Sowing Seeds Successfully: Why It Fails, and What Fixes It

A Seed Sowing Never Fails at Random

The first reflex is to blame the seed packet. That is almost always wrong, and more importantly it is unproductive: until the failure has been named, it gets repeated. A seed is a threshold device. It clears each threshold or it does not, and there are only a few of them: continuous water, a temperature range, a depth compatible with its reserve, sometimes light, then an environment that does not kill it during the days when it has no defences.

The tipping point is worth knowing, because it explains half of all invisible failures. Germination unfolds in three phases. The first is imbibition: the seed drinks quickly, its membranes rehydrate, it repairs damage accumulated in its DNA and proteins during storage, and cellular respiration restarts. The second is a plateau, where water uptake slows and reserves are mobilised. The end of this second phase is the emergence of the radicle, and at that precise moment the seed loses its tolerance to drying out. Until it has broken through, an ordinary seed can be drained and dried again without harm. Once the radicle has emerged, a tray that dries out for a few hours kills everything that had germinated, and nothing shows on the surface. The gardener sees a tray that never comes up and blames the seed.

Hence the one move that replaces every guess: wait for the stated delay, let it pass clearly, then dig up five or six seeds with the tip of a knife and look. An intact, hard, unswollen seed: it never imbibed, either the substrate was too dry, or the seed was dead. A swollen, soft, browned seed with no radicle: it drank, then rotted, the substrate was waterlogged or too cold for too long. A white radicle out, with the stem bent over just under the surface, stopped partway: sown too deep, the reserve ran out before reaching the light. Seedlings that came up and then toppled within hours, the stem pinched at substrate level: damping-off. Seedlings that came up pale and leggy: light arrived too late. Five diagnoses, five different fixes, and one minute of digging to tell them apart.

The Temperature That Matters Is the Substrate’s, Not the Room’s

A room thermometer has never made a seed germinate. The number that decides is the substrate’s, at sowing depth, and it runs lower than the room’s. A moist substrate constantly loses heat through evaporation: it behaves like a cooler and settles below the temperature of the surrounding air. Set the tray on a tiled floor, on a windowsill, or against a single-glazed window, and the gap widens further. A kitchen probe pushed into a cell and read first thing in the morning, at the coldest moment, costs a few euros and eliminates half of all failures.

What a few degrees cost can be measured. J. F. Harrington’s data, from the University of California, Davis, remain the reference: they give the number of days before seedlings appear for seeds sown at 1.3 cm, under controlled conditions, substrate by substrate. A pepper comes up in 25 days at 15°C, 12 days at 20°C and 8 days at 25°C. A tomato takes 43 days at 10°C, 14 days at 15°C and 8 days at 20°C. Aubergine goes from 13 days at 20°C to 5 days at 30°C. The delay is not just an inconvenience for the calendar: every day spent under the surface is a day of exposure to soil-borne fungi. Germination speed is itself a health measure.

The upper limit is the side of the story nobody tells, and it traps at least as many sowings as the lower one does. Hotter is not better. Lettuce still germinates at 0°C, in 49 days, and refuses to germinate above 30°C, earlier still in some varieties: that is not a weakness, it is an active block, thermo-inhibition. A flash of red light lifts this block up to around 28°C, which is precisely why lettuce is among the seeds that need light. Spinach follows the same slope in reverse of most vegetables: its germination rate falls as the substrate warms up. Parsnip takes 14 days at 20°C, 32 days at 30°C, and no longer comes up beyond that. Onion comes up in 4 days between 25 and 30°C, but falls back to 12 days at 35°C. Sowing lettuce or spinach in a hot conservatory is failing from excess, with exactly the same symptoms as failing from cold.

A heat mat solves the lower-limit problem, on two conditions. The first is a thermostat and a probe in the substrate, because models differ widely in how much they add above room temperature, anywhere from a few degrees to ten or so. The second is removing it as soon as half the seedlings are up. The temperature that suits the seed is higher than the one that suits the seedling, and a warm substrate under weak light is the exact recipe for a leggy plant.

