An Example from Northwestern New Mexico, USA

The Importance of Sample Context in Dendroarchaeological Interpretation: An Example from Northwestern New Mexico, USA

Item Type Article

Authors Towner, Ronald; Grow, Dave; Psaltis, June; Falzone, Alice

Citation Towner, R., Grow, D., Psaltis, J., Falzone, A. 2001. The importance of sample context in dendroarchaeological interpretation: an example from northwestern New Mexico, USA. Tree-Ring Research 57(1):75-88.

Publisher Tree-Ring Society

Journal Tree-Ring Research

Download date 28/09/2021 21:48:08

Link to Item http://hdl.handle.net/10150/262558 TREE -RING RESEARCH, Vol. 57(1), 2001, pp. 75 -88

THE IMPORTANCE OF SAMPLE CONTEXT IN DENDROARCHAEOLOGICAL INTERPRETATION: AN EXAMPLE FROM NORTHWESTERN NEW MEXICO, USA

RONALD TOWNER, DAVE GROW, JUNE PSALTIS, and ALICE FALZONE Laboratory of Tree -Ring Research University of Arizona Tucson, AZ 85721, USA

ABSTRACT Archaeological tree -ring samples provide environmental, chronological, and behavioral information about past human use of the landscape. Such information, however, can only be fully exploited if sample pro - veniences and beam attributes are completely documented. This paper discusses implications of the sample proveniences, beam attributes, and dates from McKean Pueblito, an eighteenth century Navajo site in northwestern New Mexico, USA. Although the date distribution suggests at least three different interpretations of the site construction history, the contextual data indicate that the site was built in AD 1708 and remodeled in 1713. Areal contextual and tree -ring data from McKean Pueblito and other sites in the area are used to discuss larger scale Navajo behavioral and demographic adaptations to the changing physical and social environments of eighteenth century New Mexico. These examples illustrate how sample context at various spatial scales can significantly enhance interpretations of tree -ring data. Keywords: tree -ring dating, dendroarchaeology, Navajo archaeology, Navajo pueblitos, Dinétah, beam attributes.

INTRODUCTION tree -ring samples and dates for evidence of past Archaeological dendrochronology contributeshuman behaviors such as species selection, wood three types of information to the study of the past:harvesting and modification methods, beam stock- chronological, behavioral, and environmentalpiling, structure repair and remodeling, dead wood (Dean 1996). In order to properly interpret threeuse, or other factors. Anomalous dates are typi- types of information contained in archaeologicalcally identified at the feature, room, or structure tree -ring data, however, samples must be evaluatedlevel of analysis. Behavioral information, which is on several contextual levels. Environmental infor- imparted by species distributions, date clusters, mation (tree species exploited and climate recon-dead wood use, beam reuse, and construction, re- structions from archaeological samples) must bepair, and remodeling episodes is also often restrict- related to local and regional biotic and atmospher-ed to room or site contexts (cf. Graves 1982; ic parameters. Chronological information is de-Towner and Dean 1992). With a large sample and rived from sample attributes and tool marks and site database, however, the behavioral context can crossdated archaeological materials, but elicitingbe expanded to address broader anthropological more precise chronological information necessi-questions concerning past human adaptations to tates the identification of "anomalous" dates; thatnatural and cultural environments. The purpose of is, dates that do not identify the event of interestthis paper is to document the chronological and because of exterior ring removal by natural or cul-behavioral information from McKean Pueblito, an tural processes (Dean 1978). Behavioral informa-eighteenth century Navajo site in the Dinétah, and tion is partially incorporated in the environmentaluse previously collected data to discuss the broad- and chronological realms, but is most fruitfully ex-er context of eighteenth century Navajo adaptation ploited by examining the context and attributes ofin the area. In doing so, we demonstrate that ar-

