by Richard B. Firestone & William Topping
Volume 16, Number 2 March 2001
The Paleoindian occupation of North America, theoretically the point of entry of the first people to the Americas, is traditionally assumed to have occurred within a short time span beginning at about 12,000 yr B.P.
This is inconsistent with much older South American dates of around 32,000 yr B.P.1 and the similarity of the Paleoindian toolkit to Mousterian traditions that disappeared about 30,000 years ago.2
A pattern of unusually young radiocarbon dates in the Northeast has been noted by Bonnichsen and Will.3, 4
Our research indicates that the entire Great Lakes region (and beyond) was subjected to particle bombardment and a catastrophic nuclear irradiation that produced secondary thermal neutrons from cosmic ray interactions.
The neutrons produced unusually large quantities of 239Pu and substantially altered the natural uranium abundance ratios (235U/238U) in artifacts and in other exposed materials including cherts, sediments, and the entire landscape.
These neutrons necessarily
transmuted residual nitrogen (14N) in the dated charcoals
to radiocarbon, thus explaining anomalous dates.
For example, at the Gainey site in Michigan a 2880 yr B.P. radiocarbon date was reported, while the thermoluminescence date for that site is 12,400 yr B.P.5 Other anomalous dates found at Leavitt in Michigan,6 Zander and Thedford in Ontario,7 Potts in New York,8 Alton in Indiana,9 and Grant Lake in Nunavut10 are summarized in Table 1.
The Grant Lake Paleoindian site is most
remarkable because its 160 [rc] yr B.P. age is nearly contemporary,
while adjacent and deeper samples give ages of 1480-3620 [rc] yr B.P.
In a pioneering study of the Paleoindian site at Barnes, Michigan, Wright and Roosa observed that Paleoindian artifacts were deposited before the formation of spodosols ceased in this area about 10,000 yr B.P.14
This conclusion was based on observing that cemented
sediments on artifacts, found outside their original context,
defines their original stratigraphic position.
The C sediment horizon is clearly recognized by its transitional color and confirmed by elevated concentrations of potassium and other isotopes. Color and chemistry are key indicators of this very old soil11,12,13,14 derived from parent materials and associated postglacial runoff.15 At Gainey, large quantities of micrometeorite-like particles appear to be concentrated near the boundary between the B and C sediment horizons.
They can be separated with a magnet and are identified by the presence of chondrules and by visual evidence of sintering and partial melting. These particles, dissimilar to common magnetites, are found in association with a high frequency of "spherules." The depth profiles for potassium and particles at the Gainey site are compared in Fig. 1.
Minor vertical sorting of
particles is apparent, with a shallow spike of particles near the
surface probably resulting from modern agricultural or industrial
activity. Total gamma-ray counting of sediment profiles in the
various locations invariably showed increased radioactivity at the
B-C boundary consistent with enhanced potassium (40K) and
possibly other activities.
The depth of penetration into the artifacts implies that the particles entered with substantial energy.16 Field simulations with control cherts for large particles (100-200 microns) suggest an entrance velocity greater than 0.4 km/s, and experiments at the National Superconducting Cyclotron Laboratory indicate that the smaller particles left tracks comparable to about 526 MeV iron ions (56Fe) in Gainey artifacts.
Similar features are not observed in later-period prehistoric artifacts or in bedrock chert sources. Track angles were estimated visually; track densities were measured with a stage micrometer; track depths were found by adjusting the microscope focus through the track.
These data are summarized in Table 1.
Track and particle data in Table 1 suggest that the total track volume (density times depth) is highest at the Michigan, Illinois, and Indiana sites and decreases in all directions from this region, consistent with a widespread catastrophe concentrated over the Great Lakes region.
The nearly vertical direction of the tracks left by
particle impacts at most sites suggests they came from a distant
Exceptions include 14N, which captures a neutron and emits a proton to produce 14C; and 235U, which mainly fissions into two lighter elements.
The relative size of isotopes in chert is shown in figure "A neutron's view of chert."
Significant variations in the isotopic ratio do not occur because of chemical processes; however, a thermal neutron bombardment depletes 235U and thus alters the ratio. Solar or galactic cosmic rays interacting with matter produce fast secondary neutrons that become thermalized by scattering from surrounding materials.
Thermal neutrons see a target of large
cross section (681 barns) A for destroying 235U, compared with a
target of only 2.68 barns for neutron capture on 238U. Therefore,
despite the low abundance of 235U, about 1.8 times as many 235U
atoms are destroyed as 238U atoms by thermal neutrons.
239Pu produced during the bombardment will also be partly destroyed by thermal neutrons with 1017 barn cross section.
Assuming 239Pu doesn't mobilize, it will decay back to 235U (half-life 24,110 yr), partially restoring the normal abundance.
Paleoindian artifacts from Gainey, Leavitt, and Butler, and two later-period artifacts from the same geographic area of Michigan were analyzed for 235U content by gamma-ray counting at the Phoenix Memorial Laboratory, University of Michigan.
