What to plant next to tomatoes, and what is just folklore

Search for "what to plant next to tomatoes" and you will find dozens of companion planting charts, all identical, none sourced. They have circulated for so long that people mistake them for agronomy.

The reality is more interesting and less categorical. Some pairings have a documented mechanism. Others rest on a single old experiment copied for fifty years. And the factor that most affects your tomatoes is almost never their neighbour.

What has a real mechanism

Tagetes (French and African marigolds). This is the best-established case in the vegetable garden. Certain Tagetes species reduce root-knot nematode populations in soil. The mechanism is understood: the plant acts as a trap crop and releases nematicidal compounds. The nuance matters, though — the effect comes mainly from a dense stand of Tagetes occupying the ground before the tomatoes, not from three ornamental plants at the edge of the bed. With no nematode problem, the benefit becomes decorative.

Short-cycle ground cover. Lettuce, rocket and radish at the foot of tomatoes do not "protect" anything. They occupy space while the tomato is still small, which suppresses weeds and slows evaporation. That is a real, measurable gain in land use, not a chemical interaction.

Flowers that feed beneficial insects. Borage, phacelia, calendula and flowering dill draw in hoverflies, ladybirds and parasitoid wasps whose larvae eat aphids. The benefit is not tomato-specific — the whole plot gains — but it is documented and it costs nothing.

What is doubtful despite its reputation

Basil. Tomato and basil is an absolute classic. It works beautifully on a plate. In the garden, the claim that basil repels tomato pests or improves the flavour of the fruit rests on very little: a handful of trials on the repellent effect of volatile compounds, inconsistent results, and no convincing demonstration of any effect on taste. Grow basil because you eat it — an excellent reason. Do not rely on it as protection.

Garlic and onions. They are credited with a fungicidal effect against blight. Garlic extracts do show antifungal activity in the laboratory. A garlic plant thirty centimetres from a tomato does not reproduce those conditions. That said, garlic and tomatoes coexist without competing, so the pairing costs nothing either.

The border of "herbs that repel everything". Most of those claims come from trials on essential oils applied at high concentration. A whole plant in a garden releases nothing comparable.

What is genuinely worth avoiding

Potatoes. This is the strongest exclusion on the list, for a precise reason: tomatoes and potatoes are both Solanaceae and share their diseases, late blight in particular. Planting them side by side, or one after the other in the same spot, builds a bridge for the pathogen. That is not companion planting, it is rotation.

Florence fennel. It is reputed to inhibit the growth of many neighbours. Fennel's allelopathic effect is described, though not precisely quantified in garden conditions. Since it grows perfectly well on its own, the doubt is enough to keep it apart.

Anything that casts shade. Sweetcorn, sunflowers, climbing squash. There is no mystery here: tomatoes need direct light. A neighbour two metres tall on the sunny side of your plants will cost more crop than any pairing can win back.

The factor that matters more

If you take away one thing: spacing, rotation and airflow weigh far more than the choice of neighbour.

Tomato blight develops when foliage stays wet for long periods. Crowded plants, dense foliage, watering over the leaves, a plot sheltered from wind — that is what loses harvests. No companion plant compensates for thirty missing centimetres of spacing.

In the same way, returning Solanaceae to the same bed three years running accumulates inoculum in the soil. Rotation is boring, it has none of the charm of companion charts, and it is markedly more effective.

How Marmanda handles this

Marmanda shows crop pairings with what they usually lack: the observed effect, the mechanism where one is known, the conditions of the trial, the territory it applies to, and an honest assessment of how strong the evidence is. A popular but weakly supported pairing appears as exactly that. When no admissible data exists for a pair, Marmanda leaves the box empty rather than filling it.

Placement, meanwhile, accounts for real spacing, aspect and the space you actually have — the factors that genuinely decide your harvest.

In short

Plant marigolds if you have nematodes. Plant flowers because beneficial insects work for free. Plant basil because it tastes good. Keep potatoes away. And give your tomatoes air: that is the one decision that shows up in the yield.