Polyculture Pairings: What the Three Sisters Got Right

The Iroquois three sisters — corn, beans and squash grown together in a single mound — has become the most-cited example of polyculture in popular gardening writing. The example is genuinely good, but it has been so romanticised that the broader principle gets misunderstood. Three sisters works for specific botanical reasons: corn provides physical structure for the climbing beans, beans fix nitrogen that the heavy-feeding corn benefits from, and squash leaves shade the soil and suppress weeds. The combination is not magical agricultural folklore. It is an applied ecology lesson, and the same principles can be extended to many other plant combinations once you understand what you are doing.

This piece walks through which polyculture pairings actually work in temperate climate gardens, what the research evidence supports, and which traditional companion-planting claims deserve scepticism. The aim is to be specific about what is and is not supported by either traditional practice or contemporary research.

What polyculture actually does

Polyculture, the practice of growing two or more crops together rather than in single-species blocks, can produce several documented benefits when the species are chosen well. The most consistently demonstrated effects are: reduced pest pressure (because pests adapted to one species cannot easily find their host plants in mixed plantings), improved nutrient cycling (when nitrogen-fixing legumes are paired with heavy nitrogen feeders), more efficient use of space (when species occupy different layers, like deep-rooted and shallow-rooted plants together), and yield stability across seasons (because conditions that disadvantage one crop may benefit another).

The 2020 meta-analysis by Iverson and colleagues, published in Frontiers in Plant Science, reviewed 552 polyculture studies and found that well-designed polycultures averaged approximately 4 to 8 percent higher total productivity than equivalent monocultures, with substantial variation by specific combination. The strongest effects were in legume-cereal pairings and in three-or-more-species combinations.

What polycultures cannot reliably do is produce dramatic pest suppression of the kind sometimes claimed in popular companion-planting books. Most « secret » pest-deterrent claims (carrots and tomatoes, marigolds protecting everything, basil repelling all aphids) have weak research support. The benefits are real but more modest and more general than the marketing suggests.

The three sisters in detail

The Iroquois three sisters polyculture, traditional across many indigenous North American cultures, combines maize, beans (typically pole beans or runner beans) and squash. The mechanism is well-understood:

Maize grows tall and provides physical support for the climbing beans, eliminating the need for trellises or stakes.
Beans fix atmospheric nitrogen through symbiotic Rhizobium bacteria in their root nodules. Some of the fixed nitrogen becomes available to neighbouring plants, particularly through root exudates and decaying root material.
Squash sprawls across the ground, with large leaves that shade the soil, suppress weeds, retain soil moisture, and provide some pest deterrence through prickly leaf surfaces.

For home gardens, the three sisters works best in a relatively large planting (at least 3 by 3 metres) because the corn needs sufficient population density for proper pollination. In smaller spaces, modified versions using runner beans on a tripod trellis, with squash at the base, capture some of the principles without requiring corn.

European polyculture traditions

Europe has its own polyculture traditions, less famous than the three sisters but in some cases longer-running. The medieval European cottage garden routinely combined vegetables, herbs, fruit trees and flowers in mixed beds rather than monoculture rows. The French jardin de curé tradition, the British cottage garden, and the central European Bauerngarten all developed sophisticated mixed-planting practices over centuries.

The brassica-allium pairing is one of the most reliable European polyculture combinations. Onions, leeks, garlic, chives or shallots planted between or around brassicas (cabbage, kale, broccoli, cauliflower) provide measurable suppression of cabbage root fly and several other brassica pests. The mechanism is partly olfactory (the sulfur compounds in alliums interfere with pest host-finding) and partly through allelopathic effects on certain soil-borne pathogens. Multiple studies, including a 2014 paper in Crop Protection, support measurable benefit.

The carrot-leek pairing is another well-established European combination. Leeks deter carrot fly; carrots may deter leek moth. The pairing is documented in horticultural literature going back to nineteenth-century French gardening manuals.

Tomatoes and basil are perhaps the most romanticised pairing in popular gardening writing. The pest-deterrent claim is weakly supported in research, but the spatial pairing is genuinely useful: basil and tomatoes thrive in the same conditions, basil’s lower habit fills space at the base of staked tomatoes, and the harvests are usefully synchronised in the kitchen. The benefit is structural and culinary rather than primarily protective.

Pairings that probably do not work

Several traditional companion-planting claims have weak or contradictory research support. It is worth flagging them explicitly to set realistic expectations.

Marigolds protecting everything is partially correct and frequently overstated. Tagetes patula and Tagetes minuta produce thiophenes from their roots that are demonstrated to suppress some plant-parasitic nematodes, particularly Meloidogyne species. The effect on above-ground insect pests is much smaller and less consistent than commonly claimed. Marigolds are useful in nematode-troubled beds; they are not a general-purpose pest deterrent for unrelated insects.

