Many artifacts and skeletal remains can be dated by reference to an established typological framework. The chronology of such a framework, except within a well-documented historical period, is ultimately based on a chronometric technique such as radiocarbon, thermoluminescence, or dendrochronology. These and other techniques are also used for placement within the framework. The essential characteristic of nearly all these is that they are “absolute,” in the sense of giving a date that is based on currently measurable quantities; hence, they are objective, being independent of existing chronologies. The techniques outlined below are: (1) radiocarbon, (2) dendrochronology, (3) luminescence, (4) electron spin resonance, (5) uranium-series, (6) potassium-argon, (7) fission tracks, (8) amino acid racemization, (9) obsidian dating, (10) other chemical techniques, (11) archaeomagnetism, and (12) pollen analysis, climatic change, oxygen-isotope stages, and the Milankovitch timescale. Error limits are quoted at the 68 percent level of confidence (i.e., 1 sigma).

Radiocarbon Dating.

The carbon in the carbon dioxide in the atmosphere consists mainly of nonradioactive carbon-12 and carbon-13; there is also a minute amount of weakly radioactive carbon-14 resulting from the interaction of cosmic-ray neutrons with the nitrogen in the upper atmosphere. Carbon dioxide mixes rapidly throughout the atmosphere, and by photosynthesis enters plant life and hence into animals; atmospheric carbon dioxide also enters the oceans as dissolved carbonate, so that this too contains carbon-14—as do any shells and deposits formed from it. The totality of atmosphere, biosphere, and oceans is known as the carbon-exchange reservoir. The concentration ratio between carbon-14 atoms and nonradioactive carbon atoms is approximately the same (about one in a million million) throughout the reservoir; it also stays approximately constant with time.

In organic matter that is no longer exchanging its carbon with the reservoir, the carbon-14 lost by radioactive decay is not replenished; this is the case, for instance, in the cellulose molecules of wood. From the time of formation, the concentration ratio decreases at a rate determined by the immutable 5,730-year half-life of carbon-14; this means that wood 5,730 years old will have a carbon-14 concentration that is one half of its value at formation; for wood that is 11,460 years old, the concentration will be one quarter of the formation value, and so on—or, the concentration decreases by 1 percent every eighty-three years. Hence, comparing the concentration measured in an ancient organic sample with the assumed value at formation can determine the amount of time that has elapsed.

In practice, the method is useful for the period four hundred to four thousand years ago—although, with special measurement installations, some sample types can reach substantially further back, to about seventy thousand years ago. The error limits on the calendar-year date (i.e., after calibration; see below) are typically in the range ± 50–150 years, being smaller in some millennia than others.

Measurement.

The carbon-14 concentration in a sample can be determined by converting the carbon of the sample into a gas or into benzene and measuring the emission rate of beta particles in an appropriate detector system. Since the late 1970s, direct measurement of the concentration by means of accelerator mass spectrometry (AMS) has been increasingly employed. Although such an installation is very expensive, it has the important advantage that the size of the sample required is less by several orders of magnitude—a few thousandths of a gram of carbon instead of a few grams.

Sample types.

An essential characteristic of a sample is that over the centuries of burial it should not have acquired any fresh carbon from the atmosphere (e.g., by fungal growth). A minute amount of modern carbon can cause the date that is determined to be substantially too recent, particularly for samples near the forty-thousand-year limit. Conversely, the incorporation of “dead” carbon at formation can cause the opposite effect. Although the sample material itself may have high integrity, intrusive contamination may have been acquired during burial; the humic acids carried in percolating groundwater are an example. It follows that the extent to which a sample is reliable is bound up with the stringency of the laboratory pretreatment that can be applied. The severity that a laboratory can afford to use is dependent both on the size of the initial sample and on the amount of carbon required by the measurement facility; the severity needed depends also on the age.

For wood the use of extracted cellulose avoids lignin and humic acids. For precise dating, a serious problem with wood and charcoal is estimation of the extent to which the wood's formation predated the archaeological event of interest. For bone, use of the extracted protein fraction (collagen, gelatin) is necessary. Unfortunately, the amount remaining decreases with age and in some burial environments, among them the Near East, too little is left in bones from periods earlier than, say, the Neolithic. Charred bone is a good alternative. Grain can be dated reliably, particularly if charred. Single grains can be dated with the AMS technique, but site association is a severe problem. Among other datable sample types are shell, peat, sediment and soils, ivory paper and textiles, straw in mud bricks, and traces of charcoal in iron objects. The golden rule in sample collection is prior discussion with the relevant laboratory.

