The Real Plumbing of a Forest

Pull up a seedling and you get the picture you were taught: a stem, a spread of pale roots, dirt. It looks like a self-contained machine that mines the ground for what it needs. It isn’t. Somewhere in the last millimetre before the soil, the plant hands the job over — to threads far thinner than any root, built by an entirely different kingdom of life. The word for the joint where that handover happens is mycorrhiza, which is Greek for, more or less, fungus-root. And once you know it’s there, the honest diagram of a forest stops being roots in soil and becomes roots plugged into fungus, fungus in soil.

We’ve told the origin of this partnership already — how a rootless green scum crawled out of the water about half a billion years ago with fungi doing its mining, and how the two of them ground the first soil out of bare stone. This is the sequel, and it’s the less romantic, more mechanical question: what does the fungus actually do down there, today, for a living tree? The answer turns out to be measurable, and a little humbling for the tree.

Not a special case — the default

Start with how normal this is. A survey of 3,617 land plant species across 263 families, compiled from 659 papers, found that “80 and 92% of surveyed land plant species and families are mycorrhizal” Established. Going without is the deviation, not the rule.

Two honest asterisks on that, because the number gets quoted more confidently than it deserves. The word doing quiet work is surveyed: the denominator is 3,617 species somebody had already gone looking at, not every land plant there is — and species get looked at partly because someone expects to find something. And “land plants” includes the mosses and liverworts, where the rate is far lower. The generalisation is much safer for vascular plants than for green life in general.

Sort those — the vascular plants — by partnership type and they split roughly like this: about 72% arbuscular mycorrhizal, at least 2% ectomycorrhizal, 1.5% ericoid, 10% orchid — and just 8% completely non-mycorrhizal, plus another 7% inconsistent, sometimes colonised and sometimes not Established. (Rounded figures, not precision instruments: the same paper’s body text gives 71% arbuscular across all vascular plants and 72% across flowering plants, and its own diagnostic warning is that older literature over-reports mycorrhizal status, because a few stray hyphae in a root do not make a mycorrhiza.) Hold onto that 2%. It’s about to do something disproportionate.

Why a root can’t reach its own dinner

To see why plants outsource, follow the phosphorus. Plants take phosphorus up as inorganic phosphate, and phosphate is a miser’s nutrient: it “has low diffusion rates in soil,” and much of what’s there is “sequestered by soil minerals such as Fe and Al hydroxides” or “bound to organic matter” Established.

Slow diffusion has a brutal consequence. A root absorbs the phosphate in the thin film of soil actually touching it — and then nothing much arrives to replace it, because phosphate doesn’t travel. Within days the root is sitting in a hole it dug itself: a depletion zone, a halo of stripped soil that the root has no way to reach past. It is a mouth surrounded by an empty plate.

Someone measured exactly how far that reaches, and how far a fungus gets instead. In a 1991 experiment, white clover was grown in pots divided by a 30-micrometre nylon net — a mesh too fine for roots to cross and wide open to fungal threads. In the plants grown without a fungal partner, phosphorus depletion “extended about 1 cm into the outer compartment.” In the mycorrhizal plants, “a uniform P depletion zone extended up to 11.7 cm (the length of the hyphal compartment) from the root surface” Established.

Read that parenthesis again, because it’s the best line in the paper. The length of the hyphal compartment. The fungus had stripped phosphate all the way to the far wall of the box. The experiment ran out of room before the hyphae ran out of reach.

Roughly three-quarters of the phosphorus, arriving from soil the root never touched. That is not a supplement. That is the supply line.

Two ways to build the joint

The two dominant partnerships solve the same handshake problem with opposite architecture, and both are worth picturing properly.

In an arbuscular mycorrhiza, the fungus goes inside. It burrows through the outer root layer, spreads between the cortical cells, and then pushes into individual cells to build an arbuscule — a tiny repeatedly-forked tree of fungal tissue, described in the literature as the “branched structures where the bidirectional exchange of signal molecules and nutrients occurs between the two partners” Established. It looks like an invasion and functions like a radiator: all that branching is surface area, and surface area is the whole point of a trade counter.

In an ectomycorrhizal root, the fungus stays out of the cells entirely. It wraps the root tip in a dense sheath — the mantle — and then threads a lattice called the Hartig net into “the apoplastic space between root cells” Established. No cell is ever entered. The fungus simply fills every gap in the masonry until the root tip is less a root than a shared organ.

And here’s where that 2% pays off. Only about 2% of vascular plant species are ectomycorrhizal — but in temperate, Mediterranean and boreal ecosystems the main tree families sit inside that sliver: Fagaceae, Betulaceae, Pinaceae and Salicaceae — the oaks and beeches, the birches, the pines, the willows Established. As the review we’re quoting puts it, “the majority of timber, softwood lumber and construction wood traded worldwide derives from this low percentage of vascular plants.” A northern forest — the whole cathedral of it — is built almost entirely out of what looks like a rounding error in the plant kingdom’s statistics. And every root tip in it ends in a fungus.

