a new site, the archaeologist immediately requires information about its
age in order to set it in context with other sites. In research into our
heritage the conservationist or architect may be able to date the general
period of a building he is working with from either the situation, materials
of construction, type of timber joints or other stylistic features. Almost
certainly the century or portion of a century when it was built may be
assigned with some certainty. However, as more and more work is done and
increasing numbers of structures with complex constructional phases are
encountered, the general features may not be sufficient to give the accuracy
in dating that is currently required. If research into other sources of
information also fails to throw light on the building's history, resort
may be made to the various scientific methods of dating. This article
outlines three of the most important methods currently used for dating
buildings or, in a complex situation, the order of construction within
the building. These are: dendrochronology (or 'tree-ring' dating), radiocarbon
dating and thermoluminescence dating.
Each method has a distinct role in the investigation of historic buildings. None is infallible and before embarking on an extensive dating survey, due thought must be given to what might be achieved and which methods might be the more successful. If necessary, seek advice.
Whilst earlier types of wooden joints may be copied in later buildings and earlier styles may be reintroduced in later periods to confound the conservationist or historian, any reuse of older materials should become obvious by the use of the chronometrical methods described here. The incorporation of ancient bog oak into a building, no matter how intricately carved or jointed, would immediately become obvious to the chronologist, as would timber renovations.
Dendrochronology is the oldest method, having been introduced over a century
ago by an American astronomer, Professor A E Douglass. He wanted to know
whether the number of sunspots affected weather on Earth. If this were
so, the width of the annual growth rings would show changes in synchronism
with the sunspot numbers. He established a laboratory in the university
of Arizona, at Tucson, to study tree-rings. Unfortunately, after many
years of analysis he was not able to confirm the correlation he sought.
Nevertheless, the laboratory was able to demonstrate many interesting
properties of ring widths and their relationship with various aspects
of climate and other natural phenomena and, of course, their use in the
accurate dating of timber. His laboratory is still one of the leading
centres in world dendrochronology. It was not until 1939 that the science
was taken seriously in Europe, mainly through the efforts of Professor
Huber in Germany, and not until after World War II that such studies became
established in the UK. The main centres in Britain researching this field
are located at universities in Belfast, Cambridge, East Anglia, London,
Nottingham and Sheffield with several freelance practitioners. Whilst
most dendrochronological research is still concerned with climatic change,
where the precise dating provided by the growth rings is of vital importance,
all units in this country are proficient in performing dendrochronological
surveys of buildings.
Oak is the species of prime interest and it is possible to date wood back to over 7,000 years with a precision, in appropriate circumstances, of a single year. This is most impressive and makes dendrochronology the main dating method for structures containing oak timber. The method relies upon the response of trees to the weather during the growing season, which runs from March to October. In a 'good' growing season the trees within a large climatically homogeneous region all respond by putting on a wide growth ring within the cambium which separates the sapwood from the bark. In a 'poor' growing season the trees all respond so that only a very narrow growth ring is formed. In more typical growing seasons a ring of intermediate width is produced. (It should be noted that there is no direct linear relationship between ring width and, say, sunshine, or other weather components) Thus a 'good' or 'poor' growing season is defined with reference to the amount of growth produced. For example, the year 1976 had a gloriously hot, long summer with most rainfall arriving in autumn but the trees did not appreciate it and all oaks produced a distinctively narrow ring. Again the summer of 1915 was cold and wet, quite different from 1976, yet the trees also produced a distinctly narrow growth ring. So it will be seen that seasons that are hot and dry as well as those that are cold and wet will produce a narrow ring so that such a ring is not diagnostic of the weather. Year by year the trees throughout the region produce a similar pattern of wide and narrow rings in response to the weather changes. It is this pattern that allows the accurate dating. The pattern of ring widths on a specimen taken from a building is matched, using a computer with a 'master chronology' often several centuries long for the particular area. This regional chronology will have been painstakingly built up from many thousands of measurements and by cross-matching many overlapping patterns of timbers. The youngest patterns are obtained from living trees, where the felling date of the final ring is known. Progressively older patterns are obtained from trees in recent buildings, older buildings, archaeological sites and ancient bog oaks. Because of local, non-climatic causes of change of growth width, the chronologies around the country vary somewhat, and the best dating match is always obtained from a local regional master chronology.
The dendro-date is thus the year in which the final ring of the specimen grew (the year in which the tree was felled, but not necessarily the year in which the building was constructed). In order to obtain an accurate match and hence a date, it is important to have at least 80 rings on the specimen that is to he dated. With fewer rings the pattern might have repeat matches at different points in the time scale and so give rise to multiple possible dates. This has implications for some vernacular structures in which rapidly grown, wide-ringed oaks, 30 to 40 years old, were used. In such instances it might be possible to date the wall plate which often contains far more rings. In practice it is found that 100 or 120 growth rings are most likely to provide a unique match. However, because of the local ecological, non-climatic effects on the tree ring, it is not possible to guarantee that any particular specimen will give a date. In order to have greater certainty it is important to obtain several samples, in the form of cores drilled from the timber, and to construct a 'site chronology' for the building. The number of cores required will depend upon the complexity of the structure, but some ten cores per building phase is preferred. These are normally taken by the dendrochronologist in co-operation with the historian and the position of the cores is carefully marked on the building plan for future analysis of the results. The core leaves a small hole in the timber of about 15mm in diameter which may be plugged with a timber dowel.
