|
Post by Admin on Jun 26, 2022 17:48:26 GMT
Most of the handaxes are now stored at the British Museum, while smaller collections are held at regional museums, including Canterbury and Herne Bay. Ashmore [20] reports that, from the original more than 330, only 223 are currently accounted for, with 189 of these being stored at the British Museum. We can add an additional handaxe to this tally, having found a single example within the collections of Canterbury Museums Trust (figure 1). An unknown number of bifaces, flakes and cores entered private collections or were lost as part of the aggregate output. The handaxes are typically described to be narrow, thick and of ‘ovate tradition’ through to ‘pear-shaped’ [2,4,15], although there is notable variation and the presence of ficrons and other more pointed forms [20] (table 1). They are also noted to be ‘rough’ in their form, display irregular edges, and to retain substantial portions of cortex in many instances (although there are exceptions; figure 2). Little is known about the flake artefacts or the presence of retouched tools. Table 1. Descriptive morphological data from 224 of the Acheulean handaxes recovered from Fordwich Pit during the 1920s, and now stored at the British Museum (n = 189), Herne Bay Seaside Museum (n = 34) and Canterbury Museums Trust (n = 1). Data for all artefacts are derived from Ashmore [20], except for the Canterbury Museums Trust handaxe. Note that some ratios were incorrectly calculated by Ashmore [20]; we provide corrected values. length (mm) width (mm) thickness (mm) point of max. width from tip (mm) elongation ratio (width/length) mean 142.4 74.3 47.2 89.9 0.53 min 58.0 41.0 19.0 17.0 0.31 max 230.0 113.0 77.0 163.0 0.98 s.d. 29.5 118.0 125.0 23.6 0.09 CV (%) 20.7 15.9 26.4 26.2 17.6 Figure 2. Four of the handaxes recovered from Fordwich Pit during the 1920s. Note that three display thick side profiles, elongated forms, and are relatively lightly worked (reduced) character (i.e. a, c and d could be considered ‘crude’). The fourth (b) is more heavily worked.
|
|
|
Post by Admin on Jun 26, 2022 20:27:10 GMT
3. Material and methods 3.1. New excavations at Fordwich Pit Discoveries made during the first fieldwork season at Fordwich form the focus of the present article and represent the start of the first major archaeological work at the site in its 100-year history. Artefacts were recovered from two contexts. The majority come from two 1 × 1 m trenches dug into a portion of preserved gravel terrace on the edge of the quarry. Owing to the COVID-19 pandemic, work during 2020 was initially limited to these two small test trenches (‘A’ and ‘B’; figure 3). Other artefacts were recovered through ad hoc field walking, or from surface finds in the immediate vicinity of the two trenches. The primary goal of the initial excavations was to confirm whether these sediments retained evidence for the presence of hominins in the form of lithic artefacts. To this end, three-dimensional point-plotting of the artefacts that were discovered was not undertaken. Instead, sediments were removed in 50 × 50 cm quarters with a spit depth of 10 cm. These were sieved through a 6 mm mesh screen with artefacts recovered being assigned to a particular 10 cm stratigraphic level, while a smaller number were recorded in situ. Where a spit cut across a sediment unit change, it was assigned to a layer as per table 3 depending on which layer contained the majority of the spit's depth (meaning that there is up to 4 cm of interchange between named layers in terms of artefact assignment). 3.2. Identification and analysis of lithic artefacts Due to the fluvial deposition of the sediment [2,20] we anticipated there to be occasional challenges when identifying artefacts. Thus, a highly conservative approach to artefact identification was followed. All stone objects displaying features suggestive of anthropogenic modification were collected through sieving or in situ excavation. These objects were then subjected to analysis by three experienced Palaeolithic archaeologists (TP, MP and AK), two of whom have considerable experience identifying lithic artefacts from high-energy fluvial sediments, including in British Quaternary sequences. For an artefact to be considered for inclusion in subsequent technological analyses, agreement on its anthropogenic origin between all three analysts was required. Artefact condition (degree of rolling) was subjectively assessed on a scale of 1–4, as is common in lithic archaeology (e.g. [10,41]). All artefacts were considered in natural lighting and made use of low-magnification optical aids. Once assessed, each artefact was grouped and compared with others of the same scale for verification. Methods to quantify differences in edge curvature (i.e. rolling) on lithic objects do exist [42,43], but for the present purposes these techniques were not deemed necessary. For reference, flakes ‘C’ and ‘D’ in figure 6 were scaled at level 4, flake ‘H’ was scaled at level 2, while the right flake in figure 8 was scaled at rounding level 1 (i.e. it was considered exceptionally fresh with little to no taphonomic damage). 