The Snow House: How the Iglu Was Made

Status: Researched draft — cultural review pending

How to read this entry: Documented fact = published scholarship (Sources). Tradition — the people's own account = Inuit oral history / Inuit-organisation teaching, labelled as such. Debated / Claim = disputed point. No ceremonial or restricted content is described. Scope note: this entry is a history, not a building manual. Cross-reference: the Inuit entry in this collection, whose rule this entry follows — the snow house was not universal Inuit practice, and its regional limits are part of its history.

The Word and Its Limits

• A house, not only a snow house. Documented fact: iglu (plural igluit) means a house of any material — tent, sod, driftwood or modern. Several dialects (Inuinnaqtun, Natsilingmiutut, Kivalliq, North Baffin) use it for snow houses as for other buildings; English "igloo" alone narrowed to the snow dome (Dorais, 2010).
• The snow house specifically. Documented fact: in the Igloolik region it is igluvijaq (plural igluvijait), iglu being kept for other buildings; Mathiassen writes it iglo (Mathiassen, 1928, Material Culture of the Iglulik Eskimos, Fifth Thule Expedition, Vol. VII).
• Regional limit, stated first. Documented fact: the snow-block dome was traditionally the winter house of the Central Arctic — Iglulingmiut, Aivilingmiut, Natsilingmiut (Netsilik), Inuinnait (Copper) and neighbours in what is now Nunavut — and of Qaanaaq, Greenland. Other Inuit used snow chiefly as insulation over other houses (Boas, 1888; Mathiassen, 1928).

Inheritance: What Came from the Thule

• The migration. Documented fact: the Thule, ancestors of today's Inuit, spread east from Alaska to Greenland, dated c. 1000–1200 CE as material culture (McGhee, 1996). Tradition — the people's own account: Iglulik accounts ascribed the older stone-house ruins to a prior people, the Tunit (Mathiassen, 1928).
• What the Thule house was. Documented fact: the excavated Thule winter house was not a snow dome but a large, oval, semi-subterranean house: flagstone or driftwood floor, raised rear platform of stone slabs, stone and whalebone walls, sod roof, long entrance tunnel (McGhee, 1978, in Hodgetts & Eastaugh, 2017). Its plan — sunken cold-trapping passage, raised rear platform, lamp at the platform's edge — is the plan the snow house also follows.
• Debated / Claim — the dome's antiquity. Snow survives archaeologically nowhere, so no dome can be dated. Mathiassen (1928) thought the broad snow-knife form in Boas "seems to be connected with the Thule culture" — an inference from a tool, not a dated house. That the Thule built domes is Debated / Claim, not asserted here.

Finding the Snow, and the Knife

• The right drift. Documented fact: building began with a bank of "sufficiently deep, firm and homogeneous snow," preferably about half a metre deep so blocks could be cut vertically. Too loose could not be used; too hard was avoided. Snow houses could be built from October (Mathiassen, 1928).
• Testing it. Documented fact: snow was tested with a probe (aputsiuk), a thin wood, bone or antler rod — examples at 81 and 84 cm — pushed down so the builder felt the snow by resistance (Mathiassen, 1928).
• The snow knife (pána). Documented fact: by the 1920s Iglulik builders used single-edged steel knives of European manufacture, self-fitted handles — two-handed, thin, knobbed as a hand-stop; examples run 37–50 cm. The earlier knife was ivory, fairly slender (old Pingerqalik specimen, 30 cm; pierced Amitsoq blade, 37 cm; similar in Parry, 1824, and Boas, 1888); bone and antler knives are also recorded before steel (Mathiassen, 1928; Dawson, 2002).

