A personal perspective on an Outrageous Hypothesis

I first became aware of the idea of some sort of late Pleistocene “cosmic event” in the late Spring of 2007 when, like may other scientists, I heard about a symposium on the topic at the American Geophysical Union meetings in Acapulco where the hypothesis was essentially unveiled to the scientific community. I heard that this “event” was responsible for late Pleistocene extinctions, the demise of the Clovis “culture” and dramatic climate change in the final millennia of the Pleistocene. I was interested in the hypothesis because I’ve spent many years investigating topics that the hypothesis addressed, including Paleoindian archaeology, and late Pleistocene paleoenvironments and paleogeography. I was skeptical because, well, we are supposed to be skeptical of new hypotheses in science. More specifically, however, several things bothered me:

In no particular order...

1) The cause of the late Pleistocene extinctions has been debated for decades. Significant points that often gets lost in the debate are that: i) it is not clear that all or even most of the mammals facing doom in North America survived until the “event” at 12.9k and ii) the extinctions were global and the timing and composition of extincted fauna varied from continent to continent. Invoking a comet to do the deed simply makes no sense.

2) I was pretty familiar with stratigraphic records on the Great Plains and in the Southwest, and very few provided any indication of any particular “event” at 12.9k.

3) I was very familiar with Paleoindian archaeology and geology, especially on the Great Plains, which has the highest concentration of in situ and dated Paleoindian sites on the continent. I knew of no archaeological or stratigraphic “discontinuities” at 12.9k.

4) There was no consideration of the depositional environments that yielded the impact indicators.

5) I was struck by the absence of data or even opinion that cast doubt on the hypothesis. I felt like I was being “stampeded” by the proponents of this hypothesis, and that always raises questions in my mind.

Then, more or less by coincidence, I was going to be in Lubbock, Texas, in the summer of 2007 about the same time as Jim and Doug Kennett, who are part of the Clovis Impact team. We arranged to meet at the Lubbock Lake site, where I’ve worked on both the archaeology and the geology since the 1970s. The Paleoindian stratigraphy at Lubbock Lake is very similar to that at the Clovis site in New Mexico (aka Blackwater Draw site), about 120 km to the northwest. They had some data from Clovis that supported the impact hypothesis. Alan West, another member of the Impact Team, had been to Lubbock Lake earlier and found a lot of magnetic material in the Paleoindian deposits and declared that it, too, had a record of magnetic particles (considered key impact indicators by the Impact Team at the time) that supported their hypothesis. I was skeptical of that claim, too, because magnetic particles are common in all sorts of sediments, including those in and around the Lubbock area. Nevertheless, we agreed that the Lubbock Lake stratigraphy was worth further investigation.


In the Fall of 2007, I proposed collecting samples across the “Impact Zone” at 12,900 years BP at Lubbock Lake and also at the San Jon site. San Jon is a Paleoindian site in one of the thousands of small playa basins that dot the High Plains surface. Based on a lot of field work I conducted during the 1980s and 1990s, I knew that the sediments that fill the playa basins are a homogeneous mud that began accumulating before the time of the purported impact. And most everything that was deposited in the playa basins stayed there. Further, San Jon is only 80 km north of Clovis. If an airfall record of an impact was to be preserved anywhere, it should be in the playa basins.

I also proposed collecting samples from both Lubbock Lake and San Jon, because I had worked out most of the stratigraphy and geochronology at both sites, and would submit them blind to the Impact Team (i.e., they would not know the stratigraphic sequence or radiocarbon age of any of the samples). The idea was 1) to see if we could produce the same results at those sites as they claimed to have from other similar sites, and 2) to see what sort of results the Impact Team would get from these samples. Would they support the Impact Hypothesis? and would their results be the same as ours? Reproducibility of results is a fundamental aspect of hypothesis testing. I also got in touch with Todd Surovell at the University of Wyoming. I knew he had an interest in this hypothesis, and he said he would process samples to look for the magnetic grains and magnetic microspherules using the procedures supplied by the Impact Team. Having two independent teams, one not in any way affiliated with the Impact Hypothesis, working on splits of the same samples was also another important way of assessing reproducibility. Alan West confirmed Todd’s identifications of magnetic microspherules from Lubbock Lake.

During 2008, Todd Surovell began contacting other investigators and obtained samples from other localities; some investigated by The Impact Team, others not previously studied. In the Fall of 2008, I also collected a section at the Clovis site (Blackwater Draw Locality #1) that was previously collected by the Impact Team.


In the meantime, I came across a book by Firestone, West and Warwick-Smith: The Cycle of Cosmic Catastrophes: Flood, Fire, and Famine in the History of Civilization (2006). It was the first publication to present the Clovis Impact hypothesis but had gained little attention until after the AGU conference. When I first picked it up, I casually flipped through it and kept coming across comments that I knew to be misstatements if not grossly in error. For example, my own research on playas was completely misstated, and what were purported to be my conclusions were essentially the opposite of what I said! They state (p. 216) that because I had a suite of radiocarbon dates from “the underlying formation” (i.e., from below the fill in the playa basins, which they mistakenly refer to as “salty salinas”), then these depressions must have formed at about the same time as the Carolina Bays (according to their discussion elsewhere, at about 12.9ka), and they must have been formed by the Impact Event. The problem with this interpretation (as well as interpretations of the Bays - see below) is that my dates were from the playa fills, which is very clear throughout the paper, and therefore the depressions must be older than 12.9k. How could something so clear and straightforward be so stunningly misinterpreted and misrepresented?

See Holliday, V.T., et , 1996. Stratigraphy and Geochronology of Playa Fills on the Southern High Plains. Geological Society of America Bulletin, 108: 953-965.

