The permaculture apple-tree guild looks the same on every blog. Some of it is real, some has nothing behind it - here's the whole thing graded, claim by claim.
Plan a fruit tree guild at your place →
What you need
Search for a fruit tree guild and you'll find the same picture over and over: an apple tree in the middle, a ring of comfrey, a ring of daffodils, some garlic, a nitrogen-fixing shrub, and a set of confident labels saying what each one does for the tree. It's a genuinely appealing idea - a small permanent community of plants that looks after itself.
The picture isn't wrong, exactly. But the labels are doing very different amounts of work. Some name a mechanism with real evidence behind it. One names a plant that will escape your garden and is banned in several states. And at least one names an effect that, as far as anybody has measured, doesn't exist.
This guide takes the whole thing apart and says how much is actually known about each piece, because that's the only honest way to hand someone a design that's meant to last forty years.
Almost every guild article starts with the plant list. The plant list is the least important part. These two come first, they're both as well established as anything on this site, and getting either wrong is expensive in a way no plant choice is.
Mechanism: Tillage destroys the crown. Rotation moves the annuals; the perennial cannot move.
Evidence: settled without a trial - mechanics.
Split perennial members into a permanent border; rotate the annuals around it.
This is the one that catches people who are already gardening well. You've got beds, you rotate them properly, and a guild sounds like a nice thing to add to one. It isn't - it's a contradiction. A rotation works by moving crops; a perennial works by staying put and building a crown. Put the two in the same ground and each season's digging destroys the thing that needed a decade. The fix isn't clever, it's just deliberate: perennials go in permanent ground, and the annuals rotate around them.
Mechanism: A tree's root zone expands for a decade and extends well past the canopy edge.
Evidence: settled without a trial - root architecture.
And this is why the tidy circular diagram misleads. A guild isn't a bed with a tree in it. Its real extent is the drip line and beyond, it moves outward every year for a decade, and it doesn't have corners. The planner treats it that way - not as a rectangle you fill, but as rings around a centre:
Plan the space the tree will need in ten years, not the space the whip needs now.
The only genuinely non-negotiable member, and the one real gate on it is climate: your hardiness zone decides which fruit tree survives the winter where you live, which is the one job zone is actually for. Apple is the classic choice here; pear, plum, peach and sour cherry all fill the role, and the colder your zone the shorter that list gets.
Mechanism: Resource competition and physical occupancy of the root collar zone.
Evidence: settled without a trial - competition.
This one is real and it's well established, and the mechanism is refreshingly unmysterious: turfgrass at the trunk competes hard with a young tree, and a dense ring of bulbs physically takes up that ground and grabs those resources first. Nothing exotic is being claimed. Something's in the way of the grass.
The second claim about the bulb ring is weaker, but it isn't nothing:
Mechanism: Narcissus bulbs and tissue contain lycorine and related alkaloids, which are toxic to eat. Rodents avoid them accordingly — field mice rarely touch daffodils even where they readily eat tulips and crocus — so a ring of narcissus tends to be left alone rather than gnawed. It is an avoidance effect, not a barrier: it discourages feeding on the bulbs, not a guarantee against every rodent.
Evidence: confirmed against 3 sources in the research-based literature.
Univ. of Illinois Extension, 'Bulbs', extension.illinois.edu/flowers/bulbs (read 2026-07-11): field mice 'most often attack tulips, crocus and gladioli and rarely eat daffodils, alliums or colchicums.'
Univ. of Illinois Extension, 'Knowing poisonous plants can save a life', extension.illinois.edu/blogs/know-how-know-more/2022-07-05-knowing-poisonous-plants-can-save-life (read 2026-07-11): 'Daffodil, hyacinth, and narcissus contain the toxin lycorine, which, if eaten, typically causes nausea.'
Clemson HGIC, 'Planting Deer Resistant Bulbs', hgic.clemson.edu/planting-deer-resistant-fall-bulbs (read 2026-07-11): daffodils 'are also poisonous, so deer are not tempted to eat them.'
A good hunch - and note what the evidence is. Extension services report that voles and mice largely leave daffodils alone, and daffodils are toxic, so the avoidance is at least plausible as a mechanism and not just an observation. That's a good deal better than most companion claims manage. It's still not a trial showing your tree suffered less rodent damage, which is why it advises rather than promises.
And then there's the third one, which is where the guide stops being polite:
Evidence: none exists for the underlying claim - that absence is the point.
The weakest standing the corpus has, and for the bluntest possible reason: what's recorded against this claim is that no evidence for it exists at all - not weak evidence, not conflicting evidence, but nobody has shown this. Planting garlic around a fruit tree to repel borers is a thing the internet says. If you like garlic, plant garlic; it's a fine plant and it costs you nothing. Just don't plant it instead of watching the trunk for borers, because that's the trade you'd actually be making.
Notice the rule wasn't deleted. A claim that turns out to be empty stays in the record, marked inert, so the next person who meets it on a blog finds out here that it was looked into.