chaeological tree -ring data can provide importantforked -pole hogan villages located in the pinyon - information regarding past human behavior in Di-juniper woodland to a combination of forked -pole nétah and elsewhere, and that sample context is anhogans and sandstone masonry structures, known important aspect of dendroarchaeological analysis.as pueblitos, located on mesa rims, buttes, and boulders (Marshall 1991). Throughout much of the twentieth century, this architectural and settlement BACKGROUND AND PREVIOUS pattern transformation was attributed to a massive RESEARCH influx of Pueblo Indians fleeing the Spanish re- conquest of the Rio Grande area in 1692 -1696 for Although not well defined, the area known as the relative safety of the Navajo country farther the Dinétah, or ancestral Navajo homeland, is lo- north and west. In the past decade, however, the cated around Cañon Largo in northwestern New immigration itself, and resulting impacts on Na- Mexico, USA (Figure 1). It is an area of extreme vajo society, have been questioned on both histor- temperature variation with cold, snowy winters ical (Hogan 1991) and dendroarchaeological and hot, dry summers; most precipitation falls as grounds (Towner 1992, 1996, 1997). winter snow, with the exception of violent mon- Hogan's (1991) reanalysis of historical docu- soon rains that typically occur during July and Au- ments suggests that most Puebloans who emigrat- gust. The area is part of the vast pinyon juniper ed from the Rio Grande went to other pueblo vil- woodland of the Colorado Plateau physiographic lages, such as Hopi and Zuni. Towner (1996) sug- province (Brown 1994). Locally available trees in- gests, on the basis of dendroarchaeological data, clude various species of juniper (Juniperus spp.), that masonry pueblitos were a purely Navajo re- pinyon (Pious edulis), a few ponderosa pine (Pi- sponse to increased Ute raiding, and that the num- nus ponderosa), and widely spaced Douglas -fir ber of Puebloan refugees in the Dinétah has been (Pseudotsuga menziesii), all of which were avail- vastly overstated. Finally, Towner et al. (2000) able for construction and fuel use. have noted chronological and topographic vari- The Dinétah is the location of many events in ability in pueblito and defensive site locations, and Navajo oral traditions, and is considered the home suggest that different threats at different times of many Navajo deities and other holy people prompted different Navajo responses. It is to this (Matthews 1897). It is also the location of the ear- last issue that McKean Pueblito and other eigh- liest known Navajo archaeological sites (Hancock teenth century Navajo defensive sites on Blanco 1997). These factors have made the Dinétah, and Mesa speak most eloquently. the archaeological sites it contains, critical in interpretations concerning the routes and timing of the Navajo entry into the Southwest, and ethno- McKEAN PUEBLITO genesis of the tribe. The entry of the Navajo into the Southwest has been discussed extensively else- McKean Pueblito (LA 112641) is located on the where (Brown 1996; Schaafsma 1996; Wilcoxeast side of Blanco Mesa atop a 336 m2 sandstone 1981) and is beyond the scope of this paper. Al-knoll between Big Rincon and Heifer Canyon in though this paper does not directly discuss NavajoSan Juan County, NM, at an elevation of 2105 m ethnogenesis, the site population dynamics in-(Figure 2). The knoll itself is located on a broad ferred from the tree -ring data have important im-sandstone bench above both canyons, and stands plications for ethnogenetic models. The Navajoapproximately 15 m above the bench. The elevated undoubtedly occupied the area by 1540, and prob-topographic setting of the site provides an expan- ably by the mid -1400s (Brown 1996; Hancocksive viewshed, particularly to the east and along 1997). They abandoned the area in the mid- 1750s, the Largo drainage system. Unlike most, but not and migrated to the south and west (Hester 1962).all pueblitos, McKean Pueblito does not appear to In the late seventeenth and early eighteenth cen- have been associated with a forked -pole hogan. turies, the Navajo changed the style of their hab-Despite significant survey efforts around the knoll itationfrompredominatelywooden,conicalby ourselves and others, no habitation site remains Sample Context in Dendroarchaeology 77

Project Area

New Mexico

0 km 10 Navajo Reservoir

`.A

-1 ( r Blanco l..---- -t Canyon

Blanco Mesa A

Cañon Largo Palluche Canyon f

Figure 1. Map of Dinétah area. have been identified. In addition, there is also veryented in two rows. Rooms 1 -5 are located on the little refuse associated with this pueblito. Thesewest side of the knoll, approximately 1 -2 meters two factors suggest that the pueblito was used in-below the top of the knoll. Rooms 1 -5 have as- frequently and for short durations. sociated wood beams, but only Room 4 has an There are 10 sandstone masonry rooms in theintact roof. Rooms 1 -3 and 5 contain beams in structure (Figure 3), and approximately 51 -75 per- various states of preservation, but none of the cent of the structure is still intact (Copeland et al.beams are in situ. Their inferred function as roof 1995). Nine of the rooms in the structure are ori-primary beams is based on their large size. Rooms 78 TOWNER, GROW, PSALTIS, and FALZONE

Figure 2. Photograph of McKean Pueblito looking northeast (Photograph courtesy of J. Psaltis).

6-10 are identifiable only by rock outlines, and sections from all pieces of wood judged suitable none contain any roofing elements; they are di for crossdating and that exhibited characteristics of rectly on top of the knoll in a row adjacent to a "true outside"; that is, samples whose appear Rooms 1-5 (Figure 3). Rooms 6-9 face Canon ance in the field suggested minimal exterior ring Largo and because there are no extant roof beams loss. There are at least five beams in Room 5 and and no architectural evidence that the rooms were twelve in Room 2 that were not sampled because ever roofed, we infer that they may have served weathering and erosion have removed their exte only as windbreaks or to block light that came rior rings. Of the 27 samples collected, 24 are from from inside the enclosed rooms from view of the the pueblito itself, and three are from metal ax-cut Largo. Room 10 is on the south side of the knoll stumps below the knoll. Towner et al. (1998) have and is not adjacent to the other rooms. It is asso suggested that culturally modified trees and stumps ciated with two circular stone features and a pos around early Navajo sites can provide important sible wall, but its function is unknown. Contrary chronological information pertaining to the site oc to the evidence suggested by the site map (Figure cupations, but at McKean Pueblito this proved not 3), there are no identifiable bond-abut relationships to be true; none of the sampled juniper stumps for any of the rooms, and therefore, the relative dated because their erratic ring series do not match construction sequence for the rooms cannot be in the local master chronology. ferred. Table 1 shows the species, provenience, dating, and attribute information for each sample collect Tree-Ring Data ed. The room provenience information is valuable Dendrochronological sampling of the structure for chronological interpretations (see below), but involved procuring lh-inch cores or wood cross the number of samples from each room is un- Sample Context in Dendroarchaeology

McKean Pueblito LA 112641 Rm 1 Edge of fboulder

"WINDOW'

