7 Discussion of individual taxa

In the following descriptions the emphasis is on fruits and vegetables. Plants from other categories are also mentioned where some aspect of their distribution, preservation or identification was considered worthy of discussion. The eight groups - berries, other fruits, vegetables, pulses, herbs and flavourings, oil plants, cereals and nuts - are somewhat arbitrary.

7.1 Berries
Please Note: + indicates additional, poorly dated material.

7.1.1 Fragaria vesca

The earliest record of wild strawberry is from Mesolithic Westward Ho! 83-4 [1055]. There are two Bronze Age records: Oakbank Crannog [1374] and Runnymede 78 [1638]. From the Roman period onwards strawberry seeds are common in deposits with faeces.

7.1.2 Ribes uva-crispa

Despite its probable native status in Britain (Stace 1991, 373),gooseberry does not appear in the archaeobotanical record until the late Saxon/medieval period at Frewin Hall, Oxford [0071], and the medieval period at North Mains 78-9 [1071] and Worcester Barrel 75 [1601]. Smith (1976, 309) suggests that domestication of the wild plants did not take place until the 16th century. There is, however, evidence of its use in the 13th century (Harvey 1981, cited in Greig 1988, 60).

7.1.3 Rubus fruticosus agg. and R. idaeus

Blackberry seeds occur commonly and throughout the record, percentage site presence rising to nearly 50% in the medieval period. Not all these records necessarily represent food, however, bramble often growing on neglected areas of settlement sites and the seeds being likely to be 'sown' into deposits by passing birds. The seeds are relatively robust and are often found where little else has preserved in the deposit. Raspberry seeds occur throughout the record but are less common than blackberry.

7.1.4 Rubus caesius and R. chamaemorus

There are surprisingly few records of dewberry. It may be that the seeds have been recorded as Rubus fruticosus agg. and are therefore under-represented. The only prehistoric record is from Iron Age Collfryn 80-2 [1546]. Cloudberry has an upland distribution; the one prehistoric (charred) record from a late Iron Age settlement site [1243], Dubby Sike, may not represent food; it could have been incorporated into the charred assemblage by chance.

7.1.5 Sambucus nigra

Like blackberry pips, elderberry seeds are robust and are often the only plant material to be preserved in some deposits. They also survive quite well in non-waterlogged soils and frequently appear as modern contaminants in samples taken from near the modern ground surface. The preservation mechanism for recorded elderberry seeds is 79% waterlogged, 9% mineralised and 12% charred. This is calculated on the basis of the numbers of records in the database, not on total numbers of individual seeds.

7.1.6 Sorbus aucuparia

Concentrations of the seeds of such species as rowan might be found when they had been used for jellies and jams, as perhaps at 13th-14th century Queen St Aberdeen [1135], where 130 rowan seeds as well as an equal number of Vaccinium sp. seeds were found in a 1 litre subsample. None of the other records are for more than a few seeds, or they are given as 'present' only.

7.1.7 Vaccinium spp.

This includes Vaccinium myrtillus (bilberry) although most workers do not identify the seeds to species level. In the particularly well-preserved deposits in York, Coppergate 76-81 [1920], epidermis fragments and tori (the discs at whose centre the pistil sits) from these fruits have also been found. The latter allows a certain identification of Vaccinium myrtillus to be made.

Although there is a variety of edible wild fruits and berries available in the native flora of the British Isles, several are strongly or mildly poisonous. Prunus padus, for example, tastes bitter and contains cyanide (Lang 1987, 110) and certainly would not have been consumed without dire consequences. Several others which are only mildly poisonous can be eaten after boiling e.g. Sorbus aucuparia (Lang 1987, 77). Sambucus nigra berries are mildly poisonous if unripe or undercooked (Lang 1987, 173). It is interesting to imagine the experimentation of the earliest hunter-gatherers with tasting and cooking poisonous berries to make them palatable.

7.2 Other fruits

7.2.1 Cydonia/Pyrus sp.

The are no records of quince seeds from the archaeological record. The identification of some stone cells from the flesh of the fruit from faecal deposits in Watergate St, Chester 85 [1017] and Queen Anne House, Shrewsbury [data not yet added], may be of pear or quince (see Greig 1988, 61). For further discussion of pear see under apple below.

