Pre-Clovis archaeological sites of the Americas 3: Meadowcroft Rockshelter Pennsylvania

Site importance

Meadowcroft Rockshelter in Pennsylvania has become one of the most celebrated archaeological sites in the Americas. The now accepted date of initial occupation of approximately 16000 years BP is the oldest in North America.

Meadowcroft holds the distinction of demonstrating the longest occupational sequence of humans in the Americas^1&2. The numerous occupation layers contained over 2,000 stone flakes and tools, 150 fire pits and 1 million animal and plant remains. A heavily cut and burnt deer antler base, dated to 16,175±975 years BP was the oldest bone found at the site³. Additional finds included basketry, the area's earliest corn maize dating to the 300's B.C.E., the area's earliest squash and pottery, dating back to about 1000 B.C.E.. Brian Fagan former lecturer at both the Smithsonian Institution, and the Getty Conservation Institute says “Meadowcroft is the single most complex archaeological excavation I have ever seen,” said world-renowned author and New World habitation expert Brian Fagan. “It’s a classic example of the very best in stratigraphic observation and meticulous recording in the field. The superb excavation methods give one great confidence in the important evidence for the first Americans found in the bottom layers of the site.”⁴

View of rockshelter during excavation
Photo credit: Simon Fraser University Museum of Archaeology & Ethnology

View of Meadowcroft seen from the opposite side of Cross Creek April 1975

Photo credit: Mark McConaughy¹⁴

Discovery

The story of the discovery of the site’s significance is an interesting one. Albert Miller, a native of Washington County whose family have owned the area around the cave since 1795 was an amateur archaeologist, naturalist, local historian and farmer.

Born in 1911 he witnessed first-hand the revolution in mechanised farming which swept through the United States in those years, it was these huge changes and the loss of traditional ways of life that inspired him to begin documenting rural life. He wrote later that he “…became enamored of the idea of how old rural and farm building[s], farm machines, tools and methods might be preserved. I felt that the day would come when demonstrations of the past would fascinate people of all ages.”⁵

The idea set in motion the creation of the future museum at Meadowcroft and the non-profit Meadowcroft Foundation, of which Delvin and Albert were president and vice president respectively. Incidentally Meadowcroft was conceived of as a name by Albert and his brother, Delvin from a contraction of their two adjacent farm names: Meadowlands and Bancroft.

 
In 1955 whilst out with his dogs Miller noted that a large groundhog hole in the vicinity of what would later become known as Meadowcroft Rockshelter had what appeared to be prehistoric debris in the excavated earth, reaching further in he retrieved a pre-historic flint knife. Further evidence came to light in 1967 when Miller enlarged a Badger hole and found lithic flakes, shells and animal bones. To him this proved the pre-historic occupation of the shelter. He thus reported his find to local archaeologists. Five years later, in 1973, a young University of Pittsburgh anthropology professor named Dr. James Adovasio, visited the farm and was shown the original pre-historic flint knife. Astonished by the find, and realising its significance, Adovasio made an application to the Meadowcroft Foundation (essentially Albert and his brother Delvin) to begin excavations. These begun in the summer of 1973 and continued for six years.

Site description and finds

The Meadowcroft Rockshelter is a cave site c. 48km southwest of Pittsburgh near the village of Avella in Washington County, Pennsylvania. The cave itself is on the north bank of Cross Creek, a tributary of the Ohio River.
The surface rocks of this region are layered sedimentary rocks of Middle to Upper Pennsylvanian Age (Casselman Formation). The predominant lithologies are shale, quartz sandstone, limestone, and coal in decreasing order of abundance.
Geologically, Meadowcroft is located in the unglaciated portion of the Appalachian or Allegheny Plateau, west of the valley and ridge province of the Appalachian Mountains, and northwest of the Appalachian Basin.
Present topography was probably generated during the Pleistocene when increased precipitation and runoff caused extensive downcutting. Since the Wisconsin Glacial boundary only extends southward to northern Beaver County (some 40 odd km north), the Cross Creek Valley and Meadowcroft Rockshelter probably existed in nearly their present configuration well before the close of the Wisconsin, ca. 11,000 B.P1¹.

It is notable that the area was not a glaciated one but within close proximity to it lay the Laurentide Ice Sheet. This undoubtedly affected the paleoclimate to a lesser or greater extent.