The dating of the Carolina Bays (thousands of elliptical depressions scattered along the Atlantic Coastal Plain) was also largely ignored or misstated (p. 127). One minute spent searching the internet turned up three GSA abstracts with OSL dates clearly showing that the sand rims around some Bays date to between 15,000 and 40,000 years BP and between 70,000 and 80,000 years BP. Some rims were active “during multiple phases over the past 100,000 years.

See A.H. Ivester et al, 2004. The timing of Carolina bay and inland dune activity on the Atlantic Coastal Plain of Georgia and South Carolina. Geological Society of America Abstracts with Programs 36(5):69

Aerial Photograph of Carolina Bays

But also on p. 127 of The Book, the following statement is presented: “All of the evidence fits our theory that the rims and bays formed all at the same instant [i.e., by an “extraterrestrial event” at around 12.9k]. In support of that, Ivester and coworkers (2003) dated two bay rims to 11,300 and 12,630 years ago using OSL... We used the same technique to date two levels of...Bay rim sand... the [OSL] Dating Laboratory at the University of Washington reported that the ‘highest age (11,400+/-6100 years) is close to the age of Clovis...’”

This passage contains so much misleading and misunderstood information that it is hard to know how to start sorting it out. Luminescence dating produces ages in calendar years, so OSL dates of 11,300 and 12,630 are too young for the “Event” at 12.9k. The mean of the date determined by the Impact Team at UW is far too young for the “Event” but, moreover, is absolutely meaningless given that the standard deviation is over 50% of the mean age! But the grossest distortion is the reference to the work of Ivester et al. In that paper, they clearly state that they are looking at multiple rims formed around some Bays. “Four concentric rims along the margin of one Bay... selected for dating have ages of 35,660+ or -2600; 25,210+ or -1900; 11,160+ or -900; and 2,150+ or -300 years ago...The trend of younger sand rims toward the bay center indicates that the bay has shrunk in area over the last 36,000 years... An additional date of 20,390+ or -1600 years documents eolian reworking of sediment associated with an adjacent bay to the southwest. Another new luminescence date from the Carolina bay rim bordering Arabia Bay in southern Georgia shows the rim was active 12,630+ or -1000 years ago. These dates indicate bay rims were periodically active well after the maximum advance of the Wisconsin ice sheet.” A rather remarkable twisting of words. The dates cited by Ivester et al. clearly do not pertain to the initial formation of any Bay.

See A.H. Ivester et al, 2004. Concentric sand rims document the evolution of a Carolina Bay in the middle coastal plain of South Carolina. Geological Society of America Abstracts with Programs 35(6):169

The Paleoindian archaeological record in the southeastern U.S. is described as “well dated” (p. 113), and I knew that not to be the case. Some components are reasonably well dated, but the older Paleoindian occupations such as Clovis (and Clovis archaeology was an important part of the story) are not dated at all.

The book further claims (p. 113) that “in the Southeast, well-dated [Paleoindian] sites provide good evidence for continuous occupation.” That is simply not true. Further, they quote archaeologist Al Goodyear as saying (p. 113) “... I’m noticing a big drop in the incidence of spear points dating from right after that time” (13,000 years ago). In actuality, numerical age control for post-Clovis Paleoindian artifact styles is almost non-existent.

In looking at the stratigraphy at the Clovis site, they claim that "18 inches" above the "Event" zone is a ledge "jammed with spears, tools, and bone." (p. 73). I’ve spent a lot of time at the Clovis site, much of it involving stratigraphic work. I know of no such “ledge.” Further, "no humans had visited Blackwater Draw for more than 1000 years" (p. 73). There is simply no evidence for that claim whatsoever. The considerable work at the site by my colleague C.V. Haynes apparently was ignored.

After simply paging through the book and seeing all this twisting of scientific data, distortions of scientific fact, and facts easily verified in the published scientific literature, my skepticism began to emerge


In the Fall of 2007, a summary article by Firestone et al. was published in PNAS.

Firestone, R.B., A. West, J. P. Kennett & 23 co-authors, 2007. Evidence for an extraterrestrial impact 12,900 years ago that contributed to the megafaunal extinctions and the Younger Dryas cooling. Proceedings of the National Academy of Sciences 104:16016-16021.

This paper remains the key source in this entire debate because it comes closest to a comprehensive statement of the hypothesis concerning the “YDB” layer (Younger Dryas Boundary). But after reading it, I noted more problems of “facts” and data.

p.16016 (opening sentence of paper): “A carbon-rich black layer, dating to 12.9 ka (12,900 calendar years B.P.)... has been identified by C. V. Haynes, Jr.... at 50 sites across North America as black mats, carbonaceous silts, or dark organic clays.... The age of the base of this black layer coincides with the abrupt onset of Younger Dryas (YD) cooling.”

But a quick glance at the radiocarbon dates in Table 1 of Haynes’ paper shows that base dates for this zone vary widely.

p.16017: “Ten Clovis and equivalent-age sites were selected because of their long-established archeological and paleontological significance, and, hence, most are well documented and dated by previous researchers.”

In fact, very few of these sites could be considered to have “long-established archeological and paleontological significance.”

- The Clovis site and Murray Springs are arguably the only two.

- Morley has no archaeological or paleontological significance

- For Topper the archaeological data are only now emerging.

- Daisey Cave is an important archaeological site, but as indicated in the SI, was not occupied
before 11.5ka

- Gainey is probably an important site, but is poorly published.

- Chobot is very poorly published

- Lake Hind has minimal archaeological significance, no paleontological significance, and is poorly known.

