C14 dating for sediments from F44, F45, F46 and Find 14

Richard Tipping, Department of Environmental Science, University of Stirling. 2000.

Summary of the Report

The results of C14 dating from the bases and 'tops' of F44, F45, F46 and Find 14 are presented.
F44, F45 and F46 began receiving organic sediment at separate times in the Devensian Lateglacial and in the early Holocene.
The C14 assay from Find 14 shows the lake to have been replaced at the dated site in the early Holocene.
There are no grounds for rejecting the C14 assays, although there is potential for 'hard water' error making the C14 assays 'old'.
A test of the C14 assays through biostratigraphic correlation is recommended as an immediate measure.
If correct, the C14 assays suggest that F44 - F46 are natural features caused through collapse of subterranean structures, and they are not now seen as anthropogenic features.
C14 assays from the highest points in the stratigraphies of F44 - 46 show that stratigraphic integrity is maintained until the later prehistoric period.
The small sizes of these features and their longevity provide an exceptional and exceedingly rare opportunity for fine-detailed, spatially precise reconstructions of Holocene environmental change that can be linked securely to the archaeological record at Nosterfield.

Introduction

On 30th April 2000 the case for and up-to-date costs of commercial C14 dating of seven sediment samples from F44-46 and Find 14 were presented. Samples for C14 dating were submitted to Beta Analytic Inc. (Florida) on the 17th May 2000. This report will (a) present the results of these assays, received on July 5th 2000, (b) discuss the significance of the assays for interpretation of the sequence, (c) evaluate the significance of these sediments in relation to archaeological and palaeo-environmental records, (d) recommend additional analyses that should be undertaken following this construction of chronologies and (e) advise on costs for this.

Invoices for C14 dating and preparation of materials will be sent by separate post. Agreement was reached with Mr. Copp that I should pay for the C14 dating initially. (Beta Analytic Inc. will thus invoice the 'University of Stirling') but prompt payment of the invoices would then be appreciated.

C14 Samples

Sediments from the monoliths and cores from F44, 45 and 46 had previously been identified (Long and Tipping: 4th November 1998) as sufficiently organic for C14 dating. The samples were submitted were:

F44 Core III 239-240.0cm
Monolith I 0.0-1.0cm
F45 Core III 280.0-281.0cm
Monolith I 0.0-1.0cm
F46 Core II 219.0-220.0cm
Monolith I 8.0-9.0cm

The sediment sequence from Find 14, a 25.0cm monolith sample, had to be analysed further to define (a) organic content and (b) carbonate content. Data are not presented here, but this work defined a depth within the tin of 9.0cm (39.1 Sm OD) when organic content by loss-on-ignition rose over 4.0cm from 10% to 73% and carbonate content, also by loss-on-ignition (Tipping: 30th April 2000) fell from 53% to 3%. This is regarded as a reliable indicator of the transition from calcareous marl to peat, and a 1.0cm slice at 9.0-10.0cm was submitted for C14 dating.

C14 Assays & Interpretations

The assays reported by Beta Analytic Inc. are given in Table 1. Samples were assayed by standard AMS treatment.

C14 assays from the basal organic sediments at F44, 45 and 46 all indicate that the predominantly organic sediment infills of these features began to accumulate in the Devensian Laterglacial or early Holocene. There is no correlation between feature depth and age. F44 seemingly began to receive sediment at the end of the Lateglacial Interstadial (Beta-143455), F45 at the end of the Loch Lomond (Youger Dryas) Stadial (Beta-143456) and F46 within the early Holocene (143457). There is no evidence available to question these assays, although uncertainties arise over sources of carbonate within the sediment introducing 'hard water error'. The sites lie on Magnesian limestone at depth, and the gravels surrounding the features contain Carboniferous limestone clasts. Pre-treatment by two acid washes (Beta Analytic Inc. pers. comm.) removes free carbonate but cannot isolate carbonate preserved within plant fragments which would induce 'ageing' errors. The only guides currently available are the _13C contents of assays, and these do not indicate contamination. Tests of the C14 dating results using palynological data are suggested below, but there is currently no reason to reject these assays.

There is no stratigraphic significance to the series of dates on the highest organic sediments in each feature. The tops of the features have undergone weathering, pedogenesis and mineralisation. These assays measure the ages of the youngest peats that have remained unaltered by soil-formation, and this can be expected to be dependent on original peat thicknesses within the features and differential processes of ancient and recent truncation and drainage. These assays do not measure the ages at which peat-infilling ceased: this cannot be known. There is no pattern to the C14 ages but one is not expected. F44 loses stratigraphic 'integrity' in sediments younger than the late Bronze Age (Beta-143453); F45 in sediments younger than the mid-late Iron Age; F46 has no reliable stratigraphy after the early Bronze Age. Again there are at present no grounds for rejecting these assays.

