Brysbaert, A.: May 1998

A LATE BRONZE AGE SICKLE FROM SHINEWATER PARK: THE TREATMENT OF A WATERLOGGED COMPOSITE

Ann Brysbaert
bruno@staf.planetinternet.be
Archaeological Conservator/Archaeologist
Willem De Roolaan 68A, B-8620 Nieuwpoort, Belgium

Abstract

A late bronze age sickle was found at Shinewater Park, Sussex, England in 1995. A thoroughly researched conservation plan was required in order to meet the display conditions of the receiving museum and to meet the high standards required by the specialists involved in the treatment of this unique sickle. The sickle was lifted in a block of soil and then cleaned while still waterlogged. This was the initial step of treatment before the proper investigation and condition assessment could be initiated. After thorough discussion and research on possible treatment options, the alcohol-ether treatment was decided upon, a three-dimensional mould was produced and a design for a display mount was suggested. The necessary communication with specialists from several fields was very enriching, as was the depth of research which went into every step of the treatment process of the sickle.

Introduction
A waterlogged hafted sickle (Figure 1) was excavated in the summer of 1995 at the site of Shinewater Park (close to Eastbourne, southeast England) by the Southeastern Archaeological Service under the direction of Chris Greatorex (Figure 2). The site was discovered during the excavation of a lake which forms part of a new community park being developed by Eastbourne Borough Council. The object was found in a peat environment and was blocklifted from site in this soil (Figure 3) (Woodcock 1995).

Figure 1: Line drawing of the hafted sickle by Jane Russell. (109K)

Figure 2: Location of Eastbourne on the map of Great Britain. (40K)

Figure 3: The hafted sickle as received, blocklifted in peat. (50K)

Some of the remains found on site include post alignments and other wooden structures, a skeleton of a child, antler artefacts and several copper alloy objects, including the sickle. The latter finds and the post alignments have suggested to some a comparison with the site of Flag Fen in Cambridgeshire, according to M. Taylor, an independent wood specialist (personal communication). Together with the other copper alloy objects, the sickle was recorded to have lain horizontally in the acidic peat. The pottery from the site gives a date around 800-600 BC, which is Late Bronze Age or Early Iron Age.
The finds were sent to the English Heritage conservator for South East of England, Adrian Tribe, who is based at the Institute of Archaeology, University College London. After formal agreement with the owners and relevant specialists, conservation and investigation of the sickle was assigned to the author.
Problems which were expected in the conservation of the sickle can be described briefly as follows:

The sickle was a composite (wood-metal) waterlogged object which could not be taken apart.

Its age made it a very special object; it was the first ever found of its kind.

Several specialists were involved in the study and processing of the sickle; careful communication had to be carried out before each step could be made.

In view of these problems, the aims of the project can be outlined as follows:

to meet and facilitate the agreed steps of research and treatment in order to render the object stable and presentable for display

to learn about the conservation, investigation, and analysis of composite and waterlogged objects

to experience how to work and function in the framework of an ongoing project, involving other conservators, find specialists, and experts on different materials

Object description and technology
The object could be partially described before it was taken out of its surrounding peat, and more details became available after lifting and cleaning procedures were carried out.
The sickle consisted of two main materials. The metal of the blade was considered to be a copper alloy and the wood of the handle was identified through a sample as Field Maple (sample taken and identified by M. Taylor).
Its sizes:

length metal blade: 11.2 cm

length wooden handle: 16.5 cm

length metal socket: 6.2 cm

length curved part of handle: 2.5 cm

diameter of socket: 2.7 cm

weight of sickle (wet): 175.12 g

weight after treatment (dry): 89 g

It is not entirely clear how the slightly curvy metal blade and hollow socket with two rivet holes would have been shaped. It may have been cast in an open, single piece mould. Ridges ran along the blade on both sides.
The grain of the wooden handle ran along its length. The end of the handle formed a clearly cut curved part, nicely rounded and worked. This curve probably stopped the hand of the user from sliding off the instrument during usage. This curved part also gave an idea of how the object could have been used: a straight cutting movement was made towards the body of the user. A small hole went through the thickness of the handle about 2.5 cm below the metal socket. The hole seemed to be deliberately made because of its square section. Taylor suggested that if it was purposely made, a rope could have been put through it to enable the owner to carry the sickle on his belt. But the option that the hole was the result of animal activity during burial was not excluded.
The connection between the wooden handle and the metal socket was achieved by the use of two rivets. However, since the diameter of the wood was hardly smaller than the diameter of the metal socket, the wood might have been slightly cut at the socket end in order to fit in the metal piece as an extra way of fitting both parts together tightly. The wood obviously had also swollen because it was buried and became waterlogged.
Condition assessment and investigation
The metal part of the object was complete but some of the corrosion became detached during the removal of peat. These pieces were kept as a sample for XRD analysis. Some areas showed green, red, and pinkish corrosion products, and in other areas the copper seemed to have re-deposited on top of the existing red corrosion layer. According to J. Spriggs of York Archaeological Wood Centre, this was caused by the fact that when an object becomes buried, it initially corrodes fast. When, however, the burial deposit becomes anaerobic and thus creates a reducing environment, the copper reduces back to metal again and deposits over the existing layers. This was shown quite clearly on the blade and around the socket. The metal surface was pitted and etched.
The wooden part looked intact but some damage (pits) was observed. On the most damaged side, some pits had been caused during the excavation of the object, but some pointed stone chips had apparently been partly pushed into the wood during burial. From the analysis of the samples taken by M. Taylor, the wood seemed to be in a very good condition. J. Watson of the Ancient Monuments Laboratory, English Heritage, did not see any trace of fungal attack or resinous deposits (personal communication). The pits in the vessels seemed to be intact and so was most of the spiral thickening. At the end of the handle towards the curved part, small roots grew out of the wood. On the most damaged side of the handle, the curved end suffered most loss. The handle showed metal stains close to the socket (Figures 4, 5).

Figure 4: Interface metal socket-wooden handle: copper salts drawn into the wood (photomicrograph). (53K)

Figure 5: "Waves" of copper salts from the socket into the wood of the handle (photomicrograph). (55K)

X-radiographs showed that the density difference between the rivet holes and the metal is very high. This suggested that these holes were filled with wood rather than with metal pins. This suggestion was agreed upon by M. Taylor who came across more of these wooden pegs. She confirmed that very often sapwood was used for pegs of this kind. When the object was examined under higher magnification (65x), only tiny fragments of wood could be seen in the rivet holes.
The connection between the wooden handle and the metal socket seemed to be quite strong but care was taken during handling because the interface was considered the weakest place of the object, merely due to the weight of the metal part versus the wooden part.
The handle was dated by P. Pettitt of the Radiocarbon Accelerator Unit, Oxford, using accelerator mass spectrometry. The result was 2655 ± 50 BP. This accorded well with the British Museum dates for the timber platform, and confirmed the use of the site in the 9th century (calibrated), according to S. Needham from the British Museum Department of Prehistoric and Romano-British Antiquities (personal communication).
Because of its uniqueness and high archaeological value, treatment of the sickle had to be approached with well established methods with known long-term results. There was no scope for any error, so the following issues had to be taken into consideration:

both the wood and the metal parts had to be stable and had to stay together during and after treatment

no novel methods could be considered because of their unknown impact in the future

the chosen treatment had to be in line with health and safety regulations

although chemical and physical changes in the wood were expected after treatment, they had to be absolutely minimal

everybody involved in the project or the treatment process had to agree on the chosen method

the moulding material, chosen to make the three-dimensional record of the sickle, had to meet specific requirements to suit its purpose

environmental conditions after treatment for both storage and display had to be optimal for both the metal and the wood

special attention had to go to the design of a supportive but attractive display mount to allow the stable sickle to show its value in the museum

Treatment procedures
After the necessary literature research, the following tests were carried out:

different methods for peat removal and initial washing were investigated

moulding materials were tested, to find which would set in wet conditions, would not leave stains or residues on the object, would take fine enough an impression to enable a cast, and would come off the object after setting without exerting damage to the object

For the lifting and cleaning process, spoons, sponges, soft brushes, Melinex and mini trowels were selected to remove the bulk of the peat out of the box (Figure 6). Because the object was uncovered for quite a long time, it needed to be wetted very regularly in order to avoid drying out. Once the peat was removed and the sickle was cleaned, it was lifted and stored in a box which was padded with wet foam and filled with water (Figure 7).