  Location of Meadowcroft Rockshelter with respect to the maximum extent of the Laurentide Ice Sheet c. 20000ya. Credit: Adovasio 2002⁶.

The excavation slowly uncovered 11 stratigraphic levels with stratum XI being at the top, or current ground surface, whilst stratum I was found to be the lowest. Finds were made in all levels, including level I. From Adovasio 1975:
“Extent: This, the basal stratum at the site, has been exposed thus far only outside the drip line in a limited section of "deep-hole." It lies directly beneath Stratum II and is presumed to be continuous across the site. The large roof fall on the western edge of the site rests on the surface of this stratum.
Composition: Stratum I consists of badly decomposed, foliated sandstones (possibly roof fall?) interbedded with lenses of charcoal. Compositional analysis of this unit is presently incomplete.
Associations: The upper 1,0 cm of this unit contains 5 lithic artifacts which may represent intrusions from Stratum IIb, otherwise this unit is sterile.”

Crucially sufficient charcoal was discovered for radiocarbon dating. Lithic artifacts (stone flakes) were also found. In this context ‘lithic artifacts’ means the excavator believes that the flakes recovered are of human manufacture. Charcoal lenses can be read as hearths or fires.

 
Of stratum II the report says:
“Stratum II Extent: Thickness of this unit varies from 40-120 cm. The upper 50 cm are distinguished as Stratum Ha which is continuous across the site. IIa is separated from the lower deposits (lIb) by massive roof fall. IIb has been exposed thus far only outside the drip line in the "deep-hole," Stratum II lies above Stratum I and directly beneath Stratum III. Associations: Lithe materials, bone, firepits and fire floors, floral remains, shell.”
 
The lithics and other artifacts excavated contained some fine specimens:

Selected Bifaces from Paleoindian levels. Mungai Knife is the centre biface.

Photo source: Mark McConaughy¹⁴

Lamellar Blades from Paleoindian levels.
Photo source: Mark McConaughy¹⁴

Lithic Assemblage from Meadowcroft Rockshelter.
The now famous Miller Lanceolate point, named for Albert Miller, is at left.
Photo source: Mark McConaughy¹⁴ (Original source: Heinz History Centre)

Large Basketry Fragment photographed in situ at Meadowcroft Rockshelter.
Photo source : Meadowcroft a pre-Clovis site¹⁵ (Original source Meadowcroft Rockshelter and Historic Village Collection. Heinz History Centre)

Corn Cob Fragments from Meadowcroft Rockshelter
Photo source : Meadowcroft a pre-Clovis site¹⁵ (Original source Meadowcroft Rockshelter and Historic Village Collection. Heinz History Centre).

Bone Artifacts from Meadowcroft Rockshelter
A good number of bone artifacts were recovered at Meadowcroft. These were made from the bones of animals the occupants killed for food. The awl far left has been grooved at one end to accept a cord, so that it could be placed around the neck, ready for use in sewing tasks.
Photo source : Meadowcroft a pre-Clovis site¹⁵ (Original source Meadowcroft Rockshelter and Historic Village Collection. Heinz History Centre).

A summary of the finds per stratum were included in Adovasio’s table 2¹:

Note the reference to Ash/Charcoal lenses (2) AND Lithic Concentrations (1) for stratum I



Later in the report Adovasio produces a table (table 3) showing the radiocarbon dates obtained from the Charcoal for most of the early (lower) strata:

 

The shocking part was that FIVE of the radiocarbon ages pre-dated the accepted age for Clovis of about 11500BC! At first reaction was muted, but as Adovasio continued to publish on Meadowcroft^7,8 serious objections began to be raised about the site’s dating. The main objections raised were by Vance Haynes⁷ and centred around the possible contamination of the charcoal samples used to date the site by lignite or coal. Adovasio calmly and consistently rebutted these claims of contamination⁹ and published greater deatil¹⁰. However the attacks just kept on coming. Particularly prominent were the criticisms by Dincauze (1981)¹¹, Kelly (1987)¹² and Tankersley, Munson and Smith (1987)¹³ to name but three. In fact the situation became so bad and relations within the archaeological community so fraught that Adovasio coined the phrase ‘Clovis Mafia’ to describe those constantly making, what he saw as, unwarranted and baseless criticisms. In 1990, Adovasio²  attempted to firmly and finally quash the contamination, paleoenvironmental and stratigraphic issues raise by his critics by publishing new AMS radiometric dates obtained by four different laboratories. The dates are shown in his 1990 figure 1 below:

Original caption reads:
Plot of 52 Meadowcroft Rockshelter radiocarbon dates showing one standard deviation. Four reversals (asterisks) are noted: SI-2056 is actually from the lowest one-third of stratum IIb; SI-2363 is from the upper one-third of stratum IV; SI-2363 is from the upper one-third of stratum IX. Both SI-1681 and 2056 respectively were small heavily diluted samples. Adapted and and expanded from Stuckenrath et al. (1982: Figure 2).


