About
A free reference for germinating seeds: native plants, prairie perennials, rock garden species, trees and shrubs.
What it does
Germination protocols for these species are published in about five incompatible notations, mostly abbreviations, and none of them tell you a date. This site stores each protocol as ordered steps — an action, a duration, a temperature — and walks them backwards from your last frost to say when to start.
That is the whole product. Everything else is navigation.
Three rules it keeps
- Every protocol shows the text it came from.
- Under "Show the original text" on each card is the source's own sentence, verbatim. Competing references ask you to trust an uncited letter code. If a card and its quotation disagree, the quotation is right and the card is a bug — tell me.
- Nothing is averaged.
- Where one source reports several treatments, or two sources disagree, every protocol is shown separately with its own dates and its own citation. A blended number would belong to nobody.
- Unreviewed records are not published as protocols.
- A record a person hasn't checked against the source shows its quotation and no steps, no temperatures and no dates — 0 of the 1149 species here are in that state. People act on a number whatever the warning next to it says, so there is no number.
Confidence
Every card carries one of two marks. Verified means a person read the original source and confirmed every field. Derived means it was read from the source by machine and spot-checked by a person. A third state, unverified, exists in the data and never reaches a page.
The name
Deno’s Garden is named for N. C. Deno, whose Seed Germination Theory and Practice is where most of what is here comes from. He put that work in the public domain, which is why it can be quoted in full on every card.He did not make this site and it is not published by him or by his estate — the name is a credit, not a byline. Errors here are mine, not his, and the quotation on each card is how you tell the difference.
Sources
| Source | Licence | Protocols | Facts |
|---|---|---|---|
| Extension Gardener Plant Toolbox | Consulted for factual values; nothing quoted | 0 | 23 |
| GBIF Backbone Taxonomy | Consulted for factual values; nothing quoted | 0 | 133 |
| Germination records | Grower record — an observation, not a published source | 0 | 35 |
| PLANTS Database | Public domain (US government work) | 0 | 339 |
| Plant Finder | Consulted for factual values; nothing quoted | 0 | 564 |
| Plants Poisonous to Livestock | Consulted for factual values; nothing quoted | 0 | 20 |
| World Flora Online | Consulted for factual values; nothing quoted | 0 | 4 |
| Seed Germination Theory and Practice, 2nd ed. | Public domain | 1589 | 39 |
| Seed Germination Theory and Practice, 1st Supplement | Public domain | 539 | 0 |
| Seed Germination Theory and Practice, 2nd Supplement | Public domain | 479 | 0 |
| Seed Information Database | Creative Commons Attribution — reused with credit | 9 | 0 |
Where a species page credits Deno’s Garden Records, it means the grower's notebook listed above — Deno’s Garden is Amir Abed, and observations made here are labelled as observations rather than dressed up as a citation. Everything else is quoted from a published source and named in full on the card that uses it.
Protocols are used under the licences above. The site's own text and code are separate from the sources it quotes.
Where the dates come from
Steps are walked backwards from the day seedlings go outside. Two numbers in that calculation are estimates rather than data, and both change every date on the site:
- 28 days from germination to a seedling big enough to plant out. Flat across every species, because sources almost never state it.
- 21 days for germination itself, used only when the source gives no figure. A stated value always wins.
Frost dates come from NOAA's 1991–2020 US Climate Normals, joined to Census ZCTA centroids — 32,883 zip codes, both sources public domain. The date shown is the median: half of years see a frost later than it. The planner names the station each figure came from and how far away it is, because in hill country elevation moves these dates more than distance does. With no zip code, pages fall back to a national average of May 3, which is wrong for nearly everywhere.
What this is not
It is a reading of published protocols, not horticultural advice and not a guarantee. Seed lots differ in age, provenance and viability; the same protocol run twice gives two answers. Sources disagree, and where they do this site shows the disagreement rather than resolving it for you. Treat every date as the middle of a range and sow more seed than you need.
There are no accounts, no newsletter, no tracking beyond a page-view count, and nothing here is for sale.
Something wrong?
Errors are the point of the mailto links. The most useful reports, in order:
- A card that disagrees with its own quotation — that is a parsing bug and it is the worst kind.
- A protocol that is wrong in the source itself, with a citation for the better one.
- A species that isn't here. Those emails decide what gets read next.
The data is edited as text files in git, deliberately. There is no admin interface and no way for anyone to submit a protocol through the site — every record passes through a person first.