Illustrating with examples, outline the principle of either: Dendrochronology OR C14 Radiocarbon
dating -describe the advantages/limitations of the method.
Dendrochrononolgy utilises the feature in some species of tree floral to lay down a distinctive cell pattern, know as a tree ring, at the rate of one only per year and in each and every year. It is vital for this method of study to be effective that the ring formation occurs every year without fail and that one ring only is laid down per year. The reasons for this are simple, should a tree miss a year of ring formation or should it add extra rings within a year, the methods of establishing a chronology from the data would be invalid due to unpredictable irregularities.
In Britain it is fortunate that the Pedunculate or English oak (Quercus robur) , exhibits such characteristics. The occurrence of ancient oaks in this country vastly exceeds that of any other western European country, and oaks can lived for well over 500 years, especially if pollarded. This and their utilisation throughout much of British history as building material, makes them an ideal and useful candidate for dendrochronological study. Many buildings and other structures have be subjected to dendrochronological survey, such as Donnington le Heath Manor House, Leicestershire, Lincoln Cathedral, the Sweet Track in Somerset. These studies necessitate several samples being taken to build up a site chronology, in order to give the most accurate hypothosis possible.
The basis of dendrochronology is simple. However the accurate interpretation of the raw data becomes more complex.
The Oak has a growing season within each year, this starts in Spring and ends towards the end of Summer/beginning of Autumn. The actual length and cell productivity of the trees over this period will vary depending on the environmental and climatic factors involved , which in turn can vary from year to year.
It is these variations in tree ring production that allow comparisons to be made from tree to tree and eventually a relevant chronology established.
In the Spring the tree is triggered to cell production by various positive environmental and climatic changes, whereby it starts producing distinctive massive cell build up, which can be seen visually as a lighter coloured band withing the tree structure. As the season progresses the tree settles to a less massive cell production which is exhibited but a visually darker band. This darker band has a tendency to be greater influenced by the climatic conditions present and a greater variation can be seen in its width than in the early growth width, which whilst existing is not so distinctive.
In years where the climate is favourable the cellular growth will be greater, whereas in poor years the cell growth is minimised. In addition it should be noted that each tree has its own individual environ, influenced by factors such as, soil type, proximity to its neighbours, position within the tree group, dieseases or faunal infestation and so on. As a tree ages the width of each ring laid down also tends to decrease. However, the ratio of ring measurement when subjected to statistical enhancement can give and pattern of ratios of growth, comparable to other trees in a meaningful way. It should not be assumed that one can take a group of samples and match them up visually. The raw data variables tend to be too great for simple analysis, although in very small samples of trees from a small area it may be possible.
More important to dendrochronology than the good years are the bad. Bad climatic conditions tend to be shared by all trees both with good environs and bad. It appears unlikely that the same pattern of good and bad years would be repeated exactly over a period of years, and the greater the number of years the less likelihood of exact repetition there is. Therefore it can be assumed that if a pattern over a fifty year or greater period is established, that will hold true for only that one fifty year period, within specified geographical boundaries. However where a single sample is present a pattern showing in excess of fifty years would be preferred, whereas the length of pattern may be reduced where several samples from the same site are available.
As trees grow their outer rings under the bark are the active sap wood, whereas the inner wood or heartwood, is inactive. The sapwood tends to contain 15-40 rings with the bark being outside the ring being formed, or if in winter the last ring formed. When bark is present it is possible to be sure that the last ring was laid down just prior to the tree being felled, however the presence of sapwood only may still be useful in establishing a range of felling dates due to the reasonably consistency of amount of sapwood rings.
To establish a chronology it is necessary to obtain some samples of a known felling date. This may be possible from samples with a historically recorded felling of usage date , or may be from recently felled trees. Tree rings from trees that are felled in say 1997, can be compared to earlier timbers matched by the ring data which often in the earlier unknown sample can take the chronology further back in time from the first sample. By using this overlap technique and by constantly adding new data from timbers the chronology can be further refined both in date information and geographical information. Whereas initially a chronology was only possible for the whole of Britain, due to the scarcity of data, more localised chronologies such as the East Midlands chronologies are now possible giving a far more accurate basis for each local geographical area.
