Where do people come from?

Published 30 October 2023

Back to writing
Contents (click to toggle)

6,748 words • 34 min read

When someone is born, chemical parts once scattered across the world converge to form a human.

Days, years, and decades before, those atoms were always somewhere in the world. I recently got curious about this: where do people come from? Or rather, looking ahead — where are future people?

I thought to write this out of sheer curiosity, but I like the picture that emerged. It’s a picture in which we’re not optionally relying on the products of the vast industrial and agricultural processes humans built — we are almost entirely made of them.

By elements we are mostly oxygen, carbon, then hydrogen and nitrogen; and by molecules we are mostly water (65%), then various proteins and lipids. That’s by weight. Since water molecules are much simpler than most other molecules, ~99% of the molecules in our body are H2OH_2O. So let’s start with water.


Of course during gestation the mother is the source of water and everything else. Assuming most of this derives from (sidenote: The water in bottled drinks is often and perhaps mostly just tap water, despite marketing language.), then most proximally (up to around a day before) that water will have been flowing through a network of water distribution pipes.

Then the possibilities branch out depending on how water treatment and storage works for the area. It might have sat in a water tower for a while, but likely (very roughly a day before) it will have passed through a water treatment plant.

Before that, the water was being stored in some large (sidenote: If you’re near the coast, why not pump water from the sea and remove the salt? Mainly because it’s almost always more expensive than letting evaporation do the desalination for you, and importing the water. Crudely speaking, salt dissolves in water really well: since the salt molecules aren’t much bigger than water molecules, you can’t filter it out. Alternative methods (leading among them being a kind of ‘reverse osmosis’ technique) are energy-intensive. Still, Israel gets the majority of its drinking water from desalination, and costs are decreasing.) — possibly an artificial reservoir, possibly a natural surface water source like a lake or river, or possibly underground in an aquifer.

If it came from a reservoir or a river, the water likely spent weeks to months there between arriving and being processed. If from a lake, closer to a century. And if the water came from underground, then it might have been there for millennia before resurfacing through a spring or well.

But virtually all these water sources were ultimately fed by rain: rain that flowed along streams to be collected in rivers and lakes; or which soaked through the earth to become groundwater.

And rain falls from the sky, so at some point in its journey, the water was hanging around in the atmosphere; typically for around a week. In turn, most rain comes from water evaporated from surface water, and most surface water comes from the oceans.

So tap water almost always comes from surface water, which was seawater before that. A year before it comes out the tap, in most parts of the world, my guess is that most of that water was part of an ocean. Which means: a couple years before they are born, most of the molecules that make up a person are part of an ocean — as are most the molecules which will compose people born a year or two from now.

Where in the ocean? In a process called ‘upwelling’, coastal winds running over the surface of the water displace shallow water, which is replaced by deeper waters. As a result, the world’s oceans are slowly and constantly churning, drawing up (sidenote: When ocean-dwelling life (like phytoplankton) dies, it falls a long way to the ocean floor, bringing its nutrients with it. But if nutrients are always sinking, how are they replenished? Upwelling is the answer!).

But upwelling is a small part of a bigger process. Before water emerges to the surface of the ocean, it was part of a slow, planet-spanning (sidenote: Thermo- as in temperature, and -haline as in salt.) — the ‘global conveyor belt’. Surface currents are driven by the wind, but these deep ocean currents are driven only by fractional temperature and density gradients, and so they inch along, sometimes at split-centimetres per second.

The ‘global conveyor belt’ isn’t a perfect loop — there are a few junctions and overlaps, with big loops reaching up from a kind of central donut around Antarctica to form a loop around the Pacific and another in the Atlantic (the ‘Atlantic meridional overturning circulation’). But broadly speaking, one full circuit takes around 500 years to complete. 500 years for the Earth’s waters to ‘turn over’: from shallow water to the depths and back, drawing up nutrients as it returns.

