Ten stages, and what each one actually means.
A crop does not grow steadily. It goes through named, dated, measurable stages, and an agronomist can tell you the yield ceiling of a wheat crop in February and the yield itself by the end of May. This is that sequence, and what we tell you at each point in it.
The season, stage by stage.
The Zadoks scale is the standard used by every agronomist in Britain. We use it here because it is what actually gets written on the spray record, and inventing a friendlier one would only make the updates less useful.
-
Between 5 and 20 October
Drilling
Seed goes in at around 350 seeds per square metre, three centimetres deep, a fortnight later than the textbook says. Every day of delay in October is one fewer blackgrass plant in May.
The drill date, the seed rate, the variety and the field, logged the same evening.
-
Twelve to twenty days later, late October
Emergence — GS09
The coleoptile breaks the surface and the rows become visible from the gate. This is the first point at which anything can be judged, and the first point at which slugs matter.
A photograph from the same fence post we will use all year, so the sequence lines up.
-
November to February
Tillering — GS21 to GS29
The plant puts out side shoots. Every tiller is a potential ear, and the number it makes over winter sets the ceiling on everything that follows.
A tiller count per square metre in early February, and what it implies for the ceiling.
-
Late March into April
Stem extension — GS30 to GS32
The first node becomes detectable at the base of the stem and the growing point lifts above ground, where frost can finally reach it. The plant stops making tillers and starts losing them.
The date GS31 was reached, because every input decision after this hangs off it.
-
Early MayNow
Flag leaf — GS39
The last leaf emerges. It does something close to forty-five per cent of the work of filling the grain, which is why it gets the most expensive fungicide of the year.
The flag leaf date, the septoria pressure, and what was sprayed on it.
-
Mid to late May
Ear emergence — GS51 to GS59
The ear pushes out of the boot. Flowering follows within days, and a hot dry week here costs grains per ear that no amount of later sunshine puts back.
The ear count per square metre — the number that predicts the yield better than anything else.
-
June into mid-July
Grain fill — GS71 to GS77
Watery ripe to soft dough. The grain lays down starch, and the crop is now living almost entirely on what the flag leaf and the ear can photosynthesise.
A thousand-grain weight estimate, and the first honest yield guess of the year.
-
Mid to late July
Ripening — GS87 to GS92
Hard dough to caryopsis hard. The green goes out of the straw, moisture falls, and the whole farm starts watching a forecast it cannot influence.
Grain moisture readings, twice a week, published whether they are good or not.
-
Any dry window in August
Harvest
The combine goes in when the grain is under fifteen per cent moisture and the forecast holds for two days. In 2025 that window barely opened, and the durum came in under two tonnes a hectare against an expected six.
The cut date, the yield in tonnes per hectare, and the moisture at the header.
-
Within a day of cutting
In the shed
Weighed over the weighbridge, dried if it needs it, stored at 14.5 per cent moisture and sampled for nitrogen — which decides whether barley is malting or merely feed.
The weighbridge ticket number, the sample result, and the date your allocation was set aside.
Every stage on this list fires exactly one notification. Ten a season, capped, and nothing between nine at night and eight in the morning.
Why the tracker is the product.
People do not buy a quarter acre of barley because they need forty-eight bottles of lager. Forty-eight bottles of lager can be bought in about ninety seconds, more cheaply, from the brewing pages of this same site.
They buy it because between March and August something happens to a piece of ground with their name on it, roughly once a fortnight, and they get told about it. That is the whole product. The lager is the receipt.
Which is also why the tracker publishes the bad stages as loudly as the good ones. A notification that says the tiller count is six per cent below the five-year mean is more interesting than one that says everything is fine, and considerably more honest.
What each stage predicts.
| Stage | When | What it tells you |
|---|---|---|
| Emergence, GS09 | Late October | Whether the establishment is good enough to bother with. Under 200 plants/m² and the field gets redrilled. |
| Tillering, GS21–29 | November to February | The ceiling. You cannot have more ears than you have tillers, so this is the maximum the crop can ever be. |
| GS31 | Late March | The date every subsequent input decision hangs off, and the point at which frost becomes expensive. |
| Flag leaf, GS39 | Early May | About forty-five per cent of the eventual yield lives on this one leaf, which is why it gets the dearest fungicide. |
| Ear count, GS59 | Late May | The single best predictor of yield available before the combine moves. |
| Grain fill, GS71–77 | June to July | Thousand-grain weight, and therefore the difference between a good year and a merely acceptable one. |
| Moisture, GS92 | Late July | Whether we can cut. Under fifteen per cent and the combine moves; over it and we wait and pay for drying. |
What the tracker sends, and what it does not.
Ten notifications across eleven months. The marketing one is off by default and we are not going to quietly turn it on.
Capped at ten a season, and nothing between 9pm and 8am.
The stage nobody expects to care about.
Tillering. It happens in the dark half of the year, nothing visible changes for three months, and it is the stage that decides the whole crop.
A wheat plant makes side shoots over winter. Each one can become an ear. Come February an agronomist walks the field, counts tillers in a quarter-metre quadrat, multiplies up, and knows the maximum yield that crop can ever produce. Everything after that is loss management.
So the February notification is a number and a comparison against the five-year mean for that field, and it is the one that people reply to.
- 742
- tillers/m², B2, Feb 2026
- +6%
- against the five-year mean
- 498
- ears/m² in the end
See it on a real patch.
Twenty-eight quarter acres, nine of them already gone.
