With the enormous growth in bioarchaeological studies in the past two decades, the need for databases in which to store large bodies of data and from which to recover information quickly and flexibly has become vital. This paper offers a brief introduction to one such database, the ABCD, and illustrates its potential by means of a survey of one category of evidence available from it.
The ABCD is a relational database containing information about records of macrofossil plant remains from archaeological deposits throughout the British Isles. It was compiled by PRT between 1989 and 1994 (though most of the data included were published prior to 1992) and is currently being maintained at the Environmental Archaeology Unit (EAU), University of York, by ARH. The data are stored in a series of 14 tables linked by certain common fields; at present, the software primarily intended for use in interrogating and manipulating the data on a PC is Paradox, but the tables may be easily exported as ASCII files for use by word-processing or spreadsheet software.
As far as possible, copies of original reports whose data have been abstracted for the ABCD are held in a paper archive at the EAU. It is hoped that this will eventually form the basis of a National Archive of Environmental Archaeology.
In writing this paper, we have tried to bear in mind both the expert and non-expert user and will no doubt have failed to satisfy either! Some aspects of the background to archaeobotany have, necessarily, been dealt with very summarily. Our choice of illustrations has been limited by the material we had to hand when the paper was commissioned; inevitably it represents only a biased selection of remains relevant to the main subject of the paper. It should not be assumed that such material is likely to be found on all archaeological sites, although it is our contention that delicate food remains (of the kind illustrated) preserved by anoxic waterlogging' may be more frequent than the number of records suggests -- they are easily overlooked and often require considerable expenditure of time in preparation of both fossil and modern reference material before identification is possible.
Direct evidence for the plants which were used for food in the past comes from the study of a variety of archaeobotanical materials. Grain, seeds, fruit stones, nutshells and other more delicate parts such as fragments of epidermis may be preserved on archaeological sites by several mechanisms. Although the route that is taken by these materials through the stages of use, disposal/deposition, preservation and subsequent excavation is very complex and sometimes difficult to interpret, a wealth of information has been obtained, especially in the last 20 or so years. The mechanisms of preservation of plant remains are:
Inevitably, these operate to different degrees on different sites, depending on the conditions within the deposit both at the time of formation and during subsequent diagenetic changes. Moreover, different kinds of plant materials are more likely to be preserved under some conditions than others: it is well known that on 'dry' occupation sites with well-drained deposits preservation will be by charring and that the kinds of remains preserved will primarily consist of cereals, pulses and associated crop weeds.
Where there is extensive 'waterlogging', on the other hand, a much wider range of taxa may be preserved, including delicate plant tissues such as leaf and stem fragments, though cereals and pulses may be relatively poorly represented.
Mineral-replacement typically occurs in latrine-pit fills and other contexts where a high concentration of calcium salts obtains (discussed by Green 1979, 280); here plant tissues are largely or wholly replaced by these salts and the resulting structures may sometimes be casts or moulds, as seen in 'conventional' fossils from sedimentary rocks, rather than simply the original fossil.
Desiccation is a mechanism of preservation rarely seen in British material, but commonly encountered in arid regions (eg van der Veen 1996).
It is over ten years since Greig reviewed the evidence for food plants in the past in northern Europe (Greig 1982) and Green compared the archaeological evidence for food plants with that from historical documentary sources (Green 1984). Zohary and Hopf (1994) describe the origin and spread of cultivated plants in West Asia, Europe and the Nile but do not discuss the presence of food plants in Britain in any detail, only giving results published by Jessen and Helbæk (1944) and Helbæk (1952). It is therefore timely to review the evidence again and more fully. This paper draws on the results of many recent excavations and discusses the development of new techniques which have added considerably to our knowledge. Illustrations of some of the food plants discussed, especially vegetative materials, identified from excavations in York are also presented (Figures 3-10)
The choice of taxa for inclusion in this analysis is not entirely straightforward because some species which are now regarded as weeds might have formed an important part of the diet in the past. Plants such as certain Chenopodium spp. (eg fat-hen, C. album and good King-Henry, C. bonus-henricus), Urtica sp. (nettles) and Stellaria media (chickweed), remains of whose fruits and seeds occur frequently on most types of site, may well have been a source of food in the prehistoric period and perhaps later, if only during times of famine. These species are very common in what are interpreted as the 'disturbed ground' component of plant macrofossil assemblages from archaeological deposits so it is impossible to distinguish them as food plants unless they occur in faeces or as stomach contents (see below and especially Hillman 1986). Such plants have not been included in this analysis. The somewhat fluid relationship between certain crop plants and weeds is interesting. Some species, such as Bromus secalinus ('chess'), may have been cultivated in early periods, but survive today only as weeds. Others, such as Camelina sativa (gold-of-pleasure or false flax), first appeared as a weed of another crop, in this case flax, before it was cultivated in its own right in the medieval period.