DNT-1022 DNT-1035 DNT-1036 DNT-102

Rm 4 DNT-102

DNT-1028

DNT-1034

DNT-1033 DNT-1037 DNT-1038 DNT-1032 DNT-1030 DNT-1 31

0 2m

Figure 3. Plan map of McKean Pueblito showing sampled beam locations (based on field map courtesy of J. Copeland). doubtedly skewed by our sampling strategy. Thely from the valley bottom. Thus, the species ex- species sampled at the site include 14 juniper (Jun - ploited indicate tree harvesting in the immediate iperus spp.), 12 pinyon (Pinus edulis), and onevicinity of the site and also from a different en- cottonwood (Populus spp.) Certainly our subjec-vironmental zone in the area. tive sampling criteria influenced the observed spe- Eleven of the 27 samples collected (40.7 %) cies distribution, but data from other pueblitos in-yielded dates, but only two samples (DNT -1030, dicate that pinyon and juniper are the most com-DNT -1033) provided cutting (tree death) dates, mon species used in pueblito construction, and thatand a single sample (DNT -1035) yielded a near cottonwood was used infrequently (Towner 1997). cutting date. The poor dating success rate of ju- The pinyon and juniper trees were undoubtedlyniper (7.1 %) is due to the opportunistic growth procured within a few hundred meters of the site -strategy of many junipers in this area; such erratic hence the presence of metal ax -cut stumps in thegrowth produces multiple false and missing rings, area. The cottonwood timber, however, must have and makes crossdating difficult if not impossible. been procured from a more mesic microenviron-Pinyon, on the other hand, rarely produces false ment, probably either a nearby spring or more like- or locally absent rings, and therefore we were able Table 1. Dates and attributes of collected McKean Pueblito tree -ring samples. 0 DNT -1018-1017NumberSample SpeciesJUNPNN Room 1 RoofLoose Primary Log Provenience Outside Date BrokenBroken, end burned ends Tool Marks SmokeCharred blackened Other Attributes DNTONT -1021-1022-1020-1019 JUNPNNPOP Room 321 RoofLoose Primary logLog 1692vv inc MetalNoneBroken ax ends -cut endend, broken end Weathered DNT -1025-1023-1026-1024 JUNPNN Room 3 RoofLoose Primary Log 1707 +vv inc NoneMetalBroken ax end -cut end Weathered DNTDNT -1029-1028 -1027 PNNJUN Room 4 Roof Primary 1706vv1705 inc +vv inc MetalBroken ax limb shaping,debarking, trim, metallimb limb trim,ax trim, -cut and andend beam beam end end SmokeSmoke blackened blackened DNT -1031-1032-1030 PNNJUN Room 4 Roof Primary 17081713B + comp+v MetalBroken ax limb shaping, trim, andmetallimb beam trim, ax -cut enddebarking, end SmokeNot smoke blackened blackened DNT -1035-1034-1033 JUNPNN Room 4 Roof SecondaryPrimary 17071713G +B comp comp NoneMetal ax debarkingand and beam limb end trim NotForkSmoke smoke blackened blackened DNT -1039-1038-1037-1036 JUNPNN Room 54/54 Roof Door PrimarySecondary Lintel 17071706 +vv inc MetalNoneNone ax limb trim WeatheredNot smoke blackened DNT -1041-1043-1040-1042 JUNPNN AxRoom Ax-cut -cut5 Stump Roof Stump Primary 1704 +vv inc Metal ax shapingcut (flattened) WeatheredFork latewoodringDate near symbols cells.outside (after of sample;Robinson ++ et = al. absent 1974): rings B = necessitate bark present; a ring v = count;cutting comp date in= terminalopinion ofring analyst; on sample vv = containsnoncutting latewood date in band;opinion inc = terminal Metal ax cut of analyst; + = possible locally absent Weatheredring on sample exhibits no Sample Context in Dendroarchaeology 81

to crossdate 10 of the 12 (83.3 %) pinyon samples. single sample is uncertain at best. At McKean Pue- The single cottonwood sample (DNT -1020) lacksblito, only Room 4 retains enough architectural in- ring -width variability, contains only six rings, andtegrity to yield information from smoke blackened was not crossdated. beams. In Room 4, six of the eight primary beams The physical attributes of the samples, in com-exhibit smoke blackening. Sample DNT -1034 bination with the dates of the beams, can be usedlacks smoke blackening, probably because it is the to elicit behavioral and chronological informationouter most beam and has been exposed to weath- from the samples. For example, the presence ofering. As will be discussed below, the absence of metal ax -cut beam ends, limb trimming, and de- smoke blackening on beam DNT -1031, located in barking indicates that the site postdates the Span-the middle of the room, has important implications ish colonization of New Mexico and introductionfor interpreting the tree -ring dates from the room. of metal tools. Thus, the beam attributes and tool marks can be used to broadly date the site to the late seventeenth or early eighteenth centuries evenChronological Interpretations in the absence of absolute chronometric data. The The eleven dates from the site range from presence of many broken beam ends and limbs1692vv inc to 1713B comp, but the only cutting suggests that dead wood was procured for manydates cluster in 1713 and the only near cutting date architectural elements. Interestingly, none of theis 1707 +B comp (`comp' indicates that the ter- beams that exhibit broken ends or limbs producedminal ring on the specimen exhibits a complete absolute dates. Because as trees die a natural death latewood band and `inc' indicates no latewood is they exhibit a higher proportion of locally absentpresent in the terminal ring). All of the timbers and missing rings (Ahlstrom 1985), and because itwere apparently procured prior to the beginning of is easier to break dead wood than living wood, thethe 1714 pinyon growing season. The two samples absence of dates and high frequency of broken el-that date at 1713 (DNT -1030, DNT -1033) were cut ements suggest that dead wood was used frequent- sometime after the 1713 growing season ended but ly at McKean Pueblito. Whether this procurementbefore the initiation of growth in the spring of strategy reflects a paucity of available metal axes,1714. The sampled dated 1708 +B comp (DNT - expedient construction episodes, or some other1035) was cut sometime between the end of the factor is unknown at this time. 1708 growing season and start of the 1710 grow- Other important attributes recorded for the sam-ing season (the " +" indicates a possible locally ples include the "condition" of their surfaces. absent ring near the end of the ring sequence). The Most beams from McKean Pueblito were weath-noncutting dates (1692vv- 1707vv) lack evidence ered to some degree. Beams collected from rooms of true outsides and presumably have suffered with collapsed roofs tend toexhibit extremesome exterior ring loss due to weathering and ero- weathering, and the resultant loss of exterior rings sion. As discussed below, however, the small clus- yields only noncutting dates. The presence or ab- ter of noncutting dates in the first decade of the sence of smoke blackening on beams, the result of 1700s (DNT -1024, 1027, 1029, 1037, 1038, 1039, nondestructive cooking or heating fires in rooms, and 1040) suggests that exterior ring loss was min- was also recorded. Timbers in collapsed rooms ex- imal. posed to weathering rarely exhibit smoke black- The distribution can be interpreted in at least ening because the evidence has been eroded.three different ways depending on the use of sam- Beams in protected areas, such as completelyple contexts. One possible interpretation is that the roofed rooms, retain smoke blackening and canentire site was constructed in the fall of 1713 or contribute important information concerning repairspring of 1714 using a mixture of freshly cut tim- or remodeling episodes and changing room func-bers and possibly some dead wood. This interpre- tion. Sample DNT -1017 from Room 1 has beentation is based on the fact that the latest dates, and extensively charred, which may indicate cata-only cutting dates (1713B comp), from the site are strophic burning of the room, but evidence from afrom in situ roof primary beams in Room 4. Both 82 TOWNER, GROW, PSALTIS, and FALZONE samples exhibit complete terminal rings, which in-ments as roof primary beams is unlikely from an dicates that they were cut after the 1713 pinyonarchitectural perspective as well. Finally, the ab- growing season, but before growth started in thesence of any suitable structures in the area, com- spring of the 1714. Dead wood use may be indi-bined with the abundant construction materials in catedbysample DNT -1032 which datedthe adjacent pinyon - juniper forest, indicate that 1708+ +v. Although the field assessment and lab-beam reuse was minimal, if practiced at all. The oratory analysis suggest that the sample has a truearea surrounding the knoll had enough vegetation outside, the " + +" symbol indicates several lo-to harvest new wood for the construction and cally absent rings near the outside of the samplebringing the wood from other structures would not that necessitated a ring count past the mid- 1680s.have been necessary. As pinyon trees slowly die, the frequency of lo- A third possible interpretation of the tree -ring cally absent rings increases, and Ahlstrom (1985)data is that McKean Pueblito was constructed in suggests that " + +" dates are indicative of dead1707 or 1708 using freshly cut timbers, and that wood use. Room 4 (at least) was remodeled in 1713 using Several factors, however, suggest that this sitefreshly cut timbers and a piece of dead wood interpretation may not be valid. If 1713 is the date (DNT -1032). The inference of a 1707 -1708 con- of initial construction, all but one of the remainingstruction date is based on two 1707 +vv dates, one samples must have consistently eroded betweenfrom Room 4 and one from Room 3,the five and nine rings. We believe it is highly unlikely1708 + +v near cutting date from Room 4, and the that erosional processes removed approximatelygenerally tight clustering of the 1705, 1706, 1707, the same number of rings from all of the remaining and 1708 noncutting dates in Room 4. Although samples, particularly DNT -1024 (1707 +vv) andthese samples do not show evidence of true out- DNT -1040 (1704 +vv) which are located in differ- sides, we believe the clustering indicates tree har- ent rooms. Finally, it must be remembered that thevesting in the same year. The provenience of the "vv" designation indicates that there is no evi-beams in Room 4 support this inference. All of the dence of a true outside on the sample; it does notsamples dated 1705 -1708 are either roof second- indicate that there is unequivocal evidence that ex- aries or primaries located at either end of the room. terior ring loss has occurred (Robinson et al. Interestingly, there is a large hole in the north 1974). Thus, the small cluster of dates in the firstwall of the structure that was initially thought to decade of the 1700s may actually indicate tree har-be a window. Closer analysis of field photographs, vesting at that time, probably in 1707 or 1708. however, suggests that this hole -which opens Another possible interpretation is that the siteonto Room 3 -was created in order to facilitate was constructed in 1713/1714 with freshly cut tim- the replacement of four primary beams (DNT - bers and reused wood from other structures or1030, 1031, 1032, 1033) under the secondary from dead trees. Sample DNT -1035 cannot havebeams (DNT 1035, 1036) in the center of Room been cut in 1713/1714 and in this interpretation4. Digital photographs show that this hole is irreg- must be a reused beam. Sample DNT -1034 did notular in shape and retains neither upper nor lower reveal an outside date, but it did have a forked end. window lintels. Our interpretation is that because The forked end suggests that it may have been useof the use of a different mortar type during the previously used as part of a forked -pole hogan.replacement event, the hole eroded out of the wall Numerous other beams had at least one end bro-after the site was abandoned. These four beams ken, which could only be accomplished once theinclude the two 1713 cutting dates, the piece of wood had been dead for a period of time and dried. dead wood, and an undated sample; the undated Although identified in a few sites, the reuse ofsample may be an even later repair, as it is the beams is not common at pueblito sites (Towneronly roof primary in the center of the room that 1997). Most pueblito masonry rooms needdoes not exhibit smoke blackening. straight, relatively long roof primary beams to The dating of this pueblito resulted in three dif- span the walls; therefore, the reuse of hogan ele-ferent interpretations, and we acknowledge that Sample Context in Dendroarchaeology 83

others are possible. In general, we suggest thesite, and Towner (1997) collected an additional 13 most parsimonious explanation of the dates andsamples. Based on 28 dates, Towner infers that the beam proveniences is that the structure was builtsite was built in the late 1730s and remodeled in in 1708 and Room 4 was remodeled in 1713. Ex-the mid- 1740s. trapolating the data from a single room to the en- The two upland sites southeast of Palluche Can- tire structure is somewhat problematic, but we be-yon, Compressor Station ruin and the Overlook lieve the minimal tree -ring data from Rooms 3 andSite, were both built in the late 1720s. The Over- 5 support our conclusions. look Site consists of a five -room masonry building, four forked -pole hogans, a masonry -based hogan, and a windbreak (Powers and Johnson 1987). McKEAN PUEBLITO IN REGIONAL Towner (1997) collected 10 samples from the site CONTEXT and NLC researchers collected 22 samples (Correll McKean Pueblito is one of five known pueblitosand Kurtz 1957). A total of 27 samples yielded and /or defensive hogan sites located on Blancodates, including 16 cutting or near cutting dates. Mesa or Superior Mesa overlooking Cañon Largo.Towner (1997) indicates that the structure was The other upland sites are the Cottonwood Dividebuilt in the fall of 1727 and extensively remodeled Site (LA 55829), Star Rock Refuge (LA 55838),in the spring of 1728. Interestingly, the windbreak and Pork Chop Pass (LA 5661) on Blanco Mesa,associated with the site yielded a date of 1742 +G. and Compressor Station Ruin (LA 5658) and theCompressor Station Ruin is a four -room pueblito Overlook Site (10732) across Palluche Canyon onassociated with an isolated slab -lined hearth that Superior Mesa. In addition to the upland sites,was probably part of a hogan, a pile of burned rock three valley bottom sites, Foothold Ruin (LAindicative of a sweatlodge, and a forked -pole ho- 9073), Twine House (LA 127737), and 42 Pueblito gan. Towner (1997) collected thirteen samples (NLC- E- CL- UL -U), are located along Palluchefrom the site, all of which yielded dates. He infers Canyon near its confluence with Cañon Largothat the entire structure was built in the spring of (Figure 4). 1728. The Cottonwood Divide site comprises five ho- The valley bottom sites were built during dif- gans, six wall segments, and two log- covered crev- ferent time periods, probably in response to dif- ices (Powers and Johnson 1987). Towner (1997)ferent stimuli. Twine House is a three -room ma- collected 10 samples from the site, and sevensonry structure on a boulder top. Ababneh et al. yielded dates; unfortunately, all of the dated tim-(2000) collected 18 samples from the site, twelve bers are either vv or + +vv dates indicative ofof which provided dates. They suggest that the site exterior ring loss, deadwood use, or both. Usingwas constructed in the late 1720s, abandoned for Ahlstrom's (1985) criterion that the latest noncut-most of the 1730s, and reoccupied in the early ting date from a site that lacks cutting dates may1740s. Based on the Navajo Land Claim investi- indicate the occupation period, Towner (1997)gations in the 1950s, 42 Pueblito consisted of a suggests that the site was occupied shortly aftertwo -room building and two hogans. Ababneh et 1715. Star Rock Refuge consists of a two -roomal.'s (2000) resampling of the site confirmed the masonry building, several associated free standingprovenance of previously collected samples, and walls, and two sealed crevices (Powers and John-provided a new date for the site as well. They sug- son 1987). Towner (1997) collected 11 samplesgest that the site was constructed in the late 1720s, from the site, but only one sample dated. Thisand repaired in 1741 or later. It is possible that the 1708vv date from a loose log suggests the site wassite was abandoned during the 1730s. Foothold occupied after 1708. The Pork Chop Pass site con- Ruin consists of a a masonry room located on top sists of a three- or four -room masonry structure,of a boulder and two additional masonry rooms at three associated middens, and a deteriorated wood-the base of the boulder; at least three forked -pole en corral approximately 50 m east of the pueblito.hogans were recorded by Navajo Land Claim re- NLC researchers collected 15 samples from thesearchers (Correll and Kurtz 1957), but the hogans TOWNER, GROW, PSALTIS, and FALZONE

Ensenada Mesa

Star Rock Refuge

Compressor Pork Station Chop Ruin Pass

Overlook Sites

42 Pueblito

Superior Mesa

0 2 km Twine House

New Mexico

Project Area

Figure 4.Map ofthe Cañon Largo -Palluche Canyon area showing pueblito sites discussed in text. no longer exist. Towner (1997) collected 21 sam-poral distinctions between the pueblitos in this ples from the site, and previous NLC efforts col-area. The three upland sites located in inaccessible lected 18 samples. A total of 30 samples providedlocations on the northern portion of Blanco Mesa dates, but most are noncutting dates. Ababneh et(McKean, Cottonwood Dive, and Star Rock Ref- al. (2000) suggest that the site was built in the lateuge) all date to the first two decades of the eigh- 1730s or early 1740s, but acknowledge that a 1720teenth century. The three upland sites near the cutting date collected from a loose log by the NLCmouth of Palluche Canyon (Compressor Station suggests an earlier occupation is possible. Ruin, Pork Chop Pass, and the Overlook Site) Thus, there appears to be clear spatial and tern-were built in the late 1720s or late 1730s. All of Sample Context in Dendroarchaeology 85 the valley bottom sites (and The Overlook Site)peditions into the Dinétah (Hendricks and Wilson show some evidence of construction in the late 1996). These expeditions typically consisted of a 1720s and repair or remodeling in the early 1740s large group of Spanish soldiers, citizen militia, (Table 1). None exhibit unambiguous evidence ofPueblo Indian auxiliaries, and associated support tree -harvesting activity in the 1730s, and all maypersonnel and livestock. The mobility of such a have been abandoned for some period of time dur- force was necessarily restricted to avenues ame- ing that decade. Apparently, none of these sitesnable to moving large numbers of troops and live- was contemporaneous with the Blanco Mesa up-stock. One such avenue is the floodplain of Cañon land sites; indeed, they may have been built afterLargo, a wide sandy river valley penetrating the the upland sites were abandoned. Dinétah. Pueblitos built overlooking Cañon Largo would have two distinct advantages over sites located nearer the floodplain: (1) increased visibility DISCUSSION along the canyon would have allowed forewarning Jacobson et al.'s (1992) computer analysis in- of an approaching force, and (2) the location above dicates that Cottonwood Divide cannot be seenthe canyon floor would have provided protection from Star Rock Refuge; during our field investi-from a large invading force that could not move gations, however, Cottonwood Divide was clearlymen, equipment, and livestock up steep slopes visible from McKean Pueblito; thus, there is atquickly. A second possible reason for the site lo- least some intervisibility between the contempo-cations on the east side of Blanco Mesa, instead rary upland sites on northern Blanco Mesa. Mc-of the western escarpment farther away from the Kean Pueblito also provides unobstructed views route of Spanish incursions, is related to geology. upstream and downstream along Cañon Largo, andTopographically, Blanco Mesa slopes gently to the probably could be seen from several sites in thatnortheast towards Cañon Largo. Any rainfall on drainage, although no contemporaneous valleythe mesa infiltrates the sandstone bedrock and sites are known at this time. Observation of theflows slowly towards Cañon Largo. Perhaps by lo- locations of the other upland sites indicates that all cating the pueblitos along the downslope edge of of those sites, except the Overlook Site, share thethe mesa, the inhabitants could take advantage of same type of expansive viewshed as McKean. The the supply of water emanating along the edge of valley bottom sites, however, do not have expan- the mesa in the form of more abundant seeps and sive viewsheds; although Foothold Ruin and the springs. Overlook Site can be seen from each other, they In 1716, a truce was negotiated between the do not command views of anything except the val- Spanish and the Navajo (Reeve 1959), effectively ley floor. Interestingly, the three sites northwest ofremoving the threat of Spanish intrusion into the Palluche Canyon with expansive viewsheds dateDinétah. Subsequently, the Navajo may have re- between 1708 and 1715, but the Overlook site andturned to the valley bottoms to take advantage of the three valley bottom sites, which do not havebetter agricultural soils and a more dependable wa- expansive viewsheds, date between 1727 andter source along the stream channels. Although the 1741. Spanish threat was removed, another different We suggest that the northern upland sites (Mc-threat soon materialized. Ute raiding parties intent Kean, Cottonwood Divide, Star Rock Refuge)on the acquisition of Navajo slaves, especially were located on Blanco Mesa for two reasons.children, for sale in the Rio Grande Valley, soon First, following Towner et al. (2000) we interpretthreatened the Dinétah Navajo. Indeed, one of the their locations as defensive strategies designed toobjectives of the Spaniard's peace offering was to counter Spanish entradas. Before 1716, the pri-create a buffer between the Utes and the Spanish mary threat to Navajo security was the Spanish -pueblos of the Rio Grande area. Instead of a large led militia in the Rio Grande Valley. On severalinvading force, however, the Ute threat was com- different occasions during the early eighteenthposed of smaller, highly mobile, horse -mounted, century, the Spaniards sent large, well - armed ex-raiding parties similar to "light cavalry" forces 86 TOWNER, GROW, PSALTIS, and FALZONE

