Epaves Bay, Great Northern Peninsula, Newfoundland, Canada

L'Anse aux Meadows

Early 11th century (a 2021 study's best fit is the year 1021)

A long sod-roofed hall with timber door frames and white smoke hoods, beside a gravel path, with the sea on the horizon. — A replica hall at the historic site. The photographer identifies it as Hall A of the A-B-C complex, built north of the buried ruins. The turf and the timber door are the reconstruction. Wallace uses full-scale replicas here when she estimates how soon a sod hall would need repair. TravelingOtter, CC BY-SA 2.0. Shown unedited.

George Decker, a local man, leads Helge Ingstad to a line of grassy ridges at the northern tip of Newfoundland. Digging from 1961 to 1968 turns the ridges into sod buildings, a furnace, and a handful of Norse objects. A study published in 2021 dates three cut trees to a best-fit year of 1021. The open question is what those buildings were for, and what name, if any, the sagas already had for them.

Why this is extraordinary

In 1960 a Norwegian writer, Helge Ingstad, was searching the coast from New England northward for a place where Norse sailors might have landed. Parks Canada's culture page says that at L'Anse aux Meadows a local man, George Decker, led him to a group of overgrown bumps and ridges. Ingstad's wife, the archaeologist Anne Stine Ingstad, directed the digging that followed. The same page says the two of them led an international team, from Norway, Iceland, Sweden, and the United States, through the next eight years. A later brochure counts seven expeditions between 1961 and 1968, and adds Canadian archaeologists to the roll.

What those seasons recorded, in the culture page's account, were the lower courses of eight buildings. Walls and roofs had been sod over a timber frame, of the kind used in Iceland and Greenland around the year 1000. Out of the rooms came a bronze ring-headed pin, a stone the page calls an oil lamp, a soapstone spindle whorl, a bone needle, a small gilded brass fragment, a whetstone, a great deal of iron slag, and a large number of iron boat nails. Parks Canada's later map of the same ruins sorts the eight into three halls, smaller huts, and a smelting hut across the brook, and it reads one lean-to as a place where a boat was repaired. Birgitta Linderoth Wallace gives the halls their sizes, from the Parks Canada excavations that followed the Ingstad seasons. A review of the Ingstads' book dates that later digging to 1973–1976. Wallace reads the whole terrace as a winter base.

A 2021 study in Nature, Wallace among the authors, reports a tighter clock. Cosmic rays in the year 993 left a spike of radiocarbon in the air, and trees took that spike into a single ring. In three pieces of wood from this terrace, cut with metal tools and still carrying the outer growth edge, the study's best match places the felling in 1021. Other years still pass the study's own test for one piece or another. The study argues that the agreement of three trees on one year is the surprising part.

What the place was is a separate argument, and it is more than one argument. Wallace argues for a winter base and a gateway, the sagas' Straumfjord, with grapes and lumber gathered farther south, where butternuts grow. Helge Ingstad's reading, as a review of the Ingstads' book states it, is that this was Leifsbúðir, Leif Eriksson's camp, and the main settlement. A sun sentence in the saga of the Greenlanders has been turned into latitudes that sit south of this terrace. How many winters people stayed, and why the halls were left, are published on more than one side. So is a second Newfoundland point, at Point Rosee in the Codroy Valley, tested in 2015 and 2016 and written up on 8 November 2017. The find lists for this terrace do not include a rune. Stones read as runes, farther south and west, are weighed in the Runestones in America room.

Why it matters

If the terrace is Leif Eriksson's camp, or Vinland itself, the sagas and the ruins are the same story, and the grapes in the text have to be explained at this latitude or explained away. If it is a gateway and a winter base, Vinland is a wider coast, and this is the door. If the stay was a few years, Greenland's small population is part of the reason the door closed. If it was longer, the thin middens have to be read again. A second site at Point Rosee would have changed the map. The 2017 report is that side's later account of the same ground.

The pin is the cluster of hall remains above Epaves Bay, on the Strait of Belle Isle. UNESCO describes eight timber-framed turf buildings, three of them dwellings, one a forge, and four workshops. Wallace groups the same eight into three dwelling complexes and an iron hut. The Greenland settlement the sagas treat as the starting point, and the century in which its writing stops, are in the Hvalsey room. The rune stones are a different question, in Runestones in America.

Photo tour

5 photographs, unedited. On a phone, swipe the row. Choose one to look closely.

A wide green terrace and a curving bay, with a rocky island offshore, paths through the grass, and low buildings along the far shore. — The terrace and the bay on 27 August 2022. The photographer's description names Hay Cove to the left and Norstead, a separate modern reconstruction, off to the right. The buried halls are on this ground. Their outlines are not a readable plan in the frame. UNESCO says the archaeological remains were reburied to protect them.Gordon Leggett, CC BY-SA 4.0

A long wall of stacked turf under a grass roof, with wooden smoke hoods along the ridge and poles leaning by a path. — Turf courses and smoke hoods on a reconstructed building. The photographer's note says these are recreations of the halls dug in the 1960s and then covered. The stacked sod is the feature Wallace and the Ingstad reports both start from. The hoods and the path are the modern site.Eric Titcombe, CC BY 2.0

A grass-covered hall behind a wooden picket fence, with the sea and low hills behind it and a person standing in the doorway. — A replica hall, a fenced yard, and the sea. The person in the doorway is part of the historic site's presentation. Wallace reports that archaeological tests found no livestock enclosures and no sign of grazing. The fence in the photograph is the reconstruction's.D. Gordon E. Robertson, CC BY-SA 3.0

Inside a timber room, a person in a green tunic sits on a raised wooden platform playing a wooden flute, with a red shield, horns, and pots around him. — The platform along the wall is the kind of bench Wallace describes for sleeping and sitting. The clothes, the shield, the horns, and the pots are the reconstruction's furnishings. Her reading of the departure is that household inventory was taken away.D. Gordon E. Robertson, CC BY-SA 3.0

No openly licensed photograph of the bronze pin, the spindle whorl, the iron nails, the butternuts, or the butternut burl was found. Those pictures are Parks Canada's.
No unedited Commons photograph of the 1961 to 1968 excavation, with the walls still open, is used. UNESCO says the remains were reburied.
A Commons photograph of a modern wooden boat inside a shed at the site was left out, so a display hull would not be read as an excavated ship. Wallace's boat reading rests on cut nails and on a floor plank, not on a surviving hull.
No openly licensed photograph of the Point Rosee trenches is used. The 2017 report is the account of that ground.

