The Modular Tropical Agroforestry Recipe Book

Synthesis model: GLM 5.2

A Super-Synthesis Integrating Architecture, Belowground Mechanisms, Mycorrhizal/Microbial Symbiosis, and Mushroom & Pest Ecology

---

Provenance

This document is the capstone meta-synthesis of a four-scope research project on tropical agroforestry, conducted in the Open Knowledge Tree repository default (UUID 3028db69-f2bd-49f8-a7b7-f2c7cc705278).^{1:rel6:rel7:rel8:rel9:rel} The research gathered ~1,300 sources (from a starting base of <300), producing 100,000+ new facts across 16 investigations. Four parallel synthesizer agents produced sub-synthesis documents:

1. Tropical Polyculture Architecture & Companion Guild Design — classic models (homegardens, cabruca, shaded coffee, Inga alley, syntropic, breadfruit), companion-planting mechanisms, stratification, succession, purpose-cells.
2. Belowground Mechanisms — root architecture, hydraulic lift, shade↔coffee leaf rust, N-fixation & transfer, P-mobilization, mulch decomposition, chop-and-drop, the "fertilizer tree" debate.
3. Mycorrhizal Networks & Soil Microbiome — AMF/ECM compatibility, the contested common mycorrhizal network ("wood wide web") debate, rhizobia, biochar, soil food web, inoculation management.
4. Mushrooms & Pest Ecology — tropical mushroom species, agroforestry-derived substrates, the Trichoderma dual-role, the termite–Termitomyces–Bacillus three-way link, push-pull, the pest/benefit duality the user explicitly requested.^{10:rel11:supp12:supp13:supp14:supp}

This super-synthesis integrates them into an operational recipe book. It is NOT a summary — it cross-pollinates findings across scopes into a higher-order design framework with modular cells the user can assemble.

A note on epistemics: Every claim below traces to the evidence in the four sub-syntheses. Where the evidence is contested, both scenarios are held in tension rather than collapsed.^{15:rel16:rel17:rel18:rel19:rel} Where a stakeholder's institutional, ideological, or commercial incentives materially shape a claim, those incentives are named. Neutral labels ("alternative", "contested", "unverified") are used throughout — never loaded ones ("fringe", "pseudoscientific", "debunked").^{20:rel21:rel22:rel23:rel} The user explicitly asked that pest organisms not be treated as black-and-white; this document honors that.

---

Part 1 — The Integrative Framework

1.1 The four layers and why they cannot be designed separately

A tropical agroforest is not a collection of plants; it is four interlocking biological layers that must be co-designed because each layer's function depends on the others:^{24:supp25:supp26:supp1:rel27:rel}

Layer	What it does	Key organisms	Depends on
Architecture (aboveground)	Captures light, structures habitat, produces yield	Trees, shrubs, herbs, crops	Root space, shade transmission, pollinators, pest balance
Belowground mechanisms	Partitions water & nutrients, fixes N, mobilizes P, decomposes mulch	Roots, exudates, decomposers	Mycorrhizae, soil structure, mulch input
Mycorrhizal/microbiome	Symbiotic nutrient & water uptake, defense priming, N-fixation	AMF, ECM fungi, rhizobia, PSB, Bacillus, soil food web	Root exudates, mulch carbon, compatible hosts
Mushroom & pest ecology	Decomposes lignocellulose, produces mushroom food, balances pests & beneficials	Pleurotus, Termitomyces, Trichoderma, Beauveria, insects, birds	Substrate from mulch stream, habitat from architecture, mycorrhiza-induced defense

A design that optimizes any single layer in isolation will fail. Maximizing mushroom production invites Trichoderma contamination that also destroys the plant-biocontrol function of the same genus.^{28:rel29:rel30:rel31:supp32:supp} Maximizing shade for pest suppression can increase coffee leaf rust under the wrong shade-tree species. Maximizing a legume's N-fixation by not pruning it deprives the system of its mulch stream and creates belowground water competition.^{33:rel34:rel35:supp36:supp37:rel} The recipe book is therefore modular but coupled — each cell is a unit, but the cells share the mulch stream, the mycorrhizal network, the pollinator pool, and the pest-natural-enemy balance.

1.2 The cross-scope bridges (the integration that makes this a super-synthesis)

These are the load-bearing connections across the four scopes. Each is an actionable design constraint.

Bridge 1 — Mycorrhizal compatibility gates species combinations. Most tropical trees form arbuscular mycorrhizae (AMF), but Dipterocarpaceae (Southeast Asian dominant family), some Fabaceae (Dicymbe, Aldina in Amazonia), and some Myrtaceae (Eucalyptus) form ectomycorrhizae (ECM).^{38:supp39:rel40:supp41:supp42:rel} Eucalyptus is dual: AMF as a seedling, ECM at maturity. A cell mixing an AMF crop (cacao, coffee, banana) with an ECM tree (Dipterocarp, certain legumes) may function poorly because the two mycorrhizal types build different soil communities and the plants do not share a common mycorrhizal network.^{43:rel44:rel45:rel46:supp47:contr} Design rule: group plants by mycorrhizal type within a cell; do not mix AMF and ECM hosts in the same root zone unless the system is explicitly designed for it (e.g., Eucalyptus transition management).

Bridge 2 — The mulch stream connects every cell. Prunings and leaf fall from the architecture layer (Scope 1) become mulch (Scope 2), which feeds mycorrhizae and decomposers (Scope 3) and becomes mushroom substrate (Scope 4).^{48:rel49:supp50:supp51:supp52:supp} Spent mushroom substrate returns as compost. This is the circular flow that makes the system a closed loop. Coffee husk/pulp is the single strongest documented substrate bridge — it is simultaneously mulch, mushroom substrate (Pleurotus), and a shaper of edaphic predatory-mite densities (a pest-management effect).^{53:supp54:supp55:supp56:supp57:supp} One byproduct, three cells. Design rule: locate the mushroom cell adjacent to the coffee/cocoa cell so the byproduct stream flows directly.

Bridge 3 — The shade↔pest↔mycorrhizae triangle. Shade architecture (Scope 1) shapes pest pressure — coffee leaf rust (CLR) increases under certain shade-tree species (Chloroleucon eurycyclum reduces spore wash-off) but decreases under others (shade supports natural enemies that suppress berry borer and other pests).^{58:contr59:supp60:supp61:supp62:supp} Shade also shapes mycorrhizal colonization (Ethiopian homegardens show highest AMF spore abundance). And shade shapes mulch moisture (which drives decomposition).^{63:rel64:supp65:supp66:supp67:supp} These are not independent decisions. Design rule: match the shade-tree species to the local CLR pressure, natural-enemy community, and mycorrhizal needs — never prescribe "more shade" or "less shade" as a blanket rule.

