Macro fossils

The aim of the article is to discuss how the composition of Bronze Age macrofossil samples reflects different aspects of daily life like diet and cooking. The article argues that the increasing weed content in the Late Bronze Age macrofossil samples should partly be regarded as a new resource that was used in the cooking process. The contemporaneous increase in hulled barley at the expense of naked barley and wheat, might reflect a diminished interest in baking leavened bread and a stronger preference for cooked cereal-based dishes. These changes in the domestic sphere should be regarded as intimately connected with changes in the Late Bronze Age cosmology, in particular with the development of the Umfield culture.


INTRODUCTION
The aim of this article is to discuss how the Bronze Age macrofossil samples can be regarded as a cultural phenomenon, intimately linked to choices and decisions made in connection with the cooking process. From this point of view the macrofossil samples demonstrate different aspects of cultural traditions, for example, the type of plant food that could be eaten and how to prepare the cereal-based dishes. I will also argue that the composition ofthe macrofossil samples is part of an ideology expressing the cosmology of the Urnfield culture.
During the last ten years changes in the composition of Bronze Age macrofossil samples from southern and central Sweden have been discussed intensively. The debate has focused on how the changing composition ofthe macrofossil samples may reflect alterations in the agricultural system (Engelmark 1992; Gustavsson 1995Gustavsson , 1998. The Ystad project that was concluded around the year 1990 demonstrated that there was a change in the cultivated crops at the transition between the Early and the Late Bronze Age. During the Early Bronze Age emmer, Triticum diococcunt, spelt wheat, has been explained as a result of the introduction of manured cultivation. Fat hen has a great demand for nitrogen and can therefore be seen as an indicator of manuring (Engelmark 1992:372;Gustavsson 1998:67).The change from naked barley to hulled barley has been given a similar interpretation. Hulled barley is thought to have greater requirements for easily soluble nutrients, especially nitrogen, and is thus benefitted by manuring (Engelmark 1992:372;Gustavsson 1998:67f).
Against this, Per Lagerås and Mats Regnell have questioned whether the change in the agricultural system was as dramatic as has been proposed. They have argued that there is no straightforward connection between the occurrence of hulled barley and fat hen on the one side, and the introduction of manured fields on the other (Lagerås k Regnell 1999).
Common to the debate concerning Bronze Age macrofossil samples is an interest in those activities that took place in the field. Little interest has been paid to the activities carried out around the hearth. Therefore, I will argue that there is a need for a broader debate. Issues like diet, cooking and cosmology are important factors if one wants to understand the changing composition of the macrofossil samples. In the following these matters will be touched upon.

DIET
The Danish scholar Hans Helba:k wrote in the 1950s several articles in which he stressed diet as an important factor, if one wants to understand the composition ofthe macrofossil samples (Helbaek 1951a(Helbaek , 1951b(Helbaek , 1955(Helbaek , 1958. Helba:k argued that the weed seeds were intentionally collected and used in the cooking process. He proposed that harvesting was carried out with a sickle, that only the ears ofthe cereals were cut, and that the weeds were avoided. Thereafter the fields were fallowed and the weeds were intentionally collected in the fallowed fields (Helba:k 1958:107).
Contrary to this, Karin Viklund has stressed the importance of the cleaning process (Viklund 1998).She proposes that the crop was cut close to the ground, in a way that resulted in a mix of cereals and weeds (Viklund 1998:40ff). To get a clean cereal product that could be used for consumption and seed corn, the harvest had to undergo a series of threshing and cleaning processes. The purpose of cleaning was to separate the weeds and chaff from the threshed grains (Viklund 1998:60ff).
The cleaning could be carried out in different ways (Viklund 1998:60ff).To "winnow" meant that the mixture of grains, chaff and weeds that was left after the threshing, was placed in a winnowing-through and tossed in the air. By this action the lighter particles, like chaff and most of the weed seeds, blew away and the heavier grains and weed seeds fell back on the floor. Pouring the cereals from one container into another, outdoors in the wind, would give a similar result. Threshed cereals could also be flung by using a shovel. The lighter chaff and seeds were deposited closest to the person, while Contrary to Helba:k and Welinder, Viklund suggests that the sample from house No. 11 that contains weed seeds, represents a threshed cereal sample which has not yet been cleaned (Viklund 1998:56ff).If the crop was cut close to the ground and had not yet been cleaned, this would explain the high proportion of weed seeds. According to Viklund, the wooden tub that stored the seeds is similar in shape to the winnowing-through known from historical times. Thereby the find probably represents a vessel of seeds that was about to be winnowed at the time of the conflagration.
The other macrofossil sample in house No. 11, which was totally devoid of weed seeds, might represent a product where the cleaning had just been finished.
