The Outside, Brought In

Is house dust mostly dead skin? It is one of the most repeated facts about the home, and we could find no measurement behind it. What there is instead: a ceiling that takes one line of arithmetic, and a careful published model of floor dust that puts about three fifths of it outside the house to begin with. Here is that model, rebuilt from its own equations so you can move its dials, including the places where the paper's numbers do not agree with each other.

1. The claim, and the one thing you can say without a model

The popular version comes in several sizes: "most", "70 to 80 percent", or on Wikipedia today, "Dust in homes is composed of about 20–50% dead skin cells," cited to a 1981 book on house-dust biology that we could not read. The Museum of Hoaxes called the strong form "a widely repeated factoid" as long ago as 2009.

One ceiling needs almost no model. Dust scientists split dust into an organic part and a mineral part by heating it until the organic matter burns away (the "loss on ignition"). Skin flakes are organic tissue, so they burn. So skin can never be a larger share of dust than the dust's whole organic share, and it is really a smaller one, because the organic share also holds everything else organic: in the list of the paper this page rebuilds, "lint, skin particles, organic fibers, food debris, etc." For homes, that paper puts the organic share at about 40 percent:

“A prominent feature of indoor dust is its OM content, with levels of about 40 wt% in residential housing.”Layton and Beamer (2009), Environmental Science & Technology 43, 8199

That is a typical figure, not a law (some homes will be higher, and heating also drives a little water and carbonate out of minerals, which slightly overstates the organic part). But on it, "70 to 80 percent" is impossible and "most" is impossible. Everything below is about where the other 60 percent comes from, and how big the skin's piece of the 40 might be.

2. The floor's ledger

In 2009 David Layton and Paloma Beamer (of the University of Arizona) published a mass balance for the dust on a home's floors, built to trace how arsenic and lead from outside end up where children crawl. It treats the house as two boxes, the air and the floor. Soil arrives on shoes. Outdoor particles leak in with the air and settle. People and their things add organic matter. Walking kicks floor dust back into the air, where most of it settles again and some leaves through the gaps. Cleaning takes the rest away.

Its two defining equations, fed the paper's values for homes in six Midwestern states, reproduce the eight dust-mass numbers the paper prints from them (the first table under the instrument). That is the model you are holding.

At the paper's values, 50.1% of what reaches the floor is soil walked in, 12.5% is outdoor air particles, and 37.4% is the indoor organic input. The outside total is 62.6%, which matches the ACS press release for the paper ("over 60 percent of house dust originates outdoors") and Beamer's own words to NPR: "Two-thirds comes from both soil tracked in, and the outdoor air particles."

Now the part that matters for honesty. That 60-odd percent is not a second finding on top of the 40 percent organic. It is the same finding. The paper did not measure organic matter in these homes: it took the typical 40 percent as a reference and worked its organic input and its track-in back from that. And the model has no indoor source of mineral dust (no plaster, no brick, no cat litter), so whatever is not organic must have come from outside. Set the organic share at 40 percent and the outside share has to land near 60, plus the little organic matter that soil and outdoor particles carry in with them. What the model adds is the split of that outside share into soil and air, and that split is not settled:

The paper's equations 15 and 16, which work out the organic input and track-in from arsenic measurements, give 0.033 and 0.035 g a day when fed the arsenic values printed in its main text, not the 0.074 and 0.099 it prints. Its equation 6 gives an arsenic input from outdoor air of 0.37 µg a day, not the 0.67 behind its abstract's "nearly 60%" (the rebuild's share is 69% with the recomputed fluxes and 44% with the printed ones). Its Supporting Information, which we did not read, may hold the inputs it actually used. Every reading puts the outside total between 62.6% and 66.0%, because the 40 percent pins it. The soil share runs from 37.6% to 50.1%.

Now press little soil walked in. In the model, a home at the 10th percentile of track-in, with everything else unchanged, has an outside share of 40.0% and an organic share of 63.0%, so the ceiling on skin sits above half. The model cannot tell you how much of the organic part is skin, which is why the skin dial is labelled a guess: move it to 100 and you have assumed that nothing else organic ever reaches the floor, no lint, no fibres, no crumbs.

3. How much skin do you shed? Nobody agrees

If you shed a lot, perhaps skin fills the organic share. The published figures do not settle it: they spread over a factor of about two thousand, and part of the spread is that they count different things. Some count the whole outer layer of skin coming away; others count only the flakes that get into the air, and one of those (Dillon) cites the other (Clark). The cell counts below are turned into grams using one squame (a shed skin flake) at about 2.5 nanograms, the figure Weschler and colleagues give.