Depth: Two to Three Times the Diameter, and the Seeds You Must Not Bury

The rule fits in one line: a seed is buried at two or three times its own diameter, never more. The reason is a matter of accounting. The reserve held in the seed is finite, and it has to fund the entire trip up to the light, since until the cotyledons have emerged, the seedling produces nothing. Sown too deep, a seed germinates perfectly and dies underground, its reserve spent three centimetres short of the goal. It is the one case where nothing turns up on the surface even though everything worked. The second cost is a health one, and agricultural extension services state it explicitly: the longer a seedling takes to reach the surface, the longer it stays exposed to the organisms behind damping-off. Depth and disease are the same problem.

The opposite excess also exists, and it mostly concerns seed sown directly outdoors. A barely covered seed is at the mercy of half a day of dry wind. Carrot piles up both difficulties: it needs 5 mm at most, it takes 14 to 21 days to come up outdoors, against a week to ten days under controlled conditions, and the seedbed has to stay moist for that entire stretch. What stops it is almost never the depth but the surface crust, formed by a heavy downpour or a rose-fitted watering can, which the seedling cannot break through. A row cover of fine, non-clogging material laid over the row solves this better than any amount of extra watering.

Finally, some species need light to germinate, and covering them is exactly what kills them. The verified list is short, and it should not be stretched by guesswork: lettuce, dill, summer savory, petunia and begonia are set on the substrate, pressed in with the flat of the hand, and never covered, the summer savory note applying to the annual, not the perennial. Celery falls into the same category: it needs diffuse light, is happy with a very thin covering, and its germination improves markedly when its nights run 5 to 8°C cooler than its days. Onion, on the other hand, has no business on this list: it does not need light, which actually inhibits its germination, and it is sown at about 1.3 cm. The trap mostly closes the other way round, because people assume a fine seed must be sown on the surface: basil is covered with a few millimetres and light does not govern its germination, nasturtium is covered generously, and carrot does not need light either, whatever gets repeated. For the species that need light, a clear cover keeps humidity up without blocking the signal, whereas a newspaper or a piece of card laid over the tray blocks both.

Two details finish the job. Contact matters as much as depth: a seed resting over a pocket of air does not imbibe, hence the value of firming it down lightly, with a board or the palm of the hand, rather than just sprinkling it on. And what gets sown is not always a single seed. In beetroot, the grain sold is a dry fruit that commonly contains several embryos: three seedlings from one sowing spot is not a dosing mistake, and thinning is not optional there, except with monogerm varieties.

Seed-Starting Medium Is Fine and Low in Nutrients, and That Is Deliberate

Two features set a seed-starting medium apart from multi-purpose compost, and neither one is a matter of quality. The first is fineness. A seed one millimetre across needs continuous contact with moist particles to drink; a half-rotted fragment of bark resting against it is a pocket of air, in other words a seed that does not imbibe. A finely sieved medium gives even imbibition, and therefore even emergence, which matters more than it appears to: a tray that comes up over fifteen days is unmanageable, because the first seedlings are already demanding light and air while the last ones still need warmth and a cover.

The second is low fertility, and it is intentional. Dissolved salts in a growing medium lower its water potential: the seed has to drink against that gradient, and the more fertile the medium, the slower and more incomplete imbibition becomes. This is not a theory, it is a measured effect, and lettuce sown in a medium rich in soluble salts germinates more slowly and in smaller numbers. Professional benchmarks put a good-quality potting mix at 1.5 to 3 mmhos per centimetre in conductivity, and a pH of 5.5 to 6.5; a medium meant for germination is simply less fertile than a potting mix. A seedling lives off its seed’s own reserve until the first true leaves appear: before that stage, fertility in the growing medium does it no good, and actively harms it.

That leaves structure, and it explains why garden soil in a tray is the worst possible choice. Germination is an aerobic process: from the moment of imbibition, cellular respiration restarts and consumes oxygen. A compacted, saturated medium no longer supplies any. Garden soil compacts when watered, crusts over as it dries, and throws in the organisms behind damping-off as a bonus. Perlite and vermiculite feed nothing, they keep pores open, which is all that is asked of them.

