24 Journal of the North Atlantic Special Volume 2 1National Museum of Denmark, Danish Middle Ages and Renaissance, Frederiksholm Kanal 12, DK-1220 Copenhagen K, Denmark. 2Greenland National Museum and Archives, Box 145, GR-3900 Nuuk, Greenland. 3KUAS, The Heritage Agency of Denmark, H.C. Andersens Boulevard 2, DK-1533 Copenhagen V, Denmark. *Corresponding author - jette.arneborg@natmus.dk. Introduction The Hvalsey Fjord farmA has all the elements of a high-status farm. First and foremost is the impressive church, but there is also the feasting hall, where the farmer entertained his guests, the store house by the shore where commodities for foreign trade were kept, and the large byres for the prestigious cattle. Still, compared to high-status farms such as the Bishop’s see Garðar in Igaliku and Brattahlið in Qassiarsuk (Arneborg n.d.), the Hvalsey Fjord farm is small in terms of the number of ruins, and that may have been the reason why the Danish architect Aage Roussell in 1935 suggested that the neighbouring group of ruins at Ø83a must have been a part of the Hvalsey farm, calling them “the dairy farm,”B while the main cluster he accordingly called the “home farm.” In 2004, in connection with new surveys at the Hvalsey Fjord farm and church (Ø83), the “the dairy farm” (Ø83a) was also surveyed briefl y, and four small test pits were made in two of the ruins. The primary aim was to assess the state of preservation of the ruins and to recover samples for AMS radiocarbon dating.C In the following, we report on the 2004 test excavations at the “dairy farm” and discuss Roussell’s conception of a “home farm” and “dairy farm.” The Site The site Ø83a is ca. one kilometer east of the Hvalsey Fjord farm and church, and close to the entrance of the inlet Tasiusaq (Fig. 1). Today, the site has been affected by the activities of a sheep-farm in the same location. When Roussell (1935) was there in 1935, he described the site as one of the most luxuriant he had ever seen in Greenland. And it was one of the largest as well. The farm buildings were built on small dry rises on the north side of two small lakes, and with one exception, all faced the lakes. Research History In total, Roussell (1935) recorded eight Norse ruins (numbered 17–24), three outdoor cooking pits, an Inuit naanngisat (“hopping stones”), and “numerous” Inuit ruins on the site (Fig. 2). Already in 1964, ruin 24 and the Inuit naanngisat had disappeared (Thorvildsen 1964), and in 2004, we were not able to fi nd the ruins 18, 19, and 23. The area of the three last-mentioned ruins has been cultivated recently, and the ruins very likely have been removed in this connection. In addition to our work in 2004, Vésteinsson (2008:24–26) described the site in 2005. In 1935, Roussell excavated in ruins 20, 21, and 22. He reports that they excavated “quite a lot” in ruin no. 20 to fi nd out if it was the dwelling of the farm. According to Roussell, the walls were made of rather small stones and solid turf walls, and the conclusion was that the eastern part of the ruin might have been that of a small, separate house. The western part of the house was a byre—a stone slab used as a stall partition was still standing in situ—and on the fl oor there was evidence for drainage. In front of the house, on the south side, there was a fairly large area paved with fl agstones. Apart from the above mentioned, Roussell could not establish room separations, and he concluded that the house was a byre with a barn. In a minor “test pit” in the “lower end” of ruin 21, Roussell unearthed a fl agged fl oor, and the same was the case in ruin 22, were he dug “some test pits.” Immediately north of ruin 20, two large cooking pits were excavated. After excavation, the most southerly and the larger of the two was—according to Roussell—the “nicest,” and was therefore reported in detail. The pit was almost circular with a diameter of about 140 cm at the rim. The diameter at the fl at base was about 65 cm. The sides were straight and partly lined with fl at slabs. The bottom of the pit was The “Dairy Farm” of the Hvalsey Fjord Farm Jette Arneborg1,*, Fuuja Larsen2, and Niels-Christian Clemmensen3 Abstract - Based on his archaeological investigations in 1935 at the Hvalsey Fjord farm and church (ruin group