Radiocarbon samples, dating, and site contexts

The ¹⁴C dates employed are either new dates, all UCIAMS, published here (S1 Table), or previously published dates [39, 65]. No permits were required for the described study, which complied with all relevant regulations. The new UCIAMS samples were analyzed by the W.M. Keck Carbon Cycle Accelerator Mass Spectrometry Laboratory (KCCAMS) at the University of California-Irvine [101]. At KCCAMS, bone samples were decalcified in 0.5N HCl, gelatinized at 60°C and pH 2, and ultrafiltered to select a high molecular weight fraction (>30kDa) [102]. Charcoal and maize samples were subjected to the standard acid-base-acid (1N HCl and 1N NaOH, 75°C) pretreatment. δ¹³C values were measured at KCCAMS to a precision of <0.1‰ relative to standards traceable to PDB with a Thermo Finnigan Delta Plus stable isotope ratio mass spectrometer (IRMS) with Gas Bench input. Details on KCCAMS dating protocols are available on the facility’s website (https://www.ess.uci.edu/group/ams/facility/ams).

The new samples for dating from the Garoga, Klock, and Smith-Pagerie sites were recovered from hearths or large pit features. The size and shape of the large pit features (S1 Table and S1 File) and in many cases the presence of charred or partially charred grass and/or bark linings suggest their original function was storage of agricultural crops [103]. The linings and any crop macrobotanical remains located on and/or immediately above the linings most likely represent the original uses of the pits. Therefore, radiocarbon dates on these elements are likely to represent time during each feature’s original period of use. Fill above these elements represent subsequent activities including purposeful filling episodes, fill from sheetwash and erosion of pit walls, as well as subsequent secondary use [103, 104]. Unburned bone recovered from hearths is unlikely to have been deposited while the hearths were in primary use. As such, radiocarbon dates on material recovered from these known pit contexts represent terminus ad quem for the pits’ original functions. We focused especially on identifying wood charcoal samples from specific loci with a number of tree-rings preserved out to original waney edge/bark. The terminal tree-ring in these cases therefore provides a close date estimate for the human use of this sample in the episode represented in the context. The availability of a series of tree-rings leading to this event means that we can employ the ¹⁴C ‘wiggle-matching’ approach [66, 67] to constrain and better define the likely date. Even short-sequence wiggle-matches can usually usefully inform chronological analysis [63, 64, 68, 98, 105].

We note that cases of reported inaccuracies for short-sequence wiggle-matches at high-precision [106, 107] are in fact relatively minor in scale in most cases and typically occur when the time-intervals of the samples dated (i.e., the specific number of tree-rings dated, and so calendar window comprising the dated samples) are much shorter than the integrated (smoothed) IntCal ¹⁴C Calibration curve resolution available. Inaccuracies may also occur at certain periods when the total length of the time series is sufficiently short that it may not be able to permit discrimination between multiple periods with similar atmospheric ¹⁴C levels and trajectories (i.e. reversals or wiggles in the calibration curve). Over the last 1000 years, such parallel instances of similar ¹⁴C levels and trajectories are typically distinguishable once time-series are greater than ~50 years. The latter issue of potential ambiguity is potentially relevant in the context of this article especially to the periods around ~1490–1535 and ~1600–1635. In such cases, additional constraints may be necessary to achieve satisfactory discrimination for very short time series. The former issue applies when, for example, single, annual, tree-rings only are dated (potentially capturing high-frequency annual-scale variation in atmospheric ¹⁴C—attested in available annual resolution ¹⁴C time series: e.g. [108–110]), whereas the calibration record, while largely based on annual data for the second millennium CE, is in fact used as a smoothed curve removing very high-frequency (i.e., annual-scale) variation [111, 112]. For this reason in our study, we sought to date samples comprising several tree-rings in most cases where possible (5 rings in six cases, 2 or 3 rings in 4 cases each, 6 or 9 rings in 2 cases, and a single ring in just 2 cases: see S1 Table), in order to damp-down any very high-frequency variation and so achieve data likely more similar to the smoothed IntCal curve. Nonetheless, we suggest that we do observe an issue in the case of the Klock Ulmus sp. wiggle-match (see below and Fig 3). Here short-term variation or another offset may apply. The consistency of results across different analyses of the same/similar samples in this case suggests perhaps the relevance of real short-term higher-frequency atmospheric variation–especially noticeable in the relevant mid-later 16^th century period where there is a plateau or moderate reversal in the calibration curve accentuating any ¹⁴C offset/difference. As noted by [113], any instances of inter-laboratory variation/offsets will further exacerbate any such problems at high-precision and where only a short sequence with limited data is wiggle-matched. This observation suggests that care needs to be taken with high-resolution dating in the mid-later 16^th century, especially if considering annual resolution samples or short sequences. But, in our specific case, it is important to observe that this apparent high-frequency variation (or other offset) is not occurring at a time where there is a potential alternative calendar placement (contrast the case if the series was lying around either ~1490–1535 and/or ~1600–1635 when there could be an ambiguity). Instead, the mid to later 16^th century calendar placement for this Klock time series is the only available placement for the series regardless of the evident variation, noise or offset.