7.2.2 Ficus carica

Fig seeds are particularly common in Roman and medieval cess pit deposits but there are only two well-dated records for the 'SE' period. These are from Saxo-Norman London [1736] and Anglo-Scandinavian York [1920]. At both sites the number of seeds is very low compared with some medieval cess pits where fig seeds are found in tens if not hundreds. Despite the fact that they are waterlogged, it is possible that the records of fig from the early contexts at Coppergate represent material reworked from underlying Roman material. There is some residual Roman pottery which provides evidence of reworked deposits (this is discussed briefly by Kenward and Hall 1995, 511, 513).

As can be seen from Table 5, most of the Roman records of fig are from founded towns, military sites or 'Romanised' villas (this classification is explained in the Glossary). The only site which has been classified as 'native' is from a wooden cess-filled tank in St Anne's Lane, Nantwich 85 [1382]. This is dated by dendrochronology to a felling date of mid-summer AD 132 (three planks still had their sapwood intact) which suggests that it was very likely a military installation for the salt-working industry, although it would still be classified as native. Was this a 'Portaloo' of the 2nd century?

Bearing in mind that 94% of all fig records (for all periods) are for waterlogged or mineralised preservation, it is perhaps not surprising that there are no other records from 'native' sites. There are relatively few reports for contexts with a waterlogged preservation regime from native sites (10 sites excluding Nantwich i.e. 20% of all Roman sites with waterlogged preservation). All of these are either ditch or well fills, not particularly the types of deposit that are likely to contain faecal material in rural areas. Bearing in mind the way faecal material would have been disposed of in rural areas, at most periods, are we ever likely to find good evidence of Romano-British faeces?

7.2.3 Malus sylvestris

There are no certain records of Malus sp. from the Mesolithic period (merely two uncertain and poorly dated ones) but by the Neolithic there are enough records to suggest that crab apples were a common part of the diet.

It seems unlikely that archaeobotanical work will provide evidence of the first use of domesticated varieties of apple in Britain because it is not considered possible to distinguish the wild and cultivated forms on the basis of the seeds alone. It seems that an assumption has been made that the early records are crab apple and the medieval and later ones are more likely to be domesticated apple. Bakels (1991), for example, gives 'M. sylvestris' in her list of prehistoric finds, mentions 'Malus sp.' for the Roman period and lists 'M. domestica' for the medieval period, without any explanation. Schweingruber (1979) compared the measurements of a random sample of wild apples collected in Switzerland with finds of Neolithic apples from Lake Village sites. The modern fruit was charred to make it comparable with the archaeological material. He found that the fossil material was a similar size to the modern, concluding that nearly all the Neolithic material was of wild crab apple size. There have been occasional records of whole cores but these are rarely preserved in a state that can be measured. Some whole apples of 'crab' size have been found from the prehistoric period, notably, a cache of dried apples from Tankardstown South [1164] from the foundation trench of a Neolithic house. These apples were interpreted as a store of dried fruit, which had become charred when the building had burnt down (Monk 1988).

7.2.4 Pyrus sp.

It seems that it is difficult to distinguish the seeds of Malus and Pyrus with certainty, although some authors do so and certainly modern reference material is quite distinctily different. If endocarp is also preserved then the presence of apples is certain since pear lacks this horny layer around the seed. If stone cells are preserved together with pips then Pyrus or Cydonia are inferred (see above). Pyrus communis records comprise material from the late medieval Dominican Priory, Oxford 76-83 [1305] and two Roman sites - Chew Park Well 54 [1420] and Tiddington [1035] - but for none of these is there a published description of the seeds which provides a convincing argument for the presence of pear. Whether or not pears were cultivated in Britain in the Roman period (there is documentary evidence of them being part of the military diet), by Domesday old pear trees were noted as boundary markers so had probably already been cultivated for some time Roach (1985, 120).

7.2.5 Phoenix dactylifera

Date stones, being large and woody, are robust, suggesting that they would preserve well: the type of material which would be found from large-scale bulk sieving programmes on urban sites. There are only two records, however, one of charred stones, from early Roman Colchester, Lion Walk 71-4 [1522] and the other in the fill of a waterlogged garderobe pit from 14th century Hull, High St and Blackfriargate 73-7 [1938]. It is unlikely that in the past dates were imported into Britain de-stoned as they are sometimes today. The conclusion must be that dates were a particularly rare and exotic food in Britain in the past, even at periods when other fruits likely to have been traded in a dried form from the Mediterranean, Near East or N. Africa were regularly consumed.