Of particular note are that the dates for the lowest stratum, stratum I are confirmed and retained.

Having dealt with the stratigraphic and paleoenvironmental objections in separate publications^16-18 he turned his attention to the contamination of the radiocarbon samples issue. Adovasio has this to say:

“With regard to particulate contamination, we repeat that there is no coal seam in or near the rockshelter. ..There are small, isolated, discontinuous fragments of vitrinized Pennsylvania-age wood west of the back (north) wall of the rockshelter, but these in situ fragments are circumscribed occurrences and are not represented en bloc within 7m of the hearths that produced the earlies Meadowcroft dates. ..Every radiometric sample from the Pleistocene-age levels was examined for coal particles using both optical and scanning-electron microscopy. No coal particles were ever identified by four radiocarbon laboratories or by independent researchers.. We forcefully reiterate that during these scrutinies, selected lower and middle stratum IIa samples underwent not only reflectance analysis but also paleobotanical examination for Densosporites and other commonly occurring spores of Pennsylvanian-age coal. In all cases, the results were negative..”

Adovasio goes on, acidly, at length about the absurdness of the claims of particulate coal contamination. Given all he had put up with it is worth seeing what he had to say:

“For criticisms about particulate contamination (or stratigraphic perturbation) to beto be credible, the mechanical introduction of vitrinized Pennsylvanian-age wood into earlier Meadowcroft fire pits would have required some unknown (and unspecified) mechanism so precise and of such duration that it nevertheless resulted in consistent stratigraphic order among the early dates. Then about 12,800 years ago, any mechanism that was making the early dates too early would have to have discontinued so that all the successive younger (more recent) dates were valid.. The absence ofany objective evidence for the lack of particulate contamination and lack of any mechanism for the selective injection of contaminants solely into the pre-12,800-year-old samples renders the argument for particulate contamination unconvincing.”

 
In other words, to his critics “You are making fools of yourselves making weak criticisms. Shut up!”.
 
With regard to non-particulate contamination of the charcoal samples, Adovasio says this:

“1. Vitrinized wood cannot be dissolved in ground water..

2. If vitrinite is not the source of any dissolved contaminent, there are no other candidates as the underlying shale (stratum I) that is the basal unit is not carbonaceous.”

 
Here Adovasio is making a direct jibe at Vance Haynes who famously kept a piece of vitrinized wood in a beaker of water on his desk for months.. nothing happened!
 
Adovasio continues:

“3. Assuming the presence of an as-yet-undiscovered potential source of contamination, there is no viable mechanism for its transport as the present water table lies nearly 5m below the deepest occupation surface at the site and it was never higher prehistorically..

4. Most basic to all questions of nonparticulate contamination is a misreading of the available data..”

 
Adovasio explains his detractor’s confusion to them. Tellingly, this matter is never raised again.

 
 Adovasio disavows his earliest 32000 year old cultural materials

Now we come to an extremely important passage regarding the charcoal with associated lithic assemblage, dated to 32000 year ago, obtained from stratum I. Regarding this Adovasio says:

“The last, diminutive sample from the pre-cultural levels at the site was sent for processing in the Oxford AMS system. The sample already had been scrutinized for particulates and Densosporites spores, but the Oxford lab also examined it and detected no contaminants. The solid fraction was extracted with no need to arrest the reaction and subsequently date, as was the soluble fraction. The results for the solid fraction was 31,400 +/- 1.200 years, 29450 BC (OxA-363). The soluble fraction was dated at 30,900 +/- 1,100 years, 28,950 (OxA-364) (see Gillespie et al. 1985:241). These dates conform very closely to the previously calculated Smithsonian lab date of 30,710 +/- 1,140 years 28,760 B.C. for this level, which to repeat, is not associated with any artifacts and apparently dates to well before the initial occupation of the site by humans (Adovasio 1975: 16, Table 3).” See below for more on this issue.

