Evidence vs. Lore: How We Rate Every Companion-Planting Claim

Search for almost any two vegetables together and you'll find a chart somewhere claiming they're either best friends or mortal enemies in the garden. Most of those charts trace back to the same handful of decades-old gardening books, copied from site to site until nobody can say where the original claim came from or whether anyone ever tested it. We didn't want to build another one of those charts. Every pairing on this site carries a plain label - evidence-backed, mixed, or traditional-lore - and this page explains exactly what each one means and how we decide.

The Honest Starting Point: Most of This Has Never Been Tested

It's worth saying plainly, before anything else: the large majority of companion-planting folklore has never been through a controlled trial. Home garden intercropping is a hard thing to study properly - it requires replicated plots, a monoculture control, consistent conditions, and enough seasons to rule out a fluke year - and most of the pairings repeated in gardening books simply haven't had that kind of scrutiny applied to them. That doesn't mean they're wrong. It means we don't know, and we think a reference site should say so rather than assign every pairing the same confident tone regardless of how much is actually behind it.

The Three Labels

Evidence-Backed

This is our top rating, and we use it sparingly. A pairing earns it one of two ways: either a real intercropping trial measured an outcome - reduced pest damage, higher yield, better germination - comparing the two crops grown together against each grown alone, or the underlying mechanism is basic, settled plant physiology that doesn't need a companion-planting-specific study to confirm. Legume nitrogen fixation is a good example of the second kind: it's textbook botany, not a gardening claim, so a pairing built on it starts from solid ground even before you ask whether anyone tested that exact combination.

Mixed

Most pairings on this site land here, and it's the label that needs the most explaining because it covers a genuinely wide range of situations. Sometimes it means the underlying chemistry is real and documented in a related context, but nobody has run the trial for this specific pair of crops - a scent-masking compound that's been studied against one pest, borrowed here for a different one. Sometimes it means a real mechanism exists but works against a competing consideration, like a legume's nitrogen contribution to a companion that actually does worse in overly rich soil. Mixed isn't a hedge to avoid picking a side - it's the honest answer when a claim is neither confirmed nor baseless.

Traditional Lore

This is where most inherited garden wisdom actually belongs, and we don't treat it as an insult. A pairing gets this label when the reasoning behind it is a general, hard-to-test idea - a strong-smelling herb assumed to confuse pests, a flower assumed to improve a neighbor's flavor - with no controlled comparison behind it that we're aware of, and no settled chemistry pointing toward a specific measurable effect either. Plenty of traditional-lore pairings are still worth trying: they're usually low-risk, and gardening advice that's survived generations of repetition earned that longevity somehow, even if the reason isn't the one usually given for it.

"Untested" Is Not the Same as "False"

This distinction matters enough that it deserves its own section. A traditional-lore rating tells you what we know about the state of the research, not what will happen in your garden. Absence of a trial is not evidence that a pairing fails - it's evidence that nobody with the resources to run a proper trial has bothered to test something that a home gardener can try for the cost of a few seeds and a season. We'd rather tell you honestly that a claim is unconfirmed than dress it up with false confidence, and we'd rather you try a traditional-lore pairing knowing exactly how much weight to put on it than believe it's been proven when it hasn't.

What an Actual Trial Looks Like

When we do cite something as evidence-backed because of an intercropping trial rather than settled physiology, the study design generally follows a similar pattern: two or more crops grown together in a plot, an identical planting of each crop grown alone as a control, matched soil and weather conditions, and a specific measured outcome - pest counts, damage percentage, or yield - compared between the two setups. That's a meaningfully higher bar than an anecdote from one gardener's one season, which is why we reserve the top rating for pairings that have actually cleared it, rather than extending it to every claim that sounds mechanistically plausible.

The Same Mechanism Can Land at Different Ratings

One thing that surprises people browsing this site: two pairings can share the exact same underlying mechanism and still carry different ratings. A scent-based pest deterrent built on genuinely documented chemistry might be evidence-backed for the specific pest-and-crop combination that's actually been trialed, and only traditional-lore when the same chemical family is credited with protecting an unrelated crop nobody has tested it against. We rate the claim, not the chemical - a real compound doing something proven in one context doesn't automatically transfer its evidence rating to every other pairing that borrows the same reasoning.

What This Means When You're Using This Site

Practically, here's how we'd suggest reading the labels: an evidence-backed antagonist pairing is worth taking seriously and planning around, the way you'd treat a real disease risk. An evidence-backed beneficial pairing is worth actively implementing rather than just reading about. A mixed rating means there's a real, defensible reason behind the pairing, even if it isn't proven - reasonable to try, not something to rely on exclusively. And traditional lore means exactly what it says: this idea has been passed down for a long time, it's almost always harmless to test in your own garden, and we don't know whether the reason usually given for it is the real one.