The single assay from Find 14 dates the earliest occurrence of terrestrial (fen) peat over lake sediment (marl). This contact is sedimentologically conformable with alternating bands of marl and peat. The peat is of early Holocene age (Beta-143458).

Table 1

Feature	Depth	Altitude	Sediment	Wet wt. (g)	Lab no.	C14 age±σ	δ¹³C	Calib age BC
F44	239-240.0cm	36.66mOD	Peat	1.87	Beta-143455	11140±60	-26.2	11005-10960
	0.0-1.3cm	39.10mOD	Organic clay	2.92	Beta-143453	3110±30	-28.2	1405-1260
F45	280.0-281.0cm	37.52mOD	Organic silt	2.36	Beta-143456	10180±60	-26.2	10370-9605
	0.0-1.0cm	39.32mOD	Peat	1.92	Beta-143452	2330±40	-29.0	395-200
F46	219.0-220.0cm	36.88mOD	Peat	1.05	Beta-143457	8900±50	-27.4	8220-7780
	8.0-9.0cm	39.13mOD	Peat	1.66	Beta-143454	3930±40	-28.1	2470-2210
Find 14	9.0-10.0cm	39.15mOD	Peat	1.68	Beta-143458	9380±50	-28.4	8705-8440

Test of the C14 Assays

A test of the veracity of the C14 dates is recommended as an urgent measure. For this the sediments infilling F45 will be pollen-analysed rapidly ('skeletal' counts of 150 total land pollen) at 4.0cm intervals between 281.0cm and 200.0cm (20 subsamples). The base of this sequence is C14 dated to the earliest Holocene. The lowermost sediments should, then, contain a clearly identifiable pollen sequence depicting the postglacial migration of tree taxa from before 10000 and 7000 C14 BP (Birks 1989). Regional biostratigraphic correlations will confirm or refute the C14 dating of the base of this sequence.

Interpretation of the Features

The features are distinctive in their circularity and their high depth:diameter ratios. The features are presumed also to have near-vertical sides, although this was not demonstrated from excavation or probing. The shapes of the features allowed the suggestion that these were anthropogenic features, and on a gravel substrate, that they may have been wells.

This interpretation is rejected here. Although not conclusive in itself, the ages of the basal sediments infilling the features suggests an anthropogenic origin to be unlikely. They can still be of anthropogenic (Late Upper Palaeolithic and early Mesolithic) origin, but their antiquity and variation in age of formation probably preclude this. The 'deliberate' cuts in the gravels argued for in the November 1998 report must be challenged, and it is likely that the major product of collapse was the wide funnel-shaped entrances to the shafts, which were considered late in the sequence in earlier interpretations.

The features are probably natural features within the already-formed fluvioglacial gravels bordering the River Ure. Their origin is thought to lie as small collapse features within the gravels as a result of cavern-collapse in underlying limestones. This interpretation can explain the circularity of the features, and their close concentration may reflect the collapse over time of one large cavern. The interpretation does not wholly explain the depth and narrowness of the shafts but these may be determined by the nature of the collapse beneath. There appears to be no discernible environmental trigger for the collapse. The C14 dates show that collapse occurred over c.3000 cal. years in a range of climatic conditions. There are insufficient data to argue more than that the timing of separate collapses is governed by chance. Alternative explanations of origin are rejected on morphological or geomorphological grounds. The features are not periglacial in origin; nor can they be scours induced by flowing water.

Collapse is assumed to be followed closely in time by sediment accumulation. Sediment fills recorded in the report of November 1998 (Long and Tipping) are quite variable, though all indicate deposition in water-lain (pond) or waterlogged (peat) environments. Rates of sediment accumulation must have been highly variable, and no attempt is made here to interpolate ages between the basal and top C14 dates in the fills. The sequences may indeed, over this long a period, contain depositional hiatuses.

However, the C14 dates allow some appreciation of the variations in sediment fills. In F44 between 229-138cm is a structureless minerogenic clay which may now be seen as a product of deposition within the Younger Dryas Stadial (11000-10000 C14 BP). Such minerogenic sediments are absent at both F45 and F46 which formed at the beginning of and within the Holocene and were infilled with much more organic sediment.

