The Archaeological Potential of Secondary Contexts

Module 7 - Introduction

The goals of the module were the characterisation of the potential of the secondary context geoarchaeological resource for the reconstruction of riverine palaeo-environments, with particular reference to:

1. The range of biological data sources
These currently include vertebrates (including large and small mammals, fish, amphibians and birds), molluscs (non-marine), ostracods, coleoptera, pollen, and plant macro-fossils. It is also apparent however that new sources of palaeo-environmental data are becoming available, reflecting new technological and methodological developments. These include Cladocera (a freshwater crustacean – the water flea), Chironomids (non-biting midges), and testate amoebae (Evans & O’Connor 1999: 145; Charman 2001).

2. The relative potential of the different data sources
Based on the relative fragility of the different materials and their variable survival potential during re-working episodes, it was apparent that mammals and molluscs are more likely to be recovered from archaeological secondary contexts than any other type of palaeo-environmental data. This reflects their greater durability and assumes (see module 1) that high energy sediments (e.g. fluvial gravels) constitute the bulk of the deposits within these secondary contexts. However, it is also clear that where fine-grained and organic sediments are preserved, a wide range of biological data may be recovered (although taphonomic factors are still important). Finally, it was clear that local sedimentary regime variations are of greater significance than sampling strategies with regard to the diversity of biological data recovered.

Bridgland’s (1994) synthesis of the Quaternary of the Thames provided an excellent resource for assessing the relative abundance of different biological data within fluvial secondary contexts. The presence/absence of different biological data categories at each of 39 sites was documented, along with the main sediment types recorded at each site. Although it is clear that the preservation of different data categories can be strongly influenced by soil and sediment types, the 39 sites cover a wide range of geographical localities and this minimises (without wholly removing) the effects of localised preservational biases.

The most common categories of recovered biological evidence are large mammals, molluscs (primarily non-marine species) and pollen. Small mammals, ostracods, coleoptera and plant macro-fossils are also relatively common, while birds, fish, amphibians and reptiles were only rarely recovered.

What was clear from the review was the recovery of certain categories of biological data from specific sediments and deposits. For example, pollen, plant macro-fossils and insects (e.g. coleoptera) were predominantly recovered from organic deposits and other fine-grained, low energy sediments (e.g. pollen has been recovered from the Swanscombe lower loams (Bridgland 1994: 193-218), laminated silts, sands and clays at Purfleet (Bridgland 1994: 218-228), and (along with plant macro-fossils) from the organic silty-clays at Wivenhoe gravel pit (Bridgland 1994: 313-317). Small mammals and other small vertebrates (e.g. birds, fish, amphibians and reptiles) were also typically recovered from organic and fine-grained deposits (e.g. organic clays and silts at Great Totham), reflecting their vulnerability to physical and chemical destruction in higher energy fluvial deposits.

By contrast, only mammalian fauna (typically large mammals) and molluscs were recovered from the coarser-grained (and higher energy) fluvial gravels sediments, and in many of these cases, the molluscs were actually recovered from silt and sand lenses within those gravels. This is particularly well illustrated at the Long Hanborough gravel pit, where non-marine mollusc assemblages were sampled from distinctive silt and fine-sand lenses within the gravel unit (Bridgland 1994: 49-58). By contrast, mammalian fauna has often been recovered from gravel units, such as the Stanton Harcourt gravel, which yielded a cold-climate assemblage including Mammuthus primigenius, Coelodonta antiquitatis, Bison priscus and Equus ferus (Bridgland 1994: 65-79).

In summary, it is clear that the recovery of the different categories of biological evidence reflects a combination of several factors: soil and sediment chemistry; the depositional environments (e.g. high and low energy settings); the preservation of organic and/or fine-grained sediments within fluvial sequences; and the nature of the archaeological sampling. Given these variables, it should be clear that the potential range of biological data will vary considerably, but a case study synthesis of Bridgland (1994) does suggest that large mammal fauna and molluscs may be the most commonly encountered types of data within purely coarse-grained, fluvial gravel deposits and sequences. Naturally, the presence of finer-grained and organic sediments within those sequences will potentially increase the range of biological data, although regional soil conditions can still limit the available evidence, as at Broom (see module 2).

3. The spatio-temporal resolution of the biological data, mapped against different scales of hominid behaviour
The review indicated two contrasting scales of data resolution:

1. High-resolution, reflecting biological material that is highly sensitive to palaeo-climatic change and other aspects of palaeo-environmental variation (e.g. coleoptera and molluscs).

