The residents
They Made You
A specimen in a natural history museum carries a label, and somewhere on that label is where the thing was found: a bay, a ridge, a depth, a date. It is the least glamorous line and it is the one that makes the object evidence rather than decoration.
Every label in this room has one. They all name the same locality.
How to read this wall. The room is ordered by how far in the organism got — the surface, then the wet linings, then the tube, which is where nearly all of them are. The last two sections are not collection localities in the same sense, and the wall says so at the point it stops being one.
Before the labels, the count
How many of them, how many of you
Counted, 2016
About 1.3 bacterial cells for every human cell
3.8 × 1013bacterial cells3.0 × 1013human cellsand about 0.2 kg of them all told
Somebody finally counted. Thirty-eight trillion bacterial cells, thirty trillion human ones, and the two numbers are the same size.
The other half of that paper is stranger and gets quoted less: about nine in ten of your own cells are blood cells. Weigh it all and the bacteria come to roughly two hundred grams — less than one of your lungs.
A count for one 70 kg reference body, not a measurement of any particular person. Both numbers move with body size, and the bacterial figure is dominated by the colon — the rest of the body contributes a rounding error to it.
Asserted, 1972 — retired
Ten bacterial cells for every human cell
Drawn at the length it claims. It continues for about 4 times the width of this frame.
You have almost certainly met this one. It is in the popular books, in the talks, and on the walls of other museums: you are only ten per cent human.
It is drawn here at the length it claims, which is why it leaves the frame.
Retired, not disproved — there was never a measurement to disprove. The figure traces to a 1972 estimate of bacteria per gram of gut contents, set beside a round guess at the number of human cells, and it was repeated for forty years by people who assumed somebody had checked.
The surface4 labelstechnically outside
Two square metres of salt, oil and dead keratin, cooling and flaking away underneath them. Everything here is holding ground against something else that wants it.
Collectedthe sebaceous follicle, a few millimetres down
Cutibacterium acnes
It farms the oil, in the dark, at the bottom of a pore
~106cells per square centimetre of oily skin
Human skin is a desert with an oil field under it. Salty, dry, and cooling, and almost nothing can make a living up there — but every follicle is a shaft leading down to a gland that pumps sebum continuously, out of the air, at body temperature. That is the habitat, and this is the organism that took it.
It is an anaerobe. It sits below the reach of oxygen, splitting the fats in sebum into free fatty acids and propionic acid, which is where its old genus name came from. The acidity of human skin — the thing every cosmetics counter calls the acid mantle — is substantially its exhaust.
One species name covering strains that behave differently enough to be worth separating; the lineages on clear skin and the lineages in an inflamed follicle are not the same, and the name predates anyone knowing that.
Known since 1897Described in 1897 and named for the least interesting thing it is involved in. It has changed genus twice since — Propionibacterium in 1946, Cutibacterium in 2016 — and the current name is the first one that says where it lives.
Collectedthe forearm, and the inside of a nostril
Staphylococcus epidermidis
It evicts its own cousin from the premises
Two staphylococci want the same warm, salty, protein-rich surface. One of them builds a film to hold ground. The other secretes an enzyme that takes the film apart, along with the proteins the film needs to attach in the first place.
Skin commensals turn out to carry a whole armoury of these compounds, aimed at close relatives, made for reasons that have nothing to do with us. We are the ground being fought over, and the outcome of that fight is one of the better things that happens to a person in a day.
Not every strain makes Esp, and the effect is territorial rather than medical — this is one organism dismantling a competitor's construction, and the benefit to the person it is happening on is a side effect of that.
Known since 2010Shown in 2010: people whose nasal Staphylococcus epidermidis secretes a serine protease called Esp are markedly less likely to be carrying Staphylococcus aureus at all. Applying the protease clears an established film.
Collectedthe scalp, and every other oily surface
Malassezia
It gave up making its own fat, because ours was reliable
This one is a fungus, which makes it the odd resident on a surface otherwise dominated by bacteria — and it is the dominant fungus on nearly every adult human being.
