Companion planting — what research shows and what is only tradition

Companion planting — placing plants next to each other so they help one another — is one of the most repeated topics in gardening guides. The trouble is that some of these recommendations rest on solid science, while others are folklore copied from book to book for a century, never tested in a controlled trial. That is why, at Atlas-Flora, we attach a level of evidence to every neighbour relationship: research, observation or tradition. This guide explains what genuinely works and what is just a good-sounding myth.

What plant companionship means

Plants affect one another in several measurable ways. They compete for water, light and nutrients. Some fix atmospheric nitrogen and enrich the soil for their neighbours. Others release compounds into the soil that suppress competitors’ growth — that is allelopathy, a well-documented phenomenon, not folk belief. Insect-attracting plants can draw predators into the bed that eat pests. And species from the same botanical family share diseases and pests, so planting them together raises the risk of an outbreak. These are real mechanisms — the problem starts when a specific “pair” is credited with merits that science does not confirm.

Why we label the level of evidence

Companion planting suffers from a typical problem of folk knowledge: it sounds convincing, so no one checks it. “Basil improves the flavour of tomatoes” is repeated so often that it looks like a fact — yet not a single controlled study stands behind it. Instead of faking a certainty we do not have, we state the basis for each relationship. It is the same principle we apply across the whole atlas, described in detail on our methodology page. The three levels are: research (a controlled experiment, a known mechanism), observation (repeatable gardener experience, but without hard data) and tradition (handed down over generations without scientific proof).

What research confirms

Marigolds and soil nematodes. This is the flagship example of companion planting with genuine scientific backing. The roots of Tagetes marigolds release alpha-terthienyl — a compound toxic to root-knot nematodes of the genus Meloidogyne. Note the nuance: the effect appears only when marigolds grow densely as a cover crop over a whole season on the infested bed, acting like a soil-cleansing green manure. A few plants tucked between the tomatoes will not do the job — which is exactly what most guides advise.

Attracting beneficial insects. Plants with small, flat umbel flowers — like dill — provide easily accessible nectar for hoverflies, ladybirds and small parasitic wasps. Their larvae eat aphids and other pests. This is a documented mechanism of conservation biological control: the plant does not “repel” anything itself; it recruits allies. Flowering calendula works similarly, its orange flowers drawing in pollinators and predators.

Allelopathy. That one plant can chemically suppress another is solidly researched. Allium species — including garlic and onion — suppress the growth of legumes. The walnut tree releases juglone from its roots, highly toxic to many vegetables under its canopy. This works the opposite way to “good” companionship: it tells you what not to plant together.

What is observation, not certainty

The classic pair carrot and onion rests on scent masking: the strong smell of onion is meant to make it harder for the carrot fly to find the carrot, and vice versa. Some field trials show that mixed cropping does reduce infestation — but the effect is often moderate and depends on the ratio of the plants. It is a sensible, worth-trying practice, so we mark it as observation rather than hard research. The scent of garlic is likewise linked to fewer aphids on neighbours — plausible, but more weakly proven in the field than in the lab.

What is more likely a myth

“Basil improves the flavour of tomatoes” is the most repeated companion-planting myth — it sounds lovely, but no controlled taste study backs it. Basil has real merits (it attracts pollinators and pairs with tomato in the kitchen), just not the ones folklore claims. The second common error is confusing calendula with Tagetes marigold and crediting it with a nematode effect — that belongs to Tagetes, not Calendula. And as a general rule: most ready-made “companion charts” circulating online mix a handful of confirmed pairs with dozens of untested ones — treat them as a starting point, not an oracle.

How to use this

The point is not to abandon companion planting — some of its advice genuinely works. The point is to know which. For a practical vegetable-neighbour plan see the guide companion planting for vegetables, and every plant card in Atlas-Flora has a “Companion plants” section listing proven neighbours and cautions — each entry with a visible evidence level, so you can judge for yourself how much to trust it.