2. Low-resolution, based on biological material that is less sensitive to palaeo-climatic change and other aspects of palaeo-environmental variation (e.g. large mammal fauna and pollen).

These scalar differences function in both spatial and temporal spheres. For example, not only do beetles and molluscs change their geographical ranges rapidly in response to environmental change (high resolution spatial data) but they also have rapid generation times and population turnover rates, thus providing high resolution temporal data.

However, with respect to palaeo-environmental reconstructions, the robusticity of biological data and their susceptibility to destruction during fluvial transport is also a critical factor. Thus molluscs, beetles, ostracods and small mammals tend to be preserved in close proximity to their death environment and provide valuable data for local habitat reconstruction, while large mammals and pollen are far more susceptible to long range transport and are therefore often only reliable as indicators of regional and sub-regional palaeo-environments.

Based on the assessment of the different biological data types, we propose two classifications for their spatial and temporal resolution:

SPACE

1. Macro – large mammal fauna, pollen (the regional component) and arboreal macro-fossils (esp. trunk material).

2. Micro – small mammals, small vertebrates, molluscs, ostracods, pollen (the on-site component), coleoptera, and plant macro-fossils.

In this context, macro refers to regional spatial scales, as represented by river system catchments. While these can obviously vary in size (e.g. compare the River Thames with the River Axe), the key points are that the material has derived from a wide range of fluvial palaeo-habitats and from a comparable catchment to derived archaeological artefacts. The incorporation of large mammal fauna, arboreal macro-fossils, and the regional component of pollen assemblages within the macro-scale group reflects the robusticity of this material and their potential for long distance transport by water (fauna and wood) and air and water (pollen). By contrast, micro refers to ‘site’-based spatial scales. While it is impossible (and fruitless) to try and discuss specifically-sized spatial areas, these scales can be viewed as reflecting specific micro-habitats and landscapes (e.g. the Barnham channel and floodplain (Ashton et al. 2000) or the Boxgrove landsurface (Roberts & Parfitt 1998)). These data support the high-resolution reconstruction of individual palaeo-habitats, which may or may not contain the archaeological evidence of hominid behaviour. The inclusion of small mammals, small vertebrates, molluscs, ostracods, coleoptera, pollen (the on-site component) and plant macro-fossils reflects the general fragility of this material and the taphonomy of their deposition.

TIME

1. Macro – large mammal fauna

2. Intermediate – pollen (the regional component), molluscs

3. Micro - small mammals, small vertebrates, ostracods, pollen (the on-site component), coleoptera, and plant macro-fossils.

In this context, macro refers to marine isotope stages, or marine isotope sub-stages, reflecting the broader environmental tolerances of large mammal species, low turnover rates and generation times, and the relatively coarse biostratigraphical signatures of large mammal associations (Schreve 2001a, 2001b). Moreover, the presence within high energy fluvial contexts and varied physical conditions of large mammal fauna indicates their ability to withstand extensive re-working and derivation. The occurrence of large mammal fauna within a specific fluvial sediment (especially coarse-grained gravels) is therefore not a high-resolution temporal indicator (i.e. death did not necessarily occur immediately prior to the deposition of the sediment). This contrasts markedly with the fragile micro-fauna (e.g. coleoptera, ostracods, small mammals and other small vertebrates), whose intact preservation tends to indicate a short time period (and limited transport and derivation) between death and sedimentary accumulation.

The classification of pollen and molluscs within the intermediate group reflects a range of factors. For pollen, the robusticity of the material indicates its potential for surviving re-working episodes, and ultimately being deposited in pollen assemblages where the ‘fresh’ pollen is centuries or even millennia younger. In the case of molluscs, their apparent occurrence within coarser-grained gravel and sand deposits (e.g. Bridgland 1994) suggests a degree of robusticity and the potential to survive re-working episodes within high energy contexts.

These classifications are clearly intended as guidance rather than a hard rule. For example, certain elements of large mammal fauna are far more robust than others (e.g. limb bones compared to cranial material), while small mammal fauna may be preserved over several phases of derivation and re-working in exceptional circumstances. However, the physical conditions of this material (e.g. fauna, shells, pollen exine) can often be employed as an indicator of atypical levels of derivation and re-working in time and space.