Figure 6: The sickle in its peat during the lifting process. (45K)

Figure 7: The cleaned sickle stored in water. (55K)

Xantopren L was chosen to be the best material for the moulding process because it met all pre-set requirements (Larsen 1984). The process of moulding was carried out as follows:

all holes and undercuts were filled with wet acid free tissue to prevent the moulding material from creeping in the wood

since the metal was much heavier, it had to be suspended onto fine nylon thread to stop it from disappearing in the moulding material when the first half was made (Figure 8)

Figure 8: Half of the sickle immersed in Xantropen L while suspended on nylon thread on wooden sticks. (43K)

in order to be able to separate both halves afterwards, Vaseline cream was applied onto the set first half before the second half was poured in the basin; when this was set, both halves were separated successfully and a very fine impression was achieved

Following a review of the literature on stabilization methods for wet composites, (Cook, et al. 1984; Hawley 1987; Starling 1987; Selwyn, et al. 1993), the following methods were considered:

diabetic sugar treatments

non-corrosive alternatives to poly ethylene glycol (PEG) referring to poly amine glycols

acetone-rosin method

modified PEG solutions with extra protection for the metal

alcohol-ether method

After much debate, the alcohol-ether method was selected. In this method, the wood is de-watered through various baths of alcohol and then through ether (ethoxyethane). In the last stage, a consolidant can be required as additive. The method has well known and good long-term effects, it is affordable and does not take very long. The chief disadvantage is that health and safety can be a serious point of discussion, because the flash point of ether is below room temperature (-32° C). The use of ether in a non-spark proof and non-isolated laboratory or fume cupboard is thus not permitted. Suitable facilities were made available at the Ancient Monument Laboratory, English Heritage.
Treatment was carried out as follows: a solution of water and industrial methylated spirits (IMS) was made up to gradually replace all the water in the object. The percentage of the IMS went up over a period of time until 100% IMS was reached. This solution was then replaced by ether, in which the object was immersed. After 24 hours, it was taken out of the ether and allowed to air dry in a closed fume cupboard. Initially all went well, but after one hour cross-grained cracks started to appear. Fortunately, these all closed up when the object was placed in a normal environment again. A few smaller cracks, however, remained visible at the curved end of the handle, but they were very small.
During further investigation of the metal under high magnification (65x), striations on one side of the blade were visible and were possible signs of use wear. Some corrosion deposits were removed mechanically. Most of the surface was seriously etched and little was left of the "original surface." The etched areas were not touched during the investigative cleaning process.
Since the future museum environment was not known, the blade was stabilised against further corrosion by brushing on a solution of 3% benzotriazole (BTA) in IMS. Brushing was chosen, in order to keep the rivet holes with the wood uncontaminated. Protective and consolidative layers of Incralac, containing some BTA, finished off the metal (Figure 9).

Figure 9: The finished dried sickle. (26K)

A display mount had to be made which included a firm but softly padded support for both the handle and the metal. Because the metal part was much heavier, the wooden handle and the wood-metal interface were the most vulnerable areas and had to be supported best. Furthermore, a mini case with removable base was thought to be a good option. It would stop people from handling the object but would remain completely visible from all sides.
The object needed proper packing for transport to the museum and storage. This packing consisted of a polyethylene box in which the sickle was held into cut sheets of Plastazote (expanded polyethylene foam) to prevent it from moving and to absorb possible vibration. These sheets were wrapped in acid free tissue.
Health and safety
Throughout all treatment steps and handling, gloves were worn. Benzotriazole, which has been mentioned to be a possible carcinogen, especially in powder form, was made into a solution in the fume cupboard while a dust mask and gloves were worn. The same precautions were applied during the application of this chemical by brush and during the application of Incralac.
The most important issue in terms of health and safety, however, were the precautions associated with the use of ether, to avoid either inhalation or fire. This was achieved by using spark proof fume cupboards in a lockable area with restricted access, and by using as little ether as possible.
Considerations for the future
The following recommendations were made for safeguarding the object in the future:

light levels should not exceed 50 lux to prevent damage to the wood

temperatures of 18-20° C are ideal and should be kept stable at all times

the relative humidity (RH) should be kept stable. This is very important because of the composite nature of the object. Copper alloys and wood require different levels of RH when they are separate. Since this is not possible here, a compromise RH of between 25% and 46% should be set (Erhardt & Mecklenberg 1994). Below 25%, the wood would suffer from being stored too dry and would crack (a potential danger for the sickle because cracks appeared already earlier during the treatment but fortunately closed again). Above 46% copper alloy is prone to active corrosion.

Conclusion
The first aim of the project was to adopt a conservation process that would render the object stable and presentable. A second aim was to learn about the conservation, investigation and analysis of composite and waterlogged objects. A last one was to experience how to work and function in the framework of an ongoing project, involving other conservators, find specialists and experts on different materials. All aims were met and from that perspective, the conservation project was successful.
Acknowledgements
I would like to thank the Conservation staff of the Institute of Archaeology, especially Adrian Tribe for supervision, advice, and help, and Stuart Laidlaw for technical help with the photography. I owe thanks to various specialists who were so kind to give me advice and share their knowledge with me. Special thanks go to: J. Watson (English Heritage), A. Ray (British Museum), D. Sully (Museum of London), and also M. Taylor for taking the wood sample and doing a wood species identification. Special thanks go to the conservation staff at the Ancient Monument Laboratory who allowed me to carry out the alcohol-ether treatment with their facilities, and to C. Slack who carried out the XRD-analysis on the corrosion products of the sickle blade.
Materials
Xantopren L (silicone rubber) - dental moulding material, Cotterill and Turner, UK
Incralac (74% Paraloid B44, 20% toluene, 5% ethanol, 0.5% benzotriazole, 0.5% epoxidised soya oil)
Acetone and Ether - Merck, UK

References
Cook, C., Dietrich, A., Grattan, D.W. and Adair, N. 1984. "Experiments with aqueous treatments for waterlogged wood-metal objects." In: Waterlogged Wood. Study and Conservation. Proceedings of the Second ICOM Waterlogged Wood Working Group Conference, Grenoble. 28-31 August, 1984. ICOM, 147-159.
Erhardt, D. and Mecklenburg, M. 1994. "Relative humidity re-examined." In: Roy, A. and Smith, P. (eds.), Preventive Conservation. Practice, Theory and Research. IIC. Preprints of the Contributions to the Ottawa Congress, 12-16 September 1994, London. London: IIC, 32-38.
Hawley, J.K. 1987. "A synopsis of current treatments for waterlogged wood and metal composite objects." In MacLeod, I. (ed.), Conservation of Wet Wood and Metal. Proceedings of the ICOM Conservation Working Groups on Wet Organic Archaeological Materials and Metals. Fremantle: ICOM, 223-243.
Larsen, E.B. 1994. Moulding and Casting of Museum Objects. Using Silicone Rubber and Epoxy Resin. Copenhagen: Konservatorskolen Det Kongelige Danske Kunstakademi.
Selwyn, L.S., Rennie-Bisaillion, D.A., and Binnie, N.E. 1993. "Metal corrosion rates in aqueous treatments for waterlogged wood-metal composites." Studies in Conservation 38, 180-197.
Starling, K. 1987. "The conservation of wet metal/organic composite archaeological artefacts at the Museum of London." In: MacLeod, I. (ed.), Conservation of Wet Wood and Metal. Proceedings of the ICOM Conservation Working Groups on Wet Organic Archaeological Materials and Metals. Fremantle: ICOM, 215-219.
Woodcock, A. 1995. "A Late Bronze Age waterlogged site at Shinewater Park near Eastbourne in East Sussex, England." NewsWARP 18, 7-9.

The methods, techniques, and conclusions found in individual papers are the work and responsibility of the author of the paper, and should in no way be thought to represent the opinion or endorsement of either the Journal of Conservation & Museum Studies, the Institute of Archaeology, or University College London. No liability or contract is accepted or implied by the publication of these data.