See the evidence at the source

These pictures are not openly licensed, so they are not copied onto this page. Each link opens in a new tab.

Parks Canada's photographs of the ringed pin and the siteThe culture page shows the bronze pin. The licence on that photograph is Parks Canada's, so the picture stays on their page.

What is still unexplained

These sit beside the consensus account, with the same visual weight. None of them is marked solved.

The plans agree, the uses are readings Eight buildings, and what each room was for UNESCO's listing describes a complete 11th-century settlement of eight turf structures on a narrow terrace above a peat bog and a small brook, at the shore of Epaves Bay. Parks Canada's site map names them. Hall A is a large hall with a private room and shared quarters, one room littered with slag. House B is lower-status housing and a place for roasting bog ore. Hut C is the smallest housing, "serving perhaps as quarters for several slaves." Hall D is the smallest of the three halls, with a storeroom and a room of woodworking debris. Hut E is read as a women's workshop, and the stone weights inside "may have been part of a loom." Hall F is the leader's hall, the largest, with a private room, a central room the map calls a banquet hall, two storerooms, and a lean-to for boat repair. Hut G is a workshop and housing for a lower-status crew. Hut J, across the brook, is the smelting hut. Wallace's measurements, from the same walls, are these. Hall A is 29 metres on the outside and about 102 square metres inside, which she sets beside Erik the Red's last house in Iceland, about 16 metres and about 50 square metres. Hall F has about 160 square metres of floor, about 80 of them for living, six rooms and a lean-to. Hall D is 20.2 metres outside and about 88 square metres, with a living room of about 62. House B is about 15 square metres. From the sleeping platforms she estimates room for 60 to 78 people in the halls, two or three in house B, and five to 11 more in huts C, E, and G, something like 70 to 90 people in all. The 2021 study's supplement uses that same 70 to 90, and describes hall walls 1.5 to 2 metres thick. Parks Canada's map makes a different size comparison for hall F. It says the hall is twice the size of Erik the Red's home in Greenland. The two comparisons are not the same house. A few objects have two published names. The culture page describes a stone oil lamp. Wallace describes the same kind of beach stone, found at a doorway in hall F, with no char in the hollow, and she reads it as the socket for a door post. Anne Stine Ingstad's 1977 report had called it a lamp. A shoulder bone first published as a domestic pig is, in later identifications Wallace cites, a seal. A small building dug into the old shoreline was published as a sauna. Wallace argues the fireplace is the wrong kind, and that a slate feature associated with that reading has a radiocarbon intercept of 1170 plus or minus 70, which she places after the Norse layer. The stones in hut E are loom weights on the site map. Wallace leaves them as weights that could have held a loom or a fishing net, and she notes birch-bark rolls outside the hut of the kind used to wrap net sinkers. The jasper fire-starters are Wallace's clue to where people had come from. She counts nine worn pieces. Four are from western Greenland, all of them in the hall F middens. Three are from western or southwestern Iceland, and two from southern Borgarfjord in western Iceland, both of those from hall D. Of the Icelandic pieces besides Borgarfjord, two lay in the A-B-C complex and one in hall F's long fire. Parks Canada's map says the same thing in fewer words. Icelandic jasper in A and D, Greenlandic jasper in F. A bronze pin about 10 centimetres long, of a type Wallace calls Hiberno-Norse and dates as especially common around 950, within a span of about 920 to 1050, is in a pit. The culture page puts a ring-headed pin in the cooking pit of a large dwelling. Wallace puts this pin in a forge pit in hall D. A soapstone spindle whorl of 16.9 grams, cut from a sooted cook-pot, and a small needle hone lay in the lean-to of hall F. A bone needle with a drilled head lay in hall D's long fire. Wallace reads the textile tools as women's work, and the buildings as a camp that was mostly men. The culture page draws the same conclusion from the whorl and the needle, and says the settlers were not all men. A clear glass bead a little over a centimetre across was recorded outside hall D. Wallace writes that it was lost again in a Parks Canada laboratory, and that photographs are what remain.

Repair, or a new small boat A furnace, cut nails, and a lean-to Hut J is the clearest building on the terrace, because the waste explains it. Wallace describes a small stone shaft set over a pit, the gaps chinked with clay, charged with layers of bog ore and charcoal. She gives the heat needed to win iron from that ore as at least 1100 to 1250 degrees. The ore forms along the brook. A charcoal pit stood nearby. Smelting slag sits around the furnace and down the slope. Smithing slag sits in room III of hall A, which she reads as the smith's place. Parks Canada's map says the amount and kind of slag suggest a single smelt, and very little iron, enough for about 100 to 200 nails. Wallace's phrase for the yield is a few kilograms, made once, then worked. The nails are the other half of the reading. Wallace counts 99 fragments and one whole nail, nearly all the iron from the site. About 66 percent of the fragments lay in the F-G complex, with a concentration in room VI, the lean-to on the inland side of hall F. She reports that X-ray images show the nails cut the way a boat nail is cut, with a chisel between the planks, and that some tips were snipped as a new nail was clenched. On 65 Norwegian farm sites, she notes, excavators found 15 nails in total, because nails were for boats and for the most expensive buildings, not for ordinary sod halls. The one whole nail, unused, came from carpentry waste in the bog and is higher in phosphorus than the others. The ore and the slag on the site are also high in phosphorus, so she allows that this one nail may have been made here. Boat repair is the reading on the site map and in UNESCO's sentence about woodworking "likely used for ship repair." Wallace also records a second reading, from a Norwegian boat historian, Olaf Engvig. In clinker work it is easier to build a new boat than to replace a single plank. Irregular post traces in room VI, she writes, could be the remains of a boat-building frame, and any boat shaped there would have been small, no longer than eight metres, a craft for landings rather than the ocean crossing. She thinks the crossing itself took larger ships, probably at least three, one for each dwelling complex. That number of ships is her inference from the three complexes. It is not a count of keels in the bog. The bog beside hall D kept the carpentry. Wallace describes post ends, chips, a barrel lid, a floor plank from a small boat, an auger bow, a birch-bark cup, and treenails. Most of the wood is local. Balsam fir, pine, larch, birch, alder. Two pieces are Scots pine, a European tree she notes was introduced to North America after 1500. A carved finial, and the peg in the boat plank, are that wood. She reads them as objects made in Europe and carried over. The waste is thickest outside the D-E complex, which is why she puts the woodworking there, and why the site map puts wood debris in hall D.

