obtained through additional excava-tions and analysis may provide evidence to support oneor a combination of these hypotheses.

Macrobotanical Remains

A principle goal of the 1998 field season was to ob-tain additional botanical remains for both ethnobo-tanical analysis and for radiocarbon dating. Althoughmacrobotanical returns were generally low from the1997 field season, the deeper deposits were the mostproductive. For this reason, the 1998 excavation strat-egy placed emphasis on recovering flotation samplesfrom deposits 50 cm or more below the surface. Infact, individual terraces were selected for testing in1998 based, in part, on their the potential for yieldingdeep deposits.

In 1998, we took 163 float samples during excavationand floated them in the field lab, using the decantingmethod for field flotation. When water is added to thesample, the organic material floats to the surface, and

the water is then poured through a .5 mm geologicalsieve. The light fraction is then carefully removed fromthe screen and allowed to dry. The heavy fraction,containing gravels, was retained to examine for smallfaunal remains and microlithics. Fifty light fractionswere submitted to Karen Adams of the Crow CanyonArchaeological Research Center for analysis. The fol-lowing results are based upon that analysis.

The 1998 flotation samples were far more productivethan the 1997 samples in that ten out of twelve fea-tures excavated produced charred reproductive plantparts. While the 1997 data implied that corn may havebeen less important at Cerro Juanaquea than indig-enous small seeded species, 1998 findings suggest thatmaize (Zea mays) agriculture played a significant rolein the diet. Five charred cob fragments, four cob seg-ments, and six kernel fragments were recovered fromthe screen during excavation in 1998, and similar re-mains, including maize cupules, were found in 42 of50 flotation samples (84 percent, Table 8). Of the flo-tation material, 13 maize cupules and five kernel frag-ments came from Cerro los Torres and Cerro Vidal.The maize is a 12-rowed variety consistent with maizefound at archaeological sites throughout the South-west and northern Mexico, including other Late Ar-chaic sites. The close proximity of the Rio CasasGrandes suggests that the inhabitants practiced flood-plain agriculture.

As shown in Table 8, the next most common plantsrecovered were chenopodium or amaranth (cheno-ams), which occurred in 20 of 50 1998 flotationsamples (40 percent). Other charred seeds found inthe 1997 or 1998 samples include wild gourds(Cucurbita digitata or C. foetidissima type), uniden-tified grasses (Gramineae), bulrush (Scirpus sp. type),chia (Salvia sp. type), plains lovegrass (Eragrostisintermedia type), milk-vetch (Astragalus nutullianatype), globe-mallow (Sphaeralcea sp.), and Monocoty-ledon. Additional taxa found in 1998 that were notrecovered in 1997 include Trianthema, Ferocactus,Physalis, and Euphorbia sp. type. Most of these taxaare known to be potentially economic seed plants, in-dicating that nondomesticated species played an im-portant role in the diet as well.

25

In her preliminary analysis of the cheno-ams, GayleFritz, Associate Professor at Washington University, hasidentified charred amaranth and chenopodium. Whilethe chenopodium from Cerro Juanaquea isnondomesticated, scanning electron microscope workindicates thin seed coats on some of the amaranth speci-mens from Cerro Juanaquea, Cerro Vidal, and Cerrolos Torres, suggesting that at least some of the ama-ranth may represent a domesticated variety. Domesti-cated amaranth has been documented from a numberof Hohokam sites throughout the American Southwestand in Late Archaic contexts at Fresnal Shelter in south-ern New Mexico (Fritz 1984; Tagg 1996). However, ifsome of the amaranth from Cerro Juanaquea proves

to be domesticated, it would be the first documentedcase of domesticated amaranth in northwestern Mexicoat such an early date. These results are preliminary, andDr. Fritz is continuing her analysis.

In addition to charred reproductive parts, flotation andexcavation also produced wood charcoal from plantssuch as mesquite (Prosopis sp. type), saltbush (Atriplextype), Leguminoseae, Compositae type, ocotillo(Fouquieria type), and creosote bush (Larrea type).These taxa are listed in Table 9. The most commonwood charcoal was mesquite, which makes a superiorfuelwood. Also, ocotillo is commonly used as a build-ing material.