P. 16019: “The YDB at the 10 Clovis- and equivalent-age sites has been well dated to 12.9 ka.“

This is a key point because the hypothesis fundamentally rests on a demonstration that the layers in question with the purported impact markers are all of exactly the same age, or at least as close to “exactly” as modern numerical dating methods (chiefly radiocarbon) can get. But in fact few of the layers are “well dated to 12.9 ka.” (This is clearly indicated in the SI to the PNAS paper). The North American sites are:

- Murray Springs, AZ

- Blackwater Draw, NM No dates directly linked to sampled section

- Daisey Cave, CA

- Wally’s Beach, Alberta “None of the Paleoindian points recovered was in situ and therefore it is not possible to directly link the points with the [dated] faunal remains” (Kooyman et al. 2001 American Antiquity, 687).

- Gainey, MI No Black Mat, no dates, no obvious indication of a 12.9ka level

- Topper, SC No Black Mat and no dates

- Chobbot, Alberta No dates

- Lake Hind, Manitoba

- Morley, Alberta No dates

- 15 Carolina Bays No dates

To summarize: 1) Several of these sites are “dated” by presence of Clovis artifacts, but that provides no precise indication of the 12.9ka level because the Clovis occupation was at least several centuries long. So using the archaeology as an age indicator, given the necessity for precise dating, is circular reasoning. 2) Five of the nine sites (over 50%) have no numerical age control whatsoever or no direct numerical age control on the YDB layers; no age control of any kind is reported for the 15 Carolina Bays (0%).


Firestone et al. (200) PNAS, Figure 1 (Link)

p. 16017: “Each of the 10 Clovis-age sites displays a YDB layer (average thickness of 3 cm).”

This is impossible to verify because sampling intervals and stratigraphic descriptions have never been provided. The comment that the average thickness of the “YDB layer” is 3 cm is significant in light of subsequent critiques of sampling by others (see below). As of June, 2011, the raw data for all samples have yet to become available despite requests to Allen West and Jim Kennett.

p. 16017: “We further suggest that the catastrophic effects of this ET event and associated biomass burning led to abrupt YD cooling, contributed to the late Pleistocene megafaunal extinction, promoted human cultural changes, and led to immediate decline in some post-Clovis human populations.”

The extinction issue is very complicated, and in fact no recently published data shows a synchronous extinction. At the recent (2010) AMQUA meetings, Russ Graham and Tom Stafford (who has also co-authored with some of the Impact proponents) presented a paper with the latest radiocarbon dates showing that most fauna was gone by 12.9k and that some mammoth survived after 12.9. And there is also the work of Gill et al. (2010) showing that mammoth and other herbivores in the Midwest were on the decline long before 12.9k (and that work followed other work in the Northeast showing the same thing [e.g., Robinson et al, 2009]).

Jacquelyn L. Gill, John W. Williams, Stephen T. Jackson, Katherine B. Lininger, Guy S. Robinson, 2009. “Pleistocene Megafaunal Collapse, Novel Plant Communities, and Enhanced Fire Regimes in North America” Science 326:1100-1103.

G. S. Robinson, L. P. Burney, D. A. Burney, 2005 “Landscape Paleoecology and Megafaunal Extinction in Southeastern New York State” Ecological Monographs 75, 295-315

Moreover, the reference to post-Clovis human population decline is based on a two-page paper on Redstone artifacts, which as indicated above are presumed to be post-Clovis in age (Goodyear, 2006), but in fact are not dated at all. A detailed look at the Paleoindian archaeological record provides no evidence for a population decline.

Vance T. Holliday and David J. Meltzer “The 12.9ka Impact Hypothesis and North American Paleoindians” Current Anthropology. V 51, p. 575-585

David J. Meltzer and Vance T. Holliday, “Response to comments on ‘The 12.9ka Impact Hypothesis and North American Paleoindians’” Current Anthropology.

p. 16018: “Charcoal displays peaks in the YDB at eight of nine Clovis-age sites and is present in 15 of 15 Bays, reaching peaks in four Bays with paleosols.”

And from the Supplemental Information: “The Bays have poorly stratified, sandy, elevated rims (up to 7 m) that often are higher to the southeast. All of the Bay rims examined were found to have, throughout their entire 1.5- to 5-m sandy rims, a typical assemblage of YDB markers (magnetic grains, magnetic microspherules, Ir, charcoal, soot, glass-like carbon, nanodiamonds, carbon spherules, and fullerenes with 3He).”

The sandy Bay rims are described as “poorly stratified” and yet some have “buried paleosols.” Which is it? And all Bay rims sampled have YDB markers throughout, including, presumably, the buried soils? What does that mean? The Bay rims can’t be used as evidence if they contain no discrete impact marker layer. The comment suggests that YDB markers can be found outside of discrete contexts, negating their significance.

p. 16019: “At Murray Springs, Haynes... first reported the presence of glass-like or ‘vitreous’ carbon in the black mat. In addition, he chemically analyzed the black mat layer, concluding that it most likely resulted from the decomposition of charred wood and/or a prolonged algal bloom, both of which could result from event-related processes (e.g., climate change and biomass burning). Some black mats have no algal component, only charcoal.”

Haynes clearly describes the black mat as an algal layer (and this is so stated in the SI to the 2007 PNAS paper). How does an algal bloom result from “event-related processes”? Algal blooms occur all the time on the Earth’s surface and almost all in the absence of any extraterrestrial event.

p. 16020: “if multiple 2-km objects struck the 2-km-thick Laurentide Ice Sheet at <30°, they may have left negligible traces after deglaciation… [perhaps] limited to enigmatic depressions or disturbances in the Canadian Shield (e.g., under the Great Lakes or Hudson Bay)”

An obvious flaw with that speculation is that, by 12,900 years ago, only the Lake Superior basin was still under glacial ice, a fact well-known and documented for decades.