Another find, also containing a lot ofweed seeds, was made at Gerding in Denmark in 1949 (Helba:k 1951a). From a burned-down settlement, dated to the pre-Roman Iron Age, Peter Skoglrrnd a jar with a volume of about one litre was found. The jar was complete and identical to those jars found in contemporaneous graves.
The size ofthe jar and the find context suggest that it is not a storage vessel, but a jar used in the cooking process. The jar contained 95 cc of cereals and weed seeds (table 2) The main contents were naked barley, fat hen, black bindweed, Polygonum convolvulus, pale persicaria, Polygonum lapathifolium, and corn spurrey, Spergula arvensis. In addition to these seeds the jar also contained small amounts of a further 17 species. Helba:k proposes that the high proportion of fat hen, black bindweed and corn spurrey means that these species have been intentionally collected, and thereafter used in the cooking process (Helba:k 1951a:68fl).
The Gerding find has also been discussed by Viklund, who argues that the plant material could be interpreted in terms of selection, a product obtained by cleaning the crop (Viklund 1998:78). The length, breadth and thickness of the grains indicate that it is a cleaned fraction from which the biggest, and perhaps also the smallest, grains have been removed. The abundance of the rather large seeds of black bindweed suggests that it does not represent a sieved fraction. Since the size Table 2. The main coniponents ofthe Gtrrding find.
In addition to the main components, 17 species were represented with I -16 seeds, which makes a total of a firrther 86 seeds (afier Hefbtek 1951a:68). of the grain is so restricted, the sample probably represents a product obtained by flinging, probably tail grain or mid-grain.
Viklund's idea, that the weeds were harvested together with the cereals, and that weeds and cereals were thereafter sorted in different manners, seems quite convincing. At the same time, I would argue that it is possible to combine this notion with Helba:k's idea, that finds containing a lot of weeds were sometimes meant to be eaten.
These interpretations, together with the find circumstances, have favoured the idea that they are odd and spectacular finds, and that they could not be used for general conclusions. For example, the Danish archaeologist P. V. Glob argued that the bog corpses have been sacrificed, and that their stomach contents represent special meals eaten just before the sacrifices (Glob 1969:140).However, Helba:k, who studied the stomach contents of the Tollund man and the Grauballe man, was convinced that the macrofossils found in the stomachs of these bog corpses reflected normal meals. He came to that conclusion by comparing the stomach contents with the composition of macrofossil samples from contemporaneous Early Iron Age sites (Helba:k 1951b:325ff, 1958:106ff).
A common feature for the stomach contents of those three Danish bog corpses that have undergone a careful examination, is the dominance of vegetarian food. The stomachs contained a broad spectrum of seeds including Current Swedish Arehaeotogy, Vot. 7, 1999 a high proportion of weeds (Brandt 1951;Helba:k 1951bHelba:k , 1958. The stomachs of the Tollund man and the Borremose man contained seeds from about 20 different species respectively, while the stomach of the Grauballe man contained seeds from about 60 different species. In the stomach of the Tollund man the bulk contents were seeds from hulled barley, flax and black bindweed (and others from the same genera) Polygonum sp. , mixed with small amounts of gold of pleasure, fat hen, corn spurrey and field pansy, Viola arvensis (Helba:k 1951b, 1958).The main ingredients in the last supper of the Grauballe man were seeds from hulled barley, black bindweed (and others from the same genera), and soft brome, Bromus hordeaceus, mixed with small amounts of wheat, oat, fat hen, corn spurrey, sheep's sorrel, perrenial rye-grass, Lolium perenne, Yorkshire-fog, Holcus lanatus, ribwort plantain, Plantago lanceoiata, and greater plantain, Plantago major. The stomach also contained small amounts of animal fat (Helba:k 1958).The last supper of the Borremose man was composed of seeds from black bindweed (and others from the same genera) and corn spurrey mixed with fat hen, sheep's sorrel and Yorkshire-fog (and others from the same genera), Holcus sp. (Brandt 1951).
Helba:k concludes that the last supper of the three bog men seems to have been a gruel or porridge based on cereals and weed seeds mixed with some fat. The fat in the porridge of the Tollund man was vegetarian fat, while the fat in the porridge of the Grauballe man was animal fat (Helba:k 1958:106). The stomach contents of the bog corpses have also been discussed by Jimmy Jonsson (Jonsson 1991).He argues that the bog men's last suppers were made of flung seeds, like the tail grain or the mid-grain that often contain weed seeds and chaff. The composition of these fractions seems quite similar to the mixture of chaff, weed seeds and grains that were found in the stomachs of the bog corpses (Jonsson 1991:24ff).Another alternative is that the seeds found in the stomachs had been winnowed. By repeating the winnowing it is quite easy to remove the chaff.