What is actually known: Symphytum has deep roots and produces large potassium-rich biomass that decomposes fast. That is a good mulch plant. The specific mineral-mining claim has very little behind it.
Evidence: confirmed against 1 source in the research-based literature.
Cornell Small Farms Program (Cornell Coop. Extension), 'New Findings Further the Study of Dynamic Accumulators' (Tyler & Zarro, 2022), smallfarms.cornell.edu/2022/04/new-findings-further-the-study-of-dynamic-accumulators (read 2026-07-11): an on-farm study 'has shown Russian comfrey to be a dynamic accumulator of both potassium and silicon' (tissue concentrations above the set thresholds). NOTE: this establishes high nutrient CONCENTRATION in comfrey biomass (supporting the 'potassium-rich biomass / good mulch' point), not the deep-root subsoil-MINING mechanism the rule refutes — accumulation reflects uptake from available soil, not sourcing of otherwise-inaccessible subsoil.
Comfrey makes excellent chop-and-drop mulch. The mineral-mining story is unsupported.
This is the most interesting entry in the guild, because the popular claim and the useful reality have been welded together for so long that pulling them apart sounds like an attack on the plant. It isn't. Comfrey is genuinely excellent here: deep-rooted, enormous fast-growing bulk, potassium-rich tissue, and it breaks down quickly when you cut it and drop it. Chop-and-drop comfrey is good practice, and this guide recommends it.
What has very little behind it is the specific dynamic accumulator story - that comfrey's roots mine minerals from subsoil other plants can't reach and deliver them to the surface. The Cornell work that gets cited for it confirms comfrey tissue is high in potassium and silicon, which is a statement about concentration, not about sourcing. High uptake from soil that already has it available isn't the same as mining a layer nobody else can reach, and the second is the part of the claim doing the persuading.
So: plant the comfrey. Cut it three or four times a season and drop it under the tree. Just know you're adding good mulch, not running a nutrient pump.
This is a claim the planner refuses to enforce — recorded here with what is actually known, rather than dropped.
What is actually known: It fixes nitrogen. It is also bird-dispersed, forms dense thickets, and is listed as invasive or noxious across much of the eastern and midwestern US.
Evidence: confirmed against 2 sources in the research-based literature.
NC State Extension Gardener Plant Toolbox, 'Elaeagnus umbellata', plants.ces.ncsu.edu/plants/elaeagnus-umbellata (read 2026-07-11): 'This plant is an invasive species in North Carolina'; 'It becomes quite competitive even in poor soils by fixing nitrogen in its roots'; 'The fruits are taken by birds and small mammals, which disperse the seeds ... a large plant may produce hundreds or thousands of fruits. Autumn olive threatens native species by out-competing them'; 'Many states have banned its sale.'
USDA Forest Service, Fire Effects Information System (Munger 2003), 'Elaeagnus umbellata', fs.usda.gov/database/feis/plants/shrub/elaumb/all.html (read 2026-07-11): symbiosis with soil actinomycetes 'permits the fixation and subsequent utilization of atmospheric nitrogen'; 'Seeds are dispersed by frugivorous birds'; listed 'as a noxious weed in several West Virginia counties'; USFS Region 8 lists it a 'category 1 weed' whose introduction 'is prohibited on National Forest System Lands.'
Autumn olive will escape your yard. Use white clover, wild lupine, or a native fixer, and check your state noxious weed list before planting any woody nitrogen fixer.
Read that card again: well established, fatal if ignored, and blocked outright. This isn't a "consider alternatives" note. Autumn olive is a real nitrogen fixer and it's recommended for exactly that reason all over the permaculture web - and it's also bird-dispersed, forms dense thickets, and is listed as invasive or noxious across much of the eastern and midwestern US, with several states banning its sale outright. It won't stay in your yard. Birds will take it into everyone else's.
White clover fills this role everywhere. Wild lupine fills it as a native. Both are in the planner as the standard and the alternative fixer, and neither carries this problem.
The wider habit is the point, and it holds beyond this one shrub: check your state's noxious weed list before planting any woody nitrogen fixer. The trait that makes a plant good at colonising poor ground on your behalf is the same trait that makes it good at colonising ground you didn't offer it.
Yes - but for different reasons than the ones usually given.
What's load-bearing is the structure: permanent ground that never gets turned, planned at the size the tree will grow into, with the grass kept off the trunk. Those are well established, the mechanisms are plain, and they're the parts that decide whether the tree thrives.
What's good practice but oversold is the middle tier: comfrey mulch (real, useful, wrong story), daffodils against rodents (plausible, unproven).
And what's folklore is folklore - the borer-repelling garlic, and the specific mineral-mining story - held in the record marked inert so nobody has to re-argue them from scratch.
A guild built on the first tier and honest about the rest is a very good design. A guild built on the third tier is a picture.
Next in What to grow → Why is my squash not setting fruit? Pollination for the home gardener
The rules this guide leans on: