Eleventh-century shift in timber procurement areas for the great houses of Chaco Canyon

Abstract

An enduring mystery from the great houses of Chaco Canyon is the origin of more than 240,000 construction timbers. We evaluate probable timber procurement areas for seven great houses by applying tree-ring width-based sourcing to a set of 170 timbers. To our knowledge, this is the first use of tree rings to assess timber origins in the southwestern United States. We found that the Chuska and Zuni Mountains (>75 km distant) were the most likely sources, accounting for 70% of timbers. Most notably, procurement areas changed through time. Before 1020 Common Era (CE) nearly all timbers originated from the Zunis (a previously unrecognized source), but by 1060 CE the Chuskas eclipsed the Zuni area in total wood imports. This shift occurred at the onset of Chaco florescence in the 11th century, a time with substantial expansion of existing great houses and the addition of seven new great houses in the Chaco Core area. It also coincides with the proliferation of Chuskan stone tools and pottery in the archaeological record of Chaco Canyon, further underscoring the link between land use and occupation in the Chuska area and the peak of great house construction. Our findings, based on the most temporally specific and replicated evidence of Chacoan resource procurement obtained to date, corroborate the long-standing but recently challenged interpretation that large numbers of timbers were harvested and transported from distant mountain ranges to build the great houses at Chaco Canyon.

Keywords: Ancestral Puebloans; archaeology; dendrochronology; human–environment interactions; timber origins.

Fig. 1.

Aerial view of Pueblo Bonito, the largest of the Chaco Canyon great houses. Image courtesy of Adriel Heisey.

Fig. S1.

An example of sourcing a great-house beam via tree-ring-width methods. (A) An individual beam (black line), the ponderosa pine JPB-88 from Pueblo del Arroyo, and the Chuskas chronology (red line). (B) Bivariate plot comparing the ring-width indices of the beam with the source-area chronology. Correlation (r) and t values are provided. (C) The same beam (JPB-88) with the southern Jemez chronology (blue line). (D) As in B, but with JPB-88 and the S. Jemez chronology. In this case, the beam clearly matches better to the Chuska area.

Fig. S2.

Evaluation of dendroprovenance in the San Juan Basin. (A–F) Each tile provides a different test for a set of modern trees (green triangles). Circle sizes are proportional to the number of trees (labeled in the circle) sourcing to a given location. Site information is provided in Table S2; a description of the test is presented in SI Text.

Fig. S3.

Strength of sourcing by (A) probable location and (B) tree species. The t values for individual beams are shown as small dashes, and the long black lines give the mean of each category. The dotted lines provide the overall median t value in each graph. Polygons indicate the general spread of the t values as probability distributions. These are produced by a Gaussian kernel density function with a bandwidth equal to the average SD across all categories in each graph (48).

Fig. 2.

Source locations for great-house timbers (n = 170). The sizes of the green dots are proportional to the percent of beams sourcing to that location; values provide the percentages. Locations of outlier sites and prehistoric road segments come from Mills et al. (43) and Kantner and Kintigh (44), respectively.

Fig. S4.

Sourcing results by tree species. Dot sizes are proportional to the percent of beams sourcing to that location; values provide the percentages. (A) Ponderosa pine, (B) piñon and juniper, (C) Douglas-fir, and (D) spruce and fir.

Fig. 3.

Time series of great-house timber origins. Both plots are on the same time scale. (A) The cumulative sum of beams sourcing to the Zuni (green) and Chuska (blue) Mountains. (B) Distribution of cutting dates by 5-y bins for each source location. The bars are plotted on the center year of each bin.

Fig. 4.

Time series of the Chaco great-house archaeological record. Percentages for chipped stone and pottery represent the amount of Narbona Pass Chert (31) and Chuskan pottery (1) in those records. Curves for construction timbers are the same as in Fig. 3A. Highlighted time periods for construction and importation of pottery come from the Chaco Synthesis Project (1).

Fig. S5.

Temporal distribution of beams with cutting and near cutting outside dates. (A) All great-house timbers that met our four criteria described in the text. (B) Timbers we analyzed. Spruce and fir are grouped because we did not differentiate these genera.

Fig. S6.

Comparison of more conservative subsets of great-house timbers. For each plot, the lines show the cumulative sum of source results for the Chuska Mountains (blue) and Zuni Mountains (green), as in Figs. 3 and 4. The above plots show consecutively more conservative subsets of our results. (A) the full dataset with series ≥30 rings as in the main text, (B) only series ≥75 rings, (C) only beams with t ≥ 6, and (D) beams that have ≥75 rings and t ≥ 6.