a contribution to the study of the medieval ......medieval icelandic literature and it was interest...

BRUNO BERSON

A CONTRIBUTION TO THE STUDYOF THE MEDIEVAL ICELANDIC FARM:

THE BYRES.

Since the end of the nineteenth century numerous farm sites dating from the middle ageshave been excavated in Iceland. More than 50 have now been partially or fully excavated(Margrét Hermanns-Auðardóttir 1989). The focus has always been primarily on thedwellings, and while a number of animal shelters have been excavated their characteristics,development and significance remains poorly understood. This paper gives an overview ofarchaeologically investigated medieval byres in Iceland, discussing the nature and limita-tions of the evidence, the principal characteristics of the buildings and their value forresearch into medieval Icelandic society and economy.

Bruno Berson, 60 rue Jules Charpentier, 37000 TOURS, France(Email: [email protected])

Keywords: byres, mediavel Iceland, animal husbandry

ARCHAEOLOGIA ISLANDICA 2 (2002) 37-64

The sitesThe first excavation of a byre tookplace in Lundur in Borgarfjörður in1884. The antiquarian SigurðurVigfússon took interest in the ruinwhen he heard local traditions to theeffect that it was a heathen temple. Hecarried out a hasty excavation andbecame convinced that he had foundthe ruins of a Viking age temple(Sigurður Vigfússon 1885). More thanfifty years later, in 1939, aScandinavian archaeological expedi-tion investigated a number of sites inÞjórsárdalur and Borgarfjörður. Thesite at Lundur was chosen for re-exca-vation, led by the Finnish archaeolo-gist Jouko Voionmaa (1943).

Voionmaa came to the conclusionafter carefully excavating the struc-ture that it was not a temple but anordinary farmstead dwelling. Basedon the excavation of a similar struc-ture at Gröf, which clearly was a byrewith attached barn, Kristján Eldjárnproposed in 1964 that the structure atLundur was neither a temple nor adwelling, but a byre connected to abarn (Kristján Eldjárn 1964). Thisidentification has since been generallyaccepted.

The structure was built of stone-linedturf walls and divided into fourrooms. The byre was in the south partof the building, measuring 9 x 3,5 m

38

BRUNO BERSON

Fig. 1: Sites discussed in the text. (Bruno Berson)

on the inside. It had three aisles, thecentral one being fully paved from thenorth to the south and sloping towarda doorway on the south end gable. Thestalls were along the walls on eachside of the paving, but divisionsbetween individual stalls were notclearly visible. At the northern end thebyre was divided from a smaller roomby a turf and stone wall. This roommeasured 3,8 x 2,2 m, opening intothe barn to the north. The barn floorwas sunken, cut directly into the slop-ing hillside. It measured 9,75 x 3 m onthe inside. It had two rows of post-holes dividing it into three aisles. Incontrast to the byre the floor-surfacein the barn was flat. An additionalroom was to the side of the small cen-

tral room, through which a secondentrance lay into the building. Thewhole structure was built on a slopewith a marked sloping of the floorfrom the north end of the small roomin the middle to the doorway on thesouth end of the byre. This doorwayopened on to a slope dropping steeplyfrom the gable-end of the building.The sloping of the floor in the byre,the location of the building on the topof a steep hill-side overlooking thefarm-site proper, the doorway on thegable end and the lack of a fire place– and the very small number ofobjects retrieved – all support theinterpretation of this building as abyre with an attached barn. The datingof this site as medieval is based on its

39

A CONTRIBUTION TO THE STUDY OF THE MEDIEVAL ICELANDIC FARM: THE BYRES

Fig. 2: Layout of the farm at Lundur. After Voionmaa, 1939.

40

BRUNO BERSON

Fig.3: The byre and the barn built together at Lundur. After Voionmaa, 1939.

41


Fig. 4: Layout of the farm at Þórarinstaðir. After Eldjárn 1949.

42

BRUNO BERSON

shape, in particular by analogy withthe 14th century farm of Gröf, but noother dating evidence has beenretreived.

In 1945 Kristján Eldjárn carried outan excavation at Þórarinsstaðir inHrunamannaafrétt (Kristján Eldjárn1949). This site, which is in the high-land margins, some 300 m above sealevel, was first described in 1895 andwas partly excavated by Daniel Bruunin 1897 (Bruun 1928). Unusually forits time the 1945 excavation revealednot only the dwelling complex butalso several detached animal sheltersand a smithy. It appeared that thedwelling had been covered by a thicktephra layer before the roofs collapsedinto the individual rooms and this wasconsidered to account for the remark-able preservation of the ruins. Thedwelling complex is unique amongmedieval farmsteads in Iceland in thata byre and a possible barn are con-nected directly to and accessed fromthe inside of the complex.