Ø83) and at the neighbouring ruins (ruin group Ø83a), the Danish architect Aage Roussell linked the two sites together, stating that the Ø83a buildings had functioned as a subsidiary “dairy farm” to the main farm at Hvalsey (Ø83). Excavations in 2004 showed that the buildings at Ø83a had been in use for a short period only, and artifacts found in 1935 together with the layout and design of the houses point to a date in the early period of the Norse settlement. Roussell did not identify a dwelling at Ø83a, but we argue that one of the houses at Ø83a could have been a dwelling and that the two sites—the Hvalsey Fjord farm (Ø83) and Ø83a—were two individual farms. We suggest that the Ø83a-farm most probably was the predecessor of the Ø83 Hvalsey farm. For unknown reasons, the earlier farm was moved to the Ø83 site. 2009 Special Volume 2:24–29 Norse Greenland: Selected Papers from the Hvalsey Conference 2008 Journal of the North Atlantic 2009 J. Arneborg, F. Larsen, and N.-C. Clemmensen 25 covered with a 5 cm thick layer of charcoal. Apart from the charcoal, there were no fi nds in the pits. A smaller pit south of ruin 21 also contained a few animal bones. Because he could not identify a dwelling, Roussell concluded that the site could not have been an independent farm. The 1935 excavations produced only eight artifacts:D A spindle-whorl of soapstone (Fig. 3), a line-sinker for fi shing (Fig. 4), a pad of soapstone for repairing soapstone vessels, and a sherd from a soapstone vessel. In addition, a few small pieces of iron nails were collected. The sinker is of soapstone and has vertical groves for the fi shing line. Linesinkers are rare in the Norse Greenlandic context, with only two others known. A similar weight was found in the hall (room I) in the living house of ruin group Ø29a at Eric the Red’s Brattahlið in Qassiarsuk (Danish National Museum register), and another one was found in the midden at the supposed nunnery, ruin group Ø149, at Narsarsuaq in Uunartoq Fjord (Vebæk 1991a:79, fi nd no. 179; Vebæk 1991b:13). The 2004 Excavations In 2004, four small trenches were dug, one trench (unit 20.1) in ruin 20, and three trenches (units 22.1–3) in ruin 22E (Fig. 5). In ruin 20, a 1-m wide and 4-m long trench was cut through the northern wall in the central part of the building. The wall was ca. 1.2 m thick and built of two rows of medium-sized stones with turf-blocks in-between. In this location, Roussell had very clearly excavated the inside of the house down to natural subsoil, leaving no undisturbed cultural layers. Outside the wall, there was a ca. 30-cm thick layer of turf collapse with very few stones. The subsoil consisted of coarse grey to yellow gravel, and on top of that and under both the wall and the collapse layer there was a thin, unbroken vegetation layer representing what seems to have been the original surface. No signs of activity on top of the vegetation layer outside the house were identifi ed. Three small trenches were placed in ruin 22. Unit 22.1 (1 x 4 m) was dug inside the building up against the north wall. As in ruin 20, the ca. 1.5-m thick wall was constructed of two rows of medium-sized stones with an inner core consisting of a mixture of subsoil (coarse grey to yellow gravel) and bits and pieces of fossil vegetation layer. The level of the fl agged fl oor was ca. 30 cm below the base of the wall, and evidently the house had been dug into the slope, and the material from the inside of the house used Figure 1. The central part of the Norse Eastern Settlement. to construct the walls. 26 Journal of the North Atlantic Special Volume 2 Unit 22.2 in ruin 22 cut through the ca. 1.5-m thick south wall. The trench was 1 x 4 m. Here too, the wall was constructed with a frame of mediumsized stones. The core of the wall was a mixture of soil (fossil vegetation layer and subsoil) and ca. 25 x 25 cm large stones. The fl oor of the room had been paved with large fl ags. Unit 22.3 was located in the eastern part of ruin 22. It was 1 x 3 m and touched the inner side of both the northern and the southern walls. The house had been Figure 2. Aage Roussell’s survey plan from 1935. Ruin 17: ca. 10 m long and 4 m wide. Inside measurements. The house gable towards the fjord is open. Roussell identifi es the ruin