a. The laboratory stated data wiggle-matched against IntCal20. Two measurements for RY1001-1002 are combined as a satisfactory weighted average and there are then dates on RY1010-1018 and RY1021-1025. The likely non-modelled calibrated locations do not accommodate the known tree-ring spacing in detail. The two dates on RY1001-1002 identify 1527–1552 as the most likely placement (purple box). But this is not compatible with the other two dates given the tree-ring spacings (hence the expected, if the likely non-modelled dates on RY1001-1002 are correct, pink box is left floating in free space away from the IntCal20 curve). The wiggle-match, trying to accommodate all the dates therefore moves the modelled RY1001-1002 dates (solid green distribution) to a more recent placement (largely missing IntCal20 and hence achieving a poor OxCal Agreement value: 44.6<60) but allowing the other modelled dates (solid gray and solid cyan distributions) to be associated with IntCal20 (although the overall Amodel value is poor: 31<60). b. The raw IntCal20 dataset from which the modelled curve in a. is derived [111] highlighting the ETH dataset [108]–these raw data indicate greater possible variation which might allow the observed Klock data to fit better, including earlier, similar to the purple and pink boxes in a.

Tree-ring sequences are described in terms of Relative Years (RY): the first measured (innermost) tree-ring (after pith, if present) is RY 1001. In this case we sought to have more than one such sequence for each site with the view that achieving multiple compatible results reinforces confidence in the dates obtained. The samples from Brigg’s Run, in contrast, are from an avocational archaeologist’s collection donated to the New York State Museum with minimal associated documentation. Thus we make no intra-site history assumptions in this case. The key issue for this site, rich in European trade goods including types that, according to all understanding, can only date from the early 17^th century ([39] at pp.252-259), is whether the ¹⁴C dates are consistent with placement in the first decades of the 17^th century.

In this study, aimed at high-resolution dating of the Garoga, Klock and Smith-Pagerie sites, we were specifically interested in whether samples reflected original functions or subsequent fill, and, if subsequent fill, whether this could be related to the closure or end of use of the feature and likely the wider site occupation, or even post-occupation activities. To that end, site-specific excavation and laboratory records in the New York State Museum’s archaeological archives were consulted to determine the origin of samples dated from within the various pits (S1 File). Written records, often made by field-school students, were of varying quality and detail. However, profile drawings combined with these notes made it possible in some cases to either determine specifically where samples were recovered or their most likely location of recovery within a feature’s fill. Determinations of likely location of recovery are provided in S1 Table with discussion and additional information in S1 File. Here we provide descriptions of the pits from Garoga, Klock, and Smith-Pagerie from which the wood charcoal samples we selected for dating were recovered.

Garoga site Feature 36 was a cylindrical pit measuring 122 cm in diameter and 147 cm deep within Longhouse 9 ([76] at p.99). The pit’s profile indicates a complex series of fill units (S1 File). Field notes indicate that a “lobe” of fill was present at the top of the feature along the south edge. This fill contained fire-cracked rock, charcoal, ash, and a few sherds. This is the only fill unit from which charcoal is specifically mentioned and is the most likely source of the charcoal samples. We therefore regard this context as after the main use of the pit–a terminus ante quem (TAQ) for the earlier use and for the site as a whole.