The nuts are here, the route is a reading Three butternuts, and a coast farther south In the carpentry waste Wallace records three butternuts and a burl of butternut wood cut with a metal knife. The tree is Juglans cinerea, white walnut. She writes that it has never grown in Newfoundland. The northern edge she gives is about 47 degrees north, in the inner Miramichi of northeastern New Brunswick, along the Saint John River, and in the St. Lawrence valley west of Baie St. Paul, Quebec. Wild grapes, on her account, grow in those same northern stands. UNESCO's listing says the artifacts include indications that the people who used the camp voyaged farther south. The 2021 supplement says butternut was found much farther south in the Americas and is not a Newfoundland tree. Wallace's reading of the nuts is a journey. The terrace is the base. The south shore of the Gulf of St. Lawrence, and the Miramichi in particular, is Hóp, the sagas' summer camp, where lumber and grapes were gathered and brought back. She quotes a sixteenth-century description of grapevines climbing poplars and elms, and she reads the sagas' "grape trees" as those trunks, felled and shipped, not as vineyard stakes. The nuts, for her, are the reason the wine in the name can be wine. A page in the Great Unsolved Mysteries Vinland project, arguing the same case, says the name in Flateyjarbók is spelled with a double i, Vijnland, and that a meadow name would have been built as Vinjaland or Vinjarland. That page treats the long í as wine. The other reading of the name is the short-i word for a meadow, which the same page describes and then sets aside. The 1970s brochure tells Ingstad's search this way. He treated the grapes in the saga as possibly large red berries, and he argued that the sailing times reach northern Newfoundland and not the grape coasts farther south. The nuts do not, by themselves, name the beach where they were picked, or say whether the people who cut the burl had stood under the tree. Wallace argues they did. A second excavated Norse camp at the butternut's northern edge is not part of the published record. The nuts are the southern evidence this terrace has.

1021 is the best fit, not the only year that passes Three clocks, and a year The brochure from the years after the Ingstad dig says 16 samples of bone, turf, and charcoal gave an average near the year 1000. Wallace's later count is about 50 radiocarbon dates for the Norse layer, inside a site total of 148. The 2021 paper says more than 150 dates exist, 55 of them tied to the Norse occupation, and that once calibrated they stretch across the Viking Age and past it. The supplement puts those 55, at 95 percent probability, between 600 and 1200. Wallace sorts them by material, which is the useful part. Dates on worked wood run, in her probable figures, from about 640 to about 980, because a plank can be cut from the heart of an old spruce. Charcoal runs from about 670 to about 990. Twigs and branches, short-lived wood, fall between 990 and 1050, and she gives the mean as 1014. Two posts are about 900 and about 980. Sod and peat run from about 680 to about 1150, and she warns that a thick turf sample mixes centuries of plants. A whale bone looked like 1060 or 1080 until the marine reservoir is considered. She cites 500 to 600 radiocarbon years for Newfoundland waters, and she does not place the whale in the Norse layer. The 2021 study is a different use of radiocarbon. A cosmic-ray event in 993 raised the radiocarbon in the air by about 9 per mil, which the paper describes as a shift of about 70 radiocarbon years. That spike is in tree rings worldwide. Match it in a piece of unknown age, count the rings out to the bark edge, and the felling year is a calendar year. The team measured 127 samples, 115 at Groningen and 12 at Mannheim, from 83 rings in four wooden objects. One object, find 4A 70 B5-14, has only nine rings and does not contain the 993 spike, so it drops out. The other three are different trees, and at least two species. A balsam-fir log about half a metre long, 4A 59 E3-1, with clean cuts from a low-angle metal edge. A fir stump, 4A 68 E2-2, the stems slashed. An earlier date on that stump, TO-119, was 896 to 1114 at 95 percent. And a branch more than a metre long, 4A 68 J4-6, possibly juniper or thuja. All three still have the waney edge, the last ring under the bark. A first match against the calibration curve puts each waney edge, at 95 percent, somewhere between 1019 and 1024. A second test, against a year-by-year reference, gives 1021 as the best fit for every piece. The paper is plain that other solutions still pass at 95 percent. 1022 for the log. 1022 for the stump. 1019, 1020, and 1022 for the branch. The fit they prefer lines the sharp drop up so that ring 29 in from the edge is the year 992. One branch was cut in spring. The stump was cut in summer or autumn. The log's season could not be read, because it had been soaked in polyethylene glycol. The study argues that a preserved waney edge makes driftwood unlikely, since a trip in the water would have worn that surface off, and that fresh fir was growing near the site, so deadwood was not needed. It also argues that metal-tool cuts of this kind were not part of Indigenous woodworking in the area at the time. That is the study's case for a Norse felling in 1021. The years that also pass the test are in the same paper. The sagas, as Wallace summarizes them, put the voyages shortly after the year 1000. She also cites the saga scholar Ólafur Halldórsson, who argued the voyage could not have been much before 1020. A best fit of 1021 sits beside that second reading of the saga clock, and a little after the round year 1000. The study says the older radiocarbon could not choose among those years. The new one chooses, and then records the near misses.

Years, or a longer stay How long, and why the halls are empty Wallace's case for a short occupation is the thinness of the living. The middens are small, mostly ash and burnt food bone. She counts 125 unburnt bone fragments and 315 burnt ones. Phosphate tests in the bog and on the terrace did not find a vanished heap of bone and ash. The carpentry scatter reads to her as one episode. Floor deposits inside the buildings are slight. Nothing was repaired or rebuilt. Full-scale replicas at the historic site, she writes, suggest that a sod hall would have needed real renovation after about 15 years, and the originals show none. The 2021 supplement adds the absence of a burial ground, and a comparison with the refuse of Icelandic and Greenlandic farms. Parks Canada's map says the smelt happened once. Wallace's phrase is "a matter of years rather than decades." The sagas, in the same supplement, say the expeditions returned to Greenland after a short period whose length is not given as a number. Wallace quotes a translation of Erik's saga in which the party spends a third winter at Straumfjord. The saga of the Greenlanders, in her account, sends Leif, then his siblings, then Thorfinn Karlsefni, and they use houses Leif would lend and would not give away. Several voyages and a third winter are more than a single season. They are still not a string of decades. The labour of the buildings is Wallace's own figure for how much work a short stay could hold. About 1,100 cubic metres of sod for the three halls, at least 86 trees between 10 and 50 centimetres across for the roof posts, and a rough equivalent of 60 people for two months or 90 people for a month and a half. She thinks early 11th-century Greenland had about 400 to 500 people, so 20 to 30 of them on this terrace would have been 5 to 7 percent of that colony. A daughter settlement, on that arithmetic, was more than Greenland could spare. The culture page's word for the place is still "colony," and "settlers." Why they left is a list of readings, not a letter. Wallace sees an orderly departure. Household tools and substantial belongings are gone. She does not see a disaster in the ground. She adds the sagas' conflict with the people those texts call Skrælingjar, the small number of ships, the difficulty of judging longitude, and a Greenland economy that needed walrus ivory from the north more than it needed grapes from the south. Lumber, she notes, was also available in Markland, which she places in Labrador, closer to Greenland than the Gulf is. The supplement's historical note is simply that the sagas have every expeditionary group go home. A longer reading takes the eight halls, the 70 to 90 berths, and the third winter, and asks whether "years rather than decades" has room in it for the whole saga sequence, or whether the sequence is several crews in one short span. The ground, on Wallace's account, has the removal and the missing repairs. It does not have a sentence that says why.