Consider what happens to wastewater in the east coast of the US. It might pass from treatment plants to nearby rivers or bays, and then into the ocean. The water will mix with surface currents and join the Gulf Stream to flow until the North Atlantic. There, some of the water will freeze, leaving salt behind, and becoming denser as a result. That water will sink to the ocean floor to become North Atlantic Deep Water, collecting in a kind of great underwater basin in the evocatively named ‘midnight’ and ‘abyssal’ zones, many kilometres deep. From there, deep southwards currents move the water into the southern hemisphere, rising again around Antarctica to join the Antarctic Circumpolar Current, joining more than 30 million cubic metres per second. The water joins the South Atlantic Current, now flowing north along the Atlantic seaboard — to pass along for another loop, or to evaporate and rain, and be collected as drinking water, centuries after the whole thing began.

So here is a wild guess: if you live in the UK or eastern US, then in the year 1800, most of the water in your body (and hence most the molecules in your body) was in the deep ocean, likely the Atlantic, though perhaps somewhere nearer to Antarctica.

Importantly, most of all the water on the planet is deep ocean water, meaning water (sidenote: I believe there are estimates of water volume by layer, but a way to see this is to notice that almost all water in the world is in oceans, and then to notice that the average depth of the world’s oceans is deeper than 2000m.) — the so-called ‘midnight’ (bathyal), ‘abyssal’, (sidenote: The etymology is what you’d guess. Oceanography goes hard.) zones. No sunlight penetrates these depths. Therefore: I think it’s the case that most of the molecules that will make up future biological humans, for the next few millennia, currently reside in these zones: deep in the ocean.

Exceptions can be found in areas fed by groundwater, where the water that will make up future people is more likely to be presently underground, since water can spend centuries in aquifers before being pumped. In the US, that includes Mississippi, Kansas, and California (sidenote: https://pubs.er.usgs.gov/publication/cir1441).

So compared to the US average, more of the water that will make up future people born in California a few decades from now is underground, while in the northeastern US much of it will come from the Great Lakes, and in the northwestern US it’s currently mostly in the Pacific (before it evaporates). There is something fitting to the sometimes mythic place of great bodies of water in the American self-image that Chicagoans are disproportionately drawn up from Lake Michigan; Washingtonians from the Potomac; New Orleanians from the Mississippi.


That’s the water. Carbon is the next most significant element by weight. The carbon in our bodies comes from food, either by plants directly, or by eating animals (…which ate animals…) which (sidenote: With a small number of exceptions, like the ‘leaf slug’ which can indirectly photosynthesising by absorbing chloroplasts from the algae they eat.). Plants take carbon from the air, where it’s initially bound to oxygen as carbon dioxide, and ‘fix’ it into useful organic compounds. Thus effectively all our carbon atoms are drawn down from the air, via plants which break its bond with oxygen.

Here’s a question: as a great oak tree grows, the material for the tree came from somewhere. And just like how animals eat with their mouths, trees draw up nutrients from the soil. So why doesn’t a 10 ton oak tree form a big crater in the soil where its solid biomass came from? Where did all that tree come from?

This question moved the chemist Jan Baptist van Helmont to experiment. He grew a willow tree in a container, taking measurements of the soil in the container as it grew, as well as the water he was adding to it. After five years the tree was the weight of a man, but the soil had lost less than 60 grams. Van Helmont deduced that almost all the mass of the tree therefore came from the water he was adding.

But this was incorrect. A tree is about 50% carbon atoms by mass, and mostly carbon by dry mass. So what gives?

Richard Feynman nicely captures the answer: “People look at trees and they think it comes out of the ground […] but you ask, where does the substance come from? Trees come out of the air.” Trees come out of the air!

And if plants grow from the air, and you get all your fixed carbon from plants, then so do you — aside from water, most of your remaining weight is carbon.

But what about the autobiography of a typical carbon atom from before it ended up in a plant?

