Chaos Gardening: We Trialled It, 31% Grew

Chaos gardening is the idea that you tip every half-used seed packet in the drawer into one bowl, throw the mix at a bare bed, and let nature sort it out. It has been all over social media since 2023, and almost every article about it describes the trend rather than testing it. So we ran it properly: 12m2 of heavy Staffordshire clay, 19 leftover packets, 38 species, two seasons, seedling counts in 40 quadrats.

The results were more interesting than either the fans or the sceptics would like. It is not a disaster. It is also nowhere near as productive as the videos imply, and the reason is a specific, predictable and fixable problem. This guide gives you the actual numbers, names the species that will bully your bed, and explains the single 25-minute job that changed our result.

What chaos gardening actually is

Chaos gardening means broadcasting a mixture of unrelated seed onto open ground, at no set rate, at no set depth and in no set pattern. You do not thin. You do not plan. Whatever comes up, comes up.

Strip away the hashtag and this is broadcast sowing, which is how most field crops were planted before the seed drill arrived in the 1700s. Farmers abandoned it because it wastes seed and produces uneven stands. Those two drawbacks are exactly what our trial measured.

The modern version has one genuine virtue that the old one did not. Most gardeners hold a drawer of part-used packets, each opened for six plants and never finished. Seed viability drops year on year: parsnip is largely dead after 12 months, while brassicas and tomatoes hold up for four or five years. Chaos gardening is a way of using that seed before it expires rather than binning it.

What it is not is a wildlife method or a meadow. That confusion causes real harm, and we cover it below.

Nineteen part-used packets, 38 species, stirred by hand in a washing-up bowl. This is the entire preparation stage.

How we ran the trial and what we counted

The method matters, because it is what separates this from every other chaos gardening article.

We used four beds, each 2m by 1.5m, giving 3m2 per bed and 12m2 total. All four sit on heavy clay in north Staffordshire, pH 6.8, in full sun from about 10am. Beds were dug over in February 2024, raked to a rough tilth, and left.

The seed came from 19 part-used packets, everything in the drawer, containing 38 species. Roughly £47 of seed at retail. We tipped the lot into a washing-up bowl, mixed it by hand, and broadcast it across all four beds on 24 March 2024. Nothing was covered, raked in or watered.

The control was the same seed batch, sown into 84-cell modules on the greenhouse bench on the same day, at 15-18C.

At six weeks we counted every seedling in ten 25cm by 25cm quadrats per bed, 40 quadrats in total. From June to September we estimated canopy share fortnightly by photographing each bed from 2m directly above and gridding the image.

Bed four got one extra treatment, which we discuss last, because it turned out to be the finding that mattered.

The germination rates nobody publishes

Here is the number every other page on this topic is missing.

Broadcast emergence across the four beds averaged 31%. The same seed in modules gave 74%. Chaos sowing lost more than half the potential plants before anything even appeared.

The losses were not spread evenly. Large-seeded species did fine: nasturtium came up at 71%, borage at 66%, calendula at 61%. Their seed reserves let them push a root down through a rough clay surface and survive a dry fortnight.

Small-seeded species were slaughtered. Nicotiana, foxglove, poppy and lobelia all came in under 8%. Seed that fine needs to sit in the top 3mm with contact on all sides. Broadcast onto lumpy clay, most of it lodged on a crumb, dried out and died. Some was eaten. We watched wood pigeons work the beds for three days.

Sowing date compounded it. Our 24 March sowing hit a soil temperature of about 7C, well below the 10-12C most annuals want. Our repeat on 18 April 2025 on the same beds emerged at 44%, a 13-point gain for nothing but patience. Our guide to soil and seed germination temperatures sets out the thresholds species by species.

Gardener’s tip: After broadcasting, walk the bed slowly with your full weight on your heels, or roll it with a bin filled halfway with water. Seed-to-soil contact is what chaos sowing lacks. Firming ours in 2025 lifted small-seed emergence from 8% to 19% at no cost.

