Time for Collaboration: A. E. Douglass, Archaeologists, and the Establishment of Tree-Ring Dating in the American Southwest
Nash, Stephen E., Journal of the Southwest
Archaeologist Fay-Cooper Cole of the University of Chicago wrote in 1934 that "chronology is the soul of archaeology," yet at that time the study of temporal relationships within the archaeological record had an exceedingly short pedigree. The first papers that explicitly considered North American archaeological chronology had been published only two decades previously (e.g., Kroeber 1916; Nelson 1914, 1916; Spier 1917a, b). While stratigraphic analysis and artifact seriation help archaeologists establish the relative order of events in a given site or region, these techniques cannot be used to establish absolute, Christian calendar dates for prehistoric sites. Thus, while they were slowly becoming interested in chronological relationships, archaeologists working during the first quarter of the twentieth century had no chronometric techniques with which to date their relative sequences. On June 22, 1929, this situation changed forever as archaeologists working for astronomer Andrew Ellicott Douglass of the University of Arizona (figure 1) discovered tree-ring specimen HH-39 at Whipple Ruin in Show Low, Arizona (figure 2). HH-39 successfully "bridged the gap" between the dated, modern portion and the undated, archaeological portion of Douglass' tree-ring chronology, thereby allowing him to accurately and precisely date for the first time some forty prehistoric sites across the American Southwest (Douglass 1929).
[Figures 1-2 ILLUSTRATION OMITTED]
Archaeologists have long marveled at Douglass' contribution to their discipline (e.g., Bannister 1962; Haury 1935, 1962; Nash 1997), but published histories (Douglass 1929, 1935; Haury 1962; Judd 1962, 1964, 1968) of the development of archaeological tree-ring dating focus on the formally named First, Second, and Third National Geographic Society Beam Expeditions of 1923, 1928, and 1929, respectively, and the effort to derive dates for Neil Judd, who had recently completed excavations at Pueblo Bonito in Chaco Canyon (figure 3). Recent research suggests that these histories oversimplify a complex, comprehensive, multi-institutional, interdisciplinary, and ultimately successful fifteen-year-long effort that involved no fewer than seven such "Beam Expeditions" as well a great deal of laboratory work between 1914 and 1929 (Nash 1997). This paper examines, through the analysis of archived research documents and correspondence,(1) the development of archaeological tree-ring dating in the American Southwest to illuminate archaeologically and historically important events that have traditionally been overlooked.
[Figure 3 ILLUSTRATION OMITTED]
Around the turn of the twentieth century, Douglass became interested in Southwestern tree rings because he wanted to find a natural, terrestrial, long-term record of sun spot activity (Bannister 1963; Douglass 1937; Glock 1933; Webb 1983). By 1904, he had examined a number of tree stumps in the area around Flagstaff, Arizona, and had begun to formulate some of the basic principles of dendrochronology, the science of applying Christian calendar dates to growth rings in trees. Most important among these is the principle of crossdating (Douglass 1909; Stokes and Smiley 1968), which requires that the dendrochronologist "utilize the presence and absence of synchrony from different [samples] to identify the growth rings that may be misinterpreted," such as missing or double rings (Fritts and Swetnam 1989:131). It was not until 1911, however, when Douglass found a characteristic Flagstaff ring-width pattern in tree stumps in Prescott, Arizona, some sixty-seven miles to the southwest, that he realized the full climatological possibilities of crossdating (Douglass 1914, 1921). Douglass reasoned that if trees found in different microenvironments, at different elevations, many air-miles apart showed the same pattern of narrow and wide ring sequences, they must be responding to a common climatic variable. …