Although this method is capable of dating to the individual year, in practice several factors conspire to reduce the precision in dating the construction, sometimes drastically, and it is important to be aware of the limitations. Whilst in the middle ages it was the practice to use the timber 'green' - usually within a year of the felling date - in more recent times the timber is usually allowed to dry out, sometimes for decades, before use. Furthermore, carpenters, aware of the effects of insect attacks, would deliberately remove the sapwood and even some heartwood. The number of sapwood rings may vary between 15 and 50 years, depending on the position in and the age of the tree. Thus the year of the last ring dated could be misleading to the construction date and be underestimated by an unknown number, possibly 60 years. Sapwood may be found on at least some of the timbers in the dendrochronological survey and the site master chronology will lead to a more reliable date than an individual core.
Whereas tree-ring dating is limited in this country to oak structures,
radiocarbon dating may be used for any wood species and, indeed, for any
other organic based materials found in buildings such as: wattle and daub;
straw used for insulation; hair used in plaster; leather wall hangings;
and, perhaps surprisingly, mortar. The range of radiocarbon dating reaches
back to 60,000 years. For the last few thousand years it can have a precision
of a few decades and may, in certain circumstances, be comparable with
tree-ring dates. The method was conceived by American Professor Willard
F Libby of Chicago in 1947. The laboratory at Cambridge here in England
was among the first six to be set up anywhere in the world. There are
now several radiocarbon dating laboratories in Britain including those
at Belfast, Cambridge, East Kilbride, Oxford and Swansea, as well as a
commercial unit near Harwell.
Radiocarbon dating is based on the element carbon, the basis of all life on earth. The atoms of this element are of three different types or 'isotopes'. They are identical chemically but have slightly different physical properties, particularly in mass. The isotopes are respectively 12, 13 and 14 times as heavy as the common hydrogen atom (the base unit by which the weight of other elements is measured). The isotopes C-12 and C-13 are stable and make up the bulk of the element, but the C-14 isotope, which is mildly radioactive, is extremely rare. The instability of radiocarbon results in half of it disappearing in 5,730 years (its 'half-life'). This instability is the basis of the dating method. All creatures have the same concentration of radiocarbon in their cells while they remain alive. This level is maintained constant by a sequence of events affecting the food web. It starts with photosynthesis in green leaves of plants, whereby atmospheric water vapour and carbon dioxide, containing the radiocarbon, are combined in the presence of sunlight to produce sugar. The plant biological process converts this to the myriad of substances required for life. These substances are shared via the food network to all animals including man. For our purposes it may be assumed that the amount of radiocarbon in the atmosphere is constant over time.
Once the creature dies the food chain is broken and the concentration of radiocarbon in the cells falls away. By measuring the residual C-14 concentration in the material the date of its death may be calculated. In the case of tree-rings the food chain is effectively broken at the end of the growing season and the radiocarbon concentration immediately begins to fall. Thus, in principle, the age of each growth ring may be measured. In practice, the measurements may resolve differences of about 20 or 30 years.
Samples from a building for radiocarbon dating should be taken with care and due regard to provenance. For timber specimens, samples should be obtained as near to the bark as possible, as for dendrochronology. Samples such as leather, cloth, food residues or straw represent a year's growth and so a point in time. Thatch, whether straw or rush, will date the last repair and not necessarily the construction date. Mortar is made by heating limestone to over 850°C to form quicklime. When slaked and used as mortar between layers of bricks it dries by absorption of contemporary carbon dioxide from the air and so may be used to date this event.
Thermoluminescence, or TL, was first used in the 1950s for the measurement
of radiation exposure, and underwent a period of difficulties before being
applied to dating; the first dates it produced being too young. Upon resolution
of the technical problems the method was used for dating pottery and burnt
flints from archaeological sites with a precision of about 7-10 per cent.
Subsequently it has been used in the investigation of recent geological
formations reaching back to half a million years. In its most common form
it may shed light on the age of fired clay and quartz based materials
but approaching the present no closer than about a thousand years. A modern
variant on the technique is able to date far more recent fired clay material.
TL depends upon minute levels of background radiation in the clay matrix, a tiny fraction of which is absorbed and stored as a charge at imperfections in the crystal lattice of quartz inclusions. The firing of pottery removes the inherited geological TL and sets the dating clock to zero. In the laboratory grains of quartz are extracted from the pottery and heated in light-tight apparatus at a constant rate to around 400°C . Superimposed upon the red-hot glow, a tiny flash of light is produced as the stored energy is released (hence 'thermoluminescence') and the flash is recorded by computer. The quantity of light produced is proportional to the length of time since it was last fired. Unfortunately, problems remain since all samples do not have the same sensitivity to the radiation and background radiation levels vary. Furthermore, the results are sensitive to water content. Thus many measurements must be made in order to obtain a date.
Recently this method has been improved. The flash of light is released by scanning the sample with an energetic green laser beam and light-emitting diodes are used as detectors. This form of the method, known as 'optically stimulated luminescence dating', enables objects which are not more than a few hundred years old to be dated to within a few decades. Hence it is far more useful than the original TL technique in dating buildings. The requirement remains that the sample should have undergone some heating event to set the clock to zero. It also requires that a dosimeter be left undisturbed in situ at the site for some months in order to discover the natural radioactivity permeating the samples. These must be inorganic and contain some light transmitting materials. Pottery artefacts and certain bricks might be suitable specimens, and often TL provides the only way to distinguish medieval or Tudor bricks and chimney pots from Victorian reproductions.
There are several laboratories capable of this sort of measurement in this country which include the Geology Department, Aberystwyth; the British Museum; the Godwin Institute, Cambridge; the Department of Archaeology, Durham; Environmental Sciences; the Institute of Archaeology, University College, London; and the Research Laboratory for Archaeology and the History of Art, Oxford.
There are thus now various ways in which chronologists are able to help conservators, historians and architects in their endeavours in dating our heritage.