Overall therefore, macro-scale (as defined above) palaeoenvironmental reconstructions are primarily dependent upon large mammal faunal assemblages within secondary contexts. These species associations have been widely employed to reconstruct broad-scale environments and climatic conditions. For example, in the Aveley (Sandy Lane) mammal assemblage zone (MAZ) the predominance of species such as woolly mammoth and horse suggests an open grassland environment, while decreasing numbers of straight-tusked elephant and Merck’s rhinoceros suggest a reduction in woodland coverage (Schreve 2001b: 1701-1702). These reconstructions inevitably cover large spatial areas, reflecting the broad ecological tolerances of many large mammal species, the large ranges of individual species, and their wide geographical distributions (Schreve has traced her MAZ’s into western and central Europe (Schreve & Bridgland 2002)). The reconstructions also span long time periods (e.g. MIS stages and sub-stages), reflecting the difficulties in assessing the time depth of derived fossils, and the relatively stable, non-ephemeral nature of large mammal communities. In effect therefore, these reconstructions are essentially time-averaged, and yield a reliable overview of the mammal fauna (associated with a specific terrace unit for example), although it must be remembered that it will not detect micro-variations in space and time. This has been demonstrated by Schreve (1997, 2001a, 2001b, & Bridgland 2002), whose development of mammal assemblage zones has documented the repetitive occurrence of distinctive faunal assemblages and species associations. These species associations minimise the dangers associated with the employment of single indicator species, which are particularly prevalent in secondary contexts subject to re-working.

Macro-scale palaeo-environmental reconstructions can also make potential use of pollen assemblage zones from organic deposits within secondary context sequences, with specific reference to the reconstruction of regional vegetation development, palaeo-climatic conditions, and broad scale climatic change. However, in light of the major taphonomic complexities associated with pollen assemblages, and the typically low-resolution, fragmented pollen sequences from fluvial secondary contexts (Thomas 2001), it is suggested here that regionally-derived pollen data from fine-grained deposits within fluvial secondary contexts will typically be strongly time and space-averaged. This is also true for arboreal macro-fossils occurring within both fine and coarse-grained fluvial sediments. Overall, these data may reveal broad patterns in palaeo-climate and regional vegetation types, although temporal trends in vegetation development will be difficult to detect. The nature of fluvial sequences and the problems of pollen taphonomy also highlight the dangers of utilising single indicator species to document palaeo-climatic change.

In contrast, micro-scale palaeo-environmental reconstructions are primarily dependent upon a range of biological data (small vertebrates, molluscs, ostracods, coleoptera, pollen), from a wide range of coarse and fine-grained secondary contexts. These data can provide a wide range of extremely high-resolution data, indicating local environment conditions and short-term change (e.g. fish species can indicate aquatic temperature and flow conditions, beetle distributions vary in response to vegetation, substrate types, hydrology and/or micro-climate, while terrestrial mollusc species vary in response to vegetation cover and soil saturation levels). However, it is apparent from the extant literature that the time-spans associated with particular micro-habitat conditions and/or with climatic change are rarely explicitly stated (if they are known at all).

These high-resolution biological data sources may be used within biostratigraphical models (e.g. Thomas 2001; Keen 2001; Preece 2001; Coope 2001). These may potentially assist in macro-scale palaeo-environmental reconstructions, both through indicating broad palaeo-climatic patterns (e.g. the summer temperature contrasts between MIS-5e and MIS-7), and by facilitating comparisons with high-resolution data from other sites.

Having highlighted the spatio-temporal resolution of the available biological data, the module also sought to map hominid behaviours against the reconstructed palaeo-environments. Four types of hominid behaviour are identified, at two key scales:

1. On-site activity (micro-scale behaviour). This primarily covers tool production and subsistence activities (e.g. carcass butchery), reflecting those activities that are most commonly recognised in the archaeological record (e.g. Roberts & Parfitt 1998; Ashton et al. 1998; Ashton et al. 2000; Gowlett & Hallos 2000). It is of course recognised that these activities do not represent the full range of hominid daily existence.

2. Technological change (macro-scale behaviour). This relates to the appearance of new technological innovations in the archaeological record (e.g. the appearance of Levallois technology in late MIS-9 and early MIS-8) and/or changes in technological practise (e.g. the shift from Clactonian to Acheulean during MIS-11 and MIS-9 (White & Schreve 2000). It is recognised that technological change also occurs at the micro-scale (the scale of individual technological innovations), although this is relatively difficult to detect. White’s (1998a) analysis of changing biface shape at Swanscombe in response to landscape transformation and variations in raw material supply operates at the sub-marine isotope stage level, but does still not reach micro-scales as defined here.

