Grauballemandens sidste Måltid


  • Hans Helbæk



Grauballe man, Last meal, gravballe manden, sidste måltid


The Last Meal of Grauballe Man:

An Analysis of the Food Remains in the Stomach.

In the spring of 1952, the corpse of a man was found in a peat bog at Grauballe in central Jutland. Various aspects of the find have already been discussed by specialists in a series of papers published in Kuml 1956, and, of these, two of the most interesting were treatises dealing with the dating of the body. Results arrived at by the Carbon 14 method 33) and by pollen analysis 24) gave a date between the third and fifth centuries AD.

In the present paper, the contents of the intestines of the corpse are described. The publication, eight years ago, of Tollund Man - a body of approximately the same date, found under similar circumstances 10) - makes it possible here to include in the discussion of Grauballe Man, a comparison with the former. To this end the list of plants found in the stomach of Tollund Man is repeated here. In both cases the alimentary canals were intact, and the last meal of both men was preserved in its entirety, but while the stomach contents of Tollund Man was 275 ccms., there were 610 ccms. of food remains in Grauballe Man.

The food remains appear as a fine-grained brown mud mixed with larger particles consisting of seeds and fragments of seeds, internodes of cereal spikes, husks, sand, a couple of pebbles, and a number of small bone fragments. Very many seeds, particularly of Polygonum lapathifolium, and grains of wild grasses, were preserved intact, and in certain cases even the starch and the albuminoid matter were found to have kept their structural and chemical properties unchanged.

In the list, four dots denote the principal components of the meal; three dots indicate that over fifteen specimens of the species were found; two dots indicate a number from five to fifteen specimens; and one dot indicates less than five specimens. Complete seeds, as well as major, easily recognisable portions of such, are counted as specimens, but although, for instance, Holcus grains occur by the hundred, these tiny fruits do not amount to any significant volume and thus the species is represented by only three dots. On the whole, it must be admitted that the representation of the original proportions of the meal depends largely on the subjective judgement of the examiner, since the different degrees of disintegration in the various seeds and grains make an accurate estimate of these millions of particles quite impossible. An X indicates parasites as also the evidence of meat the proportional nutritive value of which cannot be estimated.

Very good evidence was found for the identification of the cereals. Fragments of the integument (seed-coat) of wheat were present in a modest number (Pl. I a), sometimes overlaid by the almost obliterated traces of the chlorophyll cells (Pl. II c). Although this organ is very similar in rye, the identity could be established in one large fragment by the presence of a fringe of hairs corresponding to wheat but deviating from the typical rye grain hair by its thick wall and narrow lumen (Pl. III f). There was evidence for rye, on the other hand, in a number of hairs with thin walls and wide lumen, which were found detached from their own seed shells and fortuitously stuck to a Brome grass integument (Pl. III e)2). Specification of the wheat was made possible by the occurrence of spikelet parts. One whole spikelet and several glume bases testify to the presence of a slender type of Spelt, Triticum spelta, (Fig 2 g, h), and a number of spikelet bases, glume bases and internodes, belong to Emmer, Triticum dicoccum. (Fig. 2 a, b, c).

Numerous fragments of the seed-coat of barley were found. Normally only the integument was preserved, determinable by the two layers of thin-walled cells generally orientated the same way (Pl. I d), very often attached to rather large areas of aleurone cells, characterised by their granulate plasma and distinct nucleus (Pl. I b). While the aleurone cells of wheat in this find vary from 28 to 72 µ, those of barley are only 20 to 29 µ, and, as opposed to wheat, they are often arranged in two or three layers. In many cases even the seed shell (pericarp) was partly preserved in connection with integument fragments. This is probably due to schorching by cooking, as the cells of this organ in the cereals are much less resistant than are the integument tissues. Large patches were observed, showing two or three transversely arranged layers, of the typical barley chlorophyll cells, (Pl. II d), sometimes even covered by the remnants of the distal hypodermis, large, fairly thick-walled longitudinal cells (Pl. II d). Neither epidermis nor hairs were encountered.

Many details of the barley spike were available to enable identification of variety and form. A series of internodes with attached remains of glumes, pales, and rachillas, make it possible to determine whether the grains were naked or hulled and whether the spike was erect or nodding, (Hordeum hexastichum or H. tetrastichum). The principles followed in determining whether the barley grain was of the naked or hulled form are based upon the disintegration of the spike by threshing. In Naked barley the pales and the rachilla will remain attached to the internode; in Hulled barley they will not. Thus, when, as in some cases in the Grauballe material, we find internodes with one or more rachillas and remnants of the pales attached we may take it that the floret base, or pedicel, was not broken by threshing but that the grain was thrown out of its husk in the naked state (Fig. 2 d, e). On the other hand, if the pedicel is broken and no traces of rachillas or pales are left, it may be concluded that these were beaten off attached to the grain which was consequently of the hulled form (Fig. 2 f). As it happens, base portions of the glumes were preserved in more than one example of inter­nodes of the hulled form. (Fig. 2 f).