Calibration.

A raw age from a laboratory is given in conventional radiocarbon years BP (“before the present,” defined as 1950). This is not the same as the age in calendar years (i.e., sidereal years) because the carbon-14 concentration in the biosphere varies slightly from decade to decade and century to century: thus, wood formed at some time in the past does not necessarily have the same carbon-14 concentration as wood formed today. The variation occurring over the last ten thousand years has now been established by measuring samples of wood for which the age has been determined by dendrochronology. By using this “calibration curve,” ages in radiocarbon years can be converted into ages in calendar years or calendar dates—denoted by cal BP, cal BCE, or cal CE. Extension of the calibration curve further back in time is being undertaken by intercomparison with the uranium-series technique using samples of coral.

The terminology just given, BP and so forth, is specific to radiocarbon dating. Most other techniques yield ages directly in calendar years and the use of BP is then confusing (because, it implies radiocarbon years rather than calendar years); exceptions to this are amino-acid dates and obsidian dates calibrated by reference to radiocarbon.

Dendrochronology (Tree-Ring Dating).

For some species of trees growing in some climatic conditions, visually recognizable rings form annually with a climatically determined width pattern. This width pattern allows cross dating between trees felled in different periods (the pattern shown by the inner part of a recent tree can be matched with that shown by the outer part of a tree that was felled earlier). In this way a master chronology can be established for a region and building timber of unknown age can be dated by discovering where its pattern fits onto the master chronology. If sapwood is present on the timber, high accuracy is possible; even so, a degree of uncertainty remains as to how many years elapsed between felling and usage on the site concerned. To obtain a reliable fit onto the master chronology, the piece being dated needs to contain a minimum of about fifty rings.

The Californian bristlecone pine chronology and the European oak chronology have made possible the conversion of radiocarbon years to calendar years. The trees concerned are remarkably long-lived and by leapfrogging back from recent trees to successively older fossil trees, as indicated above, it has been possible to cover the last ten millennia.

Luminescence Dating (LD).

Thermoluminescence (TL) and optical dating (OD) comprise luminescence dating (LD); OD utilizes optically stimulated luminescence (OSL). These are based on the cumulative effect of nuclear radiation on the crystal structure of certain minerals, mainly quartz and feldspar. The nuclear radiation is provided by trace amounts of potassium-40, rubidium-87, thorium, and uranium that are naturally present in the sample and its surroundings. The dating signal is obtained, in the case of TL, by heating grains extracted from the sample and measuring the light they emit by means of a highly sensitive photo-multiplier; in the case of OSL, the measurement is obtained by shining light on the sample; lasers and halogen lamps can be used, as well as infrared-emitting diodes—in which case the luminescence may be termed IRSL.

The TL technique was first applied to pottery, the event dated being the firing by the ancient potter (thereby setting to zero the previously accummulated TL). It was subsequently extended to stalagmitic calcite (which has zero TL at formation) and to burnt flint (heated accidentally by having fallen into fire, or deliberately). The latter application, along with the ESR dating of tooth enamel, has had a particular impact on hypotheses concerning the relationship between anatomically modern humans and Neanderthals. The TL and OSL signals from quartz and feldspar can also be set to zero by exposure to daylight. Thus, windblown sediment (such as loess) is set to zero during its transportation, as is waterborne sediment, though in this case the zeroing process is less effective and the OSL method has a strong advantage over TL.

For either method it is necessary to evaluate the dose rate of nuclear radiation that the sample was receiving during burial. This is done partly by radioactive analysis in the laboratory and partly by on-site measurement (of the penetrating gamma radiation that reaches the sample from surrounding soil and rock, up to a distance of about 0.3 m). It is also necessary to estimate the water content of the sample and soil/rock because any water present attenuates the nuclear radiation. Uncertainty about water content (averaged over the burial period) can limit the accuracy obtained; sites which have always been bone-dry have an advantage.