as a boathouse. Ruin 18: ca. 11 m long and 6.5 m wide. Roussell describes the ruin as very indistinct. Not found again in 2004. Ruin 19: ca. 7 m long and 5 m wide. Not found again in 2004. Ruin 20: ca. 50 m long and 6 m wide. Exterior measurements. The eastern part of the ruin may represent a small separate square building. The western part of the building was a byre with a stall partition of stone still standing and signs of a drain in the fl oor. Ruin 21: ca. 18 m long and 5.5 m wide. Rather indistinct. Possibly more than one room. In 2004 a cultivated fi eld touches the southwest corner of the building. Ruin 22: ca. 20 m long and 5 m wide. Exterior measurements. Three rooms. Test excavations revealed fl agged fl oors. Ruin 23: ca. 11 m long and 5 m wide. Two rooms. Ruin 24: ca. 15 m long and 13 m wide. Exterior measurements. A pen/enclosure, probably with a small building inside it. Situated in grassy meadow. Not found again in 1964. Just north of ruin 20 there were two pits identifi ed by Roussell as cooking pits. Southwest of ruin 24 was an Inuit naanngisat 24 m long with 26 stones still in situ and distinct marks after another 14 stones. Not found again in 1964. 2009 J. Arneborg, F. Larsen, and N.-C. Clemmensen 27 fl agged and seems to have been rather narrow. The distance between the two walls was only about 2 m. In summary, even though nothing was left of the interior, our profi le in ruin 20 showed that the house has one construction phase only and the same applies to ruin 22. There were no indications of the houses having been in use for long. Dating Apart from tiny pieces of charcoal—insuffi cient for AMS-dating—we unfortunately did not recover anything in our trenches which could be used for dating. We therefore have to pin down the date of the site on the basis of the layout and state of the buildings and the archaeological fi nds from 1935. The houses at Ø83a all belong to the group that Roussell classifi ed as “the long-house” and which he dated to the fi rst period of settlement in Greenland (Roussell 1941:202ff.). In spite of deserved criticism of his house typology (Albrethsen 1982; Andreasen 1981), Roussell’s dating of this specifi c group of buildings still stands, and since Roussell brought forward his typology, at least two new houses—the dwelling at GUS and the dwelling at the so-called landnam farm in Narsaq, ruin group Ø17a—can be included in the group to give further evidence to the early dating of the “long-house” (Albrethsen and Ólafsson 1998, Arneborg et al. 1998, Vebæk 1993). The construction of the walls of the Ø83a houses is comparable to the Narsaq (Ø17a) dwelling. Here, the sill consisted of a frame of stones with an inside fi lling of soil and turf. On the sill were courses of “… alternate layers of stone and turf.” (Vebæk 1993:14 and fi g. 9). The fl oors were partly covered with slabs (Vebæk 1993:14–15). The Narsaq house is radiocarbon dated within the period AD 885–1155 (± 1 s.d.). At GUS, the frame of walls was entirely of turf with an inside fi lling of soil and turf fragments (Albrethsen and Ólafsson 1998:19). The GUS house is radiocarbon dated within the period AD 1020–1280 (±1 s.d.) (Arneborg et al. 1998:27). Outdoor cooking pits are known from Viking Age farms in Iceland (Fornleifastofnun Íslands Annual Report 2005:28, 2006:34f; Hermanns- Auðardóttir 1989:107). In Greenland, in addition to Ø83a, a large (5 m diameter, 2 m depth) outdoor cooking pit has been recorded at the bishop’s farm Gardar in Igaliku. The cooking pit is located in one of the large enclosures of the farm and is impossible to date (Nørlund 1929:89). Among the artifacts from 1935, the line sinker also points to an early date. The Narsarsuaq line sinker cannot be dated. However, having been found in the hall, the Brattahlið sinker may very well belong to the early settlement period (Arneborg, Saga trails: 25). Among others, at Jarlshof, Shetland several comparable sandstone weights have been found in one of the houses assigned to the Viking Age (Hamilton 1956:160, 182; Steane and Foreman 1988:152), and at Bornais, Outer Hebrides, stone weights have been found in an early 11th-century house (Sharples 2004:267). In Iceland, a line sinker of lava was found in the Ísleifsstaðir hall (Stenberger 1943:168f), which