They were compared with identical chert types representative of the source materials for the artifacts. Control samples were extracted from the inner core of the purest chert known to be utilized by prehistoric people.
The Paleoindian artifacts contained
about 78 percent as much 235U as the controls and later-period
artifacts, suggesting substantial depletion. Depletion of 235U
necessarily indicates that thermal neutrons impacted these artifacts
and the surrounding prehistoric landscape.
McMaster ran additional calibration standards and has considerable expertise analyzing low-level uranium.
This analysis was sensitive to a few ppb for 235U and 0.1-0.3 ppm for 238U, more than sufficient to precisely analyze the uranium-rich chert samples (0.7-163.5 ppm).
Most samples were depleted in 235U, depletion increasing geographically from the southwest (Baker, Chuska chert, 17 percent) to the northeast (Upper Mercer, 77 percent), as shown in Table 2.
This is consistent with cosmic rays focused towards northern latitudes by Earth's magnetic field. Only a very large thermal neutron flux, greater than 1020 n/cm2, could have depleted 235U at all locations.
Samples of unaltered flakes from Taylor and sediment originally adjacent to Gainey artifacts showed 235U enriched by 30 percent.
Both samples were closely associated with the particles described
above. The position of these samples appears to be related to the
enrichment, which cannot be explained by thermal neutrons from the
bombardment. To test this, we bathed another Taylor flake in
48-percent HF at 60°F for ten minutes to remove the outer 70 percent
of the sample and the attached particles. Analysis showed the
"inner" flake depleted in 235U by 20 percent, consistent with the
other depleted cherts.
This neutron flux can then be used to estimate the amount of additional 14C that would have been produced in charcoal by neutrons colliding with 14N (14N cross section = 1.83 barns).
The corrected radiocarbon age can then
be estimated by comparing the current amount of 14C in the dated
charcoals, determined from their measured radiocarbon age, with the
amount of 14C that would have been produced by the bombardment. For
these calculations we assume that charcoal contains 0.05 percent
and that initial 14C concentrations were the same as today (one 14C
atom for 1012 12C atoms).
This mobility is demonstrated at the Nevada Test Site, where plutonium, produced in nuclear tests conducted by the U.S. between 1956 and 1992, migrated 1.3 km.20
It has also been shown that atoms produced by radioactive decay or nuclear reaction become weakly bound to the parent material and pass more readily into solution than isotopes not affected.21 Both 239Pu and 235U are thus expected to be mobile, complicating any analysis. This is consistent with the enrichment of 235U in the two external samples where migrating 239Pu or 235U may have been trapped, thus enriching the relatively uranium-poor outer regions. Alternatively, excess 235U may have been carried in by the particles. Radiocarbon produced in situ by irradiation should also be mobile.
If 14C is more mobile than
239Pu, then the dates calculated above should be decreased
The proposed date for the Gainey site
also falls closer in line with the radiocarbon date for a
Lewisville, Texas, Paleoindian site of 26,610 ± 300 yr B.P.22,23
and radiocarbon dates as early as c. 20,000 yr B.P. for Meadowcroft
the Lewisville and Meadowcroft sites were likely exposed at the same
time to thermal neutrons, we estimate that their dates should be
reset to c. 55,000 yr B.P. and c. 45,000 yr B.P., respectively.
(The Gainey thermoluminescence date of 12,400 yr B.P. is probably a result of the heat generated by the nuclear bombardment at that time, which would have reset the TL index to zero.)
The modified dates for Paleoindian settlements suggest that the timetable for glacial advance sequences, strongly driven by conventional radiocarbon dates, should be revisited in light of the evidence presented here of much older occupations than previously thought."
Magnetic excursions occur every 10,000-20,000 years when the
Earth's magnetic field becomes weak, and the poles may even reverse
for a short time.
The global Carbon Cycle suggests that 14C produced by cosmic rays would be rapidly dispersed in the large carbon reservoirs in the atmosphere, land, and oceans.26
We would expect to see a sudden increase in radiocarbon in the atmosphere that would be incorporated into plants and animals soon after the irradiation; after only a few years, most of the radiocarbon would move into the ocean reservoirs. The 14C level in the fossil record would reset to a higher value.
The excess global radiocarbon would then decay with a half-life of 5730 years, which should be seen in the radiocarbon analysis of varved systems.
Fig. 2 plots 14C from the INTCAL98 radiocarbon age calibration data of Stuiver et al. for 15,000-0 yr B.P.27 and Icelandic marine sediment 14C data measured by Voelker et al. for 50,000-11,000 yr B.P.28
Excess 14C is indicated by the difference between the reported radiocarbon dates and actual dates. Sharp increases in 14C are apparent in the marine data at 40,000-43,000, 32,000-34,000 and c. 12,000 yr B.P
These increases are coincident with
geomagnetic excursions B that occurred at about 12,000 (Gothenburg),
32,000 (Mono Lake), and 43,000 yr B.P. (Laschamp),29
when the reduced magnetic field would have made Earth especially
vulnerable to cosmic ray bombardment. The interstitial radiocarbon
data following the three excursions were numerically fit, assuming
exponential decay plus a constant cosmic ray-produced component. The
fitted half-lives of 5750 yr (37,000-34,000 yr B.P.), 6020 yr
(32,000-16,000 yr B.P.), and 6120 yr (12,000-0 yr B.P.) are in good
agreement with the expected value.