A long Indigenous use, and a gap Wallace argues for Before the sod walls, and after Parks Canada's culture page opens with a longer story than the Norse one. In the past four to five thousand years, many people lived at this place, some of them longer than others, and among them was a small group of Norse sailors. Wallace's list of the occupations on the southern shore of Epaves Bay, close to the water, names Maritime Archaic, Groswater, Middle Dorset, a pre-Norse Indigenous occupation with affinities to the Cow Head complex, and a later one with affinities to Point Revenge and Little Passage, which she calls proto-Beothuk. Her argument about the Norse years themselves is that no Indigenous group was on the site at the same time. That is a reading of the layers, not a statement that the peninsula was empty. The sagas' meetings with Skrælingjar, in her geography, belong at the southern camp, not on this terrace. One object sits outside that sequence and gets its own caution. A late Dorset soapstone lamp, the only late Dorset piece on the site, lay in the surface sod above the collapsed furnace hut. Charcoal inside it gave a twentieth-century date. Wallace notes that late Dorset sites are known in northern Labrador, not here, and she treats the lamp as something brought south long after the Norse had gone. Inuit traders were in nearby Quirpon and Noddy Bay in the eighteenth century, she writes, and Innu from Labrador camped among the ruins in the eighteenth and nineteenth. The lamp's date and the campers are her case that this one bowl is not a Norse-period meeting. The Point Revenge and Little Passage material is the case for people here after the halls had fallen in. Wallace's plan of the A-B-C complex marks a point of that tradition among the Norse finds on the drawing. She does not read it as a shared household. A reader who wants overlap looks at that point on the plan, and at the saga meetings, and asks whether "not on the site at the same time" is the only way the layers can be stacked. Wallace's sentence is that the groups on the shore are before and after, and that the Norse layer in the bog is a band of its own, with metal-cut wood in the middle and stone-cut wood above and below. The 2021 supplement repeats that stratigraphy.

A second site was tested, and the report changed the reading Point Rosee, and the 2017 report Point Rosee is a point of land in the Codroy Valley, about 2.5 kilometres west-southwest of Millville, on Newfoundland's southwest coast. It is not this terrace, and it is not a second pin. In 2014 Sarah Parcak read satellite images for surface and vegetation patterns that might be buried features, and a magnetometer survey that year found concentrations that might be iron. In 2015, under a separate permit, four places were tested. Two were possible longhouse shapes from the imagery. One was a possible hearth from the magnetometer. One other test, the 2017 report says, turned out to be natural. The remaining areas produced sub-surface features that lined up with the vegetation marks, and the 2015 work, as the later report describes it, posited sod or turf features and possible human processing of bog iron. Those results were the reason given for going back. The report of 8 November 2017, by Parcak and Greg Mumford with Gerald Penney Associates, is the account of the 2016 season and of the laboratory follow-up. The executive summary says no diagnostic Norse artifacts were found. Further tests on the bog iron "determined that they were likely natural deposits with no evidence of having been further processed." Turf features that had been read as human "have also been reinterpreted as natural." In the discussion, a re-examination of the possible roasting hearth and the turf-like bands, led in the field by Karen Milek, "deemed both areas to be far more likely, if not all but certainly, to represent natural features." The recommendations say there are no clear findings of human activity before 1800, that the posited turf walls are not human-made, and that the 2015 and 2016 seasons "found no evidence whatsoever for either a Norse presence or human activity at Point Rosee prior to the historic period." The areas opened, the report says, are not promising enough to justify a return in search of a Norse settlement. The same pages keep several threads open, and they are the report's own. A square of darker vegetation, visible from the ground and in the satellite images, and picked up in early magnetometer work, still had no matching feature under it. The authors call that square an intriguing riddle, with soil analysis then still pending. A more detailed look at the bog iron is described as pending. An optically stimulated luminescence result puts the last placement of a boulder, the one once read as a hearth stone, around the year 200 or a bit later. The report says that date is interesting and that, once the roasting feature is read as natural, the boulder's story is not particularly significant. It allows that some human agency, perhaps much earlier, might have moved the stone, or that slope drift might have done it. Along the Great Codroy River, two Norse specialists on the team saw landscape of a kind they associate with Norse sites, and they did not have time to test it. The report's sentence is that the Codroy Valley region remains a good candidate for early European visits, and for older Indigenous sites, even though Point Rosee itself is not the settlement that was hoped for. Birgitta Wallace visited the 2016 dig, the participant list says, to observe and to help with interpretation. Her name on that list is not a second verdict. The verdict in the recommendations is the one quoted above. A 2016 sun calculation by Martin Ekman uses 47.8 degrees for Point Rosee and finds that the midwinter sun at the azimuth he takes from the sagas would stand about 1.2 degrees above the horizon there. He treats the point as a probable Norse settlement, which is the reading the 2017 report does not keep. The arithmetic of the sun at that latitude is a separate sentence from the trenches. It is taken up with the other sky readings below.