For pottery and burnt flint, a half-dozen fragments not less than 30 mm across and 10 mm thick are required. For sediment, about half a kilogram is typically needed. Great care must be taken to avoid exposure to daylight, particularly if the OSL method is being used. Hence, the sample collection of sediment should be done by laboratory specialists with appropriate equipment (e.g., a short steel tube which is driven into the section; the material at the exposed ends is discarded in the laboratory); otherwise, the samples should be collected at night and secured in thick, opaque plastic bags. Even slight exposure to light may cause the evaluated age to be too recent.

Age range is highly dependent on sample type and site; several hundred thousand years is a typical upper limit for flint, calcite, and sediment—although with the latter there is a tendency (if some types of feldspar are present in the sample) for the age to be underestimated, particularly beyond fifty thousand years. The lower limit is on the order of one thousand years. With pottery, the lower limit can be on the order of a decade in some cases; the upper limit is usually beyond ten thousand years. The accuracy attainable is usually in the range 5–10 percent of the age (i.e., ±150–300 years at 1000 BCE, ±300–600 years at 4000 BCE). Hence, the main strength of the technique is for sites where there are no samples suitable for radiocarbon and on Paleolithic sites beyond the range of radiocarbon.

Electron Spin Resonance (ESR).

Another way of measuring the cumulative effect of nuclear radiation is with electron spin resonance (ESR). Its principal archaeological applications have been to tooth enamel and stalagmite calcite. ESR reaches back several million years; its lower limit is a few thousand years. The above remarks about dose-rate evaluation apply here, but with additional complexity in the case of tooth enamel because of uranium migration. As a result, two ages are quoted, based on the assumptions of early uptake (EU) and linear uptake (LU), respectively (in some cases the EU age is substantially lower than the LU age). Large teeth (elephant, mammoth) are preferred; usually, separate determinations on about a dozen samples are averaged.

Uranium-Series Dating.

Stalagmite calcite and other forms of that mineral and also some products from volcanic eruptions lend themselves to uranium-series dating. At formation, only uranium is present in the sample material. With time, radioactive daughter products (thorium-230, protactinium-231) build up slowly; determination of the ratio of daughter product to parent uranium allows the age to be evaluated. Measurement is conventionally by alpha spectrometry, but recently thermal ionization mass spectrometry (TIMS) has been used, which gives much higher precision and needs less of a sample. With TIMS, the age range is approximately 500–500,000 years; with alpha spectrometry, 2,000–350,000 years. In good circumstances, the error limits using TIMS can be as low as ± 1 percent of the age; with alpha spectrometry the limits are usually in the range of ±5 percent to ±10 percent.

Potassium-Argon Dating.

Of crucial importance in the study of early hominids, potassium-argon dating is a technique based on the accummulation, in volcanic lava, for example, of argon-40 produced from the slow decay of potassium-40. During a volcanic eruption, previously accumulated argon (a gas) is released, which is the event which is dated. Obviously, to be useful, human occupation must be reliably associated with volcanic products—for example, between two lava beds. Its age range is enormous: from ten thousand years to about four hundred million years.

Fission Track Dating.

The main impact of fission track dating has been on the study of early hominids. Its one or two applications to Neolithic obsidian artifacts are to be regarded as exotic rather than useful. The technique is based on the counting of the miniscule tracks left in some minerals by nuclear fragments immediately after the occurrence of spontaneous fission undergone by uranium nuclei present as impurity.

Amino-Acid Racemization.

Dating by amino-acid racemization is based on the slow conversion within a protein molecule of an amino acid (such as aspartic) from its L form, at formation, to its D form, until an equilibrium mixture of the two is reached. Epimerization of isoleucine is another reaction utilized, notably in the dating of ostrich eggshells. These processes are strongly influenced by environmental conditions—by temperature in particular; site-by-site calibration against radiocarbon is usual, along with extrapolation for samples that are beyond the limit of the latter technique.

In its early application the technique acquired a reputation for unreliability, particularly for poorly preserved bone. However, good reliability is now being obtained when sample types are selected carefully (such as ostrich eggshell, well-preserved bone, and tooth enamel), when strict attention is paid to the validity of the radiocarbon dates used for calibration, and when stringent laboratory procedures are followed. Using aspartic acid the last 50,000–100,000 years can be covered; using the epimerization of isoleucine, reaching a half million years is feasible. In addition to the extension of age range, another advantage over radiocarbon is the comparative ease of measurement and, hence, the lower cost. An archaeologist can undertake a much wider sampling of site with amino-acid racemization than could be afforded using radiocarbon only.