Vésteinsson (2006:118) dates to the mid- or late 10th century. Having failed to identify a dwelling, Roussell categorized the Ø83a site as a “dairy-farm” belonging to the Hvalsey Fjord farm (Ø83). By using the term “dairy-farm,” Roussell perhaps had in mind a shieling or some sort of pasture/summer farm. However, in Danish, he used the word “avlsgård” (Roussell 1935), which implies that it was at Ø83a Figure 3. Spindle whorl of soapstone found 1935 at ruin site Ø83a. Diameter: 3.9 cm. Photograph © National Museum of Denmark. Figure 4. Line sinker of soapstone found 1935 at ruin site Ø83a. Length: 13.9 cm. Photograph © National Museum of Denmark. 28 Journal of the North Atlantic Special Volume 2 that the actual farming took place, away from the living quarters and main buildings at the Hvalsey farm (Ø83). This assumption is at least partly contradicted by the byres, stables, and barns at the Hvalsey farm (Ø83) Roussell himself described. Vésteinsson (2008:25) suggests that the large ruin 20 at Ø83a may have included both dwelling and stables. In Norse Greenland, byres/stables/barns have not been found in the same buildings as dwellings, but often in separate buildings lying very close to each other (e.g., Vebæk 1992:30, fi g. 31), and the same may have been the case here. Roussell reported that the eastern part of ruin 20 seemed to be that of a separate house, though small. However, we agree with Vésteinsson that due to the limited build-up of cultural deposits and the poor preservation of organic materials, Roussell most probably missed the dwelling of the farm. “Dairy-farms” (i.e., avlsgårde) are not known from other sites in Norse Greenland, and the size and number of the houses, the cooking pits, and the presence of a byre also point towards permanent habitation. Therefore, despite the lack of conclusive evidence, the possibility cannot be excluded that the ruins 21 and/or 22 also may have been dwellings. Consequently, the Ø83a farm should be considered no less a farm than other Greenlandic farms. By linking the two sites Ø83 and Ø83a, Roussell implicitly claimed that they had functioned at the same time. The archaeological excavations, however, show that the Ø83a-site was short-lived and that it most probably belongs to the early period of settlement, in contrast to the Hvalsey Fjord farm (Ø83), where the recorded ruins seem to belong to fi nal period of settlement (e.g., Roussell 1941:240f). Still, the question remains about the relationship between the two sites. Right now, we have no indications about when the farm at Hvalsey Fjord (Ø83) was established. Both sites may have been settled in the landnám period, with the Hvalsey farm (Ø83) prospering at the expense of the other (Ø83a), leading to its early abandonment. The atypically close proximity between the two sites, however, suggests that Vésteinsson is correct when he proposes that the Ø83a farm may have been a precursor to the Hvalsey Fjord farm Ø83 (Vésteinsson 2008:25), and that the farm for yet unknown reasons was moved from the secluded site on the inlet Tasiusaq to the more prominent and spectacular site at the head of fjord nowadays known as Qaqortup Imaa. Conclusion The plan of site Ø83a, the layout of its buildings, and the results of minor archaeological excavations there in 1935 and 2004 suggest that the site was established in the early period of Norse settlement in Greenland and for a rather short period Figure 5. Ruin group Ø83a (inset; aerial image © 2009 Digital Globe) and survey plan 2004. only. Having failed to Measured by N.-C. Clemmensen. 2009 J. Arneborg, F. Larsen, and N.-C. Clemmensen 29 identify a dwelling in his excavation in 1935, Aage Roussell coupled the buildings at Ø83a with the nearby Hvalsey Fjord farm (site Ø83), arguing that the buildings at Ø83a were the economy buildings— a dairy farm or avlsgård—of the “manor” at Ø83. On presently available evidence, it is not possible to establish the chronological connection between the two sites, and we argue for the possibility that the two sites constitute two independent farms that were either established simultaneously in the early period of settlement or—more probably—that the old farm at Ø83a was the landnám farm and predecessor of the later and better-known Hvalsey