The lists are long, and a rune is not on them No rune in the published finds The culture page's objects are the pin, the lamp, the whorl, the needle, the gilded fragment, the hone, the slag, and the boat nails. Wallace's personal items add the glass bead and the door socket, and her waste lists add the nails, the slag, and the wood. UNESCO speaks of about 800 objects of wood, bronze, bone, and stone, and of iron production. None of these accounts includes a runic inscription. A camp with buildings and iron, and with no rune in the published catalogue, is one fact pattern. Stones in Minnesota, Oklahoma, Maine, and Rhode Island, read by some as Norse runes and by others as later cuttings, are a different fact pattern. They are weighed, with this camp as part of the evidence on both sides, in the Runestones in America room. A rune that was never in the ground cannot be produced by a longer catalogue. A rune that was cut and has not survived would not appear in the lists either. The lists are what the room has.

A cosmic-ray clock, and a saga sentence about the winter sun Sky readings No azimuth, no horizon height, and no solstice or star alignment of these halls is published in the sources used here. The buildings face the bay and the brook. Nobody in these reports has measured a wall against a sunrise. The sky material that does exist is of two other kinds, and both of them are used to tell the time or the latitude, rather than to aim a door. The first is the year 993. The 2021 study describes a burst of cosmic rays that raised radiocarbon in the atmosphere, worldwide, by about 9 per mil. Trees built that extra radiocarbon into the ring that was growing then. You would "see" it in a laboratory counter, not in the night sky over Epaves Bay, though the cause the study names is a sky event. Once the spike is found in a ring, the rings are counted outward to the last ring under the bark. The study's best fit for that count, on three trees, is a felling in 1021. The same paper records other years that still pass the test. What the spike dates is the cutting of those pieces of wood. It does not, by itself, say how many winters the halls were roofed, or which saga name belongs on the terrace. A cosmic-ray clock is a serious dating tool. It is not an alignment of the settlement. The second is a sentence in the saga of the Greenlanders. Martin Ekman translates it, from a text in Mats G. Larsson's edition, this way. The days there were of more equal length than in Greenland or Iceland, and on the shortest day the sun had eyktarstaðr and dagmálastaðr. Those two words are times of the Norse day, tied to meals and to the end of work. Ekman, following Gustav Storm and Hans Geelmuyden in 1886, reads a passage in the Icelandic law collection Grágás. Eykt is when the sun has passed two thirds of the southwestern eighth of the horizon and has one third left. He calculates that direction as 52.5 degrees west of south, and dagmál as 52.5 degrees east of south. You would observe it by watching where the sun stands on the shortest day, against a coast you already know how to divide. Geelmuyden's result, as Storm's Vinland study reports it, is that the place in the Flateyjarbók sentence cannot lie north of 49 degrees 55 minutes. Ekman's own recalculation of the same limit is 49.9 degrees. At the latitude he uses for this terrace, 51.6 degrees, he finds the midwinter sun at that azimuth 1.6 degrees below the true horizon, and 1.2 degrees below it once refraction is allowed for. On that arithmetic the sun is not up at eykt here, even if the direction is uncertain by a degree or two, which is the uncertainty he takes from a medieval Icelandic sun table. He calculates the contrast the saga is pointing at. At Brattahlíð in Greenland, latitude 61.0 in his figures, the sun at that azimuth is below the horizon from 11 November to 31 January. At Thingvellir in Iceland, 64.3 degrees, it is below from 5 November to 6 February. A winter day on which the sun is still up at the afternoon mark would have been a different country from the one the crews had left. A second way of reading the same words treats them as hours on a clock. Storm's study reports Jakob Aall's result from that approach. 41 degrees, 24 minutes, and 10 seconds. Ekman's azimuth and Aall's clock hour cannot both be what the sentence was doing, unless the carver of the prose meant a direction that also happened to be a named hour. Each one is a full use of the line. The azimuth reading puts the observation at or south of about 50 degrees, which is south of this terrace and still far north of Aall's New England figure. Wallace's Vinland is the Gulf of St. Lawrence, from the Strait of Belle Isle south to the Northumberland Strait, which includes coasts on both sides of Geelmuyden's line. Her Hóp, in northeastern New Brunswick, is south of it. Her Straumfjord is this terrace, north of it. If the sun sentence belongs to Leif's camp, and if eykt is Geelmuyden's azimuth, Ekman's arithmetic is a difficulty for putting that sentence here. If the sentence belongs to a summer camp farther south, the same arithmetic is what Wallace's split between base and Hóp can use. If Aall's clock reading is the right use of the words, both this terrace and the Miramichi are too far north for the sentence. The saga does not say that anyone built a marker on the ground to watch it. The use, in the text, is to locate a land.

Location and dating

Location: The pin is the cluster of mapped hall remains above Epaves Bay, at 51.595 degrees north and 55.533 degrees west, on the tip of the Great Northern Peninsula, facing the Strait of Belle Isle. UNESCO's point for the whole property, 7,991 hectares inscribed in 1978, is 51 degrees 35 minutes 05 seconds north, 55 degrees 33 minutes 00 seconds west. The buried ruins and the reconstructed halls a visitor walks among are the archaeological cluster, not that property centroid. Point Rosee is a different coast, in the Codroy Valley about 2.5 kilometres west-southwest of Millville. Ekman's sun calculation uses 51.6 degrees for this terrace and 47.8 degrees for Point Rosee.
Period label: Current consensus — open to revision Parks Canada describes Indigenous use over the past four to five thousand years, with a Norse episode inside it. Wallace places that episode in the early 11th century, with twig dates from 990 to 1050 and a mean of 1014, and with saga voyages shortly after the year 1000, or, on Halldórsson's argument, not much before 1020. The 2021 study's best fit for three fellings is 1021, with other years still inside its test. The culture page's architectural comparison is the decades just before and after the year 1000. The pin type Wallace cites runs about 920 to 1050.
How this dating was done: The wide net is radiocarbon on charcoal, wood, twigs, sod, and bone, which Wallace reads material by material because old heartwood and thick turf pull a date earlier than the day of use. The year-resolution clock is the 993 radiocarbon spike inside a ring count, which is dendrochronology with radiocarbon as the marker. Architecture and the bronze pin are typology, set beside Icelandic and Greenlandic houses and a dated pin series. The sun sentence is a latitude argument from a text, the kind of sky reading archaeoastronomy would measure if a wall or a horizon had been surveyed. No such survey of these halls is in the sources used here. Ekman's figures are a calculation from the saga and from a medieval law passage, not a measurement taken on the terrace.
What could change it: A waney-edge date on a fourth timber, with the 993 ring in it, would thicken or shift 1021. A radiocarbon measurement on the butternuts themselves would tie the southern cargo to the same year or set it loose. A published horizon survey, with an azimuth, would give the halls a sky reading they do not yet have. At Point Rosee, the soil work and the bog-iron reassessment the 2017 report left pending would speak to the square of dark vegetation and to the ore. They would not, by themselves, rewrite the terrace at Epaves Bay.