Other plants, such as blackberry, Rubus fruticosus, and elderberry, Sambucus nigra, would have been a source of berries available as a wild resource at any period but may have been more significant as weeds growing around habitation sites. Bramble and elder seeds are often found in archaeological deposits where few other taxa are preserved, and sometimes as high concentrations of seeds. This may be explained by their relative robustness in preservation or by the fact that they were growing around the site at the time of deposition, perhaps in an area of abandoned ground.
Some of the seeds found in archaeological deposits may represent plants used for a wide variety of purposes other than food, such as medicines, or textiles, or they may have been brought to the site by chance with materials. The hazel (Corylus avellana) nuts found in very large numbers on the floor of the recently excavated Rose Theatre in London (not yet published), dating to the 1590s, may have been the remains of nuts eaten during performances. Another interpretation, however, is that they were a by-product from elsewhere and thrown on the floor to improve the drainage. Although this latter interpretation is perhaps less likely, it provides an example of the problems of interpretation.
Thus not all the records of 'food' plants necessarily represent plants actually used for food at the time and place indicated by their archaeological context and, as discussed below, they do not all represent the remains of food material actually consumed. It is often quite impossible to distinguish such non-food plants when there is no clear evidence of faeces. Even when faeces are present, in most cases there is also a mixture of other rubbish material derived from other sources. The interpretation of introduced species such as figs, grapes and cereals as food plants is perhaps more straightforward, however.
Very few workers ever suggest any doubt that the food plant remains from occupation sites represent human food. Where humans and domestic animals were living in close proximity, it seems very likely that waste material from both sources would be incorporated into site accumulations, middens and pits. Sometimes the material can be shown to be from domestic animals - for example, the food plants, such as cereals, which were incorporated in organic accumulations in Roman levels at the General Accident site (24-30 Tanner Row), York 83-4 [1616]. The range of evidence from the biological materials in these accumulations indicated that it was largely stable manure mixed with a range of other organic refuse (Hall and Kenward 1990, 352). It may be difficult to interpret equivalent assemblages where the preservation is not as good as at this site.
Material from faecal concretions and latrine deposits can be shown to be specifically human when it is associated with human intestinal parasite eggs (Hall et al. 1983). However, the eggs of species of intestinal parasites of other host animals, such as horses and pigs, have also been found (Jones et al. 1988), and it is clearly important to be able to distinguish parasites of humans from those of their domesticates if evidence for foods is to be apportioned correctly by consumer. The recycling of human food waste (and faeces) by a variety of animals on habitation sites merely adds another layer of complexity here!
The definition of a food plant, for the purpose of this paper, also depends on the part of the plant which is preserved in the archaeological deposits. Apples, for example, are only counted where the records are of pips and 'cores' or endocarp fragments, not of Malus sp. wood or charcoal. Similarly, remains of parts of plants which are edible and likely to have been eaten may have originated through some other process; seeds of flax (linseed, Linum usitatissimum), for example, may indicate waste from processing of flax stems for fibre rather than seed consumption - here the archaeological context and the sum of other bioarchaeological evidence may be decisive in distinguishing a use in food.
A particular interpretative problem concerns taxa where the part recovered as a fossil is the propagule (fruit, seed) but the part used for food is vegetative. An example of this is the carrot, Daucus carota, for which the only fossil records are of the mericarps (fruits); although perhaps used at times as a flavouring (the fruits are aromatic like those of fennel and dill), it is much more likely that the fossils recovered originated in calcareous grassland where carrot grows wild (and reached the archaeological deposit via hay or in herbivore dung).
Sometimes the 'secondary' evidence for the use of a plant as food seems less ambiguous. The field bean (Vicia faba) pod fragments from Anglo-Scandinavian occupation deposits at 16-22 Coppergate, York (Figures 7-9), seem unlikely to have been discarded except as waste from food preparation. This is a case where the primary evidence of the plant concerned is especially likely to be limited since remains of field bean seeds are rather rarely preserved by waterlogging, the main means of preservation in the Coppergate deposits.
To summarise, the remains of different species of food plants which are found in archaeological contexts will not have arrived there by the same taphonomic process. There are many variables:
(i) The foodstuffs may have been:
(ii) The plant remains may represent
(iii) It is only in exceptional circumstances that the food taxa found in any one sample might be shown to represent one meal, perhaps the contents of an individual's stomach, the best preserved examples coming from bog bodies such as Lindow Man (Hillman 1986). The fill of a latrine may represent faeces from a particular group of people (although this is difficult to prove). But most often the information comes from material of mixed, heterogeneous deposits where the results can only be tied to a relatively broad group such as the Anglo-Scandinavian folk living in several tenements at Coppergate, York.
For simplicity, the use of the word 'seed' has been used throughout this paper rather than the specific botanical terms carpel, achene, nutlet, and so on. Other parts of plants, such as fruitstones and nutshells are specified, where appropriate. There is also a glossary that can be consulted. The following abbreviations for the broad chronological units are used throughout this paper (see date codes in the glossary for further period divisions):
Figure 1: Map showing the location of archaeological sites with food plants from charred contexts ('m' type published reports only). Sites with work still 'in press' not shown.
