Every seed packet in the world gives its sowing advice the same way: sow six weeks before the last frost. That is excellent advice and almost useless on its own, because nobody’s diary has a week in it. What you need is a date — a day you can write in a calendar, put a tray of compost on, and actually act on. This takes the day of the year your last spring frost usually falls on, counts back by however many weeks the packet says, counts forward again for transplanting, and puts all three on a single line you can read at a glance.
Counting back from the frost
Seed packets talk in weeks. Diaries need a day you can write down.
Use a long-term average. Day 120 is 30 April in a normal year; day 135 is mid-May.
Straight off the seed packet. Tomatoes 6 to 8, peppers 8 to 10, brassicas 4 to 6.
Tender crops go out after the frost. Hardy ones can go out before it — use a negative situation by setting this low and moving the frost date instead.
Fine tuning
Seven to ten days is enough for most seedlings; longer for anything going into exposed ground.
Result
Why the answer is a day number and not a date
Working in day-of-the-year is not affectation. A calendar date has to know whether the year is a leap year, which month lengths apply and, in half the world, which of two calendar systems is in use. A day number does not. It is a plain count from the first of January, it converts to a date in about two seconds once you know the month lengths, and it makes the arithmetic — which is all addition and subtraction of sevens — trivially checkable by hand.
It also matches how the source data is published. Frost statistics come as averages over thirty years, usually quoted as a date, and converting a date to a day number once is far less error-prone than doing calendar arithmetic in your head every time you read a new packet. Once you have the number for your area, it does not change: you reuse it for every crop, every year.
Counting backwards is the whole method
Seed packets are written backwards from the frost date because that is the thing that actually limits you. The frost date is a hard constraint imposed by the weather; the sowing date is a flexible one imposed by how long a particular crop takes to reach transplant size. So the packet tells you the offset and leaves you to supply the date, and the offset differs between crops because the growth rate differs between crops.
Tomatoes need six to eight weeks indoors because they have to be large enough to flower by the time they go out. Peppers need eight to ten because they are slower still and will not set fruit until they have size. Brassicas need four to six and are often planted out before the frost rather than after, because they do not care about a light frost at all. The offset is not a difficulty rating; it is a growth-rate statement.
| Crop | Sow indoors | Plant out | Why the offset |
|---|---|---|---|
| Tomatoes | 6–8 weeks before | 2 weeks after | Needs size before it will flower |
| Peppers and chillies | 8–10 weeks before | 2–3 weeks after | Slow to germinate and slow to bulk |
| Brassicas | 4–6 weeks before | 2–4 weeks before | Frost hardy, so they go out early |
| Lettuce | 4 weeks before | 2–4 weeks before | Fast, and bolts if held too long |
| Courgettes | 3–4 weeks before | 1–2 weeks after | Outgrows its pot in five weeks |
| Hardy annuals | 6 weeks before | On the frost date | Sown to flower, not to size up |
The last fortnight is the part people skip
A seedling raised indoors has lived its whole life at a steady temperature, out of the wind, under diffuse light. Putting that plant straight into a garden is a shock: the temperature swings fifteen degrees between day and night, the wind removes water faster than the roots can replace it, and the sunlight is several times stronger than a windowsill. Plants survive it, but they stop growing for two or three weeks while they adjust, which wipes out the head start the indoor sowing was supposed to give you.
Hardening off avoids that. A week to ten days of increasing exposure — a sheltered spot in the day, back under cover at night, then out overnight for the last few nights — builds thicker cuticle, sturdier stems and a root system that can keep up with transpiration. It is the cheapest week in the whole process and the most frequently skipped.
The hardening-off window on the chart runs backwards from the plant-out date, because that is how it has to be scheduled: you know when the plants go in, and you need them acclimatised by then, so the exposure period has to finish on that day.
If a cold snap is forecast during the hardening window, put them back under cover for the nights. Being hardened off is not the same as being frost proof, and a late frost on a half-hardened plant does more damage than on one that never went outside.
How far off a frost date can be
A published last-frost date is an average, and averages hide the shape of the distribution. In most temperate areas the last frost falls within about two weeks of the average in two years out of three, and outside three weeks in roughly one year in ten. That is why the advice is always to treat the frost date as a planning tool rather than a promise, and why the plant-out date should stay flexible by a week in either direction.
The practical consequence is that the sowing date is more robust than the planting date. If you sow on the calculated day and the frost comes late, you hold the plants for a week — slightly inconvenient, and easily managed by potting them on. If you plant out on the calculated day and the frost comes late, you lose the plants. Sow on time; plant out on the weather.
Sowing too early is worse than sowing late
There is a temptation to add a week or two to the packet's offset, on the theory that a bigger plant is a better plant. It is not. A seedling held in a tray past its transplant size becomes root-bound, and a root-bound plant takes far longer to establish than a smaller one — the roots circle the inside of the cell and have to be persuaded to grow outwards into the soil, which can cost three weeks.
