Most of the naturally derived antibiotics in clinical use were made by Actinobacteria.
Not discovered in a lab first. Made by organisms that had been manufacturing those molecules for millions of years before modern antibiotic screening. Actinobacteria produce about two-thirds of the naturally derived antibiotics in current clinical use (settled). Many of those compounds come from soil-dwelling Streptomyces. Streptomycin, the first drug that worked against tuberculosis, came out of a soil organism in 1943.
The point is not that nature is nice. The point is that a working part of medicine has a physical address, and for many of those drugs the address is underground.
one health is one system, not four fields
One Health is the position that human health, animal health, plant health, and environmental health are one interdependent system rather than four separate disciplines. The idea has roots in One Medicine and comparative medicine. It later became a way to track diseases that jump between animals and people, and has since widened to include plants and environments.
In 2023, Samiran Banerjee and Marcel van der Heijden made the sharper claim in Nature Reviews Microbiology: soil is a cornerstone of One Health. They list more than 40 soil microbiome functions that contribute directly or indirectly to soil, plant, animal, and human health. They also identify microorganisms shared between One Health compartments, and conclude that soil, plant, and human microbiomes are "perhaps more interconnected than previously thought." Soil is also a reservoir of pathogens and antimicrobial-resistance genes. One Health includes that half. It does not make soil a pharmacy with a warning label taped on.
Karen-Beth Scholthof made the companion argument in 2024: plants and plant microbes belong inside One Health, not next to it.
And soil is where the diversity actually is. A 2023 review in PNAS estimates soil is home to 59% (±15%) of all species on Earth — the most biodiverse single habitat we know of. That is an estimate with wide error bars, not a census (supported).
what soil life makes, and how sure we are
Every mechanism below carries a confidence label.
| what soil life makes | mechanism | confidence |
|---|---|---|
| most naturally derived antibiotics in clinical use | secondary metabolites from Actinobacteria, many of them soil Streptomyces | settled |
| plant nutrition and yield stability | mycorrhizal symbiosis, nitrogen cycling, suppression of soil-borne pathogens | settled |
| soil structure, water holding, carbon storage | microbial aggregation and organic matter dynamics | settled |
| immune exposure and training | contact with diverse environmental microbes diversifies commensal microbiota and shifts immunoregulatory markers | supported |
| a named health outcome from restoring a named parcel | — | not established |
That last row is the claim this post will not make.
the one body claim, and then the land
Some of the same microbial genera turn up in soil, in plant roots, and in human guts. Banerjee and van der Heijden identify organisms shared between those compartments (supported). Shared taxa are not the same as a personal transfer.
Here is the one clinical foothold, and then the sentence ends on land.
In 2020, researchers in Finland added forest floor, sod, and plants to urban daycare yards. Seventy-five children, ages three to five, were studied across greened yards (36), standard yards (16), and nature-oriented daycares (23). After 28 days, children in the greened yards had a more diverse set of certain skin bacteria, and a blood marker tied to immune regulation moved in a healthier direction. The change was a whole yard — soil, plants, and play — not soil alone.
That is one small trial, 28 days, immune markers rather than diagnoses. It is not a cure, and the authors did not claim one.
The direction runs one way in this post, on purpose: body, then soil, then place. Never body, then product.
the maker can stop, and biology sets the clock
Soil life is not ambient. It is a standing population doing work, and it can be reduced.
The practices that reduce it are specific and well documented: repeated tillage, fumigation and broad-spectrum sterilization, drainage of wet ground, compaction, and long runs of a single crop. Each is rational at the field scale under current prices. Together they simplify the community that produces those 40-plus functions.
Recovery does not run on a fiscal calendar.
- Microbial composition can shift within a season after a practice change (supported)
- Soil structure and organic matter often take years to decades (supported)
- New mineral topsoil can take centuries (supported)
Those clocks are different processes. They do not move together, and they are place-dependent.
So protection and restoration are different trades. Protecting soil that still works is usually cheaper and faster than rebuilding soil that has been simplified. That asymmetry is the argument for acting while the maker is still running.
the money is pointed at the wrong half
The usual number here is the aggregate one — far more capital flows against living systems than toward them. True, and not specific enough.
The specific misdirection is this. FAO, UNDP, and UNEP, averaging 2013–2018, put global support to agricultural producers at about $540 billion a year, and found 87% of it — roughly $470 billion — price-distorting or harmful to nature and health. A large share of that support is price incentives and subsidies tied to a particular commodity or input.
Tied support does not write a line-item for killing soil biology. It pays for the practices that often simplify it, and it does not price the loss. The pathway from a subsidy to a hospital bill is not automatic — resistance, runoff, and yield volatility have other causes too — but the money is still pointed at volume first.
The payor is whoever already carries the downstream cost of simplified land: a utility filtering what bare tilled ground releases, a food company buying yield stability it does not grow, a government paying for both. A health system has a reason to care that soil remains a source of antimicrobial chemistry and not only of resistance genes. That is a research and stewardship case, not a claim that protecting an acre treats infection.
protect and invest in a maker's condition, on a place
ensurance exists to point money at the condition of living systems on named ground, before loss, instead of paying out after it.
The unit is a place: a named soil, a watershed, a habitat. RealValue, our accounting engine, makes selected stocks, flows, and condition on a parcel legible to capital — Healthy Soils, Medicinal & Genetic, and Pest & Disease Control among the flows. It does not sequence a microbiome. Field evidence is still required for any claim about what that parcel's soil life is doing.
That number is a bridge, not a verdict on worth. It makes a living process legible to a treasury. It does not say the price is the value. Instruments here are based on the condition of living systems, never backed by them.
There is no ensurance product for your body, your gut, or your microbiome. There are no microbiome tokens. No one here will tell you which acre to buy for your health. The object of protection is land, water, and habitat.
the filter: not every maker is load-bearing
All living systems make conditions. That does not make all of them something a treasury or an insurer must treat as load-bearing. The filter is narrow, and it should be:
- non-substitutable — no engineered alternative at the scale required
- irreversible — loss cannot be undone on a human timescale
- cascading — failure propagates into other systems
Working soil biology can meet this test on particular land. That is a place-level determination, not a slogan, and it is not a determination to make from outside the Indigenous nations, local communities, and rights-holders of the place.
Some dimensions of a soil's value are infrastructural. The category stays selective; it does not claim the whole. Criticality is a functional test, not a ranking of worth.
what this post claims, and what it does not
Claims: soil, plant, animal, and human microbial life share organisms and functions across compartments (supported to settled, depending on the function). Most naturally derived antibiotics in clinical use come from Actinobacteria, many of them soil-dwelling (settled). Environmental microbial exposure can shift immune markers (supported). Soil biological function can be lost faster than it rebuilds (supported).
Does not claim: that restoring a parcel treats a named disease. That soil contact substitutes for medicine. That the causal pathway from any given acre to any given body is known.
Two neighboring posts make different claims on purpose. the body knows before the mind does is about sensing a landscape — a different verb and a different literature. nature access and health equity is about parks and who can reach them — an access argument, not a microbial one. For how condition gets measured as an investment variable, read the investment value of ecosystem condition.