Bridge 4 — The nitrogen-fixation triangle. A legume (Scope 1) fixes N via rhizobia (Scope 3), transfers it via root exudates and common mycorrhizal networks (Scopes 2 & 3), but pruning reduces nodulation (Sanginga 1990: successive cutings reduce nodulation) — the very pruning that delivers mulch suppresses the N-fixation that justifies the legume.^{68:rel69:rel70:supp71:supp34:rel} Young Inga fixes little N (Grossman 2006). The "fertilizer tree" is therefore a system property (legume + rhizobia + mycorrhizae + appropriate pruning regime + time), not a species property. Design rule: do not assume a legume is fixing N; verify nodulation, allow establishment time before relying on N transfer, and balance the pruning frequency against the nodulation cost.^{34:rel72:rel69:rel73:supp74:rel}

Bridge 5 — The Trichoderma tension. Trichoderma is simultaneously the worst competitor of cultivated mushrooms (green mold disease of Agaricus and Pleurotus) AND one of the most widely used biocontrol fungi against plant pathogens. A cell cannot simultaneously maximize mushroom production and Trichoderma-based plant biocontrol without spatial or temporal separation.^{75:supp76:rel32:supp77:rel78:rel} The termite–Termitomyces–Bacillus symbiosis (Bridge 7) suggests Bacillus may be the organism that lets a system hold this tension by suppressing Trichoderma where it is unwanted without eliminating it system-wide. Design rule: separate the mushroom cell spatially from any deliberate Trichoderma deployment; consider Bacillus co-inoculation as a Trichoderma-management tool.^{79:supp75:supp80:supp81:rel31:supp}

Bridge 6 — Mycorrhiza-induced plant defense. AMF colonization (Scope 3) primes plant defenses against insect pests (Scope 4) — mycorrhiza-induced resistance (MIR) is documented (Jung 2012, 1313 cites) and applicable to woody plants. The belowground symbiosis is a pest-management instrument.^{82:rel83:contr84:supp85:supp86:supp} Design rule: managing for high AMF colonization (low disturbance, diverse hosts, native inoculum) is also a pest-management intervention.

Bridge 7 — The termite–Termitomyces–Bacillus three-way cell. This is the densest cross-scope integration in the entire evidence base.^{87:supp88:rel89:rel90:supp91:rel} The fungus-growing termite Macrotermes natalensis (a structural pest) cultivates Termitomyces (a prized edible mushroom) and harbors bacillaene-producing Bacillus strains that inhibit antagonistic fungi (Pseudoxylaria and Trichoderma) without harming the mutualistic fungus (Um et al. 2013, Scientific Reports). One fact, five organisms, held in tension. This is the basis for Cell 10 — the experimental termite-managed mushroom cell — which is the most innovative and least-tested recipe in this book. Design rule: this is a research target, not a proven recipe. Flag as the most promising direction for further investigation.

Bridge 8 — Hydraulic lift ↔ common mycorrhizal networks (both contested). Deep-rooted trees may lift water hydraulically (Scope 2) and share it via CMNs (Scope 3) — but the CMN-mediated transfer evidence is itself contested (Warren et al.^{92:supp93:supp94:rel95:supp96:rel} 2008 caution their result could be a hyphal-growth artifact, not true CMN transport), and the agronomic benefit of HL to a companion crop at field scale is unproven. Design rule: treat HL water-sharing as plausible upside, never as load-bearing.^{97:rel98:rel99:rel100:rel101:rel} Choose deep-rooted companions for subsoil nutrient capture and shade; treat any HL water transfer as a bonus.

1.3 The circular economy of a tropical agroforest^{1:rel2:rel102:rel3:rel103:rel}

The system's defining property is that it closes loops. The mulch stream flows from architecture → belowground → mycorrhizae → mushrooms → compost → soil → back to plants. A well-designed agroforest is a biological circular-economy machine:^{104:rel105:rel106:rel107:rel108:supp}

Architecture (pruning, leaf fall)
      ↓
Mulch layer (Scope 2)
      ↓ (decomposition, nutrient release)
Mycorrhizae & soil food web (Scope 3) ←→ Soil nutrients
      ↓ (lignocellulose byproducts)
Mushroom cell (Scope 4) → food
      ↓ (spent substrate)
Compost → Soil organic matter → Roots → Architecture

Every cell in Part 3 below specifies what it consumes from and contributes to this circular flow.

---

Part 2 — Cross-Cutting Tensions (Held, Not Resolved)

The evidence does not resolve these debates.^{109:supp110:supp111:supp112:supp15:rel} A recipe book that pretended otherwise would be dishonest. Each tension is presented with both scenarios at full strength.

2.1 Common mycorrhizal networks ("wood wide web")

Scenario A (transfer is real): Simard 1997 (1388 cites, foundational birch–Douglas fir carbon transfer), Teste 2008, Gorzelak 2015, Klein 2023, Luo 2023, Bock 2025 (Dark Septate Endophyte CMN), Babikova 2013 (aphid defense signal transfer), Song 2014 (herbivore defense signal), Lehmann 2025 meta-analysis, the Mother Tree Project.^{113:supp114:contr115:supp116:contr117:contr}

Scenario B (claims overstated): Karst 2023 (228 cites — positive-citation-bias and misinformation critique), Robinson 2024 ("perils of plant personification," 66 cites), Barto 2012 (questioned), Fitter/Robinson 1999 (small magnitude), Bücking 2016 (absence of transfer / fungal bargaining power), Popkin/NYT journalism critique, Selosse 2006 (balanced review). Simard 2025 published a rejoinder.

The tropical-application gap: Almost all CMN research is temperate/boreal (Douglas fir, birch, tomato, potato).^{118:supp40:supp119:rel120:rel121:rel} No direct tropical multistrata CMN study appears in the evidence base. Tropical application is by inference, not direct evidence.^122:rel This is a finding, not a dismissal.

Recipe implication: Do not design a cell assuming CMN-mediated resource sharing is load-bearing. Manage for mycorrhizal compatibility and diversity; treat any CMN-mediated transfer as a possible bonus.

2.2 Shade ↔ coffee leaf rust

Scenario A (shade increases CLR): Avelino et al.^123:rel 2020 — the shade tree Chloroleucon eurycyclum promotes CLR by reducing uredospore wash-off by rain (the canopy intercepts throughfall). Liebig et al. 2019 — interactive effects of altitude, microclimate, and shading.^{124:rel125:rel126:rel127:rel} World Coffee Research explicitly frames the question as unresolved.

