The Mycorrhizal Network

Inter-organism substrate-current bridges, the forest as substrate-coherent coherence cell, and the Wood-Wide-Web with appropriate hedges

A mycorrhiza is a symbiotic association between a plant root and a fungus, in which fungal hyphae penetrate the root (intracellularly in arbuscular mycorrhizas, apoplastically in ectomycorrhizas) and extend outward into the soil at \sim 10^2–10^3 times the volumetric reach of the root system alone, sampling water, phosphorus, nitrogen, and trace minerals at hyphal-diameter (\sim 2–10\;\mu\text{m}) resolution across soil pores too fine for roots (\sim 100\;\mu\text{m} minimum) to enter. The plant supplies the fungus with carbon (sucrose, lipids — the latter established as a major flow only since Keymer et al. 2017 and Luginbuehl et al. 2017); the fungus supplies the plant with mineral nutrients sampled from soil volumes the root system cannot directly reach. Roughly \sim 80\% of vascular-plant species form mycorrhizal associations, with arbuscular mycorrhizas (AM, Glomeromycota) the dominant form across grasses, herbs, and most tropical trees, and ectomycorrhizas (EM, Basidiomycota and Ascomycota) the dominant form across temperate and boreal forest trees (Pinaceae, Fagaceae, Betulaceae, Salicaceae). The association is ancient — AM-fossil evidence in early land plants from the Rhynie chert (\sim 410\;\text{Ma}) and molecular-clock estimates pushing AM-plant symbiosis back to \sim 450\;\text{Ma} (late Ordovician) make mycorrhizal symbiosis as old as land plants themselves, and probably enabling of land-plant colonisation at all.

Beyond the symbiosis between one plant and one fungus, a fungal individual can simultaneously host symbioses with many plants — and a single common mycorrhizal network (CMN) can therefore physically link the root systems of multiple plants, including plants of different species, across a forest patch. Suzanne Simard’s 1997 Douglas-fir / paper-birch CMN paper (in Nature) was the first direct experimental demonstration of inter-plant carbon transfer through a CMN, using ^{14}C and ^{13}C radiolabelling. The architecture has since become known popularly as the “Wood Wide Web”, with claims of ecological significance — mature “mother trees” preferentially transferring carbon to kin seedlings, inter-tree warning signals about herbivory, network-mediated competitive dynamics — that have generated both major popular attention and (since Karst, Jones, and Hoeksema 2023 in Nature Ecology & Evolution) substantial professional pushback on the strength of the published evidence for the larger ecological-significance claims.

The framework’s claim about this organisation is direct and architecturally clean. The mycorrhizal hyphal network is biology’s chemistry-side implementation of an inter-organism substrate-current corridor, with arbuscules and Hartig nets as the substrate-coherent boundary-matching transport zones at the plant-fungus interface, hyphae as the substrate’s smallest inter-organism conduits, common mycelial networks as inter-modon coupling at forest scale, and the forest as a substrate-coherent coherence cell at 10^1–10^4\;\text{m}. The architecture is real, the documented chemistry-side mass-flow through the network is real (water, ^{32}\text{P}-phosphate, ^{15}\text{N}-amino acids, and at least some labelled carbon all move between connected plants in controlled-tracer studies); what remains contested is the quantitative significance of inter-plant flux relative to fitness, and the adaptive interpretation of “the network as a communicating community”. The framework’s position on the contested layer is the same as its Penrose-Hameroff position at the brain rung: take the architecture and the documented flux, and leave the strongest functional claims dependent on whichever way the data falls out as more evidence accumulates. The substrate-coherent inter-organism corridor exists; the framework does not need to commit on whether Simard’s mother-tree-carbon-transfer narrative is the right reading of the corridor’s ecological role to make its architectural claim about what the corridor is.

This chapter develops AM and EM as two substrate-coherent symbiotic architectures, hyphae as the substrate’s smallest inter-organism conduits, the common mycelial network as inter-modon coupling, the forest as substrate-coherent coherence cell, the Wood-Wide-Web question with appropriate hedges, and the closing of the plants-section five-scale stack with forward links to the Gaia chapter.

Arbuscular and Ectomycorrhizal Architectures: Two Inter-Organism Corridors

The two dominant mycorrhizal types are architecturally distinct but substrate-mechanically parallel.