Somewhere in its evolution it dropped the enzyme complex that builds fatty acids from scratch. That is not a small loss. It is one of the central pieces of metabolic machinery, and living things do not discard it unless the alternative is dependable. The alternative was a mammal. It now secretes lipases, takes the fats it needs from the sebum, and cannot survive anywhere a mammal is not.
A genus, not a species — which Malassezia is on a given head varies by person and by site. Its presence on adult human skin is close to universal in the populations that have been surveyed, which is not the same as universal.
Known since 2007The genome, published in 2007, showed the missing piece: no fatty acid synthase. It cannot build its own lipids at all, which is why it is found on warm-blooded animals and essentially nowhere else on Earth.
Collectedthe base of an eyelash
Demodex folliculorum
An animal, halfway through becoming a part of us
0.3millimetres long — the largest organism in this room, and still invisible
An eight-legged animal, a third of a millimetre long, spends its entire life head down in a follicle at the root of a human hair. It is an arachnid — a relative of spiders and ticks — and at 0.3 mm it sits right at the edge of what an unaided eye can resolve, which is the boundary Powers of Ten draws with a dashed ring.
The genome is the interesting part. It has one of the smallest gene complements measured in its group: it has been shedding machinery — for stress responses, for building certain body parts, for making its own supplies — because a human follicle is a stable, warm, well-provisioned place and the genes were no longer earning their keep. Reduction on that scale is what a genome does on its way into a permanent arrangement. It is the same road the organism two sections down took, and finished.
Prevalence figures depend entirely on the sampling method, and a face that yields none may only mean the sample missed. That the transition to symbiosis is under way is the paper's argument from gene loss, not a settled result.
Known since 2022Described in 1842 from material taken out of a follicle. The genome, sequenced in 2022, is what changed the story; the paper's title is *Human Follicular Mites — Ectoparasites Becoming Symbionts*.
The wet surfaces3 labelstechnically outside
Warmer, and stocked. On a mucous membrane the host is not merely tolerating a population — it is running supplies out to it.
Collectedthe plaque on a molar
Corynebacterium matruchotii
It builds the scaffolding, and a city grows on it
The film on your teeth is not a smear of bacteria. It is a structure, and it has a plan.
Long filaments of this organism radiate outward from a dense core like the spines of a hedgehog, and other species occupy fixed positions along them. Oxygen users at the tips, where oxygen is. Anaerobes packed at the base, where it is not. A ring of one genus part-way out, a shell of another at the surface. Two hundred microns across, on a tooth, arranged.
Sequencing a sample tells you who is present. It took someone taking a picture to show that they were standing in formation.
The hedgehog is one arrangement among several, described from plaque scraped off healthy volunteers. How general it is across mouths, ages and diets is precisely what the imaging is now being used to ask.
Known since 2016In 2016 a Boston group stained dental plaque for nine taxa at once and photographed it where it sat instead of grinding it up first. The architecture had been there the whole time; nobody had looked at it intact.
Collectedthe back of the tongue
The nitrate reducers of the tongue
They perform a chemical step your own cells cannot
No human cell can reduce nitrate to nitrite. Plenty of bacteria can, and some of them live on the back of your tongue.
What makes this remarkable is the plumbing. Nitrate from vegetables is absorbed in the gut, circulates, and is then actively concentrated by the salivary glands and sent back up into the mouth at many times its concentration in blood. It is delivered to the bacteria. They reduce it, you swallow, and the nitrite becomes nitric oxide — the signal that relaxes the muscle in blood vessel walls.
A human body evolved a delivery route to a chemical step it cannot perform, because something living on it could.
A consortium rather than a species — Veillonella, Rothia, Actinomyces and others — and which of them matters varies between mouths. The blood pressure effect is a few millimetres of mercury, which is real, modest, and about the size of a mild drug.
Known since 2013Shown in 2013 by taking them away: rinse with an antiseptic mouthwash for a week and blood pressure rises, along with a measurable collapse in the nitrite in saliva and plasma. The step that goes missing is bacterial.