3.3. Sample collection for infrared-radiofluorescence dating Sixteen samples were taken for infrared-radiofluorescence (IR-RF) dating (figure 3). The samples were collected from freshly cleaned outcrops using light-tide steel tubes. The tubes were hammered into the wall and (after removing them) immediately sealed with light-tide tape and stored in black plastic bags. Eleven were collected from the two excavated trenches (samples 1 to 11) while a further five were collected from the exposure investigated by Bridgland et al. [7] (samples 13 to 17; [note that sample 12 was purposefully missed). Within the trenches, samples were principally collected from sand lenses (samples 1, 2, 5, 6, 9, 10, 11) or the only visible band of sand (samples 4, 7, 8; figure 3). It was not possible to collect samples from the lowermost gravels as no sand lenses were exposed. Repeat samples were taken from several locations. Three samples could not be dated as there was only a very small amount of feldspar with the right grain-size (90–250 μm; samples 3, 6 and 8; table 2).
|
|
|
Post by Admin on Jun 27, 2022 17:20:27 GMT
To establish a chronological framework for the fluvial aggradation at Fordwich, we used IR-RF dating [44–46]. This method is a powerful tool for determination of the point at which feldspar grains were last exposed to sunlight, and hence to obtain information on the burial age of the deposits. The IR-RF signal saturates at doses mostly in between 1200 and 1500 Gy and thus allows the successful dating of Middle-Pleistocene sediments [47,48]. It is known that the IR-RF signal is more difficult to bleach if (for example) compared with the quartz optically stimulated luminescence (OSL) or infrared stimulated luminescence (IRSL) signals measured at lower temperatures (e.g. [44,49]). This makes the application of IR-RF challenging for young Holocene and mid-to-late Weichselian samples. Nevertheless, dose residuals should play only a minor role for Middle Pleistocene samples. Since Fordwich represents a fluvial aggradation of previously disposed stone tools the method will deliver the minimum age of the artefacts. Artefacts displaying considerable edge abrasion and rolling are likely to be older than the IR-RF dates indicate. Sample preparation was conducted under subdued red light in the luminescence and infrared radiofluorescence laboratory at the Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology (Leipzig, Germany). First, the outer 2 cm of sediment in the cylinders was removed to avoid any light-contamination. Subsequent treatment to separate the coarse K-feldspar grains included (as is standard) sieving and chemical treatment with 15% HCl, followed by the destruction of organic matter using 30% H2O2. The mineral-separation was done using sodium polytungstade. A density of 2.58 and 2.53 g cm−3 was used to separate the K-feldspar rich fraction. For De-measurements either the 90–250 µm or the 90–125 µm fraction was used. The wide grain size range (90–250 µm) was needed for most samples with respect to the limited amount of available coarse grain K-feldspar. All De-measurements were carried out on a Lexsyg research system with a freshly calibrated 90Sr/90Y beta ring source with a dose rate of approximately 0.51 Gy s−1. Figure 4. IR-RF curves obtained from sample L-Eva 2286. The red signal marks the natural IR-RF signal, measured for 5000s at 70°C. The green signal shows the regenerated IR-RF curve, that was detected during beta-irradiation after the aliquot was bleached. During the prior bleaching procedure, the IR-RF-related traps were emptied, and hence, the signal intensity was raised to its maximum level (high trapping probability per time). The IR-RF signal was detected through a Hamamatsu PM tube and the signal was filtered through a Chroma D850/40 interference filter. The IRSAR protocol [50] was used for all equivalent dose (De) measurements and 3–13 aliquots of 5 mm diameter were measured per sample, depending on the available amount of coarse grain K-feldspar. For data analysis the R software package (v. 3.5.1) was used including the R-luminescence package [51]. The age calculation was done with the Adele software (ADEL v. 2107). The mean De-value with its standard error was used for age evaluation respectively. Figure 4 shows representative IR-RF curves (natural and regenerated signal) from sample L-Eva 2286 and figure 5 shows the corresponding De-distribution of that sample.