Arbuscular mycorrhizas (AM). Glomeromycota fungi (a single phylum of \sim 300–400 named species but possibly \sim 1500–2000 true species) form intracellular symbioses with \sim 80\% of vascular-plant species. The hypha enters through the root epidermis, grows through the root cortex apoplastically, and at a target cortical cell penetrates the cellulose wall but not the plasma membrane — the PM invaginates around the entering hypha and engulfs it in a substrate-coherent extension called the periarbuscular membrane. Inside the engulfed space, the hypha branches into a small tree-like structure (the arbuscule) with \sim 10^2–10^3 terminal branches, presenting an extraordinary membrane-contact surface for nutrient exchange. The plant PM-derived periarbuscular membrane is compositionally distinct from the rest of the plant PM, with arbuscule-specific phosphate transporters (PT4 family), ammonium transporters, and sugar/lipid efflux carriers populated at the membrane. Exchange happens across the periarbuscular membrane: fungus delivers phosphate, ammonium, and amino acids; plant delivers sucrose and lipids; both flows cross the same substrate-coherent membrane interface. Arbuscules are short-lived — each turns over in \sim 4–10 days, with the periarbuscular membrane reabsorbed and a new arbuscule forming elsewhere, in a continuous symbiotic-interface-renewal cycle.

Ectomycorrhizas (EM). Basidiomycota and Ascomycota fungi (many independent lineages, \sim 6000–10000 named species) form extracellular symbioses with \sim 2\% of vascular-plant species — but those species include almost all dominant temperate-and-boreal forest trees. The architecture has two characteristic components. The mantle is a sheath of densely packed hyphae enclosing the root tip externally, \sim 10–40\;\mu\text{m} thick, which displaces the root cap and acts as the soil-facing interface for the symbiosis. The Hartig net is a network of hyphae growing between the epidermal and outer-cortical cells of the root, threading through the apoplastic space but not penetrating cell walls or plasma membranes. Exchange happens across the plant cell wall at the Hartig-net interface — fungus delivers nitrogen (especially as amino acids, with EM fungi specifically capable of accessing organic-N pools through their secreted proteases), phosphate, and water; plant delivers sucrose, with EM fungi notably not receiving plant lipids in the way AM fungi do. The Hartig net is long-lived relative to AM arbuscules, persisting across the root tip’s lifespan (\simweeks-to-months), with the symbiotic interface stable across seasons.

The framework reads AM and EM as two substrate-coherent symbiotic architectures populating the substrate-mechanical space of plant-fungal interfaces at two distinct topological positions. AM is the intracellular expression — the substrate-coherent boundary-matching transport zone is inside the plant cortical cell, on a periarbuscular membrane that is itself a substrate-coherent extension of the plant PM, with the arbuscule’s \sim 10^2–10^3-branch fractal architecture maximising membrane contact surface area. EM is the apoplastic expression — the substrate-coherent boundary-matching transport zone is between the plant cells, on the cell-wall interface, with the Hartig-net’s lateral hyphal extension distributed across the cortical-cell array. Both architectures solve the same substrate-mechanical problem (maximise substrate-coherent contact surface between two organism modons for substrate-current exchange) by two different chemistry-side strategies (intracellular fractal branching with engulfing PM, vs. apoplastic lateral extension along the cell-wall interface). The framework predicts the architectural-mode population is exhaustive — these are the two substrate-preferred symbiotic-interface architectures, and ericoid / orchid / monotropoid mycorrhizas are variations on the AM-intracellular template rather than a third categorically distinct architecture.

Hyphae as the Substrate’s Smallest Inter-Organism Conduit

A fungal hypha is a tubular cell wall of chitin and \beta-glucan, \sim 2–10\;\mu\text{m} in outer diameter, growing by apical extension through soil pores at \sim 0.1–1\;\text{mm h}^{-1}, branching at \simmm intervals, and extending the hyphal network at \sim 10–100\;\text{m} of hypha per gram of soil in mycorrhizal-active forest topsoil. Hyphae are coenocytic or septate (most Glomeromycota are coenocytic with no internal cross-walls; most Basidiomycota and Ascomycota are septate with regulated dolipore septa) and cytoplasmically continuous internally, supporting bidirectional protoplasmic streaming at \sim 10\;\mu\text{m s}^{-1} along the hyphal axis. Long-distance corridors are sometimes organised into rhizomorphs — aligned hyphal cords of \sim 0.1–1\;\text{mm} diameter with hundreds of hyphae running parallel, providing \simcm-to-m-scale low-resistance transport channels within the network.