Fertility earns its place back at one precise moment, the appearance of the first true leaves. From then on, the seedling is self-sufficient, its reserve is used up, and it needs a medium that feeds it: either it gets pricked out into a richer mix, or it gets topped up with a light, diluted feed. Not before.

Water: From Below, and Never a Permanent Puddle

Overhead watering is where three failures converge. The first is mechanical: a jet of water, even through a fine rose, shifts surface seeds and buries them. A correct depth turns into a wrong one, and a seed that needed light ends up covered. The second is structural: falling water packs the surface down and lets it crust over as it dries, a crust that fine seedlings cannot break through. The third is a health issue, and it is the costliest: it keeps the collar wet, which is exactly the zone where damping-off attacks.

Bottom watering removes all three. Water goes into the tray holding the pots, capillarity draws it up into the root ball, and the excess gets tipped out after twenty to thirty minutes. The gradient this produces is exactly the right one: a drier surface where the fungi live, a moist root zone where the roots live. What must never happen is a tray left sitting in standing water for several days, which recreates exactly the suffocating conditions the technique was meant to avoid. Watering has to be enough without the plants sitting in it.

Timing and water quality close the subject. Watering early enough in the day for the surface to be dry by nightfall denies the fungi the damp, cool, still window they are waiting for. And for a seed tray, clean water is what to use: agricultural guidance explicitly rules out pond water and runoff that may have been in contact with soil carrying the organisms behind damping-off. A rainwater butt falls into that category for seed sowing, even though it remains excellent for the rest of the garden.

Last comes an apparent paradox, because it is what sets the whole watering calendar. A seed that is imbibing must never dry out, or it dies the moment its radicle has emerged. A seedling that has come up must never stay waterlogged, or it damps off. The switch happens exactly at emergence: constant moisture and a cover before, cover removed and the surface allowed to dry out between waterings after. Keeping the cover on for one week too many after emergence is the single most common mistake in the whole process.

Damping-off Cannot Be Cured, Only Prevented

Damping-off is not one disease but a group of convergent attacks. The genera actually responsible are Pythium, Rhizoctonia, Fusarium and Phytophthora. The detail that changes how to respond: Pythium and Phytophthora are not true fungi but oomycetes, water moulds, and their spores are mobile. They carry flagella and swim through a film of water, cell to cell across a tray. That is why losses spread in patches rather than plant by plant, and why a tray set in a full tub contaminates every neighbouring tray through the watering water.

The disease takes two forms that look nothing alike. Before emergence, the seed or the germ is killed below the surface: what shows up then are empty circular patches in the tray, and this is the form almost always blamed on the seed packet. After emergence, the seedlings look normal for a few days, then the base of the stem browns or blackens right at substrate level and the plant topples. Rhizoctonia leaves a distinct, sunken lesion, reddish-brown to dark brown, at soil level or just below it; on sturdier stems, it produces wire stem, thinned and discoloured, which twists or bends without breaking.

The conditions each one favours are not the same, which rules out working from a single rule of thumb. Pythium thrives in a cool to cold substrate, waterlogged and poorly aerated. Rhizoctonia tends to settle in under warmth and in less sodden conditions. So there is no single temperature to avoid: what they share is stagnant humidity and a seedling that stays vulnerable for a long time. Hence the most counter-intuitive conclusion on the subject, and yet it is the explicit agronomic recommendation: the best prevention is speed. Enough warmth, enough light, sowing not too deep, so the seedling gets through its fragile period as fast as possible.

What actually prevents it comes down to five moves. Fresh growing medium, never the bottom of an open bag stored straight on the ground. Clean containers: brush them, then soak for an hour in a solution of one part household bleach to nine parts water, and rinse thoroughly, and the same goes for benches and trays reused season after season. This bath is mixed wearing gloves, in a ventilated room, and never combined with any other product. Sparse sowing, so air moves between the seedlings. Each tray set in its own full saucer rather than a shared tub, so water carries nothing from one tray to the next. And a small fan gently stirring the air: it dries the surface, and the mechanical stimulation shortens and thickens the stems, an effect documented under the name thigmomorphogenesis.