The lake adjacent to the features was 'terrestrialised' in the early Holocene. Although this C14 date correlates with evidence elsewhere in NW Europe for a marked phase of aridity and lake-level fall (Tipping 1996), little can be said with confidence on this from a single assay. No further work is recommended at this site since the lacustrine history is not relevant to the archaeological record.

Potential of the Sequences for Palaeoenvironmental & Archaeological Records

If the test of the C14 dates (above) confirms the antiquity of the basal assays, as is anticipated, then the features have received organic sediments for a very considerable period. Although hiatuses are perhaps likely they are not detectable from current measures, and broadly these features contain sediments from before the earliest Mesolithic to, at F45, the late Iron Age. This changes the potential of these features for archaeological and palaeoenvironmental interpretation, from their being seen as single-period records to near-complete prehistoric sequences, but it does not reduce their significance.

The infills of these features represent quite exceptional sediment records of landscape change and human activity which can be defined through the use of palaeoecological techniques. The most important aspect of these features for such interpretations is, curiously, their small diameter: the shafts are <3m across. What makes these sites exceptional requires a diversion to state-of-the-art concerns in the key palaeo-ecological technique, pollen analysis (palynology). Of over-riding importance in current work is the need to define landscape change at spatial scales relevant to human populations. Most pollen diagrams do not do this because the size and type of pollen site (large lakes or large peat-bogs) mean that pollen originates from very large, vague and unspecified sources kilometres away. The current failure to establish close links between vegetation change, human impacts and archaeological records is partly because the sizes of landscapes we each measure are different. Pollen analysts have learnt to refine the spatial resolution (pollen source area) of sites such that, in simple terms, basin diameter provides a good estimate of the scale of landscape being depicted: the smaller the pollen site, the smaller the area depicted.

The Nosterfield features will in wooded conditions have received pollen from 50-70m around the features. To be able to state this using empirical models indicates the power that this control provides. At sites like these we can operate at a scale where human beings become part of the landscape. Woodland disturbance, the use of fire, clearance, settlement and agriculture all become tangible and linked to archaeological records and chronologies. Scales of activity can be defined, and the ecological consequences of human activities explored. Such sites are exceptionally rare and this cannot be stressed enough.

This spatial precision was always the real palaeoecological value of the sites, but we had not appreciated the huge timespan covered by the deposits. Given continuity of sediment accumulation, it is possible to understand the Neolithic finds at Nosterfield, as we hoped, but also now to define earlier, Mesolithic, impacts and later prehistoric developments.

In the next section different approaches to environmental reconstruction are briefly reviewed and recommended. Here I review the needs to analyse all three sediment fills. The three features at Nosterfield cover essentially the same time-span, and all the features reflect the same small pollen source area. There is no advantage to analysing the three sediment fills. The choice of site is determined by (a) the interests of the investigation. (b) timespan covered by the sediments and (c) complexity and information content of the sediments. F44 contains Devensian Lateglacial sediments. Although interesting in themselves, the concern of the work is to explore environmental changes related to the archeological record from the area, which is entirely of Holocene age, and so analysis of F44 is not recommended. F45 represents the longest sequence of sediments, reaching to the mid-late Iron Age before the stratigraphy is disrupted, much younger and more complete than F46. Palaeoecologists often seek the simplest sediment sequences to provide undisturbed contexts, but the most interesting work comes from complex sequences which link ecological, geomorphic, hydrological and archaeological change: F45 provides the best context for this. F45 is recommended for full analysis.

Palaeoenvironmental Analyses Recommended on the fill of F45

281.0cm of sediment spans the time from the earliest Holocene to the late Iron Age. The sediments were initially formed in standing water with clays and bands of amorphous organic matter. Later peats are on occasions interbedded with standing-water clays (Long & Tipping: 30th November 1998), and there is clearly a complex relation between the water-table, peat growth and sediment inwashing. The following analyses are seen as cost-effective yet comprehensive approaches to the integration of palaeo-environmental data. Costs are considered in the next and final section.

X-ray analysis of the cores: X-ray analysis has become an important non-destructive technique for defining the complexity of macro- and micro-scale sediment stratigraphies. These are often not visible. The complexity of the sequence at F45 suggests that much subtlety will emerge from X-ray analysis, and this is indispensable to defining the changes in depositional environments, defining the subsampling positions of palaeoecological samples and their interpretation.

X-ray analysis will be undertaken at the purpose-built facility at the British Geological Survey in Edinburgh. The time taken is 3.0 days for preparation of samples and photography. Materials are charged at £90.00 per exposure, and the three cores and one monolith may require three exposures at different settings to obtain maximal clarity.