Figure 2: Map showing location of archaeological sites with food plants from contexts of waterlogged and mineralised preservation regime ('m' type reports only).
The two maps (Figures 1 and 2) show the location of archaeological sites with 'm' (macrofossil) type published reports containing food plant remains (ie reports which deal only with samples of cereal grains ('g') or 'spot' identifications of other food taxa ('i') are not shown; definitions of 'm', etc. are given in the ABCD User Guide). Figure 1 shows sites with contexts with charred preservation regimes. The concentration of research on sites in southern England is clearly shown. Figure 2 shows sites with contexts of waterlogged and mineralised preservation regimes. Here, there is a reasonable distribution of sites in England (although there are obviously clusters in the towns, particularly Oxford, York, London, Norwich and Southampton).
The results of the analysis of the occurrence of food plants in the ABCD are presented in three tables.
Before discussing these results, some explanation of the methods used for the production of the tables are necessary. The species are listed in the order in which they first appear in the archaeobotanical record. They have been sorted on the basis of closely-dated records only. The numbers represent the total number of sites where evidence for each food plant has been found. The numbers only give the total figure for the broad period (such as prehistoric, Roman and so on).
Before comparing the numbers between periods it is important to note the total number of sites which have produced food plants for each period (see Table 1). It should also be noted that no indication of the quantification of the different species in particular samples or ubiquity in samples from individual sites has been given here. For the purposes of this broad review this was felt to be less important, especially as some taxa are always over-represented in comparison with others because of the variation in number of propagules produced by one fruit. For example, the hundreds of seeds produced by one fig, the single stone in a date, or the small number of pips in a grape, cannot easily be compared in terms of the quantity of food consumed or the relative importance of different fruits in the diet.
As mentioned above, prehistoric and Roman records of food plants are included here for comparative purposes in order to place the taxa recorded from later periods in perspective in terms of introduction and use. It is not the intention to provide a full discussion of these periods or of every taxon listed.
The range of wild foods collected in the Mesolithic period were, of course, commonly collected throughout all periods (see Table 2). Sloes (Prunus spinosa), for example, are particularly commonly recorded from Anglo-Saxon/Anglo-Scandinavian cess pits. A recent thorough review of food plants in the Neolithic period (Moffett et al. 1989) suggests that collected wild foods were as important as cereals throughout this period. Although this evidence is based only on charred plant remains, the increase in 'bulk sieving' over the last decade has altered the conventional view of the Neolithic economy which was derived mainly from cereal impressions in pottery (eg Helbæk 1952).
Relatively more settlement sites have been excavated from the Bronze Age and Iron Age than the Neolithic period. Perhaps not surprisingly, the range of species gradually increases. Apart from the cereals, other introduced, cultivated, plants which occur during the prehistoric period are flax, beans and peas. Opium poppy (Papaver somniferum), which first appears in the Bronze Age, was perhaps part of the cultivation assemblage too. There is one record of grape from the Neolithic period (Jones and Legge 1987). It remains to be seen whether further work on the prehistoric period will extend the record for grape and other introduced species. Further discussion of these taxa will be found below.
The Roman period saw the introduction of a great variety of herbs, spices and vegetables (see Table 3). There are some taxa which disappear from the record in the immediate post-Roman period and do not re-appear until the medieval period (eg date, parsley, stone pine) or until the post-medieval period (in the case of cucumber). Some species do not occur again after the Roman period. With the exception of olive, these are mostly taxa represented by very rare records.
Evidence for plant remains from the immediate post-Roman period is very thin (see Table 4). There have been relatively few excavations and few which produced plant remains. The information is also very patchy in geographic terms as there are some early Saxon sites in southern England which cannot necessarily be compared in cultural terms with activity further north.
It is not until the Anglo-Saxon/Anglo-Scandinavian period, around the mid 9th century that the way people were living and depositing their rubbish around them in the rapidly developing towns began to produce ideal conditions for the preservation of plant remains, namely, the widespread practice of depositing faecal material in cess pits. Coupled with this is the increase in the excavation of these well-preserved, waterlogged, deposits and the extensive sampling programmes which have been carried out on these towns. This work has produced a wealth of evidence, showing the range and variety of plant material used by ordinary people (traders and craft workers) living in towns.
In the medieval period, evidence from some sites can be related specifically to the people living in particular houses (eg the cess pit from the Merchant's House in Tenby [1363]). There are reports from a variety of sites such as castles and monasteries as well for urban deposits, but there are far fewer of these and the preservation is not so good - the drains and sewerage systems of most monastic sites, for example, were so well organised that finds of faecal material or kitchen waste are rarely found to be preserved when the fills of such features are excavated and analysed. Dickson, however, has recently found a 'diversity of plant remains' from a 15th century sewer leading from Paisley Abbey, Scotland which includes Ficus carica seeds (Dickson in press). It is also difficult to compare evidence from rural and urban sites because far fewer rural sites have produced well-preserved archaeobotanical evidence of a comparable nature to that from urban organic accumulations.
© Internet Archaeology
Last updated: Thu Jun 27 1996