Figure 3: A picture of the block in situ, in trench SC24.
Figure 4: Sawing the block at the end of the excavation (NM and CC).
The block of sediment to be excavated in 2007 at Star Carr was taken from SC24, the only trench open at the time with wet deposits. It was not removed from the section but was left within the trench and excavated around in order to tie up the fieldwork observations with those in the laboratory (Fig. 3). This was particularly important at the base of the archaeological sediments, which contained large quantities of wood, including some large timbers that extended into the block. The block was first carefully sawn into four pieces (Fig. 4), first vertically down the centre on a N–S alignment, and then through the middle horizontally. This was carried out in order to test the impact on the archaeological remains of cutting through the peat, and to see how feasible it was to match together deposits. The blocks of peat were then lifted onto metal sheets, wrapped in plastic and labelled.
The block was transported to the laboratory in York (Fig. 5). It was approximately 0.6m square and 0.5m in height. It was set up on a level sump tray in a made-to-measure plywood crate with two glass sides. The glass sides were marked with a grid at 1cm intervals to facilitate the measurement of finds in three dimensions.
Difficulties were encountered reassembling the four block sections. Firstly, it was discovered that large quantities of mineral concretions and white moulds had formed on the outsides. The blocks were also stuck to the metal sheets on which they had been carried back from site. Once they were carefully sliced from the sheets, the block sections had to be positioned in the box. Some shaving of the sides had to be carried out to facilitate this and the upper block sections were much harder to position, particularly the upper west section which went in last: it was impossible to move the blocks without some material falling off the surfaces. This was particularly the case on the interface between the upper and lower pairs of blocks. As a result, the upper west section was not level with the upper east section, being approximately 1cm higher because of the presence of loose material at the interface.
An angle-poise lamp with magnifying glass was set up to facilitate the recognition of particularly small remains, as was used for the excavation of the Cambridge block. The block was kept moist throughout the period of excavation by wetting it with water that was taken from the trench at the end of the excavation, using a large 50 litre plastic container (to avoid affecting pH by the use of York tap water).
A camera and lights were mounted above the block, enabling photographs to be taken of each exposed layer and any interesting finds or wood that emerged. A number of set-ups were attempted before a reasonably satisfactory lighting arrangement was found. The magnifying angle-poise provided satisfactory working light, but only over a limited area. Two photographic lamps were therefore tried: firstly, one of the 650W halogen lamps from a copy stand, followed by a 150W-equivalent energy-saving bulb. Although the first bulb was just sufficiently bright for photography and, given the angle at which it was set, would have provided good shadows for showing the texture of the surface, it 'blew' after a few days, but because replacements were only available from Germany and because such bulbs are only designed to last around 48 hours and not intended to be left on for long periods, it was decided to abandon it. The energy-saving bulb was not bright enough to facilitate photography and barely aided the angle-poise as a working lamp. The best compromise was achieved, finally, by the use of a commercial hot-shoe-mounted flash unit directly attached to the camera. The disadvantages of this was that it made it difficult to see the texture of the sediment and it reflected off the sediment if it was particularly moist, and off the glass sides of the box, causing odd sections of brightness.
The excavation was carried out by only one person, PH, to ensure consistency, under the supervision of NM. He excavated by repeatedly slowly cleaning back with bamboo and plastic tools where the sediment was sufficiently robust, and modelling tools were used around the delicate wood. The sediment was removed and recorded at spits of 1cm thickness. Each layer was labelled BL## (Block Layer) and 53 layers were excavated in total (with the first layer being a cleaning layer and hence not recorded below). Any finds that were greater than 1cm in thickness were left in place until they were fully exposed and then were removed for analysis.
Initially, all the wood from each layer was recorded. Each piece of wood was given an ID number, recorded on paper and later on a spreadsheet, and then bagged individually. The location of each piece was recorded by printing out a photo of each layer and labelling it by hand. This process continued for the first five layers. These layers were slightly less than 10mm each, approximately 30mm overall, and this excavation took 244 hours. As a result, it was agreed that the process was far too time-consuming and needed to be rethought. Therefore, from layer BL06 downwards any wood fragment less than 5mm in diameter and approximately 50mm long was retained in the sediment sample, but any larger pieces or chips or unusual pieces were bagged individually and recorded as before. Sediment was retained, bagged and labelled for each spit. About a quarter of each layer was analysed for plant macrofossils. The rest of the sediment was analysed for mineral content using loss-on-ignition. Residues from both processes have been retained.
