Clovis Technology is not Unique to Clovis

ABSTRACT

We previously showed that stone-tool technological attributes thought to be unique to the Clovis period were present in a radiocarbon and OSL dated middle Holocene-age stratum at Goodson Shelter, Oklahoma (Eren et al. 2018a. “Is Clovis Technology Unique to Clovis?” PaleoAmerica 4:202–228). Consequently, we argued that technological attributes alone should not be used to assign assemblages to Clovis times. Huckell, Haynes, and Holliday (2019. “Comments on the Lithic Technology and Geochronology of the Goodson Rock Shelter.” PaleoAmerica 6:131–134) proposed two alternative hypotheses: that material we identified as Clovis-like was not, or that it was Clovis but had been mixed with younger deposits. They called for more information on the Clovis-like assemblage at Goodson, and additional dating of the site's lowest deposits. We provide that information, which confirms that stone-tool technologies ostensibly unique to Clovis were indeed in use in the middle Holocene.

KEYWORDS:

• Clovis
• technology
• caches
• blades
• evolutionary convergence

Acknowledgements

We are grateful to Mark Bateman, Donald Grayson, Charles Frederick, Susan Mentzer, Tammy Rittenour, and Christopher Roos for comments and advice, and to Anna Zherebnenko for help in the 2019 OSL sampling at Goodson Shelter. Fieldwork and laboratory analysis of Goodson Shelter was supported by the Quest Archaeological Research Fund, Department of Anthropology, Southern Methodist University.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 These and other matters are in our detailed report of the excavations at the site, which we had hoped would be completed in time to refer to it in our 2018 article, which would have addressed a number of the questions raised by Huckell et al. (2019). Unfortunately, it was not. That manuscript is completed and being revised for publication (Andrews et al., in revision).

2 Although perhaps less widely known today than in his time, Chamberlin was America’s preeminent glacial geologist of the late 19th/early 20th century. In addition to successive academic appointments (including at the Universities of Wisconsin and Chicago), he was the decades-long Chief of the USGS Glacial Division. Among his many substantial contributions was his extensive mapping in the early 1880s of what he referred to as the “Kettle Moraine,” which he identified as the terminal moraine of “the Second Glacial Epoch” (Chamberlin 1883 – that ‘epoch’ was soon to be known as the Wisconsin period). This was the definitive demonstration that the North American Pleistocene involved multiple glacial episodes (for additional biographic information on Chamberlin, see Meltzer 2015, 88–92).

3 Chamberlin published another version of the paper in the Journal of Geology in 1897, which was reprinted on a couple of occasions. The original (1890) version was reprinted in Science in 1965.

4 We previously (Eren et al. 2018a) reported there were ∼600 projectile points from the site. That was an error in our preliminary tallying.

5 Based on our initial examination, we suspected this might be a crested blade; after further study, we now interpret it as a channel flake. Either way, it is a Clovis look-alike.

6 Two of the blades we originally identified came from levels 106 and 109. They were made of the same lithic raw material as seen in Stratum 1.

7 Huckell et al. (2019) neglect to acknowledge that our discriminant function analyses assigned the Green and Keven Davis blades, and two-thirds of the Dickenson blades, to Clovis (Eren et al. 2018a, table 3), and did so regardless of whether the discriminant functions were based on the Gault or the Green blades (Eren et al. 2018a, table 3).

8 Although their inference as to the age of the Dickenson cache seems reasonable, it is nonetheless just an inference that may or may not be correct.

9 The index of curvature is based on two measures: the length of the plane connecting the proximal and distal end points of a blade, and the maximum perpendicular distance between that plane and the blade’s interior surface of the blade. The index is calculated by taking the ratio of the latter to the former, and multiplying by 100 (Collins 1999, 86–87). Although the index can be 0, we excluded all 0 values from our sample, since it was not always evident in the data we compiled that a cell left blank in the index of curvature column was intended to convey a 0 or missing data.

10 We previously did not separate the basal red sands and underlying cobble bed load into two units. We do so here, as it helps clarify the shelter’s origins and geomorphic history (also, Andrews et al., in revision).

11 Huckell et al. (2019) surmised that Stratum 1 must be thicker than the measured value we stated, on the basis of their assumption that the intervals on the prism pole shown in our Figure 3 (Eren et al. 2018a) were in 50 cm increments. That is wrong: the prism pole increments are in the English system, so each is 12 in. or 30.48 cm. Also, they apparently did not realize that the floor of the excavation shown in the photograph was just atop the basal gravels, so virtually all of Stratum 1 was indeed visible, contra Huckell et al. (2019). Stratum 1, as we stated previously and again here, is ∼18-28 cm in thickness.

12 Ages calibrated in Calib 8.20, based on the IntCal20 calibration curve. Consequently, these calibrated ages differ slightly from what we published in 2018, since those were calibrated based on the IntCal13 calibration curve.

13 That Huckell et al (2019) were skeptical was not unreasonable, since our 2018 paper had not provided sufficient details on the provenience of the OSL samples.

14 Bateman (2020) provides the full details of the analysis and results. Briefly summarizing that report, quartz grains were extracted and cleaned following the procedure in Bateman and Catt (1996). Samples were measured with a Risø DA-15 luminescence reader. Paleodose values were determined using the single-aliquot regenerative (SAR) approach (Murray and Wintle 2000) in which an interpolative growth curve is constructed using data derived from repeated measurements (five) of a single aliquot which has been given various laboratory irradiations. Values from individual grains were only accepted if they exhibited an OSL signal measurable above background and showed good growth with dose, among other criteria (Bateman et al. 2007). Up to 1200 grains were measured for each sample; ∼5% of the grains per sample passed the acceptance criteria. All samples possessed generally good luminescence characteristics with a rapid decay of OSL with stimulation and OSL signals dominated by a fast component and OSL signal that grew well with laboratory dose. Concentrations of naturally occurring potassium, thorium, uranium, and rubidium, the main contributors of dose to sedimentary quartz, were determined by inductively coupled plasma mass spectrometry. These elemental concentrations were converted to annual dose rates using data from Guerin et al. (2011). Calculations took into account sediment grain sizes used, density and paleo-moisture (with present-day moisture applied as the average paleo-moisture level with an uncertainty of ± 5%). The contribution to dose rates from cosmic sources was calculated following Prescott and Hutton (1994, table 2). The calculated dose rates are based on analyses of corresponding bulk sediment samples (one sample for OSL 2019-1 through 2019-5; one for OSL 2019-6). It is assumed that the present-day values reflect those since burial.

Additional information

Notes on contributors

Metin I. Eren

Metin I. Eren is Associate Professor of Anthropology at Kent State University and a Research Associate in Archaeology at the Cleveland Museum of Natural History. He co-directs the Kent State University Experimental Archaeology Laboratory.

David J. Meltzer

David J. Meltzer is Henderson-Morrison Professor of Prehistory at Southern Methodist University. His research interests are in the peopling of the Americas, Paleoindians and paleoenvironments, and the history of American archaeology.

Brian N. Andrews

Brian N. Andrews is an Associate Professor in the Department of Psychology and Sociology at Rogers State University in Claremore, Oklahoma. He has conducted research throughout the Great Plains and Rocky Mountains, examining questions of mobility, settlement, technology, spatial patterning, and social organization.