Dyke, et al., 2003. Geological Survey of Canada, Open file 1574

16021: “For humans, major adaptive shifts are evident at 12.9 ka, along with an inferred population decline, as subsistence strategies changed because of dramatic ecological change and the extinction, reduction, and displacement of key prey species.”

This was news to me and to any other Paleoindian specialist. But sometimes you don’t see what you are not looking for, so several of us revisited to the Paleoindian literature so see if these claims could have any merit. As noted above, this notion was initially based on Paleoindian artifact data from the Southeast U.S. This was a surprise because there is almost no good stratigraphic or radiocarbon record for Paleoindian archaeology in that region. Much of our work has been on the Great Plains, which has the best dated regional stratigraphic record of Paleoindian occupation in North America, so we decided to test the hypothesis with data from the Great Plains. We see no evidence of any sort of occupation hiatus at 12.9ka. The end of the Clovis point style tells us nothing about an impact, and in any case the style persists after 12.9ka. Arguments that stratified sites with a post-Clovis occupation hiatus misstate the archaeological and geological records. At sites with multiple Paleoindian occupations, “sterile” layers between occupation zones are the norm, whether they separate Clovis from Folsom zones, Folsom from other Folsom occupations, or any combination of occupations you care to mention. Moreover, out of >150 Paleoindian sites we looked at in the literature, over two-thirds are single occupation sites (see below). So whether they are Clovis, Folsom, or late Paleoindian features, there is no occupation above. Absence of a post-Clovis occupation is not a mysterious “hiatus,” it is the norm at most Paleoindian sites.

Another general question about the data from impact markers in the PNAS paper: Why the multiple peaks among the various indicators? e.g., double carbon spherule and double charcoal peaks at Chobot; the magnetic grain and spherule peak higher than the main carbon spherule peak at Chobot; two Iridium peaks and one carbon spherule peak matching neither Ir peak at Lake Hind; and a variety of spikes that don’t match up at Topper. How exactly did that happen? A single “event” should sprinkle its traces across the continent at the same time (the proponents make this point over and over). Yet they rarely occur together in the sites. I know of no sedimentological or weathering process that could so discretely vertically sort the various indicators.

To date (October, 2010), the 2006 book and the 2007 PNAS paper are the most comprehensive arguments in support of a YD Impact. Significantly, none of the authors is a specialist in late Quaternary terrestrial geology or biology nor in Paleoindian archaeology.


In 2008 and 2009 a series of papers came out, directly contradicting several implications of the PNAS paper. These skeptical papers were written by a wide array of scientists from different disciplines who are leaders in their respective fields and who, for the most part, worked independently of one another.

Nicolas Pinter and Scott E. Ishman, 2008. Impacts, mega-tsunami, and other extraordinary claims.” GSA Today, 18(1) 37-38.

The first critical paper was written before the PNAS paper came out and published just a few months later. Pinter and Ishman (January, 2008) published a brief paper in GSA Today and took issue with The Book by Firestone et al, along with a series of meeting abstracts. They were the first to note that magnetic grains and in particular magnetic microspherules are everywhere on the Earth’s surface and throughout the geologic record. They were also the first to point out problems linking both megamammal extinctions and the “black mat” with some sort of 12.9k years BP “extraterrestrial event.”

Briggs Buchanan, Mark Collard, and K. Edinborough. 2008. “Paleoindian demography and the extraterrestrial hypothesis.” Proceedings of the National Academy of Sciences 105:11651-11654.

Buchanan and colleagues used the frequency of radiocarbon dates from archaeological sites as a proxy for population and concluded that there was no population “gap” just after 12.,9k I don’t believe that numbers of radiocarbon dates tell us much, if anything, about populations; too many variables of preservation, visibility, and sampling come in to play. Nevertheless, the authors see no gaps, so there were people around.

Marlon, J. P. J. Bartlein, M. K. Walsh, S. P. Harrison, K. J. Brown, M. E. Edwards, P. E. Higuera, M. J. Power, R. S. Anderson, C. Briles, A. Brunelle, C. Carcaillet, M. Daniels, F. S. Hu, M. Lavoie, C. Long, T. Minckley, P. J. H. Richard, A. C. Scott, D. S. Shafer, W. Tinner, C. E. Umbanhowar, and C. Whitlock. 2009. Wildfire responses to abrupt climate change in North America. Proceedings of the National Academy of Sciences 106:2519-2524.

This paper dealt with other issues besides The Impact, but the researchers did look to see if they saw any charcoal spikes dating to 12.9ka in lake cores, which is something predicted by The Impact Hypothesis. They saw no evidence for the predicted charcoal layer (see discussion of 2010 AMQUA meeting, below).

More papers followed in later 2009 and in 2010.

Todd Surovell, Vance Holliday, Joseph Gingerich, Caroline Ketron, C Vance Haynes, Ilene Hilman, Daniel Wagner, Eileen Johnson, Phillippe Claeys, 2009 "An Independent Evaluation of the Younger Dryas Extraterrestrial Impact Hypothesis" Proceedings of the National Academy of Sciences v. 106, p. 18155-18158.

An attempt was made to reproduce the evidence for an impact by analyzing samples for magnetic grains and magnetic microspherules. The samples came from several sites reported in the original 2007 PNAS paper (Blackwater and Topper) plus a number of other dated stratigraphic sequences that contained sediments dating to 12.9k. In all cases, no peaks in grains or spheres were detected.