|
|
|
Post by Admin on Jun 28, 2022 17:43:20 GMT
Figure 5. IR-RF De-distribution obtained from sample L-Eva 2286. The concentrations of U, Th and 40K were measured on the dried (bulk) sample material using low-level (high-resolution) gamma-ray spectrometry (N-type detector) at the VKTA (Felsenkeller laboratory) in Dresden, Germany. The specific activity of 238U was determined based on the activity of 234Th. 214Bi and 214Pb were used for 226Rn, 228Ac for 228Ra, and 212Pb and 208TI to get the activity of 228Th. The specific activity for 232Th is based on the weighted mean of the values for 228Ra and 228Th assuming an equilibrium for the 232Th decay-chain. To account for the cosmic dose rate contribution, the height a.s.l., geographical coordinates as well as the sampling depth below surface were considered. The evaluation of all dose rate components was done with the Adele software package (v. 2017) and the details on calculation can be found in Degering & Degering [52]. A value of 0.067 ± 0.012 [53] was used and the internal potassium content was estimated to be at 12.5 ± 0.5% [54]. The dosimetry data can be found in electronic supplementary material, table S1. 4. Results 4.1. Infrared-radiofluorescence results The IR-RF ages and their 1 sigma error outlined in table 2 show that there are three age clusters with estimates that overlap in their error range. Two clusters, one at approximately 372 ka and another at approximately 542 ka, are located in our recent 2020 excavation, while the third approximately 437 ka date cluster is from the 1998 exposure investigated by Bridgland et al. [7] (figure 3b). These multiple clusters point to several periods of fluvial aggradation and/or reworking of the deposits, probably interrupted by pauses in fluvial activity or incision, followed by later reworking of the uppermost sediments by other processes, probably related to the valley-side slope (electronic supplementary material, figures S1 and S2). The 347 to 385 ka ages for the upper main gravel and intermediate coarse flints at our excavations correlate with MIS 10/11. The lower main gravel broadly correlates with MIS 14. We interpret the younger age for the upper part of the sequence to be the result of later reworking of the uppermost part of the fluvial aggradation, as already described. Fluvial deposition at Fordwich is interpreted as having occurred during cold climatic periods with high-energy fluvial activity, potentially in a braided river system. Indeed, braided gravel-bed rivers point to periglacial climates in NW Europe, as does evidence of substantial gravel supply and deposition [24,55–57]. In combination with the abraded nature of many of the lithic tools we interpret the artefacts as having most likely been discarded by hominins during MIS 15. Note that MIS stage interpretations use central values and occasionally move one MIS stage up or down when more extreme error range values are considered. The Bridgland et al. [7] exposure returned ages of between 379 ± 21 and 455 ± 24 ka for the upper sand layer and loam, and a date of 433 ± 23 ka for the middle sand layer; stratigraphic levels that are not present at our recent excavations. Caution should be applied to the age interpretation of the middle sand layer due to the lack of repeat sampling.
|
|
|
Post by Admin on Jun 28, 2022 18:26:51 GMT
4.2. Archaeological units and stratigraphy
Two clear stratigraphic levels were present in each trench. The upper level is composed of topsoil intermixed with fine sand and flint gravel. There is a slight incline to the surface of the next (lower) level, which comprises dense gravel, running from trench ‘B’ to ‘A’ (left to right in figure 3) such that the first ‘soil’ section is 60 cm deep (Trench B) and the second is 48 cm deep (Trench A). This reflects an incline observed in the underlying sand (Lambeth Group sands) which affects the profile of all layer comparisons between the two trenches (i.e. levels in A are higher than in B) (electronic supplementary material, figure S1). Toward the bottom of the soil layer a thin band of loose gravel retains high levels of organic matter. Below this is the upper limit of the main gravel mass (i.e. the first geological unit), which continues for a further 190 cm.