The framework reads hyphae as the substrate’s smallest inter-organism corridor architecture, structurally parallel to the microtubule cylinder at intracellular scale, the desmotubule at inter-cell scale, and the xylem/phloem conduit at intra-organism scale — but now lifted one rung further, to inter-organism scale. The substrate’s closed-cylindrical-conduit family extends across at least five rungs: desmotubule (\sim 15\;\text{nm}) → microtubule (\sim 25\;\text{nm}) → axoneme (\sim 200\;\text{nm}) → fungal hypha (\sim 2–10\;\mu\text{m}) → vascular conduit (\sim 10–500\;\mu\text{m}). The hypha at \sim 2–10\;\mu\text{m} sits at the substrate-preferred rung between cell-scale and tissue-scale corridors, with the diameter set by the substrate-mechanical demands of apical-growth tip-region cytoplasm + chitin/glucan wall thickness + capacity for cytoplasmic streaming. The framework predicts hyphal-diameter clustering at this substrate-preferred rung across fungal phyla (Glomeromycota, Basidiomycota, Ascomycota, Mucoromycota), distinct from continuous variation with growth rate or substrate.

Rhizomorphs are biology’s chemistry-side implementation of bundled hyphal corridors at \simmm-scale, structurally parallel to the bundling of \sim 10^3–10^4 axons into a peripheral nerve in the animal body — same substrate principle (parallel substrate-coherent conduits aligned to share boundaries and run at higher aggregate throughput), different chemistry-side material (chitin and glucan vs. lipid bilayer and protein cytoskeleton). The framework reads the hypha-rhizomorph two-scale bundling architecture as substrate-preferred for inter-organism mass-and-information transport, with hyphae as the finest-scale individual conduit and rhizomorphs as the long-distance corridor.

Common Mycelial Networks as Inter-Modon Coupling

A single fungal individual (technically a single genet, with one genome shared across the entire hyphal network) can simultaneously host symbioses with multiple plant individuals — sometimes hundreds or thousands of trees across a forest patch, sometimes plants of different species linked through the same fungal genet, with the network persisting across decades or centuries. The largest documented mycorrhizal-fungus individuals (Armillaria solidipes in Oregon at \sim 8.9\;\text{km}^2 and \sim 2400 years old; also as bracket-fungus genets and yeast colonies) make mycorrhizal fungal individuals among the largest and longest-lived organisms on Earth. The architectural fact is robust: CMNs are real, with controlled isotope-tracer studies (since Simard 1997, replicated in many independent labs including Teste, Klein, Lerat, Selosse, Beiler, and others) documenting that water, nitrogen, phosphorus, and at least some carbon move from one connected plant to another through the network.

The framework reads CMNs as inter-modon coupling at forest scale. The plant body is one substrate-coherent modon from root tip to canopy (the architectural reading the prior plant chapters established); the mycorrhizal network couples plant modons into a connected substrate-coherent system at the forest rung. Structurally, this is parallel to the role synapses play at the brain rung: chemistry-side regulated bridges between substrate-coherent modon individuals that maintain the partners’ individual modon-coherence while creating a substrate-coherent inter-modon coupling at a larger rung. Synapses bridge neurons into a brain; mycorrhizal interfaces bridge plants into a forest. Two chemistry-side implementations, same substrate-architectural commitment.

The substrate-current that flows across the inter-modon coupling is different at the two rungs. At the brain rung, the substrate-current is ms-scale electrochemical (neurotransmitter release, postsynaptic depolarisation, action potential propagation). At the forest rung, the substrate-current is hours-to-days-scale chemical and hydraulic mass-flow (water, nutrients, signal molecules, and at least some carbon, moving through hyphal protoplasm and apoplastic-fungal interfaces). Both are biology’s chemistry-side implementations of substrate-current circulation at the inter-modon coupling rung; the timescale difference reflects the substrate-channel timescales each system tracks (brain: predator-prey ms-scale; forest: seasonal-and-multi-year hydrological-and-mineral-cycle scales).

The “Wood Wide Web” with Appropriate Hedges

The popular Wood-Wide-Web framing makes three distinct claims that need to be separated and weighed on their own evidence.