Since the stomachs ofthe bog corpses contain chaff, it is reasonable that the seeds had only been winnowed once or twice. Jonsson concludes that the bog men had eaten dishes made of seeds that were badly cleaned (Jonsson 1991:19ff).
It has been suggested that weed seeds and chaff were only mixed with cereals in years of bad harvest or during famine years (Engelmark 1992:373; Rowley-Conwy 1988:189;Viklund 1998:105).However, Jimmy Jonsson, who studied a total of eight North-European bog corpses, concluded that seven out of eight stomachs contained high proportions of weed seeds and six out of eight stomachs contained chaff. It does not seem probable that all these persons should have died in a famine year (Jonsson 1991:35).It is more likely that people did not care very much whether the food contained some weed seeds and chaff. From historical times it is well known that peasants added weed seeds and chaff to the meal, to make the stored seed last longer. This was not only done in years of bad harvest but also in ordinary years (Keyland 1989:I I I ff).
The stomach contents of the bog corpses favour the idea that people did not only eat the cleaned products, but also the fractions made up of a mixture of grains, weed seeds and chaff. Therefore, returning to the Gerding find, it is quite possible that it represents a flung fraction as has been proposed by Viklund.  Bread that is to be yeast-leavened must be baked with wheat which contains gluten (Hansson 1997:II:10). Barley, oat and millet contain far too little gluten, which in any case is of the wrong chemical composition, to produce the high, light and spongy bread that is baked from wheat-flour (table 3). Since wheat was in process in the Neolithic and the Early Bronze Age, leavened bread could have been baked by mixing flour from wheat with flour from other cereals.
Leavened bread requires uniform heating and space to expand, which could be attained by using an oven. Since no ovens have yet been found in Neolithic and Bronze Age contexts, it has been assumed that leavened bread was not baked in those days (Viklund 1998:143f). However, leavened bread does not necessarily require an oven; it could also be baked in a jar. In the Faeroes in the 19th century' leavened bread was baked in an iron jar placed at the fire (Campbell 1950:7ff).It seems probable that leavened bread could have been baked in jars, or in simple ovens, that have left no significant archaeological traces.
It has also been argued that the speltoid wheats are hard-threshed, and therefore were roasted to facilitate the separation of the chaff from the grains (table 3) (Hansson 1997:II:10).
When roasted and heated the gluten proteins lose their elasticity and cannot work in the leavening process. An alternative way of Current Swedish Archaeo!ogy, Vol. 7, 1999 separating the chaff from the grains would, however, be to crush the grains and thereafter separate the grains from the chaff. There is no clear evidence that leavened bread was baked during the Neolithic and the Early Bronze Age. However, I would argue, the fact that wheat was in process strongly indicates that leavened bread was baked. But, if too many other ingredients than wheat-flour were added, the leavened wheaten bread might not rise (Hansson 1997:II:12).The change from cultivating both wheat and barley, towards a predominance for barley, could therefore be interpreted as reflecting a diminished interest in baking leavened bread and a stronger preference for baking unleavened bread or making cooked cerealbased dishes.
Ann-Marie Hansson has listed three kinds of cereal-based dishes that were in process during prehistoric times; grain-paste, porridge and bread (Hansson 1997:II:Sff).
Grain-paste and porridge are similar dishes prepared from whole, chrushed or partly ground cereals. The difference between the dishes is that grain-paste is not cooked but porridge is always cooked. Bread differs from grain-paste and porridge in that the cereals that constitute the bread are ground and thereafter roasted or baked (table 4).
When making porridge one can avoid the labour-demanding grinding process and instead crush the cereals at a grain-stamp ( fig.   2). Before the water mill was invented, grinding was a very labour-demanding process. Eighteenth-century records from the Faeroes say that one person had to grind a whole day to make enough flour to satisfy the  Cereals contain I 7% fat, while the oil content of flax and gold of pleasure could be estimated to 30-40% (Hansson 1997:48). When flax and gold of pleasure started to be cultivated in monoculture at the transition from the Early to the Late Bronze Age, this meant that they were desired crops, probably for their oil content (Hansson 1997:48).The introduction of flax and gold of pleasure favours the idea that it became more common Current Swedish Arehaeotogy, Vot. 7, l999 to make fat cereal-based dishes. Fat is the most concentrated source of food energy, and the introduction of new species rich in fat indicates that people's need for fat now to a greater extent than before was being satisfied by vegetarian fat. To summarise, it seems possible that the changes in the macrofossil samplesi.e. the decreasing amount of wheat, the replacement of naked barley by hulled barley and the introduction of new cropsreflect a greater interest in making fat cereal-based dishes where the cereals were crushed and thereafter cooked. In the next chapter I will argue that the above-discussed changes should be seen as part of the changes in the North-European Bronze Age society.