This occurrence can be dated precisely
to 12,500 ± 500 yr B.P., an average of the remarkably consistent
concentration peak centroids in the Greenland ice core data.
Significant increases at that time are not found in comparable data
for the Antarctic, which indicates that the cosmic ray irradiation
was centered in the Northern Hemisphere. Weak evidence of an
occurrence at 12,500 yr B.P. is seen in the radiocarbon record for
marine sediments near Venezuela,34
confirming that the cosmic ray bombardment was most severe in
Thus any increase in 10Be would be cosmic in origin; and the cosmic ray rate could only change if there were a nearby supernova. During the last Ice Age the 10Be deposition rate in ice at both poles was much higher than today. Gulf of California marine sediments clearly show strong 10
Be peaks at 32,000 and 43,000 yr B.P. McHargue argues
that these peaks can only be explained by a supernova.
The effect of a
supernova on Earth
Fig. 2 shows that each episode in a series produced a similar amount of atmospheric radiocarbon.
The sun lies almost exactly in the center41 of the Local Bubble, believed to be the result of a past nearby supernova event. A candidate for the reverse shock wave is the supernova remnant North Polar Spur, with an estimated age of 75,000 years and a distance of 130 ± 75 parsecs (424 light years),42 conveniently located in the north sky from where it would have preferentially irradiated the Northern Hemisphere.
Assuming the Taylor flux is average and 1,000 neutrons are produced per erg of gamma-ray energy,43 the catastrophe would have released about 1016 erg/cm2 (2 x 108 cal/cm2), corresponding to a solar flare of 1043 ergs or a gamma-flash of 1054 ergs from a supernova about 1 parsec away.
The geographical distribution of particle tracks, 235U depletion, and 239Pu concentration shown in Fig. 3 are quite consistent, although the particle tracks seem to be confined to a smaller geographic area.
They indicate energy released over the
northeastern sector of the U.S., with maximum energy at about 43° N,
85° W, the Michigan area of the Great Lakes region.
Clark et al. estimate that supernovas release 1047-1050 ergs within 10 parsecs of Earth every 100 million years.45 Brackenridge suggests that a supernova impacted the earth in Paleoindian times.46
Damon et al. report evidence from the 14C tree ring record that SN1006, which occurred at a distance of 1300 parsecs, produced a neutron shower of 2 x 108 n/cm2.47 Castagnoli et al. report evidence of the past six nearby supernovae from the thermoluminescence record of Tyrrhenian sea sediments.48
Dar et al. suggest that a cosmic ray jet
within 1000 parsec would produce 1012 muons/cm2 (greater than 3 x
109 eV) and 1010 protons and neutrons/cm 2 (greater than 106 eV) and
deposit over 1012 erg/cm2 in the atmosphere every 100 million years.49
A cosmic ray jet is also predicted to produce heavy elements via the
r-process and could be a source of 235U enriched up to 60 percent in
The presence of a nearby small and dense interstellar cloud may explain the origin of the particle bombardment.50 The size of the initial catastrophe may be too large for a solar flare, but a sufficiently powerful nearby supernova or cosmic ray jet could account for it.
It appears that the catastrophe initiated a sequence
of events that may have included solar flares, impacts, and
secondary cosmic ray bombardments.
Larger animals were more affected than smaller ones, a pattern that conforms to the expectation that radiation exposure affects large bodies more than smaller ones.54,55
Sharp fluctuations of 14C in the Icelandic marine sediments at each geomagnetic excursion are interesting; because global carbon deposits in the ocean sediments at a rate of only about 0.0005 percent a year, a sudden increase in sediment 14C may reflect the rapid die-off of organisms that incorporated radiocarbon shortly after bombardment.
Massive radiation would be expected to cause major mutations in plant life. Maize probably evolved by macro-mutation at that time,55,56 and plant domestication of possibly mutated forms appears worldwide after the Late Glacial period.
For example, there was a rapid
transition from wild to domesticated grains in the Near East after
This work mandates that conventional radiocarbon dates be reinterpreted in light of hard terrestrial evidence of exposure of the radiocarbon samples to a cosmological catastrophe that affected vast areas of North America and beyond.
A nuclear catastrophe can reset a group of unrelated artifacts to a common younger date, creating gaps and false episodes in the fossil record. Geographical variation and complicated overburdens may further confuse the interpretation.
Scrutiny of Paleoindian artifacts and the North American paleolandscape, associated stratigraphic sediments, coupled with continued radiological investigations, may provide more evidence for the cosmic catastrophe and new clues to the origin of Paleoindians.