The big numbers are for the whole outer layer. Weschler and colleagues, summarising the dermatology literature:

“Humans shed about a thousand cells/cm2/h or roughly 5 × 10⁸ cells/day. A typical desquamated skin particle, referred to as a “squame”, is roughly 40 × 30 × 2 μm, with a mass of ∼2.5 ng. This translates to a shedding rate per human of 30 to 90 mg skin flakes/h.”Weschler et al. (2011), Environmental Science & Technology 45, 3872 (reference numbers omitted)

Note the unit. Thirty to ninety milligrams an hour is 0.72 to 2.16 grams a day. Their cell count times their flake mass gives 1.25 g, inside that range. The small numbers are for skin that gets into the air: “As many as 10⁶–10⁷ of these skin particles can be dispersed from the body in 24 h,” in a 2009 review by Clark and de Calcina-Goff, which puts those particles off the body by “movement and the rubbing actions of clothing.”

Now set that beside the ledger. The model's whole indoor organic input, skin and lint and crumbs together, is 0.074 g a day for an entire house. One person shedding at the rate Weschler cites makes about 10 to 29 times that. So, on the model's central values, and taking that organic input at face value, at most 3.4% to 10.3% of the skin one person sheds could be arriving in the floor dust, even if every organic crumb on the floor were skin. The paper's main text does not give the number of people per home; move the dial and the fraction only shrinks.

Name what that rests on. The organic input was worked back from the assumed 40 percent, not measured (and recomputed by the paper's own equations it is smaller still, 0.033 g). It assumes nothing organic decays on the floor; if skin is eaten by mites or oxidised by ozone, more has to arrive to keep the floor 40 percent organic, and the bound loosens. It is a central value with no spread given. And it describes wipe samples of hard floors, so skin held in carpets and bedding is outside what was sampled.

With those named, the arithmetic still turns the question over. How much you shed was never the limit. The limit is where it goes, and the model does not say: it does not follow skin into the bed, the laundry, the shower drain or the vacuum's filter, and we found no study that weighs those paths. "We shed a lot of skin" and "the floor dust is mostly skin" cannot both be true of this model's floor.

4. The number we refused to compute

One chemical fingerprint of skin in dust is squalene, an oil that skin makes. In 2011 Weschler's group measured it in dust vacuumed from surfaces above the floor in the bedrooms of 500 children (aged 3 to 5) in and around Odense, Denmark: a geometric mean of 32 µg per gram. They cite Clark and Shirley's 1973 measurement of squalene in skin scales scraped from a person, 10,000 µg per gram (1 percent). They expected a lot of skin: in their words, skin flakes “are anticipated to be major constituents of indoor dust.”

Divide one by the other and you get 0.32%, which would say dust is a third of one percent skin. That number is at best a lower bound, and it is not on this page as an answer. Squalene reacts quickly with ozone. On the paper's own rough gas-phase arithmetic, at an average indoor ozone level of 5 parts per billion “the half-life of squalene is roughly half an hour”; if anything like that holds in settled dust, a flake that has lain for days keeps little of it. The measurement shows that skin is there, and says little about how much. The authors did not turn it into a percentage, and qualified the half-life at once: “We must be cautious regarding such an analysis.”

5. How the answer travelled

Layton and Beamer's paper was about arsenic and lead, and its abstract's headline number, “nearly 60% of the As input to floor dust”, is about arsenic, not dust. The American Chemical Society's press release for it put the dust finding plainly: “They found that over 60 percent of house dust originates outdoors.” A widely read explainer at Live Science (the version online today is dated 2022, but the Internet Archive's index lists the same address in 2013, so the article may be older; we could not open that copy) summarises the same 2009 study like this:

“According to a 2009 study of house dust in the U.S. Midwest, 60% of the components of the dust came from indoors, and 40% came from dirt and other materials tracked in from outside.”Live Science, "Is house dust mostly dead skin?" (dated 26 May 2022)

That is the finding turned inside out. In the press release and in Beamer's own words, over 60% (two-thirds) comes from outside; in the rebuild, the indoor organic input is 37.4% of what reaches the floor. The explainer did reach the right conclusion, that the 70 to 80 percent figure is "likely not true for most houses", while reversing the finding it rested on. The myth and its correction both passed through the same study, and the correction came out with the numbers the wrong way round.

6. What nobody has measured

We looked for a study that weighs the share of settled household dust that is skin, directly, and did not find one. Studies of skin in airborne particles exist from the 1970s (Clark and Shirley, 1973, which we could not read), and a 2011 paper on foot-and-mouth disease (Dillon) describes skin as “a significant fraction (1–10%) of measured indoor and outdoor aerosols and indoor dust”, citing Clark's airborne studies and two studies of dust in livestock buildings. That is not a measurement of your floor.

The model here is a model: fitted to wipe samples of hard floors in one region's homes, with a single organic flux that lumps every indoor source together, and no indoor mineral dust. Carpets hold more dust than bare floors and give more of it back to the air. An older model summarised by the US EPA put outdoor soil at 31.3% of indoor dust by weight (Calabrese and Stanek, 1992, which we did not read), against the rebuild's 37.6% to 50.1%. Homes will differ, and the dials show how much it matters when track-in or the organic input changes. What survives is the ceiling from section 1, and, at the Midwest values, the observation from section 3: the floor dust is not where most of your skin goes.