The byre measured 8 x 3,80 m on theinside and was divided into threeaisles. The central aisle was pavedand extended from a doorway leadingto a hall in the south to another door-way leading to an entranceway and asmall cell at the north end of the com-plex. The rows of stalls were clearlyvisible, demarcated by large uprightslabs. On the western side 9 stallswere identified but only 5 on the east-ern side where a doorway lead to abarn measuring 6,2 x 2,2 m. In the

field surrounding the dwellingremains of three buildings identifiedas sheep-sheds were excavated. Onewas badly eroded, situated at the backof the dwelling, at least 4 m long and1,25 m wide on the inside. Slabswhich had made up a manger alongthe eastern long-wall were observed.A second sheep shed was 20 m north-west of the dwelling complex, alsoorientated north-south. It measured10,5 x 2,4 m on the inside and hadslightly curved long walls. Remainsof mangers made of vertical flat slabswere found along both walls. Thethird sheep shed was by far thelargest, measuring 22 m in length. Itwas divided into three rooms, an 18x2m sheep stall, another 4,5x1,3 m stalland a 4,75x2,1 m barn. The stalls bothhad mangers made of slabs along oneside. Kristján Eldjárn calculated thatsome 160 sheep could have been shel-tered in these buildings.Þórarinsstaðir is the only site wheremedieval sheep sheds have been exca-vated in Iceland. The location of thefarm in the snow-heavy Icelandichighlands will certainly have forcedthe farmer to build shelters for all hislivestock, but the 1945 excavationwas also unusually thorough in inves-tigating all visible structures at thesite. The ruins were full of a tephralayer, which was later identified as theash from the eruption of Hekla in1104, suggesting an 11th century datefor these ruins. While the largestsheep-shed seemed to have beenenlarged at least twice, the structuresdid not otherwise show signs of repair

43


Fig. 5: Plan of the centralised farm and the two sheepsheds. After Eldjárn, 1949.

44

BRUNO BERSON

Fig. 6: Plan of the byre Bergþórshvoll. After Eldjárn K. & Gestsson, G., 1951.

or rebuilding and the excavator con-cluded that the site had only beenoccupied for a matter of decades.

The farm Bergþórshvoll has becomefamous as the scene of the burning ofNjáll and his family described inNjáls saga. According to the saga’schronology the farm was destroyed byfire in 1011. Njáls saga has long beenconsidered the masterpiece ofmedieval Icelandic literature and itwas interest in the saga which led tothe excavation of the farm mound in1927-1928. The results of the excava-tion were not published until 1952 andthen only in summary form. The orig-inal excavation had not revealed anytraces of fire and in 1951 an addition-al trench was added to the west of themain farm mound. The mound is sit-uated on a small rise in an otherwise

flat marshland cloase to the south-coast. The 1951 trench revealed muchless building remains than the largerarea excavated in the 1920s.Underneath a succession of threefloor-layers interpreted as dwellings afloor bearing traces of fire was found.This structure was identified as abyre. It was divided into three aislesby wooden posts. The central aislewas not paved but was lower than theside-aisles. Along the sides remains ofstall-divisions were found in the formof wooden stakes and postholes. Someof the stall divisions were indicatedby depressions made by rows ofwooden stakes. Like other buildingsat Bergþórshvoll this one was builtentirely of turf, but the farm is situat-ed far away from outcrops of stone.An entranceway was on the westerngable end and a possible doorway

45


46

BRUNO BERSON

leading to other, as yet unexcavatedbuildings, could also be seen on theeastern gable end. The roof supportsand some of the divisions between thestalls were made of birch wood. Theeastern part of the floor was level butthe western part sloped considerablyso that the entrance on the westerngable was 0,8 m lower than the east-ern part of the floor. The structuremeasured 14,2 x 4,2 m and was esti-mated to have sheltered about 30heads of cattle. An adjacent floor wasexcavated to the south of the byre,masuring 4x3,3 m and this was inter-preted as a barn which seems to havehad a separate entrance (KristjánEldjárn & Gísli Gestsson 1952).

A few years after the Bergþórshvollexcavation Gísli Gestsson startedarchaeological investigations at Gröfin Öræfi (Gísli Gestsson 1959). Thisregion had been densely settled in theMiddle Ages but was devastated by aneruption in Öræfajökull in 1362, leav-ing only a handful of farms. One ofthe farms which were abandonedcompletely as a result of the eruptionwas Gröf. The ruins were buriedunder a thick layer of pumice andwere remarkably well preserved.Three buildings were excavated, adwelling complex with 6 rooms, akiln-house and a byre with barn situ-ated some 50 m east of the dwelling.