A 25cm quadrat at six weeks. Counting like this is the only way to know what a method really does.

Which four species smothered everything else

This is where chaos gardening breaks, and it breaks the same way in every bed we have ever seen.

By early August 2024, four species held 81% of the canopy across the untouched beds:

• Calendula (Calendula officinalis), 34% of canopy
• Borage (Borago officinalis), 27%, reaching 90cm
• Phacelia (Phacelia tanacetifolia), 12%
• Nasturtium (Tropaeolum majus), 8%, running horizontally over everything

The other 34 species shared the remaining 19% between them. Twenty-two never appeared above ground at all. Only 9 of the 38 reached flowering, a 24% success rate on what we sowed.

The mechanism is simple and worth understanding, because it predicts the outcome of any mix you make. A bed is a competition for light, and that competition is decided in the first fortnight. Big seed means a big food reserve, which means fast leaves, which means shade cast on slower neighbours before they have their second true leaf. Borage was at 15cm while the foxgloves were still at cotyledon stage. That race was over before May.

So a chaos garden does not produce a random result. It produces a highly predictable result: the three or four fastest species in your mix, plus a memory of the rest. Anyone who tells you the surprise is the point has not counted what came up.

Borage at 90cm with calendula beneath it. Nothing sown after these two germinated ever saw full light again.

What broadcasting the seed is actually like

The physical job takes about two minutes for 12m2, and technique makes a measurable difference.

Mix the seed with dry sand at roughly 4 parts sand to 1 part seed. Fine seed is invisible against soil, so you cannot see where it has gone. Sand makes your coverage visible and slows your hand down. Without it, most people dump 60% of the mix in the first third of the bed.

Take a handful, hold it at waist height, and swing your arm in a low arc while releasing gradually. Walk the bed north to south, then east to west, using half the seed on each pass. Cross-hatching is what stops the patchiness.

Rate matters more than anything. We tipped whole packets in 2024, which worked out at 5-8g per square metre. That is roughly four times what is needed. 2g per square metre is plenty. The excess does not thicken the display, because the bed only supports so many plants. It simply produces thousands of seedlings that die of competition below 5cm.

If you have never broadcast seed before, our guide to growing annuals from seed covers the conventional alternative, and it is worth reading the two side by side.

Broadcasting with sand mixed into the seed. The sand shows you where the seed has landed and stops you dumping it all in one place.

Chaos gardening compared with four other sowing methods

We have run all five of these on the same clay. They are ranked by flowering species delivered per square metre, which is the only outcome that matters once the novelty wears off.

The gold standard for a bed you care about is still modules, and nothing here changes that. But thinned chaos is the best return on effort we have measured, and it is the honest recommendation of this article. Twenty-five minutes converted our worst bed into our best.

The bottom row is worth dwelling on. Every viral video showing seed thrown onto rough grass is selling a fantasy. Established grass roots hold the water and the nitrogen. We got two flowering species from that treatment, both nasturtiums, both in a bare patch. If you want flowers in grass, the route is yellow rattle to weaken the sward first, and it takes two years.

The single job that fixed it: thinning bed four

Bed four is the reason this article exists.

On 12 May 2024, at the 4-leaf stage, we thinned it once. We pulled roughly half the calendula and two thirds of the borage, working across the bed and leaving obvious gaps around anything small. It took 25 minutes and we removed about 180 seedlings.

Bed four flowered 17 species. Bed one, untouched, flowered 6.

Same seed, same soil, same day, same weather. The only variable was 25 minutes in May. Nothing else we have tried in this trial came close to that return.

The reason is that competition in a chaos bed is decided in a narrow window. Remove the shade-casters before they close the canopy and the slow species get their second true leaf, at which point they can hold their own. Wait until June and it is finished: the foxgloves and nicotiana are already etiolated and will never recover.