Somewhere in the sky, of course. Like how seawater shuffles around the global conveyor belt, similarly the world’s atmosphere circulates around the planet, in a system we call ‘weather’. And like how ocean currents turn shallow water over to deep water, planet-scale ‘cells’ turn over air in the upper-atmosphere back down to Earth, swooping along the surface as trade winds, and lifting back upwards again. The ‘Hadley cell’ lifts air at the equator, moves poleward some 15km up in the upper atmosphere, sinks at around 30 degrees of latitude, and moves back toward the equator. The Ferrel cell circulates air around the 30 to 60 degrees of latitude range, and the polar cells cover the poles. An image might help

Most sources I could find place the residence time for carbon dioxide in the atmosphere at roughly 5 years, meaning that’s the median time a carbon atom stays (sidenote: So what’s up with claims that CO2 added to the atmosphere by human emissions can stick around for thousands of years? Take this article: “Once it’s added to the atmosphere, it hangs around, for a long time: between 300 to 1,000 years.” I think the answer is that the change in atmospheric CO2 concentration does remain higher for a very long time. Any particular CO2 molecule will get re-absorbed, mostly by the oceans, in a shorter time — but the oceans will release more CO2 also. This article is relevant.). So we’ve traced carbon back through animals to plants to overturning currents in the atmosphere, where a carbon atom will stay for roughly (sidenote: Frustratingly I couldn’t find any estimate for the time it takes for air to complete a full circulation of the Hadley or other cells. Maybe that’s not a meaningful question?). But where was that CO2 molecule before it was in the sky?

This is a question about the carbon cycle. Like how water follows a great loop from oceans to skies to rain and back; carbon atoms move around phases in the air, soil, stone, plants, soils, and the oceans. There is no single answer to where our carbon atoms come from, because more than one thing emits CO2. So it’s a question of proportions, and it turns out to be a mix of three main sources.

First: some carbon gets released from burning fossil fuels, starting sometime in the last couple centuries, where it had previously been buried underground for tens or hundreds of millions of years as part of a long-dead organism. Mostly algae, bacteria, and plants; though for some reason I had always imagined fossil fuels as the remains of literal dinosaurs. Coal is still the biggest emitter, followed by oil and gas. A big fraction of the organisms which became fuel lived and died during the (sidenote: Why did this period produce so much coal? A leading explanation points to the appearance of trees with bark; specifically bearing the wood fibre lignin and the waxy substance suberin which acts like a sealant. This plausibly caused a period where decomposing bacteria and fungi had not yet evolved the enzymes which could effectively digest these new, tougher, trees. Combined with a period of low sea levels and hence a flowering of lowland swamps and forests, huge numbers of trees with bark grew, died, and had enough time to fossilise before decomposing.), just over 300 million years ago.

The Carboniferous period

There is a neat deep-time circularity to this image: that dead creatures are buried for aeons, then reanimated in a sense: pumped and processed and burned and absorbed again into living things. A kind of long-delayed comeuppance. In this vein Randall Munroe points out how plastic dinosaur toys (made of plastic refined from oil) contain some amount of actual dinosaur.

But proportionally speaking, fossil fuels are a sideshow. Very roughly an order of magnitude more CO2 is emitted from organic decomposition and microbial respiration. This is the ‘fast carbon cycle’, where carbon is fixed by plants, becomes part of an organism, and then the organism dies. By the sheer rate of CO2 fixed and emitted back, phytoplankton (photosynthesising organisms in the ocean) are the big players, followed by plants on land.

But organic respiration and decomposition aren’t the only drivers of the oceanic carbon cycle. Most carbon in the ocean isn’t part of any organism. Mostly it’s stored in bicarbonate (HCO3HCO_3-) and carbonate (CO32CO_3^{2-}) ions made from CO2 reacting with seawater. In fact carbon in the oceans outweighs atmospheric atmosphere by a factor of 50 or so, and just over 90% of all the carbon in the carbon cycle is (sidenote: Most of the Earth’s carbon is in the Earth itself, but that carbon isn’t being released for a while, so it’s not treated as part of the carbon cycle.), mostly in the deep and intermediate water.

To throw around some numbers: about 220 petagrams (gigatonnes) of carbon dioxide will be released into the atmosphere this year; around 3% from burning fossil fuels; 44% from the ocean surface; and 53% from the land. There’s about 760 petagrams of carbon in the atmosphere at any one time. And I think we can reasonably infer that these percentages describe the proportional origins of carbon for a plant that you eat. So in tracing the origin of the carbon in our bodies, we’ve taken one more step back: from food to plants to the atmosphere to the wider carbon cycle.