Samples for analysis can be taken in several ways. Firstly either from live tree or from timber within a structure or from an unknown origin such as wood or charcoal from an excavation. The physical sample may either be a complete slice, 5-150mm thick, or more often from live wood or standing buildings, a core sample. Core samples are extracted from timber using a hollow tube cored attached to an electric drill. The diameter of the corer is typically 13-15mm and may vary in length. Position and location of each sample should be recorded meticulously. Preferably 8 to 10 samples per building, from different timbers should be taken. More if the building is complex. This is to give a greater accuracy of sampling when often some will be rejected as unsuitable or undatable. From this a site master chronology should be able to be produced, this site master may not be dateable until compared with the Master chronology but will give relative dating of timbers in relation to each other. The site chronology should then be able to be match into the Master Chronologies (as available) and then included, thus refining the Master further.
From the samples accurate measurements are taken of the distances from the start of one tree ring to the next. This is done under a microscope. The data is fed into a computer where it is subjected to the equations following later(or similar). See examples of collected data in Appendix 1.
This statistical equation is performed on the raw data in order that the differences relevant to analysis are enhanced and those which are less so, are decreased. Factors that need minimising are the differences in growth rings due to the age of the tree, where younger trees produce a generally greater growth than older, good year growth , where trees with more positive factors will maximise growth better than those in a poorer environ . Factors to maximise are the growth rings in poorer years. These show as troughs in graphical representation, as opposed to the peaks of good years.
x represents a value for the tree rings, and y is from 2nd tree in the correlation.
x and y therefore represent a % of the mean of all the x & y values for example
r=the correlation coefficient of the x & y component and can be between -1 and +1.
A perfect match being represented by r = +1 and a mismatch by r=0
To adjust the value of r to a probability, it must be converted to t using
this then gives the probability of r arising by chance alone within the sample N.
This analysis can give a t value, which is a representation of correlation between sequences. The higher the t value the greater the correlation between the samples t>3.5 for example being acceptable as good date correlation. It is also understood that some claim (Baille 1990) that if the equation is made statistically more robust the values reduce and t should therefore be named 'maximised t'.
Statistic must also be capable of replication in the hands of other, preferably distance, practitioners.
It is in reality these statistical value that are matched into a chronology rather than the often portrayed, in illustrations as above, rather simplistic tree rings themselves. It is the relative differences in tree ring growth from tree to tree that give us the usable data and not the raw data.
Dendrochronolgy is a useful tool in archaeology in several ways. It can provide relative dating within a undated related sampling, or when overlapping dated samples are available, in an acceptable geographical area it can provide absolute dating. It is also a useful tool to calibrating carbon 14 dating techniques, where dendrochronological & C14 dates can be taken and the C14 variables calculated in relation to the dendrochronological dates, thus being invaluable in refining the practices of C14 dating given the variable inherent in the method.
However, it is not without its negative aspects, as with all complex statistical analysis, one must be aware that there is a possibility of the equation being made to fit the data rather than the other way around, particularly if the mechanism is not well understood by all practitioners. This is where the interpretation depends on the experience of the dendrochronologist , which is an important factor, but one which is bound to vary, as it takes dendrchronolgy away from pure data and into interpretation by individuals with inevitable differences of opinion and experience. The principle has been based on a wide geographical area and narrowed down over time, both in terms of geographical area and complete data, one assumes that in addition in including new data into the Masters, that revision of data based on new sample is also undertaken to establish if the analysis holds true when approached from all points of relevance rather than the original in isolation.
BAILLIE, M.G.L. Tree-ring dating and Archaeology Chicago (1990)
ENGLISH HERITAGE, Dendroochronology (?)
FRITT HAL Tree Rings and Climate (1976)
LAXTON,R.R, and LITTON, C.D. An East Midlands master tree-ring chronology and its use for dating vernacular buildings, University of Nottingham, Dept of Classical and Archaeological Studies, Monograph series, III (1988)
MABEY, Richard Flora Britannica Sinclair-Stevenson (1996)
http://www.sonic.net/bristlecone/dendro.html Diagram, with permission.
ITRDB DENDROCHRONOLOGY FORUM