It appears from the shape and size of the internodes, as also from the positioning of the lateral floret bases, that the spikes were six-row, and that they belonged to the nodding variety, whether of the hulled or naked form. Summing up the evidence one gets the impression that much more barley was hulled than naked, and that all of it was a form of H. tetrastichum.

The identification of oats presents a difficulty in that the cell tissue of its integument is virtually identical with that of Brome grass 25) (Pl. III a, b). Many good examples of the seed coat of the latter were isolated, and these could be identified by their shape and size alone. (They vary in length from 4.94 to 6.41 mm, and in width, from 1.83 to 2.20 mm). It is characteristic of them that the cork strip of their ventral suture always lies along the longitudinal axis of the grain: a feature due to the fact that ,the grain is almost flat in its fresh state and thus collapses without distortion when the endosperm disappears. The oats grain, on the other hand, is round in cross-section, and the cork strip may lie at the centre or edge of the flattened seed coat, according to the direction of the exterior pressure as the grain disintegrates. For the most part the integument fragments of oats were too small to indicate the size of the grain but the few whole seed coats were between 7.14 and 7.32 mm in length, 3.11 and 3.38 mm in width. Very faint traces of the disrupted transversal cells of the seed shell could be discerned on the distal side of some of the integument fragments of both Brome grass and oats (Pl. II a, b), but hairs or epidermal tissues were not observed. In one case, however, the partly dissolved pale tissues were found still attached to the dorsal surface of the lower end of a Brome grass fruit.

Judging by its size, the Brome grass may be referred to the species Bromus mollis. As to the oats, no determinable pale remains of the cultivated species were encountered, whereas three well preserved fragments of the ventral pale of Wild oat, (Avena fatua,) prove the presence of this species. It is characterized by a strong oval scar at the pale base and the rachilla. However, even without definite proof, cultivated oats may well be present, as the above dimensions of the grains are rather too large for the average Wild oat grain.

Profuse evidence was found of plant diseases. Many sclerotia of Ergot (Claviceps purpurea) were found floating in the material (Pl. IV d) and in some cases they were still found in situ in the grass florets, especially in Holcus (Fig. 3· g). Spores and simple hypha systems occurred by the million, a circumstance which caused the writer to call upon the assistance of mycologists. Professor N. F. Buchwald, plant pathologist of the Royal Veterinary and Agricultural College, Copenhagen, took care of the barley diseases and he identified the hyphae of the fungus family Dematiaceae (Pl. VII a), and Covered barley Smut (Ustilago hordei). The spores of the latter are 6-8 µ, thick-walled and smooth (Pl. VII b). In the absence of spores, the former cannot be identified by species. Professor M. Lange and Dr. A. Skovsted, of Copenhagen University, identified the spores of two other species of Smut, U. utriculosa and U. anomala. The former lives on Polygonum lapathifolium, and its spores are from 10 to 12 µ in diametre (Pl. VII d), whereas the latter may be found on P. convolvulus or P. persicaria, having spores of 11 to 14 µ in diametre (Pl. VII e).

The plant list contains eleven species of weed- or wild grasses. Apart from Agropyron caninum (Fig. 3 j), Deschampsia caespitosa (Fig. 3 c), and Poa nemoralis, (Fig. 3 b), all of them have previously been found in prehistoric contexts in this country 15). It is interesting to note that Echinochloa (PI. V a, Fig. 3 a, e) was then frequent enough to occur here, in the Tollund stomach (where, unfortunately, it was wrongly identified as Setaria pumila) 10), and in the Fjand, Early lron Age find 36). The spikelets of Phragmites are, of course, fortuitous; nobody would collect this species for food, but it may have occurred on the edge of a waterlogged field, or the fragments may have fallen into the food from the thatch of the dwelling. While the other grasses may occur as field weeds, Sieglingia (Pl. IV a, Fig. 3 f, h) is a grassland plant, and Agropyron and Poa nemoralis belong to the wooded country 4, 7). Thus these three must have been collected outside the cultivated area. Agropyron does not occur in an ecological station like the Grauballe district, and this circumstance, as well as the occurrence in the stomach of another woodland plant, Campanula glomerata, suggests that the Grauballe man did not belong to the actual locale where he was found, but came rather from the wooded strip at the same latitude, but further east along the coast.