Beans and onions is the most-cited « incompatible » pairing in companion-planting literature. The supposed antagonism is not strongly supported in research. The pairing produces neither obvious benefits nor obvious harms in most studies.

Mint repelling pests is widely repeated and weakly supported. Mint plants do produce volatile compounds with some insect-deterrent properties, but the field-scale effect is modest, and the practical problem is that mint is invasive enough that growing it widely creates more problems than it solves.

A raised wooden vegetable bed in bright sunlight with tomato plants staked vertically and basil plants growing densely at the base of the tomatoes, with mulch covering the soil around the plant bases. — The tomato-basil pairing works primarily through complementary growth habits and culinary synergy.

The structural polyculture: vertical layering

Beyond specific pest-management pairings, the most productive form of polyculture is structural — combining plants that occupy different vertical layers and rooting depths so that a single garden bed produces multiple crops simultaneously.

A typical structural polyculture might combine: tall vertical plants (pole beans, peas on trellises, sunflowers), medium upright plants (tomatoes, peppers, fennel), low spreading plants (squash, cucumbers, sprawling herbs), low upright plants (lettuce, spinach, basil), and root crops (carrots, beets, garlic) in the soil layer below. A bed of perhaps four square metres can productively support eight to twelve different crop types simultaneously when designed this way, where the same area in monoculture would produce only one or two.

The structural approach was substantially developed by the British gardening writer Charles Dowding through his « no-dig » intensive vegetable gardens, and earlier by the French jardin maraîcher tradition that influenced biointensive horticulture in the United States via Alan Chadwick and John Jeavons.

Practical pairings worth trying

For home gardeners interested in polyculture without becoming theoretical about it, the following combinations have reliably worked across the gardens I have visited and managed:

Carrots and leeks in a 1:1 alternation along rows. Mutually pest-suppressing.
Cabbage with onions or garlic at the bed perimeter. Suppresses brassica pests.
Lettuce undergrowing tomatoes in early summer. Lettuce uses the partial shade tomatoes provide as the season heats up.
Beans climbing maize (modified three sisters) with squash at the base. Classical and effective.
Beetroot and onions. The two share similar growing conditions and harvest windows.
Strawberries and borage. Borage flowers attract pollinators that improve strawberry fruit set, and the plants do not compete strongly.
Asparagus and parsley. Parsley occupies the bare soil between asparagus crowns in early summer.
Brassicas with phacelia or buckwheat as flowering interplants. Improves natural enemy populations of brassica pests.

Pest suppression: the broader picture

The most consistent finding across polyculture research is that mixed plantings reduce pest pressure not primarily through specific repellent pairings, but through what ecologists call the « associational resistance » effect: pests find single host plants more easily in monoculture rows than scattered through mixed plantings. The effect is general and applies across many species combinations.

The practical implication is that almost any well-designed mixed planting will have lower pest pressure than the equivalent monoculture, regardless of whether specific magical companion plants are included. The pest suppression is a feature of the polyculture as a whole, not of specific pairings within it.

Limitations of polyculture

Polyculture is not universally superior. It complicates harvest (since multiple species mature at different times), can complicate watering (when plants in the same bed have different water needs), and can produce competition problems when species are matched poorly (root crops and brassicas, for example, can compete for the same soil layer).

The largest limitation, in commercial agriculture, is that polyculture is harder to mechanise. Home gardeners do not face this problem at scale, but it explains why polyculture has remained marginal in commercial vegetable production despite its agronomic advantages.

A worked example: a polyculture bed across a season

To make the principles concrete, consider a 4-square-metre bed planted as a structured polyculture across a temperate growing season. The bed runs north-south to allow even sun exposure across all rows. Planting begins in late March and continues through early June.

In late March, the bed receives early peas (climbing on a 1.8-metre trellis along the north edge), early carrots in two rows, and a perimeter of garlic and overwintered onions. By mid-May, the peas are producing, the carrots are visible, and lettuce is sown between the carrots in a relay planting. By late May, runner beans replace the early peas on the trellis (after pea harvest concludes), with squash planted at the base of the trellis. By June, tomatoes are planted in the central row with basil at the base, and a perimeter row of dwarf French beans replaces the spent garlic.

The bed in this configuration produces, across a typical year: 3 to 5 kilograms of peas, 4 to 6 kilograms of carrots, two flushes of lettuce, garlic and onions, 6 to 10 kilograms of runner beans, 8 to 15 kilograms of squash, 8 to 12 kilograms of tomatoes, French beans, and basil. The total food output runs 30 to 50 kilograms from a 4-square-metre bed across the season — substantially more than the same area in any single monoculture. Total time investment is roughly 25 to 40 hours across the year, including all planting, watering, weeding and harvest.

