4.2 Farming and Feeding

This section represents surveys published of Iron Age sites within Scotland and synthesises some methodological papers examining specific aspects of Iron Age diet or food production. There are several issues in terms of the completeness of the data set for biological remains within Scotland. The first of these is related to the age of excavation. Although animal bones and marine molluscs have long been of some interest to excavators (MacNaughton 1891 and 1893), these have often been collected in a sporadic and selective manner (e.g. MacGregor 1974) and published reports often include little more than lists of species present (e.g. Graham 1968). For plant remains and bones of smaller animals (particularly small fish) this problem persisted until sieving of sediments became routine on excavations (Wheeler and Jones 1976); this has only really occurred in the last 20-30 years. The second major problem is the almost complete lack of bone preservation in most of the Scottish mainland. Barring a couple of sites in the south east (e.g. Barnetson 1982; Cussans et al. in prep), crannog excavations (mostly antiquarian) and parts of Caithness almost all excavation reports note only the presence of tiny unidentifiable fragments of burnt bone, all other traces having been lost to the acid soil.

However, areas of the Atlantic zone offer excellent preservation conditions, and a number of large scale excavations have employed routine sampling, sieving and flotation programmes to produce some excellent economic data for Iron Age Scotland. These dominate the following picture.

Whalebone mattock from Foshigarry, North Uist © NMS

Production of plants and animals

The suite of domestic mammal species found on Iron Age sites appears to be fairly uniform throughout Scotland. Cattle and sheep are almost always the dominant species with pigs playing a more minor role ; the unusual site of High Pasture Cave, with a dominance of pigs within the cave, is a notable exception to this (Drew 2006). The status of goats is less certain as they are often difficult to distinguish from sheep and are therefore probably underrepresented, (Noddle 1977, 2000) although they do not appear to have been present in Shetland or the Western Isles at this time (Platt 1956, Cussans and Bond 2010 and 2015 Mulville 1999; Smith and Mulville 2004; McCormick 2006). Horses may also have had occasional food value (Smith 1994, Cussans and Bond 2010). At a few sites domestic fowl and/or goose seem to have been present, particularly in later phases (Platt 1948; Bramwell 1977; Allison 1997; O‘Sullivan 1998c; Cartledge and Grimbly 1999).

In terms of cultivated crops hulled, six-row barley (Hordeum vulgare var. vulgare) appears to be the dominant crop across Scotland during the Iron Age. Although there is still some presence of naked barley (Hordeum vulgare var. nudum) this declines over time (a pattern that began before the Iron Age). Although some authors have likened this crop to Scottish bere (MacGregor 1974; Dickson and Dickson 1984; Boyd 1998), Bond (2003) noted a crop made up of barley of varying hulledness and grain morphology and interpreted this as a mixed ‘landrace‘ of barley. Oats (Avena spp.) have a low but persistent presence in Iron Age assemblages but often these cannot be determined to species and many may be simply weeds of the barley crop. At some sites cultivated oats (A. sativa/strigosa) have been positively identified. At many Scottish sites wheat (Triticum sp.), usually emmer wheat (Triticum dicoccum), also played a minor role. Flax (Linum usitatissimum) seeds have also occasionally been identified, and may have been used for food or oil; alternatively the crop may have been grown for the fibres. A possible identification of rye (Secale cereale) has been made at Dun Vulan (Smith 1999).

Methods of rearing animals and cultivating crops are more difficult to determine than the presence of the species themselves. In terms of animal husbandry there is limited evidence for the presence of foddering practices and possibly for stalling/penning of animals. No reference to pathologies related to stalling of animals could be found, but structural evidence has been suggested on several sites.  Woodend Farm, Johnstonebridge, Dumfries and Galloway (Banks 2000), high phosphate levels were associated with the interior of one of the posthole defined structures (Duncan 2000). Other possible provision of animal housing has been found at Dun Vulan (Parker Pearson and Sharples 1999) and at the wheelhouse sites of Allasdale and Clettraval (Young 1953; Scott 1948),where structures within the enclosure were identified as byres. The role of ring-ditches in timber roundhouses has seen some debate (summarised in Dunwell 2007, 104). They are a long-lived and varied phenomenon (theme 5.3), with some little more than shallow erosion scoops and others deep, deliberate features. One plausible formation process for them stems from use as byres, perhaps for over-wintering animals, with the scoop caused by clearing out accumulated midden material for use as fertiliser (see also discussion of soil improvements below).