Published ranges, drawn as bars. A bar is a claim about a span, not a point.

Fifty-five Norse-related radiocarbon datesContested

Wallace's twig and branch datesContested

The 2021 felling, best fit and the years that also passContested

The pin type Wallace cites

600 CE800 CE1200 CE

Fifty-five Norse-related radiocarbon dates. The 2021 supplement's 95 percent window for the older series. The study argues this spread cannot pick a year.
Wallace's twig and branch dates. Short-lived wood. She gives the mean as 1014 and treats this cluster as the closest of the older clocks.
The 2021 felling, best fit and the years that also pass. Best fit 1021 on all three trees. 1019 and 1020 pass for one piece. 1022 passes for each of them.
The pin type Wallace cites. Fanning's span for this Hiberno-Norse pin, especially common around 950. A range for the type, not a felling year.

What is actually there

Parks Canada dates the find to 1960, names George Decker, and describes eight sod buildings excavated over the following eight years. A brochure counts seven expeditions from 1961 to 1968.
UNESCO, inscribed in 1978, describes eight turf buildings, about 800 artifacts, ironworking, and woodworking it reads as likely ship repair, and says the remains correspond to the Vinland sagas.
Wallace measures hall A at 29 metres and about 102 square metres, hall F at about 160 square metres, hall D at 20.2 metres and about 88 square metres, and estimates 70 to 90 people.
She counts 99 nail fragments and one whole nail, about 66 percent of the fragments in the F-G complex, and reads the X-rays as boat nails. Parks Canada's map estimates one smelt, enough iron for about 100 to 200 nails.
Three butternuts and a cut butternut burl lay in the carpentry waste. Wallace places the tree's northern limit near 47 degrees north, in New Brunswick and the St. Lawrence valley.
Her twig dates run from 990 to 1050, mean 1014. The 2021 study's best fit for three metal-cut trees with bark edge is 1021. Other years from 1019 to 1022 pass its test for at least one tree.
A bone once called pig is later called seal. A stone called a lamp is later read as a door socket. A building called a sauna is later read as an ordinary hut, with one associated date after the Norse layer.
Parks Canada describes Indigenous presence over four to five thousand years. Wallace lists cultures before and after the Norse layer and argues the terrace was not shared during the Norse stay.
The 2017 Point Rosee report says the bog iron and the turf features are natural, and that the seasons found no Norse presence and no human activity there before the historic period. It also says the wider Codroy Valley is still worth surveying.
No published find list used here includes a runic inscription.

Open questions

Question	Status
Is this terrace Straumfjord, Leif's houses, Vinland itself, or a camp the sagas do not name?	Open. Wallace argues the first, and a gateway. Ingstad's published reading is Leifsbúðir. The sun-latitude readings put the saga's measured place farther south.
Were the three trees cut in 1021?	Open inside a narrow band. That year is the study's best fit for all three. 1019, 1020, and 1022 also pass for at least one piece.
How many winters, and why are the halls empty?	Open. Wallace argues years rather than decades, and an orderly leaving. A third winter in Erik's saga, and a series of voyages, are the longer reading.
Where did the butternuts come from, and is there a southern camp?	Open. The nuts and the burl are in the waste. Wallace places the harvest in northeastern New Brunswick. No second excavated camp has been published there.
Does the shortest-day sentence describe this latitude?	Open. Geelmuyden's azimuth reading gives a northern limit near 49 degrees 55 minutes. Aall's clock reading gives about 41 degrees 24 minutes. This terrace is near 51.6 degrees.
What should a reader do with Point Rosee?	The 2017 report's recommendation is that the tested areas are not a Norse settlement. The same report leaves the Codroy Valley, a square of dark vegetation, and a pending ore study on the table.

The current narrative

Where publication stands today. It is the best-supported reading, and it is open to revision. It is not a verdict.

The current narrative — open to revision

A winter base and a gateway

Wallace's papers, Parks Canada's site map, and UNESCO's listing describe the terrace as a Norse base of the early 11th century, used over winters, where boats were repaired and iron was made once, and from which people went south. Wallace's name for it is Straumfjord, inside a Vinland that also includes the Gulf. Hóp, the grape and lumber camp, is farther south. The 2021 study's best fit places three fellings in 1021, which she and her coauthors read as activity in that year.

What supports this

Eight sod buildings match Icelandic and Greenlandic plans of the period, with storage, a smithy, a furnace, and a lean-to where most of the cut nails lay.
There is no byre, no enclosure, and no grazing disturbance in the tests Wallace reports, and the food bone she accepts is seal and whale. A bone once called pig was reidentified as seal.
Three butternuts and a cut burl are not Newfoundland trees. Wallace places their range at the same northern edge as wild grapes.
The Strait of Belle Isle, Belle Isle, and currents she cites at 0.4 metres per second are her match for the saga's fjord of currents.
Thin middens, a single smelt, and no repairs are her case that the base was real and brief.
Halldórsson's saga argument, not much before 1020, sits next to the 1021 best fit.

What pushes back

Ingstad's reading, as Brown's review of the Ingstads' book states it, is that the site is Leifsbúðir, the main settlement, not a gate to a Vinland somewhere else.
The 1970s brochure's account of his search treats the grapes as possibly berries, so the wine-land need not lie south of this latitude.
Ekman's azimuth arithmetic puts the saga's shortest-day sun below the horizon at 51.6 degrees. Geelmuyden's limit is 49 degrees 55 minutes.
Larsson and Barnes, as Wallace notes, have read the site as a settlement the sagas do not name.
Erik's saga, in the translation she uses, includes a third winter, and the Greenlanders' saga sends several expeditions. A "few years" has to hold that sequence, or the sequence belongs to a longer stay.
The 2021 paper itself records years other than 1021 that pass the same test.

What would settle it

A new plan of the walls would not, by itself, choose among Straumfjord, Leifsbúðir, and an unnamed camp. A southern site with Norse buildings, grapes or butternut wood in place, and a date that meets the 993 ring would give Hóp a location. A felling year that moved off 1021 by more than the years already inside the test would move the clock. The room's part is to keep Wallace's package, which is the reading the site's own publications now print, beside the readings that use the same walls differently.