The byre and the barn were builttogether on a northeast-southwestaxis. The entrance was on the southend gable and lead to the byre. This

was divided into three aisles andalong the west wall 5 stalls were stillvisible, divided by 4 upright slabs. Nostall divisions could be seen along theeastern wall. The byre measured 8 x3,70 m on the inside. The aisle in thecentre part was fully paved and leadfrom the entranceway to the barn inthe northern part of the structure. Thebarn was also divided into three aislesby two parallel rows of postpads. Itmeasured 12,6 x 3,8 m on the inside.The floor of the structure sloped some2 m from the northern end of the barnto the entrance of the byre.

The site of Gjáskógar in Þjórsárdalurwas excavated in 1949, 1952 and1960 by Kristján Eldjárn (KristjánEldjárn 1961). It is situated in theupper reaches of the valley at 300 mabove the sea level – high above themain settlement area on the valleyfloor. Like Þórarinsstaðir, Gjáskógarwas a real mountain farm which musthave been snow-bound for manymonths during the winter.

The ruins were in a good state ofpreservation and the excavationbrought to light a dwelling complex, apart of the homefield and a byre. Aseparate barn was not found. Thedwelling complex is of the same typeas commonly found at medieval farmsin Þjórsárdalur. Underneath the farm-house traces of smelting debris werefound, suggesting that this was origi-nally an ironworking site. The byrewas situated 10 m to the east of thedwelling complex. Its inside measure-

47


Fig. 8: Plan of the byre and the barn at Gröf. After Gestsson, 1959.

48

BRUNO BERSON

Fig. 10: Plan of the byre at Gjáskógar. After Eldjárn.

Fig. 9: Layout of the farm at Gjáskógar. After Eldjárn, 1961.

49


Fig. 11: Layout of the farm at Hvítarholt. After Magnússon, 1967.

ments are 10,2 x 3 m. The structurewas divided into three aisles, the cen-tral one paved from the door to themiddle of the room, 6,3 m in length inall. Traces of stall-divisions were seenin the form of fallen slabs on the west-ern side of the paving. The floor of thewhole building sloped considerably,with a 1,25 m differnce in heightbetween the innermost part and theentrance on the southern gable. The

function of the inner half of the build-ing – north of the end of the pavement– was unclear. In the middle of thehouse remains of ironworking werefound, not necessarily coterminuouswith the use of the rest of the buildingas a byre. Like Stöng andÞórarinsstaðir the ruins were full ofpumice from the 1104 eruption ofHekla and the excavator concludedthat the farm had been abandoned as

50

BRUNO BERSON

result of this eruption and had onlybeen occupied for a short periodbefore that. However, a sectionthrough the byre shows that 20-30 smof aeolian soil had accumulated on thefloor before the pumice was deposit-ed, suggesting a similar set-up as atStöng, where the 1104 pumice hasbeen shown to have infilled the aban-doned structure in the 13th century asa result of erosion (Vilhjálmur ÖrnVilhjálmsson 1989).

From 1962 to 1967 Þór Magnússonexcavated the site Hvítarholt inHrunamannahreppur. It is situated ona rise just south of River Hvítá, in theupper parts of the southern lowlands.Some 10 structures were uncovered,among them three longhouses ofViking age type. Five sunken hutswere also excavated, some predatingthe longhouses. The two others build-ings were a byre with a barn (VI) anda supposed barn (II) postdating one ofthe longhouses.

The byre was situated some 30 mfrom the main cluster of buildings atHvítárholt. It measured about 10 x 5m on the inside, but precise measure-ments could not be made as the wallswere very indistinct. They were madeof turf which contained the so-calledLandnám tephra from 871±2. Thebyre was divided into three aisles, thecentral one partly paved with flatstones. Along the walls the stalls hadbeen separated by wooden planks evi-denced by traces of postholes. Theother room was the barn. It measured14 x 5 m and no traces of internalarchitecture could be detected.

Strangely enough a hearth was foundin the middle of the barn, indicatingthat the room had had several phasesof use. The excavator suggested thatthe room had been used as a dwellingduring the construction of the otherlonghouses and was maybe not a barnat all.

On top of one of the longhouses asmaller building had been erected. Ithad three distinct floor layers, all fattywith grass remains embedded in them.This led to the interpretation of thestructure as a barn. It meaured 9,9x3,8m but the north end was indistinct. Inthe earliest floor layer a row of irreg-

Fig. 12: The byre and the barn at Hvítarholt (structure VI). After Magnússon, 1967.

51


Fig. 13: Layout of the farm at Sámsstaðir. After Rafnsson, S., 1972.

ularly placed slabs and stones hadbeen placed along the middle of thesouthern part of the building. Theexcavator considered them to be tooirregular to be a proper pavement butsuggested that they had been laid onthe floor to keep it from getting toomuddy. The floor sloped somewhatfrom north to south but no distinctentrance could be discerned on thesouthern gable. Although this buildingclearly post-dates one of the long-houses it seems nevertheless to be ofViking age date, evidenced by thefinds of steatite sherds, a glass beadand a pig’s tooth (Þór Magnússon1972).