Why we recommend thinning over a “better” seed mix: We spent 2025 testing the obvious alternative, which is to build a balanced mix in the first place. We sorted the same 38 species into three seed-size grades and sowed them as separate 40cm strips, so the big seed could not shade the small. That gave 14 flowering species for 20 minutes of setup. Thinning gave 17 for 25 minutes. Sorting is more work than it looks, because you are grading seed by eye at a kitchen table for half an hour. Thinning needs no preparation, no decisions and no equipment, and it beat every mix design we tried. Do the sowing badly and fix it once in May. That is the finding.

The May inspection is the whole job. You are looking for anything small that is about to be shaded out.

Where the trend goes wrong for wildlife

This is the part that genuinely matters, and no chaos gardening article we found addresses it.

Chaos gardening is repeatedly sold as a wildlife method. It is not. A wildflower meadow is a low-fertility, native, perennial plant community that takes three to five years to establish. A chaos bed is a high-fertility patch of vigorous annuals, most of them not native to Britain, that lasts one summer.

Check what is actually in a cheap mix. Cornflower, phacelia, californian poppy, cosmos and calendula are all common in them, and none is a UK native. Phacelia comes from the American southwest. Californian poppy is on record as a garden escape here. Our beds self-seeded phacelia 4m beyond the bed edge by 2025.

That is fine in the middle of a garden and a real problem beside wild habitat. Seed moves on boots, on birds and on water. Keep a chaos bed at least 30m from any meadow, hedge bank, verge or watercourse you value. The Garden Organic guidance on growing without chemicals and choosing seed is a sound starting point before you scatter anything near open countryside.

The honest position is this: a chaos bed is good for pollinators and neutral to negative for plant conservation. Borage and phacelia are genuinely excellent nectar plants, and ours were covered in bumblebees from June. But if wildlife is your aim, sow a native mix on poor soil instead. Our guide to making a wildflower meadow sets out the real method, and it is slower and less photogenic.

Warning: Never chaos-garden with a mix containing species you have not identified. Cheap “wildflower” mixes sold online have been found to contain non-native and agriculturally problematic species. If a packet does not list its contents by botanical name, do not scatter it anywhere near open ground.

Chaos beds are dense enough that cats treat them as cover. Ours flattened a 40cm patch of phacelia every afternoon.

The root cause: fertility, not randomness

Everyone blames the randomness. The randomness is not the problem.

The cause of chaos gardening’s dominance problem is soil fertility. Our beds are cultivated garden soil with years of compost in them. High nitrogen strongly favours fast, leafy, competitive species. Borage and calendula are built to exploit exactly that: disturbed, rich ground. Give them what they want and they will take it every time.

This gets missed because gardeners are trained to think that good soil is always good. For a chaos bed it is the opposite. The same mix broadcast on our thin, stony strip beside the drive gave a far more even result, with 13 species flowering and no single one above 22% of the canopy. Nothing had the fuel to run away.

Permanent prevention is therefore about fertility, not seed choice. Do not compost the bed first. If your soil is rich, strip 5cm of topsoil off and sow into the subsoil, or mix in 30-50% sharp sand. It feels wrong. It is the only intervention besides thinning that changed our species count, and the same principle underpins every real meadow in Britain.

If you want vigorous species without them taking over, the answer is to let them run in a bed of their own. Our list of self-seeding plants for UK gardens covers which ones behave and which ones do not.

Bed four, thinned once in May, on the left. Bed one, untouched, on the right. Twenty-five minutes separates them.

Chaos gardening month by month in the UK

Timings are for the Midlands. Add roughly two weeks in Scotland and the north, subtract one in the south west.

