Sleeping seeds · onion, lotus, date
From the one-year onion to the 2,000-year date: climb the ladder of sleeping seeds, guess how many wake from the bottles buried in 1879 and catch the impostors — the mummy wheat and the 10,000-year lupine. The last rung hides a trap.
Climb the ladder Skip to the game
Drawing: a seed waking up · Ages with sources below
Seven rungs, from one year to tens of thousands. Each rung says what fell asleep, for how long, and how we know it is no impostor. Scroll and wake them one by one.
Rung 1 of 7 · 1 year
Cornell's home-storage guide says onion-family and parsnip seeds often lose about half their germination each year; last year's packet is a lottery. Their thin coats oxidise fast. Worth running the home test (see Questions) before you bin it.
Check: home-storage figures, not lab records.
Rung 2 of 7 · a few years
Wheat kept dry loses its vigour within a few years. The 1842 committee got no shoots from its controlled test; Henslow later identified Tupper's wheat as the modern Belle Vue Talavera.
Check: the verdict repeated since 1842.
Rung 3 of 7 · 80 years · 5 of 50
From Beal's bottles buried in 1879, after 80 years 5 evening-primrose seeds out of 50 still germinated. One of the last survivors: of the other species, only a dock (Rumex) and mullein still germinated that year.
Check: the 2002 Beal table — every point another bottle.
Rung 4 of 7 · 141 years · 20 of 50
The record of Beal's bottles: from the one dug up in April 2021, after 141 years, 20 mullein seeds out of 50 germinated — 40%. Nineteen were moth mullein, one a hybrid.
Check: DNA sequenced, sterile control soil grew nothing.
Rung 5 of 7 · 1,288 ± 271 years
From a dry lake bed in China, a lotus seed germinated after nearly 1,300 years — and only then was it dated. Its hard, water-impermeable coat helps preserve it.
Check: dated AFTER germinating, from dead coat tissue.
Rung 6 of 7 · ≈2,000 years
Methuselah, from the seed found at Masada, germinated in 2005 and grew normally for two years. Then six more, from 32 seeds gathered at four sites (Masada, Qumran, Wadi Makukh, Wadi Kelt) — the researchers gave them study names: Adam, Jonah, Uriel, Boaz, Judith and Hannah; ages between the 4th century BCE and the 2nd century CE.
Check: coat from the rootlets, corrected for 250–300 years of modern carbon.
Rung 7 of 7 · 31,800 ± 300 years · TISSUE, not seed
From squirrel burrows frozen 38 metres deep in Siberia, 36 Silene plants were grown — but from the placental tissue of the fruits, in glass. The mature seeds never germinated. The oldest plant brought back to life — per the 2012 study — not the oldest seed germinated.
Check: seeds beside the dated placenta; the flowers differ from modern ones.
The ladder is logarithmic: each big step adds another zero. The last rung is a different kind — hence the label.
Fifty moth-mullein seeds in a bottle, since 1879. Dig the bottles up one by one and see how many wake — each point is a bottle dug up that year.
each dot is a seed; the green ones woke
In the same 2000 bottle one more seed woke: a single mallow, after a cold treatment. Of the 25 mullein that woke then, 23 were moth mullein and 2 an undetermined Verbascum — putative hybrids.
| Year | Germinated | Percent |
|---|---|---|
| 1930 | 31 / 50 | 62% |
| 1940 | 34 / 50 | 68% |
| 1950 | 37 / 50 | 74% |
| 1960 | 35 / 50 | 70% |
| 1970 | 10 / 50 | 20% |
| 1980 | 21 / 50 | 42% |
| 2000 | 25 / 50 | 50% |
| 2021 | 20 / 50 | 40% |
Each point is another bottle: that is why 1970 drops to 10 between 35 and 21. Most other species had died within the first 60 years. Four bottles remain buried; the last comes up in 2100. The line between points is a drawing convention, not a measurement.
Plants did rise from grains sold as 'mummy wheat' — but none was ancient Egyptian wheat. In the 1840s London buzzed: wheat from the Theban tombs, growing in aristocrats' gardens. The story lasted a century. Here it is end to end, with the date it was born and the experiments that killed it.