(Towner et al. 2000). The different threat posed byvested and modified living and dead timbers with the Utes resulted in later pueblito construction ina metal ax, and repaired at least one room of the the valley bottoms and mesa tops on elevatedstructure using freshly cut timbers. All of these boulders, a defensive setting suitable to repel the behaviors influenced the tree -ring date distribution smaller, horse -mounted raiding parties. By build- from the site. ing pueblitos on small boulders, and using them The dendrochronologists who investigated the for refuges, the Navajo defenders forced the at-site also influence the date distribution and inter- tackers to dismount, thus losing the advantages ofpretations of the site. By sampling only those speed and height engendered by horses. beams that exhibited characteristics of a true out- An interesting possibility that could be testedside, we ignored timbers that may (however un- with archaeological, but not dendrochronolgicallikely) post- or predate the sampled beams; such data, is that the upland and valley bottom sitesdating would necessitate a revision of the site were constructed by the same Navajo extendedchronology and interpretations. By recording and family group. Ababneh et al. (2000) have sug-photographing the beams, however, we have been gested-that all the sites in Palluche Canyon wereable to identify anomalous tree- ring dates that reoccupied by the same group, and we are temptedsulted from the behaviors of the site occupants. to suggest that the same group used the sites on As discussed above, the date distribution from Blanco Mesa to counter a different threat a decadeMcKean Pueblito can be interpreted in several or more earlier. ways if the beam attributes and provenience are ignored. By recording the exact provenience and beam attributes of samples collected at McKean CONCLUSIONS Pueblito, we are able to infer that Room 4 was Dendrochronological fieldwork in northwesternconstructed in AD 1708 and repaired in 1713. The New Mexico has contributed significantly to ourlocation of the only two cutting dates (1713) ad- understanding of eighteenth century Navajo pop-jacent to the misidentified "window" in Room 4 ulation dynamics. Indeed, our research on McKeanindicates that the roof was repaired, not construct- Pueblito has provided dates linking all the knowned, at that time. Thus, it is the exact provenience pueblito and defensive hogan sites on Blancoof the dated beams, and not simply their temporal Mesa, and differentiating them in time from theplacement, that allows us to reconstruct past hu- later valley bottom sites in Palluche Canyon. Weman behaviors. have suggested that the date and site distributions On a larger scale, the good preservation in can be most parsimoniously explained as early Na- Room 4 allowed us to collect high quality samples vajo adaptations to different social and environ-that yielded the latest dates (and only cutting mental parameters. dates) from the site. Unfortunately, the architecture The main purpose of this paper, however, is toat the site, specifically the bond -abut relationships illustrate the importance of sample context in theof the walls, is too deteriorated for us to infer a interpretation of archaeological tree -ring dates. room specific construction sequence. Nevertheless, Dean (1996) indicates that the most important fac-the tree -ring, architecture, and preservation data tors in a date distribution are the wood -use behav-denote two occupational episodes at the site. These ior of the site occupants and the sampling behaviordates also aid in the interpretation of noncutting of dendrochronolgists. The complex behaviors ofdates from other sites on the mesa, and suggest the eighteenth century Navajos who occupied thethat all the upland sites were absolutely contem- site created the potential and utilized data for thisporary. In the absence of such a regional context, paper; additional archaeological and dendrochro- McKean Pueblito would simply be another mar- nological data still exist at the site. The Navajosginally interesting ruin on the landscape. By plac- chose a site location, selected wood and stone asing the tree -ring dates in their architectural, tem- construction materials (including tree species suchporal, and spatial contexts, however, we are able as juniper and pinyon that can be crossdated), har-to generate testable hypotheses concerning the past Sample Context in Dendroarchaeology 87 human group and individual behaviors that created by R. H. Towner. University of Utah Press, Salt Lake the site and community of which it was a part. City, Utah; pp. 47 -70. On a theoretical level, this dendroarchaeological Copeland, J., P. Gaudy, and S. McKean 1995Laboratory of Anthropology Site Record for McKean investigation of McKean Pueblito illustrates the Pueblito (LA 112641) on file, BLM Farmington. necessity of detailed tree -ring sample context data Correll, J. L., and W. F. Kurtz at the individual beam, room, site, and intrasite 1957Navajo Land Claim project site forms. Ms. on file, scales. It also shows that if dendroarchaeology is Navajo Nation, Window Rock, Arizona. to move beyond simply dating sites and into the Dean J. S. 1978Independent dating in archaeological analysis. In Ad- illumination of past human behaviors, we must ex- vances in Archaeological Method and Theory, Vol- ploit all the environmental, chronological, and be- ume I, edited by M. B. Schiffer, Academic Press, havior information inherent in tree -ring samples New York; pp. 223 -255. and their multiple spatial contexts. 1996Behavioral sources of error in archaeological tree - ring dating: Navajo and Pueblo wood use. In Tree - Rings, Environment and Humanity: Proceedings of ACKNOWLEDGMENTS the International Conference, Tucson, Arizona 17- 21 May 1994, edited by J. S. Dean, D. M. Meko, This research was conducted by the 2000 Tree - and T. S. Swetnam, pp. 497 -503. Radiocarbon, Tuc- Ring Dating in Archaeological Research class at son. the Laboratory of Tree -Ring Research, The Uni- Graves, M. versity of Arizona. Without the support of the Lab- 1982Anomalous tree -ring dates and the sequence of room construction at Canyon Creek Ruin, Arizona. The oratory of Tree -Ring Research, particularly Mal- Kiva 47(3):107 -131. colm Hughes and Tom Swetnam, the field trip Hancock, P. M. upon which this class -based research is based 1997Dendrochronological dates from the Dinétah. Paper could not have been conducted. We gratefully ac- presented at the Protohistoric Conference, Albuquer- knowledge the field and laboratory help of David que, New Mexico. Hendricks, R., and J. P. Wilson J.Street, Mary M. Prasciunus, Beth Bonham, 1996The Navajos in 1705: Roque de Madrid's Campaign James Copeland, Dave Simons, and Denny Bow- Journal. University of New Mexico Press, Albu- den III. Richard V. N. Ahlstrom, Doug Dykeman, querque. Jeff Dean, and Tom Swetnam commented on ear- Hester, J. J. lier versions of this paper and their insights have 1962Early Navajo Migrations and Acculturation in the made it much better. Of course, the authors are Southwest. Museum of New Mexico Papers in An- responsible for any errors of commission or omis- thropology, Volume 6. Museum of New Mexico Press, Santa Fe. sion. Hogan, P. 1991Navajo- Pueblo interaction during the Gobernador Phase: a reassessment of the evidence. In Rethinking REFERENCES CITED Navajo Pueblitos. Cultural Resources Series, Volume Ababneh, L. N., R. H. Towner, M. M. Prasciunas, and K. T. 8. USDI New Mexico Bureau of Land Management. Porter Albuquerque. 2000The Dendrochronology of Palluche Canyon, Dinétah. Jacobson, L., S. Fosberg, and R. Brewley Kiva 66(2):267 -289. 1992Navajo defensive systems in the eighteenth century. Ahlstrom, R. V. N. In Cultural Diversity and Adaptation: The Archaic, 1985The Interpretation of Archaeological Tree -Ring Anasazi, and Navajo Occupation of the Upper San Dates. Ph.D. dissertation, University of Arizona, Juan Basin, edited by L. S. Reed, and P. F. Reed, pp. Tucson. University Microfilms International, Ann Ar- 105 -136. Cultural Resources Series No. 9, Bureau of bor. Land Management, New Mexico State Office, Santa Brown, D. E. Fe, New Mexico. 1994Biotic Communities: Southwestern United States and Marshall, M. P. Northwestern Mexico. University of Utah Press, Salt 1991The pueblito as a site complex: archaeological in- Lake City. vestigations in the Dinétah District. In Rethinking Brown, G. M. Navajo Pueblitos. Cultural Resources Series, Volume 1996The protohistoric transition in the Northern San Juan 8. USDI New Mexico Bureau of Land Management, region. In The Archaeology of Navajo Origins, edited Albuquerque. 88 TOWNER, GROW, PSALTIS, and FALZONE