Fjord farm. Acknowledgments The excavations at Ø83a were initiated by the former head of the Qaqortoq museum, now curator at the Greenland National Museum and Archives in Nuuk, Georg Nyegaard. The Commission for Scientifi c Research in Greenland funded the excavations. Together with the authors, undergraduate Christian Thomsen, Department of Medieval and Renaissance Archaeology Aarhus University, also participated in the archaeological investigations at Ø83a. We all thank sheep farmer Kalistaaraq Karlsen for his kind hospitality. Literature Cited Albrethsen, S.E. 1982. Træk af den norrøne gårds udvikling på Grønland. Vestnordisk byggeskikk gjennom to tusen år. AmS-Skrifter 7, Stavanger:269–287. English summary. Albrethsen, S.E., and G. Ólafsson. 1998. A viking age hall. Pp. 19–26, In J. Arneborg and H.C. Gulløv (Eds.). Man, Culture and Environment in Ancient Greenland. The Danish National Museum and Danish Polar Center, Copenhagen, Denmark. Andreasen, C. 1981. Langhus - ganghus - centraliseret gård. Hikuin 7, Århus:179–184. Arneborg, J. n.d. Saga Trails. Nanortalik Museum, Narsaq Museum, Qaqortoq Museum, and National Museum of Denmark, Copenhagen, Denmark. Arneborg, J., J. Heinemeier, N. Rud, and A. Sveinbjörnsdóttir. 1998. AMS dates from the hall (XVII). Pp. 27–30, In J. Arneborg and H.C. Gulløv (Eds.). Man, Culture and Environment in Ancient Greenland. The Danish National Museum and Danish Polar Center, Copenhagen, Denmark. Fornleifastofnun Íslands Annual Report. 2005. Reykjavík, Iceland. Fornleifastofnun Íslands Annual Report. 2006. Reykjavík, Iceland. Hamilton, J.R.C. 1956. Excavations at Jarlshof, Shetland. Archaeological Report No. 1. Ministry of Works, Edinburgh, Scotland, UK. Hermanns-Auðardóttir, M. 1989. Islands Tidliga Bosätning. Studia Archaeologica Universitatis Umensis 1. Umeå Universitet. Arkeologiska Institutionen, Umeå, Sweden. Nørlund, P. 1929. Norse Ruin at Gardar. Meddelelser om Grønland Vol. LXXVI. Copenhagen, Denmark. Roussell, Aa. 1935. Hvalsey, avlsgaarden, 1. Unpublished report in the archives Antikvarisk-Topografi sk Arkiv. National Museum of Denmark, Copenhagen, Denmark. Roussell, Aa. 1941. Farms and Churches in the Medieval Norse Settlements of Greenland. Meddelelser om Grønland vol. 89(1). Copenhagen, Denmark. Sharples, N. 2004. A fi nd of Ringerike art from Bornais in the Outer Hebrides. Pp. 255–272, In Hines, J., A. Lane, and M. Redknap (Eds.). Land, Sea, and Home. Society for Medieval Archaeology Monograph 20. Steane, J.M., and M. Foreman. 1988. Medieval fi shing tackle. Pp. 137–186, In M. Aston (Ed.). Medieval Fish, Fisheries, and Fishponds in England. BAR British Series 182(i). Stenberger, M. 1943. Ísleifsstaðir, Borgarfjarðarsýsla. Pp. 145–170, In M. Stenberger (Ed.). Forntida Gårdar i Island. Munksgaard, Copenhagen, Denmark. Thorvildsen, K. 1964. Grønlandsrejse 24/6–8/8 1964. Diary in the archives Antikvarisk-Topografi sk Arkiv. National Museum of Denmark, Copenhagen, Denmark. Vebæk, C.L. 1991a. The Church Topography of the Eastern Settlement and the Benedictine Convent in Uunartoq Fjord. Meddelelser om Grønland—Man and Society 14. Copenhagen, Denmark. Vebæk, C.L. 1991b. Hunting on land and at sea and fi shing in Medieval Norse Greenland. Acta Borealia 1-1991:5–14 Vebæk, C.L. 1992. Vatnahverfi . An inland district of the Eastern Settlement in Greenland. Meddelelser om Grønland— Man and Society 17, Copenhagen, Denmark. Vebæk, C.L. 1993, Narsaq: A Norse landnáma farm. Meddelser om Grønland. Man and Society 18, Copenhagen, Denmark. Vésteinsson, O. 2006. The Building and its context. Reykjavík 871 ± 2. Landnámssýningin. The Settlement Exhibition. Reykjavík City Museum:116–121 Vésteinsson, O. 2008. Archaeological investigations in Hvalseyjarfjörður, Eystribyggð 2005. Fornleifastofnun Ìslands. FS388-0501. Reykjavík, Iceland Endnotes ADanish National Museum system ruin group Ø83, Greenland National Museum and Archives system ancient monument site 60V2-IV-646. BDanish National Museum system ruin group Ø83a, Greenland National Museum and Archives system ancient monument site 60V2-IV-645. CGreenland National Museum and Archives fi le no. KNK 2675. DNational Museum of Denmark inventory no. D12691.1-5. EFuuja Larsen, The Greenland National Museum and Archives was the head of the excavations.