The general lack of evidence for boundaries in the landscape has been discussed above; there could have been temporary boundaries or features such as hedges which are hard to identify, but taken at face value it suggests that flocks were closely tended rather than enclosed.

In terms of the provision of fodder the consumption of seaweed during the winter by sheep at the Iron Age site of Mine Howe in Orkney has been identified through the analysis of ?13C and ?18O values (Balasse et al. 2009). Dental microwear analysis (e.g. Mainland 2000) also has the potential for the identification of foddering practices. The introduction of oats into the Northern Isles during the Iron Age is likely to have made extra storable fodder available for cattle and horses (Bond 2003). Cussans (2006) linked this apparent agricultural intensification to an increase in sheep size during the Atlantic late Iron Age at Old Scatness, Shetland.

Selection of stone agricultural tools from the Northern Isles, © NMS

Evidence for the cultivation of crops comes from several strands. Evidence for fields has been discussed above, the evidence suggesting a shifting patchwork of cultivated areas. Analysis of soils through thin-section micromorphology, phosphate levels etc. can give clues as to anthropogenic soil amendments. Evidence for such amendments have been found in the Northern Isles where it is suggested that manure (Guttmann et al. 2006) and midden material (Guttmann et al. 2003) were spread onto the fields. It has also been suggested that midden heaps themselves were spread out and used for the cultivation of crops (Guttmann et al. 2004). Harvesting methods inferred from weed ecologies appear to vary from site to site; in some areas crops appear to have been harvested by uprooting (Boardman 1998; Smith 1999; Church 2000; Church and Cressey 2006) and in others by reaping either high up the stem (Milles 2000) or low down (Dickson 1994, 1999; Bond 2007b). Artefactual assemblages provide further evidence, such as ard points and sickles, (e.g. Rees 1983; Armit 1991:191; Wilson 1980), although their rarity as finds means they can provide only the most general patterns. They show diachronic variation: the sickle, for instance, is a later Iron Age phenomenon, with the reaping hook (which lacks the backward curve of the blade) a more multi-functional tool which preceded it and continued in use alongside it (Rees 1979, 450-461). There is no evidence that the scythe, introduced by the Roman army, was adopted locally. Ard points in particular show regional variation according to the availability of raw materials, with stone tips in the Bronze and Iron Age of the Northern Isles, whalebone tips in the Western Isles, and wood (sometimes tipped in iron) elsewhere (Fenton 1963; Rees 1979, 7-61). There is no evidence of plough-pebbles, representing the use of the mouldboard plough, until the post-Roman period (Fenton 1963, 276-9, Hill and Kucharski 1990). The presence of ard marks in excavated layers gives direct evidence for which areas were under cultivation (e.g. Barclay 1985; Fowler 1983, 113-7, 150-6), although these vestigial traces are not always noted in excavation. Some of these may represent the use of a heavy ard to break up ground which had laid fallow, rather than regular cultivation, as they do not seem to represent repeated ploughings (e.g. Dockrill 2007, 58-61).

Procurement of wild resources

Although Scottish Iron Age economies were clearly dominated by domestic plants and animals, hunting, gathering, fishing and fowling all had significant if minor roles in the diets of prehistoric people and would have offered welcome diversity in the diet, a variety of additional minerals and nutrients and a fallback in times of agricultural failure. The key wild mammals exploited were red deer (Cervus elaphus), seals and whales; hare (Lepus capensis) and roe deer (Capreolus capreolus) were also occasionally exploited in some areas.

The relative importance of red deer in the diet can be difficult to assess as in some cases fragments of antler (which may come from shed antlers) and bone are counted together. Studies of antler generally suggest that shed rather than butchered material was preferred, although both were used (e.g. Hallén 1994, 197;  Smith 1994, 145; Hunter 2006c, 138; Hunter et al. forthcoming). This implies a good knowledge of deer ecology, and particularly of times and places in the landscape where antlers were likely to be shed. It is notable that, among the rare representational art of the period, deer are the most common animal represented, suggesting they were seen to have a special significance (e.g. MacGregor 1976, nos 327, 329-331). There are indications from the Western Isles of the exploitation of deer varying on different islands, and suggestions that deer were seen as rather different from other animals (Mulville & Thoms 2005, 239-242).