7.2.6 Prunus spp.

Sloes (Prunus spinosa) stones are sometimes found within faecal concretions from Anglo-Scandinavian York, suggesting that they had been ingested. This seems surprising to modern tastes as sloes are highly astringent when raw; to quote Lang (1987) 'no one would eat sloes for pleasure'. Wiltshire (1995), however, has commented on ways in which their palatability may be improved. Sloe stones are occasionally preserved with the flesh (endocarp) still attached (Figure 4). Mineralised faecal concretions from York also produced occasional fragments of fruit epidermis, probably sloe or plum, which has a characteristic angular cell pattern (Figure 3) and occasional stomata.

The full range of sizes of 'plum' stones are found in archaeological deposits, from sloes (the smallest), up to the larger domesticated plums which come mostly from medieval deposits. Several workers have carried out length/breadth measurements in order to investigate morphological variations and to separate sloe, bullace, wild plum and the smaller domesticated varieties such as greengage (see, for example, Behre 1978; Moffett 1992, 280). It would be useful if all the measurements from different sites could be collated so that statistical analysis could be carried out on as large a data set as possible. Measurements of modern fruit stones are also needed; workers such as Moffett (1992, 278), for example, have only measured the archaeological material in order to 'compare any groupings with the author's subjective identifications'.

7.2.7 Vitis vinifera

There is a prehistoric record of a grape pip from a Neolithic context at Stepleton 81 [1378]. Not only is this single seed the only prehistoric evidence but it is also as early as any of the records of the domesticated grape in the rest of Europe. It has been radiocarbon accelerator-dated to the Neolithic with almost identical dates on associated cereal grains and charcoal (Jones and Legge 1987, 454), so there is no doubt about its provenance. Evidence of grapes is particularly common from Roman and medieval deposits but, like some of the other exotic taxa, such as Ficus carica, much rarer in the intervening period. There is no good archaeobotanical evidence yet which can suggest when grape was first cultivated in the British Isles, rather than being imported. Williams has discussed the macrofossil evidence for the Roman period but it is somewhat inconclusive (Williams 1977, 333).

7.3 Vegetables

7.3.1 Allium porrum

Since fragments of leek epidermis were first identified from faecal concretions from Anglo-Scandinavian Coppergate, York (Tomlinson 1991, 115) fragments, at least of Allium sp., have also been found from other Roman and medieval sites in York, from Roman Nantwich, and two sites in medieval Chester. There are several other, as yet unpublished, records. The epidermis of leek leaves is distinguished by a combination of characters. The cell are oblong and have thin, straight-sided walls, different from the more open 'bedspring' pattern of onion cells. The stomata have no subsidiary cells (anomocytic), and they are sunken in small rectangular pits, which show up more clearly in digested fragments. In fossil material these pits are typically seen in lateral view as 'pegs', using transmission microscopy (see Figure 5). There are two distinctive features of Allium porrum which can be seen in the fossil as well as the modern material. These are the cuticular sculpturing which runs down the centre of each cell and the row of rounded teeth which runs along the leaf margin, giving it a scabrid appearance when viewed with the naked eye. Only the green parts of the leaves have the stomata, sculpturing and marginal teeth. It would not be possible to identify the white parts of the leek 'stem'; fragments of epidermis without the marginal teeth can only be identified to genus (see Figure 5).

Seeds of leek have been found from Watergate St, Chester 85 [1017] in a pit filled with faecal material dating to the mid 13th century (Ward 1988). The three-sided seeds were wrinkled, but the surface had a very characteristic tuberculate/scabrate/rugose pattern. A single leek seed has been found at Sewer Lane, Hull 74 [1627] during re-examination by ARH (unpublished) of material originally reported by Williams.

7.3.2 Asparagus officinalis

The single record of a charred seed of asparagus is from Tibbet's Close, Alcester 83-4 [1302] and is dated to the late Roman period. Moffett (1988, 74) gives a possible taphonomic explanation for finding charred seeds of asparagus, namely the burning of asparagus beds in the autumn to stimulate growth. The seeds from Alcester came from a hearth and ditch fill deposits.

7.3.3 Beta vulgaris

Both chards and beetroots were well-established vegetables in classical times, but there are no pre-classical archaeological records anywhere in the Old World according to Zohary and Hopf (1994). The record of Beta from a well in Roman Lincolnshire is associated with very few other food plants and many weeds. The two Roman records from York are classified as weeds in the relevant report (Hall and Kenward 1990, 395) and they are from contexts of dumping and organic accumulations which perhaps suggest these are ruderals. The plant would not be expected to be growing here without some human intervention, as the distribution of the native sub-species (ssp. maritima) in Britain is definitely coastal (Perring and Walters 1962, 83). Triglochin maritima occurs in the same context groups as the York material of Beta vulgaris, suggesting some coastal influence in Roman York (Hall and Kenward 1990). This is perhaps from herbivore dung from animals which had been grazing on salt marsh before being brought into the town. It is less likely that such seeds were transported by the River Ouse, although this was tidal to York prior to the 18th century.