Scenario B (shade reduces certain pests/diseases): Staver et al.^{128:supp62:supp129:supp} 2001 — 35–65% shade reduced Cercospora brown eye spot and mealybug without raising CLR or lowering yield in Central America. Piato et al. 2021 — organic + shade reduces pest infestations in Robusta (Amazonia).^{130:rel131:supp132:rel133:supp134:supp} Shade supports ant and bird biocontrol of berry borer. Piato 2020 meta-analysis: shade may enhance yield, growth, biodiversity, though few studies test pest/disease effects.^{135:supp136:rel137:supp138:rel139:rel}

Recipe implication: The discriminating variables are shade-tree species identity (C. eurycyclum increases CLR; others may not), canopy architecture (throughfall vs interception), and the local pest-natural-enemy balance.^{140:supp141:rel142:rel61:supp59:supp} Avoid dense-shade, low-throughfall species in CLR-prone zones; favor open-canopy, high-throughfall species paired with ant- and bird-supporting shade diversity.

2.3 "Fertilizer tree" / Evergreen Agriculture^{143:rel144:rel145:rel146:rel147:rel}

Regenerative perspective (FOR): Faidherbia albida reverse phenology (leafless in wet season, leafy in dry — non-competitive); deep nutrient capture and P recycling (Sileshi 2016); 6.7 t/ha wheat under F. albida (Sida 2018); reduces Striga and improves AMF (Birhane 2018); FMNR landscape restoration across millions of hectares in Niger/Burkina Faso; ICRAF framing of avoided emissions and foreign-exchange savings (Haskett 2019).^{148:supp149:supp150:supp151:supp152:supp}

Conventional nutrient-budget perspective (AGAINST): 18 kg N/ha/yr from litter (Yengwe 2018) is modest vs crop demand (maize can remove 50–150 kg N/ha in grain); the 6.7 t/ha wheat figure is under-tree-canopy, not field-average; reverse phenology does not eliminate wet-season root competition. Beer 1988: N-fixation may be over-emphasized; organic-matter production may matter more.^{153:supp154:supp155:rel34:rel156:supp} Turvey & Smethurst 1983: N fertilizer is cheaper. Kou-Giesbrecht & Menge 2019: N-fixing trees could exacerbate climate change under elevated N deposition.^{157:supp158:supp159:rel}

Stakeholder motivations (symmetric): ICRAF (institutional — career/funding interest in scaling Evergreen Agriculture as a brand); regenerative practitioners (ideological/attention — flagship narrative); conventional agronomy (nutrient-budget frame — standard counter-frame). All three carry incentives; all three are engaged symmetrically.

Recipe implication: F. albida is the strongest evidence-based "fertilizer tree" for semi-arid African cereal systems with its unique reverse phenology. Outside its native range, verify actual N contribution before labeling any legume a "fertilizer tree."

2.4 Companion planting / push-pull

Mechanism-validated (FOR): N-transfer legume→non-legume (Thilakarathna 2016; isotopic ¹⁵N evidence, Snoeck 2000; tree→grass transfer, Jalonen 2009).^{160:supp161:supp162:supp163:supp164:supp} Push-pull semiochemistry (Desmodium repels stemborers & inhibits Striga; Napier grass attracts them — Rothamsted/ICIPE). Allelopathy (teak, eucalyptus, sunn hemp, velvet bean — isolated compounds documented).^{165:rel166:rel167:rel168:supp169:supp}

Observational/unreplicated (CAUTION): The broader companion-planting literature (284 facts) is predominantly temperate-vegetable-oriented (tomato-basil, carrot-onion) and characterized by the permaculture synthesis itself as "claims from guides rather than established scientific knowledge." Many repellent-volatile claims are "not experimentally supported" for aphid colonization.

Push-pull specifically: The mechanism is well-evidenced (kairomones, Striga inhibition).^{170:supp171:rel172:rel173:rel174:supp} The economic claims (ROI >2.2, 61.9% yield increase) rest on a narrower evidentiary base — the push-pull synthesis self-assesses that its sources are "internally consistent but potentially incomplete" (single advocacy article, Wikipedia, one research paper; no farmer critiques or failed adoptions represented).

Recipe implication: Use companion planting where the mechanism is documented (N-transfer legumes, push-pull in African maize, allelopathic weed suppressors).^{175:supp176:supp177:contr178:contr179:rel} Test, don't assume, where the mechanism is inferred. Do not treat a volatile's presence as efficacy.

2.5 Trichoderma — antagonist or ally^{77:rel32:supp180:rel181:rel182:rel}

Both, simultaneously, context-dependent. Trichoderma causes green mold disease (the most damaging fungal disease of cultivated mushrooms — T. aggressivum on Agaricus, related species on Pleurotus).^{183:supp184:supp185:rel} Trichoderma is also the most widely used biocontrol fungicide for plant pathogens (mycoparasitism). The discriminating variable is which Trichoderma species in which context.^{75:supp77:rel186:supp187:rel188:supp} No single experiment tests both roles in one system. The question is open by design.

Recipe implication: Spatial/temporal separation, not labeling.^{189:supp190:supp191:rel192:supp193:supp} Consider Bacillus co-inoculation to suppress Trichoderma where unwanted (per the termite–Termitomyces–Bacillus model).

2.6 Termites, leafcutter ants, "weeds" — pest or keystone (the user's explicit non-black-and-white framing)^{194:supp195:supp196:supp197:supp198:rel}

Leafcutter ants (Atta): Defoliate crops (Inga's "few pests except leafcutter ants") BUT excavate up to 23 m³ / 40 tonnes of soil per nest (massive aeration and turnover), run a decomposition pipeline via their fungal gardens, and are prey for birds, bats, and army ants (a food-web hub). Near an Inga alley they are a crop pest; in a restoration zone they are soil-builders.^199:rel Same organism.

Termites (Isoptera): Structural and crop pests BUT major decomposers, soil aerators, and — through Macrotermitinae — obligate symbionts of Termitomyces (a prized edible). Termite management is framed as a soil-productivity tool ("Managing termites and organic resources to improve soil productivity in the Sahel"), not a pest-elimination target.^{200:supp201:supp202:supp196:supp203:supp} The Um 2013 fact ties termites to Bacillus (a mushroom helper and plant biocontrol) and Trichoderma suppression. Removing the pest removes the mushroom, the Bacillus, and the soil-building.^{204:rel205:rel206:rel207:rel}

"Weeds": Host both pests AND beneficials (not simply pest reservoirs). Diversification (not weed elimination) is the pest-suppression lever documented across coffee, cocoa, maize, and the weed-beneficial literature.^{208:supp209:contr210:rel211:supp212:supp} Organic practices + shade reduce pest pressure in Robusta (Amazonia).

Recipe implication: Never label an organism simply "bad" or "good." Design for the pest/beneficial balance via system architecture, not via eradication.^{213:supp214:supp215:supp208:supp216:supp} Outbreaks are symptoms of system imbalance, not intrinsic properties of pest species.

2.7 Inoculation — commercial or native^{217:rel218:rel219:rel220:rel221:rel}

Commercial inoculants often fail: Salomon 2022 (28 commercial AMF inoculants tested globally; many fail to colonize). Basiru 2022 (invasion risk of commercial inoculants).^{222:supp223:supp224:supp225:rel226:supp}

Native inoculum preservation: Koziol 2017 (practical guide to native AMF inoculation). Cardoso 2006 (759 cites — "management better than inoculation"). Rodale (on-farm AMF). Miyasaka 2003 (Hawaii manual).

