In a recent paper, Pelton et al. (Reference Pelton, Kornfeld, Larson and Minckley2017) have presented a statistical procedure for evaluating radiocarbon (14C) dates from a complex stratified cultural sequence with numerous 14C dates at Locality I of the Hell Gap site in Wyoming. There are so many stratigraphically controlled 14C dates, some conflicting, from Locality I that interpretations can vary significantly. Their procedure provides a useful approach to avoiding, or at least reducing, subjectivity. However, before presenting a qualitative approach to assessing the same radiocarbon dates at Locality I, there are several caveats that perhaps should be mentioned.
My chief criticism of their study is that they have assumed that the earliest dated cultural strata, their Mode 11, is a Goshen level, yet all of the excavations at Locality I since those of Henry Irwin (1968) have failed to yield any Goshen diagnostics in stratigraphic context. Haynes and Hill (Reference Haynes and Hill2017) pointed out that the stratigraphic provenance of the original Goshen points is so close to that of the lower Folsom level that their relative ages are equivocal. The age of ca. 12,800 cal yr BP (11,000 14C yr BP) for Mode 11 could well be for a Clovis occupation that has yet to be defined via diagnostics.
Their assumption “that a date’s relative position within a stratum is more accurate than its absolute depth below the top of that stratum” (Pelton et al., Reference Pelton, Kornfeld, Larson and Minckley2017, 235) is, of course, true provided that deposition at both locations is during the same increment of time as is typical for floodplain overbank aggradation, whereby increments of deposition are typically made up of widespread layers deposited one after another in layer-cake fashion (Fig. 1, lower strata 1–3). This mode of deposition at Locality I is believed to apply to the lower cultural strata (E1 through E5) (Haynes, Reference Haynes2009).
Figure 1 A hypothetical stratigraphic cross section parallel to the channel of Hell Gap Creek showing laminar bedding of overbank aggradation of early strata (1–3) and lensoidal bedding for slopewash sedimentation of later strata (1–10). The microstratigraphic contacts within the lensoidal bedding of the upper zone may not be discernable in the field, making it difficult to tell that sample A is significantly younger than sample B even though both are at the same depth below the surface. No scale.
However, many of the higher strata (F through G) are derived from slopewash from the adjacent hillside whereby an increment of deposition in one area represents a different increment of time than in another area (Fig. 1, upper strata 1–10). Here increments 2 through 10 are all younger than increment 1, such that a sample from the middle of the stratum at A is in increment 6 and is considerably younger than a sample from the middle at B, which is in increment 1.
However, in spite of this caution, the fact that their statistical results show clustering within strata (i.e. their modes of figs. 3 and 4) suggests that increments of deposition making up a particular stratum are small enough and frequent enough that their assumption is valid for strata derived via slopewash at Locality I.
I do not know how it affects their statistical manipulation, but their standard stratigraphic section (SSS) at the southwest corner of the Witness Block does not contain three strata (i.e., G1, F1, and D1), present elsewhere at Locality I. However, one of their dates, 7840±62 14C yr BP (AA-65328), is from substratum F1g, which occurs only in the east wall of Locality I East, yet another date in this wall (i.e., 5719±97 14C yr BP [AA-65326]) in stratum G1 is omitted, presumably because G1 is not present at the SSS. But neither is stratum F1 present in the SSS. It seems to me that for a stratigraphic section to be a standard, it needs to contain all of the strata at that locality.
In spite of their reluctance to vet dates any more than they did, a selection based on the elimination of anomalous ones and ones of dubious quality, from my point of view, might be worth examining. Their table 1 lists, in stratigraphic order, all radiocarbon dates from Locality I determined from 1993 to 2005. What appear to be anomalous results in this list may be discerned and eliminated by qualitative reasoning, as explained next, and so indicated by an “X” placed before the date in Table 1.
Table 1 Hell Gap Locality I Radiocarbon Dates Arranged in Stratigraphic Order
* Rejected by Pelton et al. Reference Pelton, Kornfeld, Larson and Minckley2017
** Not listed in Table 2 of Pelton et al. Reference Pelton, Kornfeld, Larson and Minckley2017
† Material analyzed: C = charcoal, H = humates, R = residue of bulk sediment sample
X = Value eliminated for reasons explained at the top of page 2.