3. Demographic change (macro-scale behaviour). Evidence of demographic change is difficult to detect and verify, even at macro-scales, during the Middle Pleistocene, although recent research utilising artefact densities as a population proxy (Hosfield 1999; Ashton & Lewis 2002) has begun to suggest the presence of robust patterns.

4. Responses to climatic change and perception of environments (micro and macro-scale behaviours). It is suggested that hominid responses to environmental change (e.g. climatic, vegetational) are more likely to have operated at similar spatial and temporal scales to the large mammalian fauna, although their sensitivity to micro-changes (e.g. in flora and micro-fauna) remains a moot point.

A preliminary mapping of these behaviour types and scales against the palaeo-environmental data is therefore proposed:

A. On-site knapping and subsistence activities (micro-scale) – where the archaeological debris of these behaviours is in primary context, they can be directly mapped against any micro-scale biological data within the same sedimentary unit (which as outlined above is likely to be multi-proxy and high resolution). However, we are not dealing with archaeological primary contexts. Where the archaeological debris is derived, not only is the behavioural information of lower resolution, but it also cannot be directly mapped against the biological data. This is true whether the biological data occurs in the same sedimentary unit as the archaeology, or in other units within the secondary context sequence. The biological data and reconstructed (high-resolution) habitats can be discussed with respect to the archaeology, but only as examples of the types of environments within which the hominid behaviour may have occurred. In a similar manner, macro-scale palaeo-environmental data can also be mapped against these types of micro-scale archaeological debris, but only as an indicator of generic, rather than specific, environmental conditions (i.e. direct associations can obviously not be demonstrated, reflecting the time and space-averaged nature of both sets of data).

B. Technological change (macro-scale) – this is most commonly represented through changes in lithic technology at the marine isotope stage or sub-marine isotope stage level (e.g. Conway et al. 1996; White & Schreve 2000), and the data typically consists of time-averaged assemblages. As previously, these data cannot be directly mapped against high-resolution biological evidence, due to the spatio-temporal contrasts. The latter data can only be employed as an example of the multiple possible habitats and environmental conditions that existed during the timespan over which the archaeology was deposited. However, the archaeology can be indirectly associated with the low-resolution biological evidence, based on their complementary scales (both data sets are time and space-averaged palimpsests, encompassing regional and extra-regional space and marine isotope stage-time). The biological data provides a low-resolution, large scale image of mammalian communities and palaeo-landscapes over defined time-spans (e.g. MIS-10 or MIS-11e), against which changes in hominid technology and behaviour can be tested at the MIS and sub-MIS scale.

C. Demographic change (macro-scale) – this has been recently demonstrated through MIS variations in artefact density, within fluvial secondary contexts (Ashton & Lewis 2002). As with the technological change data (see above), three key points are evident:

C.1 These palimpsest data sets cannot be directly mapped against high-resolution biological evidence, due to the contrasts in the spatio-temporal scales.

C.2 The high resolution biological data can only be employed as an example of the multiple possible habitats and environmental conditions that existed during the timespan (MIS in the work of Ashton & Lewis (2002)) over which the fluvial secondary context archaeology was deposited.

C.3 The archaeology can be indirectly associated with the low-resolution biological evidence, utilising their shared time and space-averaged palimpsest structures. The low-resolution biological data (defined by MIS-units) provides an environmental framework against which changes in hominid demography can be tested. Such possible connections (e.g. between technology, hunting strategies, social structures, environments and biota) have been more fully explored by Ashton & Lewis (2002).

D. Responses to climatic change and perception of environments (micro- and macro-scale) – this aspect of hominid behaviour is intended to highlight elements of Palaeolithic societies which are typically ignored in favour of tool making and subsistence activities. The premise is relatively simple: that as another social mammal, hominids may well have perceived and reacted to environmental and climatic change at broadly similar spatio-temporal scales to the large mammalian fauna. Since macro-scale data (e.g. Schreve’s MAZ’s) provides a range of data regarding mammalian distributions at the MIS and MIS sub-stage scales, it is argued that hominids could potentially be mapped against these patterns to explore trends in Palaeolithic occupation and migration (e.g. the apparent abandonment of Britain during MIS-6). Finally, it is noted that hominids may also have been sensitive to the high-resolution environmental and climatic changes evident in micro-flora and fauna, suggesting that micro-scale palaeoenvironmental data may also assist in the interpretation of hominid behaviour.

4. Relationships between palaeo-environmental data and current research questions
A brief review of the extant literature regarding current British Palaeolithic research suggests a core of key themes to which the interrogation of archaeological secondary contexts is relevant:

1. The earliest occupation of ‘Britain’ during the Pleistocene (e.g. Roberts et al. 1995; Roberts & Parfitt 1998; Wymer 1999; Rose et al. 2001).