The post-Roman records for Beta vulgaris are only from Hull and York, which perhaps suggests that it is not being recognised elsewhere. Knörzer (1991) has reviewed the evidence for Beta vulgaris on the continent, where there are around 20 records (nearly half of these unpublished).

7.3.4 Brassica sp(p).

In the list of vegetables above, 'Brassica sp(p).' includes the following: B. napus (rape), B. rapa cf. ssp. rapa (turnip) and B. oleracea (cabbage). Brassica nigra is listed separately.

Only 14% of Brassica records are identified to species. The number of individual identifications in the database (excluding 'cf. Brassica') for all types of preservation mechanism for all the Brassica genus are given in Table 6. Those identified to species level are underlined. This list shows how difficult it is to give a simple answer to the question 'what evidence is there for brassicas in the archaeobotanical record?'. There are more identifications of seeds preserved by waterlogging than by charring, but it does seem to be possible to identify the charred seeds (20% are identifications of charred material).

An identification atlas published by Berggren in 1981 covers seed identification of the Cruciferae (see also Berggren 1960, cited in Berggren 1981, which deals with Brassica and Sinapis). The only distinguishing characters are the reticulate patterns on the seed surface (testa); there are no significant differences in seed size or shape which are appropriate for the identification of fossil material. This means that unless the seeds are particularly well preserved the surface patterns are very difficult to separate. Berggren's key was, of course, not designed for fossil seeds and therefore uses characters such as seed colour which are not at all appropriate in a fossil context. It is likely that better progress would be made if a scanning electron microscope (SEM) was used. Butler (1991) has shown how the variable surface characteristics of the testa of species of Lathyrus, Lens and Vicia can be clearly seen using SEM work and can be used for identification purposes. Similarly, Fraser (1981) studied Brassica using SEM. Unfortunately, the latter work has not been published except as a thesis.

There is one Iron Age record of Brassica rapa cf. ssp. rapa (turnip) from the site of Bu, Orkney [1174] but otherwise the subspecies have not been identified (this is also the only prehistoric record of the species). The other identifications may only represent the wild, weedy subspecies. Although it seems unlikely that the vegetative remains of turnip will be preserved on British sites, Hather et al. (1992) have identified parenchymatous tissue of turnip from Byzantine (Early Christian) Sparta.

There are no certain identifications of Brassica oleracea (wild cabbage). There is an uncertain ('cf.') identification from the Roman period from Church St, York [1628]. It is worth noting here an error in this particular report which is one which quite often occurs in archaeobotanical reports. The table of results (Greig 1976, 25) has the name 'cf. Brassica oleracea' listed, but in the list of 'plants useful to man' on the opposite page it is given as Brassica oleracea, giving the implication that this was a 'certain' identification.

Leafy cabbages were well established as garden plants in Roman times. Heading varieties, possibly kohlrabi, were described by writers in ancient Rome (Thompson 1976, 50). Wild cabbages thrive in the Mediterranean basin and are likely to have been domesticated in this area (Snogerup 1980, cited by Zohary and Hopf 1994).

Some recent work examining the biochemistry of potsherds from archaeological excavations claims to be able to identify lipids from the cuticular surface of cabbage leaves which had been cooked in the pottery vessels (Evershed et al. 1992). The work has concentrated on a large number of potsherds from the late Saxon/medieval excavations at Raunds in Northamptonshire. The techniques being used for the analysis are gas chromatography and mass spectrometry. The lipids are extracted using solvents from the powdered body of the potsherds. These researchers have also examined the movements of lipid substances through the pots and concentrations in different parts of the vessel (Charters et al. 1993) to show that their interpretations of cooking residues are valid. Unfortunately, although their claims to have identified Brassica oleracea are convincing from the biochemical results, they appear to have only looked at the wax component of a range of modern Brassica oleracea varieties and a modern cultivar of Brassica rapa (Evershed et al. 1991). It would perhaps be more appropriate for them to broaden their examination of comparative material to some of the wild brassicas and other members of the Cruciferae before claiming they have identified Brassica oleracea. A circular argument is being used: they state that they are only looking at varieties which have been shown from documentary evidence (my italics) to be available in the late Saxon period.

