We know edibility for 22% of atlas species

The 29 August 2026 snapshot covers 1,901 catalogue entries. The edibility field has a value other than "unknown" for 426 of them — 22.41 per cent of the catalogue. Among the 1,395 species with at least one observation, 366 have a settled edibility, or 26.24 per cent. All values come from a direct database query counting rows rather than sampling.

The methodology here is entirely our own, which is why it is worth describing precisely. The species catalogue is built automatically from the taxa appearing in observations pulled from iNaturalist and GBIF. Polish names, photographs and descriptions are drawn from Wikidata, Wikimedia Commons and Wikipedia. Edibility is taken from none of those sources — it is entered by hand after checking the literature, because it is the one field where an error can cost somebody their health.

That is where the gap between this indicator and the rest of the catalogue comes from: the automatically fed fields have coverage in the tens of per cent and grow with every run, while edibility grows only when a person sits down and fills it in. Twenty-two per cent is not an import failure. It is a measure of the editorial work still outstanding.

We publish the number because it marks the limit of what the service is good for. The "edible" filter works on a quarter of the catalogue, so a result list is never complete, and the absence of a marking does not mean inedible — it means unchecked. A species page says "unknown" outright rather than guessing. A statistic that flatters its own service is useless — this one is meant to show where it falls short.

What makes up the 426 determinations

The 29 August 2026 snapshot contains 161 species marked edible, 133 inedible, 118 toxic and 14 deadly. Together these are 426 of 1,901 entries, or 22.41%. The other 1,475 carry “unknown”. Among the 1,395 species with observations, 366 have a determination, or 26.24%.

A missing value never means either “safe” or “inedible.” Editors fill this field after checking literature, and imports do not guess it. The statistic exposes the completeness limit of the filter; it is not dietary advice or an invitation to judge a mushroom from its page alone.