In 1971 and 1972 SveinbjörnRafnsson carried out an excavation atthe farm Sámsstaðir in Þjórsárdalur(Sveinbjörn Rafnsson 1976). The sitehad been partially ecavated earlier, byÞorsteinn Erlingsson in 1895(Þorsteinn Erlingsson 1899). Like theother sites in Þjórsárdalur Sámsstaðirhas been believed to date from theeleventh century, as it was infilled bytephra from the Hekla 1104 eruption,but remains of an older house werediscovered below the main structurethought to date from the tenth century.A bridle fragment from the eleventhcentury was found in the dwellingsupporting an 11th century date forthe site. In light of more recent inves-

52

BRUNO BERSON

tigations at Stöng the dating ofSámsstaðir has however been throwninto question, allowing for the possi-bility that the buildings were in useinto the 12th or even 13th centuries.

The byre at Sámsstaðir was situatedabout 80 m to the east of the dwellingcomplex and was built together withthe barn. The byre had been badlydamaged by erosion but the wallscould be seen to have been made ofturf, lined with stone on the inside.The byre measured 9 x 3,6 m and ispresumed to have sheltered about 20

head of cattle. The room was dividedinto three aisles and sloped towardsthe doorway on the south end gable.On the eastern side 9 slots fromupright slabs separting the stalls wereobserved. The central aisle was fullypaved from the doorway to the barn atthe north of the structure. The barnwas about 4,2 x 3 m and was muchbetter preserved than the byre, withstone lined turf walls standing up to0,6 m.

Between 1967 and 1981 KristjánEldjárn carried out a series of excava-

Fig. 14: The byre and the barn at the site of Sámsstaðir in the Þjórsárdalur. After Rafnsson, S., 1972.

53


Fig. 15: Byre at Goðatættur on the island of Papey. After Eldjárn K., 1981.

tions on the island of Papey, offIceland’s east coast. The projectaimed to verify traditions that theisland had originally been settled byIrish monks but no unambiguoustraces of such a settlement were found(Kristján Eldjárn 1988). The island isonly 2 km² but many ruins of turfhouses were still visible on the sur-face. Among these is the siteGoðatættur, with two buildings. Onewas a typical Viking Age longhouse,aligned north-south. 17 m the south ofit was another building which meas-ured 12 x 4,4 m and followed a north-east-southwest orientation. Theentrance was on the west end gableand a pavement extended from itabout two thirds of the way in to theroom. Two thirds of the structure werethus divided into three aisles and inthis part of the building the floor

sloped towards the doorway. No stallswere visible along the walls. The sur-face in the northern one-third of thebuilding was covered with peatashand a hearth was visible in the middle.A substantial number of artifacts wasfound in the structure which wasinterpreted as a byre which later hadbeen used for human habitation. It ishowever not inconceivable that thenorthern most third of the buildingwas used as a dwelling at the sametime as cows were stalled in thesouthern two-thirds.

An axe of K type dating from thetenth century was found in the byre.Tephra from the 1362 eruption ofÖræfajökull covered the ruins, whichhad clearly been long abandoned andwere completely collapsed before itwas deposited. The excavator con-

54

BRUNO BERSON

cluded that the site had been occupiedin the 10th century.

The last site in Iceland where abyre has been excavated isHerjólfsdalur on the island ofHeimaey in the Vestmanneyjar archi-pelago off Iceland’s south coast. Thesite was ecavated by MargrétHermanns-Auðardóttir in a series ofcampaigns between 1972 and 1983

(Margrét Hermanns-Auðardóttir 1989,1992).

There were 11 structures on the site,among them two byres. Building VIIIwhich is considered to be the earliestat the site combined a byre and ahuman dwelling, measuring 9,8 x 3,4m in all. The byre took up a half ofthis space or 4,5 m, and was divided

Fig. 16: Layout of the farm Herjólfsdalur. After Margrét Hermanns-Auðardóttir 1983.

55


Fig. 17: The two byres at Herjólfsdalur. To the left byre no VIII corresponding to the first period. To theright byre no IV. After Margrét Hermanns-Auðardóttir, 1983.

into three aisles, the central one fullypaved. Some sloping of the flourstowards the entrance on the gable endwas observed. Upto 15 stalls wereidentified, ranging from 0,6 m to 1,6m in width. Stake holes associatedwith the stalls were intepreted as theremains of stakes to tether the ani-mals. The other byre (Building IV)was considered to belong to a laterphase of the site. It measured 8 x 4 m

and was divided into three aisles. Thepaving of the central aisle slopedtoward the door in the northern endgable and continued outside the build-ing. Two rows of stalls were on eitherside of the paving, easily identifiablefrom slots left by the flat slabs sepa-rating the stalls. 16 stalls were identi-fied, measuring between 0,6 and 1,7m in width, and as in Building VIIIstakeholes were found interpreted as

56

BRUNO BERSON

tethering stakes. No barn was associ-ated with this byre but it had a pas-sage on the western wall leading to anadjacent dwelling (Building V).

