4.6 Pollen Analysis

4.6.1 Introduction to the technique

Pollen analysis (palynology) is perhaps the most widely adopted, and arguably the most successful, of the biological techniques used in reconstructing past environments (Lowe and Walker 1997). Put simply, it is a method for investigating former vegetation by means of the pollen grains (and spores) that plants produce (Fægri and Iversen 1989; Moore et al. 1991). These sub-microfossils are typically found preserved in abundance in a wide variety of wet and acidic sedimentary deposits, such as lake muds and peats. Pollen may also collect within soils and be present in archaeological contexts (e.g. ditches, ponds, structural turves, plaggen soils, etc.). Pollen records within these materials can often be extremely difficult for palynologists to interpret, affected as they may be by problems such as poor pollen preservation or mixing (Dimbleby 1985).

Core samples for pollen analysis taken from natural peatland contexts (e.g. fens and bogs) are typically collected using a hand-operated ‘Russian’ corer such as the one pictured here, © J Schofield

Where polleniferous material is stratified and can be securely dated, it can be used to build up a picture of how individual taxa and plant communities have changed over time. This allows secondary deductions about, for example, climate change and human disturbance to the environment. The technique is flexible, enabling vegetation to be considered at a variety of geographical scales, from regional through to local (Janssen 1973; Jacobson and Bradshaw 1981). In cases where several pollen profiles have been studied across a landscape (e.g. Glen Affric, northern Scotland [Davies and Tipping 2004], and the Bowmont valley, southern Scotland [Tipping 2010]), this can allow spatial reconstructions of the vegetation mosaic, and provide greater confidence in linking cause and effect; anthropogenic impacts on vegetation are often seen as highly variable both spatially and temporally, whereas changes driven by natural factors (e.g. climate) may appear synchronous between sites.

The literature on pollen analysis is vast and extended discussion of this is impossible here. For Scotland alone, Tipping (1994) listed 239 sites providing palynological data for just the mid- to late Holocene, and this corpus of information continues to grow, providing excellent coverage for most areas (the Moray Firth being identified as one exception; Edwards 1999). A number of reviews are available which consider advancements within the technique and the role it has played in studies of vegetation change within Scotland (e.g. Walker 1984; Tipping 1994; Edwards and Whittington 2003). Consequently, the supporting references included here are necessary somewhat selective and the focus concentrated narrowly upon some of the more recent developments within the field, particularly where these relate to the interplay between people and the environment.

Pollen-based studies which consider the impact of people – and their domestic livestock – on past landscapes may now also include the analysis of coprophilous fungal spores, such as Podospora-type (HdV-368, pictured here), ©J Schofield.

Oldfield (1993, 16) comments that ‘interpreting palaeoecological data is rarely a matter of unambiguous, objective certainty’. Fossil pollen assemblages cannot, as yet, be directly translated into plant abundances, or used to produce maps of vegetation cover except at the broadest (landscape) scales, although research into modelling procedures is advancing in this direction. One central question that still needs to be addressed is defining exactly where the pollen that accumulates in any sediment has come from. Studies of modern analogues can help address this, and research into modern pollen-vegetation relationships in the Outer Hebrides (Brayshay et al. 2000) and northwest Scotland (Bunting 2003) have proved useful in furnishing information on treeless landscapes (e.g. grasslands, moorlands, heathlands and machair). Software for simulating landscape scenarios from pollen data is being developed and refined (e.g. Middleton and Bunting 2004; Bunting and Middleton 2005), and this approach could provide answers to questions often posed about the past vegetation mosaic and land use in the areas immediately surrounding prehistoric monuments and settlements. For example, Tipping et al. (2009) have already applied a modelling exercise to reconstruct the possible spatial arrangement of plant communities around a Neolithic ‘timber hall’ in northeast Scotland.

A number of supplementary techniques are commonly applied together with pollen analysis, yet there is little opportunity to go further than pass comment on these here. These commonly include analyses of site stratigraphies and sediments, microscopic charcoal (e.g. Edwards and Whittington 2000; Edwards et al. 2000; Tipping and Milburn 2000), and increasingly, consideration of non-pollen palynomorphs (NPPs; e.g. Clark 1999). The latter have always appeared in residues prepared for pollen analysis but, for many years, they received little attention from palynologists. The analysis of one group of NPPs in particular, the coprophilous fungal spores (van Geel et al. 2003; see Figure above), seems to be growing in importance given their proven value as indicators of past grazing intensity (Blackford and Innes 2006; Raper and Bush 2009).

4.6.2 Pollen analytical facilities and expertise within Scotland

Many of the universities within Scotland possess the facilities and expertise to undertake pollen-based research. Amongst those institutions which are currently active, palynologists at the Universities of Aberdeen and Stirling have long-standing research interests within Scotland, whilst the focus at the University of Edinburgh is upon vegetation reconstruction in Neotropical environments. A number of archaeological units (including Headland Archaeology and GUARD Archaeology) also offer to undertake pollen analysis on a commercial basis though preparation of samples is outsourced to other laboratories.

4.6.3 Recommendations for the future

Figure 32: Cereals such as rye, oats, wheat and barley characteristically produce larger pollen grains (typically >37 ?m) than other grasses. This makes them suitable for assessment using optimising techniques that involve the rapid scanning of pollen residues at low magnification (x100), © J Schofield

In the foreseeable future, the main applications of pollen analysis will (and should) be to continue to provide the environmental context from which discussion of the archaeological record may begin. Key themes which may continue to attract the attention of palynologists in Scotland, and which will also be of interest to archaeologists, include:

1. the visibility of Mesolithic hunter-gatherers in areas where archaeological data are limited or absent (cf. the Mesolithic in the Outer Hebrides [e.g. Bohnke 1988; Edwards 1996] and Shetland [e.g. Bennett et al. 1992; Edwards et al. 2009])
2. the continuity of upland settlement and restructuring of agriculture in response to climate change during the Late Bronze Age and Early Iron Age (e.g. Davies et al. 2004; Tipping et al. 2008b )
3. the precise timing and nature of the adoption of cereals, which may be advanced through optimizing techniques for cereal pollen detection (e.g. Edwards et al. 2005)
4. disentangling the causes that lie behind the prehistoric decline of pine woodlands in northern Scotland (e.g. Tipping et al. 2008b).

A series of other general challenges might also be usefully addressed. These include:

1. Continual improvements to pollen taxonomy, with consistent application of nomenclature between workers (cf. Bennett et al. 1994).
2. Advancing the understanding of the relationships between pollen assemblages and plant communities through modern analogue studies; this may enable, for example, tighter definition of the scale of woodland clearance events and land use activities in prehistory.
3. Improved mapping of plant communities in time and space from pollen data.
4. Clearly determining how to distinguish woodland management techniques (e.g. coppicing, pollarding, etc.; cf. Göransson 1987), and how to define key land use activities (e.g. summer transhumance, woodland pastoralism, etc.), using pollen analytical data.
5. Defining reliable measures of plant diversity from pollen data.