Raising the Dome: Spiral, Lean and Key Block

• The blocks. Documented fact: two scratches marked a block's length, 60–70 cm; the first, triangular block was discarded, later blocks cut against the vertical face so formed, loosened by a kick, often cut slightly curved, concave side in. In good snow blocks ran 60–80 × 40–50 × 15–20 cm, smaller in poor snow; about ten were cut before building began, about 35 for a house about 3 m across (Mathiassen, 1928).
• The spiral, from inside. Documented fact: the first blocks stood in a circle, leaning slightly inward, faces trimmed to fit, doorway preferably south. That ring was cut so its top rose as a spiral, and building continued along it: the builder, working from inside, lifted each block with both hands, set it on the layer below, and with the knife in his right hand cut its faces to fit while pushing it home with his left — each block resting on two obliquely cut faces (Mathiassen, 1928; Boas, 1888).
• Lean and key block. Documented fact: each row inclines more than the last until the uppermost blocks lie almost horizontal yet rest firmly. The top (key) block was pushed vertically up through the remaining hole, cut above it to shape, and let down into it (Mathiassen, 1928).
• The shape. Documented fact: the dome is not a hemisphere but a catenoid of revolution with an optimum height-to-diameter ratio; it eliminates ring tension and shell moments, carrying load in compression (Handy, 1973, Arctic 26(4)).

Chinking, Tunnel, Air, Platform — and Why It Is Warm

• Chinking. Documented fact: gaps were plugged with loose snow or small fitted blocks, joints pressed tight with the mitten; for a lasting house, loose snow was shovelled over the dome — mostly women's work (Mathiassen, 1928).
• Tunnel and cold trap. Documented fact: entry was by a low passage, often cut down into the bank, so floor and platform stood above its top; cold air fills the passage and stays low while warmed air rises (Boas, 1888) — the cold-trap principle: dense cold air held below the platform (Kershaw et al., 1996).
• Ventilation. Documented fact: a small hole near the top — qingaq, "the nose of the house" (Mathiassen, 1928); qangirn (Boas, 1888) — was closed at night with skin or a mitten, never snow, which was too airtight. Boas: cold air enters by the door, warms, rises and escapes through the hole, giving the lamps draught.
• Platform and window. Documented fact: the rear was a platform of stamped snow, sometimes a ledge of the drift left standing as blocks were cut inside the circle — igleq (main) and akit (side); in Mathiassen's Repulse Bay two-family house (4.3 m diameter, 2.65 m high) it stood 0.55 m above the floor.
• The qulliq. Documented fact: heat and light came from the qulliq (kudlik) — a shallow, typically soapstone lamp burning rendered seal blubber along a wick of moss or Arctic cotton grass, set on pegs before the platform; tending it was documented women's work (Mathiassen, 1928; Kershaw et al., 1996). Described at that published level only.
• Why it is warm — and the measurements. Documented fact: wind-packed snow is mostly air trapped among ice crystals, and the trapped air insulates. Kershaw et al. (1996), instrumenting two houses near Resolute with Inuit advisers, measured new-wall heat flux equal to an insulated modern 2 × 4 wall (about RSI 2.17 / R12); two adults burning 2.3 kg of seal fat a day hold an unlined 3 m house at about 5 °C at −30 °C outside. Mathiassen (1928) shows the same pattern: at Itibdjeriang, outside −34 °C, −3.1 °C on the platform and −6.0 °C on the floor; at Qajuvfik (skin-lined), outside −30 °C, +5.2 °C on the platform.
• Lining and decay. Documented fact: many Iglulingmiut houses (not Aivilingmiut ones) were lined with sealskins hung 10–40 cm off the wall. Boas (1888): the lining keeps warm air off the roof; lined houses ran at 10–20 °C where unlined ones rose only 2–3 °C above freezing. Lamp heat melted the innermost snow, which refroze as ice, making the house tight and strong; as the wall iced through it dripped and was left. Usual life: a month to six weeks, lined houses often a whole winter; by May roofs fell in (Mathiassen, 1928; Boas, 1888).