Seals are likely to have been exploited for their meat, blubber and skins and may have been hunted or opportunistically scavenged. It is difficult to say if whales were exploited as a food source but they were certainly valued for their bone and probably their blubber. There has been much debate as to whether the whale material found on Iron Age settlements came from carcasses of animals that had become beached rather than being actively hunted. MacGregor (1974) suggests that the low number of whale strandings recorded in Orkney in recent times (nine in 60 years) would not have been sufficient to allow the development of specific artefact types and that therefore some whales must have been purposefully hunted but this does not take account of the large quantity of material represented by a single stranding nor the changing world population of whales since the first millennium BC. An alternative model is provided by ethnographic evidence models of whaling inof Scotland in the pre-modern period, which wasere essentially opportunistic: when whales were spotted close in-shore, boats were mobilised and a ruckus was created in the hope of driving them onto the shore (Baldwin 2008). Mulville (2002), in reviewing Iron Age evidence of cetaceans, notes that many aspect of cetacean use remain unclear, and highlights the potential of DNA analysis as a way of identifying the often heavily-worked and thus undiagnostic fragments which survive. A clearer knowledge of the range of species represented, and a study of their chronology and quantity on specific sites, would allow a much better-informed discussion of procurement strategies.

Sample of fish remains from a context recovered at the Iron Age settlement of Bostadh Beach, Lewis, Western Isles. ©Ceron-Carrasco 2005.

The extent and nature of fishing activities are difficult to assess on earlier excavations due to the lack of systematic sampling and sieving (e.g. MacCartney 1984; Finlay 1991; Nicholson 1997); however, cod family (Gadidae) fish, in particular saithe (Pollachius virens), pollack (Pollachius pollachius) and cod (Gadus morhua), do seem to have been preferred (Platt 1956; Wheeler 1977; Colley 1983; Macartney 1984; Seller 1986, 1989; Finlay 1991, 1996; Locker 1994; Cerón-Carrasco 1995, 1998a, 1998b, 2005, 2006a; Cerón-Carrasco and Parker Pearson 1999; Nicholson 2007a, 2007b, 2010a, forthcoming). Fishing technologies seem to have varied from site to site; the majority of sites show a concentration on inshore or coastal fishing, whereas at a few sites offshore fishing seems to have taken place, particularly in the early Medieval period (Atlantic Late Iron Age; Late Iron Age (Colley 1983; Sharples 1984; Seller 1989; Cerón-Carrasco 1995, 1998a, 1998b; Finlay 1996; Russ et al. in prep).

Artefactual evidence of fishing has never been synthesised. Hooks are rare, but some stone weights and pumice floats could be from fishing nets while bone tool assemblages include examples of what may be netting needles (e.g. Benton 1931, 184, fig 7). Isotopic evidence is so far limited, but a study in East Lothian indicated that marine resources played little if any recognisable role in the diet (Jay and Richards 2007, 182).

The exploitation of birds seems to have been a minor but persistent activity. Bird bones are present at a large number of Scottish Iron Age sites and although the abundance of different species varies from site to site those that are most frequent tend to belong to sea birds, for example gannet (Sula bassana), cormorant (Phalacrocorax carbo), shag (Phalacrocorax aristotelis), great auk (Alca impennis) and gulls (Larus spp.)

It seems most likely that the majority of these birds were exploited during the spring and summer months during the nesting season, when they would be relatively easy to catch (Bramwell 1977; Allison 1997; Serjeantson 2007a; Nicholson 2010b). Some birds which would be autumn/winter visitors have also been identified (Bramwell 1977; Allison 1997; Hamilton-Dyer 2006). Although there is little direct evidence, ethnographic parallels suggest that  birds’ eggs would also have been exploited (Hedges 1983, 118; Nicholson 2010b).

