The Verification Venue · pointed at a thing everyone gets wrong
The Head That Was Never the Radiator
Every cold snap, someone says it: you lose 40 to 50% of your body heat through your head. You don't. The head is about 9% of your skin, and it sheds about 9-10% of your heat. The scary number was manufactured, and you can manufacture it yourself, below, in about ten seconds.
Heat leaves a bare body roughly in proportion to how much bare surface you point at the cold, not because of anything magic about the scalp. So the honest question isn't "does the head leak?" It's "what fraction of the exposed surface is the head?" Cover everything else, and the answer stops being 9%. Watch.
Everything bare except the groin: the setup behind the honest ~10% figure.
Tap a region to cover / uncover it
of total heat loss leaves through the head
This one dial is the whole trick. The better the insulation (lower %), the more the bare head dominates. Arctic suits insulate hard, which is exactly why the bare head there swallows 40%+ of the loss.
Left alone in a swimsuit, the head is what it looks like: ~9% of the surface, ~9% of the loss. Slide the insulation good enough and cover everything but the head (the Arctic survival suit preset) and the same arithmetic spits out 40-50%. Nothing about the head changed. You just took away every other place heat could leave.
The check: every % recomputed in front of you
The share of total heat loss through a region is just its bare-surface fraction, weighted by whether it's covered, normalised over the whole body:
Surface-area fractions area are the clinical rule of nines (head&neck 9%, each arm 9%, front torso 18%, back 18%, each leg 18%, groin 1%), with the hands split off at the Lund–Browder value. They sum to exactly 100. For the setup you've built right now:
| region | area % | state | weight w | area·w | share of loss |
|---|---|---|---|---|---|
| total | 100 | – | 100% |
Same offline check runs in research/heat-lost-through-your-head/verify-heat-lost-through-your-head.mjs. It recomputes every number here two independent ways and exits non-zero if any drifts.
Where the 40% came from
The trail runs back to a mid-century military cold-exposure study (popular retellings split between a 1950s cold trial and a 1970 US Army survival manual repeating the figure). Volunteers were bundled into arctic survival gear (insulated neck-down) and left bare-headed in the cold. Yes, most of their measured heat loss came out of the head. Of course it did: the head was the only part not wrapped in insulation. Put those same people in swim trunks, the BMJ authors noted, and the head is back to about 10%.
| Claim | What's actually true | Number |
|---|---|---|
| "You lose 40-50% of body heat through your head." | Only true when everything else is insulated. Bare, the head is ordinary skin. | ~9% |
| Head's share of body surface area | Rule of nines: head+neck. Lund–Browder puts the adult head nearer 7%. | 7-9% |
| Head's share of heat loss, bare (BMJ 2008) | Matches its surface share: nothing special. | ~10% |
| Arctic-suit setup (myth's source) | Reproduced by this page's own arithmetic at 12% clothing leak-through. | ~45% |
| Whole-head cold-water submersion, core cooling (Pretorius 2006) | Core cooled ~42% faster, far more than the head's share of total heat loss. The head punches above its surface weight. | +42% |
So the hat is right, for the wrong reason
Here's the part the smug debunk usually skips. The head is not quite ordinary skin. When you get cold, your body defends its core by slamming shut the blood vessels in your arms, legs and hands (vasoconstriction), so those surfaces cool toward the air and stop bleeding heat. The skin of the head and face barely does this (Froese & Burton measured the head's insulation as nearly constant regardless of how cold it got). So in real cold, while everything else throttles down, the head keeps radiating at close to full tilt, and can then lose a share bigger than its 7-9% of surface.
Flip the honest twist toggle above and you'll see the head's share climb past 9% with nothing covered at all, driven purely by the asymmetry, not by clothing. Whole-head submersion in 17°C water cooled the core ~42% faster (Pretorius 2006), much more than the head's slice of total surface heat loss would predict. So: wear a hat in the cold. Not because your head is a chimney (it isn't) but because it's the one part that stays uncovered and refuses to shut itself down while the rest of you bundles up and clamps off.
What's idealised here, and what's exactly true
Exactly true. The surface-area fractions are the published rule of nines (with hands at the Lund–Browder value) and sum to 100 by construction. The identity at the heart of the debunk is exact: if every region is equally bare, each region's share of heat loss equals its share of surface area, so the bare head's share is its area, ~9%. That is arithmetic, not a measurement.
An idealisation. "Heat loss ∝ exposed surface area" is a simplification. Real heat loss depends on skin temperature, local blood flow, air movement (convection), humidity, radiation and evaporation, not surface area alone. This page uses area as the first-order proxy the myth itself assumes, then shows that even on the myth's own terms the head is ~10%, not ~45%.
Named free choices. Two dials are not measured constants: the covered leak-through r (how much heat still escapes through clothing: default 12%, the value that lands the arctic setup near the classic 40-45%) and the cold-skin vasoconstriction factor v for the second layer (default 50%, illustrative). The direction of the vasoconstriction effect is well-supported; the exact magnitude on this page is not a claim.
Fuzzy origin. The dating of the source experiment is genuinely uncertain: sources split between a 1950s military cold trial and a 1970 US Army survival manual repeating the figure, so this page says "mid-century military cold-exposure study," not a hard date.
The practical takeaway stays true. A hat matters in cold weather. General cold-safety advice here is deliberately non-specific and not medical guidance.