Recipe implication: In tropical agroforestry, prefer native inoculum preservation and management (low disturbance, diverse hosts, organic matter) over commercial inoculants unless a specific inoculant is locally validated. Disturbance (tillage, glyphosate, fire/slash-and-burn) severely reduces AMF — manage disturbance, not just inoculation.^{227:supp228:supp229:supp230:supp231:supp}

2.8 Syntropic vs conventional — the literature-availability asymmetry

Conventional agroforestry (ICRAF, CATIE, peer-reviewed cacao/coffee) has dense, replicated, long-term coverage.^{232:rel233:rel234:rel235:rel236:rel} Syntropic agriculture (Ernst Götsch) has mostly practitioner documentation, Agenda Gotsch project material, and grey literature; independent peer-reviewed replicated trials comparing syntropic rows to conventional agroforestry at matched sites are sparse. Traditional/indigenous systems have rich ethnobotanical documentation but less yield-quantified evidence.^{237:rel238:rel239:rel}

This is a finding about the literature, not a verdict on the systems. The convergence of the ecological framework-species restoration literature, the syntropic succession diagram, and traditional breadfruit agroforestry's staged canopy on "managed succession" as a shared structural logic is an agent-noticed pattern, not a claim from any single source.

---

Part 3 — The Recipe Cells

Each cell is a modular design unit. For each: Purpose, Species by stratum, Mycorrhizal compatibility note, Successional phase, Key interactions, Circular links (consumes from / contributes to), Caveats & open questions.^{240:rel241:rel24:supp242:rel243:rel}

Cell 1 — Agroforest Health / Soil-Building Cell (the foundation cell)

Purpose: Establish soil fertility, organic matter, mycorrhizal networks, and nitrogen capital on degraded or bare land.^{244:supp245:rel246:supp247:supp248:supp} This is the cell to plant first.

Species by stratum:

• Emergent/canopy: Faidherbia albida (semi-arid Africa — reverse phenology), Gliricidia sepium, Inga edulis
• Shrub/small tree: Leucaena leucocephala (with allelopathy caveat), Erythrina poeppigiana, Sesbania rostrata, Cajanus cajan
• Herb/groundcover: Mucuna pruriens (velvet bean — green manure + allelopathy), Crotalaria juncea (sunn hemp — green manure + weed suppression), Pueraria phaseoloides, sweet potato (living mulch)

Mycorrhizal note: All listed species are AMF hosts.^{249:rel250:rel251:rel252:rel253:rel} Compatible. Include a native grass (e.g., Megathyrsus maximus) for organic-acid P-solubilization via root exudates. Consider biochar application to boost AMF colonization (Hammer 2014: AMF grows on biochar; Solaiman 2019: biochar-P dictates mycorrhizal colonization).^{254:supp255:supp256:rel257:supp258:supp}

Successional phase: Year 0–5. Pioneer/early. Inga pollarding cycle ~20–24 months.

Key interactions:

• N-fixation: Gliricidia, Inga, Leucaena, Erythrina, Sesbania fix via rhizobia. Caveat: young Inga fixes little N (Grossman 2006); successive cutings reduce nodulation (Sanginga 1990) — balance pruning frequency against the nodulation cost.
• Mulch: legume prunings are high-N, low-lignin → fast N release (Leblanc 2006 Inga; Palm 1990). Mucuna/sunn hemp add allelopathic weed suppression.
• P-mobilization: root exudates (organic acids) + AMF + P-solubilizing bacteria (Pantigoso 2023) work synergistically. Co-inoculate AMF + PSB if native inoculum is poor (but prefer native — Salomon 2022 shows commercial inoculants often fail).
• Mycorrhizae: low disturbance (no tillage, no glyphosate, no fire) preserves AMF. F. albida improves AMF spore density (Birhane 2018).
• Deep nutrient capture: F. albida and Gliricidia deep roots capture subsoil nutrients (Buresh 1996; Sileshi 2016).

Circular links:

• Consumes from: nothing (this is the foundation).
• Contributes to: mulch stream (feeds Cells 2, 4, 6); mycorrhizal inoculum (benefits all cells); N capital (benefits Cells 3, 4, 5); pollinator habitat (when Inga/Gliricidia flower — feeds Cell 7).^{259:supp50:supp260:rel261:rel262:supp}

Caveats: Eucalyptus and teak allelopathy (Kato-Noguchi 2021; Liu 2019) — do NOT combine with N-fixing understory. Leucaena allelopathy may not release at inhibitory concentrations in field (2025 study).^{263:contr264:rel265:contr266:contr} The "fertilizer tree" label should be verified against actual N contribution (18 kg N/ha/yr from F. albida litter is modest vs crop demand).^267:rel Tephrosia is a pest-repellent legume BUT can build up root-knot nematodes (Meloidogyne) — a non-black-and-white trap.

---

Cell 2 — Fruit & Canopy Staples Cell^{268:rel269:rel270:rel271:supp}

Purpose: Year-round fruit and starchy staple production from a multi-strata canopy.

Species by stratum:

• Emergent (25–40 m): Coconut (Cocos nucifera), sago palm, Brazil nut (Bertholletia excelsa — long-cycle climax), native cabruca emergent timber
• Canopy (8–20 m): Breadfruit (Artocarpus altilis), mango (Mangifera), jackfruit, avocado
• Mid-story (2–8 m): Banana (Musa acuminata), papaya, citrus, cacao, coffee (in sub-optimal variants)
• Groundcover (0–1 m): Sweet potato, pineapple, medicinal herbs, ferns^{25:supp269:rel272:supp268:rel273:supp}

Mycorrhizal note: All AMF hosts. Compatible. Breadfruit/coconut/banana/cacao form the classic Pacific four-layer polyculture (Elevitch & Ragone).

Successional phase: Papaya/pineapple as Year 0–4 pioneers; banana and breadfruit as mid-term (Year 2–10); coconut and Brazil nut as 10–20+ year climax.^{274:supp275:rel276:rel277:supp278:supp} Matches the syntropic succession diagram (Stage 1 pioneers → Stage 4 climax).

Key interactions:

• Shade: breadfruit/coconut overstory shades mid-story; PAR transmission tuned by pruning. 50% canopy density supports cacao, coffee, vanilla, black pepper, cardamom understory (NTBG McBryde).
• Pollinators: flowering shade trees (Inga, citrus, mango, banana male flowers) support bees — bridges to Cell 7. Cacao pollinated by midges (Forcipomyia) needing leaf litter + humid shade — a different guild from coffee's bees.
• Roots: coconut/mango/citrus deep-rooted (Lehmann 2003) — complement shallow-rooted cacao/coffee (Bai 2023 Canarium deeper than cocoa).
• Mulch: banana and papaya provide fast-cycling early biomass.^{279:supp280:supp281:supp282:supp235:rel}

Circular links:

• Consumes from: Cell 1 (N capital, mycorrhizal inoculum, mulch stream); Cell 7 (pollinators).
• Contributes to: mulch stream (prunings → Cells 1, 3, 6); fruit to harvest; shade microclimate (benefits understory).