Collectedthe vaginal mucosa
Lactobacillus crispatus
It holds a whole surface at pH 4
The epithelium supplies glycogen. The organism ferments it to lactic acid and holds the surface at about pH 4 — acid enough that most of what arrives there cannot establish. One species, doing chemistry, running a whole environment.
Set against other primates this is strange. Monkeys and apes that have been surveyed carry diverse vaginal communities at close to neutral pH, with no lactobacillus in charge of anything. Humans are the outlier and nobody has a settled explanation for it.
For most people this is also the first community they ever meet. An infant born vaginally arrives coated in its mother’s; one born by caesarean is colonised instead by whatever is on the skin in the room.
Lactobacillus dominance is one of several stable community types, and a community without it is not thereby unhealthy — that inference is common in popular accounts and the surveys do not support it.
Known since 1892Named from vaginal flora in 1892 and treated as a curiosity for the next century. The question of why humans are unusual in having it was not put properly until 2016.
The tube5 labelstechnically outside
Your gut is a hole that goes through you. Everything in this section is living in a cavity that is, strictly, still the outside world — and it is where almost all of them are.
Collectedthe lumen of the large intestine
Bacteroides thetaiotaomicron
It eats the part of dinner your genome cannot read
Your own genome carries a handful of enzymes for breaking down complex plant sugars. This organism carries hundreds, and it can sense which one has arrived and switch its machinery over to it.
Everything on your plate that gets called fibre passes through the stomach and the small intestine untouched by anything of yours. It is not food yet. It becomes food a metre and a half later, here, and what comes back to you is what this organism and its neighbours leave behind.
Counting sugar-cleaving enzymes depends on how the families are drawn, so the comparison is stated as an order of magnitude rather than as a number. It also switches to grazing host mucus when the plant sugars run out, which is why fibre in a diet is not only about the fibre.
Known since 2003Its genome was published in 2003, and the surprise was the size of the toolkit: page after page of enzymes for taking apart plant polysaccharides, in an organism that has never been near a plant.
Collectedthe mucus layer against the colon wall
Akkermansia muciniphila
It eats you, and that is the arrangement
The wall of your colon is coated in a layer of mucus that your own cells secrete without stopping. This organism lives in that layer and eats it.
That sounds like an attack and it is closer to grazing. The layer is renewed continuously, the grazing keeps it turning over, and a mucus layer that turns over is a mucus layer that works. There is a well-fed population living on a surface you are constantly rebuilding, and the rebuilding is the point.
Its discovery is the part worth keeping. For a century, “unculturable” was a statement about organisms. It was usually a statement about the menu.
Reported abundance varies widely between studies and populations. The associations with metabolic health are real and repeatedly observed; which way the causation runs is not settled, and the popular framing of it as a beneficial supplement runs well ahead of the evidence.
Known since 2004Isolated in 2004 by offering an enrichment culture nothing to eat but mucin and seeing what turned up. It had been one of the more abundant organisms in the human gut the entire time, and it was invisible because nobody had put the right meal in front of it.
Collectedthe gut of a breastfed infant
Bifidobacterium longum subsp. infantis
Your mother made food for it, and handed it over
~200distinct sugars in human milk that a human infant cannot digest
Human milk contains around two hundred different complex sugars, and after fat and lactose they are the third most abundant thing in it. An infant cannot digest a single one of them. They pass the stomach intact, pass the small intestine intact, and arrive in the colon exactly as they were made.
They are not for the baby. They are for this organism, which has the enzymes to take them apart, and which arrives in a breastfed gut and displaces most of what else was there.
A human body manufactures food for a bacterium, packages it in milk, and delivers it to a newborn. The sugars and the enzymes that open them have been shaped against each other for long enough that neither makes sense alone.
The count of distinct oligosaccharides depends on the mother and on the method, and the range in the literature is wide. Not every breastfed infant carries this organism, and which strains do the job best is an open question.
Known since 2008Its genome, published in 2008, carries a dedicated cluster of genes for exactly the sugars in human milk that a human baby has no enzyme for. The fit is too specific to be a coincidence and it is read as one.
