Evidence Gap Index

It exists for two reasons. A reader deciding whether to act on a figure deserves to know how well that particular kind of figure is evidenced, and we would rather show the number than be asked for it. And a gap nobody has counted never gets closed: everything below is also our own worklist, in the order we think it matters.

Nothing here is a score. Every line is a count you can check against the pages it describes.

510 of 510 maturity figures do not state their basis. Days from sowing and days from transplant differ by the weeks spent raising a seedling, and almost no seed source says which it means. Where the basis is unknown the location pages report a range rather than picking one, but the underlying figures stay ambiguous until suppliers say.

Species and Colour, shape and size class and Culinary use inherit rather than cite. Resistances, measurements, species and breeder attributions each name their own sources. These do not: they rest on the record's source list as a whole, so you cannot tell which citation supports a particular one. Closing that is ordinary outstanding work.

44 of 80 trial entries are unpublished. Across 5 trials we hold, most entries are varieties this database does not have a record for, or names that resolve to more than one variety we do hold. Their results are withheld rather than guessed at, and the trial pages list them by name with the reason.

Watermelon, Cucumber, Melon, Corn, Wheat, Perennial Ryegrass carry crop-level information and no varieties. They are real pages about real crops, but the variety depth behind tomato and pepper does not exist for them yet.

A dash means no value is documented in the cited sources. It does not mean zero, and it does not mean none. A variety with no recorded resistance has not been shown to be susceptible; it has not been shown to be anything, because nobody published a result we could point at.

Thin evidence is also not the same as bad evidence. For most disease-resistance genes the breeder does the screening and publishes it in their own catalog, and no third party re-tests it. That claim is the best evidence in existence, and it is still worth knowing that it comes from the seller.