Most of the dicotyledon seeds and fruits have been found before in Danish prehistoric finds and do not call for a description here 15, 23). Only Rumex crispus, Cerastium, Alchemilla (Fig. 3 I), Potentilla argentia and P. erecta, (Fig. 3 k), Trifolium campestre, Rhinanthus (Pl. V c), Campanula, Achillea, Matricaria inodora Sonchus asper, and Crepis capillaris are met with here for the first time and are treated in some detail. The fruits of Ranunculus acer and R. repens have both been found before, but the opportunity is taken here of giving a description of the anatomical characters on which discrimination is based.

Under the pericarp distal epidermis there is in both species a coarse paremchyma in the cells of which cubic crystals are embedded which reach into the underlying sclerenchyma, forming shallow pits. In R. acer these crystals occur in every cell and form an even pattern all over the central portion of the fruit (Pl. IV c), while in R. repens they are more dispersed and irregular. In our examples of R. repens the present surface consists of the naked sclerenchyma with the pits left by the torn-off crystals (Pl. IV b).

As mentioned above, certain non-vegetable matter was found in this stomach. Like the Tollund Man's, it contained about one table-spoon-full of fine-grain sand. How it came to be in the stomach is not readily explicable, but it is suggested that the particles may have come from the collected seeds. The fruits of Polygonum, for instance, are gathered with their glandular perianth. Earth particles thrown up by rain cling to the surface of these bracts, and this circumstance may well account for the sand. By careless handling of the food two comparatively large pebbles got into it, the largest 8.5 by 5 mm, as well as an unchewed piece of charcoal, 7 by 6 mm.

Among about a dozen fragments of bone, softened by the dissolution of the calcium phosphate in the acid bog water, only one was large enough for an attempt to be made at determining the skeletal part it came from and the size of the animal. The bone specialist Ulrik Möhl was kind enough to undertake an examination, but the evidence is too slender for an identification. However, he states that the largest piece, 7 mm long and 15 by 3+ mm in oval cross section with a thin wall of solid bone matter and filled with cancellous tissue, may come from the lower end of a rib, or from a spine or transverse process of the vertebra of a medium-sized domesticated animal. Provided this to be the case, the size of the bone in question would correspond to a piglet. Be this as it may, the main thing to us is the fact now established, that the Grauballe Man ate meat.

Some very small hairs were kindly examined by the hair specialist, P. Valentin-Jensen. Here again, we must be content with the statement that these are animal hairs; they are too small even for the fur of a common mouse, although at certain places on the skin of that animal similar hairs can be found, on nose and feet. Identification is not certain, however. (Pl. VII c).

In the Grauballe as well as in the Tollund stomach a great number of small objects occurred, unknown to the writer. Through the cooperation of the Laboratory of Freshwater Biology of Copenhagen University (Professor Kaj Berg) and Dronning Louise's Children's Hospital (Professor Oluf Andersen) it was established that they are the eggs of the intestinal worm Trichuris trichiura which is very common even in our day all over the world except in the Arctic region. They vary in size from 50 to 59 µ (Pl. III d). The worm itself, from 40 to 50 mm long, has evidently dissolved 31, 32).

Thus the last meals of the Tollund and Grauballe men prove to be virtually identical. They were composed of starch and fat. The Tollund man had not eaten animal fat, but his meal did include vegetable fat, derived from a not inconsiderable amount of linseed and seeds of Camelina 10). The Grauballe Man had eaten animal fat with his vegetable soup, and the small pieces of bone bear witness to his having gnawed the soft ends, probably of vertebrae or ribs. Only 15 fragments of linseed (Pl. V b) and no seeds of Camelina were found. (The Camelina in the list refers to a fragment of a silicle valve).

It has been suggested that these meals were ritual and therefore of a special character, on the assumption that we are dealing with sacrificially killed persons, and on the further premise (based upon the Tollund 10) and Borremose analyses 1)) that it would be unreasonable to assume that Iron Age people normally kept a purely vegetarian diet 3). In the first place, there can be no doubt that in the Iron Age very little meat was generally available in this country. On most habitation sites of this period we find conspicuously fewer bones than on, for instance, Neolithic sites. Also, from the deposits of carbonised plant food and imprints in pottery we know definitely that seeds of weeds played an important role in the domestic economy of Iron Age man 5, 9, 11, 17, 19). In the poorer districts of Jutland the land had to lie fallow for long periods, and, the arable being thus restricted, the peasants could not afford to disregard the food value of the wild plants which sprang up on otherwise unproductive land. This was ecologically conditioned by a combination of climatic changes detrimental to the existing stage of agricultural techniques and over population leading to the cultivation of unsuitable land 6).