Klock Feature 84 was a large pit measuring 107 x 152 cm in plan and 122 cm deep (S1 File) within Longhouse 1 (or House 1 = H1). A sample of the charred grass lining was taken from the bottom of the pit, which was AMS dated for the current project (UCIAMS-239714). We regard this lining sample as dating the construction or start of use of the pit. Immediately above the grass lining was a lens of charred maize including maize cobs with kernels attached. Maize kernels (UCIAMS-239713, UCIAMS-218474) and a maize cob fragment (UCIAMS-239712) were AMS dated. These samples, all subsequent to the lining, date the primary use of the pit. The charcoal samples most likely come either from a small lens labeled “burned wood” or a thin lens near the top of the pit that indicates charcoal. We therefore interpret the date of the waney edge of the charcoal sample as defining the closing use/end of use of the pit. It is thus a TAQ for the earlier use of the pit and for House 1 generally—and likely for much of the site occupation, since such deliberate final pit closure is likely from the end of use/abandonment of the site. Indeed, given that the cutting and use date available for a sample of this charcoal (see below) appears to be a few decades at least after the rest of the Klock dating evidence, there is reason to suspect that this charcoal was in fact from the lens right at the top of the Feature 84 pit and so may in fact even post-date the main Klock site occupation period.

The maize sample dated by ISGS-A0326 needs comment. This sample was reported in [65] as coming from Feature 84. This is incorrect. It is in fact sample 45157-A and is from Feature 50. Feature 50 is the other pit within Longhouse 1 with a large maize deposit. However, unlike Feature 84, the maize deposit in the Feature 50 pit is not from the bottom (and first or early use) of the pit, and so likely represents use and fill of the pit after it was no longer serving its primary purpose. Hence we assume that this maize deposit reflects later use of Longhouse 1, in contrast to the initial/earlier use represented by the material from the bottom of the Feature 84 pit.

Klock Feature 135 was a pit measuring 123 cm in diameter located between Longhouses 3 and 5. Field notes indicate there was a lining of charred “vegetal material” at the base of the pit in which maize was present. This is the only mention of maize in the field notes and is the likely source of the AMS-dated maize kernel (UCIAMS-218475) reported in [65] representing the original function of the pit. Above this lens was a massive deposit of large fire cracked rocks, which filled the pit. The field notes do not mention charcoal samples, but it is likely that the charcoal AMS dated for the present project came from the fire-cracked rock deposit—wood charcoal is not mentioned in the description of the bark lining.

Smith-Pagerie Feature 54 was a pit measuring 122 x 152 cm in plan and 135 cm deep within Longhouse 1. A dark, charcoal-stained lens extending across the pit at depths of 63.5–72.4 cm, 61 cm above the bottom of the pit, contained abundant maize. A maize kernel from this lens was AMS dated (UCIAMS-218490) as reported in [65]. AMS dates on two maize kernels also apparently from this lens (AA-6419, AA-7405) were reported by [39]. Field notes indicate that charcoal samples selected for dating were recovered from the same lens as the maize. The pit profile drawing suggests that the upper half of the pit was filled in, partially re-excavated, and refilled (S1 File). This second fill was evidently excavated into and then refilled. The dark lens was at the base of the first posited re-excavation. A 19^th century source indicates that pits were visible at sites like this at the time and would presumably therefore have been visible to ancestral Mohawk people after having been filled ([114] at p.23): “These strange aboriginal magazines or storehouses can be seen in numbers even at the present day. Some of them are in the dense forests, others lie in open fields. They are looked upon with curiosity by the country people who regard them as the graves of Indians.” Such visibility, and the knowledge of the presence of these features (with differing Indigenous versus European/settler interpretation) may well have encouraged at least occasional activity at these loci—potentially for very different reasons (Indigenous versus European/settler)—which could explain instances of apparent late or post-use pits and high-up and near plow-soil charcoal material. We therefore interpret the samples we date as belonging to activity or activities ending with the firing represented by the date of the waney edge of the charcoal samples (thus a TAQ for the earlier use of the pit) and latest maize samples, with some older material likely from earlier (to initial) use of the pit.

Smith-Pagerie Feature 60 Longhouse 5 was a small and relatively shallow pit (36" diameter and only 24" deep: [76] at p.61). Thus, we assume that this pit likely represented just one real use episode. We therefore assume the maize dated (UCIAMS-218493) gives a date for the first use of the pit. The ¹⁴C however is much older than any other dated element from the site (even treating this as an initial site context). It is a ~98% probability outlier if assumed to be the first date from the site, and a 100% probability outlier if included in the subsequent main site Phase. This date is thus excluded from the dating models, whether as an unexplained outlier, or evidence of distinct and earlier human activity at the site locus.