Calculator
Sowing and transplant dates from your last frost
Finding your own frost dates
Every date on this page is relative to two figures specific to your location: the average last frost in spring and the average first frost in fall. Everything else is arithmetic from those two.
The authoritative sources are the National Oceanic and Atmospheric Administration climate normals and your state cooperative extension service, both of which publish local frost data. The NOAA US Climate Normals give station-level freeze dates, and the USDA Plant Hardiness Zone Map gives the winter minimum band that determines what survives outdoors year round. Use those rather than a generic figure for your state, because frost dates vary substantially over short distances.
Two cautions about these figures. They are averages, which means roughly half of years will have a later frost than the average last frost date. Published dates are commonly given at a stated probability, and the 50 percent figure is not the one to plan a tender crop around. The 10 percent figure, the date after which there is only a one in ten chance of frost, is the conservative planning date.
Local topography also matters more than the map does. Cold air is dense and flows downhill, pooling in hollows and behind barriers. A garden at the bottom of a slope can be several degrees colder on a still clear night than one partway up the same slope, and a low-lying frost pocket can be a full zone different from the published figure for the area. Your own observation over a season or two beats any published number for your specific site.
Counting backwards for sowing
The reason each crop has a different lead time is that they take different lengths of time to reach transplant size, and being ready too early is a real problem rather than a harmless margin.
A transplant that reaches full size and has nowhere to go becomes root bound, and a root bound transplant is permanently checked. Its roots have circled the cell, the plant has begun to redirect energy from vegetative growth, and it never fully recovers after planting out. Sowing three weeks early produces a worse plant than sowing on time, which is counterintuitive and is the most common home propagation error.
Peppers and eggplant, 8 to 10 weeks before last frost. They germinate slowly, need warm soil to do it at all, and grow slowly afterwards. These are the first thing sown in the propagation year.
Tomatoes, 6 to 8 weeks. Faster than peppers, and tomatoes tolerate potting on into a larger container if they outgrow their cell, which gives some tolerance for error.
Brassicas, 4 to 6 weeks. Fast growers that resent being held. Broccoli, cabbage, kale and cauliflower all go backwards quickly once root bound.
Cucurbits, 2 to 3 weeks. Cucumbers, squash and melons grow very fast and dislike root disturbance, so they are sown late and planted out small. Sowing these early is the single most common wasted effort in a propagation year.
Onions and leeks, 10 to 12 weeks, which makes them the earliest sowing of all in most schedules, though they are far more tolerant of being held than anything else on this list.
How growing under cover shifts the dates
Every figure above assumes the transplant goes into open ground. Growing under cover changes the target date rather than the lead time, and the shift depends on what kind of cover.
Unheated cold frame or low tunnel: roughly 2 to 4 weeks earlier in spring for hardy crops, and a similar extension in fall. That is the honest figure, and the mechanism is 7 to 10 degrees F of overnight protection plus wind exclusion. It does not make tender crops possible earlier, because a cold frame that is 8 degrees warmer than a 28 degree F night is still at 36.
Unheated greenhouse: similar, with the addition of standing headroom and a larger, more thermally stable air volume. It also warms the soil earlier, which matters for direct sowing.
Heated greenhouse: this is a different proposition, because it removes the frost constraint entirely and substitutes a light constraint. A house held at 55 degrees F can grow tomatoes in winter in the sense that they will not freeze, but at a midwinter daily light integral below 10 mol per square meter per day they will not fruit usefully. The binding constraint shifts from temperature to light, which no amount of heating addresses. See the DLI calculator.
The practical rule is that protection buys weeks at the shoulders of the season and does not buy a different season, unless you also add light.
Succession sowing
A single sowing of a fast crop produces a glut followed by nothing. Succession sowing spreads the harvest, and under cover it is easier to sustain because the season is longer at both ends.
The interval depends on how long the crop holds in good condition once mature. Lettuce and radish hold poorly and want sowings every two weeks. Kale, chard and most brassicas hold for weeks and want sowings every four to six. Tomatoes and peppers are single-sowing crops that produce over months, so successions are unnecessary.
The interval also stretches as the season advances, because crops grow faster in warm bright conditions and slower in cool dim ones. Two sowings a fortnight apart in the middle of summer will mature only a week apart; two sowings a fortnight apart in fall may mature a month apart. Experienced growers shorten intervals in fall to compensate.
A propagation bench makes successions practical in a way direct sowing does not, because a tray started under lights is ready to go into a bed the day the previous crop comes out. That continuity is most of the value of the propagation setup.
Days to maturity, and why the number lies
Seed packets give days to maturity, and the figure is worth understanding rather than trusting.
First, it is measured from different starting points depending on the crop. For direct-sown crops it is from sowing. For transplanted crops such as tomatoes, peppers and brassicas, it is almost always from transplanting, not from sowing, which means the actual time from seed is the stated figure plus the propagation period.
Second, it assumes good conditions. A crop maturing in 60 days under trial conditions may take 90 in a cool cloudy season, because development is driven by accumulated heat rather than by elapsed time. That is precisely what growing degree days measure and why they predict maturity better than a calendar does. See the growing degree days calculator.
Third, it does not account for day length. Crops maturing in fall grow progressively more slowly as light declines, so a fall planting takes considerably longer than the same variety planted in spring. Commercial growers apply a fall factor of roughly 1.3 to 2 times the stated maturity for late plantings, and it is a real effect rather than a fudge.