Caveats: Breadfruit economic projections in the Agroforestry Guide are Hawaii-based and promotional — traditional-architecture evidence is stronger than commercial-projection evidence.^{283:rel284:rel285:rel} Varietal diversity essential (Thomson — Pohnpei 131 varieties). Spacing differs between commercial (10×8 m) and traditional dense (1–3 m) — trade-off not evaluated in evidence.^{286:supp287:supp288:rel289:supp290:rel} Cassava (in Cell 3) is allelopathic in some pairings — do not co-locate with sensitive seedlings.

---

Cell 3 — Food Staples Cell (Roots & Starchy Fruits)^{291:supp292:supp293:supp294:supp295:rel}

Purpose: Caloric staple production (carbohydrate) integrated with overstory.

Species by stratum:

• Root layer: Cassava (Manihot esculenta), yam, sweet potato (Ipomoea batatas), taro/giant taro (Alocasia macrorrhizos)
• Mid-story: Banana, breadfruit
• Overstory: timber or coconut from Cells 2/9^{279:supp280:supp296:rel297:supp298:supp}

Mycorrhizal note: Cassava and yam are AMF hosts. Compatible with Cell 2 overstory.

Successional phase: Cassava and sweet potato are pioneers (0–2 yr); taro and yam mid-phase; banana and breadfruit persist as canopy matures.^{275:rel276:rel297:supp296:rel274:supp} Cassava is a syntropic pioneer (Stage 1).

Key interactions:

• Cassava rapidly establishes biomass and root yield while shading soil.
• Sweet potato provides living groundcover.
• Root crops need loose, non-compacted soil — avoid site preparation on saturated soils (breadfruit synthesis).^{299:rel296:rel300:rel301:rel302:rel}

Circular links:

• Consumes from: Cell 1 (N capital); Cell 2 (shade overstory).
• Contributes to: caloric harvest; sweet potato living mulch suppresses weeds (benefits neighbors).

Caveats: Cassava depletes soil potassium heavily if continuously harvested.^{303:supp301:rel304:rel300:rel} Cassava is allelopathic in some pairings. This cell should rotate, not be permanent.

---

Cell 4 — Cacao/Coffee Polyculture Cell (the cash-crop cell with the shade↔pest↔pollinator triangle)^{234:rel305:supp235:rel306:rel307:rel}

Purpose: Cacao or coffee production under managed shade, integrating the shade↔pest↔pollinator triangle.

Species by stratum:

• Emergent timber: Cordia alliodora (Central America), native cabruca emergents
• Shade/N-fixing: Inga edulis, Gliricidia sepium, Erythrina poeppigiana
• Crop: Theobroma cacao (cabruca, Colombian Amazonia, Ghana) or Coffea arabica/Canephora
• Understory/flowering: insectary herbs (feeds Cell 8)^{235:rel234:rel308:rel309:supp310:supp}

Mycorrhizal note: Cacao and coffee are AMF hosts. Compatible with Inga/Gliricidia/Erythrina (also AMF).^{311:rel312:rel261:rel313:rel314:rel} Avoid Dipterocarp or ECM legume shade trees (mycorrhizal mismatch). AMF modifies cacao P acquisition (Azizah Chulan 1986).^{315:supp316:supp317:supp318:supp319:rel}

Successional phase: Inga shade establishes Year 2; cacao/coffee yields from Year 3–5; Cordia timber Year 10–20+.

Key interactions:

• Shade↔yield: non-monotonic — yield rises with moderate shade, falls under excessive shade. Cultivar-dependent (Koutouleas 2022). Ghana: yields ~double when canopy cover increases 0→30% (Asare 2019). But excessive shade lowers flower initiation and yield.
• Shade↔CLR: CONTESTED — Chloroleucon eurycyclum increases CLR (reduces spore wash-off); other shade species support natural enemies and reduce brown eye spot/mealybug (Staver 2001). Match shade species to local CLR context.
• Shade↔pollinator: bee richness increases with canopy cover and tree diversity; shaded coffee has higher fruit set. Cacao pollinated by midges needing leaf litter + humid shade.
• N-transfer: Inga/Gliricidia/Erythrina fix N and transfer to coffee (Snoeck 2000 isotopic evidence). Caveat: young Inga fixes little (Grossman 2006); pruning reduces nodulation.
• Pruning: the management lever that tunes the shade–yield balance in real time (Esche 2022). Prune, don't remove shade; high shade for seedlings, reduced shade during cultivation, high shade again for rejuvenation.^{320:contr124:rel125:rel321:rel322:supp}

Circular links:

• Consumes from: Cell 1 (N, mycorrhizae); Cell 7 (pollinators); Cell 8 (pest-balance habitat).
• Contributes to: coffee husk/pulp → Cell 6 (mushroom substrate); cocoa pod husk → Cell 6; prunings → mulch stream; shade microclimate → understory.

Caveats: Hemileia vastatrix (CLR) causes serious defoliation in BOTH shaded and unshaded conditions — shade is not a CLR solution.^{323:supp324:supp59:supp} Push-pull not directly applicable to cacao/coffee (designed for maize stemborer/Striga). Farmer perception (Ghana: "shade reduces yield") can diverge from peer-reviewed evidence — both are evidence streams.^{325:rel326:supp327:supp328:rel329:rel}

---

Cell 5 — Medicine Cell

Purpose: Medicinal plant production integrated into homegarden polyculture.

Species by stratum:

• Small tree (3–10 m): Moringa oleifera
• Shrub/understory: Noni (Morinda citrifolia), kava (Piper methysticum)
• Groundcover: Turmeric, ginger, garlic, regional medicinal flora (Tsimane' Amazonia: Para benensis, Urera lacinata, Citrus limon, Prockia crucis)^{330:supp331:supp332:supp333:rel334:rel}

Mycorrhizal note: AMF hosts. Compatible with homegarden polyculture.

Successional phase: Moringa and noni establish Year 1; kava is a 3–5 year understory perennial.^{333:rel275:rel278:supp}

Key interactions:

• Socially structured: The Tsimane' study (Díaz-Reviriego 2016) shows medicinal-plant richness is shaped by kinship and gender relations — a medicinal cell is a network of exchange among households, not just a planting.

