Dendrochronology is the technique by which we can identify the age of a piece of wood by studying the growth rings it contains.
In this article we shall examine how it works, how we know it works, and the limitations of the technique.
How dendrochronology works
It is a well-known fact that many tree genera will produce one new growth ring each year. This means that, as every schoolchild knows, you can find out how old such a tree is by chopping it down and counting the rings.
This in itself would not be particularly useful. However, it is also the case that the rings produced are of different thicknesses according to the weather in each particular year, with a good year corresponding to a thicker growth ring. This is of some interest to paleoclimatologists, but what is important from the point of view of absolute dating is that this produces a sequence of growth rings of different thicknesses which is almost as distinctive as a fingerprint. Imagine for the sake of simplicity that there are only two thicknesses of rings: large ones and small ones. Then over a mere twenty years, over a million different sequences of large and small rings could potentially form, and which one actually does will depend on the weather over those twenty years.
This still leaves one question. If we were to dig up a piece of wood with the distinctive dendrochronological "fingerprint" of (let us say) 10,000 years ago, how would we know that that was in fact the time that it was the fingerprint of? To recognize the fingerprints of a criminal at a crime-scene, we need to have his fingerprints in our files. In the same way, to recognize a 10,000 year dendrochronological "fingerprint" we would have to know what a 10,000 year fingerprint looks like. To know that, we would have to find a piece of wood which we knew to be 10,000 years old to take the fingerprint of. It seems, then, that we can't do dendrochronology unless we already have a way to determine the age of a piece of wood. If this was the case, it would be one of the more useless scientific techniques.
However, there is a way out. Suppose we take a core sample from a tree which grew between 1500 AD and the present; that gives us fingerprints for the past 500 years or so. Now suppose we find dead wood which, unknown to us, represents growth from 1100 AD to 1600 AD. This will have a fingerprint, and the last 100 years of its fingerprint will match the first 100 years of the tree we sampled. Observing the identities between these fingerprints, we can now put a date on each of the rings of the dead wood, which allows us to extend our knowledge of what the fingerprints look like back to 1100 AD, four hundred years before the living tree took us. Now if we find another dead sample which runs from 800 AD to 1250 AD, its tree-rings have a 150-year overlap with the known sequence, we can use this to date it, and then we can extend the sequence still further. By continuing this process with older and older samples of wood, we can build up data stretching back tens of thousands of years. This technique is known as crossdating; similar principles can be employed in other absolute dating methods.
Limitations of the technique
From the point of view of a geologist, tens of thousands of years is not very much. It is useful to an archeologist, but to a geologist that's just the recent past. And it seems very unlikely that the technique will ever take us much further.
The problem is that wood is not readily preserved; for it to last a long time, it must have been preserved under fairly unusual conditions; perhaps in an anoxic peat swamp, or buried under volcanic tuff. What's more, not all kinds of wood are suitable for the task. Some trees are complacent: that is, they produce growth rings of about the same thickness whatever the weather is like; whereas other kinds of trees don't produce exactly one growth ring per year, which also makes them unsuitable for denrochronology.
So while dendrochronology may be an excellent technique so far as it goes, its scope is limited by the ability of archeologists to locate the right pieces of old wood; and these are scarce and become progressively scarcer as we go back through the geological record.
Dendrochronology: how do we know?
We can check that trees, or at least the kinds of trees we use for dendrochronology, do in fact add one ring per year. We can also check that different trees do produce the same pattern of thick and thin rings. Such observations tell us that dendrochronology should work in principle.
And in practice, when we cut down a tree with a known date of planting and count its rings, we can verify that they do in fact give its age.
We can also look, for example, at the timbers in an old building of known date. If dendrochronology works, then we would predict that the dates it gives for the timbers should not be later than the date of construction.
Or we can look at the charred timbers from cities destroyed by a volcano with a known date of eruption. For example, we can look at Herculaneum, which was destroyed by the eruption of Pompeii in 79 AD, and we would predict that dendrochronological dates for the charred timbers would not post-date 79 AD. The success of such predictions confirms the accuracy of dendrochronology.
Finally, we can note that dendrochronology is in close agreement with other techniques described in this textbook; techniques which are based on completely different principles. Even if we can imagine some unusual conditions in the past that might have messed up dendrochronology in some undetectable way, we should also have to suppose that other unusual conditions messed up other dating methods in such a way that they would still concur with dendrochronology. This is an extravagant conjecture: it is more parsimonious to conclude that the reason that all the methods concur is that they all actually work.
We shall have more to say on this subject in later articles.