The soil test came back adequate. The custom blend went on. The crop still ran out of water a week early, and next spring the invoice looks the same. That is not a nutrient deficiency. That is a missing underground network — mycorrhizal fungi that used to forage what the bag now replaces.
You already pay for the partner you don't have. Phosphorus in a tank. A second irrigation set. Forage that browns sooner than the neighbor's unplowed pasture. The line item never says fungi. The missing trade still shows up as cost.
what mycorrhizal fungi actually are
Mycorrhizal fungi are the living threads on a crop's or grass's roots that forage phosphorus, nitrogen, and water from soil the root itself will never reach — and get paid in sugars the plant made that morning. The word is Greek for fungus-root. The threads are hyphae, ten to fifty times thinner than a hair. They can multiply a root's foraging area by tens to a hundred, depending on the soil and the fungus.
They are not a nutrient. They are not "organic matter" on a lab sheet. They are not the mushroom you see after rain — most of the fungi that live with food crops never fruit anything you can pick. They are a partner. When the partner thins, you buy what the partner used to fetch.
Most land plants run this trade, including most food crops. The plant sends carbon down. The fungus sends minerals and water back. The network exists whether or not anyone maps it, inoculates it, or writes a conservation plan. What underground networks actually are is the definition pillar. This post is the trap: treating a thin network as if it were an NPK problem.
Mycorrhizal fungi are a living trade on the root, not a deficiency on a soil test. Lose the trade and you start renting the work.
Plenty of growers, graziers, and Indigenous land relations already treated the ground as alive. The maps are catching up to that fact. They do not invent the partner.
the soil test came back fine
A soil test is a good tool for what it measures. Extractable nitrogen, phosphorus, potassium, pH, sometimes organic matter. Chemistry in the sample the lab received. It will not tell you whether hyphae are still running through that field. It will not tell you whether last year's partner is still paid in sugars, or has been laid off because phosphorus arrived cheaper in a tank.
That is the contrast. The test can read "adequate" while the underground is thin. The agronomy that follows the test — buy the difference — is not stupid. For fifty years it was the training. It is incomplete. The grower who inherited a disked field did not invent that training, and is not the villain of this story.
What you lose, once the partner is gone, is not a romance about soil. You lose reach: hyphae fetch water and phosphorus from pores a root hair never enters. You lose glue: those threads help hold aggregates that take rain instead of shedding it. You lose a drought week. Loss shows up as a bill you already know — diesel, fertilizer, pumping.
Food companies and public programs pay the same bill from the other side of the fence: more nitrogen, an earlier destock, a disaster payment after a dry year the ground used to buffer. The missing network is already on the ledger. It is booked as inputs and recovery, not as a living partner.
If the land also stops holding water, that is a hydrology story as well as a fungal one — drought-proofing starts in the soil, not the reservoir. Different job: there the sponge. Here the partner that helps the sponge stay built.
four objects people treat as one
Growers, buyers, and agencies use these four as if they were the same underground. They are not, and only the last one is alive.
| what it is | what you hold | what it proves | what it does not do | |
|---|---|---|---|---|
| soil test | A lab report on extractable nutrients, pH, and sometimes organic matter | A sheet of numbers for this season's field | Chemistry in the sample the lab received | Measure living hyphae, fungal richness, or whether the plant still has a partner |
| fertilizer | A purchased nutrient, especially phosphorus and nitrogen | A bag, a tank, a custom blend | That you can replace what the lab said was short | Rebuild a network. Heavy phosphorus often makes the plant stop paying the fungi |
| inoculant | Selected mycorrhizal spores in a bag or liquid | A labeled product | That spores were applied | Restore a landscape-scale living network. Colonization is not the same as a dense, functioning trade |
| living network | Hyphae trading plant carbon for nutrients and water with the crop or the grass | Nothing you can serial-number. It lives in the soil you manage | That the trade is still running on these acres this season | Show up on an NPK test, or survive conversion, a season without living roots, or a phosphorus flood unchanged |
Read the last row first. Everything above it can be useful. None of it is the last row.
Inoculant companies are not the enemy. A spore on a transplant or a field that lost its partner can be a real start. A bag is still not a network. Fertilizer is not a sin either — crops need phosphorus. The contrast is narrower: when the plant can get P from a tank, it often stops sending sugar to the fungus. You did not "kill the soil." You rented a job the partner used to do. Rent compounds.
cropland networks run about half as dense
This is now a map, not a hunch.