Matthews, W. Navajo Origins, edited by R. H. Towner, University 1897Navaho legends. Memoirs of the American Folklore of Utah Press, Salt Lake City, Utah; pp. 149 -170. Society 5. 1997The Dendrochronology of the Navajo Pueblitos of Powers, M. A., and B. P. Johnson Dinétah. Ph.D. dissertation, Department of Anthro- 1987Defensive Sites of Dinétah. Cultural Resources Se- pology, University of Arizona, Tucson. University ries, Volume 2. USDI New Mexico Bureau of Land Microfilms, Ann Arbor. Management, Albuquerque. Towner, R. H., and J. S. Dean Reeve, E D. 1992 LA 2298: the oldest pueblito revisited. Kiva 59:315- 1959The Navaho -Spanish peace: 1720's- 1770's. New 331. Mexico Historical Review 34:9 -40. Towner, R. H., H. C. Rogers, and J. M. Copeland Robinson, W. J., B. G. Harrill, and R. L. Warren 2000Changing threats, changing responses: tactical as- 1974Tree -Ring Dates from New Mexico B: Chaco -Gob- pects of the Navajo pueblitos of Dinetah. In Tactical ernador Area. Laboratory of Tree -Ring Research, Sites in the American Southwest, edited by J. Welch, University of Arizona, Tucson. and T. Bostwick. Arizona Archaeologist. in press. Schaafsma, C. E Towner, R. H., L. Sesler, and T. Hovezak 1996Ethnic identity and protohistoric sites in northwestern 1998Navajo culturally modified trees in the Dinétah. In New Mexico: implications for reconstructions of Na- Diné Bikeyah: Papers in Honor of David M. Brugge, vajo and Ute history. In The Archaeology of Navajo edited by M. Duran and D. Kirkpatrick, pp. 195 -209. Origins, edited by R. H. Towner, University of Utah Archaeological Society of New Mexico, Albuquer- Press, Salt Lake City, Utah; pp. 19 -46. que. Towner, R. H. Wilcox, D. R. 1992Dating the Dinétah pueblitos: the tree -ring data. In 1981The entry of the Athabaskans into the American Interpreting the Past: Research with Public Partici- Southwest: the problem today. In The Protohistoric pation, edited by L. Jacobson and J. Piper, Cultural Period in the American Southwest, AD 1450 -1700, Resources Series, Volume 10. USDI New Mexico edited by D. R. Wilcox and W. B. Masse, pp. 213- Bureau of Land Management, Albuquerque; pp. 55- 256. Arizona State University Anthropological Re- 72. search Papers No. 24. Tempe. 1996The pueblito phenomenon: a new perspective on post -revolt Navajo culture. In The Archaeology ofReceived 7/24/00; accepted 2/5/01.