How Long It Took, How Big It Stood

• Times, ranged as given. Documented fact: for a house about 3 m across Mathiassen (1928) gives about 1½ hours for one man alone, about 1 hour for two men (one cutting, one building), 40–50 minutes in very favourable conditions. Boas (1888) — the only earlier figure Kershaw et al. (1996) found — reported a small travelling house (1.5 m high, 2.1 m diameter, 25 blocks) built by two people in 2 hours — a different house, not a contradiction.
• The exception, flagged. Documented fact of the record: an old Aivilik, Aqaut, once built a house for four, alone, in 26 minutes — "but this is the exception" (Mathiassen, 1928).
• Sizes, ranged. Documented fact: one-family houses ran about 3–3½ m in diameter, two-family about 4.3 m; Mathiassen's largest measured 5.5 m; an unlined one 6.5 m across, 4 m high. Beyond that the record thins: Boas, citing Captain Comer, gives a west Hudson Bay house of 7 × 4 m, and Low a Fullerton house 27 ft (about 8.2 m) across, 12 ft high, that "was too big for the material and had to be supported by battens" (Mathiassen, 1928, citing Boas, 1888, and Low).

Regional Variants: The Other Houses

• The qaggiq. Documented fact: the largest Central Arctic snow houses were festival and dance houses, not dwellings — qâgi (Mathiassen), qaggiq today. Mathiassen (1928), noting they were no longer built in his time, records them from elders: at Repulse Bay two large connected houses, dancing under the joining arch; at Iglulik a large round house, no platform, lamp on a snow block in the middle. Dawson (2002) documents composite Copper, Netsilik and Iglulik houses built around such spaces for singing, dancing, drumming and games (NiCHE Canada, 2025, gives interiors upwards of 20 ft, about 6 m).
• The qarmat. Documented fact: qarmat (singular qarmaq) were the autumn and spring houses of the same world, and winter houses elsewhere — low stone, turf or mixed walls, sometimes with whale skulls, roofed with skins or a snow dome over a stone ring. The elder Aua told Mathiassen that in his father's boyhood people lived in qarmat at Iglulik and four other places. Hodgetts & Eastaugh (2017) define qarmat against the Thule winter house as the less substantial form — shallower pit, skin rather than sod roof (Mathiassen, 1928; Boas, 1888).
• The west. Documented fact: in the western Arctic (Inuvialuit region and Alaska) the substantial winter house remained the Thule-plan semi-subterranean house: excavated floor, stone, whalebone or driftwood framing, sod cover, long low tunnel, raised rear platform; snow banked it but did not form its walls.

The Settlement Era: Displacement

• Into settlements. Documented fact: after the Second World War Canada pressed Inuit into permanent settlements, for humanitarian concern and Cold War sovereignty alike in the literature. Scrap-lumber shacks first replaced snow houses and tents there; federal reporting described crowding, poor ventilation and high tuberculosis and pneumonia (Department of National Health and Welfare, 1960; Tester, 2006, in SFU, Inuit Housing, n.d.).
• Government housing, 1959–65. Documented fact: after abandoned 1950s experiments (including a Styrofoam igloo prototype), the Eskimo Housing Loan Program (1959–1965) delivered about 1,200 units: prefabricated "matchbox" houses, then the 16 × 16 ft (about 4.9 × 4.9 m) Rigid Frame house at $420 cost, 125 in fourteen communities in 1960. Most families could afford neither payments nor fuel; most units stayed rentals (Tester, 2009; Robson, 1995, in the SFU survey). Within a generation the snow house ceased to be an ordinary winter dwelling. Tradition — the people's own account: Inuit testimony to the Qikiqtani Truth Commission records repeatedly that the quality housing promised for moving into settlements was not delivered (Qikiqtani Inuit Association, in the housing literature).

