I asked a question I had heard the answer to a hundred times: is it true that half the cells in my body are foreign bacteria?

The answer turns out to be yes, roughly. Which is strange, because the version everyone repeats is that bacteria outnumber your own cells ten to one, and that version is wrong.

The numbers

In 2016, Ron Sender, Shai Fuchs and Ron Milo went and did the arithmetic properly. For a reference adult male of 70 kg:

  • 3.8 × 10¹³ bacterial cells
  • 3.0 × 10¹³ human cells
  • A ratio of about 1.3 to 1

So bacteria do outnumber your cells, by roughly 30%. About 56% of the cells in you are not you. That is a remarkable fact and it deserves to be repeated. It is simply not the fact that gets repeated.

Two details make it stranger.

The first is mass. All those bacteria together weigh about 0.2 kg, wet, which is around 0.3% of your body weight. By cell count you are barely a majority shareholder in yourself. By weight you own more than 99% of the company. Bacteria are just very, very small.

The second is what your own cells are. Around 84% of human cells are red blood cells, which have no nucleus and no DNA. When people say half of your cells are bacterial, the human half being described is mostly a bag of haemoglobin with no genome in it. The comparison was never quite the thing we imagined.

Where 10:1 came from

The 10:1 ratio does not trace back to a study. It traces back to a single sentence in a 1972 paper by Thomas Luckey, in which he did a back-of-the-envelope calculation: assume 10¹¹ bacteria per gram of gut contents, multiply by roughly one litre of alimentary tract capacity, and you get 10¹⁴ bacteria. Divide by the then-assumed 10¹³ human cells and you have ten to one.

It was never presented as a measurement. It was an estimate, of the kind a scientist scribbles to check whether an idea is plausible.

Sender, Fuchs and Milo went looking for the source and found that “all papers regarding the number of bacteria in the human gastrointestinal tract that gave reference to the value stated could be traced to a single back-of-the-envelope estimate”. They describe the result as having achieved, through sheer repetition, “the status of an established common knowledge fact”.

Forty-four years. Textbooks, TED talks, popular science books, science journalism, and my own confident retelling of it at a dinner table. Nobody rechecked, because it sounded right and it came with a number, and numbers feel like evidence.

Why this keeps happening

I wrote recently about a fake fact concerning bats, a TikTok claim that bats are terrible at echolocation and merely heal quickly from crashing into things. It was invented. This one is different in an important way, and worse.

The bat story was made up by somebody. The 10:1 ratio was produced honestly, by a real scientist, in a real journal, clearly labelled as an estimate. Everything that went wrong happened afterwards, in the citing.

A rough figure got repeated until the roughness wore off. Each citation was individually reasonable, because each author was citing a published source, and the published source was citing a published source. The chain was sound at every link and the thing at the end of it was a guess from 1972.

That is not a failure of any individual scientist. It is a property of how citation works. And you cannot fix it by being sceptical of the last person who told you, because the last person who told you was right to believe it.

So what should you say at dinner parties?

The honest, better version:

Bacterial cells in your body slightly outnumber your own, about 1.3 to 1, so a bit more than half the cells you carry around are not yours. They weigh about 200 grams in total, roughly a punnet of blueberries. Most of your own cells are red blood cells with no DNA. And the famous ten-to-one statistic was a rough calculation from 1972 that everyone repeated for four decades without checking.

It is a worse party trick and a much better story. The number was never the interesting bit. The interesting bit is that we all knew it.

Sources