Emerging narratives

Other published readings. Each may be right. None is adopted here unless the evidence carries it.

An emerging narrative

Leif's houses, and Vinland itself

Helge Ingstad's reading, as the 1970s brochure and as Stuart Brown's review of the Ingstads' book describe it, is that L'Anse aux Meadows is Leifsbúðir, the camp in the sagas, and the main Norse settlement in the west. The brochure says he took the sailing times as far as northern Newfoundland and treated the grapes as possibly large red berries. On that reading the terrace is Vinland, not a doorway to it.

What supports this

The site sits where a ship from Greenland, passing Markland, would meet the Strait of Belle Isle, a landmark Ingstad could search for.
The buildings are substantial winter houses, which the culture page compares with Iceland and Greenland around the year 1000, and the finds include women's tools as well as a smithy.
The brochure's 16 early radiocarbon samples averaged near 1000, the round year long attached to Leif's voyage.
Wallace herself allows that hall F could be the hall Leif built, even inside her gateway reading, and the supplement notes that some readers treat Leifsbúðir, Straumfjord, and Hóp as three different places while others combine them.
A meadow reading of the name, which the Vinland project page describes, would not require grapes at this latitude. The brochure's berry reading is the version tied to Ingstad's search.

What pushes back

Wallace argues the layout is a base, not a farm. No byres, no domestic animals in the bone she accepts, and butternuts from south of Newfoundland.
The Vinland project page argues that the long í, and the spelling Vijnland in Flateyjarbók, mean wine, and that a meadow name would have been formed differently.
Ekman and Geelmuyden place the shortest-day observation south of this terrace, which is a problem if this terrace is the camp the sentence describes.
The 2021 best fit of 1021 is a little later than the round year 1000, unless Halldórsson's "not much before 1020" is the saga clock one uses.
UNESCO's sentence is that the camp corresponds to the voyages, and that people went farther south. That is the gateway reading, printed on the listing.

What would settle it

The Ingstad reading would be strengthened by a grape or a butternut that belonged to this latitude in the 11th century, or by a saga passage that put Leif's houses on a strait at the tip of a great island and nowhere farther south. It would be weakened by a second camp, farther south, with the same Norse buildings. Neither of those has been published. The berries, the meadow name, and the wine name remain three ways through the same word.

An emerging narrative

Vinland farther south than this terrace

One published use of the sun sentence, and some published uses of the climate, put the Vinland of the sagas south of Epaves Bay. On that reading these halls are a gateway, or a camp the sagas never name, and the wine-land is another coast. Wallace's own Hóp is a southern camp. The difference is that she still calls this terrace part of Vinland. The readings in this card move Vinland itself off the terrace.

What supports this

Storm reports Geelmuyden's limit. North of 49 degrees 55 minutes, the shortest-day sun is not up at the azimuth they take from Grágás. Ekman recalculates 49.9 degrees, and at 51.6 degrees he finds the sun 1.2 degrees below the horizon after refraction.
Storm also reports Aall's clock-hour result, 41 degrees 24 minutes 10 seconds, which would place the sentence far south of Newfoundland.
Wallace notes that Bergþórsson and Larsson have found this coast too harsh for Straumfjord and have preferred southern New Brunswick or eastern Nova Scotia.
She notes that Larsson and Barnes have called the site an anonymous settlement, unmentioned in the sagas.
Butternuts do not grow here. Any reading that wants the nuts and the grapes in one place has to go south of the terrace to find both.

What pushes back

Wallace's parallels for Straumfjord are physical. A strait, an island at its mouth, strong currents, and a base built for winter. She argues the 11th-century winters here were mild enough, citing a warming around the year 1000, that the saga's outdoor grazing is not absurd.
The Vinland project page, and Wallace, argue the name means wine, and then put the wine at Hóp while keeping the base in the north as part of the same country.
Parks Canada and UNESCO present this site, not a southern camp, as the place that corresponds to the voyages.
Aall's 41 degrees and Geelmuyden's 50 degrees cannot both be the sentence. A reader who moves Vinland has to say which use of the words is doing the moving.
No Norse buildings have been published at the Miramichi, at Chegoggin, or at the latitude Aall calculated. Point Rosee, farther south on Newfoundland, is the tested candidate, and the 2017 report does not keep it.

What would settle it

The sun sentence can be settled, as far as a text can be, by a philological agreement on whether eykt is a direction or an hour, published against the Grágás passage and the saga manuscripts. A latitude from that agreement either includes this terrace or it does not. Archaeology would still have to find the southern camp, if the sentence and the halls are not the same place. Wallace's case is that the nuts are the southern camp's cargo, already in the bog.

An emerging narrative

A longer stay than a few years

The buildings are a lot of work, the sagas describe more than one winter, and the culture page's word is colony. A longer occupation is the reading that takes the third winter, the series of voyages, and the 70 to 90 berths as a stay measured in more than Wallace's "years rather than decades," or at least as the whole saga sequence rather than one smelt and one season of carpentry.

What supports this

Wallace's own labour estimate is 60 people for two months or 90 for six weeks, and her sleeping-space estimate is 70 to 90. That is a settlement effort, even if it was brief.
Erik's saga, in the Jones translation she quotes, spends a third winter at Straumfjord. The Greenlanders' saga sends several expeditions that use the same houses.
The culture page calls the place a colony and the people settlers, and it compares the buildings with Iceland and Greenland just before and after the year 1000, which is a settlement comparison.
Year-round construction, which Wallace stresses, is what a crew builds when it intends to stay through winters, not what a party raises for a week.
The replicas' 15-year repair cycle is an analogy from modern copies. A milder 11th-century winter, which Wallace also argues for, would change how fast a sod roof failed.

What pushes back

The middens are small. 125 unburnt bone fragments and 315 burnt ones, no phosphate concentration, slight floors, and carpentry waste she reads as a single episode.
No repair and no rebuilding. She expects renovation to show by about 15 years, and she does not see it.
One smelt, a few kilograms of iron, and a tool kit that was taken away. The supplement adds the lack of a cemetery.
The sagas, in that supplement, have the expeditions return. A third winter is longer than a season and shorter than a colony that lasts for generations.
Greenland's early population, in the figures she cites from Lynnerup, is about 400 to 500. She argues it could not spare a permanent daughter settlement.

What would settle it

A sequence of floors, with dates on the upper and the lower, would show whether the halls were lived in once or reopened. A cemetery, or a midden with the volume of a Greenland farm, would lengthen the stay in the ground rather than in the sagas. Until then the short reading and the saga sequence are both published, and they meet in the middle at "more than one winter, less than a lifetime."