Circular links:

• Consumes from: Cell 1 (soil fertility); Cell 2 (shade for understory medicinals).
• Contributes to: medicinal harvest; flowering (noni, citrus) supports pollinators (Cell 7).^{335:supp336:supp330:supp334:rel337:rel}

Caveats: Moringa is widely promoted as a "miracle tree" — the promotional framing warrants skepticism; pharmacology data is denser than field-scale agroforestry performance data. Kava requires shade and 3–5 years — not a quick cell.

---

Cell 6 — Mushroom Cell (the circular consumer)

Purpose: Convert the lignocellulosic byproduct stream of every other cell into food and medicine.^{338:rel339:rel340:supp341:rel342:rel}

Species (by substrate):

• Pleurotus spp. (ostreatus, djamor, pulmonarius, cystidiosus, tuber-regium) — coffee husk/pulp, cocoa pod, banana leaf, sugarcane bagasse, cassava peel, sawdust
• Volvariella volvacea (paddy straw mushroom) — rice straw, banana leaf, the tropical classic
• Lentinula edodes (shiitake, tropical strains) — pruned hardwood logs (totems)
• Auricularia (wood ear) — sawdust, rubberwood
• Ganoderma (reishi) — natural logs (medicinal)
• Schizophyllum commune — wood sawdusts
• Termitomyces — NOT independently cultivable; tied to termite symbiosis (see Cell 10)

Mycorrhizal note: Mushrooms are saprotrophs or ECM (some), not AMF — mycorrhizal compatibility does not apply to the mushroom crop itself, but the substrate stream comes from AMF-host plants.^343:rel

Successional phase: Can start Year 1 on pioneer biomass (banana, cassava, papaya leaves). Continuous thereafter on pruning stream.^{344:rel345:rel346:rel347:rel348:rel}

Key interactions:

• Substrate = mulch stream: coffee husk (strongest documented bridge — 18 shared facts with Pleurotus), cocoa pod, banana leaf, cassava peel, sugarcane bagasse, pruned logs. This is the most operationalizable circular link in the recipe book.
• Trichoderma tension: the mushroom cell is vulnerable to green mold (Trichoderma aggressivum). But Trichoderma is also a plant biocontrol ally. Spatial/temporal separation required.
• Bacillus helper: Bacillus subtilis controls green mold and promotes mushroom fruiting (Potočnik 2021, Stanojević 2019, Hashem 2019). Consider Bacillus co-inoculation.
• Pest/benefit duality: fungus gnats (Sciaridae) are mushroom pests AND detritivores in the wider ecosystem. Beneficial nematodes (e-nema) can suppress flies. Pseudomonas tolaasii causes bacterial blotch (century-old problem).

Circular links:

• Consumes from: Cell 4 (coffee husk, cocoa pod), Cell 2 (banana leaf), Cell 3 (cassava peel), Cell 9 (pruned logs), Cell 1 (sawdust, sugarcane).
• Contributes to: food/medicine harvest; spent substrate → compost → soil organic matter → back to all cells.^{349:supp350:rel351:supp352:rel}

Caveats: Schizophyllum commune can be parasitic on living wood — monitor. Mushroom viruses (La France disease) are a documented but separate pathogen tier.^{353:supp354:supp355:supp356:rel357:rel}

---

Cell 7 — Pollinator & Insectary Cell

Purpose: Sustain bee, butterfly, moth, fly, and beetle pollinator populations; support fruit set in Cells 2 and 4.^{336:supp358:supp359:supp360:supp361:rel}

Species by stratum:

• Flowering canopy/mid-story: Inga spp. (mass flowering), citrus, mango, banana (male flowers), Erythrina
• Insectary understory: buckwheat, alyssum, clover, Lobularia maritima, native bee-forage species, Tephrosia (pest-repellent legume — with nematode caveat)

Mycorrhizal note: AMF hosts. Flowering benefits from good mycorrhizal status (defense priming supports flower production).^{362:rel363:rel364:rel}

Successional phase: Insectary herbs from Year 0; flowering trees from Year 3–5.

Key interactions:

• Forest fragments enhance pollinator activity in nearby coffee crops. Pollinator diversity increases fruit production in rustic Mexican coffee. Bee richness increases with canopy cover and tree diversity.
• Cacao pollinated by midges (Forcipomyia) — different guild from coffee bees; needs leaf litter + humid shade. Same architectural principle (structural diversity) supports both guilds.
• Trap crops + insectary: predators and parasitoids attracted by nectar/pollen in insectary plantings concentrate where trap crops host pests. Co-design the insectary and trap-crop cells.^{365:supp366:rel360:supp367:rel368:rel}

Circular links:

• Consumes from: Cell 1 (soil fertility); Cell 2 (shade context).
• Contributes to: pollination service → Cells 2, 4 (fruit set); nectar/pollen → predators/parasitoids → Cell 8 (pest balance).

Caveats: Pollinator response is guild-specific — coffee bee results do not generalize uncritically to cacao midges.^{369:rel367:rel370:rel371:rel} Climate seasonality affects pollinator dynamics. Tephrosia pest-repellent BUT builds up root-knot nematodes (Meloidogyne) — the non-black-and-white trap.^{372:supp373:rel}

---

Cell 8 — Pest-Balance / Habitat Cell (the system-balance cell)

Purpose: NOT "kill pests" but "manage the pest/beneficial balance via architecture." This is the cell that operationalizes the user's non-black-and-white framing.^{213:supp374:supp375:supp376:supp377:supp}

Components:

• Flowering understory (overlaps Cell 7)
• Hedgerows and windbreaks
• "Weeds" managed as beneficial hosts (not eliminated)
• Mulch habitat for ground beetles (Carabidae), spiders (Araneae), predatory mites
• Leaf litter for cacao midges (Forcipomyia) and decomposers
• Termite management zones (decomposers, not eradication)

Key interactions (the pest/benefit duality, held in tension):^{378:rel379:rel380:rel381:rel382:rel}

• Leafcutter ants (Atta): pest of Inga/crops near alleys; keystone soil-builders (40 tonnes/nest) and prey hub in restoration zones. Both roles at full strength.
• Termites (Isoptera): structural pests; major decomposers, soil aerators, Termitomyces symbionts, Bacillus hosts. "Managing termites and organic resources to improve soil productivity" — not eradication.
• "Weeds": host pests AND beneficials. Diversification (not weed elimination) suppresses outbreaks. Organic + shade reduces pest pressure in Robusta (Amazonia).
• Mycorrhiza-induced defense: AMF colonization primes plant defenses (Jung 2012 MIR) — managing for high AMF (low disturbance) is a pest-management intervention.
• Entomopathogenic fungi: Beauveria bassiana, Metarhizium anisopliae — fungi deployed against insect pests (the mushroom layer produces pest-management organisms).
• Push-pull (in maize systems): Desmodium repels stemborers + inhibits Striga; Napier grass attracts. Mechanism-validated. Economic claims narrower base.