Then there’s the slow carbon cycle, when carbon gets trapped in rocks. In the final chapter of The Periodic Table, Primo Levi imagines the life of a single carbon atom that came to be part of his body:

Our character lies for hundreds of millions of years, bound to three atoms of oxygen and one of calcium, in the form of limestone […] congealed in an eternal present, barely scratched by the moderate quivers of thermal agitation.

Until our atomic protagonist is uncovered by a human tool:

It lies within reach of man and his pickax (all honor to the pickax and its modern equivalents; they are still the most important intermediaries in the millennial dialogue between the elements and man) [and] a blow of the pickax detached it and sent it on its way to the lime kiln, plunging it into the world of things that change.

In a lime kiln furnace the carbon is separated from calcium, is lifted by the wind, dissolved three times in the oceans, and evaporates again; now part of the fast carbon cycle. From there it is absorbed by a grape vine and metabolised into glucose, “a beautiful ring-shaped structure, an almost regular hexagon”. To be (sidenote: I nearly wrote “finally”; of course the point is there is nothing final happening here.) fermented and bottled and drunk and exhaled.

There you have it for carbon: great overlapping fast and slow cycles, passing between the living and the inanimate.

Consider a bowl of rice. Rice takes rice something like 3–6 months from being planted to being harvested, and then let’s give another month between harvesting and eating. So very roughly 4 months from carbon being fixed by the plant, to its grain being eaten by a person. Before that point the carbon was in the air, as carbon dioxide, for a median period of about 5 years. Each atom’s story diverges beyond that point: about half will have spent roughly decades in plants and soils; another 45% or so will have emerged from decades to centuries in the (deep) ocean; and 3% or so will have last seen sunlight hundreds of millions of years ago, perhaps in the bark of a Carboniferous tree. That’s roughly the story of the carbon in your body.


The third most significant element by weight in humans is nitrogen.

Nitrogen in its inert form (N2) makes up most of the air we breathe, but none of this is biologically useful for humans. As with carbon, all the nitrogen which makes up our bodies, like in amino acids, needs to be fixed.

Unlike with carbon, (sidenote: You may read that some peas and beans are able to fix nitrogen, but this is misinformation from Big Legume. Rather, those plants host nitrogen-fixing ‘rhizobia’ bacteria in their roots.). To become plant food, something else must turn atmospheric nitrogen into a nitrogen-containing compound like ammonia (NH3). Some bacteria and archaea — occasionally even lightning strikes — naturally fix nitrogen. But after harvest, typically more ammonia is removed in the harvested crops than those natural processes can possibly replenish in time. For people depending at all significantly on agriculture, you’ll need an alternative: some way of making artificial fertiliser in large quantities. And this warrants an historical interlude.

It happens that ammonia can be found in guano (bird doings). This appears to have been long known to indigenous Peruvians, but was only introduced to Europe in the early 1800s, by the explorer Alexander von Humboldt. The timing was good: newly industrialising nations were increasingly struggling to find enough fertiliser to keep pace with population growth around this time, and guano was the answer.

Demand for industrial quantities of the stuff triggered “guano mania”: the US was importing nearly a millions tons per year by the 1850s, and it passed the ‘Guano Islands Act’ in 1856, enabling US citizens to take possession of unclaimed islands rich in guano, and the military to enforce it. The US began to annex nearly 100 islands in the Pacific and Caribbean to set up mines. There was conflict, terrible working conditions, imperialism, and new guano-centric political alliances. By the late 19th century, the Western world was powered by bird shit.

But it couldn’t go on. By the early 20th century, the supply of natural fertilisers wasn’t growing fast enough to keep up with the population that relied on it. Either this was to become a bottleneck to population growth (euphemism for ‘food shortage’), or somebody needed to find a way to synthesise fertiliser on an industrial scale.

Around this time, the German chemist Fritz Haber was tinkering with catalysts and reactant gases at the University of Karlsruhe. A method was known to decompose ammonia in the presence of a nickel-based catalyst. But from this knowledge, Le Châtelier’s principle suggested a method existed to somehow reverse the reaction in order to produce ammonia at high temperature and pressure. Haber discovered an ingenious and efficient way to manufacture ammonia from hydrogen and nitrogen. Karl Bosch, an industrialist at a German chemical company, soon scaled up Haber’s tabletop invention into an industrial process: the Haber-Bosch process, or just ‘Haber process’.