Finally a byre at the wellknown farmStöng in Þjórsárdalur should be men-tioned (Roussell 1943). Unfortunatelythe byre and the barn had been dam-aged by erosion before the excava-tion, so that the eastern end of both

buildings has disappeared. Unusuallythe barn was built along sode the byre,sharing one long-wall but having itsown entrance. This small complext issituated 20 m east of the dwelling.The byre was 4 m wide and at least 11m long. It was divided into threeaisles, the central one neatly pavedand somewhat lower than the stalls.Five upright slabs were still in evi-dence when the excavation took place

Fig. 18: The layout of the farm at Stöng, after Roussell A., 1939.

57


but upto 20 stalls could be counted, 10on each side. The abandonment ofStöng and other sites in Þjórsárdalurwas originally dated to 1300 butSigurður Þórarinsson’s dating of awhite pumice layer which infilledmost of the ruins in the valley to 1104set the abandonment back some twocenturies. Recent work at Stöng byVilhjálmur Örn Vilhjálmsson has castdoubt on this dating, as the 1104pumice may be redeposited, infillingthe abandoned ruins as a result of ero-sion in the 13th century. The implica-tions of this re-examination of the evi-dence for other Þjórsárdalur sites thanStöng are not yet clear, but it does castdoubt on the dating of the Gjáskógar,Sámsstaðir and Þórarinsstaðir byres.

In addition to these excavated byresthere are a few sites where erosion haslaid bare the foundations of buildingsintepreted as byres by investigators.Reasonably accurate descriptions ofbyres at this kind of site exist forLaugar in Hrunamannaafrétt – closeto Þórarinsstaðir – and Áslákstungainnri and undir Lambhöfða, both inÞjórsárdalur. At Laugar the byre was apart of the farm complex, like nearbyÞórarinsstaðir, set at right angles tothe dwelling and entered from it aswell as from the outside (ÞorsteinnErlingsson 1899). The byre measured12,5 x 4 m on the inside and remainsof some 20 stalls were observed. Aroom interpreted as a barn wasentered from the back of the buildingbut was set at right angles to it. Thereis no dating avidence for this site but

it is generally considered to bemedieval, probably from the sameperiod as Þórarinsstaðir. At Áslák -stunga innri the byre is some 45 mfrom the dwelling. It measures 14 x 4m on the inside and has an entrance onthe southern gable end, Upright slabspartitioning the stalls were still in evi-dence when the site was investigatedin 1895 and it was estimated that therewas room for some 30 heads of cattlein it – no doubt a slight exaggeration(Þorsteinn Erlingsson 1899). At undirLambhöfða the byre was also some 40m from the dwelling. It measured 10 x4 m on the inside and had entranceson each gable end. It is sitated on aslight incline and the entrance on theupper end leads into a very smallroom – too small to be a barn – whichopens to the side into what could havebeen a large enclosure where hay wasstacked. Here also could uprighrtslabs be observed and it was estimat-ed that the byre had room for 18 headsof cattle (Þorsteinn Erlingsson 1899).Both Áslákstunga innri and undirLambhöfða exhibit the same arrange-ment of buildings as the more fullyexcavated sites in Þjórsárdalur, i.e.Stöng, Gjáskógar and Sámsstaðir, andfrom this typological resemblance ithas been inferred that these sites mustalso be from the middle ages, aban-doned either in the 11th or the 13thcenturies.

Limitations of the assemblageOut of some 50 excavated medievalsites in Iceland, byres have only beeninvestigated or observed in 13 cases.

58

BRUNO BERSON

Fig. 19: a) Laugar, b) Áslákstunga innri and c) Undir Lamböfða. After Þorsteinn Erlingsson, 1899.

59


Three of these were not really excava-tions and the byre at Stöng is onlypartially preserved, leaving only 9sites with fully documented byres.This limited number is not an indica-tion that byres were not used as uni-versally as one might think. On thecontrary there is every reason tobelieve that byres were found at everyfarm. The relative scarcity of exavat-ed byres is due mainly to researchconditions. On the one hand investi-gators have always been interestedprimarily in the dwellings and have inmany cases failed to excavate ruinswhich almost certainly are theremains of outhouses. Examples ofthis are found at sites like Skallakot,Ísleifsstaðir, Grelutóttir, Granastaðirand Klaufanes. On the other hand thefrequent arrangement of locating thebyre some distance from thedwellings has meant that the byreseither cannot be found because of aeo-lian deposition or subsequent buildingactivity or they have disappearedthrough erosion. Examples of the for-mer are e.g. Reykjavík, Bessastaðir,Hofstaðir, Viðey, Kúabót and of thelatter Stóraborg is the prime example.

