A culinary bundle - grown together for the kitchen, with no interspecies claim made.
At least 2.5 m². A bed smaller than that does not get this guild quietly substituted for something else — the planner greys it out, says why, and offers the nearest adaptation that does fit.
If your bed is smaller: Pizza Garden (Patio) (1 m²).
Living perennials stay in the bed between seasons, so the ground is part rotated and part not.
No interspecies mechanism is claimed. It is grouped for a reason, and that reason is written below rather than dressed up as biology.
Rotation history attaches to the ground, not the bed, so this is what the planner remembers about this patch after the guild comes out.
There's no evidence for this and no mechanism anyone has proposed. Plant basil next to tomatoes anyway — you'll want them at the same time in the kitchen.
No proposed mechanism. No trials. Extremely widely believed.
Evidence: settled without a trial - definition.
Mechanism: Tillage destroys the crown. Rotation moves the annuals; the perennial cannot move.
Evidence: settled without a trial - mechanics.
Split perennial members into a permanent border; rotate the annuals around it.
Mechanism: Zones encode average annual minimum winter temperature, which governs whether a woody perennial survives to spring. It is the wrong quantity for an annual, which never sees winter. THE CEILING IS A SECOND AND DIFFERENT MECHANISM, and until 2026-07-26 this rule enforced it without naming it. A zone too WARM does not kill a perennial; it fails to chill it. Deciduous fruit crops need a winter dormancy period to break bud and set fruit, so in a zone milder than a plant is adapted to it may survive and never crop. Until ISSUES #13 was decided the band was enforced at both ends; now (option E) only the cold end gates — it alone follows from the definition of the zone system — and the warm end is carried as an ADVISORY, not a refusal. AND THE WARM END IS A CULTIVAR'S ANSWER, NOT A SPECIES', which is the honest limit of what this rule can do. Chill requirement is published per variety and varies enormously within a species, so a species-level ceiling says "the modelled range does not reach your zone", never "no plant of this kind will crop for you". So the engine offers the plant above its band and advises choosing a cultivar rated for the zone rather than refusing it — the same move invariant 3 makes for soil, applied to chill.
Evidence: confirmed against 3 sources in the research-based literature.
THE FLOOR IS TRUE BY DEFINITION and needs no citation: the USDA system defines a zone AS a band of average annual minimum winter temperature, so 'a plant colder-hardy than your zone survives your winter' is the definition restated. This rule was `verified_trivially` on that basis alone until 2026-07-26. THE CEILING WAS NEVER COVERED BY IT - a minimum-temperature definition says nothing about a plant failing for want of chill - and that gap is why ISSUES #13 exists. The pointer below closes it.
OSU Extension EC 1303, 'Growing blackberries in your home garden' (Bernadine Strik), catalog.extension.oregonstate.edu/sites/catalog/files/project/pdf/ec1303.pdf (read 2026-07-26). THE CEILING'S MECHANISM, STATED: 'Winter cold-hardy cultivars adapted to the colder regions of Oregon (such as zones 5 to 7), may not grow normally in regions of the Willamette Valley (zone 8) where winters are warmer and plants receive less winter chill.' A plant failing in a WARMER zone for want of chill - which is what this rule's upper bound enforces and what no source said until now. THE SAME PAGE MAKES IT A CULTIVAR PROPERTY: 'Even cultivars within the same type (trailing, erect or semierect) differ in fruit quality and flavor, appearance, tolerance to pests, cold hardiness, and plant longevity', and the remedy it gives is the one this rule's rejection text now repeats: 'Minimize risk by choosing types and cultivars that are adapted to your hardiness zone.' The page then publishes a table of sixteen cultivars EACH WITH ITS OWN ZONE RANGE, which is the strongest available evidence that a species-level band is an approximation. [read]
WHAT THIS DOES NOT SETTLE. The mechanism is now named, but the corpus still carries ONE band per species where the source carries one per cultivar, and 12 of 25 perennial zones have no citation at all. So the ceiling is explained without being precise. ISSUES #13 stays open on that, and the rejection text is deliberately phrased as 'the modelled range does not reach your zone' rather than as a prediction that nothing will crop.