Circular links:

• Consumes from: Cell 1 (mulch habitat); Cell 7 (nectar/pollen for predators).
• Contributes to: pest-balance service to all crop cells; decomposition (termites, ants, fungi).^{383:supp384:supp385:supp386:supp213:supp}

Caveats: "Diversification suppresses outbreaks" is directionally supported across multiple literatures but not tested against a counter-hypothesis in any single source — a strong directional lean, not a closed question. Push-pull economic claims (ROI >2.2) rest on a narrow evidentiary base.^{387:rel388:rel18:rel389:rel390:rel}

---

Cell 9 — Timber & Fuelwood Cell

Purpose: Long-cycle timber and fuelwood integrated with food crops.

Species:

• Timber: Cordia alliodora (Central America cacao intercrop), Bertholletia excelsa (Brazil nut — syntropic climax), native cabruca emergent timber
• Fuelwood: Inga spp. (pollarded, 20–24 month cycle), Gliricidia, Leucaena (with caveat)^{310:supp272:supp391:rel277:supp392:supp}

Mycorrhizal note: Cordia, Bertholletia are AMF. Eucalyptus is dual AMF/ECM (AMF seedling → ECM mature) and creates strong plant-soil feedbacks — complex in polyculture.^{393:supp394:supp395:supp396:rel397:supp} Dipterocarpaceae (SE Asia) are ECM — do not mix with AMF crop understory without explicit design.

Successional phase: Inga fuelwood 20–24 month pollarding; Cordia/Bertholletia 10–20+ year timber.^{398:rel399:supp400:rel270:rel401:rel}

Key interactions:

• Somarriba 2014: Theobroma cacao–Cordia alliodora timber system (Central America). Glover 1981: coffee + Erythrina + Cordia factorial.
• Inga fuelwood pollarding produces mulch (bridges to Cell 1) AND substrate (bridges to Cell 6).

Circular links:

• Consumes from: Cell 1 (N capital for Inga/Gliricidia fuelwood).
• Contributes to: timber harvest; fuelwood; pruned logs → Cell 6 (mushroom substrate); mulch → Cell 1.^{400:rel402:supp399:supp310:supp2:rel}

Caveats: Eucalyptus allelopathy suppresses legume-rhizobium symbiosis (Liu 2019) — a strong design negative for eucalyptus + N-fixer combinations. Teak allelopathy (Kato-Noguchi 2021) constrains teak intercropping.^{403:rel404:rel168:supp405:rel406:supp} Cabruca timber value is landscape-contextual (depends on forest remnants), not autonomous.

---

Cell 10 — The Termite–Termitomyces–Bacillus Cell (experimental)^{407:rel90:supp408:rel87:supp409:rel}

Purpose: The most innovative and least-tested cell. Explores whether the termite–Termitomyces–Bacillus symbiosis can be partially replicated or designed into a farming system to co-produce food mushrooms, biocontrol bacteria, and pest suppression.

Organisms:

• Macrotermes natalensis (or local fungus-growing termite — Macrotermitinae)
• Termitomyces (prized edible — not independently cultivable; tied to termite colony)
• Bacillus (bacillaene-producing strains — suppress Trichoderma and Pseudoxylaria without harming Termitomyces; also promotes mushroom fruiting and plant biocontrol)^{410:supp87:supp90:supp411:supp412:supp}

Key interaction (the densest single-fact bridge):

• Um et al. 2013 (Scientific Reports): Macrotermes natalensis harbors Bacillus strains that inhibit antagonistic fungi without harming the mutualistic Termitomyces. One fact, five organisms: termite (pest) + Termitomyces (food) + Bacillus (helper) + Trichoderma (competitor/ally) + Pseudoxylaria (competitor).^{87:supp413:supp414:rel415:supp90:supp}

Circular links:

• Consumes from: lignocellulose (from mulch stream of all cells).
• Contributes to: Termitomyces edible harvest; Bacillus biocontrol (if harvestable); decomposition; soil aeration (termite mound engineering).

Caveats — FLAG AS RESEARCH TARGET, NOT PROVEN RECIPE:

• No source in the evidence base tests whether this symbiosis can be deliberately designed into a farming system rather than only observed in termite mounds.
• Termitomyces is NOT independently cultivable — fruiting is tied to termite colony activity.
• The termite is simultaneously a structural pest. Removing the pest removes the mushroom.
• This is the single most promising direction for further research that this synthesis surfaces.

---

Part 4 — System Assembly: How the Cells Combine

4.1 Successional assembly

Year	Cells active	Logic
0–2	Cell 1 (soil-building) + Cell 3 (pioneer cassava/sweet potato) + Cell 6 (mushroom on early banana/cassava biomass) + Cell 7 (insectary herbs) + Cell 8 (habitat establishment)	Pioneers establish microclimate, N capital, mycorrhizae, mulch. Mushroom cell begins consuming early biomass.
2–5	+ Cell 2 (banana, papaya, breadfruit begin) + Cell 4 (Inga shade + cacao/coffee planting) + Cell 5 (moringa, noni) + Cell 9 (Inga fuelwood pollarding begins)	Enrichment phase. Inga/Gliricidia shade establishes. Cacao/coffee planted under shade.
5–10	+ Cell 4 (cacao/coffee yielding) + Cell 2 (breadfruit, citrus yielding) + Cell 5 (kava mature)	Mid-succession fruit/cash crops yielding.
10–20	+ Cell 9 (Cordia, Brazil nut timber) + Cell 2 (coconut, Brazil nut climax)	Climax timber and long-cycle fruits.

This matches the syntropic succession diagram (Stage 1 pioneers → Stage 2 secondary → Stage 3 mid → Stage 4 climax) and the traditional breadfruit staged canopy — a convergence across conventional, syntropic, and traditional knowledge traditions that no single source states directly but the evidence supports.

4.2 Spatial assembly

• Mushroom cell (Cell 6) adjacent to coffee/cocoa cell (Cell 4) — so coffee husk/cocoa pod flows directly as substrate.
• Pollinator cell (Cell 7) adjacent to fruit cell (Cell 2) — for fruit set pollination.
• Pest-balance cell (Cell 8) as understory throughout — flowering understory, hedgerows, windbreaks woven through all crop cells, not a separate block.
• Timber cell (Cell 9) as emergent layer over cacao/coffee — cabruca/Cordia pattern.
• Termite–Termitomyces cell (Cell 10) as an experimental module — not co-located with wooden structures (termite pest risk).^{241:rel24:supp240:rel418:rel416:rel}