François S. Paquay, Steven Goderis, Greg Ravizza, Frank Vanhaecke, Matthew Boyd, Todd Surovell, Vance T. Holliday, C. Vance Haynes, Philippe Claeys, 2009 “Absence of geochemical evidence for an impact event at the Bølling–Allerød/Younger Dryas transition" Proceedings of the National Academy of Sciences, v. 106, p. 21505-21510.

These investigators attempted to reproduce the Iridium spikes reported in the 2007 PNAS paper (and also looked for other Platinum Group Elements) at five sites in North America investigated by Firestone et al (2007) plus other sites. They found no indicators of an impact.

Jacquelyn L. Gill, John W. Williams, Stephen T. Jackson, Katherine B. Lininger, Guy S. Robinson, 2009. “Pleistocene Megafaunal Collapse, Novel Plant Communities, and Enhanced Fire Regimes in North America” Science 326:1100-1103

This paper deals with the extinction issue rather than the YD Impact per se. But it shows that overall, mega-herbivores declined significantly long before 12.9ka.

Haynes, C.V., Jr., Boerner, J., Domanik, K., Lauretta, D., Ballenger, J., Goreva, J., 2010. The Murray Springs Clovis site, Pleistocene extinction, and the question of extraterrestrial impact PNAS 107, 4010-4015.

This paper was an attempt to reproduce data published by the Impact Team for magnetic grains, magnetic microspherules, and Iridium.

David J. Meltzer and Vance T. Holliday, 2010 “Would North American Paleoindians have noticed Younger Dryas age climate changes?” Journal of World Prehistory, v 23, p 1-41.

This paper did not focus directly on the Impact hypotheses but was a broader look at the Younger Dryas, primarily on the Great Plains. The paper showed that environmental changes in the terminal Pleistocene varied in timing, degree, and direction. There was no synchronous, continent-wide climate change toward significant cooling at 12.9ka. Further, some Clovis sites are younger than the hypothesized Impact Event.

Modified from Holliday & Meltzer, 2010, Current Anthropology, v 51, pp. 575-584, Figure 1.

Scott, A.C., Pinter, N., Collinson, M.E., Hardiman, M., Anderson, R.S., Brain, A.P.R., Smith, S.Y., Marone, F., Stampanoni, M., 2010. Fungus, not comet or catastrophe, accounts for carbonaceous spherules in the Younger Dryas ‘impact layer’. Geophysical Research Letters 37, L14302, doi:10.1029/2010GL043345.

This paper suggested that the carbon spheres reported by The Impact Team are not high-temperature, impact particles, but rather were indistinguishable from low temperature, buried fungal byproducts (fungal sclerotia). They also failed to reproduce a C14 sequence reported by The Impact Team from Arlington Springs, California (a key site for The Impact Team).

There have been several responses to these papers by The Impact Team. They published a variety of claims concerning Paleoindian archaeology and dating in a series of comments on Buchanan et al., but most are misguided:

Archaeological sites containing both Clovis and immediately post-Clovis material are rare... Of the 11 well-dated credible Clovis sites [Waters and Stafford, 2007], none has post-Clovis materials immediately above, suggesting a potential disruption in settlement or landscape use”

(Kennett and West, 2008, comment p. E110)

This is true, BUT most Paleoindian sites are single-component sites, i.e., just one feature; nothing below or above and this applies at non-Clovis sites as well. Out of ~150 published accounts of buried, intact sites, Holliday & Meltzer (2010, Figure 2, Supplemental data table) document fully 2/3 are single component

Modified from Holliday & Meltzer, 2010, Current Anthropology, v 51, pp. 575-584, Figure 2.

They claim that at the Clovis site (Blackwater Draw): Folsom-age materials occur above the Clovis materials but a hiatus of ~500 years is suggested by an intervening sterile deposit (10-35 cm) and radiocarbon ages of 12.4-11.8 ka. (Kennett and West, 2008, comment p. E110).

This is wrong. In the section they examined on the “South Bank,” there are no Clovis or other archaeological features. The nearest feature, as far as records allow, was ~20 m distant (Holliday and Meltzer, 2010...

A series of assertions are made by D.J Kennett, Stafford & Southon (2008, E107):

1) Only 14C dates with measurement precisions <100 years, and preferably <60 years, should be used because larger error margins blur probability distributions; many dates had precisions from 200 years to >2,000 years.

Yes, this would be ideal, but none of the sites used in the 2007 PNAS nor those in Kennett et al. 2009a,b meet these criteria either.

Kennett, D.J., Kennett, J.P., West, A., Mercer, C., Que Hee, S.S., Bement, L., Bunch, T.E., Sellers, M., Wolbach, W.S., 2009a. Nanodiamonds in the Younger Dryas boundary sediment layer. Science 323, 94.

Kennett, D.J. Kennett, J.P., West, A., West, C.J., Bunch, T.E., and 12 others, 2009b. Shock-synthesized hexagonal diamonds in Younger Dryas boundary sediments. PNAS 106, 12623-12638.

2) Only bone dates processed with modern techniques [e.g., XAD... or ultrafiltration...] are valid because of the catastrophic consequences of poor chemical preparation...

True, but dateable bone is not easy to find in many Paleoindian sites and, again, none of the sites used in the 2007 PNAS nor those in Kennett et al 2009a,b met these criteria.

3) Stratigraphic associations between radiocarbon dates and cultural residues need to be demonstrated; e.g., much of the purported pre-11,000 14C years evidence used is now discredited.