What does not work deserves to be said just as plainly. Once the collar has pinched shut, there is nothing left to repair: the conducting tissues are destroyed and the plant is standing on a dead stem. Remove the affected seedlings along with the growing medium around them, stop watering from above, open up the air, let the surface dry out, and re-sow into fresh growing medium. No homemade preparation and no product dosage has any place here: the treatment for damping-off is the hygiene of the sowing that came before it.

Legginess: What Three Days of Dim Light Do Forever

A seedling deprived of light does not simply decline, it runs a different programme. This darkness programme, skotomorphogenesis, is perfectly coherent: every resource goes into stretching the stem, the cotyledons stay closed and folded, an apical hook protects the growing point as it forces its way through the soil, and chloroplast production is put on hold, hence the pale colour. The seedling spends its reserve betting that light is a few centimetres further up. When light arrives, the programme flips: stem stretching stops, the hook opens, the cotyledons unfold, root growth speeds up.

Hence the irreversibility, which is the real issue. Light stops future stretching, it does not shorten what has already been built. An internode is only made once. A leggy plant stays leggy, and no later adjustment catches it up. Only one family truly gets away with it: tomato carries embryonic adventitious roots along its stem, so it gets potted up buried to the cotyledons and grows a fresh root system along the buried section. Pepper does not do this easily: it roots poorly along its stem and rots more readily in a cool, moist medium, so burying it deeply backfires. Burying it a little at a first potting-on gives it mechanical support, and nothing more.

That leaves the question that decides everything: why even the best-exposed window is not enough early in the season. The unit that matters is not perceived brightness but the total quantity of useful photons received over twenty-four hours, expressed in moles per square metre per day. A seedling needs on the order of 12; professional benchmarks put lettuce, cucurbits and cabbages between 10 and 15, and tomato, pepper and aubergine between 15 and 20. Yet early in the local growing season, the outdoors itself supplies only 5 to 25 depending on latitude, and a greenhouse already loses 60 percent or more of that through its glazing and structure. A window is structurally worse off than a greenhouse: it is a single opening in a wall, seeing only a slice of sky rather than the whole dome of it, and a tray set back fifty centimetres receives only a fraction of what a tray pressed against the glass gets. Full sun behind a windowpane, at the start of the season, remains a starvation ration.

The practical consequence follows the same logic. The light reaching the plants drops off very fast as the light bar moves further away, much faster than the change in distance would suggest, so the same lamp no longer delivers anything like the same daily total depending on its height. Distance gets corrected before duration, always. Duration, for its part, cannot simply be stretched at will: beyond 18 hours of continuous lighting, tomato develops interveinal chlorosis, which sets the useful range at around 14 to 16 hours. Finally, the eye is a very poor judge, because it adapts: it cannot compare a window to a lamp, and only a meter measuring plant-usable radiation can settle the question.

One last lever, free and little known: the gap between day and night temperature governs how much the stem stretches. A day warmer than the night stretches it, a day cooler than the night shortens it, by acting on gibberellins and auxin. Greenhouse growers exploit this without any product at all, by letting the temperature drop 3 to 6°C for two to four hours just before daybreak, because it is in the early morning that the stem stretches the most. At home, this simply means not heating the seedling room first thing in the morning, and above all removing the heat mat as soon as half the seedlings are up: a warm substrate under weak light produces exactly the plant nobody wants.

Seeds Too Old: Real Shelf Life, and the Test That Settles It

Seeds die on a clock that is specific to each species, and the differences are enormous. About one year for onion, parsley and parsnip, with a caveat: other references give parsley three to five years and parsnip one to three years. Two years for leek, pepper and sweetcorn. Three years for carrot, celery, spinach, bean, pea, broccoli and asparagus. Four years for tomato, aubergine, beetroot, chard, cabbages, turnip, fennel, courgette and pumpkin. Five years for cucumber, radish, melon and endive. Six years for lettuce. Pepper is the classic vegetable-garden trap: it gets mentally filed alongside tomato, when it actually lasts half as long.

