A public record, 1590–2026
Every crop formation anyone published, once.
Three archives have recorded crop formations for thirty years, and each of them records a different set, under different names, with different dates. This is all of it, folded into one record per formation — dated, placed, measured where a source measured it, and traceable back to the page it came from.
The season
every season, 1590 to 2026 →Crop formations are a summer phenomenon and a crop phenomenon: 92% of the 6,315 formations with a known day were reported between May and September, and the count collapses when the fields are cut.
The ring is every day of the year. A spoke's length is how many formations across the whole archive were reported on that day.
The busiest days on record
- 25 July
- 107 formations, all years together
- 15 July
- 103 formations, all years together
- 23 July
- 101 formations, all years together
- 22 July
- 98 formations, all years together
- 28 July
- 96 formations, all years together
Every season
the year index →One bar per year of report. The shape of this curve is coverage — who was publishing, and when — and not a measurement of how many formations there were.
The 2026 season
all 25 →The most recent season the archive holds, as it was reported.
The most recent season the archive holds, newest first. A tile with a picture says whether it is a photograph of a field or somebody's drawing of one; a tile without one is the record itself, with a circle at the formation's own recorded diameter.
What this archive is, and is not
Read this first
A reported date is the date a formation was found and reported, not the date it was made, and this archive records what three sites published, not what happened in the fields.
Nothing here is evidence that a formation was made by anything in particular. Where an author of one of the archive's documents claims otherwise, the page records that they claimed it, in their own words, and says how the claim was counted.
5,144 of 7,551 formations have no photograph in the archive at all, and 6,689 are held by a single source, so nothing about them can be cross-checked. Where that is so, the page says so.
7,197 formations carry a coordinate. None of them was published by a source as a coordinate: each is derived from the place name a source printed, and each page states how precise that makes it.