Then, in case you were becoming attached to Haber as a moral example, he turned his chemist’s talents to war. Nitrogen compounds being components of explosives, he found the Haber process could be put to use in supporting the German front during WWI. Later, and more significantly, he developed some of the first synthetic chemical weapons to be used in war. Gas mask strapped to his face, Haber personally oversaw their release in Ypres in 1915, killing more than 5,000 allied troops. Whatever the impact of his work on producing fertiliser, he is also known as the (sidenote: Being a Jew, Haber was forced to resign from his positions after the Nazis took power in 1933. But a decade earlier chemists working at his institute had developed a fumigant insecticide called Zyklon A, and Haber was powerless to stop the Nazis adapting this chemical for their own ends.).

What about that impact?

A reason this planet can sustain as many people as it does is that we have learned to grow the same amount of food with far (sidenote: Since 1961, it now takes 40% as much land to grow the same amount of cereal. Over that period, cereal production grew by 250%, and population grew by about 160%, meaning we’re growing more cereal per person (though note that more of this cereal is also now being fed to animals, so actual caloric intake per person has increased by less). This also means that the total land used for growing cereals. Why focus on cereals? They deliver about half the world’s calories, dominate arable land use, and are otherwise representative of other crops.). “With average crop yields remaining at the 1900 level”, writes Vaclav Smil, “the crop harvest in the year 2000 would have required nearly four times more land and the cultivated area would have claimed nearly half of all ice-free continents, rather than under 15% of the total land area that is required today”. Compared to the land required to feed the same world population without them, improvements in agricultural productivity have spared the use of about 1.5 billion hectares of land from being farmed — about the size of India and the contiguous United States combined. And those improvements owe significantly to (sidenote: The source of the significant majority of the world’s fertiliser, and about half (maybe most) of the nitrogen input to global agriculture.).

Of course, more likely than the planet somehow finding an extra India+US amount of arable land to farm, a world without fertilisers is just a world with fewer people. On some estimates, perhaps two billion fewer people would be alive today without the Haber process, or at least some way of producing synthetic fertiliser on an industrial scale. Fritz Haber, meanwhile, died as both a war criminal, and the inventor of one of the most significant life-sustaining inventions of the 20th century.

All this is pretext to say: to a significant extent, people are made of the products of the Haber process.

People come from plants, plants use ammonia and other nitrogenous compounds in fertilisers to grow, and those compounds come from the Haber process. Specifically, since it produces about half the nitrogen atoms in agriculture, and most people ultimately get effectively all their nitrogen atoms from agriculture, about half of the nitrogen in your body (sidenote: How can we know? This article says that “[c]hemical fertilizers contribute about half of the nitrogen input into global agriculture”, and Smil adds: “Human activities have roughly doubled the amount of reactive N that enters the element’s biospheric cycle”.
Also check out this diagram from the Smil paper:
Smil diagram

The Haber process takes inputs from the air (atmospheric nitrogen), plus hydrogen made via steam reforming from water and hydrocarbons typically from natural gas. Then it might be shipped off and stored in liquid form in a pressurised tank, before being sprayed. Give that process 100 days, and give a typical cereal crop 200 days to grow, plus a couple months between being harvested to being eaten, and we can guess that very roughly a year before becoming part of you, about half your nitrogen atoms were being stripped from the air in some great metallic reactor like the one pictured.

A Haber process reactor, I think

The human cycle

But humans don’t just hoard atoms until we die. We release and replace them, too.

You might have heard that your body’s cells replace themselves every 7 years — this isn’t quite right, since some cells (in our brain, heart, and eyes) stick around for life, while others survive only for a (sidenote: This website is a great resource.). In any case, we’re interested in the atoms themselves, not the (sidenote: A cell could die and a new cell could be made of its material; or a cell could survive but its atomic parts be replaced.).