Of the three values for stratum F2, the charcoal date of 7700±120 cal yr BP (AA-35653) is rejected in favor of the humate age of 8728±73 14C yr BP (AA-38274) for two reasons. First, the humate age is within 1 standard deviation (1σ) of the next older sample of 8820±60 14C yr BP (AA-28776) in stratum F2. The second reason is that the humate age, being significantly older than the charcoal date, is probably less contaminated than the charcoal value. Humate ages from charcoal, if different from the charcoal ages from which they are mostly derived, are usually younger because of contamination by extraneous mobile humic acids in the geochemical environment. This is especially so if the samples are from within the zone of fluctuation of the water table, which these are not. It is unlikely that a high water table with potentially older humic acids ever reached stratum F2.
The single charcoal date for stratum F1 (7840±62 14C yr BP [AA-65328]) is rejected because it is significantly younger than the 14C ages above and below. It is, therefore, probably either translocated via bioturbation or contaminated.
The third date rejected is the humate date of 8630±370 cal yr BP (AA-38210) in favor of the charcoal age of 9030±260 14C yr BP (AA-38209) at the top of stratum E5 even though they overlap at 2σ. Also, the 9920±950 14C yr BP (AA-27677) date for middle E5 is rejected because the standard deviation is more than 10% of the indicated age. The value is simply too imprecise to be useful.
For stratum E4, the humate date of 8770±120 14C yr BP (AA-13373) is rejected in favor of the charcoal age of 9250±75 14C yr BP (AA-14433) above it and the humate age of 9360±85 14C yr BP (AA-13372) below it. Also, the charcoal date 8190±100 14C yr BP (AA-27676) is rejected because it is too young by comparison with the ages above and below. Its age is more like that of upper stratum F from which it may have derived via bioturbation.
Both of the charcoal dates, 12,100±830 14C yr BP (AA-28773) and 11,250±140 14C yr BP (AA-20546), from stratum E2 are rejected as being anomalously old compared with values above and below.
Much confusion about the stratigraphic positon and age of sample 6HG93 is because of the inadvertent misplotting of it on the profile of the east wall of Locality I West (Haynes, Reference Haynes2009, fig. G.2), which is the west wall of the Witness Block. The sample is actually from a deep test pit about 2 m south of Locality I West and about 4 m to the west of the west wall of the Witness Block. This was excavated in 1993 to determine the strata below D2. A single piece of charcoal within stratum D2 was collected from the north wall of the pit and provided the charcoal and humate ages 10,955±135 14C yr BP (AA-14434) and 11,440±120 cal yr BP (AA-33671) in table 1. The position projected to near the center of the north wall of Locality I West (Haynes, Reference Haynes2009, fig. G.2) is closer to being the true positon than that plotted at the east wall. The charcoal date of 10,955±135 14C yr BP (AA-14434) is for the bottom of the upper one-third of stratum D2.
All three dates from stratum D2 in Table 1 may be too young because they would be more appropriate for stratum D3, which is not present at Locality I. There the basal contact of stratum E1 rests on an erosional surface at Locality I that would be the Clovis occupation surface if Clovis was ever present at Locality I (Haynes, Reference Haynes2009). The samples may be intrusive having come from the E1/D2 contact via bioturbation.
Both values for stratum C are probably too young in that the overlying date of 12,170±160 14C yr BP (AA-27646) for the E1/C contact is probably a minimum age for upper stratum C based on 14C dates at localities elsewhere in the Hell Gap area. This qualitative procedure for vetting, as presented here, should be done for all of the radiocarbon dates in the Hell Gap site area (Haynes, Reference Haynes2009, Appendix F: 336–340). If time permits, this will be attempted in the future.
Although the vetting process employed here may be considered to involve a degree of subjectivity, it is conducted with the sole intent of arriving, as near as possible, to the truth and may be more realistic than the statistical procedure of Pelton et al. (Reference Pelton, Kornfeld, Larson and Minckley2017).
ACKNOWLEDGMENTS
This manuscript was improved by constructive comments from Matthew E. Hill Jr. and Vance T. Holliday. Word processing was provided by Barbara Fregoso, and the line drawing of Figure 1 was prepared by Jim Abbott. Electronic submission was graciously provided by Vance Holliday.
