The remains of ancient plants can provide a wealth of archaeological information about a site, with many methods being available to the archaeologist engaged in extracting this data. Perhaps one of the most widely-known of these techniques, possibly because of its attractive nature, is pollen analysis - a technique developed in the early years of the twentieth century by, like so many archaeological techniques, a geologist -- the Norwegian Lennart van Post. To understand the technique and the uses to which it may be put, we must first examine the biological nature of the material itself.

Because of a hard outer shell - the exine - pollen is particularly resistant to chemical attack and will survive in most conditions; the only environments which are truly hostile to this shell are abrasion, such as may be the case on sandy sites, and oxidation. However, the most favourable conditions for preservation of the pollen record are acidic, anaerobic sites such as peat bogs. This high degree of survivability combines with another factor inherent in the nature of pollen - the large amount produced - to make pollen analysis one of the most important tools available to the archaeologist. Though one further factor in the make-up of pollen enhances its value, namely the wide morphological variation between pollen from different plant species, most of which can be detected and classified using normal laboratory equipment.

Pollen analysis is a technique which demands a high level of skill on the part of the excavator, scientist and interpreter to enable it to fulfil its potential. Collection of pollen samples can prove troublesome, the risk of cross-contamination is significant and efforts must be made to minimize the effect of any excavational bias. The number and ratio of pollen grains present in a sample can also be skewed by factors such as the orientation of the site and the nature of the pollen grains themselves, for example, trees such as pine produce much greater quantities of pollen than species such as oak and thus have a tendency to overrepresent themselves in the pollen record.

Once collected the pollen is extracted from the soil, usually in the laboratory to avoid contamination, and analysed using a light, or scanning electron microscope (SEM). The wide differentiation in the size, shape and colour of the pollen grains enables identification to be made down to genera level. Following identification, the individual exines in a sub-set of the sample are quantified and plotted on a pollen analysis diagram, usually as a percentage of the whole. However skill on the part of the interpreter is required here as, already mentioned, certain plant types can over-represent themselves and the transport mechanisms of different plants may also vary, some for example being dispersed by insects whilst others are wind-blown. Pollen may also be found on a site as the result of other activities such as food preparation and consumption. The question which has to be asked of any pollen report is to what extent is this a true representation of the nature of the site?

Once collected and analysed, pollen analysis has two traditional applications, dating and the reconstruction of former vegetation. Its use as a dating technique emerged out of the large-scale work done on the peat bogs of Britain and Scandinavia. The results demonstrated that over a period of about 10,000 years, several distinct zones of differentiation appeared in the pollen record representing changing arboreal environments. This work was extended by studying the ratio of tree-polen to non-tree-pollen enabling the zone theory to be further refined and extended back to the latter part of the Pleistocene. These reason for this zonation was ascribed to changing climatic conditions and subsequently used as a means of providing a dating structure for material recovered from archaeological environments. However since the development of radiocarbon dating, and its subsequent refinement by dendrochronology, the technique has become less widely used.

The second use of pollen, to reconstruct former vegetation has been more widespread however. Radiocarbon dating has provided a more accurate timescale for pollen sequences and samples for pollen analysis are now taken from lake sediments and soils as well as peat bogs, though the problem of wind-blown pollen from the surrounding area can serve to mask the local picture. Nevertheless, the analysis of the pollen record has emerged as the main method of constructing a picture of the former vegetational history of both large and small areas, being sufficiently detailed to even detect what has been interpreted as the actions of man in forest clearance during the neolithic and other periods.

One project where such techniques have been particularly successful is in the studies carried out on the Hadar settlements of the Ome Valley in Ethiopia. Here, the present dry, arid climate was found to originate from about 2.5 MYA. Before that date a much wetter and greener environment is suggested by the presence of pollen from tropical plants.

In summary, environmental data such as that from pollen can be seen to play an important part in the range of data retrieved from a site, though the expense of employing specialists to analyse and collate the samples many prove prohibitive in some cases. Equally important is that non-specialist archaeologists are thoroughly briefed on the significance and indeed the pitfalls of pollen data so that they can make informed syntheses and not represent the pollen data merely as an adjunct to the excavation report. It is clear that in many respects the data is capable of posing as many questions as it can answer and an informed approach is vital when considering the evidence from this source of data.