Gathered foods include shellfish and wild plants. Although a variety of shellfish species are present on Scottish Iron Age sites two were heavily exploited, the common limpet (Patella vulgata) and the periwinkle (Littorina littorea). However, despite the large numbers of shells found on some sites, calculations of meat weight and consequent likely contribution to the human diet are low (Evans and Spencer 1977; Barlow 1984; Howard 1996). There are also no firm conclusions as to whether these shellfish formed a regular, albeit minor, part of the human diet, if they were used as fishing bait or if they fulfilled both roles (Colley 1983; Evans 2005; Cerón-Carrasco 2005, 2006b; Cussans 2010). Other edible species present in small numbers include mussels (Mytilus edulis), oysters (Ostrea eulis), common cockle (Cerastoderma edule) and razor clams (Ensis sp.). All of the shellfish species regularly exploited are common on rocky or sandy shores and could be easily gathered in the inter-tidal zone. On a small number of sites remains of edible crab (Cancer pagurus) have also been identified; these creatures spend part of their life cycle out at sea in deep water but come close to the shore in the spring and summer months and so were probably only caught at this time of year (Evans 1983).

Many sites produce wild plants may also have been collected for food use. The most common (probably due to their robust nutshells) are hazel nuts (Corylus avellana). The crowberry (Empetrum nigrum) is also present at a few sites; however, this may represent the collection of plant material for thatch or bedding. Other plants were less certainly purposefully gathered even though they have edible parts, as many of them also occur as weeds of cultivated ground and may have been accidentally gathered along with the cereal crops. These include wild radish (Raphanus raphanistrum), Brassicas (Brassica spp.), fat hen (Chenopodium album) and corn spurrey (Spergula arvensis).

Processing of plants and animals for consumption

Processing of plants and animals prior to consumption often leave traces and therefore give clues as to food preparation methods. For animals these are most commonly signs of butchery; the majority of bones of food-forming mammals found on Scottish Iron Ages sites are incomplete and are frequently quoted as being broken to gain access to marrow (MacNaughton 1891, 1893; MacGregor 1974; Seller 1982, 1986 and 1989; MacCartney 1984; Finlay 1991; McCormick 1998; O‘Sullivan 1998a; Mulville 1999; Serjeantson et al. 2005). Cattle bones tend to be more fragmented due to their larger size and the probable need to break them down into more manageable portions; large blade chop marks also seem to be more common on cattle bones (McCormick 1998; Cussans and Bond 2010, 2015, again probably due to their large size. Knife marks have been noted on cattle, sheep, pig, horse, deer and seal bones and in some cases (particularly cattle and sheep) regular butchery patterns have been noted (Macartney 1984; Finlay 1991; Smith 1994; Cussans and Bond 2010, 2015). Butchery evidence is also present on some fish and bird bones (Colley 1983; Smith 1994; Allison 1997; Cerón-Carrasco 1998c; Cerón-Carrasco and Parker Pearson 1999; Cartledge and Grimbly 1999; Hamilton-Dyer 2006; Nicholson 2010a, 2010b). Some evidence for the preservation of meat and fish through drying or smoking has also been found (Smith 1994; Nicholson 2004, 2010a, 160, 2010b, 169; Cussans and Bond 2015) there is no evidence so far for the processing of salt (in the form  of briquetage containers) for meat preservation.

There has been little recognition (or study) so far of variations in processing practice in different areas, periods or site types. Nor has the distribution of body parts present (or the range of macroplant remains present) yet provoked extended discussions over producer and consumer sites, although differences have been noted among some site assemblages (eg Mulville & Thoms 2005, 238-242).

Long before the routine recovery of plant remains from archaeological sites the presence of saddle and rotary querns attested to the on-site processing of cereal crops. Querns are common finds at Scottish Iron Age sites (Armit 1991; McLaren and Hunter 2008), the more primitive saddle quern being replaced by the rotary quern (Caulfield 1978; Armit 1991), although the exact chronology and pattern of the change over in forms is yet to be fully understood (McLaren and Hunter 2008; see also themes 4.5 and 9.3).

MacKie (1971, 1987) has noted a basic functional difference  between the adjustable querns of the Atlantic zone, where the coarseness of the grind can be adjusted, and the fixed bun and beehive querns of southern Scotland. This ability to readily modify the coarseness or fineness of the grinding would make it easier to make (e.g.) coarse-ground grain for stews versus finer-ground meal for bread or bannocks.