Obsidian Dating.

Another technique strongly influenced by environmental conditions, particularly temperature, is obsidian dating. It is based on the slow formation of hydration rims on freshly cleaved obsidian. Until the late 1970s, the favored approach was by regional calibration using known-age samples. Since then the emphasis has been on obtaining absolute ages, independent of other techniques or archaeological chronology. This approach requires two major evaluations (in addition to measurement of the hydration rim itself): the effective burial temperature and laboratory measurement (at elevated temperature and pressure) of the rim growth rate for each type of obsidian that is dated. Dates have been reported in the age range from 200 to 100,000 years ago. Growth rates vary widely between different parts of the world; the comparatively rapid hydration rates in tropical countries allow more recent dating than in Arctic regions.

Other Chemical Processes.

The movement of fluorine and uranium into bone and loss of nitrogen are other processes by which some indication of age can be obtained. The bulk contents are strongly dependent on the burial environment; however, this dependence is very much weakened if the depth profile, as measured by a nuclear microprobe, is used instead. Nuclear techniques measuring penetration depth of nitrogen can be used to distinguish between freshly cut surfaces of jade, for example, and those cut in antiquity.

Archaeomagnetic Dating.

The direction of the earth's magnetic field (e.g., magnetic north as indicated by a compass) changes very slowly, but it is perceptible over a decade. A fossilized record of past direction is provided by the weak, but permanent, magnetization that is acquired by clay and stones on cooling down after baking and by sediment during deposition (or consolidation). The variation differs from region to region and does not follow a well-defined pattern. Hence, for each region of application it is necessary to establish a calibration curve using kilns, hearths, and silted-up ditches of known age. Consequently, the availability of a reliable archaeological chronology is a necessary prerequisite for applying the method. An alternative possibility is to make measurements on radiocarbon-dated lake sediment, obtained as a long core extracted from the lake bottom.

On a larger time scale, complete reversals of the earth's field provide a worldwide geomagnetic chronology, calibrated by means of the potassium-argon technique. The most recent event (named the Blake event) which is well established and generally accepted occurred at about 110,000 years ago; this was a comparatively brief reversal lasting not much more than ten thousand years. Earlier, at 780,000 years ago, there was the transition from the Matuyama reversed polarity epoch to the present Brunhes epoch.

Pollen Analysis and Other Climatic Indicators.

An archaeological site carries a record of past climate notably through preserved tree and plant pollen. Other indicators are faunal assemblage and soil type. This record allows the levels of a Paleolithic site to be described as cold or warm and to be related—often through the intermediate step of a locally defined climatic sequence—to the worldwide climatic chronology, based on oxygen-isotope stages. The successive cold and warm stages of this chronology are defined by the isotope ratio (oxygen-18–oxygen-16) found in marine fossil microfauna. Finest climatic detail is given by the isotope ratio variations measured in long cores drilled into the polar ice cups.

Absolute dating of the marine stages was initially based, via sedimentary remanent magnetization, on potassium-argon dating supplemented by radiocarbon and uranium series. The chronology so obtained has now been refined by reference to the Milankovitch astronomical predictions of the climatic variation resulting from changes in the earth's orbital parameters. The marine stages are now the fundamental time divisions of the Quaternary epoch.

Pollen characteristics (such as the dominant species of tree) can be used to define pollen assemblage zones (PAZ). The succession of such zones can be used as a basis for chronozones, defined in terms of radiocarbon years. For example, at a type-site the boundaries of the major PAZ are dated by radiocarbon and those dates then define the chronozones. Arboreal pollen, expressed as a percentage of total tree and herb pollen, is a useful broad indicator of climate, being high in warm/wet periods and substantially lower otherwise.

Bibliography

  • Aitken, Martin J. Science-Based Dating in Archaeology. London, 1990. Comprehensive discussion of the techniques outlined in this article.
  • Aitken, Martin J., et al., eds. The Origin of Modern Humans and the Impact of Chronometric Dating. Princeton, 1993. Papers based on a meeting held at the Royal Society, London, February 1992.
  • Aurenche, Olivier, et al., eds. Chronologies in the Near East. British Archaeological Reports, International Series, no. 379. Oxford, 1987. Papers based on an international symposium sponsored by the Centre National de la Recherche Scientifique, Lyon 1986.

Martin J. Aitken