The authors provide no support for this claim. Discredited by whom? Dates on charcoal or wood considered valid before the development of AMS-purified collagen dating are still as valid now as when they were assayed.

4) Single-component sites do not have the same credibility as multiple-occupation sites.

The sites dated by Waters & Stafford (who the Impact Team use to support some of their claims; and Stafford is one of the commentators!) include the following single-component sites: Lange-Ferguson, Anzick, Dent, Paleo Crossing, Domebo, Lehner, Shawnee Minisink, Murray Springs, Colby (of the “top” 11 sites in Waters & Stafford’s work), East Wenatchee (of the 17 “top” sites), and Wallys Beach and UP (of the top 24).

5) The potential for site discovery is not equal through time; destruction and preservation vary by region and are determined by burial depth, depositional environment, ground water geochemistry, and site type (e.g., kill, processing, or camp).

No disagreement there, but what is the point?

Kennett et al (2008, E107) conclude their comments by noting: Much more archaeological and chronological work is required to test the YDB extraterrestrial impact hypothesis. That is true for the tests by critics only if it also true for the proponents. And nowhere do they concede that more testing of the YD hypothesis is necessary to confirm it.

The authors of the original paper then published a response:

Collard, Mark, Briggs Buchanan, and K. Edinborough 2008. Reply to Anderson et al., Jones, Kennett and West, Culleton, and Kennett et al: Further evidence against the extraterrestrial impact hypothesis. PNAS 105:E112-114.

The Impact Team published no commentary on the charcoal work by Marlon et al (2009). But on the internet the word was out that Allen West has thirty some additional datasets, and The Impact Team is writing a response paper that includes these [the new cores] along with various criticisms of their chronological and statistical approach. (Email from D. Kennett to D. Meltzer, 17 Feb 2010)

Apparently the large team of charcoal and lake coring specialists somehow “missed” over 30 cores that Allen West learned about and gained access to. And The Impact Team is somehow better equipped and better trained to interpret these new cores. In any case, almost two years after Marlon et al was published, we’ve seen no formal response, except at the 2010 AMQUA meetings (see below).

In response to the work of Haynes et al. at Murray Springs, Firestone et al. (2010) noted that Haynes et al reported extremely high concentrations of Ir ranging from 31 to 64 ppb in two magnetic fractions across the YDB. They [Haynes et al] also note a peak in magnetic grains from the Clovis-age surface.

Firestone, R.B., A. West, T. Bunch, 2010. Confirmation of the Younger Dryas boundary (YDB) data at Murray Springs, AZ. PNAS v. 107, E105.

Haynes et al (2010) responded We consider our iridium analytical results of 64 ppb and 31 ppb to not be anomalous because they are less than the 72 ppb for magnetics from the modern stream bed. Regarding the magnetics, Haynes told West when he first collected at the site that magnetic grains are ubiquitous in the late Quaternary alluvium in and around the site and their local abundances are quite variable.

So over and over we see misstatements of fact and repeated accusations that everyone else attempting but failing to confirm their hypotheses or to reproducing their results must be incompetent.

No formal commentary was published on the work of Surovell et al. in trying to reproduce the magnetic grain and magnetic microspherule data. But pot-shots appeared, including a feature story in the archaeology news magazine The Mammoth Trumpet (April, 2010, v 25, n 2) and in an unreviewed article by Firestone.

In The Mammoth Trumpet (pp. 16-17), [Allen West claimed] “I believe they had good intentions, but simply failed to follow our protocol in adequate detail.” He explains that because the markers are so tiny, proper protocols must be strictly followed to produce meaningful results. “Only scientists who have previously searched for trace materials would find our protocol clearly understandable,” he says, “and to my knowledge, none of the Surovell group has ever worked with trace materials before. I don’t believe they understood the need for carefully following the directions, and felt the changes were inconsequential.”

These comments and others in the article are remarkably condescending. Further, The Impact Team seems to be saying that only they know how to get the results they want! Is that science???.

Similar comments about the work of Surovell et al., plus various other themes that keep coming up from The Impact Team (e.g., cherry picking data that suit their arguments) are on display in a paper by Firestone in Journal of Cosmology (2009, v.2, p. 256-265). This journal is a non-peer-reviewed, on-line journal.

Figure 3 in the J. of Cosmology paper suggested that there are ~15 basins scattered across the southern half of the Great Plains that all line up in directions that lead back to supposed impact sites in the Great Lakes. What isn’t mentioned is that there are thousands of small circular to elliptical basins scattered throughout this region; over 20,000 in northwest Texas and eastern New Mexico alone (Sabin & Holliday, 1995, Annals of the Assoc of American Geographers). They are randomly oriented. Fifteen basins can easily be picked out oriented in any direction one chooses.

Firestone also comes back to the Carolina Bays to make an argument that they are the result of impact from debris of a primary extraterrestrial impact in the Great Lakes area. He attempts to contradict work that shows the basins to be much older than 12.9ka (as discussed above) by impugning the dating results on sand rims around the basins: Older OSL dates at Bays studied by Invester et al. (2007), Firestone suggests, may reflect inadvertent sampling of underlying, older sediment that may have shifted over time. Though, as noted, he embraced Ivester’s work (and distorted it) in The Cycle of Cosmic Catastrophes. Moreover, he here resorts to dismissing the work as some sort of incompetence.

Further, in a comment on the work by Surovell et al. (2009), he asserts: The YDB layer deposited at many sites across North America is only a few mm thick. Turbation by wind and water can destroy the YDB layer, change its position with respect to the YDB, or even split it into multiple thin layers. At about 20% of sites no evidence of the YDB layer remained. But this process of “splitting of the YDB” is never explained. I know of no such post-depositional process that can pick out specific microscopic indicators and move them up or down. Some soil-forming processes can move clay down and form “bands” but that is the only remotely similar process I know of. What could selectively move nanodiamonds in one direction and microspherules in another???

