Analysis of the Coffin Wood by I. Tyers


Samples of timbers from the Upper Chapel supplied for dendrochronological analysis and wood identification comprised six sides of a coffin [114], five suitable for dendrochronological analysis and one suitable as a wood identification sample. The wood identification sample was successfully analysed. Tree-ring analysis identified the five boards from the Upper Chapel coffin were derived from a single, currently undatable, tree.


The dendrochronological samples were all oak (Quercus) and supplied as cross-sections. The wood identification sample was a non-oak fragment. For the wood identification analysis, microscopic cross-sections were taken from the sample in three planes (tangential, radial, and transverse); these were mounted on glass slides with cover slips. The features were then examined at up to 400x magnification and compared with illustrations and keys in Schweingruber (1978).

Each tree-ring sample was assessed for the number of rings it contained, and whether the sequence of ring widths could be reliably resolved. For dendrochronological analysis samples need to contain 50 or more annual rings, and the sequence needs to be free of aberrant anatomical features such as those caused by physical damage to the tree while it was still alive. Standard dendrochronological analysis methods (see English Heritage 1998) were then applied to each suitable sample. The sequence of ring widths in each sample were revealed by preparing a surface equivalent to the original horizontal plane of the parent tree, using a variety of bladed tools. The width of each successive annual growth ring was revealed by this preparation method. The complete sequence of the annual growth rings in the suitable samples were then measured to an accuracy of 0.01mm using a micro-computer based travelling stage. The sequence of ring widths were then plotted onto semi-log graph paper to enable visual comparisons to be made between sequences. In addition, cross-correlation algorithms (e.g. Baillie and Pilcher 1973) were employed to search for positions where the ring sequences were highly correlated (Tyers 2004). Highly correlated positions were checked using the graphs and, if any of these were satisfactory, new composite sequences were constructed from the synchronised sequences. Any t-values reported in Table 20 were derived from the original CROS algorithm (Baillie and Pilcher 1973). A t-value of 3.5 or over is usually indicative of a good match, although this is with the proviso that high t-values at the same relative or absolute position need to have been obtained from a range of independent sequences, and that these positions were supported by satisfactory visual matching.


The six coffin boards provided for analysis comprised an ash (Fraxinus) base board, and five oaks. Details of the oak boards are recorded in Table 19. The use of non-oak boards for the bases of coffins appears to be relatively routine in early modern Sheffield, further examples having been seen at the Carver St Methodist Chapel excavations. Presumably this was cheaper, and the bases could not be seen. The five rather fine oak boards are identified by tree-ring analysis as being obtained from a single tree, as summarised in Table 20. There is a very high probability that the head and foot board represent the inner and outer parts of the same original board, as shown in Figure 32 below. The composite series from these boards has been compared with English, European and some other oak reference data without successfully identifying a date for the material. It remains possible, although relatively unlikely, that the timber was derived from an aberrant local tree. It seems more likely that these boards were imported from somewhere for which currently there is no contemporaneous reference data.

Figure 32

Figure 32: Bar diagram showing the relative dating positions of the 5 sequences from coffin [114]


© Internet Archaeology/Author(s)
University of York legal statements | Terms and Conditions | File last updated: Wed Jan 19 2011