 
Adovasio, then sets out his preferred date for the site, rather tentatively, thus:

“The Meadowcroft dates suggest, rather conservatively, that humans were present at the site perhaps 2,000-3,00 years earlier than the well-established 11,500-year horizon marker”

In other words: My site is proven pre-Clovis beyond reasonable doubt! He continues:

“The frequently cited eighteenth millennium B>C> dates (SI-2060 and SI-2062) were both very small, diluted samples, one of which, SI-2060, has a very high standard deviation of 2,400 years. If the younger range of both of these dates is averaged, then the earliest possible occupation of the site may have occurred ca. 16,770 years ago.”

This date has become the accepted date for site. It seems that the weight of continued criticism had finally worn Adovasio down. He was prepared to be conservative to silence his critics.

 
Unfortunately it wasn’t to be criticisms were STILL raised throughout the 1990’s¹⁹.

It wasn’t until 1999 that an independent study²⁰ finally laid to rest the sceptics last shreds of dissent. Meadowcroft WAS inhabited by pre-Clovis humans 16,770 years ago.

Dissent of a different kind

Mark McConaughy, a member of the field crew that carried out the excavations during 1974 on the earliest levels and who personally collected many of the samples that produced the earliest dates has made some extremely persuasive comments to the effect that Meadowcroft is most likely older than current accepted age¹⁴.

Archaeologist Mark McConaughey excavates on the Rockshelter’s east wall, July 1974.

Photo source : Meadowcroft a pre-Clovis site¹⁵ (Original source Meadowcroft Rockshelter and Historic Village Collection. Heinz History Centre)

 

 

 

Here is what he had to say:

“The early radiocarbon dates from Meadowcroft have been criticized by Haynes (1980, 1991), Kelly (1987), Mead (1980), Tankersley, Munson and Smith (1987), Tankersley and Munson (1992) and others as being the result of various types of coal contamination or incongruities concerning floral and faunal associations. According to these people, either particulate coal or soluable humates and carbon ions from coal percolated through the clays in ground water at the shelter and were embedded in the charcoal collected and radiocarbon dated. This contamination supposedly resulted in the samples producing anomalous dates. These criticisms have also been addressed by Adovasio et al. (1980), Adovasio, Donahue and Stuckenrath (1990) and Adovasio, Donahue and Stuckenrath (1992). Generally, they disagree with the view that coal contamination of the samples resulted in dates that are too old or that the floral and faunal associations do not indicate a Pleistocene occupation at the site.

I personally do not believe any of the radiocarbon dates from the early levels at Meadowcroft are erroneous. I think they properly date the samples submitted and are not contaminated by coal. However, I am hardly an unbiased observer.

The earliest remains from Meadowcroft did not come from excavations within the drip-line of the rockshelter. Most came from an area excavated beside the Old Roof Fall (an immense rock that fell from the top of the shelter prior to anyone living at the site).

Two samples I recovered were dated to ca. 19,000 years ago. Unfortunately, the two 19,000 B.P. samples were not directly associated with stone tools. ..It is for this reason that the bark was only tentatively identified as a basket fragment in Meadowcroft reports, and the dates questioned as being associated with a Paleoindian occupation.”

 

With regard to the very oldest materials recovered from stratum I McConaughy says this:

“Most other charcoal samples taken and dated from the earliest levels, (and dated as 33000 years old) or the "Deep Hole" as it was named by the crew, consisted of pieces of charcoal found with ash and burned earth. These were interpreted as surface hearths that had been put out (i.e., kicked out, spreading some of the ash and charcoal around). Stone tools were found in association with these samples lending support to the interpretation of them being extinguished hearths.”

 

This directly contradicts Adovasio 1990, but is supported by Adovasio 1975. One can only, therefore, conclude that Meadowcroft is probably older than ca. 16,000 years old as currently quoted and most probably 33000 years old.

 

One can only assume that Adovasio chose in 1990 to take the road more travelled by at last, just to get Meadowcroft within the mainstream of accepted sites.