And: Tedious microstratigraphy is required to find the YDB impact layer which was often <2-3 mm wide. No such tedious microstratigraphic field work as ever been reported much less described by The Impact Team. They go to sites and simply sample where local experts tell them that the 12.9ka layer should be. How can they or anyone else know that such layers are so thin? They never documented this in any paper they published. Indeed, as mentioned above, the 2007 PNAS paper (p. 16017) notes that: Each of the 10 Clovis-age sites displays a YDB layer (average thickness of 3 cm). So is it 3cm or 3mm?

And: Broad sampling intervals near the YD layer used by Surovell et al. (2009) have diluted their results considerably. This argument played out over the internet after Surovell et al. (2009) was published. This is a red herring. Before our sampling, The Impact team said that the average thickness of the “YDB layer” is 3 cm, but nowhere are the sampling intervals published or available - I asked both Allan West and Jim Kennett and neither responded. Some of Surovell’s samples were thicker than 3 cm, some about 3 cm, some less. The amount of magnetic microspheres is as much as an order of magnitude higher than any other samples in the reported sections. How can anyone sample a zone maybe twice or three times as thick as The Impact Team did and “dilute” the spike so it couldn’t be seen? Not possible. Moreover, they report near- zero levels of spheres above and below the 12.9ka level. How could we collect more sample with near- zero levels of spheres and dilute it such that we found more spheres than they did? Maybe Surovell was creating matter with his lab methods! That’d get him untold wealth and he could tell all of us to go to Hell!

They [ Surovell et al.] also selectively searched for highly spherical shiny microspherules thus excluding the dull, less spherical, and often pitted microspherules that we reported.

Highly spherical, shiny microspherules were sought out because that is exactly what is described as “typical” in the lab methods with the 2007 PNAS paper!


The Impact Team has increasingly focused their attention on nanodiamonds as the smoking gun for an impact. They published an article on recovery of nanodiamonds in the Greenland ice sheet:

Andrei V. KURBATOV, et al., 2010. “Discovery of a nanodiamond-rich layer in the Greenland ice sheet” Journal of Glaciology, Vol. 56, No. 199, 2010 749-759.

They apparently found a dense layer of nanodiamonds (walked right to it according to NOVA) but failed to date it.

But the notion of nanodiamonds as a smoking gun for an impact has been challenged. Tyrone L. Daulton, Nicholas Pinter, and Andrew C. Scott (PNAS September 14, 2010 vol. 107 no. 37 16043–16047) present data suggesting that the materials identified as nanodiamonds are in fact terrestrial forms of carbon. Further support for this idea is provided by:

H. Tian, D. Schryvers, and Ph. Claeys, 2011. PNAS, v. 108 no. 1, pp. 40-44. “Nanodiamonds do not provide unique evidence for a Younger Dryas impact”

This caused a stir among The Impact Team. In a typical reaction, they called the study "fundamentally flawed science" (Nature on-line, 31 August 2010 | doi:10.1038/news.2010.441) and complaints about Daulton et al. focus on sampling, sample processing, and interpretation. Impact proponents level similar criticisms at other outside studies of traces of impact and fire. Kennett sees little evidence that Daulton and his colleagues actually sampled the YDB, an often thin and sometimes hard-to-recognize layer....“There’s been a real problem of data quality,” Kennett sums up. (2010 vol 329 Science p. 1141).

So once again the skeptics failed to find the thin and sometimes hard-to-recognize layer.... but the Impact Team never explained how they themselves found it. And, of course, the skeptics are incompetent.


One of the arguments used to support the Impact Hypothesis is evidence for cratering in the Great Lakes basin. In the Journal of Siberian Federal University: Engineering & Technologies , Firestone et al. (2010) state that “if multiple 2-km objects struck the 2-km-thick Laurentide Ice Sheet at <30°, they may have left negligible traces after deglaciation… [perhaps] limited to enigmatic depressions or disturbances in the Canadian Shield (e.g., under the Great Lakes or Hudson Bay)” (Firestone et al. 2007:16020). An obvious flaw with that speculation is that by 12,900 years ago only the Lake Superior basin was still under glacial ice (Dyke et al. 2003) (sSee map above). They now suggest “deep holes” beneath four of the Great Lakes could represent impact craters (Firestone et al. 2010:57-58). They dismiss the possibility that these holes were the result of glacial erosion, citing the latest edition of Dawson’s Acadian Geology, a book published more than a century ago (Dawson 1891). Evidently, they believe subsequent generations of glacial and Quaternary geologists working in the Great Lakes failed to notice the holes’ extraterrestrial origin. Yet, if these holes were caused by an impact 12,900 years ago (and they provide no evidence the holes are that old), it is curious that the impacts produced elongated craters at different orientations, yet each one parallel to local ice flow in the up-ice end of its lake basin.



David Meltzer and I presented the archaeological and geoarchaeological record proving no evidence for a “catastrophic demographic collapse” (i.e., the end of the Clovis occupation) in North America. The Impact Hypothesis provides an extraterrestrial answer for an archaeological problem that does not exist.

