A polyculture - the species are claimed to help each other, and each claim below names its mechanism.
A derived composition (docs/PROPOSAL-35-combinations.md #28, adopted 2026-08-09), built around a hard structural fact most plans get wrong: tomatillo is self-incompatible and needs two plants to set fruit. The guild exists to make that requirement unmissable.
What binds this planting
The corpus partly backs this planting with 2 graded mechanisms its members share:
At least 3 m²32.3 ft². 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.
What it looks like planted
8 × rocky mountain bee plant
2 × tomatillo
1 × summer squash
The real placement, on an example 7 ft bed - the same layout the planner draws.
Between seasons
Nothing is left living in the bed over winter; the whole planting is re-established.
What it claims
an insectary beside a fruiting crop improves predator presence
A good hunch
Mechanism: The flowers supply nectar and pollen that adult predators and parasitoids feed on, so a patch of them near a fruiting crop keeps those beneficial insects fed and close by to work the crop. The effect is real but not guaranteed — an insectary patch does not always suppress pests.
Evidence: confirmed against 1 source in the research-based literature.
the source
UC Statewide IPM Program (UC IPM), 'Insectary Plants', ipm.ucanr.edu/environmental-protection-and-pesticide-stewardship/insectary-plants (read 2026-07-11): 'Insectary plants are those grown to attract, feed, and shelter insect parasites (parasitoids) and predators to enhance biological pest control. Insectary plants provide nectar and pollen, which the adults of many natural enemy species need to consume.' (UC IPM also notes 'insectary plants may not always help' suppress pests — consistent with the 'reasonably supported' framing.)
squash foliage suppresses weeds and retains soil moisture
A good hunch
Mechanism: A living cover keeps a leaf canopy over the ground: it intercepts light at the soil surface, so weed seedlings are shaded out before they establish, and it slows surface evaporation, so the soil holds more water. In a Three Sisters planting this is the squash's role — its broad leaves shade the bed.
Evidence: confirmed against 2 sources in the research-based literature.
Rotation history attaches to the ground, not the bed, so this is what the planner remembers about this patch after the guild comes out.
Nightshades: full load
Cucurbits: full load
Spiderflower Family: partial load
Rules that gate it
Promising
A guild deposits family history across its whole footprint.
costly if ignored → warn
Mechanism: Soilborne inoculum and overwintering pest stages disperse across the bed. A guild dominated by one family forecloses that ground for that family's rotation interval.
Evidence: confirmed against 2 sources in the research-based literature.
the 2 sources
Same soilborne-persistence basis as other rules here (read 2026-07-11): NC State Extension (Verticillium microsclerotia 'can survive in the soil without a host for up to 10 years'; Septoria/early blight 'overwinter on infected tomato debris and solanaceous weed hosts') and UMN/Ohioline (squash bug overwinters 'as adults in sheltered places ... they fly to growing cucurbit plants'; squash vine borer 'overwinters in the soil as a pupa').
Those sources establish that inoculum and pest stages persist in the ground and are not confined to the exact spot a plant occupied; this rule applies that persistence across the whole guild footprint. The whole-footprint attribution is the corpus's rotation geometry (history attaches to ground, not to the individual plant — see the data model), not a claim from an external source.
remedy
Rotate the whole bed the guild occupied, not just the rows that family grew in.
Self-incompatible species require 2+ genetically distinct plants.
costly if ignored → warn
Mechanism: Many fruit trees carry self-incompatibility alleles at a single locus: when a pollen grain shares an allele with the pistil it lands on, the pistil rejects it and no seed sets. A plant's own pollen always shares its alleles, so it cannot fertilize itself — and neither can a second tree of the identical variety, which is genetically the same. Fertilization needs pollen from a genetically distinct plant.
Evidence: confirmed against 2 sources in the research-based literature.
the 2 sources
Penn State Extension, 'Pollination Requirements for Various Fruits and Nuts', extension.psu.edu/pollination-requirements-for-various-fruits-and-nuts (read 2026-07-11): 'Apple: Cross-pollination is always needed to produce adequate fruit crop'; 'Cherry: Sweet: Most varieties are self-unfruitful'; 'Those fruits listed as self-fertile will set fruit with their own pollen.'
'Pollen-Pistil Interaction in Response to Pollination Variants in Subtropical Japanese Plum (Prunus salicina Lindl.) Varieties', PMC9692414, pmc.ncbi.nlm.nih.gov/articles/PMC9692414 (read 2026-07-11): 'Japanese plum exhibits gametophytic self-incompatibility (GSI), a prezygotic reproductive barrier where genetically related pollen is rejected by the pistil ... governed by a polymorphic locus (S) ... Self-incompatibility occurs when the same S-allele is expressed in the haploid pollen grain and the diploid pistil.'
remedy
Plant at least two, and where the species needs a second cultivar make them different varieties — two of the same will not pollinate each other.