Claim 1 — the architecture exists. Hyphal networks physically connect root systems of multiple plants across a forest patch; isotopic tracers move between connected plants through the network. This claim is well-established. Genetic tracking of fungal genets across forest plots (Beiler et al. 2010, 2015 on Douglas-fir EM networks) directly maps the architecture; isotopic experiments (Simard 1997 and many successors) directly document the flux. The framework treats this as data.

Claim 2 — significant flux happens. The flux of carbon, water, and nutrients between connected plants through the network is large enough to be ecologically relevant for the receiving plant’s fitness. This claim is contested. The Karst-Jones-Hoeksema 2023 systematic review of \sim 800 Wood-Wide-Web claims in the literature found positive citation bias (papers with supporting findings cited disproportionately more than null findings), overinterpretation of single-experiment results, and a much weaker pattern in the underlying data than the popular narrative implies. Their headline finding: of \sim 75 studies measuring inter-plant carbon transfer through CMNs, only a minority found statistically significant transfer of biologically relevant magnitude, and most found small (~ 0.1–1\% of recipient-plant carbon) or non-detectable transfer. The flux is real — but the magnitude relative to other carbon-flow channels (soil pore water, atmospheric exchange, intra-plant phloem) is small in most studies.

Claim 3 — the network functions as a communicating community. Mature “mother trees” preferentially transfer carbon to kin seedlings; the network mediates warning signals about herbivory; inter-tree resource sharing has been selected by community-level evolutionary processes. This claim is more strongly contested still. Most of the specific Simard-style “mother tree” claims are based on a small number of original experimental studies that have not been independently replicated at scale, and the broader claims about network-mediated community function rest on functional inferences that have generally not been tested directly. The Karst et al. critique points out that these claims have been amplified through popular-science channels (notably Suzanne Simard’s 2021 book Finding the Mother Tree and Peter Wohlleben’s The Hidden Life of Trees) to a degree the underlying data does not yet support.

The framework’s position is the same as at the Penrose-Hameroff rung in the brain walk. Claim 1 (the architecture) is the substrate-framework-loadbearing claim, and it is well-established. Claim 2 (significant flux) is a quantitative empirical question that the framework predicts should resolve in favour of substrate-coherent magnitude clustering (see prediction 3 below), but does not require to be large for the substrate-architectural reading to hold. Claim 3 (the network as a communicating community) is a chemistry-side adaptive-interpretation question that the substrate framework does not need to take a position on; the substrate-coherent inter-organism corridor exists regardless of whether Simard’s specific adaptive narrative survives further evidence. The framework reads Claim 1 as the architectural existence-proof for inter-organism substrate-current coupling in plants; Claims 2 and 3 are layers on top of the architectural commitment that may or may not survive the empirical literature’s accumulation, and the substrate-physics reading is robust either way.

The Forest as Substrate-Coherent Coherence Cell

A forest — operationally, a contiguous patch of vascular plants linked by a continuous mycorrhizal-and-soil-microbial network, with a closed canopy supporting microclimatic regulation of light, temperature, and humidity, and a substrate-coherent soil environment maintained by the network’s own chemistry-side activity — is at 10^1–10^4\;\text{m} scale the largest substrate-coherent biological assembly the plants section reaches before forward-linking to the Gaia chapter. The framework reads the forest as a substrate-coherent coherence cell at the inter-organism rung, structurally parallel to the cortical column-and-hypercolumn at organ scale and to the cell modon at cellular scale. Coherence cells are the substrate’s preferred tiling of substrate-coherent integration zones at every scale at which biology operates — at \sim 100\;\mu\text{m} cortical columns inside a brain, at \sim 10–100\;\mu\text{m} cells in a tissue, and at \sim 10^1–10^4\;\text{m} forest patches across a landscape, the same architectural pattern repeats: substrate-coherent integration over a characteristic spatial scale, with chemistry-side machinery regulating the integration and chemistry-side bridges connecting one coherence cell to its neighbours.