7.3.5 Cucumis sativus

The cucumber is thought to have been taken into cultivation in northern India (Jeffrey 1980, cited by Zohary and Hopf). The single record from Roman Britain is slender evidence from the archaeological record, but, of course, cucumber is documented to have been introduced at that period (Wilson 1973, 326) when it was grown in villa gardens 'along with other vegetables such as marrows, garden asparagus and cardoons' (this last a form of globe artichoke). The post-medieval record is from 18th century Queen Anne House, Shrewsbury [data not yet added] but it is unfortunately not identified to species level.

7.3.6 Cucurbita pepo

One of the few indications from the archaeobotanical record of introductions from North America is a single seed and a few fragments from mid-17th century Dudley Castle [1934] of the pumpkin, marrow, or squash. There are, however, some other unpublished records mentioned by Moffett (1992). The relatively low numbers of samples examined from post-medieval waterlogged contexts is partly a function of the widespread disregard for this period by excavating archaeologists. Davey (1987, 78) has pointed to the missed opportunity (caused by the truncation of archaeological sequences for arbitrary reasons) to study the continuity of environmental remains which could provide evidence for food consumption at least up to the 19th century when urban rubbish disposal methods changed. The few post-medieval latrines which have been properly sampled have produced interesting results, notably those at Dudley Castle [1934] and Tudor Merchant's House, Tenby [1363]. There are published botanical reports from seventeen post-medieval sites in Britain where the preservation mechanism is waterlogged or mineralised but only half of these are samples from latrines or cess pits.

7.3.7 Daucus carota

Only the seeds of Daucus carota have been found in archaeobotanical samples. Direct evidence for the consumption of carrots is unlikely to be found because roots and tubers are not often found preserved on British sites and charring seems the only mechanism likely to operate in this case, although it is also likely that root and tuber fragments are being overlooked; the potential for identifying such material has been emphasised by Hather (1991; 1993).

7.4 Pulses

Seeds of legumes are undoubtedly under-represented in the archaeological record. They are not generally found preserved by anoxic waterlogging, but are more often found in a mineralised condition or as charred material. In waterlogged or mineralised material often only the hilum, sometimes with part of the testa, is preserved (see Figure 6), but as the size and shape of the hilum is diagnostic this is useful.

7.4.1 Pisum sativum

There is one middle Bronze Age record of pea from Grimes Graves 71-2 [1625] and one certain identification from Iron Age Hengistbury Head 70-85 [1389], but the other prehistoric records are only uncertain identifications.

7.4.2 Vicia faba

There are prehistoric records of the field bean as early as the Bronze Age. Note that all the archaeological records, including the post-medieval ones are of the relatively small, rounded 'Celtic' or horse bean. Broad bean Vicia faba var major, the large-seeded type, is not known to have existed anywhere until about AD 500 in China (Bond 1976, 181). It is not clear when broad bean first arrived in Britain. The use of the name broad bean in this context (e.g. in Zohary and Hopf 1994) is liable to be confusing. There are relatively more records of field bean from the 'SE' period. It is possible that this is a factor of preservation but it may be that these rather unappetizing beans were not favoured in the cuisine of Roman and 'post-Conquest' medieval people.

Figure 6 shows the mineralised hilum of a field bean from Coppergate 76-81 [1920]; the overall size and the length to breadth ratio distinguishes it from other legume hila. Some waterlogged fragments, also from Coppergate, have been identified as field bean pods. The microphotographs (see Figures 7-9) show both the modern and fossil cell patterns with the characteristic short, blunt, bulging hairs.

7.5 Herbs and flavourings

7.5.1 Aframomum sp.

Material identifed as Aframomum (which includes two spices, A. melegueta (melegueta pepper) and A. Granum Paradisi (grains of paradise) has been identified by Greig (in press) from medieval Worcester [1601]; it had previously been published as cf. Borago officinalis. Greig has also recorded Aframomum from post-medieval Taunton and Shrewsbury.

7.5.2 Allium sativum

A rare find of five whole garlic bulblets (cloves), which were preserved by partial charring in a waterlogged, ashy, context came from Eastgate, Beverley 83-6 [1926], and dated to the late 13th century. The cloves were found in the floor deposit of a timber building. There were a few hazelnut shells but no other biological remains were recorded from this context. The cloves were very well preserved, but started to deteriorate when they began to dry out. They were therefore preserved in polyethylene glycol (PEG) which did prevent further fragmentation. Because they were only partly charred, the cell pattern of the outer scales of the bulblets could be examined under the transmission microscope and compared with modern reference material.