Many of the charred plant assemblages examined that contain principally chaff and weed seeds are likely to represent crop processing residue (Dickson and Dickson 1984; Dickson 1994; Smith 1999; Milles 2000), but remains from many sites are too small or poorly preserved to determine such practices (Boardman 1995a, 1995b; Fairweather 1996). Some deposits of fully cleaned grain are also found (van der Veen 1985; Boardman 1995a, 1998; Holden 1998b; Summers and Bond forthcoming), probably burnt during drying or storage. There appears to be a shift in grain processing and storage practices during the Iron Age from small scale day-to-day processing to larger scale processing (Bond 2002; Summers and Bond forthcoming) and possibly central storage (Kendrick 1995; Holden 1998b; Dockrill 2002; Dockrill and Batt 2004; Dockrill et al. forthcoming). One of the key pieces of evidence in this argument is the presence of a corn-drying kiln at the site of Old Scatness, Shetland with its final phase of use dating to the 2nd to early 5th century AD (Dockrill et al. forthcoming). Corn-drying kilns seem more generally to be a phenomenon of the post-Roman period except in Roman military contexts  (Holden 2006; Cook & Dunbar 2008).

In terms of structural evidence, four- and six-post structures have been seen as granaries (on fairly slender evidence), while evidence from Scalloway (Shetland) hinted at grain storage within the broch, preserved by destruction in a fire (Sharples 1998, 31).

Souterrains may have been used for food storage (Hingley 1992, 35; Anderson and Rees 2006, 53-4), and it is possible that pits were widely used for storage of foodstuffs, although direct evidence rarely survives; there is no equivalent of the consistent, repeated digging of grain storage pits found in Wessex). Another line of evidence probably relating to both cooking and food storage are residues found on pottery sherds (Campbell et al. 2004; Brown and Heron 2004; Craig et al. 2005), an area meriting more work.

Utility/Consumption of plants and animals

The use of domestic mammals on archaeological sites is usually determined from examining the age and sex structure of the herds. The extensive fragmentation of bone often found on Scottish Iron Age sites, greatly hinders this. Sex is almost never determined for cattle and sheep and age structures can often only be suggested. At sites where better data (particularly for ageing) has been obtained, this suggests mixed economies where meat, dairy products, wool and possibly traction were all important elements (Seller 1982; Smith 1994; McCormick 2006; Bond 2007a; Cussans and Bond 2010, 2015). The association of large numbers of neonate cattle bones with dairy production has been much debated (see Mulville et al. 2005) but the availability of dairy products in the Western Isles has now been confirmed through analysis of pottery residues (Craig et al. 2005) although the intensity of such practices could not be determined from this method (not least because the coating of pots with milk, attested ethnographically, may bias the sample). It is now generally commonly accepted that high numbers of neonate cattle are related to dairying practices (e.g. Serjeantson et al. 2005; Mulville et al. 2005; Bond 2007a; Cussans and Bond 2010, 2015). However Mulville et al. (2005) point out that the subject is still very much under debate; for instance, McCormick (2006) found high quantities of cattle neonates at the site of Cnip, Lewis but here the interpretation was that this was the result of poor grazing (McCormick 2006, 167). The use of animals for traction has been interpreted at some sites from pathologies, usually found on the bones of the feet (e.g. Bond 2007a).

A key aspect for future work is the nature of inter-site differences in terms of the domestic animals and crops produced and consumed in different circumstances.

Characterising agricultural practice requires the combination of techniques and methodologies. This would allow researchers to build up a full picture of the procurement, use and consumption of plant and animal resources. How well, for example, do the results of isotopic studies coincide with data from animal bones and plant remains? Innovative approaches are required for topics such as dairying, the nature of specific meals, and how foods were combined together.

The extent to which wild resources were exploited, the role they played in Iron Age diet and the way they were thought of requires fuller consideration. Specific examples are the nature of the exploitation of resources such as deer (whether for antler or as prey) and cetaceans (hunted or expedient use).

Interpretation of age structures in e.g. cattle in terms of agricultural practice remains a topic of debate

There is a need to explore the nature of foddering practices, the over-wintering of animals and the provision of animal housing, including targeting the function of ring ditches when sites with appropriate evidence are encountered.