Tree rings can pinpoint the source of wood as well as how old it is. This method has now been used to identify the sources of timber used by the Native Americans who constructed the pre-Columbian 'great houses' of Chaco Canyon.
The largest buildings erected by Native Americans in North America before European arrival were those at Chaco Canyon in what is now New Mexico, where the Ancestral Puebloan culture thrived between about AD 850 and 1140 (refs 1, 2). One of the unanswered questions that those buildings pose is: where did the Ancestral Puebloans obtain large logs for their buildings in a desert with few trees? Writing in Proceedings of the National Academy of Sciences, Guiterman et al.3 have used the method of tree-ring sourcing — not dating, but sourcing — to identify the origin as mountain forests more than 75 kilometres away and 1,000 metres higher up.
Modern visitors to Chaco are astonished that Ancestral Puebloan agriculture supported a complex society of thousands of people in a fragile dry environment where few try to farm today. Yet this culture succeeded for centuries, using sophisticated methods of managing water run-off from brief downpours. Their largest buildings, termed 'great houses', rose up to six storeys and contained hundreds of rooms; they remained North America's tallest buildings until steel construction finally permitted them to be topped by Chicago skyscrapers in 1885.
Great houses were built primarily of sandstone masonry, but they depended on wood for the beams of roofs, doors and windows (Fig. 1). Around a dozen great houses were constructed over three centuries, which required huge quantities of wood: about 240,000 trees were used, yielding beams up to 5 metres long and weighing up to 300 kilograms4. But big trees are not abundant near Chaco today, and were probably not in Ancestral Puebloan times either. Furthermore, some trees used in the construction were spruce and fir, which grow only in mountain forests at elevations much higher than Chaco.
Previous studies had identified the wood as coming from the Chuska Mountains and Mount Taylor, which are both more than 75 km in a straight line from Chaco. These origins were determined by comparing the ratios of the strontium isotopes 86Sr and 87Sr in great-house beams with the ratios in trees from local mountain ranges5,6. (The ratio varies between trees from different mountains, depending on differences in the age and mineral content of the underlying rock.) The studies assumed from local palaeoecology that the Chaco Basin itself was unforested during Ancestral Puebloan occupation. But the isotope-analysis results and that assumption were subsequently criticized7.
Guiterman et al. turned to tree-ring sourcing for further evidence. In areas with strongly seasonal climates, tree wood displays annual growth rings. Tree growth, and hence ring thickness, differs from year to year because of annual differences in temperature and rainfall. This process underlies the familiar method of dendrochronology, better known as tree-ring dating: comparing the tree-ring pattern in an archaeological wood sample with the pattern in a tree sampled in a known year. But tree growth can also vary locally in a given year, because of local differences in climate and topography. That fact enables tree rings to be used to identify the source of an archaeological wood sample, by controlling for date and then comparing the sample's tree-ring pattern with patterns of wood from different local sources. Tree-ring sourcing has been used in Europe to identify sources of wood for ships, musical instruments and paintings8, but the method has received less attention in North America.
The authors assembled tree-ring patterns from eight mountains located in a circle around Chaco. The patterns at the sites differed enough for wood to be identified to individual sources. The researchers then compared patterns from those sites with patterns in 170 beams of 6 tree species from 7 great-house structures.
It turns out that the patterns of most beams matched those of trees from the Chuska Mountains (42%) and the Zuni Mountains (28%). Sourcing patterns differed somewhat between tree species. For spruce and fir, the species used in the first strontium-based studies5,6, the source deduced by Guiterman et al. agreed with that previously identified. Interestingly, the main source shifted with time, from the Zuni Mountains before AD 1020 to the Chuska Mountains thereafter. The strontium studies had not sampled modern trees in the Zuni Mountains because of their greater distance from Chaco, but they agree with the tree-ring-sourcing results that the closer Chuska Mountains were the main source of ponderosa pine, the species used in most Chaco beams. No Chaco beams match the ring patterns of isolated stands of ponderosa pine nearby at Chaco's elevation.
Although Guiterman and colleagues' study solves one mystery about Chaco, it brings others into focus. The first concerns the distance of the wood sources. Archaeologically identified roads radiating from Chaco were presumably the transport routes, but how did people without draught animals transport 5-m-long, 300-kg beams down from mountain forests 1,000 m higher than Chaco, and then 75 km across land, while leaving almost no scratches on the wood?
A second quandary stems from our knowledge of other ancient construction processes. To build the enormous tall dome of the medieval cathedral in Florence, Italy, heavy weights were raised using a machine consisting of winches, pulleys, gears and wheels, and turned by oxen9. The Ancestral Puebloans lacked all five of these, so how did they raise long, heavy roof beams to heights of six storeys?
Third, why did the wood source switch from the Zuni Mountains to the Chuska Mountains a century before Chaco was abandoned by its inhabitants? The switch coincided with a building boom at Chaco, when seven new great houses were erected, and pottery and stone tools also began to be imported from the Chuska Mountains. Might the shift have been caused by deforestation of the Zuni Mountains, by the closer location of the Chuska Mountains to Chaco or by social developments in the source areas?
The seasonal climate, dry conditions and good preservation of archaeological wood in much of western North America make the region well suited for tree-ring-sourcing studies. Now that Guiterman et al. have demonstrated the method's value at Chaco, we may hope for many more sourcing studies by archaeologists and historians.Footnote 1
Lekson, S. H. (ed.) The Archaeology of Chaco Canyon (School for Advanced Research Press, 2006).
Plog, S. in Chaco Revisited: New Research on the Prehistory of Chaco Canyon, New Mexico (eds Heitman, C. C. & Plog, S.) 3–29 (Univ. Arizona Press, 2015).
Guiterman, C. H., Swetnam, T. W. & Dean, J. S. Proc. Natl Acad. Sci. USA http://dx.doi.org/10.1073/pnas.1514272112 (2015).
Betancourt, J., Dean, J. & Hull, H. Am. Antiq. 51, 370–375 (1986).
English, N. B., Betancourt, J. L., Dean, J. S. & Quade, J. Proc. Natl Acad. Sci. USA 98, 11891–11896 (2001).
Reynolds, A. et al. J. Archaeol. Sci. 32, 1061–1075 (2005).
Wills, W. H., Drake, B. L. & Dorschow, W. B. Proc. Natl Acad. Sci. USA 111, 11584–11591 (2014).
Bridge, M. J. Archaeol. Sci. 39, 2828–2834 (2012).
King, R. La Cupola di Brunelleschi (BURR Saggi, 2015).