Another limitation of the data isrelated to the uneven distribution ofthe byres. All the excavated byres arefound in the southern part of the coun-try, a large part in a small area inupland Árnessýsla. Again this reflectsthe choices of excavators more thananything else – it reflects the centralposition of Þjórsárdalur in Icelandicarecaheology as well as the state ofcommunications in Iceland for much

of the 20th century which made largescale excavations far from Reykjavíkdifficult or well nigh impossible. Forthis reason the analysis of Icelandicbyres presented here really only holdsfor the southern part of the country.

Plan and shape of the byreIn discussing the byre he had excavat-ed at Gjáskogar, Kristján Eldjárncommented that it had the typicalshape of a medieval Icelandic byre(Kristján Eldjárn 1965). What is thistypical shape?

All the byres know from Icelandshare a range of similarities. They arealways oblong structures ranging inwidth from 3,5 m to 4 m. The onlyexception is Hvítarholt where the byrewas 5 m in width, altough the excava-tor did warn that the line of the wallswas quite indistinct. A division intothree aisles is also a characteristicshared by all the excavated byres. Thecentral aisle is almost always pavedand in at least 5 cases it slopes downtowards the entrance which is alwaysin one of the end gables. This centralpaving allowed for easier removal ofthe dung and urine. On either side ofthe central aisle there are rows ofstalls which in most cases have someevidence of division into stalls, eitherby upright slabs or wooden partitions.In some cases the excavators com-ment that the surface of the pavedcentral aisle was lower than the stalls.Where such evidence survives it seemthat the roof was supported by a dou-ble row of posts situated along bothsides of the central aisle.

While there is remarkable uniformityin the basic shape and construction ofthe byre there are some differences asto its relationship with the barn whichis the building most closely associatedwith the byre.

The most common type of arrange-ment can be seen at Hvítarholt, Gröf,Sámsstaðir and Lundur. At these sitesthe byre and the barn form a longbuilding in which the two rooms arebuilt end to end. There is only onedoor leading to the byre abd the barncan only be accessed through thebyre, except at Lundur where therewas a small room with a side entrancebetween the byre and the barn. A sub-type is found at Þórarinsstaðir andLaugar where the barn is also onlyacessed through the byre but is set atright angles to it. In neither of thosecases is there sufficient evidence toshow that the barn was roofed andthese two sites are the only oneswhere the byres form a part of thedwelling complex. The other maintype is found at Herjólfsdalur,Gjáskógar, Goðatættur, Áslákstungainnri and undir Lambhöfða. At thesesites there is no distinct barn, eitherconnected to or independent off thebyre. At some of these sites, esp.Herjólfsdalur (VIII) and Goðatættur,it cannot be precluded that hay wasstored in that part of the byre wherethe central aisle was not paved. In thecases where no barn has been found itis possible that hay was stored differ-ently, i.e. in turf-covered stacks (OrriVésteinsson 1989). At Stöng the byre

and the barn are built side by side andeach room had its own door facing thefarm. It is possible that there was apassage between the two buildings atthe back – which has now erodedaway. Bergþórshvoll may also belongto this type although the nature of thesmall room to one side of the byre isreally not well established.

Hvítarholt is the only example of aseparate purpose-built barn but it isnot certain whether the barn is reallycintemporaneous to the rest of thsesite, nor indeed whether it is really abarn.

The layout of the farmThere are two basic types of relation-ship between byre and dwelling. Morecommonly the byre is situated somedistance from the dwelling, typically20-40 m and usually slightly upslope.Lundur and Sámsstaðir stand outbecause of the great distance betweenthe dwelling and the byre, 70 and 80m respectively.

In all these cases the byre seems tohave been inside the homefield andone explanation that suggests itselffor this arrangement is that the incon-venience of having to walk a long wayto the byre was offset by the benefitsof not having to carry the manure along distance from the byre to spreadit on the field. In those cases wherethe byre may have been at the edge oreven just outside the homefield it ispossible that keeping the animals’grazing and trampling away from thehomefield may have been a consider-ation.

60

BRUNO BERSON

At Herjólfsdalur and Bergþórshvollthe byres are closely associated withthe dwelling and at Þórarinsstaðir andLaugar the byres actually form a partof the dwelling complexes. The lattersort of arrangement is common onNorse sites in Greenland and was alsoprevalent in many parts of Iceland inearly modern times. At Herjólfsdalurthe byre (IV) and the dwelling (V)were connected by a passage and atBergþórshvoll the byre may have beenconnected to the dwelling complex.The benefits of this arrangement aremainly ease of communicationbetween byre and dwelling, allowingeasy access for milk maids and cow-boys to go about their tasks. It is alsopossible that the preservation of heatfrom the animals played a role, espe-cially if there were second storeys onthese buildings – an eventuality whichthe evidence cannot preclude but doesnot support either.