There are two other certain identifications of whole garlic cloves, both from Germany. One is from Neuss in (Knörzer 1975) and the other from Laufen (Karg 1991). Murphy (1985) describes the find of a single 'charred clove of garlic (Allium sativum) or another species of the Liliaceae' identified by Gordon Hillman from a context broadly dated to the 16th century at Alms Lane, Norwich 76 [1446].

7.5.3 Coriandrum sativum

There are numerous records of coriander in the database from the Roman period (from 20 sites), but none earlier. The earliest post-Roman records are from Anglo-Scandinavian York, from Coppergate 76-81 [1920], where it was rather more common in the earliest period (mid 9th to late 9th/early 10th century) than later in the Anglo-Scandinavian levels (Kenward and Hall 1995, fig. 192(b)). It is relatively easy to identify, even when poorly preserved or in fragmentary form in faeces (Dickson 1989, 153), but perhaps it is only preserved under very good conditions such as were found in many deposits at Coppergate. With the exception of one mineralised find, all the records of Coriandrum sativum are of waterlogged seeds or mericarp fragments.

It was therefore surprising to find that some of the general literature on food plants in earlier periods (e.g. Renfrew (1985, 23); Wilson (1973, 192)) suggests clearly that coriander was present in Britain in the prehistoric period. It appears that this is all based on the evidence from a single seed from a 'late Bronze Age hut' site excavated at Minnis Bay, Kent, in 1941, noted by Godwin (1975, 223) [1667]. It is clear, however, that the dating evidence for this site is extremely tenuous. There is no published excavation report and no specific dating evidence is provided in the botanical report (Conolly 1941). The site seems to have been located on the foreshore. Three samples were examined, the one containing the Coriandrum seed was from 'a layer of peaty vegetable matter from the surface of hut site no. 11'. The material was obviously derived from a mixed assemblage. Some of the wood fragments were worn and rounded, suggesting coastal erosion, whereas the seeds (including the coriander) were 'in an excellent state of preservation'. There were also 'numerous fragments of Bryozoa and Hydroids', indicating a mixture of marine materials. There is, therefore, no reason to believe that this material was necessarily of Bronze Age date. The coriander could just as well be medieval, or possibly, modern. It was therefore recorded in the ABCD as 'undated'. Subsequent discussion with the author (Ann Conolly) has revealed that this material was, indeed, modern contamination, since another seed, not identified at the time, has now been determined as Lycopersicon esculentum (tomato). One possible source of this contamination is that the seeds came from a ship wrecked on the Goodwin Sands in 1936 (Lousley 1953, 143) but it is perhaps more likely that it came from raw sewage from one of the nearby towns such as Margate.

The point to be made here is not the problem of contamination, nor the difficulties pertaining to the publication of botanical material in the absence of an archaeological report, but the worrying way in which the identification of one seed from an insecurely dated deposit has, apparently, become evidence for the use of this plant in prehistoric Britain.

7.5.4 Humulus lupulus

Green (1981) noted that hops were not recorded from material earlier than that of the very late Saxon period from Wessex and the present analysis seems to support this. Perhaps hops were imported rather than being collected locally although it is a common hedgerow and fenland plant in England (Stace 1991). There is some evidence for the transportation of hops from the cargo in the Anglo-Saxon Graveney Boat 70 [1009]. Records of hops are very common throughout the Anglo-Scandinavian period at Coppergate 76-81 [1920]. Murphy (Ayers and Murphy 1983, 42) mentions the possibility of water dispersal of hop seeds at the waterfront site of Whitefriars Norwich 73 [1447]. What is without doubt is that hops were in use (for whatever purpose) in Britain before the 16th century, the date they are traditionally thought to have first been imported for brewing (for a discussion of the early history of hops in Britain see Wilson 1975).

7.5.5 Theobroma cacao

The one record of a single cocoa bean is from the cellar of an 18th century house by the quay in Dublin (Geraghty in press). Records such as this are important in substantiating the information from documentary sources concerning the introduction and early use of species from the New World.

7.6 Oil plants

7.6.1 Camelina sativa

Interestingly, the only records for gold-of-pleasure from Britain are both from London; two from the Roman period (from Southwark) and two Saxon (from the City). It is possible that the plant was imported into London at both periods. Schultze-Motel (1979, 277) suggests that it was probably first cultivated in the late Bronze Age in Europe, having been spread initially as a weed of flax. Its cultivation increased in the Iron Age, especially around the North Sea and Baltic. It was cultivated fairly widely in the medieval period in Europe, but there is no evidence of its cultivation in Britain at any time.

