History of dendrochronology
Dendrochronology came of age in Europe, during the 1980s, with a full oak chronology in Ireland, dating back to 4989BC, created by Baillie, in 1988 (Baillie 1995, 11, 18). This is the main reason for the absence of precision dating in Hewett’s various published works (Hewett 1967, 1969, 1980) and, one of the main reasons why this thesis aims to apply these known dates, to Hewett’s chrono-typologies, that were unavailable to him when he conducted his research. The basis of dendrochronological dating (or “tree ring dating” as it is commonly called) is that trees of the same species, growing during similar time frames, in localised habitats, will produce similar growth-ring patterns (English_Heritage; Miles 2005; Taylor 2005). These patterns, of varying growth-ring widths, are unique to the period of growth, similar to a human finger print (Baillie 1995, 17) and can be matched against a “master chronological sequence”, of known tree-ring dates, with 95% certainty (Millard 2002, 137). Each year, a tree gains another ring as it grows, by adding a layer of cells; the thickness of this ring depends on the amount of growth in that year. These cells grow in the cambium layer, directly under the bark. Thus the older rings are located toward the heart of the tree and the younger rings in the sapwood, near the bark (Tyers 1999, 2). The heartwood is recognisable as being much darker than the sapwood, because it is essentially dead wood and, much harder than the softer sapwood (Grenville 1999, 10; Wilson & White 1986, 13). English oak will turn from sapwood to heartwood in around 15-50 years (Hillam et al. 1987). Sapwood tends to remain at a constant width as the tree grows, while the heartwood continues expanding in size, as the sapwood dies (Wilson & White 1986, 14). New growth takes the form of widely spaced cells, formed in the spring and, closer, smaller cells during the summer (Taylor 2005). During years of ideal growing conditions, trees will produce a constant sized ring, whereas in a year with poor conditions, such as too much rain, tighter growth rings will form (Grenville 1999, 9). Trees growing in similar regions are likely to display the same general chronological growth pattern which tends not to reflect any localised ecological variations but rather the climatic variations (Miles 2005; Taylor 2005). Thus, over the life of a tree, various sized rings will create a “fingerprint” unique to that tree, but common to all other trees in that area, subject to the same weather patterns (Miles 2003, 220). When a tree is felled, or dies, the rings no longer grow, the final year of growth is recorded using the outmost ring, directly under the bark (Miles 2003, 220). In the United Kingdom, oak (quercus robur and quercus petraea) provides the best examples for dating, though elm and beech can also be dated (Baillie 1982, 45; Grenville 1999, 9; Miles 2003, 221). This is of particular use to medievalists as the majority of timber framed structures were made from Oak (Miles 2003, 221). Counting the rings will give the age of the tree, but, unless the rings can then be matched against a known chronology, they cannot provide a method of dating the tree (Tyers 1999, 2). In order to obtain a dendro-date, a collection of core samples need to be taken, by skilled technicians, from several timbers within the structure (Miles 2003, 220; Morriss 2000, 142). The core is usually extracted by drilling the timber with a hollow drill bit approximately 10mm in diameter (fig 1). However, the original method would have been to take a slice of wood; impossible from a standing structure (Baillie 1982, 93). This produces a cylindrical core, the length of which depends on the thickness of the timber, which contains a sample of the inner heartwood, out to the sapwood (Morriss 2000,142). When taken back to the dendro-laboratory, the sample core is mounted and sanded smooth so that the rings can clearly be seen, fig 3. The rings are then measured under a microscope, and the data entered into the computer. The computer then compares the new data with existing chronologies from an established “master chronology” dataset, within a specialist statistical software package and, statistical analysis run, to test for significance and correlation coefficient (Baillie 1995, 20-1; Miles 2003, 220; Morriss 2000, 142). “The correlation coefficient is calculated at every position of overlap between the specimen ring pattern and the master pattern” (Baillie 1982, 86-92; 1995, 17). Providing the cores have a good date range and, more importantly, some sapwood rings, a date range can be arrived at and, in some cases, where the sapwood rings are intact, a specific felling date can be given (English_Heritage; Miles et al. 2005; Miles 2003, 220; Morriss 2000, 142). Haddlesey, R 2008. "Dendrochronology". British Medieval Architecture. (online) www.medarch.co.uk/dendrochronology.html