Related to this may be those byres,both those connected to the dwellinglike Herjólfsdalur VIII and those farremoved from it like Goðatættur andHvítárholt, where there is evidencefor human habitation in one end of thebyre. In none of these cases is the evi-dence for human habitiation extensiveand on all these sites are there othermore substantial dwellings for thehumans. It would therefore be rash tointerpret these houses as combineddwellings and byres of the type com-monly found in Iron and Viking agecontexts in Scandinavia. The similari-ties are however striking and theysuggest a degree of affinity and/or

continuity of a construction heritagebetween the Icelandic settlements ofthe Viking and high middle ages andthe Scandinavian Iron age. It is possi-ble that the phasing of theHerjólfsdalur site could be seen as anindication of a development from acombined byre-dwelling (as in struc-ture VIII) to separate but connectedbyre and dwelling (as in structures IVand V). As most of the sites where thebyre is separated by scores of metresfrom the dwelling are dated to the11th century or even later it might beargued that this represents an evenmore recent development. If that isso, some particular Icelandic condi-tions must have dictated this develop-ment – the importance of manuringinfields being the one that springsmost readily to mind. The exceptions– Þórarinsstaðir and Laugar – wouldthen be explained by their marginallocations and high altitude whichmeant that the farms were snow-bounds for many months every yearmaking it troublesome in extreme tolocate the byres far away from thedwellings.

It must be stressed however thatthis idea of a gradual distancing ofbyres from dwellings in the course ofthe first centuries after the settlementof Iceland is not based on extensive orsound evidence. For most of the sitesthe datings are insecure and in somecases there is dispute over the age to adegree of more than 200 years.

Phases of occupationMany archaeological sites in Iceland

61


62

BRUNO BERSON

and especially farm sites have evi-dence of several periods of occupa-tion. Different phases of dwellings areevidenced by thick floor layers andsigns of repair and rebuilding of indi-vidual houses. Unlike the dwellingsthere are no known cases of differentphases of use of a byre. Is this differ-ence significant?

It is true that in many cases theinattention or lack of interest onbehalf of the excavators may be theexplanation. Many of the excavatedbyres received only the most cursoryinvestigation – in most of theÞjórsárdalur sites the excavators let itsuffice to brush loose sand and tephraoff the byre ruins but did not reallyexcavate the buildings in such a waythat eventual earlier phases wouldhave come to light. It should also bepointed out that the assemblage isbiased towards sites with a shortoccupation. However at sites likeBergþórshvoll and Lundur, both withcontinuous occupation for 1000 yearsor more, the byres seem only to havea single phase.

This might indicate that byres were– unlike the dwellings – not as a rulerebuilt in the same place but a newplace was found for it everytime arebuilding became neccesary. Thismight tie into the idea that the loca-tion of byres some distance from thedwellings was occasioned by the needto spread manure over the homefield.This would make perfect sense if thebyres were moved regularly aroundthe homefield in order to intensivelymanure new parts of the field.

It should also be considered thatbyres were may not have been regularor permanent features of every farm-stead. It is possible that the byres sig-nify periodically intensive dairy farm-ing but that in different times andplaces purpose-built byres were notneccesary because the milch-cowswere few in number and could bestalled in other buildings and the restof the cattle was able to graze outsidefor the majority of the year and/or haddifferent sorts of shelters.

This also raises the question of howthe herds developed. Zooarchaeolog-ical studies at Hofstaðir in Mývatn(McGovern et al. 1998) suggest thatthe numbers of cattle increased rela-tive to sheep and goats in the 10th and11th centuries and also that the over-all the number of sheep and goatbones decreased. These changesshould be reflected in the animal shel-ters – which have as yet no beenfound at Hofstaðir.

Future workIt is clear that the number of excavat-ed byres in Iceland is too small tobase any firm conclusions on and thatmost of the questions posed here needto be further researched. The first taskis to increase the reliability of the databy opening new excavations of byres,not least in the north part of Icelandwhere no byres have yet been investi-gated. Such investigations wouldneed to throw light on all the issuesraised here, i.e. the form of the byre,its relationship with the barn andother buildings, its location in the

63

ReferencesUnpublished references:Berson, Bruno (1998), L’élevage en Islande au

Moyen Age, Master, University of Tours,

Tours.

- (1999) L’élevage en Islande au

Moyen Age, les textes et l’archéologie,

DEA, University of Lille III, Lille.

Published references:Bruun, Daniel (1928) Fortidsminder og

Nutidshjem paa Island, København.

Gísli Gestsson (1959) “Gröf í Öræfum”, Árbók

hins íslenzka fornleifafélags 1959, 5-87.