Why fungus rather than more root? Geometry. The threads run about 2 to 5 micrometres across Established — and small diameter is exactly the trick. It lets the fungus “access small soil cores for P, and achieve greater P influx rates for a given surface area” than a root can Established. A root is a pipe shoved through soil. A hypha is a thread that goes between the grains. For the same investment of carbon you can buy a great deal more contact with the world.

What the tree pays

None of this is charity, and the bill is not small. Estimates of how much of a plant’s fixed carbon goes to its fungal partners range from 1–20% of net primary production for ectomycorrhizal fungi and 1–30% for arbuscular fungi in pot studies, with field estimates of 8–17% in arctic tussock tundra, 27–34% in a temperate mixed forest, and 4–35% across temperate stands Established.

We want to be careful with those numbers rather than pick the loudest one — and so, to its credit, is the review they come from. It says outright that “plant C allocation to mycorrhizae is one of the most uncertain parameters of ecosystem-scale C budgets and simulation models of C cycling,” and that the published estimates are “sometimes difficult to compare because some refer to fungal biomass production only, whereas others include respiratory costs, a distinction that is not always clearly reported.” They’re also shares of net primary production, not of everything the leaves fix. What’s solid is that the payment is real and often large. The precise fraction, for any given forest, is not a settled figure, and anyone quoting you a single tidy percentage is rounding off a genuine argument.

What keeps the deal from collapsing into freeloading is that both sides can apparently haggle. In a 2011 experiment, researchers reported that “plants can detect, discriminate, and reward the best fungal partners with more carbohydrates,” while “their fungal partners enforce cooperation by increasing nutrient transfer only to those roots providing more carbohydrates” — concluding that “unlike many other mutualisms, the symbiont cannot be ‘enslaved’” because “control is bidirectional, and partners offering the best rate of exchange are rewarded” Established. That’s a controlled experimental system, not a forest, and the word market is a metaphor we’re laying on top of it. But the finding itself is a nice antidote to both bad stories: neither party is a selfless nurse, and neither is a parasite with a hostage.

The miners who used to be rotters

Here’s the detail that made us want to write this one, and it hooks straight back into the kingdom that invented rot.

Arbuscular fungi cannot rot things. Genomically they have “a decreased repertoire of genes encoding for plant cell wall–degrading enzymes, with no genes encoding cellobiohydrolases or β-1,4-glucosidase,” most hemicellulose and pectin gene families are missing too, and “no genes involved in lignin degradation, such as class II peroxidases, have been found” Established. Four hundred million years of living inside roots and being handed sugar, and the toolkit for eating the world went into the bin.

The ectomycorrhizal fungi went the other way, and they got there from somewhere unexpected: they “evolved multiple times from a diverse group of saprotrophic ancestors, including brown-rot, white-rot, and other saprotrophic fungi,” and consequently “have retained distinct suites of lignocellulolytic genes” Established. Several of the big, familiar forest genera — Cortinarius, Russula, Lactarius — still carry peroxidase genes, which the same review says is “potentially allowing them to access N sequestered in complex polyphenolic substrates” Preliminary. That hedge is the author’s, not ours: carrying the gene is not the same as demonstrating the job in a real forest floor, and we’re holding it where they held it.

But the framing is what lands, and it comes with a warning label attached in the source itself: those enzymes “are used for nutrient acquisition, not for securing C and energy” Established. So the mushroom you find at the foot of an oak is descended from wood-rotters, and it kept some of the chemistry — but it is not eating the forest floor for a living any more. It’s mining the forest floor for nitrogen and drawing its energy from the tree. Somewhere back in deep time, a lineage of decomposers stopped hunting and took a salary.

And sometimes it’s a bad deal

One more thing before the honesty box, because “symbiosis” gets read as “friendship” and it shouldn’t. The standard framing in the field is a mutualism–parasitism continuum — a theory established by Johnson and colleagues in 1997, “postulating that many mycorrhizal associations may shift from beneficial to detrimental for the plant, with the outcome of the symbiosis being related to plant developmental stage and edaphic or climatic growth conditions” Established. That the outcome varies is not in dispute: in one 2016 experiment, Hieracium pilosella was “consistently mutualistic” while “pronounced parasitism was observed in C. canescens” — same phenomenon, opposite sign, in two grassland plants grown side by side.

But here’s the part we’d have got wrong if we’d only read the framing sentence. That same study set out to test whether shifting the climate and the soil would slide a plant along the continuum — and it didn’t. The authors report that “environmental and edaphic conditions did not markedly affect the cost:benefit ratio of the mycorrhizal symbiosis in both species,” and concluded that such conditions “appeared to be of minor importance” Contested. So: the continuum is real and the sign really can flip, but the tidy just-so version — drought turns your fungus into a parasite — is a hypothesis that one careful test declined to confirm. The plumbing is real. It is not automatically kind, and it is not automatically legible either.