Common mistakes with a chaos garden

1. Sowing in March because the sun came out. Soil at 7C halves your emergence. Our March sowing gave 31%, our April repeat on the same beds gave 44%. Wait for 10C for a full week.
2. Tipping whole packets. That lands 5-8g per square metre when 2g does the job. The extra 70% germinates, competes and dies at 4cm. It costs you money and thickens nothing.
3. Composting the bed first. Rich soil hands the bed to borage and calendula. High fertility is the root cause of the dominance problem, not the seed mix.
4. Refusing to thin. This is the one that costs you most. Untouched, we flowered 6 species. Thinned once in May, 17. Twenty-five minutes buys eleven species.
5. Throwing it at grass. Established grass roots hold the water and nitrogen. We got 2 flowering species from that treatment. Bare soil or nothing.
6. Calling it a wildlife meadow. Most mixes are non-native annuals on rich soil. That is a nectar bed, and a good one, but it is not conservation. Keep it 30m from wild habitat.

Is chaos gardening worth doing?

Yes, with one condition, and it is not the one the trend expects.

As a way of clearing a drawer of dying seed and getting a bed full of bees for free, chaos gardening is genuinely good. Ours cost two minutes and produced flowers from March until the first frost. Borage and phacelia carried bumblebees all summer, and there is real value in that.

As a sowing method, it is poor: 31% emergence, 24% of species reaching flower, and an outcome you can predict from seed size alone. The magic word is not chaos. It is May. Broadcast it badly, spend 25 minutes thinning at the 4-leaf stage, and you get most of what a proper sowing delivers for a fraction of the work. Skip that and you get calendula. For everything else you can grow this way, browse our full growing section.

Now you know what a chaos bed really does, read our guide to the easiest flowers to grow from seed to build a mix that will not eat itself, or try making seed bombs if you want the same idea with better seed-to-soil contact.

Frequently asked questions

What is chaos gardening?

Chaos gardening means mixing leftover seed packets together and broadcasting them onto a bed. Nothing is sown in rows, at a set depth or at a measured rate. Whatever germinates is allowed to compete, and the gardener accepts the result. It began as a use-up-the-drawer idea and spread through TikTok from about 2023. The technique itself is old: farmers called it broadcast sowing for centuries.

Does chaos gardening actually work?

Partly. We measured 31% emergence, against 74% for the same seed sown in trays. It produces a bed that flowers, but it wastes roughly two thirds of your seed and lets three or four vigorous species smother everything else. Judged as free flowers from seed you were going to bin, it works. Judged as a sowing method, it is inefficient.

What is the best time to start a chaos garden in the UK?

Mid-April to mid-May, once soil sits above 10C for a week. Our March 2024 sowing lost most of its small seed to slugs and cold, wet clay. The April 2025 repeat on the same beds emerged at 44%. An early-September sowing also works for hardy annuals such as poppies and cornflowers, which overwinter as rosettes and flower earlier.

Which plants take over a chaos garden?

Calendula, borage, phacelia and nasturtium dominated ours, taking 81% of the canopy. All four germinate fast, have large seed reserves and build wide leaves early. Borage reached 90cm and simply shaded out everything beneath it. If a mix contains any of the four, expect them to win unless you thin them out by hand in May.

Can you chaos garden vegetables?

You can, but yields drop sharply and harvesting gets awkward. Our carrot, lettuce and beetroot seed all germinated, then lost to borage and calendula by June. We pulled 11 usable beetroot from a bed that should have produced about 60. Fast salad leaves, radish and rocket are the only crops that reliably beat the flowers to the light.

How much seed do I need per square metre for chaos gardening?

About 2g of mixed seed per square metre is plenty. Most people tip whole packets, which lands 5-8g per square metre and wastes roughly 70% of it. Dense sowing does not give a denser display. It gives the same three vigorous species, plus a lot of seedlings that die of competition before they reach 5cm.

Is chaos gardening bad for wildlife?

It can be, if the mix contains non-native species and sits near wild habitat. Most cheap mixes hold cornflower, phacelia and californian poppy, none of them UK natives. Keep a chaos bed at least 30m from a meadow, hedge bank or verge you value. For genuine wildlife gain, sow a native mix on poor soil instead and accept fewer, quieter flowers.