We found no evidence of Smith's [2] clear band of sand above the main mass of gravel, nor the upper thinner layer of gravel and loam (figure 3). We are confident that this is not due to past quarrying as the presence of in situ Neolithic remains immediately superior to the excavated gravels demonstrates it to be undisturbed in recent times (see below). Moreover, it matches Smith's [2] description of an incline in the depth of the gravel from east to west. The only other remaining quarry exposures at Fordwich Pit do, however, closely match the description by Smith [2] in all ways (figure 3). These were the exposures investigated by Bridgland et al. [7] and clearly demonstrate the soil and loam overlying an upper, thinner band of gravel, which itself is overlying a sand layer. Beneath is the main gravel mass which becomes interstratified with sand toward its base.
The upper limits of the sediment at the 1998 exposure and 2020 excavation are broadly level, meaning that the stratigraphic units in the former are considerably deeper (both altitude and unit thickness) than in the latter (measured to the Lambeth Group sands) (electronic supplementary material, figure S2). Again, this is in line with the gravel depth (thickness) increasing toward the west of the pit as the base declines in line with the underlying slope. This means that the gravel at our slightly more easterly 2020 excavation rests on higher Lambeth Sands than the 1998 exposure. Notably, both the 1998 exposure and our new excavations are toward the extreme west of the pit where a letter contemporary with the original quarrying activity suggests most handaxe artefacts were found [4]. Our excavations are approximately 100 m north and approximately 30 m east of the Bridgland et al. [7] exposures, and closer to the edge of the valley. The Bridgland et al. [7] exposures are further from the valley edge and are positioned immediately before (i.e. marginally easterly to) a small fluvial gulley that demarks the edge of the pit. We interpret the absence of the loam, upper gravel layer and sand band at our 2020 excavation as little other than the result of lateral facies variation within the terrace sediment body. Indeed, the MIS 12 sediments at the Bridgland et al. [7] exposure probably result from later fluvial activity aggregating gravel above portions of the MIS 14 sediments (electronic supplementary material, figure S2).
At a depth of 50 to 80 cm from the surface of the gravel, flint sizes become noticeably larger in both excavated trenches, regularly reaching 10–15 cm in maximum diameter. Several sand lenses were visible within or just above this approximately 30 cm deep intermediate coarse flint layer (figure 3a). We term the finer gravels above this intermediate coarse layer the upper main gravel, while the finer flints below it are termed the lower main gravel. Clast size varies in these gravels, but flints are typically 1 to 7 cm in diameter and are supported in a cemented matrix of fine-grained clasts containing iron (as is typical for gravels of this age; electronic supplementary material, figure S6). In Trenches A and B, the intermediate coarse flint layer is underlain by an approximately 15 cm thick band of sediment with a high sand content. This could potentially align with the lower sand layer from the original Smith [2] description and later Bridgland et al. [7] exposures (figure 3b; electronic supplementary material, figures S1 and S2). Although the substantial (greater than 100 m) distance between the locations introduces some uncertainty. Despite only being approximately 20 m apart, there is variation in the sand and loam layers between the two 1998 exposures (figure 3; electronic supplementary material, figure S5). Such lateral variation is entirely to be expected in a fluvial depositional context; especially one interpreted to be part of a braided river system. The lack of clays and silts in the excavated gravels, along with evidence of bedding structures, supports our interpretation of a fluvial origin for these sediments (i.e. flowing water) as opposed to other process (e.g. solifluction). Variable clast sizes and the presence of a high sand-content layer point to diversity in fluvial energy through the sequence.
Roughly 145 cm from the top of the gravel in Trenches A and B sediment was noticeably looser with a higher sand content. Beneath this approximately 15 cm of looser, sandier gravel there was a continuation of the tightly packed gravel for a further 70 cm (electronic supplementary material, figure S1). Again, this matches the lower stratigraphy described by Smith [2] and exposed by Bridgland et al. [7]. With the exclusion of the loam and upper gravel/sand layer, the gravels at our recent excavations therefore broadly match Smith's [2] description of the main gravel mass from which the Fordwich handaxes were most likely recovered. Future work is planned to gain additional insights into the complex fluvial architecture of the site. This will provide a better understanding of the distinct MIS 14 and MIS 12 periods of fluvial aggradation, along with the erosion and/or reworking processes. Further IR-RF samples will be collected to aid this process.
|
|