Dr. Robert Kelly: More Questions on the First Americans

On January 29, Dr. Robert Kelly gave an excellent presentation on the Pleistocene colonization of North America at the CU Boulder campus. His talk addressed several issues in interpreting existing published data on our earliest evidence of human settlement on the American continent. It was, in part, a reality check on what we really know about the very first “Americans”, while also providing a very good update on the most recent information on colonization.

Part of Dr. Kelly’s talk involved countering a theory that was, at least at the time of his presentation, not yet published. While it may seem bit unfair to counter an argument before it‘s been officially presented, Dr. Kelly would have been negligent to discuss North American colonization without addressing the alternate theory to some extent. The theory in question is the Solutrean Hypothesis. The hypothesis suggests that the first North Americans came from Europe and landed on the Atlantic shore. As evidence, its authors note the lack of evidence for Clovis points (the earliest known cultural evidence for humans on this continent) in the Northwest, which should contain the first (and supposedly the most plentiful) evidence if the first colonies did indeed come across the Bearing Strait from Asia. Clovis and European Solutrean technological similarities, coupled with a relatively high instance of Clovis discoveries in the Eastern U.S., suggest a European origin.

Dr. Kelly discussed four potential alternative explanations for the relatively high Eastern density of Clovis:

- The currently very high population in the Eastern U.S. has lead to many more accidental discoveries of Clovis.

- The Eastern U.S. may have been more environmentally productive, leading to a much higher prehistoric, post-colonization population density.

- Greater use of agriculture in the East leads to more accidental tilled field discoveries.

- More archaeological field work in the East may also skew the data.

Mary Prasciunas’ study of fluted point densities showed that the first three of the above possibilities could be demonstrated. (Unfortunately, the greater amount of archaeological field work actually done in the Midwest was not enough to alter the overall balance of the data.) When compensating for these factors, the highest apparent density appears to be in what is now Colorado, actually near the center of the continent. For the most part, Clovis appears to be fairly evenly distributed throughout North America. There was some discussion about the mining process producing a possible increase in accidental discoveries, and Kelly pointed out that compensating for this possibility might only show a more even distribution of Clovis across the U.S.

Kelly also discussed Waters and Stafford’s 2007 re-evaluation of existing Clovis point data, much of which is fairly old and utilizes an outdated carbon dating technique. As a result, only 11 dates were used to determine a date range for Clovis, which produced a potential life span of only 200 - 450 years. Waters and Stafford argued that Clovis was both too short lived and too expansive to have migrated across the country through settlement, therefore Clovis must have been culturally transmitted through existing populations. Mary Prasciunas and Todd Surovell have concluded that it would take at least 750 years for colonists to spread over the North American continent. However, an analysis of date range statistics with limited samples shows that a sample size of 11 within a 750 year total range would most likely give a range somewhere near 230 to 580 years. Dr. Kelly concludes that reasonable accuracy in this application requires at least 100 samples.

Kelly did, however, suggest that a pre-Clovis technology most likely did exist in North America. Although it’s only one example, the Five Mile Cave site contains human coprolites and pre-dates Clovis. Five Mile Cave was discovered exactly where we should find the “real” first colonizers, in the Pacific Northwest. This lead to a discussion on the lack of Clovis finds in rock shelters, which suggests that early colonizers who would most likely have frequently utilized the caves most likely did not use Clovis technology.

Kelly concluded that the Solutrean Hypothesis is definitely not supported by an accurate interpretation of existing Clovis point data. Further, due to a lack of accurate data the Clovis age range is unknown, but most likely considerably wider than the suggested 450 years; and that the rate of rockshelter use does not support the popular Clovis colonization model.

But the main conclusions regarding research design were also quite evident. Kelly’s presentation showed how interpretations based on pre-existing archaeological data can be skewed by assumptions about the nature of the data collected (e.g. that it is represents an evenly distributed geographic sample, that data collection methods are the same and equally reliable over time, etc.), and how small samples can actually bias the data towards a particular interpretation.