7.6.2 Cannabis sativa

All the Roman records of hemp 'seeds' are from York; despite records of pollen attributed to the Iron Age (cf. Bradshaw et al. 1981), there are no prehistoric finds of the seeds. Records from the post-Roman period are almost invariably from urban sites - hemp is quite likely to have become a ruderal in the vicinity of habitation sites, as it is today, and it is notable that most of the records are for very small numbers of remains (the large numbers entered in the ABCD for two lists from Anglo-Scandinavian Coppergate [1920] are composite values for combined data from many samples).

7.6.3 Linum usitatissimum

The seeds of flax are quite common in food plant assemblages from the Roman period onwards. The figures shown above do not include records of Linum usitatissimum fibres or the remains of scutching waste, or shives, which are occasionally found (Kenward and Hall 1995, 773). Whole or fragmentary capsules are often recorded. The seeds have also been found, evidently chewed (and characteristically holed in the middle of the flat face), in faecal deposits and concretions (see Figure 10).

7.6.4 Olea europaea

There are several records of olive stones in the Roman period but none later, even in the post-medieval period. There is a similar pattern in western continental Europe, although there is one 18th century record from Amsterdam (Paap 1983). The lack of post-medieval records in Britain perhaps reflects the under-sampling of deposits of this period.

7.6.5 Papaver somniferum

As mentioned earlier in this paper, the occurrence of a particular plant in an archaeological context does not necessarily indicate its use. If it is not a native British species then its presence on an archaeological site at the very least indicates its introduction by human agency which may not necessarily be deliberate.

Opium poppy seeds have been found in central Europe from the middle and late Neolithic (Zohary and Hopf 1994) but, because they have not been found this early in south-eastern Europe and the Near East, it has been concluded that domestication was from the western Mediterranean area where the wild poppy is distributed (Bakels 1982). Opium poppy is grown for its oil-rich seeds which are used directly as food, or the oil can be extracted for food or industrial purposes (Duke 1973). It is also well known for its narcotic and medicinal properties. Opium is obtained from the latex which is released by the plants by gashing their unripe capsules, but not in the British climate.

From the British Isles there is one good late Bronze Age record, from Wilsford Shaft [1331], and two from the Iron Age [1392], [1579]. Robinson (1990) suggests that the Watkins Farm opium poppy [1392] represents a disturbed ground plant rather than a crop. This is borne out by the fact that the seeds occurred in a waterlogged assemblage, not with the charred cereal grain assemblage. It seems rather surprising that there are so few prehistoric records of this plant in Britain, especially as it is easily preserved by charring.

Green (1981, 143) states clearly that opium poppy 'has been recorded from a wide range of sites from the late Iron Age onwards (in Wessex)' but many of his results are based on unpublished work and it is not clear to which Iron Age sites he is referring.

7.7 Cereals

Cereals are by far the most frequently recorded food plant remains on archaeological sites. There are several reasons for this. They survive particularly well when charred, and in fact survive so well in archaeological sediments that they may be redistributed into deposits of a later, or earlier, date (e.g. van der Veen 1992, 71). It has been suggested that cereals are more likely to come into contact with fire than other food plants, partly because of the need to parch the hulled varieties if pounding or milling is not sufficient. They will often have been stored in buildings or granaries which were susceptible to fire. There are also several other taphonomic routes by which charred cereals become incorporated into archaeological deposits (e.g. from straw used for thatch, animal bedding and from burnt animal dung used for fuel).

The distribution of cereals in prehistoric and Roman times has been well discussed by previous authors since Helbæk's work in the 1940s (Helbæk 1952). The early wheat species (einkorn, emmer and spelt) were much rarer by the Anglo-Saxon period, but records of bread wheat, particularly the compact form, and of barley, rye and oats are common throughout the post-Roman period. It seems that the free-threshing wheats (Triticum aestivum) may be relatively under-represented in all periods relative to the hulled wheats (TT. monococcum, dicoccon, and spelta) because their diagnostic parts (glumes and spikelet forks) are more likely to have become charred. There are over two hundred records of grains identified as 'Triticum sp.'. The cereal bran which often forms the matrix of mineralised faecal material is arguably much better direct evidence of the use of cereals as human food (Hall et al. 1983) but it has so far only been identified to 'wheat/rye'.

Records of Secale cereale are rare in prehistoric Britain and occasional from the Roman period onwards. The only well-dated Bronze Age record is from Runnymede [1638], where the rye amounted to less than 3% of the total number of cereal grains recovered. It was thus interpreted as a weed of the wheat and barley crops (Greig 1991, 254). Rye is, relatively, much more common in the medieval period.