Vance T. Holliday and David J. Meltzer “The 12.9ka Impact Hypothesis and North American Paleoindians” Current Anthropology, v. 51, p. 575-585

David J. Meltzer and Vance T. Holliday, “Response to comments on ‘The 12.9ka Impact Hypothesis and North American Paleoindians’” Current Anthropology, v. 51, p. 599-602


As noted earlier, samples I collected at Lubbock Lake in the Fall of 2007 were sent to both Todd Surovell and to Jim Kennett. By the Fall of 2009 Kennett and his team completed their analysis. As of May, 2011, I can not reveal the results of the test, at the request of Jim Kennett. Suffice to say, the results (for recovery of magnetic microspheres and also nanodiamonds) did not meet the expectations of the Impact Team. The immediate response was to 1) question my sampling and labeling (I double checked the samples and based on my 25+ years working at Lubbock Lake confirmed that samples were correctly identified and labeled in th field); question my understanding of the site microstratigraphy; and question the dating (but this never came up before I sampled or before they got the results they didn't like). The obvious more obvious issue - never raised by The Impact Team - is that there is a problem with their hypothesis! I hope the results will soon get published. In the meantime, Jim Kennett and I agreed to resample the section at Lubbock Lake. Making arrangements for this dragged on until publication of a piece by Rex Dalton in the magazine Miller-McCune (see link at end of this story). We finally re-sampled in May, 2011. Stay tuned....

AMQUA 2010

As of this writing (March, 2011), the culmination of the debate over the Clovis Impact Hypothesis was at the biennial meeting of the American Quaternary Association, held in Laramie, WY, in August, 2010. The meeting organizers invited three members of The Impact Team and three of the skeptics to present their cases, followed by an open Q&A session. The representatives of The Impact Team were James Kennett, Allen West, and Ted Bunch. The skeptics were Nicholas Pinter, Todd Surovell, and Mark Boslough.

Both sides gave it their best shot, I suppose. The Impact Team made some good points and had some interesting new data. But these data were not published and simply dropped on the skeptics and the audience with no real opportunity to digest it. The skeptics pointed out the flaws in logic in the impact argument, the lack of reproducibility of all previous impact data by independent groups, and distortions of other people’s data. For example, the work on charcoal and fire history from lake cores published by Jennifer Marlon et al. in PNAS in 2009 was criticized and then statistically “massaged” (though the statistical maneuvering was never described) to show that in fact there are charcoal peaks at 12.9k in the lake cores. One of the co-authors happened to be in the audience, however, and pointed out that visual and microscopic examination of the core showed no evidence for a charcoal spike. The Impact Team simply ignored the crux of the argument (the absence of charcoal in field samples).

One of the skeptics showed how the various aspects of the impact story kept changing as their story came under attack. In particular, he showed how the lab protocols for extracting magnetic grains and magnetic microspherules changed after a paper was published showing how the spherule data could not be reproduced. He showed a photo from the Firestone et al 2007 PNAS paper and protocols with "typical spherules" - nice little shiny ball bearings. Then he showed a photo of "typical spherules" from the protocols they sent out after Surovell et al. was published in PNAS in 2009. The typical spheres were all kinds of irregular little blobs. There was an audible gasp in the audience when he showed that slide.

When Alan West was up, he attempted to show where Todd went wrong with his protocols for extracting spherules. He had several PowerPoint slides devoted to showing how the little spheres would settle out of a sample bag, so if Todd (or anyone else) didn't shake the bag, they'd get an undercount. If West actually tested this mechanism, he showed no evidence of those tests. So this whole debate revolves around proper shaking of a sample bag? Anyone who has worked with sediment samples in a lab is routinely moving them and usually shaking them and is careful sampling to get a representative sample. So will the next edition of lab protocols discuss proper bag shaking? How many times is enough? 3? 17? Is that up and down or side to side? Can you turn the bag over while shaking? Or between shakes?

When questioned about the work of Richard Firestone, all three proponents of the Impact hypothesis scrambled to disassociate themselves from him. “Rick is out there on his own” one of them was heard to say. And “He is going his own way.” (Though I note that just this year Allen West was publishing with Firestone in the Journal of Siberian Federal University: Engineering & Technologies (Firestone et al. 2010), as noted above.

So it goes...

Nic Pinter, who co-authored the first critique of the Impact Hypothesis (2007, noted above), also co-authored an exceptionally thorough review, summarizing over 3 years of debate, claim, and counter-arguments.

Nicholas Pinter, Andrew C. Scott, Tyrone L. Daulton, Andrew Podoll, Christian Koeberl, R. Scott Anderson, Scott E. Ishman, 2011. The Younger Dryas impact hypothesis: A requiem. Earth-Science Reviews, doi:10.1016/j.earscirev.2011.02.005

This “requiem” is well-titled. Though some proponents of the Clovis Impact Hypothesis will not be deterred, most objective readers will likely realize that the case supporting the hypothesis was a house of cards.

Key point: No comprehensive statement on the YD Impact hypothesis has been published in a peer-reviewed venue by the proponents. In response to criticisms of their work, lack of reproducibility of data, and failure to confirm their hypotheses, The Impact Team invariably ends up accusing all others (largely an array of specialists in their respective fields, and most working independently) of incompetence.

Was there an “extraterrestrial event” somewhere on or above North America at 12.9k years BP? I have no idea. Ultimately, I don’t care whether there was a 12.9k impact. I just want to know what was going on then. If there was, I see no reason to believe - at this point - that it had an impact on people, flora, fauna, or climate. We could be dealing with an extraterrestrial impact that had no terrestrial impact!

Just because an hypothesis is outrageous doesn’t make it true.

Also, for further reading, see Rex Dalton's article in Miller-McCune. (link)

(Thanks to Nicholas Pinter for his comments on this personal narrative).

© 2007 Department of Anthropology, University of Arizona. All Rights Reserved.