Companion Planting: What Actually Works (And What's Just Folklore)

Companion planting might be the most repeated and least verified topic in gardening. Every gardening book has a chart. Every website has a list. Tomatoes love basil, carrots love onions, never plant fennel near anything. These claims get copied from source to source, decade after decade, often without anyone checking whether the original claim was based on research, observation, or someone's hunch in 1976.

I've spent years trying to separate the signal from the noise. Some companion planting principles are well-supported by research and field observation. Others are pure folklore that's been repeated so many times it feels like fact. Here's what I've found.

The Science Behind Plant Interactions

Plants interact with their neighbours in measurable ways. Understanding the mechanisms helps you evaluate which companion planting claims are credible and which are wishful thinking.

Allelopathy is the production of biochemicals that affect neighbouring plants. Black walnut produces juglone, which inhibits the growth of many species within its root zone. Sunflowers, brassicas, and some grasses also produce allelopathic compounds. This is real, well-documented chemistry, not folklore.

Nitrogen fixation is the process by which legumes (and a few other plant families) convert atmospheric nitrogen into plant-available forms through symbiotic bacteria in their root nodules. This is the most reliably beneficial companion planting interaction. Growing beans near heavy-feeding crops genuinely improves nitrogen availability in the soil, though most of the benefit comes when the legume's roots die back, not while it's growing.

Trap cropping uses sacrificial plants to lure pests away from your main crop. Nasturtiums attract aphids away from brassicas. Blue Hubbard squash draws squash vine borers away from zucchini. The research on trap cropping is solid. It works, but only when the trap crop is more attractive to the pest than your main crop and when you manage the trap crop (remove heavily infested plants before pests move on).

Volatile organic compounds: the scent of certain herbs and alliums can confuse or deter pest insects searching for their host plants by smell. Interplanting onions with carrots can reduce carrot fly damage because the onion scent masks the carrot foliage smell. Timing these interplantings with your frost-date planting schedule ensures both crops are in the ground when it matters. This mechanism has research support, though the effect size varies significantly.

Physical effects are often overlooked. Tall plants shade shorter ones (good or bad depending on what you're growing). Dense plantings suppress weeds. Climbing plants use others as support. Deep-rooted plants break up compacted soil for shallow-rooted neighbours. These mechanical interactions are straightforward and reliable.

Combinations That Hold Up to Scrutiny

These pairings have either research backing, consistent anecdotal evidence from multiple climate zones, or clear mechanistic explanations:

The Three Sisters (Corn + Beans + Squash)

This is probably the most well-documented polyculture in history, practised by Indigenous peoples across the Americas for millennia. The corn provides a climbing structure for the beans. The beans fix nitrogen that feeds the corn (a heavy nitrogen feeder). The squash shades the ground, suppressing weeds and retaining moisture. Each plant solves a problem the others create. It works. The catch: timing and spacing matter enormously. Plant the corn first, then the beans when corn is 15cm tall, then the squash when beans start climbing.

Alliums + Almost Everything

Garlic, onions, chives, and leeks are genuinely useful companions. Their sulphur compounds deter a range of pests including aphids, carrot fly, and Japanese beetles. Interplanting alliums with roses, fruit trees, carrots, and brassicas has consistent positive reports. The main exception: alliums can inhibit the growth of beans and peas.

Legumes as Nitrogen Support

Any legume (clover, beans, peas, lupins, vetch) genuinely improves soil nitrogen for neighbouring plants. The caveat is timing: most nitrogen becomes available when legume roots decompose, so chop-and-drop or turn under the legume before planting your heavy feeder for maximum benefit. Cover cropping with clover between seasons is one of the most evidence-based soil improvement strategies available to home gardeners. When deciding where to place nitrogen-fixing companions, permaculture zone planning helps you match maintenance intensity to the right part of your property.

Aromatic Herbs as Pest Confusion

Basil near tomatoes, dill near brassicas, rosemary near beans. These combinations have moderate evidence. The mechanism (volatile compound masking of host-plant scent) is real. The practical effect depends on planting density and proximity. A single basil plant three metres from your tomatoes will do nothing. Dense interplanting of herbs throughout your beds has a measurable, if modest, pest reduction effect.

For a deeper look at how to build a functional plant community around a fruit tree using these companion principles, see my guide to fruit tree guild design.

Common Claims That Don't Hold Up

"Tomatoes and Basil Grow Better Together"

This is the poster child of companion planting and it's mostly a flavour pairing, not a horticultural one. Some studies show basil's volatile compounds may reduce certain tomato pests (whitefly, aphids), but there's no evidence that either plant grows better near the other. They don't harm each other, they have compatible water and sun needs, and they taste great together. That's enough to plant them near each other, but "they help each other grow" is overstated.

"Marigolds Repel Everything"

French marigolds (Tagetes patula) do produce compounds toxic to root-knot nematodes, and this is well-researched. But the effect requires dense planting of marigolds for a full season before planting your crop, not just tucking a few marigolds between your tomatoes. As a general pest repellent? The evidence is weak. Marigolds attract beneficial insects, which helps indirectly, but "marigolds repel pests" as commonly stated is oversimplified.

Rigid "Never Plant X Near Y" Lists

Most "bad companion" lists are based on nutrient competition (plants with similar root depth competing for the same resources) or allelopathy claims that haven't been replicated. Yes, black walnut inhibits many plants. Yes, fennel produces compounds that can affect nearby species. But "never plant peppers near beans" or "keep cucumbers away from sage". These rigid prohibitions are rarely supported by controlled research.

A Better Approach: Observe and Document

Rather than following a chart, I'd encourage you to think about companion planting as a set of experiments to run in your own garden. The mechanisms are real (nitrogen fixation, pest confusion, allelopathy, physical support) but how they play out depends on your soil, climate, variety selection, and management.

Plant intentionally based on the mechanisms above. Then observe what happens. Do your carrots actually have less pest damage when interplanted with onions? Does the clover under your fruit trees make a visible difference in foliage colour by year three? Are your nasturtiums actually drawing aphids away from the brassicas, or are both getting hammered?

This is what I built PatternBase for. Every plant in the library has a companion planting profile with bidirectional relationships, not just "these two are companions" but what the documented interaction is and which direction the benefit flows. More importantly, the observation system lets you record what's actually happening with your specific combinations in your specific conditions, so you're building your own evidence base over time instead of relying on someone else's chart.

The best companion planting guide is the one you write yourself, one season at a time. You can look up companion profiles with bidirectional relationships for 2,000+ species in the PatternBase plant guide.

---

PatternBase is a free permaculture garden design tool launching March 2026. Create your free account to start documenting what actually grows.