Polyculture and pollinator support

Beyond direct food production benefits, polyculture beds support pollinator populations more effectively than monocultures. The diversity of flowering times in a mixed planting means that nectar and pollen are available across a longer window than in any single-species planting. Phacelia, borage, calendula, comfrey and most herb flowers can be added to vegetable polycultures specifically to attract pollinators, with measurable improvement in pollinator abundance documented in several studies including a 2019 paper from Reading University.

The benefit feeds back into vegetable production. Squash, beans, tomatoes and many other vegetables produce better fruit set when pollinator populations are abundant, which means that flower-rich polyculture beds typically produce higher per-plant yields than equivalent vegetable-only plantings. The improvement is modest in good pollinator years and significant in years when wild pollinator populations are stressed.

Misconceptions about companion planting

Several persistent misconceptions about companion planting deserve correction. The first is that companion planting is a fully systematised practice with reliable rules. It is not. Most companion-planting books contain a mixture of well-supported recommendations, traditional practices with weak evidentiary support, and outright folklore. The 2014 review by Parker and colleagues in Annual Review of Phytopathology found that approximately one-third of commonly cited companion-planting recommendations had moderate to strong research support, one-third had weak support, and one-third had no published support beyond traditional practice.

The second misconception is that companion planting is primarily about pest deterrence. The strongest research effects of polyculture come from pest dilution, structural complementarity, and pollinator support, not from specific pest-repellent pairings. The popular framing of « this plant deters that pest » is largely an oversimplification of more general ecological dynamics.

The third misconception is that polyculture works equally well in all garden contexts. It does not. Very small gardens (under 5 square metres) often produce more total food in carefully chosen monocultures because mixed plantings sacrifice efficiency at small scales. Large gardens (over 50 square metres) gain proportionally more from polyculture design. The middle range, 10 to 30 square metres, is where polyculture design typically pays off most clearly.

The fourth is that polyculture means random mixing. The most productive polycultures are carefully designed structures, not random mixtures. The classical three sisters, the European brassica-allium pairing, and the structural vertical layering described above all involve deliberate spacing, timing and structural relationships that random mixing would not capture.

Comparative analysis: polyculture across climate zones

Polyculture design varies meaningfully across climate zones. In Mediterranean climates with hot dry summers, polyculture beds emphasise water-conservation pairings: deep-rooted plants alongside shallow-rooted ones to use different soil moisture layers, and ground-cover plants to reduce evaporation. The Italian orto tradition combines tomatoes with squash and basil in patterns that work because the squash leaves shade the soil during the hottest hours.

In temperate maritime climates including the UK, northern France and the Low Countries, polyculture emphasises light optimisation: vertical layering to capture diffuse light efficiently, and timing to ensure that plants reach peak growth at different points in the season. The British cottage garden tradition is largely an applied polyculture system optimised for these conditions.

In continental climates with hot summers and cold winters, polyculture design emphasises seasonal succession: spring greens replaced by summer fruiting crops replaced by autumn brassicas, often within the same bed across the year. The German Bauerngarten tradition formalised this approach with a four-bed rotation system that has been continuously documented since the medieval period.

In tropical and subtropical climates, polyculture design typically incorporates perennial trees and shrubs alongside annual vegetables, in the agroforestry-style designs popularised by Bill Mollison and David Holmgren in the original Australian permaculture work. Most of the principles transfer to temperate gardens but require adjustment for the seasonal differences.

Tracking polyculture results across seasons

For gardeners wanting to learn from their own polyculture experiments, simple record-keeping produces substantial knowledge over multi-season periods. A garden notebook recording planting dates, specific cultivars, polyculture combinations, observed pest pressure, harvest yields and notable problems builds a personal database that becomes increasingly valuable as it accumulates years of data. The most experienced polyculture gardeners typically maintain such records across decades and use them to refine their planting decisions.

Several digital tools can supplement traditional notebook records. The free Garden Plan Pro app and the more comprehensive Smart Gardener platform allow detailed planting and rotation records with bed-by-bed tracking. The Permaculture Plot mapping system supports more elaborate polyculture design with documentation. None of these tools is essential, but they can supplement notebook records for gardeners who prefer digital tracking.

The single most useful piece of information to record is what worked unexpectedly well or unexpectedly badly. Standard expected outcomes (the carrots grew, the tomatoes produced) require little documentation. Surprising failures (a normally reliable variety failing for unexplained reasons) and surprising successes (a tentative pairing producing exceptional results) provide the data that genuinely improves future planning. Across five to ten years of records, patterns emerge that no general gardening literature could provide for a specific garden’s specific conditions.