Taxa1998 Ubiquity

(from flotation)1997 Ubiquity

(from flotation)Features (both flotation

and screen)Zea mays

0.84 0.05

T10,T97,T126,T163, T222,T290,T297, T387, T413,T537, CV-T20, LT-T1

Cheno-am

0.4 0.2

T6,T10,T97,T126, T163,T222,T387,T537, CV-T20, LT-T1

Gramineae 0.14 0.07 T10,T222,T387,LT-T1Scirpus 0.06 0.02 T10,T222,T387Trianthema 0.04 T10,T387Ferocactus 0.02 T97Salvia 0.02 0.07 T163,T222Monocotyledon 0.02 0.02 T222,T387Eragrostis intermedia 0.06 0.07 T222,LT-T1Physalis 0.06 LT-T1Euphorbia 0.02 LT-T1Astragalus nuttalliana 0.05 T222Cucurbita digitata T222Sphaeralcea type 0.02 T222

Table 8. Charred Reproductive Parts

Taxa 1998 Ubiquity (from flotation)

1997 Ubiquity (from flotation)

Features (both flotation and screen)

Prosopis 0.02 0.22 T6, T163, T222, T387, LT-T1Atriplex 0.02 0.05 T222, T387, LT-T1Leguminoseae 0.02 0.17 T6,T222,T537, LT-T1Compositae type --- 0.02 T222Fouquieria type --- 0.05 R250,T6,T222Larrea type --- 0.1 T222,T537

Table 9. Wood Charcoal

26

Human Remains

For the first time, in 1998 we recovered two or threefragments of human bone from T10. These pieces in-clude a cranial fragment, and a tooth (lower premo-lar). All were recovered from a single level in T10.These remains, although few, suggest that burials arepresent at Cerro Juanaquea.

Pollen Analysis

Pollen samples were submitted for analysis with twoprincipal objectives: 1) recognition of environmentalchange, and 2) identification of economic activitiessuch as cuiltivation and/or processing of food plants.Ten samples were submitted to Dr. Suzanne Fish, Ari-zona State Museum, University of Arizona for analy-sis (Table 10). Five of these samples were from datedsoil horizons or depositional units located at variousdepths in the floodplain (Table 10, Samples 2 through6). Three samples were pollen washes from a stonebowl, a metate, and a mano recovered during excava-tions on Cerro Juanaquea. Another sample was takenfrom midden deposits behind one of the terraces onCerro Juanaquea. The tenth sample was taken fromthe surface of the floodplain, in an old field whichhad been fallow for several years. This sample pro-vides an indication of pollen results which modernconditions are likely to yield.

Results of the pollen analysis are presented in Table11 (Fish 1999). One of the samples had insufficientpollen for analysis (see Table 10, Sample 5 fromBHT16, dated 1980 B.P.). Another of the floodplain

samples (see Table 10, Sample 2 from BHT6, dated590 B.P.) yielded results not directly applicable to lo-cal environmental reconstruction. Although only about600 years old, this sample had high frequencies ofpine, oak, and juniper pollen, species which were al-most certainly not common in the vicinity of Janos atA.D. 1400. Geological evidence indicates that thissample comes from alluvial deposits which were laiddown during a time of rapid deposition. We believethat the pine and oak pollen originated in the headwa-ters of the drainage, was transported in water, and wasthen deposited on the floodplain along with othersupended sediment. Such water transport and deposi-tion of pollen is a well-documented phenomenon. Forthis reason Sample 2 is disregarded in the followingdiscussion.

Prevailing models of paleoenvironmental conditionsin Chihuahua are based primarily on analysis ofpackrat middens (Spaulding and Graumlich 1986; VanDevender et al. 1987; Van Devender 1990) supple-mented by information from other sources (Metcalfeet al. 1997; Krider 1998). All of these models agreethat climatic conditions approximating those of mod-ern times were established by about 4500 B.P. For thisreason we expected that the analyzed pollen wouldreflect plant species which are present in the area to-day. However, it did seem possible that relatively mi-nor changes in climate which could have affectedvegetation might have occurred in the past 3000 years(Krider 1998; Van Devender and Worthington 1977;Van Devender 1995). Human activities may also haveaffected vegetation in ways which could have alteredpollen rain. For example, with increasing cultivationwe might expect increased frequencies of cheno-ams,

composites, and other plants whichgrow in disturbed areas. It alsoseemed possible that centuries offuelwood gathering could have re-duced arboreal species such as cot-tonwood and mesquite.

Pollen samples from soil horizonsburied in the floodplain span the past3000 years, and seem to reflect con-ditions very similar to those whichprevail today in anthropogenic land-scapes. Cheno-ams and composites

Table 10. 1997 Pollen Samples

Number Location Context Age (BP)1 Floodplain Modern Surface, Jorge Terrace 02 BHT6 C5b, Alluvial deposits 5903 BHT10 Ab2, Viejo Paleosol 14204 BHT12 Ab, Janos Paleosol 18605 BHT16 Ab, Janos Paleosol 19806 BHT2 AbK, Trincheral Paleosol 31407 T6 Stone Bowl 3000 (est.)8 T6 Metate 3000 (est.)9 T6 Mano 3000 (est.)10 T167 Midden 3000 (est.)

27

dominate all of these samples, accounting for between51.5 percent and 83 percent of the pollen. It is inter-esting that the frequency of these plants is lowest inthe sample associated with a date of 3140 B.P. (Sample6), which has a corresponding increase in grass pol-len (Gramineae). This sample location may represent

a more mesic microhabitat favoring grass or the gro