The rare, uncertain, identifications of Triticum turgidum and Triticum durum are rather tantalising. It is not clear how distinctive these species are from the other free-threshing wheat; further work is clearly needed. Moffett (1991) has discussed the growing archaeobotanical evidence for rivet wheat from British medieval contexts.

Cultivated oats (Avena sativa) are difficult to distinguish from other non-crop species unless the diagnostic features of the floret base are preserved. It is therefore likely to be under-represented as a species in the record as it is rather rare to find the floret intact. The earliest certain cultivated oat grain identifications are dated to the Iron Age from southern England. In waterlogged assemblages, such as from faecal material, Avena sp. bran is easily distinguishable from the 'wheat/rye' type bran (Hall et al. 1983; Dickson 1989, 138 and plate 1a). Only 18% of Avena records in the database are identified to species, but half of these are the cultivated species Avena sativa.

The two records of Panicum miliaceum, one Roman and one medieval are highly doubtful identifications. There is also a single, extremely doubtful, 1918, record of cf. Sorghum vulgare from Romano-British Kenchester [1764]. Early records, and those where the identifications are not substantiated with descriptions of the seeds or illustrations, should be accepted only with caution. It is easier to point to erroneous identifications when there are only one or two of them and they are, therefore, likely to be anomalous. It is far more difficult to spot inaccurate identifications when the appearance of a record of a particular species is not surprising. When a new taxon, which has not been identified before, is found, a full description with illustrations is essential. It is also important to provide details of the storage location of the fossil material. As mentioned above in passing, material from Sewer Lane, Hull [1627] has been re-examined and one grain of millet was found but may well have been modern contamination as it was very fresh in appearance. The only other record, also uncertain, is from St Thomas Street, Southwark [1653].

Several taxa which occur fairly frequently in medieval deposits in western continental Europe are notable for their rarity or complete absence in Britain. Panicum miliaceum occurs regularly from the Bronze Age onwards on the Continent (Bakels 1991) but, as shown above, there is certainly no evidence of it as a crop in Britain. Fagopyron esculentum, buckwheat, is another example. It occurs quite regularly in medieval and later sites in the Low Countries. The earliest macrofossil record, from the site of Dommelen (Bakels 1991, 293), is dated to the mid 12th century on the basis of the pottery associated with it, but up to two hundred years earlier on the basis of the 14C-dated charred grain found in the same context. There are several early pollen records, too. Documentary evidence suggests that buckwheat was not introduced into Europe until the early 15th century, arriving in Germany first (Campbell 1976). This only emphasises the tenuous nature of such evidence. From Britain there is an uncertain identification from Roman Newstead [1138] and Greig (unpublished) has found capsule valves from 18th century deposits at Shrewsbury [data not yet added]. The husks are a by-product of buckwheat which was used as packaging material in the past (Howes 1974). An example of such use is to be found amongst the plant remains from the Monte Cristi shipwreck excavations. This post-medieval English or Dutch trading ship wrecked off the Caribbean coast was carrying a cargo which included crate(s) of clay pipes. The packaging material used for these pipes has been identified as buckwheat husks (Tomlinson, unpublished report to the Monte Cristi Shipwreck Project), the 18th century equivalent of polystyrene 'chips'. There are two other, similar, records of buckwheat from packaging in ships' cargoes found on underwater wreck sites (van Rooij and Gawronski nd; see also Pals 1987, 74).

7.8 Nuts

After cereal grains, hazel nuts are probably the commonest food plant to be found on archaeological sites of all periods. Not all these records of hazel nut shells are likely to be direct evidence of human consumption, however. For example, a hoard of nut shells from underneath a bed in one of the Viking houses at Fishamble Street, Dublin (Geraghty in press) were all clearly mouse gnawed although they might originally have been intended for human consumption.

7.8.1 Pinus pinea

The stone pine was brought to Britain in the Roman period, not only for food but also for use apparently in ritual ceremonies. Pine nuts were held to be sacred by the ancient Greeks (Howes 1948, 165). One of the records, which include kernels as well as cones, is from a religious site at Carrawburgh [1198]. There is also a 12th century record from Queen Anne House, Shrewsbury [data not yet added] and an exceptionally well-preserved cone and cone bracts from 13th century levels at Wood Quay, Dublin (Geraghty 1993). As most records are of 'spot' finds, hand collected by the excavators, it is likely that this type of material is comparatively over-represented in the archaeobotanical record, especially as the cones are large and easy to recognise.

© Internet Archaeology

Last updated: Tue Aug 27 1996