MT visited the laboratory twice, and with PH examined every piece of wood that had been recorded and bagged in order to identify any pieces that may have been worked or showed evidence of woodworking, such as wood chips: an important part of the recording method is to identify by-products of woodworking so that they can be characterised. The species of the pieces selected as interesting by MT were identified by AH. The wood was generally rather soft, but in only one case was an identification impossible because hand-cut sections could not be secured (and one case where root wood is suspected to have been present).
All possible worked wood was recorded in detail using pro-forma wood sheets. These forms have been developed to aid the analysis of prehistoric woodworking and are used in conjunction with a computer database currently being prepared to go on line by Michael Bamforth of L–P Archaeology in London. The forms were originally used in 1982 (Taylor 1998), but have subsequently been further developed and modified.
AH analysed the sediment samples for plant macrofossils and the even-numbered layers were examined. Material was submitted in the form of wet residues from disaggregation of the samples. Methods broadly followed those described by Kenward et al. (1980). Disaggregation of the residues was generally far from complete: students engaged on this task were encouraged not to be too rigorous in breaking samples down manually (in any case, the peats proved rather intractable!). This resulted in both reduced damage to the sometimes very soft wood fragments (some of which may bear evidence for 'working' by some agent) and the recovery of clasts of undisaggregated detritus peat up to 10mm or more which were then available for closer examination using low-power microscopy during recording of plant remains.
For the assessment of the wet residues, the bagged residues were gently resieved to 0.3mm and a subsample of about 100cm³ selected randomly from the material in the sieve using a spoon and a graduated beaker. These subsamples were then resieved to 4, 2, 1 and 0.3mm and all of each fraction (apart from the finest) checked under the low-power binocular microscope. Notes on the nature of the material present and a list of plant taxa and other components were recorded directly to a database on a personal computer. Time constraints meant that a detailed assessment of preservation of the plant remains, along the lines of that recently suggested by Jones et al. (2007), could not be attempted, and their methodology is not, in any case, concerned with recording the nature of the matrix, merely to note it contained plant macrofossils and pollen as proxies for overall preservation.
A high-resolution series of loss-on-ignition tests were undertaken in the hope of better understanding the lake-shore peat formation process. It was hoped that a high-resolution strategy might be able to enrich our knowledge about possible silting events or flooding and drying events when married with the other techniques undertaken. Further, it was hoped that any large peaks or troughs in organic content throughout the sequence might be able to help explain any associated anomalies in the series of experiments being undertaken with other specialists. This sequence should give an idea as to how stable the environment was over time.
Samples of the sediment for loss-on-ignition were processed by AN. These were taken from each of the block layers and a subsample of each was air-dried for four days at room temperature. Once dry, samples were ground using a pestle and mortar so as to ensure the sediment was of uniform size prior to loss-on-ignition testing. Two samples, BL24 and BL41, having been overlooked during the drying procedure, were dried separately for 24 hours in a drying cupboard at approximately 40°C. This was to avoid delay and ensure efficiency when using the oven. Organic material, most notably fragments of wood, larger than the 'size of a fingernail' were omitted from the pestle and mortar to preserve material that might prove useful in other modes of analysis in the future, and so as to analyse the sediment and not to bias the loss-on-ignition results.
About 2g of sediment was taken from each dried and ground sample and weighed to three decimal places with an APX-153 Denver Instrument. These samples were then placed in labelled crucibles which were fired for two hours at 850°C. The samples were removed from the oven and placed immediately in a desiccation chamber and left until entirely cool, a process that took about 20-30 minutes. Samples were then weighed again using the same scales and degree of precision.
© Internet Archaeology/Author(s) URL: http://intarch.ac.uk/journal/issue28/2/3.html
Last updated: Thu Mar 25 2010