The second problem is light. Indoors, in late winter, a seedling on a windowsill receives perhaps a tenth of the light it would get outside. It responds by stretching: long thin stems, pale leaves, a plant that falls over when you water it. Every extra week indoors in low light makes that worse, and unlike root-binding it cannot be undone by potting on.
Do
- Use a thirty-year average frost date from a local source, not last year's weather.
- Follow the packet's offset; it is a growth-rate figure, not a suggestion.
- Harden off for a full week before planting out.
- Pot on once if the plants are ready before the weather is.
- Keep the plant-out date flexible by a week in either direction.
Avoid
- Adding two weeks to the offset to get bigger plants.
- Raising seedlings on a windowsill with no supplemental light and expecting sturdy growth.
- Moving plants from a heated room straight into the garden.
- Planting out on the calculated day regardless of the forecast.
- Using one year's frost date because it is the one you remember.
What to do with the dates once you have them
Three dates is a schedule, and a schedule is only useful if the rest of the plan hangs off it. The sowing date tells you when to have compost, trays and labels ready. The hardening date tells you when to clear a sheltered spot outdoors. The plant-out date tells you when the ground has to be prepared — which is the one that catches people out, because ground preparation has its own lead time.
Quantities come next. The seed quantity calculator turns the area and the sowing rate into packets, which is the figure you need before the sowing day rather than on it. Where plugs are going in instead of seed, the plant spacing calculator gives the count, and the row spacing planner turns that count into a bed layout — both worth doing while the seed is still germinating, because the layout determines whether the transplant day is a twenty-minute job or a two-hour one.
Area is the input all three of those need, and the garden area calculator is where that comes from. If the seedlings are going into a protected structure rather than open ground, the greenhouse volume figure is what determines how many trays you can actually harden off under cover at once — usually fewer than people assume. Feeding after transplanting is a separate calculation, and the fertilizer calculator handles it once the plants are in the ground.
Where the numbers stop being true
This is a day-count from a single frost date. It assumes the frost date you enter is meaningful, which means a long-term average from somewhere nearby, and it does not attempt to model the detail that makes real frost prediction hard: cold air pooling in a valley, a wall that radiates heat overnight, a courtyard that stays five degrees warmer than the field two hundred metres away. Microclimate routinely shifts the effective frost date by a fortnight in either direction.
It also assumes germination happens when the packet says it will, which depends entirely on temperature. A tomato germinating at eighteen degrees takes about a week; at fourteen it can take three, and the whole schedule slides by a fortnight without anything being wrong. Where the timing matters, work in days from germination rather than days from sowing — write the germination date down when you see the first seedling and recalculate from that.
Finally, this works on a single sowing. Successional sowing — the same crop every two or three weeks for a continuous harvest — is a different exercise: the same arithmetic run several times with the frost date held constant and the sowing date stepped forward.
How the figures are produced
Everything here is addition and subtraction of sevens. The sowing day is the frost day minus seven times the number of weeks the packet specifies. The transplant day is the frost day plus seven times the number of weeks after. Days under cover is the difference between the two, and weeks under cover is that divided by seven. The hardening-off date is the transplant day minus the hardening period, so that the exposure finishes exactly on the day the plants go out. No calendar library, no leap-year handling, no rounding — which is what makes it checkable by hand in a few seconds.
Questions people ask
How do I turn a day number into a date?
Add the month lengths up. Day 120 in an ordinary year is 30 April: January 31, February 28, March 31 gives 90, and another 30 lands on 30 April. In a leap year every date after February shifts by one, which is exactly the kind of thing the day number is designed to avoid.
My packet says sow direct. Does this still apply?
It does, with the offset usually shorter or zero. Direct-sown crops go in when the soil is workable and warm enough, which usually means on or after the frost date. Enter the weeks-before figure as zero and the sowing date falls on the frost date.
What if my area does not get frost?
Use the date when night temperatures reliably stay above about ten degrees instead. The arithmetic is identical; only the meaning of the reference date changes. Many maritime and Mediterranean gardens effectively have no frost constraint at all.
Should I use the first or last frost date?
Last spring frost for sowing, first autumn frost for working out whether a crop will mature. They are different numbers and confusing them is the single most common error in autumn planning — a crop needs days to maturity counted forward from transplanting, checked against the first autumn frost.
Can I sow later than the packet says?
Yes, and it is usually safer than sowing early. A plant put out two weeks younger establishes faster and overtakes a plant that was held too long. The cost is a later harvest, not a worse one.
Does hardening off really matter for hardy crops?
Less, but yes. Brassicas and hardy annuals tolerate the transition better than tomatoes, but a week of exposure still prevents the check that costs you a fortnight of growth. Even a hardy plant stops growing while it adjusts.
Why does the chart start before the sowing date?
So the sowing mark is not sitting on the edge of the axis with no context. The window is padded by a fortnight at each end, which is roughly the amount the whole schedule can slide on a cold spring.