Collectedthe colon, up against the lining
Faecalibacterium prausnitzii
The cells of your colon run on its waste
Nearly every cell in your body is fed from the bloodstream. The cells lining your colon are not. They take most of their energy from butyrate — a short-chain fatty acid produced by bacteria fermenting the fibre in the lumen — absorbed directly from the cavity they face.
It is a waste product. The organism has finished with it. The wall it happens to be sitting against burns it for fuel, and the arrangement is old enough that the wall has stopped providing for itself.
Raise a mouse with no bacteria at all and those cells go hungry in a gut full of food.
Butyrate comes from a guild of fermenters and not from this species alone; it is named here because it is among the most abundant of them in a human colon. The direct evidence is the mouse experiment, and the human case is inferred from it.
Known since 2011In 2011 the colon cells of germ-free mice were found sitting in an energy-starved state, running autophagy, and butyrate on its own put them right. The lining had been running on a bacterial by-product all along.
Collectedthe colon
Bacteroides fragilis
It finishes an immune system that arrives unfinished
A mouse raised with no bacteria has a defective immune system. Not an idle one — a malformed one: the lymphoid tissue is underdeveloped and the balance of T cell types is wrong in a way that persists.
Colonise it with this single organism and the balance corrects. Purify one polysaccharide from the organism’s coat and give it that instead, and the polysaccharide alone does it.
Which says something about what an immune system is. It does not merely tolerate the residents. It arrives expecting them, waits to be instructed by them, and without the instruction it does not finish assembling.
The mouse work is unusually clean for this field. Carrying it across to human immune development is inference, and how far it carries is the open question rather than a settled finding.
Known since 2005In 2005 a single molecule off its surface — polysaccharide A — was shown to correct the immune defects of a germ-free mouse by itself. One organism, one purified sugar, a measurable result.
The wall crosses a line here. Everything above is standing on a surface. Skin is outside. A mouth is outside. The gut is a tube that runs through a body, open at both ends, and its contents have never been inside anything. twelve of the seventeen labels in this room are organisms that could, in principle, be washed off.
The remaining five cannot. They are not lodgers, and there is no locality to give for them beyond the one they became.
Inside the cell3 labelsnever left
One organism crossed the line, about two billion years ago, and the terms of the arrangement have never been renegotiated.
Collectedthe cytoplasm of every cell you have
An alphaproteobacterium
It got in, and it never left
37genes still on its own chromosome, kept apart from your other twenty thousand
Around two billion years ago a bacterium ended up inside another cell and was not digested. Every organism in the portrait hall is descended from that cell. So is every plant, every fungus, every animal, and you.
It never fully surrendered. It still keeps its own chromosome — a closed circle of 16,569 bases carrying 37 genes, sequenced in 1981 — separate from the genome in your nucleus, replicated on its own schedule. It still divides by pinching in two. Its inner membrane still has the lipid that bacteria use and your other membranes do not.
The applied gallery ends on this organism, because it is the last thing on the list of what these things have built. This room begins there, because it is also the first thing that ever got inside. The timeline corridor puts the event two billion years back.
Which alphaproteobacterial lineage it came from, and how complex the host already was, are both live arguments. That the event happened, and happened once, is not one of them.
Known since 1967Proposed in 1905 and dismissed. Argued back into the mainstream in 1967 by Lynn Margulis, whose paper was turned down by something like fifteen journals before it ran. Settled afterwards, by sequence.
Collectedinside the mitochondrion, alongside its chromosome
The mitochondrial ribosome
Still bacterial enough that antibiotics find it
Your cells build proteins on one kind of ribosome. Your mitochondria build a handful of their own proteins on a different one, which they brought with them.
It is still close enough to its origin to be recognised by the wrong things. Several classes of antibiotic work by jamming the bacterial ribosome, and in a human cell they can find this one too — the resemblance turns up as a side effect, in a laboratory result, in a caution about which compound to use in a cell culture. The family likeness is not a metaphor. It is a drug-binding site, conserved for two billion years.
"Bacterial" here means descended from the bacterial ribosome and still recognisably it — not identical to it. The human version has lost most of the RNA a bacterial ribosome uses and replaced it with protein.