Should the current view be reevaluated?

Open to revision

The ridges, the eight buildings, the pin and the whorl and the nails, the butternuts, the 993 spike, and the 2017 sentences about Point Rosee are all published. So are a gateway reading, a Leifsbúðir reading, a sun sentence with two latitudes, a short stay, a third winter, and a find list with no rune. The 2021 study's best fit is a year. It is not, by itself, a name for the terrace or a reason the halls were emptied. The room stays open because a reader can grant the buildings and still have to choose, with the evidence on the page, among the uses and the durations.

What drives this

Eight sod buildings, a single iron smelt, and about a hundred cut nails, published as a Norse camp of the early 11th century.
A 2021 best fit of 1021 for three trees, with 1019, 1020, and 1022 still passing the study's test.
Wallace's gateway and Straumfjord, Ingstad's Leifsbúðir, and sun-latitude limits at about 50 degrees and at about 41 degrees.
Three butternuts from a tree Wallace says does not grow in Newfoundland, and no second excavated camp at the tree's northern edge.
A short-stay argument from thin middens, and a longer argument from a third winter and a series of voyages.
A 2017 report that reads Point Rosee as natural, and that still calls the Codroy Valley worth surveying.
Find lists that include iron and a bronze pin, and do not include a rune.

This is the museum’s own weighing of the evidence. It is not a visitor vote, and it is separate from the slider.

Weigh it yourself

Anonymous. Nothing here is stored except on this browser. Your lean does not change the assessment beside it.

A winter base and a gateway Unsure Leif's houses, and Vinland itself

If the contested reading is right

What would follow if a reading that is not the consensus turned out to be true, and what could settle it. This explains the stakes. It does not adopt the reading.

These are Leif's houses, and Vinland itself

The sagas and the ruins would be the same camp, and the grapes in the text would be berries, a meadow name, or a memory attached to this coast. The butternuts would be cargo from a trip that did not require a second base. The sun sentence would need a reading that allows a latitude near 51.6 degrees, or it would belong to a different paragraph than the houses.

Why it is still worth a test

A manuscript reading of eykt that keeps the winter sun up at this latitude, published against Grágás, and a botanical case that butternut or wild grape reached the Great Northern Peninsula in the 11th century.

Vinland, as the sun sentence locates it, lies south of this terrace

The halls would be a gateway or an unnamed camp, and the wine-land would be a coast still missing its buildings. Wallace's Hóp in the Miramichi would become the place the sentence might fit. Aall's 41 degrees would move it farther again. Point Rosee would not fill that gap. The 2017 report is the account of why.

Why it is still worth a test

The philological choice between a direction and an hour, and a survey of the southern Gulf aimed at turf walls, bog iron, and metal-cut wood of the same kind as the three dated pieces.

The stay ran longer than Wallace's few years

Greenland would have supported a farther base for more than a season or two, and the empty halls would be a slow ending rather than an orderly close after one episode of building and repair. The third winter would be the minimum, not the surprising extra.

Why it is still worth a test

Dated floors and a refuse study that can be compared, quantity for quantity, with a short-lived Icelandic booth and with a Greenland farm.

Point Rosee was a second Norse site, as the 2015 season proposed

Norse ironworking would have a second Newfoundland location, on a coast where Ekman's sun arithmetic has the midwinter sun up at eykt. The 2017 report reads the ore and the turf as natural and recommends no return for a Norse settlement. That report would have to be set aside by the pending analyses it names, or by finds the trenches did not produce.

Why it is still worth a test

The bog-iron reassessment and the soil work the 2017 report left unfinished, published in a form another laboratory can run.

Sources

Culture and history, L'Anse aux Meadows Parks Canada
Maps and brochures, the Norse buildings Parks Canada
L'Anse aux Meadows National Historic Site UNESCO World Heritage Centre
The Norse in Newfoundland: L'Anse aux Meadows and Vinland Newfoundland and Labrador Studies, Birgitta Linderoth Wallace
Evidence for European presence in the Americas in AD 1021 Nature, 2021
L'Anse aux Meadows, a site brochure Parks Canada, about the 1970s
The Viking Discovery of America, a review Newfoundland and Labrador Studies, Stuart C. Brown
Wine and Vinland, in "Is L'Anse aux Meadows Vinland?" Great Unsolved Mysteries in Canadian History
The midwinter sun at Viking settlements in North America Martin Ekman, Small Publications in Historical Geophysics 29, 2016
On the meaning of eyktarstaðr in the Flateyjarbók Vinland account Storm and Geelmuyden, Arkiv för nordisk filologi, 1886
Studier over Vinlandsreiserne Gustaf Storm
Point Rosee, Codroy Valley, NL (ClBu-07), 2016 test excavations Parcak, Mumford, and Gerald Penney Associates, 8 November 2017

Image credits

A replica hall at the historic site. The photographer identifies it as Hall A of the A-B-C complex, built north of the buried ruins. The turf and the timber door are the reconstruction. Wallace uses full-scale replicas here when she estimates how soon a sod hall would need repair.. TravelingOtter, CC BY-SA 2.0. Shown unedited.
The terrace and the bay on 27 August 2022. The photographer's description names Hay Cove to the left and Norstead, a separate modern reconstruction, off to the right. The buried halls are on this ground. Their outlines are not a readable plan in the frame. UNESCO says the archaeological remains were reburied to protect them.. Gordon Leggett, CC BY-SA 4.0. Shown unedited.
Turf courses and smoke hoods on a reconstructed building. The photographer's note says these are recreations of the halls dug in the 1960s and then covered. The stacked sod is the feature Wallace and the Ingstad reports both start from. The hoods and the path are the modern site.. Eric Titcombe, CC BY 2.0. Shown unedited.
A replica hall, a fenced yard, and the sea. The person in the doorway is part of the historic site's presentation. Wallace reports that archaeological tests found no livestock enclosures and no sign of grazing. The fence in the photograph is the reconstruction's.. D. Gordon E. Robertson, CC BY-SA 3.0. Shown unedited.
The platform along the wall is the kind of bench Wallace describes for sleeping and sitting. The clothes, the shield, the horns, and the pots are the reconstruction's furnishings. Her reading of the departure is that household inventory was taken away.. D. Gordon E. Robertson, CC BY-SA 3.0. Shown unedited.

← Canada World map