Kristján Eldjárn (1949) “Eyðibyggð á

Hrunamannaafrétti”, Árbók hins íslenzka

fornleifafélags 1943-48, 1-143.

home-field and its developmentthrough time. In fact the last issuecould be dealt with by small scaleinvestigations of some of the sitesalready excavated, as many of theÞjórsárdalur byres are still accessibleand it would take only minor trench-ing to establish if they are indeed sin-gle phase buildings or not. A fullrerunderstanding of the Þjórsárdalur datacompared to a detailed study of aregion in the North would undoubted-ly go a long way of resolving some ofthe issues raised here, as well as gen-erating new questions.

As mentioned above the detailedanaylisis of faunal assemblages atintensively investigated sites likeHofstaðir pose many questions aboutherd sizes, herd management andgrazing strategies, some of whichcould be illuminated if the animalshelters can be found. The collabora-tion of zooarchaeology and excava-tion of byres is therefore very impor-tant. Zooarchaeological analysis pres-ent data on the relative numbers ofdifferent animal species present in thearchaeological deposits. Such resultsimmediately gain more meaning whenthey can be compared to the buildings

the animals were housed in. AtHerjólfsdalur, which is the onlyIcelandic site which has both byresand an analysed faunal assemblage,the cattle to Sheep/goat ratio wasabout 1:1, indicating a heavy empha-sis on cattle raising, supported also bythe large and alaborate byres. Doesthis mean that those farmsites whichhave large byres were equally dedicat-ed to cattle raising and that at siteslike Hofstaðir where the ratio is morelike one head of cattle to 5 or 6 sheepor goats, the dairy farming was somarginal that we should not expect tofind a purpose-built byre. Or does itonly mean that Hofstaðir had so muchmany more sheep?

In order to better understand theeconomic and social structure ofmedieval Iceland we need to under-stand these sort of questions better.Livestock formed the base of theIcelandic economy and without pay-ing attention to the structural evidencefor its keeping we will not be able toappreciate its role in the economy, andwith out such an understanding ourgrasp of medieval Icelandic societywill remain poor.

64

BRUNO BERSON

- (1961) “Bær í Gjá-skógum í Þjórsárdal”,

Árbók hins íslenzka fornleifafélags 1961, 7-46.

- (1964) “Athugasemd um fornar tóftir á Lundi

í Lundarreykja-dal”, Árbók hins íslenzka

fornleifafélags 1964, 102-110.

- (1965) “Two Medieval Farm Sites in Iceland

and some Remarks on Tephrochronology”,

The Forth Viking Congress, York, August

1961, (Aberdeen University Studies

149):10-19.

- (1988) “Papey, Fornleifarannsóknir 1967-

1981”, Árbók hins íslenzka fornleifafélags

1988, 35-188.

- & Gísli Gestsson (1952) “Rannsóknir á

Bergþórshvoli”, Árbók hins íslenzka forn

leifafélags 1951-52, 5-75.

Margrét Hermanns-Auðardóttir (1989) Islands

tidiga bosåttning, (Studia Archaeologica

Universitatis Umensis 1), Umeå.

Margrét Hermanns-Auðardóttir (1992) “The

beginning of settlement in Iceland from an

archaeological point of view”, Acta Borealia 2

McGovern, Thomas, Mainland, Ingrid &

Amorosi, Thomas (1998) ”Hofstaðir 1996-

1997. A Preliminary Zoo-archaeological

Report” Archaeologia islandica 1, 123-28.

Orri Vésteinsson (1989) “Mygluskán og hálf

blautur ruddi. “Mygluskán og hálfblautur

ruddi. Hvernig geymdu menn hey til

forna?”, Sagnir 10, 18-26.

Roussell, Aage (1943) “Stöng, Þjórsárdalur”, in

Stenberger ed. 1939, 72-97.

Sigurður Vigfússon (1885) ”Rannsókn í

Borgarfirði 1884.” Árbók hins íslenzka forn

leifafélags 1884-1885, 61-138.

Stenberger, Mårten ed. (1939) Forntida gårdar

i Island. Nordiska arkeologiska under

sökningen i Island 1939, København.

Sveinbjörn Rafnsson (1976) “Sámsstaðir í

Þjórsárdal”, Árbók hins íslenzka forn

leifafélags 1976, 39-120.

Vilhjálmur Ö. Vilhjálmsson (1989) ”Stöng og

Þjórsárdalur-bosættelsens ophør.” Hikuin

15, 75-102.

Voionmaa, Jouko (1943) “Lundur”, in

Stenberger ed. 1939, 171-90.

Þorsteinn Erlingsson (1899) Ruins of the Saga

Time: Being an Account of Travels and

Explorations in Iceland in the Summer of

1895, London.

Þór Magnússon (1972) “Sögualdarbyggð í

Hvítárholti”, Árbók hins íslenzka forn-

leifafélags 1972, 580.

a contribution to the study of the medieval ......medieval icelandic literature and it was interest...

Documents