2. Patterns of colonisation and demography during the late Middle Pleistocene (e.g. Hosfield 1999; White & Schreve 2000; Ashton & Lewis 2002).

3. Palaeolithic technology and hominid behavioural repertoires (e.g. Roberts & Parfitt 1998; White 1998a, 1998b; Gamble 1999; Wenban-Smith 1998, et al. 2000; White 2000, & Schreve 2000).

4. The geochronological frameworks of the Middle and Late Pleistocene (e.g. Maddy et al. 1998; Bridgland 1998, 2000, 2001; Maddy et al. 2001; Schreve 2001a, 2001b; Bridgland & Schreve 2001; Current & Jacobi 2001; Preece 2001; Coope 2001; Keen 2001; Schreve & Bridgland 2002; Schreve et al. 2002).

This review of palaeoenvironmental data sources has indicated that they cover a wide range of spatial and temporal scales, from stenoptic beetles to continent-ranging mammoths. These scales mirror the variable and wide-ranging resolutions evident in archaeological material from the Middle and Late Pleistocene. Their applications therefore range from high-resolution, small-scale and multi-proxy reconstructions of micro-habitats and local environments (e.g. Roe 2001; Schreve et al. 2002), to the low-resolution, large scale mammal assemblage zones of Schreve (1997, 2001a, 2001b). In the case of the ‘earliest occupation’ and ‘geochronological frameworks’ themes outlined above, the applications of the palaeoenvironmental data are:

A. Earliest occupation theme: linking biostratigraphically significant species (or assemblages of species) to the occurrence of otherwise undated archaeological materials within both primary and secondary archaeological contexts (e.g. the presence of Stephanorhinus hundsheimensis, Ursus deningeri, and Arvicola terrestris cantiana in the archaeological sediments at Boxgrove (Roberts & Parfitt 1998)).

B. Geochronological frameworks theme: linking biostratigraphical models to existing frameworks (e.g. absolute dating and/or terrace stratigraphy models). While it is clear that different schemes are sometimes contradictory (e.g. mammalian and coleoptera are at variance with pollen biostratigraphies (Keen 2001)), these biostratigraphical approaches offer multi-proxy data and potential sub-MIS stage resolution (e.g. Schreve 2001a).

However, the application of these data to themes 2 (colonisation and demography) and 3 (technology and behaviour) is more difficult to identify. This is primarily due to the significant contrasts between the spatio-temporal scales of the various data sets. As discussed in section 4, it is suggested here that the direct applications of high-resolution palaeoenvironmental data are limited with respect to secondary context archaeology. While the derived archaeological data provide valuable time-averaged insights into technology and demography, the cultural debris cannot be demonstrably associated with the reconstructed habitats. In other words, while the archaeology provides direct evidence of hominid presence (within coarse timescales), that presence cannot be related to specific local habitats and environments. Those environments can only be presented as examples of the range of environments and habitats that were present, and which may or may not have been encountered by hominids. This last issue is specifically highlighted, since it is often not made explicit within multi-proxy investigations of Quaternary sequences.

Nonetheless, low-resolution palaeoenvironmental data can be indirectly linked at the scalar level with secondary context archaeology. This reflects their comparable scales of magnitude (e.g. river system catchments incorporating derived artefacts and fluvially transported mammal remains into time-averaged fluvial sedimentary sequences). These data are inevitably low in resolution and coarse-scale, but provide robust indications of broad trends, both in hominid technology and mammal/tree species distributions, although it is stressed that this data mapping does not assume that encounters between hominids and specific flora/fauna associations occurred. Such data are nonetheless extremely valuable (they can be employed to test multi-MIS changes in hominid demography, extra-regional colonisation, and technology, and potential relationships between hominid behaviour and the environment) and highlight the importance of regional data sets and macro-analysis approaches.

It is clear that the spatio-temporal resolutions of the biological data varied markedly by species and type, and that secondary context archaeology cannot be directly mapped against high-resolution palaeoenvironmental reconstructions. It is argued here however that secondary context archaeology and low resolution palaeoenvironmental data occur at similar scales of magnitude, and can be employed to generate comparable, low resolution models of different aspects of the Pleistocene environment. Finally, it is stressed that these relationships must be made explicit (irrespective of the specific questions), since there are fundamental scalar contrasts between the archaeological and biological data.

Module 7 Outline

Module 7 Interm Report

References

Return to top