 

One final thought from Adovasio 1975¹: Stratum I consists of badly decomposed, foliated sandstones (possibly roof fall?) - perhaps this indicates that the bedrock was not actually reached and further unexcavated levels remain below…

 

References

1. Excavations at Meadowcroft Rockshelter, 1973-1974: A Progress Report

By J. M. Adovasio, J. D. Gunn, J. Donahue, and R. Stuckenrath

1975. Archaeologist 45(3), pp.1-30.

2. The Meadowcroft Rockshelter Radiocarbon Chronology 1975-1990

J. M. Adovasio, J. Donahue and R. Stuckenrath

American Antiquity Vol. 55, No. 2 (Apr., 1990), pp. 348-354

3. Dr. Richard Shutler, Department of Archaeology, Simon Fraser University. Retrieved 07/12/15 from http://www.sfu.museum/journey/an-en/postsecondaire-postsecondary/meadowcroft

4. Quoted in Tom Imerito   Pittsburgh Quarterly Fall 2012 online

“Ancient dig discovering man in America: Archaeological finds at Meadowcroft Rockshelter turned back the clock on humankind’s first migration to the new world

Retrieved 07/12/15 from http://www.pittsburghquarterly.com/index.php/Discoveries/ancient-dig.html

5. In Albert Miller: Renaissance Man April 27, 2015 by Mark Kelly

Retrieved 06/11/2015 from http://www.heinzhistorycenter.org/blog/discover-meadowcroft/albert-miller-renaissance-man

6. The First Americans: In Pursuit of Archaeology's Greatest Mystery

by James M. Adovasio with  Jake Page. 2002. Random House.

7. Haynes, C. V.1980 Paleoindian Charcoal from Meadowcroft Rockshelter: Is Contamination a Problem? American Antiquity 45:582-587.

8. Mead, J. I. 1980 Is It Really that Old? A Comment about the Meadowcroft Rockshelter "Overview." American Antiquity 45:579-582.

9. Adovasio, J. M., J. D. Gunn, J. Donahue, R. Stuckenrath, J. E. Guilday and K. Volman1980 Yes Virginia, It Really is that Old: a Reply to Haynes and Mead.

American Antiquity 45:588-595.

10. Carlisle, R. C. and J. M. Adovasio, eds. 1982. Meadowcroft: Collected Papers on the Archaeology of Meadowcroft Rockshelter and the Cross Creek Drainage. Department of Anthropology, University of Pittsburgh, Pittsburgh.

11. Dincauze D. F 1981. The Meadowcroft Papers. Quarterly Review of Archaeology :3-4

12. Kelly, R.L. 1987 A Comment on Pre-Clovis Deposits at Meadowcroft Rockshelter. QuaternaryResearch 27:332-334

13. Tankersley, K. B., C. A. Munson and D. Smith. 1987 Recognition of Bituminous Coal Contaminants in Radiocarbon Samples. American Antiquity 52:318-330.

14. Mark McConaughy. Retrieved from:

 http://people.delphiforums.com/MCCONAUGHY/meadowcroft/pbiface1.htm

15.Retrieved from:  http://donsmaps.com/meadowcroft.html

16. Meadowcroft Rockshelter: A 16000 Year Chronicle. In Amerinds and Their Paleoenvironments in Northeastern North America, edited by W. S. Newman and B. Salwen, pp. 137-159. Annals of the New York Academy of Sciences 288. New York.

17. Adovasio et al. 1984. Meadowcroft Rockshelter and the Pleistocene/Holocene Transition in Southwestern Pennsylvania. In Contributions in Quaternary Vertebrate Paleontology: A Volume in Memorial to John E. Guilday, edited by H.H. Genoways and M. R. Dawson pp. 347-369. Special Publications No. 8. Carnegie Museum of Natural History, Pittsburg.

18. Adovasio et al. 1985 Paleoenvironmental Reconstruction at Meadowcroft Rockshelter, Washington County, Pennsylvania. In Environments and Extinctions: Man in Late Glacial North America edited by J. I. Mead and D. J. Meltzer, pp 73-110. Peopling of the Americas Edited Volume Series, Centre for the Study of Early Man, Orono, Maine.

19. Tankersley, K. B. and C. A. Munson 1993 .Comments on the Meadowcroft Rockshelter Radiocarbon Chronology and the Recognition of Coal Contaminants. American Antiquity 57: 321-326.

20. Goldberg, P and Arpin, T. L. 1999. Micromorphological Analysis of Sediments from Meadowcroft Rockshelter, Pennsylvania: Implications for Radiocarbon Dating

Journal of Field Archaeology Vol. 26, No. 3 (Autumn, 1999), pp. 325-342