4.3 The circular economy in operation

Cell 1 (legume prunings) ──────→ mulch stream
Cell 4 (coffee husk, cocoa pod) → Cell 6 (mushroom substrate)
Cell 2 (banana leaf) ──────────→ Cell 6 (mushroom substrate)
Cell 3 (cassava peel) ──────────→ Cell 6 (mushroom substrate)
Cell 9 (pruned logs) ───────────→ Cell 6 (mushroom logs)
                                    ↓
                              mushroom harvest
                                    ↓
                              spent substrate → compost → soil
                                    ↓
                              mycorrhizae & soil food web
                                    ↓
                              all plant cells (nutrient uptake)
```[349:supp][352:rel][419:supp][351:supp]

The system closes its loops. The only external inputs are sunlight, rain, and initial inoculum/seed. The mulch stream feeds the mushroom cell; the mushroom cell produces food and returns compost; the compost feeds the mycorrhizae; the mycorrhizae feed the plants; the plants produce the mulch.[349:supp][420:supp][421:supp][422:supp][52:supp] This is the design ideal; achieving it requires the cells to be co-located and the management (pruning, inoculation, pest-balance) to be synchronized.

---

## Part 5 — Honest Assessment

### 5.1 What the evidence strongly supports (mechanism-validated, multi-site)[423:rel][424:rel][425:rel][426:rel]

- **Mushroom substrates from agroforestry byproducts** — coffee husk/pulp, cocoa pod, banana leaf, cassava peel, sugarcane bagasse, pruned logs are all documented substrates across many independent studies and regions. This is the strongest circular link in the recipe book.
- **Trichoderma as both mushroom competitor and plant biocontrol** — supported by separate but robust literatures. The dual-role is not a hypothesis; it is a documented fact.
- **Pseudomonas tolaasii as bacterial blotch agent** — century-old, multiply confirmed.
- **Bacillus promotion of mushroom fruiting and suppression of competitors** — Potočnik 2021, Stanojević 2019, Um 2013, Hashem 2019 (1189 cites).
- **Push-pull mechanism** (kairomones, Desmodium Striga-inhibition) — consistent across sources.
- **Dipterocarpaceae ECM symbiosis** — well-established tropical ECM ecology.
- **Forest fragments enhancing coffee pollination** — multiple studies.
- **Legume N-transfer to non-legumes** — isotopic evidence (Snoeck 2000); mechanism via CMN and rhizosphere pathways.
- **The lignin:N mulch-quality axis** (Melillo 1982, Palm 1990) — foundational decomposition model.
- **Faidherbia albida reverse phenology** — unique and well-documented.
- **Mycorrhiza-induced resistance** (Jung 2012, 1313 cites) — applicable to woody plants.

### 5.2 What relies primarily on institutional authority or single-community repetition

- **Push-pull economic claims** (ROI >2.2, 61.9% yield increase) — attributed to an unnamed study via Wikipedia; convergence is within the Rothamsted/ICIPE community.
- **CMN transfer magnitude** — the "FOR" evidence (Simard school) and the "AGAINST" evidence (Karst/Robinson) are each internally cohesive but draw from different research communities; the tropical-application gap means neither directly applies to tropical multistrata without inference.
- **Syntropic agriculture** — practitioner/grey literature dominated; no independent peer-reviewed replicated trials at matched sites in the evidence base. This is a literature-availability asymmetry, not a verdict.
- **Faidherbia/Evergreen Agriculture scaling claims** (avoided emissions, foreign-exchange savings) — ICRAF institutional framing with policy-relevant metrics.[427:rel][428:rel]

### 5.3 What is widely repeated but under-tested

- **"Diversification suppresses outbreaks" as a universal principle** — directionally supported across coffee, maize, cocoa, and weed-beneficial literatures, but not tested against a counter-hypothesis in any single source. Strong directional lean, not a closed question.
- **The termite–Termitomyces–Bacillus cell as an operationalizable recipe** — the bridge fact (Um 2013) is real and well-sourced (Scientific Reports), but no source tests whether this can be deliberately designed into a farming system. This is the single most promising research direction.
- **Hydraulic lift delivering meaningful water subsidy to agroforestry companions at field scale** — HL is real; the CMN-mediated transfer and the agronomic benefit are both contested/unproven. Treat as upside, not load-bearing.
- **Companion-planting repellence generalized beyond documented mechanisms** — the repository contains both positive and skeptical evidence; the skeptical evidence explicitly notes many claims are unreplicated.[177:contr][429:rel][175:supp][210:rel][430:rel]

### 5.4 Open questions the evidence does not resolve

- What spatial/temporal separation lets a system use Trichoderma for plant biocontrol without contaminating the mushroom cell?
- Can the termite–Termitomyces–Bacillus symbiosis be partially replicated or designed into a farming system without retaining termites as structural pests?
- Which companion-plant repellence claims hold in which tropical systems, and which are context-dependent?
- Does CMN-mediated resource transfer operate in tropical multistrata systems at agronomically meaningful scales?
- Can the "fertilizer tree" framing extend beyond Faidherbia (with its unique reverse phenology) to other legumes without over-claiming?
- What is the actual field-scale N contribution of a young Inga/Gliricidia shade system, accounting for the pruning–nodulation trade-off?[431:rel][432:rel][433:rel][434:rel]

### 5.5 The literature-availability asymmetry as a structural finding

Conventional agroforestry has dense peer-reviewed coverage but little practitioner narrative.[435:rel][436:rel][437:rel][107:rel][438:rel] Syntropic/regenerative has rich practitioner narrative but sparse peer-reviewed replication. Traditional/indigenous systems have rich ethnobotanical documentation but less yield-quantified evidence.[237:rel][439:rel][440:rel] **This asymmetry shapes what can be confidently claimed, and it is a property of the literature, not a verdict on the systems.** A recipe book that privileges only peer-reviewed evidence would systematically exclude syntropic and traditional knowledge; one that privileges only practitioner experience would systematically exclude the null results and caveats. This document treats all three as evidence, evaluated on the same terms, with symmetric skepticism applied to institutional (ICRAF), ideological (regenerative), commercial (inoculant companies), and cultural (traditional knowledge) incentives alike.[441:rel][442:rel][443:rel][444:rel][445:rel]

---

## Closing

This recipe book is a set of design hypotheses assembled from ~1,300 sources and 100,000+ facts. It is not prescriptive — every cell carries caveats and open questions. The user asked for modular groups of organisms that work together, with attention to roots, light/shade, insects, mulch, nutrients, mushrooms, helpers/antagonists, and the non-black-and-white nature of pests. The ten cells, the circular mulch economy, the cross-scope bridges, and the parallel-scenario tensions are the structural answer to that request. The strongest parts (mushroom substrates, Trichoderma dual-role, push-pull mechanism, lignin:N mulch quality, Faidherbia reverse phenology, mycorrhiza-induced defense) rest on robust evidence; the most innovative part (the termite–Termitomyces–Bacillus cell) is a research target.[446:rel][447:supp][409:rel][448:rel][449:rel] The system is modular but coupled — the cells share the mulch, the mycorrhizae, the pollinators, and the pest balance, and they must be co-designed to function.

---

*End of the Modular Tropical Agroforestry Recipe Book.*