Wilkinson brings a jar from the Theban tombs to the British Museum; Pettigrew breaks it open. That same winter, the first 'mummy seeds' of the botanical literature grow in Bohemia.
Martin Tupper receives 12 grains from Pettigrew, out of Wilkinson's jar.
Tupper sows the 12 grains in four pots of sifted soil. A single one rises — two ears, 27 grains. In 1841 he sows again: over a hundred ears.
Faraday shows one ear at the Royal Institution, after a lecture, before some 700 people. Prince Albert receives a specimen too.
The British Association committee tests mummy wheat under controlled conditions: not one shoot rises.
Henslow shows Tupper's plant is the modern Belle Vue Talavera breed — and Wilkinson's sample contained maize, a New World plant. The mix marks a contaminated sample.
Budge sows at Kew grains from a model granary out of a tomb of about 1200 BCE, under coloured glass. After three months the grains had turned to dust.
Budge's Times challenge is tested at Cambridge, at the Official station: 50 grains in sterile sand at 17 °C. Zero germinated. Same at the Smith brothers.
The story's roots are unclear — perhaps several beginnings. The years 1833–34 are the first appearance we can catch in documents.
Each with its source. Think before you tap — the third is a trap.
1 A date seed about 2,000 years old germinated and grew normally.
2 Wheat from Egyptian tombs was successfully grown in England in the 1840s.
3 A 32,000-year-old seed germinated.
4 A lotus seed about 1,300 years old germinated.
5 Lupine seeds 10,000 years old grew from the Yukon permafrost.
The story does not end in the past: today seeds sleep in gene banks — and Romania's gene bank has sent part of its collection to Svalbard.
At Svalbard, in the Spitsbergen permafrost, sleep 1,401,285 samples (June 2026, the 70th deposit). The global backup for the collections gene banks deposit there.
At Suceava, the 'Mihai Cristea' Plant Genetic Resources Bank keeps over 25,000 accessions from 430 species. The first shipment to Svalbard left in early 2020; in October 2024, 158 accessions from 36 species followed.
A Judean date stone about 2,000 years old (Methuselah: 351 BCE–21 CE), germinated in 2005 and grown normally. Six siblings joined it, from 32 seeds planted.
Silene from Siberia: 31,800 ± 300 years old. The plants were grown from the fruits' tissue, in glass — the mature seeds never germinated. The oldest plant brought back to life (per the 2012 study), not the oldest seed germinated.
Wheat kept dry dies within a few years. The 1840s samples were contaminated: Henslow identified Tupper's wheat as the modern Belle Vue Talavera breed, and maize turned up in Wilkinson's sample, a New World plant. The controlled tests of 1842, 1897 and 1934 gave zero.
With radiocarbon from spare scraps: coat left on the rootlets (date), dead coat after germinating (lotus), seeds beside the cultured placenta (silene). Every age carries its error bar, printed on the page.
Viable means alive — it can germinate in good conditions. Germinated means it did. The tetrazolium test stains living tissue red, even when the seed sleeps deep.
Yes: a few seeds on damp paper, under the packet's conditions, count how many rise. The more, the better the packet. The method mirrors the page's tests, in miniature.
The ages come from radiocarbon dating, each with the lab's error (±): lotus 1,288 ± 271, silene 31,800 ± 300, the date on century ranges. BP counts years back from 1950.
Viable does not mean germinated: a living seed can sleep, and tissue cultured in glass (silene) is no germinated seed. The page never mixes the three.
Each case carries its check beside its story (look for 'Check'). The golden rule: date the seed you grew, not its neighbours.
The Beal curve's points are different bottles, dug up in different years — that is why the curve wobbles (1970 drops between 1960 and 1980). The line between points is a drawing convention, not a measurement.
What we do not print: hearsay ages (we found no dated primary source for the '600-year' canna stories), Vault figures beyond the June 2026 count and its 100-year experiment — that is the sister page's subject.
Every age on the page comes from the rows above; doubts stand beside the claim they limit.