Image analysis of X-ray photographs: X-ray exposures can be further interpreted by quantitative measures of the light-dark contrasts on X-ray plates by computer-driven image analysis. The University of Stirling has this facility. Quantification of X-ray patterns should define similar deposits and depositional conditions, and refine sedimentological interpretations. This application is novel and experimental, but is rapid and non-destructive. 2.0 days are allotted to this with equipment/consumable costs waived.

Magnetic Susceptibility: contrasts between mineral and organic sediments are further defined by differences in ferromagnetic iron minerals, and this is measured by volumetric magnetic susceptibility on intact cores (the technique is non-destructive) on a Bartington Instruments MS meter. 1.0 day is needed.

Further work on magnetic properties can refine source areas for inwashed sediment (topsoils; subsoils; substrates; imported (e.g. non-local) material) if volumetric measurements suggest sensitivity in the signal. This work will use the mineral residues after loss-on-ignition, so is destructive. This will require 5.0 days work at the Chemistry Department, University of Edinburgh, with equipment/consumable costs waived.

Sediment Properties by Loss-on-Ignition: the first destructive technique will be the definition of (a) water content by oven-drying at 1050C for 8 hours, (b) organic contents by loss-of-ignition after furnacing at 5500C for 4 hours and (c) carbonate content by further furnacing at 9500C for 2 hours on 2.0cm thick subsamples (140 samples). These are needed to characterise depositional environments. The techniques are routine and will be performed by technical staff over 5.0 days.

Geochemical Analyses: Inwashed sediments are derived from the immediate surroundings of the site and will retain evidence of the rates of soil development and deterioration. Standard laboratory X-ray diffraction analyses of Ca, Al, K and Fe will be obtained on mineral bands to define these. Mound 40 analyses may be needed, undertaken by technical staff.

Pollen Analyses: pollen analyses to understand vegetation change, woodland disturbance, clearance and agricultural practices need not be defended, but pollen data on aquatic and wetland taxa will also define the presence-absence of standing water and water-depth. Measures of pollen concentrations allows insights into rates of sediment accumulation too subtle to be defined by further C14 dating (below). Associated analyses of pollen preservation can identify water-table fluctuations and periods when the shaft dried out. Analyses of microscopic charcoal provide extraordinary insights into fire regimes within undisturbed and anthropogenically altered landscapes. Counts of sulphide spherules produced in anoxic environments allow an understanding of water-quality and depth.

These separate and independent analyses come as a package: all are extracted from the same microscope slides. The delicate sediment slices sampled for pollen analysis (<0.5cm thick slices) mean that decisions of temporal resolution have to be made; broadly, how often do you subsample the time-sequence - at decadal intervals or at coarser levels? This of course affects the time, labour and costs. This estimate is difficult because it presupposes the patterns to emerge from the analyses. Natural woodland dynamics need to be examined at time-intervals of 150-200 years; post-Mesolithic activities are perhaps best analysed at 50-75 year intervals (e.g. 1-3 human generations) in order to establish settlement continuity/change; Mesolithic signals can, however, be missed at this temporal resolution.

These estimates have determined the estimates of numbers of subsamples. Between the basal sediments (10200 cal. BC) and around 4000 cal. BP (to be defined by further C14 dating; below) analyses will be at 200-year intervals, or around 30 subsamples. Between c.4000 cal. BP and c.300 cal. BC the temporal resolution will increase to c.75 year intervals, requiring around 50 subsamples.

Pollen analyses will be to 500 total land pollen, and will include taxonomic recording to the highest resolvable level, measurements of pollen concentration, preservation, microscopic charcoal and sulphide spherules. The c.80 subsamples have to be processed which will take 20.0 days, and consumables costs are £80 in total. Analyses will take 80.0 days and will require analysis at post-doctoral level.

C14 Dating: two C14 dates will not suffice to describe the chronology of these complex sediments, and eight further analyses are costed through Beta Analytic Inc.

Report Preparation: the preparation of a report will combine and synthesise all the physical, geochemical and palaeoecological analyses described above, and will incorporate archaeological data from Nosterfield and the surrounding area following consultations. The writing of the report will take 10.0 days.

References

Birks, H.J.B. 1989. 'Holocene isochrome maps and patterns of tree-spreading in the British Isles' Journal of Biogeography Vol 16: 503-540

Tipping, R. 1996. 'Microscopic charcoal record, inferred human activity and climate change in the Mesolithic of northernmost Scotland' in Pollard, A. and Morrison, A. (eds.) The early prehistory of Scotland (Edinburgh)

archaeological planning consultancy > thornborough > c14 dating for sediments from f44, f45, f46 and find 14