But that ‘7 years’ myth comes from some great science. In the 50’s and 60’s — before the 1963 Limited Test Ban Treaty — nuclear weapons were tested above ground. This caused a spike in atmospheric carbon-14 (a naturally occurring radioactive isotope of carbon), through free neutrons from nuclear fission mimicking cosmic rays by converting nitrogen-14 into carbon-14. Here’s the trick: when a cell divides, it makes copies of its DNA. Unlike other parts of a cell, DNA is molecularly stable: the molecules that make it up stay the same as long as it’s around. Knowing this, plus the atmospheric levels of 14C^{14}C at different times, you can figure out when a cell was created based on the level of 14C^{14}C in its DNA. This is exactly what some researchers did, and found the average age of cells in human to centre around 7–10 years.

Life expectancy of cells

Again human activity left its mark on the very physical stuff we are made out of: weapons so powerful they changed the composition of the entire atmosphere and left a trace in our DNA.

But the atoms — the actual physical stuff — are replaced faster than the decadal scale of our cells. For instance, the water in cells is constantly being absorbed and replaced through their membranes, almost as fast as we drink and expel water. Cells are constantly metabolising molecules and expelling waste products; proteins within cells are built up from food and broken down to be excreted on the scale of weeks. So it’s fair to say that our atoms have a faster turnover rate and shorter average residence time than our cells, the same way countries tend to last longer than their citizens.

Collecting a few estimates, my guess is that a given atom in your body has been there for an average of a few months, with a wide distribution between a couple days (water) to a lifetime (stable molecules in e.g. DNA from infancy). And so I’d guess most of our weight in atoms is recycled at least every few weeks.

In any case: to me, the picture this paints is one in which we are not so much solid, isolated objects; and more like patterns surviving (sidenote: Consider tattoos. Skin cells last about a month from being born to being shed, but tattoos last for decades. The explanation is that cells absorb the ink until they die, then release it to be absorbed by newer cells, “throwing the ink in a perpetually revolving prison”. So too with the rest of us — cells swapping in and out of something like a revolving human-shaped prison (is the kind of thing I’d write in my Camus phase).).

As we eject atoms, they enter back into the flow from which other people absorb them. This raises a final question: how many atoms do we share in common with past people? How much of us will end up composing future people? I believe the following claim is true: for effectively everyone who has ever lived, and even most living adults, your body contains at least one atom that has been part of them, such as air they breathed or water they drank. The key here is just that the basic molecules that make up most of our mass are very small (citation needed). Take air as an example: the “fraction of the atmosphere that contains water vapour breathed out by a given person” is a small number, but it is larger than the number of all the water molecules in our body (on the order of 102710^{27}), meaning we likely contain many, many molecules breathed out by that person.

Very likely even water that was part of your body as a kid got entangled with irrigation, reservoirs, and water filtration systems years ago, and recently got reunited with you.

Levi again:

The number of atoms is so great that one could always be found whose story coincides with any capriciously invented story. I could recount an endless number of stories about carbon atoms that become colors or perfumes in flowers; of others which, from tiny algae to small crustaceans to fish, gradually return as carbon dioxide to the waters of the sea, in a perpetual, frightening round-dance of life and death, in which every devourer is immediately devoured[.]

The physicist Max Tegmark suggests a picture of life as a ‘braid’ in space through time.

Spacetime braid

If that’s right, then the threads of the braid that makes up a person run through skies, seas, other life, and through the machines of industry. A person’s braid isn’t isolated through their life, but ephemeral; swapping threads with the wider world until the point of death, where the threads that now make up the braid — few of the originals remain — disperse again.

If you want me again look for me under your boot-soles.

Appendix on sharing breath with historical figures

In this example, a simple model here would be to treat the Earth’s atmosphere and oceans as a connected control volume which mixes a person’s breath negligibly quickly.

We then need to consider three volumes:

We’re assuming that the water in your body is ≈ entirely coming from water in the oceans and atmosphere (realistic!), and that the water vapour breathed out by a given person is perfectly mixed into this volume before you drank it / breathed it in (I think realistic for long times ago, less realistic for more recent people):

I’m sufficiently surprised by this that I put a decent weight on having made a rookie error — maybe someone can check my working!

For a less idealised model, especially when dealing with people close together in time, you’ll need to look to a more sophisticated model of fluid parcel residence times. I’ll leave that as an exercise for the reader.

Back to writing