Continuance: The Iglu as Taught Practice

• Taught, not lived in. Documented fact: Inuit today live in modern houses; snow houses are built on the land for travel and emergency shelter and taught. Canada's History (2024) records Inuit teachers passing building skills to children. A nine-day Nunavut land programme (Strochlic, 2016, National Geographic) documents elders teaching young adults Mathiassen's same sequence: probe, circle with the pana, spiral, door last.
• The qaggiq built again. Documented fact: for Qaggiq 2021 a festival qaggiq of about 700 square feet was raised at Sylvia Grinnell Territorial Park, Iqaluit, by a team led by master instructors Solomon Awa and Jacopoosee Tiglik, spending two to three days cutting and hauling blocks and several more building (Qaggiavuut!, 2021).
• Its place in school. Documented fact: such teaching sits inside Inuit Qaujimajatuqangit (IQ), which Nunavut's Education Act (2008) makes the foundation of education, igloo building among its organised activities (Government of Nunavut, 2008). Its history thus continues as taught knowledge, on the land where it was developed.

Sources

• Boas, Franz (1888). The Central Eskimo. Bureau of American Ethnology.
• Canada's History (2024). "Ataatasiaq's Legacy."
• Dawson, Peter C. (2002). "Space Syntax Analysis of Central Inuit Snowhouses." Journal of Anthropological Archaeology 21.
• Department of National Health and Welfare, Canada (1960). Arctic housing reporting (in SFU, Inuit Housing: A Survey of the Literature, n.d.).
• Dorais, Louis-Jacques (2010). The Language of the Inuit.
• Government of Nunavut (2008). Education Act.
• Handy, Richard L. (1973). "The Igloo and the Natural Bridge as Ultimate Structures." Arctic 26(4).
• Hodgetts, Lisa M. & Eastaugh, Edward J.H. (2017). Advances in Archaeological Practice 7. (Citing McGhee, 1978.)
• Kershaw, G. Peter, Scott, Peter A. & Welch, Harold E. (1996). "The Shelter Characteristics of Traditional-styled Inuit Snow Houses." Arctic 49(4).
• Mathiassen, Therkel (1928). Material Culture of the Iglulik Eskimos. Fifth Thule Expedition Report, Vol. VII, No. 1, Ch. VII.
• McGhee, Robert (1996). Ancient People of the Arctic.
• NiCHE Canada (2025). "Building Anew: Ephemerality and the Igluvigaq (Igloo)."
• Parry, William Edward (1824). Journal of a Second Voyage. (In Mathiassen, 1928.)
• Qaggiavuut! (2021). "What You Need to Know About Qaggiq 2021."
• Qikiqtani Inuit Association (2010s). Qikiqtani Truth Commission testimony (in the housing literature).
• Rasmussen, Knud (1929; 1931). Fifth Thule Expedition Reports, Vols. VII–V.
• Robson, Robert (1995). Housing in the Northwest Territories (in the SFU survey, n.d.).
• Simon Fraser University (n.d.). Inuit Housing: A Survey of the Literature.
• Strochlic, Nina (2016). "How to Build an Igloo." National Geographic.
• Tester, Frank J. (2009). "Iglutaasaavut (Our New Homes)." Journal of Canadian Studies 43(2).

Flagged claims / weakest-sourced points: (1) Dome's antiquity — no archaeological snow house exists; the Thule link rests on a knife-form inference and plan continuity, labelled Debated. (2) Largest houses — Boas's 7 m figure is second-hand (via Comer), Low's 27 ft house needed battens, Stefansson's reported 30 ft "clubhouse" (via Handy, 1973) is unused in the body. (3) Aqaut's 26 minutes — a single anecdote Mathiassen marks as the exception; never quote as typical. (4) Temperatures — Boas, Mathiassen and Kershaw et al. differ in conditions; no single "igloo temperature" is asserted. (5) Qaggiq 2021 — Qaggiavuut! also gives a diameter inconsistent with its area figure; only area, builders and build time are used. (6) Settlement era — partly secondary summaries (SFU survey), corroborated by Tester (2009). (7) Teaching today — named examples only; no count of builders exists in the sources, and none is asserted.