Stewart et al. (2026, Science), working with the Society for the Protection of Underground Networks (SPUN), published the first global picture of arbuscular mycorrhizal hyphal density and biomass — about 1.10×10¹⁷ km of living hyphae in the top 15 cm, weighing about 300 ± 60 Mt. Plain-language door: A Hidden Infrastructure.
On that map, large-scale cropland is predicted to run at about half the hyphal density of wild systems. Read the verb. Predicted. Associated. The paper does not indict a named practice on a named farm. It does not say your no-till corn is the same as your neighbor's disked beans. It says that as a land-cover class, cropland comes out thinner. SPUN's own write-up is explicit: the research does not link particular farming practices to mycorrhizal health. A 1 km pixel is about 100 hectares — a regional screen, not a soil probe, not a verdict on your field.
Plot work supplies mechanism, not guilt. Tillage severs hyphae. A season without living roots starves the partner. Fungicides can hit the fungus directly. High phosphorus can make the plant stop paying. None of that makes a grower a villain. It describes how a trade thins when the conditions of the trade go away.
The biomass that remains is not evenly spread. Grasslands hold about 40% of global arbuscular mycorrhizal biomass and are poorly protected. That living-cover argument lives in what grasslands actually are. The point here is simpler: a lot of the underground still sits under grass, and converting that grass to cropland is one way a dense network becomes a thin one.
Van Nuland et al. (2025, Nature) found only about 9.5% of mycorrhizal richness hotspots inside protected areas. That is the conservation gap. This post is the working-land trap: the network can thin in plain sight, on land nobody would call wilderness, while the soil test still passes.
SPUN maps where the infrastructure is. It is a science and mapping partner, not a buyer. Timber, farms, food companies, and public programs are the ones already paying when the partner is gone.
the practice stack is not the network
NRCS and its cousins already speak a language that feeds the trade: minimize disturbance, keep the soil covered, keep living roots in the ground, mix the plants. Cover crops, reduced tillage, and grazing that stays grazing can give the fungus someone to trade with. Use those programs. They are cousins, not competitors.
A soil-health program is still a practice stack. It pays for residue, cover, and rotations that make a living network more likely. It does not, by itself, measure hyphal density. You can enroll the acres and still be holding a receipt. The partner is a different object.
The same split lives in another guide. Dirt is a health system is about soil as a One Health / body system. Different job. This post is about the named fungal network under the land. If you came for that health-system frame, that post is the door. Stay here if the question is the missing partner on the root.
A carbon credit is a third object again — a receipt for an estimated tonne, not the hyphae. That honesty has its own post: a carbon receipt is not a living fungal network.
frequently asked questions
what are mycorrhizal fungi?
Mycorrhizal fungi are living fungi that partner with plant roots, trading soil phosphorus, nitrogen, and water for the plant's carbon. For a grower: they are the underground foragers on the root, paid in sugars, not a nutrient you can read on a soil test.
Most food crops run arbuscular mycorrhizal fungi; many timber trees run a different trade. Type differs; the job is still a living connection — arbuscular and ectomycorrhizal are not the same underground.
why do cropland fungal networks thin?
Because the conditions of the trade get scarce: fewer continuous living roots, more physical disruption, cheaper phosphorus in a bag, sometimes fungicides. Stewart et al. predict large-scale cropland at about half the hyphal density of wild systems — a modeled land-cover association, not a ruling on any one practice or farm.
The grower who inherited that system is living inside it, not the author of it. Thinning is a receipt. It is what the missing partner looks like from above.
is a soil-health program the same as a living network?
No. A soil-health program is a set of practices and payments that can feed a living network — cover, living roots, less disturbance, more plant diversity. The living network is the hyphae themselves. You can enroll the acres and still not be holding the partner. Measure the trade, or you are holding the paperwork.
fund the living underground, not another input
The thin network is the receipt. Another input is another invoice. The work that lasts is keeping a living root in the ground long enough for the partner to stay employed — on a named place, not as a metaphor.
Natural assets are the foundation: the fungi, the soil they inhabit, the crop or grass that pays them. ensurance is how a beneficiary funds the living condition of that land. It is not the network.
After the underground is named, the instruments are ordinary. An agent stands for a place, a steward, or a purpose. A certificate is a hold whose proceeds fund that agent's present condition. A coin funds the category rather than the parcel. Live doors today, at small volume: subterranean.ensurance for the underground stock, healthy-soils.ensurance for the flow the network helps produce. Neither is land title. Neither replaces a good agronomist.
If you manage land: keep the living root, treat the soil test as chemistry, and do not confuse a bag of spores with a network. If you buy food, fiber, or risk: you already pay for the failure. Funding the land the network lives in is how that bill changes object.
what underground networks actually are →