Judging by the comparatively large objects, the Spelt spikelet, the two pebbles, and the piece of charcoal, we may take it that the man drank his food rather than ate it with a spoon. One does not disregard pebbles when eating solid food, and the charcoal is extremely friable. It is to be concluded that the food had the form of a thin soup, perhaps with the meat in it. With regard to the preparation of the meal, the scorched portions of grain shells indicate that heat was applied at some stage. It seems reasonable to assume that this was the consequence of simple boiling; if the grain had been roasted before the cooking some more or less thoroughly carbonized fragments of grain would be expected, but none were in fact found. There were however, a few small flakes of carbonized vegetable conglomerate which had evidently come off the inside of a cooking pot.

Under the circumstances, it is amazing that in some grass fruits the starch survived in agglomerate grains, and that, like fresh starch, it stained with potassium iodide. Likewise it is peculiar that the aleurone had not decomposed, even though the plasm bodies were detached from their walls in all cases but one. Although the proportions of cereals and weed seeds cannot be ascertained, the examiner is of the opinion that the seeds were, if not actually the principal component, very numerous indeed. Thousands of Polygonum fruits were preserved intact, indicating incidentally, that the grinding of the material was not very thorough.

By collating the results of the investigation of Danish lron Age grain deposits and grain impressions, with regard to the proportional frequency of Naked and Hulled barley, we may independently get a hint as to the date of the Grauballe Man 9, 11, 19). It is probable that Naked barley went out of use in this country during the first two centuries of our era, and already by 100 AD it seems to have been largely replaced by the hulled form. Since Grauballe Man ate Hulled barley in conspicuous amounts, therefore, we may settle for a date within these two centuries. The evidence for the extinction of Naked barley is universal from western Jutland to Bornholm, and thus even if we have no finds of grain from the actual environment in question, we are justified in concluding that the same was the case in eastern central Jutland.

The presence of the hairs of rye grains (Pl. III e) is another detail fixing Grauballe Man to the time after the birth of Christ since this cereal was first observed in this country in finds of the first century 9, 19).

Judging by modem conditions, the change-over from one form of barley to the other was caused by the increasing humidity which began in the first millennium B. C. While Naked barley may still yield tolerably well in this country, it is far more susceptible to diseases than is the hulled form, and after a few generations, therefore, it will degenerate. This situation was well illustrated by the Dalshöj find. (First century AD).

The occurrence of Spelt is unexpected. The existence of the cereal has been established in this country at the end of the Bronze Age 12), but no trace of it has been observed among the many deposits of Iron Age grain. The Bronze Age Spelt was not associated with the two weed grasses, Wild oat and Brome grass, but these two species which we find in the Grauballe stomach, were typical companions of the Spelt of the Early Iron Age in southern England 13). This may mean that the continental region from which the Iron Age Spelt was introduced into Denmark, was identical with that which supplied the English Spelt, but not the place of origin of our Bronze Age Spelt. The early history of this cereal in central Europe, however, is practically unknown, except in southern Germany and Switzerland 29, 30), so we cannot guess at the actual centre or radiation of these times.

It is typical of the stomach contents of Grauballe Man, as also of those of the Tollund and the Borremose men, that it comprises nothing that could not keep all year round. No seasonal vegetables such as fruit, berries, or pot-herbs were found. This circumstance suggests one common feature about the three bodies, viz., that they met their death during the winter. This, more than any other feature about the burials reinforces the theory that the corpses represent ritual sacrifices. The manner of death varies, sometimes slashing of the throat, sometimes strangulation; in some cases the corpses are accompanied by pieces of clothing, in others not 3, 34, 35). Three circumstances link our three cases together, however: they died by violence: they died during the winter: and they were not buried according to prevailing custom but were flung into a disused peat cutting. It is likely that we have here tangible evidence of the sanguinary rites of the Iron Age Midwinter Festival, when the peasants were accustomed to invoke the deities of fertility for abundance in the new born year.

This paper was made possible only by the invaluable aid of the scholars whose individual reports are included in it, and also by the generousity of the Carlsberg Foundation in supplying the high powered microscope without which the cereal analysis and the photographs, in particular, could not have been achieved. The writer wishes to extend his sincere thanks to all concerned: to the Board of Directors of the Carlsberg Foundation; to Professor N. Fabritius Buchwald; to Conservator Ulrik Möhl and Mr. P. Valentin Jensen, M. Sc. Also to Professor M. Lange and Dr. A. Skovsted; to Professor Kaj Berg and Professor Oluf Andersen.

Hans Helbaek.





Helbæk, H. (1958). Grauballemandens sidste Måltid. Kuml, 8(8), 83–116.