The forest’s substrate-coherent integration scale is set by the mycorrhizal-and-microbial network’s characteristic span and by the canopy-and-microclimatic regulation’s extent — not by the individual plant’s root or canopy reach. The forest as a single coherence cell is a network-emergent property of multiple plant individuals coupled through inter-organism corridors, structurally parallel to the brain’s organ-scale coherence cell as a network-emergent property of multiple neurons coupled through synapses. The substrate-coherent forest is not the sum of trees; it is the substrate-coherent integration zone the network of trees + mycorrhizal-and-microbial coupling sustains across the patch. A forest with the mycorrhizal-and-microbial network removed is not a forest with damaged communication; it is a collection of trees that has stopped being a forest at the substrate-coherent-coherence-cell level. This is the framework’s reading of the standard ecological observation that clearcut-and-replant monocultures persistently underperform mycorrhizally-intact forest regeneration in productivity and resilience: the architectural substrate (the network) is what makes the assembly a forest at the substrate-coherent coherence-cell level, and without it the chemistry-side machinery has the trees but not the coherence cell.

Predictions and What Would Falsify

Four predictions extend the architectural reading beyond the structural anchors.

  1. Hyphal diameter clusters at the substrate-preferred \sim 2–10\;\mu\text{m} rung across mycorrhizal fungal phyla. Glomeromycota, Basidiomycota, Ascomycota, and Mucoromycota hyphal diameters should cluster at substrate-preferred discrete rung values rather than vary continuously with growth rate, substrate, or species. Cross-phylum hyphal-anatomy datasets (since Bonfante and Genre 2010 and successors) provide the existing data; the framework predicts substrate-preferred-rung clustering at the substrate’s smallest-inter-organism-corridor rung.

  2. Arbuscule branching geometry follows a substrate-coherent fractal dimension and terminal-branch-count rung across plant-AM-fungus pairings. The fractal-dimension scaling and the terminal-branch count of arbuscules (typically \sim 10^2–10^3 terminal branches per arbuscule) should cluster at substrate-preferred discrete values across diverse plant-fungus pairings, distinct from continuous variation with cell size or symbiotic-flux demand. Cryo-electron tomography and confocal-microscopy datasets of arbuscule architecture provide the test; the framework predicts cross-pairing clustering at a small set of substrate-preferred branching geometries.

  3. Inter-plant carbon transfer through CMNs clusters at a substrate-coherent flux ratio relative to plant total carbon budget across temperate and tropical forest systems. The fraction of recipient-plant carbon derived from inter-plant transfer through a CMN should cluster at a substrate-coherent ratio (likely small, of order \sim 0.1–1\% of recipient-plant carbon for most systems, with possibly higher ratios at specific seedling-recruitment or stressed-donor conditions) rather than vary continuously across systems. This prediction is directly diagnostic of the Karst-vs-Simard question: if inter-plant carbon transfer is continuously variable and frequently high, the Wood-Wide-Web ecological-significance claims gain support; if it is substrate-coherently clustered at small magnitudes with rare condition-specific exceptions, the substrate-architectural reading (corridor real, magnitude modest, ecological significance condition-specific) is supported. The existing literature provides enough data for an initial test (Karst et al. 2023 collated \sim 75 studies); the framework predicts substrate-coherent ratio clustering rather than continuous adaptive-significance variation.

  4. Mycorrhizal-association type populates substrate-preferred biome channels exclusively, not arbitrary host-fungus combinations. Across biomes, plant-mycorrhizal-association types (AM, EM, ericoid, orchid, monotropoid, non-mycorrhizal) should populate substrate-preferred biome channels (AM in grass-and-herb-dominated systems and tropical forests; EM in temperate and boreal forests; ericoid in heath and bog; orchid in orchid-mycorrhizal symbioses; non-mycorrhizal in restricted niche conditions) rather than appear as arbitrary host-fungus combinations across biome space. Existing global mycorrhizal-association datasets (Soudzilovskaia et al. 2020, Steidinger et al. 2019, etc.) provide the data; the framework predicts substrate-preferred biome-channel populating distinct from continuous host-fungus-trait-driven combinations.

The picture is falsified if (a) hyphal diameters vary continuously without substrate-preferred-rung clustering across fungal phyla, (b) arbuscule branching geometries vary continuously without substrate-coherent-rung clustering, (c) inter-plant carbon-flux fractions vary continuously across systems without substrate-coherent-ratio clustering (in which case the Wood-Wide-Web ecological-significance claims gain support against the framework’s architectural-but-modest reading), or (d) mycorrhizal-association types populate biome space arbitrarily without substrate-preferred-channel clustering. It is supported, even partially, if any of the four ordering predictions hold against existing data.