COSMOLOGY
Contrary to the replacement of naked barley by hulled barley, which is due to regional variations, the introduction of oat, millet, flax and gold of pleasure seems to be a synchronous phenomenon in northern Europe at the transition from the Early to the Late Bronze Age. These plants have been documented in small amounts earlier, but it was now they became a substantial part of the cultivated crop (Harding 1989:174ff).This phenomenon could therefore be associated with the introduction of the Urnfield culture in northern Europe.
As Harry Fokkens has stressed, the introduction ofthe Urnfield culture should be seen not only as an economic and religious change, but as a change also involving the social organisation of society. A general decrease in the size of the houses has been interpreted as a change from multiple family households to nuclear family households (Fokkens 1997;Karlenby 1994;Skoglund 1999). These general changes in the size of the houses are highlighted at the Apalle site in central Sweden. This site offers a unique chance to study how the social organisation, and the organisation of the cooking process, changed at the transition from the Early to the Late Bronze Age (Ullén 1994).
During the early settlement phase, 1200-1000 BC, the houses were longer than during the later phase ofthe settlement, 800 -700 BC.
The earlier houses were usually divided into two rooms, one larger and one smaller. The larger room was the entrance room, and here most ofthe artefacts were found. The entrance room never had a hearth. A shallow hearth was found in the inner room of the house. The main use of this hearth was probably for heating. Cooking seems to have taken place outside the houses, where several big cooking pits were found. Central at the settlement were also two heaps of fire-cracked stones surrounded by houses ofthe earlier period (Ullén 1994:252+.  4).
As John Barrett has pointed out, there is usually a connection between ritual and everyday life (Barrett 1991:4ff). By metaphorical associations the everyday values are transformed into timeless truths, and when these values are recreated in ritual, they seem to derive from another timeless world.  (Gräslund 1983(Gräslund , 1989(Gräslund , 1994.Often it is suggested that the use of fire helped to separate the soul from the body. The central position of the hearth in the Late Bronze Age houses at Apalle, might be connected with the central role of fire in the Late Bronze Age rituals concerning death. This is underlined by the fact that quern-stones and cooking vessels are artefacts often found both at settlements and in graves from the Late Bronze Age. Thereby they reconcile the differences between the domestic sphere and the rituals concerning death. The quern-stones found in graves are often complete and were probably put in the graves intentionally (Kaliff 1997:88ff).
Pottery found in graves often contains organic materials. Sometimes the vessels containing the organic materials have been used to store the cremated bones ofthe dead person. Anders Kaliff has suggested that this reflects the tradition of using ordinary cooking vessels as containers for the cremated bones (Kaliff 1997:101f).
It seems as if tools connected with the activities carried out around the hearth, also were used in the rituals performed at the cemetery. This is probably due to the way of conceptualising the world, using daily-life concepts while thinking of religious matters or vice versa. Using a structuralistic view, it is possible to argue that the transformation of the seedsby crushing, grinding and then cooking them in a jar in the firecould have been associated with the treatment ofthe dead person, the cremation of the body and the burial of the bones in an urn (Kaliff 1997:88;Skoglund 1996:25f). The microcosmosrepresented by the house and the hearthwas thereby associated with the macrocosmos (Burström 1995:166ff;Leach 1976:60f).  Currenr Swedish Archaeology, Vol. 7, 1999 It seems possible to argue that the links between cosmology and cooking were reinforced at the transition between the Early and the Late Bronze Age. Ritual was made stronger by metaphorical associations with the work carried out around the hearth. This probably facilitated the contemporaneous changes in the choice of diet and in the cooking process.
To summarise, I would argue that the changes in the composition ofthe Bronze Age macrofossil samples are contemporaneous with the introduction of the Urnfield culture.
Changes in agriculture, diet, religion and social structures should therefore be seen as different aspects of the same new cosmology.

SUMMARY
With macrofossil samples as a starting point, this article has discussed different aspects of daily life like diet and cooking as well as the connections between household, society and rituals concerning death. The conclusions can be summarised as follows: The increasing weed content in the Late Bronze Age macrofossil samples should not only be seen as a result of technical innovations like the introduction of manuring or new harvesting methods. The weeds represented also a new resource that could be used in the cooking process. Neither should the increase in hulled barley only be seen as a result oftechnical innovations, like the introduction of manured fields. Issues like cooking and how to make cereal-based dishes must also be considered when discussing Bronze Age macrofossil samples. The Scandinavian Bronze Age agriculture is intimately connected with developments on the Continent. New ideas ofhow to organise agriculture and what food to eat, were probably part of an ideology expressing the cosmology of the Urnfield culture.
A CKNO 8'LEDGEMENTS I am grateful to Per Lagerås and Kristina Jennbert for constructive discussion and for their support in commenting on this article.