From “Narcissus” in Flowers and Their Histories by Alice M. Coats:
“The art, hobby or profession of breeding garden varieties of daffodil is not of recent origin; it had already begun in Jacobean times, and instructions for raising seedlings were given by John Rea in 1665. The majority of early gardeners, however, seem to have been content to import their new varieties from the Continent, for Philip Miller in 1724 complained that ‘in England there are very few persons who have patience to propagate any of these flowers that way, it being commonly five years before they can expect to see the fruits of their labour’. It was not until Dean Herbert of Manchester conducted his experiments in hybridizing, in preparation for his book on the Amaryllidaceae, published in 1837, that interest in daffodil-raising was really aroused in England….
“His work inspired Edward Leeds of Manchester, William Backhouse of Darlington, and finally Peter Barr, to specialize in the flower. Peter Barr was justly called the Daffodil King; for besides founding the firm of Barr and Son, and continuing the development of the strains started by Backhouse and Leeds, he travelled extensively in Spain and southern Europe in search of wild daffodils, and was instrumental in restoring to our gardens many species that had been lost since the time of Parkinson….
“He also wrote a book called Ye Narcissus, and was largely responsible for the organization of the first Daffodil Conference in 1884. He died in 1909, but his work was carried on by George Herbert Engelheart, whose honorary title was Father of the Modern Daffodil. The first Daffodil Show was held at Birmingham in 1893; and the flower has made steady progress ever since. It has now become one of the world’s most popular flowers, and specialists devote to it the attention and care once lavished on the carnation, the tulip and the auricula. In 1903 the Rev. W. Wilks (then Secretary of the R.H.S.) expressed the opinion that no further advance in daffodil-breeding was either possible or desirable; but to us, looking back, the modern daffodil seems only to have been in its infancy at that date. Since then the lovely pink-cupped varieties have made their appearance, led by ‘Mrs. R. O. Backhouse‘ in 1923; and a later achievement was a trumpet daffodil with a white perianth and a scarlet cup exhibited by her son, Mr. W. O. Backhouse, in 1953….
“We may yet see an all-red daffodil, or a white-and-green trumpet, or even a shade approaching blue.”
From “Blue Daffodils” in Blue Daffodils and Other Poems by Bryan Owen:
I saw blue daffodils
swaying in a lilac breeze
one warm afternoon in May;
a lime-green sun shone down
from a pink-striped sky
as below me
dirty black cars crawled
creepily back to their holes
in the ground.
A bee buzzed brightly,
a cat flew overhead
and it was good to be alive.
Hello!
We have arrived at the end of Daffodil Season — which tends to start in February here in the Southeast and wind its way into late March or (if we’re lucky) early April. While I have photos of other flowers (such as snowflakes, tulips, dogwoods, and quinces) in my backlog, I’ll take a couple of trips to Oakland’s gardens in the next few days to see if there are any daffodils left. Most likely, early irises — probably white ones — have started to bloom, so I wouldn’t be surprised if they get my attention even if there are a few stray daffodils remaining.
The flowers in this series are a mix of four cultivars: the large white and yellow daffodils are identified as Narcissus × incomparabilis and Narcissus pseudonarcissus; the smaller ones are Narcissus tazetta and Narcissus poeticus. Toward the middle of the galleries, I’ve included four photos showing where the large daffodils were growing, in a shaded section of the gardens where you can listen to them shift in the wind as they bloom, while you rest on a stone memorial bench.
A funny thing happened on the way to writing this blog post. As a summary of early daffodil breeding efforts, I had selected the quotation at the top (from Flowers and Their Histories by Alice M. Coats) to wrap up my daffodil series, and noticed how she mentioned the future possibility of blue daffodils toward the end. Since the book was written in the 1970s, I wondered if progress had been made toward developing daffodil blues, especially since — as is often the case when I analyze colors produced by my camera — I’m very aware that white flowers will appear to have blue tones (something I wrote about here), depending on surrounding colors, the presence of cool sunlight in shadows, and the reflective qualities of the flower I photographed. So I started hunting down the elusive blue daffodils with a simple Google search, thinking that if they did exist, I’d link to their images so you could take a look.
Here’s a screenshot of my interaction with Google on the question of blue daffodils (poke for a larger version):

“Yes, blue daffodils are real” — I am told. Except they aren’t.
Notice the confidence with which Google’s AI has informed me that blue daffodils do indeed exist, while including an image of a painting of blue daffodils for reference. If you try an image search for blue daffodils, you’ll probably see that same painting among the results, along with pictures of blue daffodils which — this should be obvious — aren’t showing a natural blue color, but one added by an image editing program. Note also that toward the end of Google’s response, it mentions Gardens Illustrated, the well-known gardening and horticultural magazine, to which it provides (at the right of the response) a link to this article…
World’s first blue daffodil finally flowers
… published on April 1, 2022.