The nuance that changes everything is rarely written down, and yet it is explicit in agricultural extension literature: vigour declines before the germination rate does. In other words, an ageing batch does not just germinate less, it produces weaker plants from the seeds that still do germinate. A packet down at 40 percent is therefore not a packet to sow more thickly, it is a packet to replace. The logic behind sparse sowing, which is also good practice against damping-off, argues the same way: twenty reliable seeds beat sixty doubtful ones crammed into one tray.

Storage follows two long-established rules of thumb, and they are easy to remember. Every percentage point less of moisture in the seed doubles its storage life, and every 5.6°C drop in storage temperature doubles it again as well. The rule of one hundred follows from this: the sum of the storage temperature in degrees Fahrenheit and the relative humidity in percent should stay under 100, with temperature contributing no more than half of it. In practice: a sealed jar, dry, dark, under 15°C, and above all let the jar come back to room temperature before opening it, or moisture from the room condenses on the cold seeds and cancels out the benefit of storage.

The homemade germination test settles the question within a few days and costs nothing. Ten seeds on damp absorbent paper, the paper folded over, slipped into a sealed bag or box so it does not dry out, labelled with the species, the date and the number of days stated on the packet, then set somewhere warm. Light is not necessary for most species; for those that need it, simply leave the box in ambient light rather than in a cupboard. At the end of the stated time, count. Eight out of ten or more: sow as normal, the published threshold being 80 percent. Seven: sow more thickly. Six or fewer: buy new seed. The test does a second, even more useful job: it gives the real germination time under your own conditions, and therefore the benchmark for knowing whether a tray has failed or is simply running late.

Pricking Out and Hardening Off: Where Already-Won Seedlings Get Lost

Pricking out happens when the first true leaves appear, the ones that follow the cotyledons, generally two to three weeks after germination, before the roots start to tangle. The technique comes down to one rule: lift the seedling by a cotyledon, never by the stem. A torn cotyledon costs almost nothing, a crushed stem is fatal, because at this stage the plant’s entire supply line runs through a single one-millimetre column that cannot repair itself. Lift from underneath with a small stick or the back of a blade, never pull. Make the hole first, lower the roots into it without folding them back, close the medium around them without over-firming it, and water immediately to chase out air pockets around the roots.

Two points close out this stage. The new container gets a richer growing medium, since the seed’s reserve is used up and the plant is now fed by its surroundings. And not every species gets pricked out: carrot, parsnip, radish and taproot vegetables more generally are sown in their final spot, because the taproot is precisely the organ being harvested, and it forks the moment it is disturbed.

Hardening off takes seven to fourteen days and cannot be skipped. A plant raised indoors has a thin cuticle, a poorly developed wax layer, stomata that are slow to close, and no mechanical conditioning at all. Outside, it meets ultraviolet light, wind and a daily quantity of light several times greater than what it grew under, all at once. The classic progression: two to three hours outside on the first day, during the mildest part of the day, sheltered from wind and direct sun, then roughly two more hours each day, up to ten to twelve hours, then one or two full nights outside before planting. What this changes physically is measurable: growth slows, the cuticle and waxy layers thicken, and water loss drops.

A plant put outside without hardening off discolours or burns within a day, and even when it survives, it spends two weeks rebuilding what one week of hardening off would have given it. Two benchmarks for timing this move without ever pinning it to a fixed date: count backward from your last frost, six to eight weeks for a tomato or pepper sowing raised indoors, and only plant out warm-season species once the soil itself has warmed up, not just the air. Hardening off is, in fact, suspended during a cold snap or a burst of wind, and resumes afterward: the weather sets the pace, never the calendar.

Common pitfalls to avoid

The care sheets to read next

Every plant mentioned has its full sheet: watering, light, soil, pests and a month-by-month calendar.

Browse every sheet in the catalogue →