Known since 2015Solved at near-atomic resolution in 2015, which showed how much of the bacterial machine is still standing and how much of it has been rebuilt out of protein since.
Collectedthe egg you came from
Your mother's mitochondria
One unbroken line, and none of it is your father's
Your nuclear genome is a shuffle of two people, and theirs of four, and so on backwards into a fog.
The genome in your mitochondria is not shuffled at all. It came from the egg. The egg got it from your mother’s egg, which got it from your grandmother’s, in an unbroken sequence of cell divisions running back past every one of those people to the cell that swallowed the bacterium in the first place. A sperm’s mitochondria are destroyed after fertilisation. That whole half of your ancestry is simply not present in this part of you.
Half of what you are made of has a family tree with one parent per generation, and it is older than the animals.
Rare cases of paternal transmission have been reported and argued over since. The rule is overwhelming rather than absolute, which is why it is worth stating as a rule.
Known since 1980Shown in 1980 by following mitochondrial DNA variants through families. The father's contribution does not show up — not diluted, not minor, absent.
Inside the genome2 labelsnever left
And then there is the sequence itself, which has been written into by things that were passing through.
Collectedabout eight per cent of your chromosomes
Endogenous retroviruses
They infected an ancestor's germ line and were inherited ever after
~8%of the human genome is sequence of retroviral origin — against about 1% that codes for protein
A retrovirus copies itself into the chromosome of the cell it infects. Do that to a sperm or an egg cell and the copy is inherited — by that individual’s children, and by everyone descended from them, for as long as the lineage lasts.
This has happened to your ancestors many thousands of times. About eight per cent of your genome is what is left of it, which is roughly eight times the fraction that codes for protein. Most of it is wreckage, mutated past working, no longer able to leave.
Some of it works.
The 8% figure is for long terminal repeat retroelements, of which endogenous retroviruses are the bulk. The true share is higher and unknowable: an old enough insertion decays past the point where it can be recognised as one.
Known since 2001The first draft of the human genome, in 2001, made the accounting possible. The retroviral fraction came out several times larger than the fraction that codes for anything.
Collectedchromosome 7, and the surface of the placenta
A retrovirus, given a job
It builds the interface between a mother and a child
The outer layer of a placenta has no cell boundaries in it. It is a single continuous sheet of cytoplasm with thousands of nuclei in it, formed by separate cells fusing together, and it is the surface across which everything a foetus receives has to pass.
The protein that fuses them is a viral envelope protein. It is what a retrovirus uses to merge its own coat with the membrane of a cell it is about to enter — the same molecular trick, the same gene, still in the genome, still switched on, and now doing something else entirely.
Mice did this with a different virus. Cats with another, cows with another, marsupials with another again. The organ that defines placental mammals has been assembled from captured viruses repeatedly and independently, which means it is not an accident that happened to us once.
If you were born, this is how.
Different placental lineages captured different viruses independently for the same job. That is the strongest part of the argument and it is also the part that says the human gene is one instance rather than the general case.
Known since 2000In 2000 the protein that fuses the outer layer of the human placenta turned out to be the envelope gene of an endogenous retrovirus. In 2009 the mouse equivalent was knocked out to find out what it was for: the placenta fails and the embryos die.
One more number
All of it is news
Every organism in this room has been doing its work in every human being who has ever lived. The labels describing that work are almost all younger than the people reading them: thirteen of the seventeen were published this century, and the median is 2005. The oldest is 1892. The newest is 2022.
There is no particular reason to think the wall is finished. One of the more abundant organisms in the human gut was grown in a dish for the first time in 2004, by somebody who worked out what it wanted for dinner.
On this room in particular. There is a version of this exhibit that goes wrong in its first sentence, and it is the version most museums build: a tour of a human body with the organisms as findings on it. It makes the visitor the subject, turns the residents into symptoms, and ends up being about hygiene.
The correction is the one the applied gallery makes: whose name is on the label. The title of this room is a sentence, and the organisms are its subject. Not one of them is doing any of this for you. They were solving their own problems — finding food, holding ground, getting out of the weather — and a human being turned out to be a reliable place to do it. You are the locality. They are the collection.