Does that date seem odd? Or maybe — if you know a minimalist amount of French like I do — you may have noticed that the blue daffodil cultivar’s name — Narcissus ‘Poisson d’Avril’ — appears to have the word “fish” (“poisson”) as part of its name. So a literal translation of that name would be “April Fish Daffodil” while a more accurate, contextual translation would be “April Fool’s Daffodil” since the historical French origins of April Fool’s Day often referenced fish instead of fools. It’s a clever pun on the part of Gardens Illustrated, one that does a nice job of obscuring the joke, and manages to combine gardening, plant breeding, history, and botany in a short but endearing article.
As someone who often can’t resist tunneling down a rabbit hole once I’ve stuck my head in, I decided to learn a little more about how prevalent the misinformation parroted (“hallucinated“) by Google’s AI was around the internet. I’m not going to link to the things I found (spam abounds under the covers) but you can certainly find these and other variations if you want to “do your own research.” And I discovered, among other things:
- an AI-generated website describing the botanical history of blue daffodils, and which daffodils were used to breed them;
- another AI-generated website which goes into detail about the chemical process used to create blue daffodils, which, after paragraphs of implausible but scientific-sounding words and phrases, points out at the end that they don’t actually exist;
- yet another AI-generated website that considered Narcissus ‘Poisson d’Avril’ a real daffodil cultivar, and used a copied page about yellow daffodils from the same site, replacing “yellow” with “blue.”
- references to their commercial availability for gardeners to purchase blue daffodils for their gardens by the year 2030;
- websites describing how to grow and propagate them (presumably not until 2031 🙂 ), with links to photographs of flowers that are actually blue irises rather than daffodils;
- repetition of content from the original April Fool’s Day article from Gardens Illustrated, with the “10 years in the making” updated to 12 years, and repeating references to the “top secret location in Wales” (one of these?) where the flowers were (not) being engineered.
- impossible images showing their presence among other daffodils in gardens, where their blue colors contrast with the more traditional yellow daffodils; and, finally:
- actual sellers on some of the world’s most well-known online shopping sites offering “blue daffodil seeds” (daffodil seeds dyed blue?) for sale that will allegedly produce blue daffodils similar to those in the fake photos accompanying the listings.
The last one is especially interesting to me because it demonstrates the commercialization of misinformation. We got there, apparently, from long-standing botanical curiosity around the idea of blue daffodils; to an April Fool’s prank; to the proliferation of the prank as a fact on AI-generated websites that didn’t get the joke; to Google’s AI assuring us that blue daffodils really do exist and it’s just a matter of a few years before they make it into our gardens; to random sellers offering “blue daffodil seeds” for sale from online shops.
Continuing further down the rabbit hole, I decided to ask three other AI tools — ChatGPT, Perplexity, and ClaudeAI — about blue daffodils too. ClaudeAI — the one I use the most — recognized blue daffodils as part of a joke right off the bat, as did Perplexity and ChatGPT. ChatGPT even had a little more fun with the joke, offering to create a photo of imaginary blue daffodils for me, so I’d have one on hand if blue daffodils someday came to be:

Of the four AIs I queried, then, only Google — the most ubiquitous source of information on the internet — seems to have been fooled. Because the idea of “dueling AIs” amuses me, I went further with ClaudeAI: I uploaded Google’s response and asked for an explanation of how this misinformation made its way around the web. Here’s what it told me:
This is such a fascinating digital folklore situation you’ve uncovered! I love that both Google’s search AI and Gemini confidently presented the April Fool’s joke as fact, complete with the future commercial availability prediction.
What you’re witnessing is essentially the birth of a modern gardening myth, spreading through AI systems. The “Poisson d’Avril” blue daffodil has become a perfect case study in how misinformation can cascade through digital ecosystems:
– A reputable gardening magazine (Garden’s Illustrated) publishes an April Fool’s joke
– Various websites and content farms pick it up as fact
– Search engines and AI systems ingest this information without recognizing the context
– The information gets repackaged and disseminated as factual content
– Eventually, opportunists create commercial listings for impossible products
The irony of multiple AI systems missing the “April Fish” joke while confidently explaining the nonexistent genetics is particularly delicious. It demonstrates one of the current limitations of large language models — they can process vast amounts of text but often miss cultural context clues that would be obvious to humans familiar with traditions like April Fool’s Day.
For gardeners and plant enthusiasts, this serves as a good reminder to approach unusual plant claims with healthy skepticism, especially when they involve dramatic color breakthroughs that haven’t been widely celebrated in horticultural circles.
So there you have it! We’ve debugged a myth about the existence of blue daffodils. The only thing that would make this even more fun is if I ordered some of the seeds and they turned out to grow daffodils with blue petals after all!
Thanks for reading and taking a look!





























